Identifying and managing psoriasis-associated comorbidities: the IMPACT research programme

Study design

Using a primary care database from the UK [specifically the Clinical Practice Research Datalink (CPRD)], a population-based cohort study was conducted that included people with and people without psoriasis. The CPRD comprises the entire medical history (demographics, treatments, clinical events, test results and referrals to hospitals) of patients registered in a general practice in the UK. The database is broadly representative of the UK population in terms of age and gender. The protocol of this study was approved by the CPRD’s Independent Scientific Advisory Committee (protocol reference number 11_134A). The study is reported according to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. 60

At the time data were extracted (September 2012), data were available for 652 practices and > 12 million patients.

Study population

An inception cohort of adult (aged ≥ 20 years) patients with psoriasis and a matched comparison group of up to five people without psoriasis were identified between 1994 and 2009. Patients with and patients without psoriasis were included if they had no history of CVD or diabetes mellitus before index date (first diagnosis of psoriasis or corresponding consulting date for comparison patients) and were registered for ≥ 2 years in the general practice before entry into the study cohort.

Comparison patients were individuals who had never received a diagnosis for psoriasis. For each patient with psoriasis up to five comparison patients were matched on age, gender and general practice. Furthermore, to ensure that comparison patients visited the practice at approximately the same time as when patients with psoriasis received their first diagnosis of psoriasis, comparison patients were also matched on index date (psoriasis diagnosis date) in a 6-month window. All patients were followed up from their respective index date (first diagnosis of psoriasis) or consulting date (for the comparison cohort) and ended at the earliest date of the occurrence of a major CV event (MI, ACS, unstable angina or stroke), transfer out of the practice, death date or end of follow-up (31 December 2011).

Definition of exposure

Patients with a first diagnosis of psoriasis between 1994 and 2009 and received a recognised treatment for psoriasis {(emollients, topical treatment, phototherapy, systemic therapy or biologics [National Institute for Health and Care Excellence (NICE) clinical guideline CG153]} were included in the cohort. 61 Patients were classified as having severe psoriasis if they had received a systemic treatment (acitretin, etretinate, ciclosporin, hydroxycarbamide, methotrexate, fumaric acid), phototherapy or a biologic therapy (etanercept, adalimumab, infliximab, ustekinumab, efalizumab); if they had received none of these they were classified as having mild psoriasis. 30

Outcome of interest

The outcome of interest was a combined CV end point, including the first event of fatal or non-fatal MI, ACS, unstable angina or stroke. In the main analysis, the outcome of interest was identified using data from CPRD by using the Read code classification, which is a hierarchical coding system used to record diagnoses in primary care. 62 To assess the robustness of the main findings to potential outcome misclassification, in a sensitivity analysis, the combined CV end point was identified using primary care data (CPRD) and the national mortality records [from the Office for National Statistics (ONS)] of those patients in practices providing linked data between CPRD and ONS. In this situation, both Read codes and the International Classification of Diseases codes were used.

Covariates

To assess whether or not there is a relationship between psoriasis and risk of major CV events, a statistical model was built including multiple risk factors (other than psoriasis). This approach will help us to understand whether or not any observed association between psoriasis and major CV events could be accounted for by other concomitant diseases/risk factors that are known to be linked to CVD. Risk factors included in the model were age, gender, depression and calendar year calculated at baseline and having or developing severe psoriasis or inflammatory arthritis (which included both diagnostic codes for PsA or rheumatoid arthritis). Diabetes mellitus, chronic kidney disease, hypertension, hyperlipidaemia, atrial fibrillation (AF), transient ischaemic attack, congestive heart failure, thromboembolism, valvular heart disease and smoking status were modelled as time-varying risk factors; therefore, an individual’s classified status could change during the study period. Psoriasis was defined as being severe from the date of first exposure to phototherapy, systemic or biologic treatment. It was considered as severe from that point onward. Smoking status was classified as ‘current’, ‘former’, ‘never smoker’ or ‘unknown smoking status’. This enabled switching from one smoking class to another during follow-up. 30 Additional risk factors such as a high body mass index (BMI) score and a high score on the Index of Multiple Deprivation (IMD),63 which is a measure of socioeconomic status, were included only in sensitivity analyses owing to the high proportion of missing data.

All code lists used for the exposure and outcome are available for download [URL: www.clinicalcodes.org (accessed 15 May 2019)]. 64

Statistical analysis

Medians and interquartile ranges were used to summarise continuous variables; proportions were used to summarise discrete covariates. The combined CV end-point events and incidence rates per 1000 person-years with 95% CIs were calculated for patients with and patients without psoriasis, and by disease severity. The estimate of the age- and gender-adjusted hazard ratios (HRs) and 95% CIs for each variable were made using Cox proportional hazard regression. Cox regression with a shared frailty model was used to estimate the fully adjusted HRs and 95% CIs. The assumption of proportionality for each variable and the model overall was tested by using Schoenfeld residuals.

Using the initial cohort identified from CPRD, multiple sensitivity analyses were performed. These included assessing whether or not a more complex or parsimonious multivariate model (by including different sets of risk factors) would change the conclusions. BMI and IMD scores contained a high proportion of missing data; therefore, they were included in the model only in sensitivity analyses. Smoking status had a small proportion of missing values; therefore, it was included in the main analysis and a category was introduced to account for missing values. Accordingly, sensitivity analyses explored the impact of using multiple imputation to estimate missing BMI and IMD scores so that they could be included in the model.

Additional sensitivity analyses were (1) including only those patients with at least one GP visit per year, (2) including only those patients with ≥ 6 months of follow-up, (3) testing for an interaction between the presence of psoriasis or severe psoriasis with age and (4) additional adjustment for patients exposed to methotrexate, ciclosporin or oral retinoids.

Finally, a subgroup analysis was performed by linking data from CPRD to the mortality data contained in the ONS. The nested cohort included patients identified from CPRD practices linked to the ONS and who had an index date (first diagnosis of psoriasis or corresponding consulting date) between 1998 and 2009; they were followed up until 2011. The outcome was the first fatal or non-fatal major CV event (MI, ACS, unstable angina or stroke) recorded in CPRD or ONS.

All statistical analyses were performed using Stata^® 12 (StataCorp LP, College Station, TX, USA).

Background

Since 2006, the possibility that people with psoriasis experience an increased rate of fatal and non-fatal CVD in comparison with the general population has been an increasing focus of discussion. 67 At the time of this study it was unclear whether such an enhanced risk was due to an increased prevalence of traditional CV risk factors (e.g. hypertension, obesity, dyslipidaemia); an increased prevalence of some of the more recently recognised CV risk factors, such as physical inactivity and depression; or the severity of the psoriasis. There was very little information on the prevalence of traditional CV risk factors in patients with psoriasis in the UK and this small body of research may have been limited by selection bias, inappropriate choice of control groups or reliance on risk factors measured for other clinical reasons.

Primary care patients aged ≥ 40 years are supposed to be offered CV risk assessment as a matter of routine in the UK. However, this is not the case for patients aged < 40 years. This could be a problem for those diagnosed with psoriasis if, as reported by Gelfand et al. ,18 they have up to three times the expected CVD risk. Without robust data it would be difficult to make a case for systematic screening of patients with psoriasis for CV risk factors over and above what might be offered to the general population. 31,67–69

Previous primary care-based studies have neither assessed how systematic screening for CVD risk factors might influence their estimated prevalence in patients with psoriasis nor estimated the potential benefits of identifying such risk factors and subsequent intervention.

Objectives

Our objectives were to:

• investigate whether or not screening for CVD risk factors in primary care could influence the estimated prevalence of CVD in patients with psoriasis

• assess whether or not the prevalence of screen-detected CVD risk factors or estimated CVD risk varies by age, psoriasis severity and the presence of PsA

• estimate the clinical benefits of normalising modifiable risk factors.

Method

Selection of general practices, participants and sample size

General practices from two PCTs in north-west England were identified. We aimed to recruit large and small general practices from deprived and affluent areas by grouping eligible general practices in quintiles by size, using the numbers of registered patients, and by level of deprivation, using the IMD (based on practice postcode). A five-by-five sampling frame was constructed (Figure 4) and practices in the nine middle quintiles were removed. Practices were then randomly selected from the remaining four quadrants: small and deprived, small and affluent, large and deprived, and large and affluent. Removing the nine middle quintiles increased the likelihood of ensuring maximum differences between the quadrants.

FIGURE 4.

Sampling frame for practices.

Forty-four eligible general practices were sent letters of invitation and information sheets and were followed up by telephone, e-mail or visit. Only five practices were recruited via this route; consequently, the recruitment strategy was widened to include three more PCTs. In these PCTs, practices were identified via the comprehensive local research network. In total, 13 practices were finally recruited and reimbursed for their participation (i.e. block payments were made to participating practices to compensate for time incurred by administrators identifying psoriasis patients in each practice).

We aimed to recruit 320 people to yield estimates of the prevalence of CVD risk factors to within ≈1.9% of the true value for rarer factors and to within ≈5.3% of the true value for common factors.

Accordingly, we aimed to include ≈80 people with psoriasis in each of four age/sex categories: males aged < 40 years; females aged < 40 years; males aged ≥ 40 years and females aged ≥ 40 years. The age of 40 years was the cut-off because routine screening for CVD risk factors is usually offered routinely once every 5 years to all patients aged ≥ 40 years. Participating general practices identified patients with psoriasis aged > 18 years using Read codes known to map to the condition and medications/topical preparations for psoriasis (see Rutter et al. 66). Exclusion criteria were severe mental health problems, no capacity to consent, recent bereavement and terminal illness.

Eligible patients were sent an invitation from their own GP to attend a CVD risk screening assessment at their general practice. The invitation included an information leaflet about the study and a reply slip on which the patient was asked to indicate if they had psoriasis and if they were interested in participating. Smaller practices mailed all eligible patients on their list; larger practices mailed an agreed number (depending on practice capacity), using a randomised numbered list to select invitees. Patients replying positively were telephoned by a researcher, who arranged attendance for CVD risk factor screening at the patient’s own general practice.

Results

Practice and participant recruitment

Practice sizes ranged from 3070 to 16,746 registered patients; IMD scores ranged from 48.79 (least deprived) to 9.77 (most deprived). Details can be found in Appendix 2.

The process of participant recruitment is shown in Figure 5. A total of 1446 people with psoriasis were invited to attend for risk factor review, of whom 287 (19.8%) completed the review.

FIGURE 5.

Flow chart of responses.

Nine practices (P1, P2, P4, P5, P6, P7, P8, P12 and P13) provided more detailed information about patient response to the screening invitation, permitting a breakdown of response rates by patient age and sex (Table 2). Acceptance rates were lower among younger (< 40 years) than among older (≥ 40 years) patients; there were no appreciable differences by gender.

TABLE 2 - Patient response rates by age and sex in nine of the participating practices (covering 85% of invitees)

Patient sex, age (years)	Invited (n)	Attended (n)	Response rate (%)
Male, < 40	231	25	11
Female, < 40	209	39	19
Male, ≥ 40	346	82	24
Female, ≥ 40	444	117	26
All patients	1230	263	21

Of those screened, 173 (60%) patients were invited to a follow-up appointment by their practice, 109 patients were not given a follow-up appointment and the outcome for five patients was unknown.

Nearly one-quarter of participants had severe psoriasis and around one-third were clinically obese. One-third of participants self-reported high cholesterol levels and high blood pressure and over two-thirds of these people were receiving medication for these conditions (lipid-lowering or antihypertensive medications). A history of coronary or cerebrovascular disease or AF was self-reported by < 10% of participants.

We found that nearly half the cohort had at least one screen-detected risk factor and one-fifth had two or more risk factors for CVD. There was no evidence that the prevalence of screen-detected risk factors varied by age, psoriasis severity or the presence of self-reported PsA. Comparison with the weighted HSE data suggests differences between our study participants and the general population on one risk factor (where p < 0.001). Hypertension was higher (p < 0.001) in study participants than in the general population. This applied to the full cohort (n = 287) and the subset of people (n = 269) who were not taking ciclosporin, which can cause hypertension.

Among the study participants receiving treatment for CVD risk factors, nearly half (46%) had suboptimal therapy levels for blood pressure and/or total cholesterol and one-quarter (26%) of participants had suboptimal levels for glucose assessed by HbA_1c.

In our study, just over one-third of participants not already considered ‘high risk’ according to NICE guidelines were assessed to be at high risk of CVD using the 10-year risk threshold of > 10% (QRISK2). 53 The mean predicted lifetime CVD risk estimated by QRISK2 was 35%. Thirteen per cent of participants had a high lifetime CVD risk of > 50%. After adjusting for age, there were no statistically significant differences in CVD risk between participants who had severe psoriasis and self-reported PsA and those without these conditions.

Discussion

To our knowledge, this is the first UK primary care-based study to report that CVD risk factor screening augments the estimated prevalence of CVD risk factors in psoriasis, and it is the first study to assess lifetime CVD risk and the potential benefits of optimising modifiable CVD risk factors. We have shown that the proportion of patients with screen-detected CVD risk factors was unrelated to age, psoriasis severity or the presence of self-reported PsA. When known and screen-detected risk factors were considered, hypertension was more prevalent in patients with psoriasis, in keeping with data from two other studies. 53,71

Although we had an adequate sample size, we did not show statistically significant differences in the age- and sex-adjusted prevalence of obesity, smoking or diabetes compared with the general population. Although more than one in three participants (37%) were at high 10-year CVD risk, only one in eight (13%) participants overall had a high lifetime risk of > 50%. Therefore, the estimation of lifetime risk does not appear to be particularly beneficial in patients with psoriasis.

We observed many patients with suboptimal management of known risk factors. Previous studies have shown undertreatment of CVD risk factors in patients with psoriasis72 and one study suggests that hypertension is more difficult to manage in these patients,73 perhaps because psoriasis or its treatment somehow increases blood pressure or reduces concordance with therapy.

A few studies have reported the prevalence of screen-detected CVD risk factors in patients with psoriasis in highly selected patient groups. 74–77 Several large studies, some of which were population based,23,78 have reported a higher prevalence of known CVD risk factors including hypertension, dyslipidaemia, diabetes, obesity, smoking, renal disease and metabolic syndrome. 23,78–82 However, control groups in these studies included people with ‘forms of dermatitis’,81 other hospitalised patients77,82 and people taking part in other research projects,77 which makes interpretation, comparison and clinical implications challenging. Other studies have reported the prevalence of known CVD risk factors in patients with psoriasis but without a control group. 83,84 Our study demonstrated that using a standard approach to CVD screening of patients with psoriasis who attend primary care is an effective means of identifying individuals with CVD risk factors who would have otherwise been missed.

Strengths and limitations

Strengths of the study include (1) we selectively sampled practices across a broad range of socioeconomic settings to ensure a meaningful sample size in four age/sex subgroups; (2) the prevalence of important CVD risk factors was established with clinically meaningful precision in these subgroups; (3) we reported how CVD risk factor screening, when added to information on known risk factors, influenced the detected prevalence of CVD risk factors; (4) we used QRISK2 to estimate the 10-year risk of CV events in patients with psoriasis not already judged to be at high CVD risk; (5) psoriasis severity was assessed; (6) we compared prevalence data to population-based controls matched for age, sex and ethnicity; (7) we assessed the benefits of optimising risk factors within 10-year and lifetime time frames; and (8) these data can be used to inform policy on primary care-based risk factor screening.

Limitations of the study include (1) risk factors were measured on one occasion, which may have led to some misclassification; (2) we did not validate diagnoses of psoriasis or PsA; (3) our cohort was largely white, limiting the generalisability of findings; (4) the participation rate was 20%; and (5) men aged < 40 years were under-represented. We suspect that patients who did not participate would be more likely to have an adverse risk factor profile than those who enrolled. Response rates were lower in men than in women, suggesting that men would be less likely to attend for routine screening, even though, being at higher risk, they are more likely to benefit from intervention.

Clinical implications

Cardiovascular risk factor screening is a means of identifying a high proportion of patients (1) at high CVD risk, (2) with screen-detected risk factors and (3) with suboptimally managed known risk factors. ‘Health checks’ have recently come under criticism for failing to improve long-term health outcomes,85,86 perhaps because of overmedication or social class biases in attendees. 87 An alternative explanation has been made by our research team in a parallel study that recorded and analysed CVD consultations (see Cardiovascular disease risk communication and reduction in psoriasis). The study found that risk information was seldom discussed with patients; instead, practitioners prioritised information collection (see Cardiovascular disease risk communication and reduction in psoriasis). 88 Thus, although the current study has shown that screening identifies modifiable CVD risk factors, caution is required before recommending this as a strategy.

Accordingly, our findings in this study need to be considered alongside those showing limited response of clinicians to identified risk before universal CVD screening for people with psoriasis can be recommended. 88

Workstream 2.i substudy: assessing the impact of psoriasis severity on artery health (pulse wave velocity)

Aims

To assess whether PWV is related to (1) the severity of psoriasis, (2) the presence of PsA and/or (3) the age at onset of psoriasis.

Method

Selection of participants

Thirteen general practices across north-west England identified patients with psoriasis and invited them for a CVD risk assessment (study 2.i). This PWV substudy recruited from the 287 people who attended the CVD risk assessments (Figure 6).

FIGURE 6.

Recruitment to PWV substudy.

Data analysis

As PWV was skewed, it was log-transformed before analysis. Linear regression was used to assess the associations between PWV and CV risk factors, with log PWV as the outcome variable. Severe psoriasis and PsA were also evaluated as potential determinants of PWV: severe psoriasis was defined as Self-Administered Psoriasis Area and Severity Index (SAPASI)102 score of > 10 or use of a disease-modifying therapy [psoralen and ultraviolet A (PUVA), methotrexate, ciclosporin, acitretin, fumaric acid esters, etanercept, adalimumab, infliximab or ustekinumab]. Psoriatic arthropathy was defined as positive responses to any three out of the first five Psoriasis Epidemiology Screening Tool questions or a positive response to the question ‘Have you ever been told that you have arthritis associated with psoriasis?’.

Relationships between arterial stiffness and measures of psoriasis were adjusted for traditional CVD risk factors, psychological variables and the presence of PsA (when appropriate).

Data analysis was performed using the Stata 13 package. Two-tailed p-values of < 0.05 were considered statistically significant.

Results

A total of 125 subjects were recruited: 66 female and 58 male. The median age was 58 years [interquartile range (IQR) 45–66 years]. There were 27 subjects with severe psoriasis and 44 with PsA. The median SAPASI score was 3.6 (IQR 1.4–6.2).

Mean PWV was 8.83 m/second (95% CI 8.08 to 9.58 m/second) in subjects with severe psoriasis and 8.80 m/second (95% CI 8.40 to 9.19 m/second) in those without; the difference was not statistically significant (p = 0.9). Controlling for CV risk factors as psoriasis-associated factors had little effect on the difference between these two groups (see Appendix 3, Table 12). PWV was significantly higher in subjects with PsA (mean 9.34 m/second, 95% CI 8.74 to 9.95 m/second) than in those without (mean 8.52 m/second, 95% CI 8.11 to 8.93 m/second; p-value for difference = 0.03). However, controlling for age and sex reduced the difference between the two groups until it was no longer statistically significant. The difference remained non-significant after controlling for CV risk factors and disease characteristics (see Appendix 3, Table 13).

There was no significant association between PWV and SAPASI score (p = 0.23). However, after controlling for age and gender, PWV increased with increasing SAPASI score (increase of 1.1% per unit increase in SAPASI score, 95% CI 0.3% to 1.8%) because the higher SAPASI scores tended to occur in younger subjects. This association was reduced to a non-significant level by controlling for CV risk factors (see Appendix 3, Tables 12–14).

There was a tendency for PWV to be higher in subjects with a higher age at onset of psoriasis (p = 0.001), but this effect became non-significant after controlling for age. Controlling for CV risk factors and psoriasis-related factors had no further impact on this association.

Conclusions

Message-framing

Message-framing – emphasising the benefits (‘gain framed’) or costs (‘loss framed’) of a behaviour or course of action – is important in risk communication116 and has been shown to affect behavioural outcomes. 118 The effects of message-framing may be different for different health-related scenarios, for example in the context of ‘prevention’ or low-risk behaviours, such as reducing alcohol, versus ‘detection’ or high-risk behaviours, such as cancer screening. 116 The effects of message-framing may also differ in relation to immediate versus future health consequences. 119 Emotional response to health messages is another factor that can also affect decision-making and behavioural outcomes. 120 These factors may be important in the context of communicating risk to people with psoriasis who might be motivated to make LBCs through a desire to either improve their immediate psoriasis symptoms or reduce their future risk of CVD.

In summary, it was unknown if, in a population of patients who are already identified as potentially psychologically vulnerable, (1) communicating CVD risk enables patients with psoriasis to understand their own personal risk and motivates or empowers them to reduce their risk rather than add to their distress or (2) message-framing affects patients’ behavioural intentions.

Study 2.ii: qualitative examination of cardiovascular disease risk communication in primary care

This study investigated how patients and primary care clinicians (GPs and practice nurses) perceived/experienced the process of assessing for and communicating about CVD risk in CVD screening consultations. The focus was on which risk communication techniques are used by GPs and practice nurses and how these affect patients’ understanding of risk in the context of psoriasis.

Study 2.iii: experimental study of message-framing

This study aimed to investigate the impact of health risk information on intentions to make healthy lifestyle changes in people with psoriasis and to determine whether gain-framed health messages (e.g. ‘if you change your lifestyle, your psoriasis will be better’) or loss-framed health messages (e.g. ‘if you don’t change your lifestyle, you will get CVD) are more effective in prompting intentions to change behaviour. It also aimed to investigate whether people’s behavioural intentions are driven by a desire to improve immediate psoriasis symptoms or reduce future risk of CVD. This work has been published. 123

Methods

Results

Practitioners in 10 of the 13 general practices participating in the overall workstream 2 screening study (study 2.i) agreed to have their consultations audio-recorded (four GPs, seven practice nurses and two research nurses). In total, 130 separate audio-recordings of CVD risk assessment consultations and 15 audio-recordings of follow-up consultations were captured. These involved 13 different practitioners and 131 patients (because one patient had the follow-up but not the initial consultation recorded). Capturing follow-up audio-recordings in particular proved to be very problematic because appointments were not conducted in dedicated clinics in the same way that assessment appointments were. Twenty-nine in-depth interviews with patients with psoriasis were then conducted (12 patients aged < 40 years; 17 patients aged ≥ 40 years). In-depth interviews with 12 primary care practitioners (six practice nurses, two research nurses and four GPs) were carried out. (Sampling for qualitative interviews was brought to a close when the data set was judged to offer sufficiently rich and detailed accounts of patient and practitioner experiences.)

Analysis focused on how risks were discussed in consultations, what methods of communicating risk information were used and if practitioners investigated patients’ knowledge of risk reduction or offered support about risk reduction/referral.

Findings

Conclusions

Screening for CVD risk provides opportunities for practitioners and patients to discuss current and future health status and enables practitioners to offer brief behaviour change interventions. Such opportunities may be missed because of a practitioner focus on data-recording rather than intervention in consultations, minimising risk and avoiding discussion with patients of the emotional impact of risk information, as well as providing standardised rather than specific/personalised advice about how to modify CVD risk.

Data collection and analysis

A community sample of participants with psoriasis was recruited to an online questionnaire to investigate the impact of health risk information on reported intentions to make healthy lifestyle changes. Ethics approval was obtained from the university REC (reference number: 13118). A two-by-two between-participant design was used to randomly allocate individuals to one of four evidence-based health message types (message frame: loss or gain; message focus: immediate psoriasis symptom reduction or future CVD risk reduction). The primary outcome measure was behavioural intentions to change. A series of two-way between-group analyses of variance tests were conducted to explore the impact of message frame and message focus on behavioural intentions.

Results

A total of 217 participants were randomised to one of the four conditions. There was a significant message frame × message focus interaction for reported behavioural intention to reduce alcohol intake (p = 0.023). Loss-framed messages (e.g. ‘if you don’t change your lifestyle you will get CVD’) were more effective for intentions to reduce future risk of CVD whereas gain-framed messages (e.g. ‘if you change your lifestyle your psoriasis will be better’) were more effective for intentions to reduce immediate psoriasis symptoms. Behavioural intention scores for increasing exercise and improving diet were not statistically significant. See Keyworth et al. 123

Conclusions

Message effectiveness about the benefits of reducing alcohol in the management of psoriasis is dependent on how the message is constructed. When presented with messages about short-term psoriasis symptom reduction, gain-framed messages were more effective than loss-framed messages in increasing behavioural intentions. Conversely, when presented with messages about long-term CVD risk reduction, loss-framed messages were more effective than gain-framed messages.

All findings from study 2.ii informed the planning and development of the IMPACT workstream 5 patient materials and practitioner training intervention, which both include risk reduction strategies.

Data collection and analysis (main patient study)

Given the evidence that people with psoriasis may disengage from health-care services for psoriasis management, the study sample was recruited from community sources (i.e. not through primary or secondary care health-care services) across Greater Manchester by distributing information through websites and in community locations such as libraries, places of religious worship, shops and community centres. Sampling was purposive for maximum variation on age, gender, ethnicity, socioeconomic background and self-identified severity of psoriasis including duration and treatment.

Data were gathered from participants in a location convenient to them. This qualitative study used face-to-face interviews to gather views about coping with psoriasis including experiences of consultations with practitioners. Ethics approval was provided by the university REC (reference number: 10325). In-depth, face-to-face, semistructured qualitative interviews were conducted using a topic guide developed from the existing psoriasis literature as well as the theoretical frameworks of illness beliefs136 and coping/appraisal. 137 These theoretical approaches conceptualise coping as an evaluative process by which individuals appraise both the source of stress and the coping resources available to them. The appraisal process accounts for individual differences in response to similar levels of adversity. Participants were asked about physical, emotional and social effects of living with psoriasis and their coping strategies (both problem- and emotion-focused, including use of social support, self-care strategies, medication use and choices as well as use of health-care services). Interviews were audio-recorded with consent, transcribed verbatim and transferred to the NVivo computer package for qualitative data management. 126 Data were analysed concurrently with data collection according to principles of framework analysis128 to code for salient themes and produce a thematic framework of key concepts using constant comparison techniques. 138

Data collection and analysis (adjunct general practitioner study)

General practitioners in Greater Manchester primary care practices were invited by e-mail to take part in an interview and purposively sampled for diversity on gender, age, ethnicity and size of practice. A topic guide was developed from the relevant literature and used to guide interviews in which GPs were asked about how they managed patients with psoriasis in the primary care setting. Interviews were audio-recorded, transcribed verbatim and, as above, also analysed using principles of framework analysis and constant comparison to generate key concepts. The NICE guideline135 (which focuses on the importance of two main elements: the assessment and the management of psoriasis) was used to identify main themes about GPs’ attitudes to managing psoriasis.

Findings (main patient study)

Participants described marked experiences of stress and distress in relation to living with psoriasis (including itching, flaking and painful skin, and lowered mood, self-confidence and self-image due to concerns about their appearance and feelings of low control over psoriasis symptoms). These demands were perceived to go unacknowledged in consultations with health-care professionals. Participants gave accounts of poor experiences in the management of psoriasis by health-care practitioners (such as perceived lack of expertise and a failure to recognise and manage psoriasis as a complex, long-term condition with appropriate monitoring, review and referral to specialist services). There were reports of coping with the demands of the condition against the perceived lack of support from services by disengaging from consulting about psoriasis or seeking alternative opinions and treatments outside formal health-care services.

Analysis also enabled identification of the function, range and use of coping/self-care strategies among people with psoriasis. Participants reported a variety of goals in relation to coping with their condition, for example reducing symptoms and psychological distress, managing uncertainty and gaining control over symptoms. Although at the outset the study had a particular focus on identifying positive coping strategies, during data collection and analysis it became clear that there were higher than expected levels of distress among the study sample and evidence of common use of less helpful coping strategies in this group (such as minimising psoriasis as unimportant, disengaging from services, stopping treatment use, avoiding social contact, bottling up emotions and acting on angry feelings as well as hypervigilance in hiding affected skin). Some effective coping responses/self-care strategies were found (such as learning to be open about psoriasis with inquisitive people, taking care of general health, seeking emotional support from trusted people, focusing on positive aspects of life, pacing activity to conserve energy and suggesting alternative treatment strategies when consulting with their health-care professionals). We acknowledge that the high levels of distress found could be due to self-selection bias, with ‘most distressed’ people responding to advertising; however, participants interviewed were often working, studying and functioning in roles as parents or carers. Moreover, there were no strong indicators that this sample was atypical.

The analysis also identified a number of culturally specific cross-cutting themes that were particularly salient for participants with a South Asian background. These were linked specifically with high levels of experienced stigmatisation due to the visible nature of psoriasis, which could affect an individual’s standing in their family and wider community. Table 5 presents illustrative data extracts from the patient study.

Findings (adjunct general practitioner study)

General practitioners reported assessment and management of psoriasis that were not in line with NICE recommendations. First, some GPs recognised psoriasis as a complex condition but most viewed it primarily as a skin complaint. No GP reported undertaking a structured assessment of the impact of psoriasis on patients to incorporate discussion of possible physical and psychological comorbidity. In particular, they reported minimising the potential emotional and social effects of the condition on people’s lives in consultations. Second, GPs did not view or manage psoriasis as a long-term condition with regular monitoring, review and appropriate referral as they would for other long-term conditions seen in primary care. Addressing lifestyle issues in relation to psoriasis was also not a priority. However, most GPs indicated low levels of expertise and confidence to manage psoriasis, citing lack of undergraduate and postgraduate training in dermatology. Illustrative data extracts are presented in Table 6.

All findings from these studies informed the planning and development of the IMPACT programme workstream 5 new patient materials and practitioner training intervention.

Study 4.i: content analysis of core training competencies relating to lifestyle behaviour change skills for health-care professionals

This study aimed to assess the extent to which LBC skills are included in the postgraduate training curricula of dermatologists, dermatology specialist nurses (DSNs), GPs and general practitioners with a special interest (GPSIs) in dermatology. 147

Study 4.ii: in-depth qualitative interview study of health-care professionals to examine how they conceptualise and manage psoriasis, including gauging attitudes to providing lifestyle behaviour change support as part of patient care

This qualitative study aimed to assess the experiences of dermatologists, specialist nurses and GPs in managing psoriasis and, in particular, providing strategies and referral routes to support lifestyle change (e.g. weight reduction, increased physical activity, reduction in smoking or alcohol use) for patients and the barriers to conducting this work in practice.

Study 4.iii: observational study of the prevalence, nature and quality of lifestyle behaviour change information available to patients with psoriasis in clinic settings

This study aimed to investigate whether or not the setting of the patient waiting room currently promotes appropriate LBC information for psoriasis patients by providing up-to-date information linking lifestyle with disease (general and specific) and high-quality links to support LBC.

Study 4.iv: survey of dermatology specialist nurses to assess views on behaviour change skills training needs

This study aimed to assess the views of DSNs on training needs in relation to supporting behaviour change in people with psoriasis. The aim was to enable planning of the level and type of training required for dermatology specialist staff who are likely to have fewer opportunities to undertake skills support in the area of LBC.

It was intended that findings from all four studies would inform the development of effective new approaches (IMPACT programme workstream 5 interventions) to help front-line staff assess and manage psoriasis more holistically and overcome any barriers to supporting patients’ LBCs.

Methods

Results

In the 11 curriculum documents analysed, 67 instances of terms related to LBC and health promotion were found. Most were found in the GP curriculum (62.7%), followed by the specialist nurse curriculum (20.9%) and dermatologist curriculum (16.4%). There were no instances in the curriculum of GPSIs in dermatology. The majority of terms were related to awareness-raising alone, with no instances linked to the skills required for long-term behaviour change facilitation (Figure 10).

FIGURE 10.

Core curricula mapped to the Prevention and Lifestyle Behaviour Change148 competence framework by professional group. Reproduced from Keyworth et al. 147 © 2014 British Association of Dermatologists. Reproduced with permission of John Wiley & Sons.

Reproduced from Keyworth et al. 147 2014 British Association of Dermatologists. Reproduced with permission of John Wiley & Sons.

Of the 67 occurrences found in the curricula, around one-third related to being aware of opportunities to introduce LBC; approximately one-fifth related to being able to identify and signpost to LBC support (however, behaviour change techniques known to enable these competencies were not included); half the instances were unable to be mapped to the Prevention and Lifestyle Behaviour Change competence framework148 and none of the core curricula related to provision of long-term support and LBC facilitation. In summary, the core practitioner training documents showed few clearly specified learning outcomes or recommendations relating to LBC knowledge, skills and attitudes. There were few references to recognised LBC techniques (see Keyworth et al. 147 for supporting data).

Conclusion

Study 4.i highlighted the lack of systematic training for practitioners who may be managing patients with psoriasis to develop appropriate skills and knowledge. Post-qualification health-care professional curricula could be improved by including more explicit LBC skills training.

Data collection and analysis

Health-care professionals were first contacted in writing through their professional organisations or via public general practice lists with snowball sampling as a second step when recruitment proved challenging (i.e. health-care professionals did not readily respond to invitations via their organisations and responded better when known peers facilitated invitations personally). In this way, individual participants were able to identify other potential interviewees to approach. This study was approved by the University of Manchester’s REC (reference number: 12017).

A topic guide developed from the literature guided semistructured interviews with health-care professionals managing people with psoriasis. Participants were asked about their attitudes and knowledge in relation to supporting lifestyle changes for people with psoriasis as part of psoriasis management, as well as their experiences of practice and strategies used, with a focus on their:

• levels of knowledge (including on alcohol consumption, smoking/smoking cessation, obesity, low activity levels, CVD risk, low mood associated with psoriasis)

• attitudes to lifestyle change support or advice-giving (clinical and managerial priorities, actual and perceived staff roles)

• understanding of barriers to engaging in behavioural change support for patients (environmental, social/cultural norms, low mood)

• perceptions of barriers for practitioners (knowledge, ‘embeddedness’ of LBC, training/skills, preservation of therapeutic relationship, patient burden)

• attitudes to the use of patient information (leaflets, booklets, or known referral routes to services such as smoking cessation)

• actual behaviours used to address LBC.

In recognition of the potentially sensitive nature of the interview questions, all interviews were undertaken by experienced interviewers trained in asking questions to probe adequately and sensitively to get beyond socially desirable answers. Interviews were audio-recorded, transcribed and analysed using principles of framework analysis,128 with a model of evidence-based factors known to influence behaviour change149 and the Self-Regulatory/Common Sense Model (SRM/CSM) of illness representations150 as frames to consider the data.

Results

In total, 23 in-depth interviews were conducted with clinicians (seven consultant dermatologists, six DSNs, five GPSIs in dermatology and five regular GPs in primary care).

Findings highlighted that, although most clinicians recognised the importance of LBC in psoriasis management, they did not see it as part of their professional role to support patients with behaviour change. Lack of time and prioritising other aspects of care such as diagnosis and medications management were cited as underlying reasons, in addition to a belief that addressing alcohol use, smoking or weight loss was a potential threat to harmonious relationships with patients. Clinicians were pessimistic about patients’ motivation to change as well as their own influence in helping patients make behavioural changes. Table 7 presents illustrative data extracts to support the analysis (see Nelson et al. 113).

In addition, practitioners held a range of different ‘personal models’ of psoriasis. Most reported working with incoherent models in mind, for example holding a ‘sophisticated’ understanding of psoriasis as a complex condition while paradoxically managing the condition in a ‘linear’ skin-focused way. Practitioners who reported working to an incoherent personal model also reported frustration in relation to managing psoriasis and satisfaction only when patients’ skin improved. Table 8 presents illustrative data extracts to support the analysis (see Chisolm et al. 151).

For the most part, limited knowledge and skills in ‘whole-person’ management, including LBC skills, underpinned these beliefs and attitudes. Nonetheless, some clinicians identified a need for training to enable them to manage psoriasis as a complex, long-term condition involving comorbidities and to support patients with LBCs.

Conclusions

There are low levels of both knowledge and skills among professionals about managing psoriasis as a complex, long-term condition, including addressing LBCs. There is also a lack of structured support in both primary and secondary care for this work. Training to broaden professionals’ conceptualisations of psoriasis and incorporate evidence-based LBC skills in consultations could enable better patient assessment and management.

Data collection and analysis

Health centres were randomly selected from a full, publicly available list. In a non-participant observation study, exploratory observational methods were used to record the prevalence and quality of leaflets and posters signposting LBC (whether general or dermatology specific) in health centre waiting areas for patients with psoriasis in both primary and secondary care. A structured observation schedule was developed to guide data collection (see Chisholm et al. 151). Ethics approval was obtained from the University of Manchester REC (reference number: 12017). Content analysis was used to identify frequency, characteristics and standard of materials. A series of quality indicators guided rating of the materials’ quality in terms of visual condition and visibility/accessibility to patients.

Results

From the 24 health centres observed, 262 sources of lifestyle information were identified. These were mainly categorised into generic posters/displays not specific to psoriasis (n = 113) and generic leaflets/flyers not specific to psoriasis (n = 98). Information was of poor quality, as well as being poorly displayed, and there was no evidence of high-quality psoriasis-specific information being made available to patients.

Conclusions

Study 4.iii found that little emphasis is given to the role of lifestyle as a health risk in patients with psoriasis. Evidence about the use of environmental cues to prompt behaviour change could inform the design and display of lifestyle information.

Methods

Results

Analysis of 77 participant responses indicated that DSNs generally expressed confidence in being able to address lifestyle change with patients with psoriasis. However, only 19% reported having knowledge of evidence-based techniques that could be used in consultations. On a scale from 1 to 7, with 7 being the highest level of confidence, the mean scores given by DSNs were between 4.3 and 4.0 for addressing smoking cessation, alcohol reduction, physical activity, diet and weight loss. There were differences in the degree to which respondents believed that different health-care professionals had a role to play in addressing LBC with patients with psoriasis: primary care-based practice nurses (100%); GPs (95%); DSNs (90%); dermatologists (73%) and GPSIs (70%).

Conclusions

Findings from this survey suggest that LBC skills training needs careful planning to tailor interventions in the most appropriate ways and deliver them in the most appropriate settings. Uptake of training may be low if health-care professionals fail to identify the relevance of acquiring such skills.

Data analyses

Descriptive statistics were collated to describe the demographic profile of the research participants, the perceived relevance of the PSO WELL materials and numerical rating scale data for self-reported changes in understanding and anxiety. Paired sample t-tests were used to assess change in illness perceptions, anxiety and depression scores before and after exposure to the materials.

Qualitative data collection and analysis were carried out using constant comparison138 and explored for similarities and differences within and across interviews using principles of framework analysis161 to generate key concepts.

Changes to illness coherence and beliefs about control

Post exposure to the PSO WELL materials, a statistically significant difference between the means was found in the primary outcome of illness coherence [16.2 to 17.51; t(55) = –3.48; p = 0.001 (two-tailed)], indicating a large effect size (η² = 0.19). This indicates an improvement in participants’ understanding of their psoriasis, on average. Notably, both personal control and treatment control domains also improved post intervention, with a large [t(55) = –2.98; p = 0.004 (two-tailed); η² = 0.14] and moderate [t(55) = –2.08; p = 0.042 (two-tailed); η² = 0.08] effect size, respectively (see table 4 in Nelson et al. 158 for significant IPQ-R domain changes).

The materials resulted in changes in understanding of specific aspects of psoriasis: 31 out of 54 (57%) participants reported not having a clear understanding of comorbidities pre intervention, which reduced to 19 (35%) participants post intervention; 31 out of 53 (58%) participants did not know about the relevance of lifestyle behaviours pre intervention, which reduced to 19 (36%) participants post intervention (similarity in reported frequencies is coincidental).

Changes to anxiety and depression scores

There were no clinically or statistically significant increases in HADS scores after exposure to the materials, implying that improvements in understanding of the nature of psoriasis occurred without a corresponding increase in either anxiety or depression. Cronbach’s alpha scores for the IPQ-R and the HADS domains range from 0.91 to 0.72, indicating high levels of internal reliability (see Nelson et al. 158).

Numerical rating scale scores

Eighty per cent of participants reported that their understanding of psoriasis increased post intervention; none reported that their anxiety increased after reading the leaflets and 16% reported reduced anxiety (see Nelson et al. 158).

Leaflet relevance

The majority of participants indicated that the content of the leaflets was highly relevant. Notably, many indicated that it would have had relevance eariler in the course of their disease (see Nelson et al. 158).

Theme 1: seeing the bigger picture – psoriasis as more than a skin condition

From exposure to the PSO WELL materials, participants reported gaining insight into aspects of psoriasis that had been unknown or unexplained before, including psoriasis as a complex long-term condition involving associated comorbidities and lifestyle links with a full range of treatments available. Elements that were perceived to encourage learning and reassurance were a plain language style, colour coding and easy navigation, ‘chunking’ of information and the leaflet’s positive/hopeful tone. Some unintended consequences of exposure to the new materials were identified, including participants’ disappointment in health-care services (and in some cases themselves) for not providing/seeking relevant information before.

Theme 2: personalising psoriasis – developing illness coherence

Participants perceived that the materials had helped with ‘making sense’ of psoriasis by linking disease aspects to their personal experience more directly (i.e. developing illness coherence). A particularly new aspect was recognising links between psoriasis, mood and lifestyle. The relevant and meaningful content of the materials was reported as an enabler of sense-making and the perceived trustworthiness/high quality of the materials also enabled participants to engage with the messages therein. Some changes in cognitive and emotional responses to psoriasis were reported from interaction with the leaflets, which appeared to encourage a sense of personal control and prompt thinking about ways to self-manage differently.

Theme 3: managing psoriasis differently – making changes?

Some participants reported having used the materials as a tool to make immediate changes to self-management, finding the activity sections useful. Others perceived the materials as useful for shared management of psoriasis in consultation with health-care professionals. A few participants were resistant to using the materials at all, especially where they had pre-existing negative attitudes to LBC in particular and could experience these messages as threatening.

Strengths and limitations

This was a within-group, pre–post test study to test the feasibility and acceptability of new patient resources for psoriasis. The intervention was not tested against a no-intervention control group; therefore, caution is needed when interpreting the results. However, exposure to the intervention was restricted to a 2-week period and this minimises the likelihood of other factors causing a change in participants’ understanding within that time frame. Participants were self-selecting and may have been more likely to be engaged with health messages. However, we measured the effects of the PSO WELL patient intervention and evaluated its process in depth to identify perceived mechanisms of action and these can usefully inform next phase of the research. Some patients perceived that the PSO WELL materials could be a useful for managing their psoriasis in partnership with their clinician. The parallel PSO WELL practitioner intervention (see Psoriasis and Wellbeing (PSO WELL): developing a patient-centred approach using motivational interviewing skills with dermatology clinicians to support healthy living in people with psoriasis) showed positive effects on clinicians’ knowledge and skills in managing psoriasis164 and the materials could be used by patients in collaboration with clinicians who have undertaken the training.

Background

The materials developed and described above aimed to address a number of patient needs. A remaining challenge was to ensure patient access to the support materials. Without the direct link that would have been afforded by face-to-face contact between patients and health trainers (see Synopsis), our team felt that there would be limited opportunities for follow-up or consolidation of changes, which may have been initiated by the leading clinician (GP or dermatologist). One way to overcome this would be the availability of an electronic health (e-health) platform that enables the patients to access and use the new online IMPACT programme materials developed in workstream 5. There have been major changes in the field of online health delivery systems since the programme started.

At the time that the original proposal was written (2010) we had anticipated production of digital versatile discs (DVDs) as the electronic resource. However, online systems for delivering interventions have changed enormously and findings from our recent research (and that of others) highlighted the need and expectation for integrated systems that would enable NHS staff to manage the links to their patients. Although there are increasing numbers of health applications (apps) available,165 usage by patients with long-term conditions drops dramatically in a very short space of time unless the process is co-managed by health-care providers. In addition, before costly electronic interventions are developed, it is important to establish if the target audience has the capacity or willingness to use the technology. Health practitioners and patients express concerns about health-related online interventions such as equity of access caused by ‘digital divides’ that may include age, education and technology ownership. Other concerns are changes to the doctor–patient relationship, the ability to process and understand complex information, and privacy and confidentiality.

The most effective systems allow clinical staff to track usage and patients to share data. We identified a commercial partner (company Z) that had already developed an online shared management support service for patients with type 2 diabetes mellitus that is currently in full (commercial) use in NHS England and was considered ‘first in class’ and ‘best in class’ by NHS England, with content certified against current NICE guidance.

The aims of this part of workstream 5 were to:

• identify whether or not patients would like to access support via online systems

• define the scope of the e-health platform for psoriasis based on company Z’s approach and on materials developed earlier in workstream 5

• specify the functional requirements of this e-health platform as a functioning service that can be integrated into current NHS systems

• build a low-fidelity prototype to demonstrate the potential for integrating this work with an existing chronic-disease software platform that has been proven to provide the highest levels of data security and privacy

• develop a full costing plan for each of the main features to assess economic feasibility

• assess acceptability to patients with psoriasis and health-care practitioners.

Methods

Findings and outputs

Pathway mapping and stakeholder identification

To inform the design of the e-health platform for people with psoriasis, detailed process mapping and stakeholder analysis was undertaken to outline the current care pathways and experiences of people with psoriasis in the UK. This process was directly informed by the data gathered in all workstreams (Figure 13).

FIGURE 13.

Pathway map illustrating the (preliminary) data-gathering approach used to identify pathways and key stakeholders.

Key functions of PSO WELL

Through the approach outlined above, a list of core functions was defined that addressed the identified needs of patients and clinicians in receiving and delivering high-quality psoriasis health care.

The plan following the end of this programme of research is that these core functions will be integrated into a dynamic system based on company Z’s digital therapeutic self-management approach, allowing a high degree of individual personalisation. This approach, in conjunction with the intended integration of PSO WELL into clinical pathways, could support patients by considering individual lifestyles and preferences to optimise self-management and potentially improve clinical engagement and response.

Data flows in dynamic and interactive systems are difficult to represent in a flow chart; however, Figure 14 illustrates the functional domains and features of the proposed system and indicates which aspects of patient care are affected.

FIGURE 14.

Functional domains, matrix and development roadmap for PSO WELL.

Costs of development of the PSO WELL e-health platform

Company Z provided the researchers with a fully costed project plan for the development of several versions of the e-health platform. Costs depend on the range and selection of functions chosen; however, all costs cover key components such as data security, data management and ongoing servicing and support costs.

Key conclusions

• New theory-based, high-quality PSO WELL materials are acceptable to patients, feasible to use and can improve understanding of psoriasis and illness coherence.

• Materials appear to broaden understanding of psoriasis as a systemic condition and increased participants’ sense of control without raising anxiety.

• Patients with a wide range of ages and years with psoriasis would value using patient materials to learn about self-management and to track their symptoms and flares using data shared with their health-care professionals.

• Patients would like both paper-based materials and access to an interactive e-health platform that links directly with NHS service delivery.

Implications

• Carefully formulated patient materials have the potential to change people’s beliefs about a condition and make self-management activities more salient.

• There are opportunities for improving psoriasis patients’ optimism about the likely benefits of effective self-management.

• There are materials available for distribution and use by health-care professionals to increase patients’ likelihood of engaging in effective self-care.

Key outputs

• A set of high-quality patient materials for use in clinical practice (both primary and secondary care settings).

• A low-fidelity prototype of an e-health platform to deliver patient materials and self-management tools that are acceptable to patients with psoriasis.

• Full costings for a range of functions that can be delivered via an e-health platform that is fully integrated with current NHS systems to enable both patient and practitioner access to support materials and individualised patient-specific data.

Intervention content and delivery

The 1-day intensive skills-based PSO WELL training programme taught an integrated patient-centred assessment consultation, conducted in the spirit of motivational interviewing (i.e. with an ethos of collaboration, evoking patients’ existing knowledge and supporting their autonomy), aimed at developing key skills in motivational interviewing processes (i.e. engaging, focusing, evoking and planning) and techniques to support patient choice and LBC. The training took place away from participants’ places of work in convenient and central locations and was delivered by IMPACT programme team members trained in motivational interviewing. A mix of presentations and skill-building activities was delivered along with opportunities for individual coaching during the session. The following areas were covered in the training package: (1) CVD risk communication, (2) assessing psoriasis, (3) motivational interviewing skills development and (4) long-term condition management strategies [see the IMPACT programme website171 for a video overview of the training: www.impactpsoriasis.org.uk/practitioners/ (accessed 5 August 2019)]. A workbook provided more detailed background information to the training components and reinforced key messages (Figure 16).

FIGURE 16.

Workbook and training materials used for PSO WELL training.

Evaluation design

A pilot before-and-after study was conducted to explore the efficacy of the PSO WELL small-group motivational interviewing-based training. We hypothesised that practitioners attending the training intervention would, following the training, display improvements in knowledge about psoriasis (including related comorbidities and associated risk factors) as well as demonstrating motivational interviewing skills to optimise behaviour change with patients. Training measures (described below) were taken immediately before and after the training session on the same day.

Participant recruitment

We invited a range of practitioners who manage patients with psoriasis to attend the training, including generalists and specialists from primary and secondary health-care settings. Recruitment sources included the British Association of Dermatologists (BAD); the Scottish and Welsh Dermatological Societies; the Primary Care Dermatology Society; primary care Clinical Commissioning Groups (CCGs) (e.g. through the North Manchester CCG bulletin); nurse specialist groups and individuals (e.g. the BDNG, the Dermatology Nurses Arena meeting and the IMPACT programme expert dermatology nurse group lead); AbbVie Inc. (North Chicago, IL, USA); NHS dermatology departments across Wales, England and Scotland; and IMPACT programme research events. All participants provided written informed consent prior to beginning the study. The recruitment advert is provided in Appendix 7. Ethics approval was obtained from the University of Manchester REC (reference number: 14223).

Before-and-after measures

Participants completed pre- and post-training measures immediately prior to and following the PSO WELL training programme as follows.

Feasibility and acceptability interviews

In addition, qualitative acceptability and feasibility interviews were conducted (either face to face or by telephone) with a subset of trainees at least 2 weeks post training to elicit views and experiences of the training and impressions of how the intervention might work in the clinical environment. Semistructured interviews were guided by questions in two key areas: (1) acceptability of the training content and delivery and (2) feasibility of practitioners’ ability to both attend and engage with the training.

Data analysis

Quantitative outcome measures (motivational interviewing skills and knowledge about psoriasis) were analysed using within-group comparisons to assess potential change over time. Depending on the nature of the data, both parametric and non-parametric tests were used. Low levels of missing data (fewer than three items on any measure) were managed using imputation of the mean value for that scale or subscale. Larger levels of missing data were managed by multiple imputation.

Qualitative interviews were transcribed verbatim and imported into NVivo 10. Data were categorised using content analysis and principles of thematic analysis. 174 Familiarisation with the data was followed by initial open coding in which any data relevant to the research question were highlighted and noted. Patterns in the data were then grouped into categories and further organised into superordinate themes. Disconfirming (i.e. atypical) cases were identified and highlighted to examine the extent of differing views identified in the sample.

Behaviour change skills

A paired-samples t-test was conducted to evaluate the impact of the PSO WELL training on clinicians’ (n = 55) motivational interviewing skills (BECCI measure). There was a statistically significant increase in BECCI scores from time 1 (mean 0.50, SD 0.47) to time 2 (mean 1.26, SD 0.71) [t(54) = 8.37; p < 0.0005 (two-tailed)]. The mean increase in BECCI scores was 0.76 with a 95% CI ranging from 0.57 to 0.94. The eta-squared statistic (η² = 0.56) indicated a large effect size. 27

Figures 17 and 18 illustrate the overall change in mean BECCI score following training, as well as the change in practitioner motivational interviewing skills across each of the four BECCI domains (domain 1, agenda setting; domain 2, the why and how of behaviour change; domain 3, the whole consultation; and domain 4, talking about targets).

FIGURE 17.

Mean BECCI score before and after attending PSO WELL training (n = 55). Max, maximum. Reproduced from Chisholm et al. 164 © 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

Reproduced from Chisholm et al. 164 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

FIGURE 18.

Mean BECCI domain scores before and after attending PSO WELL training (n = 55). Reproduced from Chisholm et al. 164 © 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

Reproduced from Chisholm et al. 164 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

As a treatment fidelity check, the extent to which participants are adopting a consistent motivational interviewing approach is measured by the ratio of clinician-to-patient talk time. Practitioners are likely to be using this approach if they speak less than the patient in the consultation overall (i.e. < 50% total consultation time). Trained, blinded BECCI raters estimated the percentage of time clinicians spoke during consultations in comparison to patients. Figure 19 illustrates the mean estimated clinician talk time, indicating an 11.18% reduction following the training.

FIGURE 19.

Mean estimated clinician talk time (% of total consultation) before and after attending PSO WELL training (n = 55). Reproduced from Chisholm et al. 164 © 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

Reproduced from Chisholm et al. 164 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

Patient actor feedback

Descriptive analyses of patient actor feedback before and after training are illustrated in Figure 20. It should be emphasised that these are impressions based on the experience of these individuals in training undergraduate medical students and postgraduate medical practitioners. These data suggest that patient actors felt that practitioners’ consultation approaches improved following training. Open-ended written feedback also indicates that patient actors were more satisfied following training. Content analysis of the patient actor feedback highlighted that actors perceived shifts in the way in which practitioners (1) approached consultations in general, (2) guided the focus of the consultation topics and (3) approached management planning. Actors reported that, following the training, they were more satisfied with consultations because practitioners adopted a more collaborative style, attended more to patients’ agendas, acknowledged their thoughts and feelings more and actively linked psoriasis to associated lifestyle behaviours. They also perceived that practitioners developed concrete action plans during consultations (which had been absent prior to training), judged to be led more by patients’ rather than practitioners’ ideas.

FIGURE 20.

Mean Likert scale scores of patient actors by item and total score before and after training (n = 46). Higher scores indicate improved outcomes. Max, maximum; Q, question. Reproduced from Chisholm et al. 164 © 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

Reproduced from Chisholm et al. 164 2016 British Association of Dermatologists. Reproduced with permission from John Wiley & Sons.

Knowledge of psoriasis

Participants’ knowledge about behavioural and mood factors related to psoriasis (e.g. obesity, smoking, alcohol use, physical inactivity, stress, low mood and anxiety) significantly increased following participation in the training. Specifically, paired-sample t-tests highlighted increases in participants’ knowledge of risk factor prevalence [t(60) = 4.30; p < 0.001] as well as the probable mechanisms linking psoriasis and risk factors [t(60) = 7.12; p < 0.001]. However, participant knowledge of psoriasis-related comorbidities (e.g. CVD, PsA and inflammatory bowel disease) did not increase significantly post training (p = 0.096).

Acceptability and feasibility interviews

A total of 18 practitioners was purposively sampled for qualitative interviews (four male, 14 female; eight specialist dermatology nurses, seven consultant dermatologists, two registrar dermatologists and one GP). Analysis of interview data highlighted five core themes illustrating practitioners’ views and experiences of attending the training and completing the associated research measures: (1) delivery of the training, (2) utility of training content, (3) skills acquisition and scope for implementation, (4) interest and ability to attend and (5) trade-offs when undergoing assessment.

In summary, practitioners valued the opportunity to share and learn from others’ experiences, practise skills and obtain coaching/feedback. Clinicians judged the training to have been engaging and the content to be transferable across clinical settings, reporting that it encouraged consideration of their own consultation style, the complexities of psoriasis and, in particular, how to adopt a whole-patient approach to care. The training provided new understanding about effective and ineffective strategies to help patients change health-related behaviours and increased clinicians’ confidence to attempt new approaches in practice. Practitioners found the 1-day course to be intensive and felt that they learned more than expected through training that was deemed relevant and appealing to a range of other practitioners working with people with psoriasis. Patient actors used as part of the skills assessment were felt to be realistic, although some clinicians found the assessments daunting. See Chisholm et al. 164 for more details.

Survey design

In line with the aim and objectives, the stated preference survey was designed to compare the intervention components only, so an opt-out or status quo option was not included in the design. A discrete choice experiment design was used to estimate the main effects for each component or attribute of the intervention. This assumes that there are no interactions between the attributes included. The survey asked participants to consider a series of choice questions (also referred to as choice sets). Each choice question gave participants two descriptions, or scenarios, of different aspects of the PSO WELL intervention and asked them to choose which they preferred. The decision to limit each choice question to two scenarios was based on discussion with the multidisciplinary team leading the design of the PSO WELL intervention and evaluation. This was to help minimise the perceived burden on participants.

Through discussion with the IMPACT programme research team, four different aspects or attributes to describe the PSO WELL intervention were identified. These comprised two attributes that described components of the intervention and two attributes that described desired outcomes of the intervention. The choice of attributes and levels, as well as how they were described, was informed by the qualitative research reported in Cardiovascular disease risk communication and reduction in psoriasis, Coping with psoriasis: learning from patients and Understanding the professional role in supporting lifestyle change in patients with psoriasis to develop the PSO WELL intervention, as well as the results of the development and feasibility studies [described in Psoriasis and Wellbeing (PSO WELL): developing patient materials to broaden understanding of psoriasis as a long-term condition, psoriasis-associated comorbidities and the role of self-management and Psoriasis and Wellbeing (PSO WELL): developing a patient-centred approach using motivational interviewing skills with dermatology clinicians to support healthy living in people with psoriasis]. Draft versions of the survey were reviewed by members of the IMPACT programme team and RUG and simplified in response to comments.

The attributes and levels included in the survey were as follows:

• Information, defined for participants as ‘general information and exercise leaflets to take away and read and work through in your own time. The information describes psoriasis and its treatment, lifestyle factors that may affect psoriasis and health, how to manage psoriasis and how to make lifestyle changes’. These can be given as printed leaflets or online. The levels for this attribute were (1) no information and exercises to take away, (2) printed information and exercises to take away and (3) printed and online information and exercises to take away.

• Clinic visit, defined for participants as ‘whether your clinician gives you personalised information (about your psoriasis and its treatment, lifestyle factors that may affect your psoriasis and your health, how to manage your psoriasis) and helps you to identify and make lifestyle changes’. The levels for this attribute were (1) no personalised information or help to make lifestyle changes, (2) clinician gives me personalised information about my psoriasis and health and (3) clinician gives me personalised information about my psoriasis and health and support to make lifestyle changes.

• Understanding, defined for participants as ‘your understanding about what psoriasis is, what the treatments are and how lifestyle choices may affect both your psoriasis symptoms and risk of other illnesses such as CVD’. The levels for this attribute were (1) worse than before, (2) same as before and (3) better than before.

• Ability, defined for participants as ‘your ability to manage your psoriasis and/or make changes to your lifestyle to improve your psoriasis and health’. The levels for this attribute were (1) worse than before, (2) same as before and (3) better than before.

A fifth attribute was added about the length of time it would take to read and use the information, manage psoriasis and make any lifestyle changes. The levels were 30 minutes, 1 hour, 2 hours, 3 hours, 4 hours and 8 hours. Time is used as a measuring rod to estimate how much participants are willing to spend or trade to get their preferred level of each attribute (see Appendix 8, Methods, Survey design: time attribute).

The number of attributes and levels gives 486 possible combinations of components of the PSO WELL intervention. It was not feasible to include all these in a single survey in this study. A fractional factorial design175,176 was used to select a sample of 18 choice sets that was efficient and orthogonal (see Appendix 8, Methods, Fractional factorial design). The 18 choice questions were randomly divided into two questionnaires of nine choice sets each. Figure 22 gives an example choice question.

FIGURE 22.

Example choice question.

The survey also included demographic questions (age, gender, age at diagnosis, types of clinician seen, treatments received) and a measure of overall health [the EuroQol-5 Dimensions, five-level version (EQ-5D-5L)].

Sample size

There is no agreed formula for working out sample sizes. One approach is to use the number of choice sets in the fractional factorial design, assuming that the true probability of choosing option 1 or option 2 is 50% and that an acceptable error for this probability is ± 2.5%. This gives a minimum sample of 85 complete responses for each choice set. We anticipated that approximately 30–40% of participants would complete all the choice sets in a questionnaire, meaning that 213 to 283 participants would be needed to get 85 complete responses per choice question. Stated preference survey guidelines suggest that a sample of 500 participants is sufficient for a statistical analysis that also accounts for heterogeneity. 177 Accordingly, we aimed to recruit 500 participants. Post hoc sample size estimates for each attribute and level were also conducted (see Appendix 8, Methods, Sample size estimate).

Participant sample, recruitment and survey administration

The target population for the survey was people with psoriasis because they represent the perspective of the intended users. The inclusion criteria were that participants had a diagnosis of psoriasis (defined as ‘told by a doctor that they have psoriasis’), were aged ≥ 18 years and lived in the UK. The criterion regarding location reflects the fact that the interventions were developed in the UK health-care context.

The survey was developed for participants to complete online. Ethics approval was provided by the University of Manchester REC (reference number: 16325). The stated preference survey was advertised via social media (see Appendix 8, Methods, Sample size estimate), specifically targeting people with psoriasis. Participants who responded to the adverts were given a link to the survey materials (see Appendix 8, Methods, Sample size estimate).

Analysis

The survey data were analysed using descriptive statistics to summarise participant characteristics and regression analysis of responses to the choice sets (see Appendix 8, Methods, Analysis). The results (coefficients) from the regression analysis give an indication of the strength of preference for levels 2 and 3 compared with the base or reference level (level 1) within each attribute. Time was used as a common measure to compare the strength of preferences between attributes. One approach is to estimate the amount of time that a person would be willing to spend (or trade) to get their preferred level of an attribute. This is known as the marginal rate of substitution (MRS) and is estimated by dividing the coefficient for each attribute level by the coefficient for the time attribute. The results of the regression analyses were used to estimate the relative amount of time participants were willing to trade for the different attributes. The primary analysis used a conditional logit model to estimate the main effects for participants who completed all the choice questions. Sensitivity analyses explored whether or not the MRS differed if a random parameters logit model was used and the impact on the results of including participants who completed one or more and seven or more choice questions. Subgroup analysis explored the impact of participant characteristics on the MRS and the relative importance of each attribute level.

Participants

Overall, 526 eligible participants were recruited and 250 (48%) of these participants completed all choice questions (see Appendix 8, Figure 71). This met our initial recruitment and sample size targets, with a slightly higher rate of people completing all choice sets. Nevertheless, post hoc sample size estimates suggest that for one attribute level (printed information) the sample size was not sufficient to identify preferences (compared with no information) at a statistically significant level.

Participants’ characteristics are reported in full in Appendix 8, Table 15 and summarised here. Participants who did not answer any choice questions (160 out of 526; 30%) also did not complete the questions about their sociodemographic characteristics or treatment, which were at the end of the survey. In total, 260 out of 366 (71%) participants who answered one or more choice questions also answered one or more demographic or treatment questions (see Appendix 8, Results, Participant characteristics and time to complete survey). Of those who answered one or more demographic or treatment questions, the majority were women (164/260; 63%). The average age of participants was 48 years [standard error (SE) 1, 95% CI 46 to 50, n = 257] and the average age at diagnosis was 22 years (SE 1, 95% CI 20 to 24, n = 252). The EQ-5D-5L health status measure was used to estimate the utility of participants. The utility score is anchored by 0 (dead) and 1 (full health). The average utility value of participants with complete EQ-5D-5L data depended on the value set used. Using the crosswalk value set, which maps the utility weights from the three-level version of the EuroQol-5 Dimensions (EQ-5D) to the five-level version, the average utility was 0.71 (SE 0.01, 95% CI 0.68 to 0.74, n = 255). Using the utility weights developed specifically for the five-level version of the EQ-5D, the average utility value was 0.80 (SE 0.01, 95% CI 0.77 to 0.82, n = 255). Of the participants who reported the types of health-care practitioner they had ever seen (n = 258), most had been cared for by both a GP and a hospital doctor and nurse (212/258; 82%) at some point. Nearly all participants had used psoriasis creams, ointments, lotions or shampoo at some time since being diagnosed with psoriasis (255/260; 98%); over half had used PUVA treatment (161/260; 62%) and just under half had been prescribed tablets (114/260; 44%). Just under one-fifth of participants had been prescribed injections (48/260; 19%).

Including all eligible respondents, the average time to complete the survey was 8 minutes (SD 22 minutes), although this ranged from < 1 minute to 416 minutes (see Appendix 8, Results, Participant characteristics and time to complete survey). Five participants had > 1 hour between the start and finish times (range 96–416 minutes). For 240 out of 250 (96%) of those who completed all choice questions, the mean time taken was lower than or within the time range given in the participant instructions (10–20 minutes).

Regression analysis indicated that there were no differences in all but one participant characteristic between those who took survey 1 or survey 2 and answered one or more choice questions. The exception to this was that participants who took survey 2 were more likely to see their GP and hospital doctor or nurse (compared with GP only) than those who took survey 1 (logistic regression; p = 0.004). The analysis also indicated that, for participants completing one or more choice questions, whether they chose option 1 or option 2 in each choice set was not dependent on their characteristics (see Appendix 8, Tables 16 and 17).

Choice data

The analyses of the choice data are described in detail in Appendix 8, Tables 18 and 19. The results are briefly summarised here. For the attributes about information, clinician visits, understanding of psoriasis and comorbidity, and ability to make lifestyle changes, participants preferred higher levels (levels 2 or 3) of each attribute rather than the reference or base level (level 1), which is in line with what was expected (Figure 23). In addition, the strength of preference measured by the coefficients was statistically significant for both levels compared with the lowest or reference level for these attributes. The exception to this was the information attribute, for which only the coefficient for printed and online information materials (level 3) was statistically significantly different from the base level of no information. The post hoc sample size estimates indicate that substantially more participants are needed to identify a statistically significant coefficient for the level of printed information materials only (see Appendix 8, Table 20).

FIGURE 23.

Coefficients for categorical attribute levels, effects coded, participants completing all choice questions, n = 250.

The coefficient for time was negative and statistically significant. This is in line with expectations that, all other things being equal, participants would prefer to spend less rather than more time managing their psoriasis and making lifestyle changes. There is some variation in this. Participants preferred to spend more time than 30 minutes (the reference level) and up to 2 hours per week. If they needed to spend more time than this, then they preferred to trade fewer rather than more hours per week (Figure 24). Although this appears to be counterintuitive, it may be plausible if the participants receive little support to help them self-manage their psoriasis and make lifestyle choices. The qualitative studies in this research programme suggest that this may be the experience of patients.

FIGURE 24.

Coefficients for time attribute, effects coded, participants completing all choice questions, n = 250.

Willingness to trade time for preferred attributes

The MRS analysis is described in Appendix 8, Tables 21–24. An overview of the results is given here. The data suggest that participants were willing to trade 3 hours (95% CI 1 to 5 hours) per week for general information and exercise leaflets to take away and read and work through in their own time if they could have the information in print and online. They were not willing to spend more than half an hour (95% CI –2 to 2 hours) to have printed information only.

Participants were willing to spend around 10 hours (95% CI 6 to 13 hours) per week to understand and manage their psoriasis and make lifestyle changes if it meant that their understanding (about what psoriasis is, what the treatments are and how lifestyle choices may affect their psoriasis symptoms or risk of other illnesses) was no worse or better than before the intervention. Participants were willing to spend a similar amount of time (9 hours, 95% CI 6 to 13 hours) if it meant that they received personalised information at clinic visits (about their psoriasis and its treatment, lifestyle factors that may affect their psoriasis and health, and how to manage their psoriasis). To have a clinic visit that also helped them to identify and make lifestyle changes, participants were willing to trade 14 hours per week (95% CI 10 to 19 hours).

Participants were willing to spend 25 hours a week (95% CI 17 to 32 hours) if it meant that their ability to manage their psoriasis and/or make changes improved.

Approach

The economic model aims to provide an assessment of the costs and outcomes of usual care and an exploratory evaluation of the cost-effectiveness of the PSO WELL intervention compared with TAU. The perspective is the UK health and social care system. This was used to define the range of costs included in the analyses. The time horizon for the primary analysis was 10 years, the maximum of previous analyses identified in the literature review. A longer horizon would stretch the limited evidence available too far, whereas shorter time horizons are assessed in scenario analyses. The key outcomes were incremental health benefit, measured in QALYs, and incremental costs. These were used to calculate the ICER, which is the primary outcome.

Key variables and data

The variables, data sources and parameter estimates are detailed in Appendix 11, Key variables; Appendix 11, Parameter estimates and data sources for the economic model; and Appendix 11, Tables 26–32. The model includes psoriasis treatment effects that were identified from the economic systematic review [Dermatology Life Quality Index (DLQI) and PASI]. These were converted to a common measure for the model using published algorithms (see Appendix 11, Key variables, Treatment effects). The model includes the outcomes of behaviour change and comorbidity interventions, which were identified from clinical guidelines61,182,183,185–187,191 and the sources used to develop the guidelines.

Psoriasis flares occur in the model, potentially causing people with mild disease to progress to severe psoriasis. A flare will cause an increase in psoriasis symptoms and/or decrease in treatment outcomes. It is assumed that even if patients with severe psoriasis have their symptoms effectively managed, they cannot return to having mild or moderate psoriasis (see Appendix 11, Key variables, Psoriasis flares).

Quality-adjusted life-years were used as the measure of total health benefit experienced by simulated patients over the course of their journey through the model. A QALY measures the length of survival and adjusts this to take into account any ill health during that time. For example, a person in full health for a year will have a QALY of 1, whereas a person who is in less than full health may have a QALY of 0.25. Death is assigned a value of zero. The model allows a simulated patient’s health and QALYs to vary as events occur and their attributes change (see Appendix 11, Key variables, Quality-adjusted life-years). Time to mortality was predicted from all-cause mortality. 192 The only cause of early mortality is fatality due to MI193 or stroke (see Appendix 11, Key variables, Mortality). 194

Various health-care resources and costs are required throughout the simulation (see Appendix 11, Key variables, Costs and resource use). These included the costs of the PSO WELL interventions and other health-care costs. The costs of the clinician training element depended on the number of clinicians trained as well as the number of psoriasis patients managed by each clinician. For the primary analysis, we conservatively assumed that each clinician managed only one patient, which increases the unit cost per patient of the training. A sensitivity analysis was used to explore the impact of increasing the number of patients treated by each clinician.

Other health-care costs included the costs of GP and dermatologist consultations, topical treatment, phototherapy, non-biologic and biologic systemic therapies and the costs of interventions to support lifestyle changes and to manage comorbidities.

All costs and QALYs were discounted at an annual rate of 3.5% to account for social time preferences as recommended for the UK. 195

Primary analysis

The detailed results of the primary and sensitivity analyses are given in Appendix 11, Tables 32 and 34, and Appendix 11, Figure 72. In the primary analysis, the model predicts that TAU is associated with 4.74 QALYs (95% CI 4.66 to 4.83 QALYs) per patient over 10 years (see Appendix 11, Table 34). In terms of costs, TAU is expected to incur a total cost of £26,242 (95% CI £25,852 to £28,452) per patient. The PSO WELL arm of the model generates 4.75 QALYs (95% CI 4.64 to 4.82 QALYs) per patient, with a mean cost of £27,801 (95% CI £24,902 to £27,466) per patient.

Overall, the model predicts a small gain in health benefit of 0.01 QALYs for PSO WELL (95% CI –0.10 to 0.13 QALYs) and with a mean net cost of £838 compared with TAU (95% CI –£1106 to £2593). This difference is smaller than the clinician intervention cost per patient (£1133) and may indicate that minor cost savings are made elsewhere to offset it. However, the 95% CI for the difference in costs crosses zero, indicating that the difference in cost could have occurred by chance.

The ICER is £72,802 per QALY. If decision-makers are willing to pay £20,000 to gain 1 QALY, then the probability that the PSO WELL intervention is cost-effective is 34%. This means that if only one patient per clinician trained benefits from the PSO WELL intervention, it is unlikely to be a cost-effective use of NHS resources (see Appendix 11, Table 34, and Appendix 11, Figures 72 and 73).

Threshold analysis

Threshold analysis was used to explore whether or not the relative cost-effectiveness of the PSO WELL intervention package improved if more patients could benefit from clinician training. This indicates that if decision-makers are willing to pay £20,000 to gain one QALY, the PSO WELL intervention may be cost-effective if a trained typical clinician group (average of 10 clinicians per group) manages ≥ 22 psoriasis patients (see Appendix 11, Figure 74). This equates to each clinician applying their new knowledge and skills to two patients in the year following training, thereby generating more health gains and costing less per patient. PSO WELL dominates TAU, simultaneously improving health outcomes and causing a net saving if 40 patients per clinician group (or four patients per clinician) could benefit from the clinician training.

Scenario analysis

Scenario analyses were used to assess whether or not the probability that the clinician training was cost-effective changed if alternative assumptions about key model inputs were changed (see Appendix 11, Table 35). For comparability of results, we assumed that 10 patients stand to benefit from training a typical clinician group of 10 clinicians.

The PSO WELL intervention is more likely to be cost-effective if the effect of training is maintained for > 1 year. If the impact of the clinician training is maintained for the 10-year time span of the model, then incremental costs fall and the probability that PSO WELL is cost-effective increases to 44%. If the number of patients benefiting per clinician group trained is increased to 16, the probability that PSO WELL is cost-effective increases to 52%.

The cost-effectiveness of PSO WELL improves if all dermatologists always act in a guideline-concordant manner and 20 patients benefit per clinician group (or two patients per clinician trained), at which point PSO WELL has a 50% probability of being cost-effective. This result suggests that a large benefit may be achieved if GPs correctly refer severe psoriasis patients to secondary care when it is guaranteed to be guideline appropriate. PSO WELL is less likely to be cost-effective over shorter model durations (suggesting that cost offsets are realised over longer durations) or if the PSO WELL development costs are included on top of delivery costs.

Summary of the health economic analysis

We undertook a systematic review of cost-effectiveness analyses, which found that there was little consensus about the cost-effectiveness of treatments and care strategies for individuals with psoriasis. The analyses were limited by a lack of quality evidence and frequently provided only a limited reporting of methods. We developed a new microsimulation model to evaluate the potential economic acceptability of the PSO WELL interventions developed. The simulation model utilised inputs from parallel IMPACT programme workstreams, the systematic literature review, clinical guidelines and expert opinion within the project team. The model was used to provide preliminary estimates of the economic outcomes, including cost-effectiveness, associated with the PSO WELL interventions. The exploratory analysis indicated that the interventions developed in this programme of work have the potential to be cost-effective and identified areas where more work is needed.

Pre-grant award PPI: patient representative co-investigators

Between 2007 and 2010, prior to the grant award, a former CEO of the Psoriasis Association, Gladys Edwards, and Toni Doyle, local NHS manager for patient experience, were co-applicants on the grant application and were involved in the development of the proposal. In fact, Gladys Edwards provided continuous encouragement and expertise during the grant writing and development process. Figure 27 shows the timeline between the initial proposal development and receipt of funding.

FIGURE 27.

The IMPACT programme timeline from proposal development to funded programme of research. PGfAR, Programme Grants for Applied Research.

Helen McAteer replaced Gladys Edwards as CEO of the Psoriasis Association in 2010 and has been closely involved in all aspects of the IMPACT programme to date. The team has therefore been committed to PPI from inception, and the contribution of patient representatives as co-investigators from this early juncture has enabled valuable lessons to be learned, which will undoubtedly improve involvement and engagement with PPI partnerships in future research projects.

Post-grant award PPI: recruiting for involvement

Research user group

Recruitment and development of the role

The three steering committee patient representatives attended initial meetings with research team representatives and the programme manager. During the meeting, we outlined the IMPACT programme aims and methods, committee meeting procedures and the expectations of the patient representative role, including the terms of reference (see Appendix 12). These define the purpose of the steering committee and representatives’ responsibilities. To facilitate the involvement and participation of each patient representative, a ‘buddy’ system was implemented. Karen Kane took on the role of managing this system and worked closely with the group before, during and after meetings. Figure 29 gives an outline of the IMPACT programme management and reporting structure.

FIGURE 29.

The IMPACT programme management structure.

Expert dermatology nurse group

In February 2013, several members of the IMPACT programme team attended a psoriasis expert group workshop facilitated by IMPACT programme co-investigators Christine Bundy and Lis Cordingley. The group comprised six specialist dermatology nurses with extensive experience in the field. Some members of the group were responsible for developing the specialist dermatology nurse’s curriculum as part of the BDNG. This expertise helped the IMPACT programme team gain valuable insights into the current training curriculum.

The workshop provided an opportunity to share knowledge and ideas on how IMPACT programme workstream 5 could progress with regard to developing training for health-care professionals managing people with psoriasis. The expert group provided vital insights around the training needs of clinicians, particularly for practice nurses and DSNs, to develop the workstream 5 training package. Discussions about curriculum development were also useful in informing the work of workstream 4 in relation to health-care professional training needs in the context of supporting lifestyle change in patients with psoriasis.

Consultant dermatologists

During 2013 and 2014 there were four opportunities to speak to consultant dermatologists about their views on how training for dermatologists could be developed. Two of these were preceptor meetings (in which expert dermatologists provide in-depth training and supervision) run from a designated International Centre of Excellence at Manchester (Salford Royal NHS Foundation Trust) to train specialist dermatologists in specific aspects of psoriasis management. They are focused largely on management of severe psoriasis. The programmes are attended by around 40 delegates, on average. Other opportunities arose at two UK dermatology national conferences at which members of the IMPACT programme research team were presenting early findings. Dermatologists were asked what key elements needed to be included in an effective training programme to broaden the assessment and management of psoriasis and comorbidities. They were also asked about the training formats that they favoured and which locations were most likely to facilitate attendance. Dermatologists indicated that they were familiar with skills-based training and that this, rather than lecture-based training, was preferred in most cases. Many identified the barriers to communication between primary care and secondary care services, including the limited information that is passed to secondary care at the point of patient referral.

General practitioners with a special interest in dermatology

During July and August 2013, 20 GPSIs practising across the UK were invited to participate in a telephone interview with one of two members of the research team: Anna Chisholm or Christina Pearce. Twelve GPSIs took part and provided insights that usefully informed the work being conducted in workstream 4 (lifestyle behaviours in psoriasis) as well as the development of workstream 5 (health-care practitioner training). In relation to supporting patients with LBCs, GPSIs reported that knowledge of the relevance of lifestyle factors in psoriasis management was increasing, but that people with psoriasis are not routinely provided with this kind of support. As this insight corresponded with findings from workstreams 2 and 4, the GPSIs were prompted during the interviews to consider how services in this area could be improved.

The GPSIs identified a range of ways in which current psoriasis services could be improved to ‘integrate’ LBC support, including (1) what services should involve (e.g. patient management plan cards, longer consultations, psoriasis update training, communication skills training), (2) how improvement could be achieved (e.g. in-house training, 3-day courses, website or e-learning, establishing champions for more ‘integrated’ services), (3) where psoriasis services should be delivered (e.g. primary care settings, social or community centres, secondary care) and (4) who should deliver services (e.g. multidisciplinary teams, specialist nurses, whole practices, psychologists or lifestyle specialists). The information gathered from this consultation exercise enabled the IMPACT programme investigators to gauge the applicability of IMPACT programme findings to current real-world contexts, and to shape the design and implementation of the workstream 5 health-care practitioner training.

The internet and social media

The IMPACT programme website [www.impactpsoriasis.org.uk (accessed 12 October 2019)] is a platform for communication, information and dissemination. Study news, updates and recently published work are regular features and it is a place where members of the public can find out more about the IMPACT programme mission. Since October 2014, when the website was relaunched, there have been > 6000 new visitors to the site from across all continents and > 1000 hits on the information specifically aimed at patients, members of the public and practitioners.

To engage with the wider social media community both in the UK and internationally, the IMPACT programme Twitter (Twitter, Inc., San Francisco, CA, USA) account, @impactpsoriasis, has been an important tool for news and networking. A range of information is publicised via this account, including study updates and psoriasis-specific information of interest to our followers, who include members of the public, health-care practitioners and individuals and organisations interested in dermatology research. We have > 1400 followers worldwide.

Manchester Psoriasis Shout Out!^® 2014: a festival bringing academics, researchers, clinicians, patients and members of the public together

The Manchester Psoriasis Shout Out!^® (a collaboration between University of Manchester, Salford Royal NHS Foundation Trust, and Manchester Academic Health Science Centre) was a week-long series of events in 2014 aimed at raising awareness about psoriasis, engaging members of the public in psoriasis research taking place in Greater Manchester and discussing opportunities for involvement. This event was also a great chance to listen to thoughts and views from members of the public and health-care practitioners and to disseminate IMPACT programme findings.

As part of the Manchester Psoriasis Shout Out!, the IMPACT programme team held a practitioner-focused networking event. This was a chance to showcase research findings directly to health-care practitioners managing patients with psoriasis and also introduce the workstream 5 PSO WELL clinician training package developed from IMPACT programme phase I (workstreams 1–4). The opportunity for practitioners to ask questions about the training package and register interest in taking part in the workstream 5 study formed part of our recruitment strategy.

Members of the RUG, the patient representatives and the local Salford Psoriasis Association group attended events and assisted in staffing (together with researchers and clinicians) the Manchester Psoriasis Shout Out! trailer during the week. Photographs from the event are included in Figure 32 and some of the RUG members took part in the films created for the event, which can be found online [www.psoriasisshoutout.co.uk/#!films/c1ql4 (accessed 12 October 2019)].

FIGURE 32.

Photographs of events undertaken as part of Manchester Psoriasis Shout Out! 2014. (a) The IMPACT workstream 5 PSO WELL launch at the Midland Hotel, Manchester. Pictured left to right: Alison Littlewood, Christine Bundy, Christopher EM Griffiths and Lis Cordingley; (b) the Manchester Psoriasis Shout Out! trailer in St Ann’s Square, Manchester. Pictured: Karen Kane; (c) the Manchester Psoriasis Shout Out! flash mob in full flow, St Ann’s Square, Manchester; (d) the Manchester Psoriasis Shout Out! poetry event, Cornerhouse, Manchester; (e) patient and public question and answer session: ask the experts, Salford Royal NHS Foundation Trust. Pictured: Christopher EM Griffiths; (f) the Manchester Psoriasis Shout Out! fashion show finale, Trafford Centre, Manchester. Permissions obtained for photographs.

Identifying and prioritising, design, and development of the grant proposal

Identifying and prioritising, design, and development of the grant proposal are grouped together as these three stages are not linear activities but an iterative, interconnected process in which earlier stages continue to be revisited until the grant proposal takes its final form. In the early stages of the IMPACT programme proposal, Gladys Edwards and Helen McAteer, representing the Psoriasis Association, were involved in identifying and prioritising research questions to be addressed in the grant application along with the design of research questions and methods. Helen McAteer reviewed participant recruitment strategies and, once the IMPACT programme was funded, supported recruitment by providing information to the Psoriasis Association membership on how to participate in workstream 3, ‘coping with psoriasis’, which yielded a good response. From the outset, there has been patient representation, with Gladys Edwards, Helen McAteer and Toni Doyle all key co-investigators throughout the development of the grant proposal. At the very start of IMPACT programme development, a Psoriasis Association survey about poor experiences of care was a major influence on the IMPACT grant application. The NHS patient experience representative ran a series of patient experience meetings (including dermatological conditions) across Salford PCT. The challenge to engaging in PPI work in the design and development stages is establishing a wider group of patient advisors in the area of research and not only those who are already members of active patient associations when few funds are available and fixed-term research staff are not yet in post because the grant is not yet awarded. 201 The RUG and the patient representatives on the newly formed steering committee in effect had less influence over the core design of the first four workstreams of the IMPACT research programme because these had already been planned during the grant application process. However, the design and content of workstream 5 was developed using both the evidence gathered from research participants in earlier workstreams as well as the input of the RUG members (including the patient representatives on the steering committee). The perception that it is not possible to influence a programme of activity that is already in motion can be frustrating. This has been reported by patient advisors recruited to projects that have already been established202 and has also been expressed by IMPACT programme patient representatives during the course of the programme. As this current programme of research ends, we have come full circle through all the research stages and can consider what should be done next. We continued to develop our ties with the Psoriasis Association and, in addition, developed a local, ‘grassroots’-level RUG group, which has continued to engage and develop.

Undertaking and managing

As depicted in Figure 29, the IMPACT programme management structure included patient representation at both programme management committee (HM and TD) and programme steering committee (HM, TD and three patient representatives) level. This was crucial to the running of the research programme.

Members of the RUG have commented and advised on several aspects of the research, including recruitment strategies, questionnaire content and design, and participant information. These are in addition to the development of the workstream 5 PSO WELL patient resources designed to support patients with psoriasis to self-manage in better ways. Prior to the formation of the group, local Psoriasis Association members commented on workstream 3 recruitment materials and an established primary care user group was contacted to comment on workstream 2 recruitment materials.

Although commenting on patient information sheets and other study documents can be seen as tokenistic (which, if there are no other opportunities for involvement, may be more likely), having patient/public input into these is a valuable adjunct to the work of the researchers. Regardless of the research team’s writing skills, testing study documents with people who have experience of the particular condition can often increase the relevance of documentation to that specific patient group, which in turn makes a successful study more likely. Such reviewing activity often plays to the strengths of some RUG members whose day-to-day life, training and employment enable them to use well-developed editorial skills to enhance research materials that are aimed at patients.

It is difficult to measure the impact of this type of input in a quantifiable way, though it is suggested that those studies that use PPI in some form are more likely to reach 90% of target recruitment. 203 Workstream 2 reached 90% of its target recruitment and all the qualitative parts of the programme recruited and sampled widely to reach data saturation on main themes. Although this cannot be attributed entirely to RUG involvement (other strategies to increase participation such as widening our geographic area of recruitment were introduced),204 the expertise of RUG members in reviewing procedures and materials for the study undoubtedly contributed to successful recruitment.

An often forgotten outcome of this type of exercise may be the enduring effect on researchers. As researchers, we learn through secondary experience (gained from the RUG group and participants in the study) what it feels like to receive, make sense of and respond to the documents we send out and the procedures we ask to be completed. This continuous exposure influences the way we construct our participant materials, collect data and think about recruitment in all future research activity.

Analysing and interpreting

Results across the programme have been presented to RUG members at regular junctures to compare the findings with their experiences. It has been an interesting process and we uncovered some knowledge gaps among RUG members, in particular with respect to the existing links between psoriasis and unhealthy lifestyle behaviours. This challenged the assumptions of the researchers that newly emerging evidence about the influences of lifestyle on psoriasis had already been cascaded to and assimilated by patients and health practitioners. This led the team to think more about dissemination and future implementation challenges for the intervention developed in the final phase of the programme.

Dissemination

Timely dissemination during the IMPACT programme has ensured that research findings are published and publicised quickly, with delivery methods aimed appropriately for our wide range of stakeholders. A quarterly IMPACT programme newsletter is distributed to all investigators, researchers, patient representatives, RUG members, professional representatives, independent expert advisory panel members, steering committee members and the research sponsor covering research progress and highlighting findings.

Patient and practitioner dissemination has taken different forms over the course of the research programme.

Implementation

The input of the RUG and the study participants has had considerable influence on the development of the patient materials, making them relevant and increasing validity. Our ongoing RUG meetings are helping the research team think about future developments and how the final interventions may be tested in a real-world situation.

Lessons learned

Several lessons have been learned along the way. Key insights include the importance of:

• Involving the RUG and patient representatives from the outset. Although IMPACT programme phase 2 afforded opportunity for wider PPI, phase 1 was under way by the time the group was formed, making it challenging for members to fully contribute in the early stages. Involving the group from the grant-writing stage would ensure full participation in the identification and prioritising of research questions. With the psoriasis RUG now established, future programmes will benefit from having close links with this expert group from the beginning.

• Ensuring an explicit PPI strategy is set from the start of the programme, with key milestones throughout each workstream, to which the PPI members and research team are fully committed.

• Implementing a more defined reporting structure between the patient representatives and the RUG members to incorporate wider viewpoints and for the patient representatives to use the RUG as a sounding board for some of their ideas.

Psoriasis and cardiovascular disease

The IMPACT programme found that there is an increased risk of CVD for patients with severe psoriasis (for definitions of severity see Epidemiology of psoriasis and its association with risk of cardiovascular disease) and that psoriasis patients in general have higher than expected levels of CV risk factors when compared with non-psoriasis populations.

However, importantly and in contrast to the highly cited and influential work of Gelfand et al. , and using the same data source, we concluded that psoriasis alone is not an independent risk factor for CVD. Indeed, our work found that co-occurrence of psoriasis and inflammatory arthritis is a risk factor for CVD. It seems highly likely that the Gelfand study neither identified nor controlled for inflammatory arthritis.

Screening for cardiovascular disease risk

Our team carried out CVD risk screening for patients with psoriasis (aged both < 40 years and ≥ 40 years) attending primary care practices in the north of England. Practitioners were asked to follow their usual CVD screening procedures, which typically involved blood tests and either one or two consultations. Follow-up consultations were usually undertaken by telephone or face-to-face appointments with nursing staff. Primary care screening of psoriasis patients identified a high proportion as being at ‘high risk’ of CVD, with potentially modifiable CVD risk factors ascertained. For example, over half of screened psoriasis patients had a very high waist circumference measurement and 29% had a raised blood pressure. A high proportion of those with previously known risk factors did not appear to be in receipt of effective interventions and those with PsA were at highest risk.

Although these data seem to indicate the potential value of screening, the parallel mixed-methodology arm of the screening study produced new insights into why CVD screening may be ineffective. There is an assumption that if a problem is identified via screening then it will lead to intervention. Our team found very little evidence of clinicians intervening, with many using screening as primarily a data-recording process rather than a ‘teachable moment’ for their patients. In the majority of interactions, health-care professionals did not expand on or tailor the risk information provided for the patient. In some cases, risk reduction was advised; however, specific advice or support was very rarely given. There were very few examples of the use of established behaviour change techniques even for those patients who actively viewed the consultation as an opportunity to discuss changes to their health risk status. Practitioners and patients had different perspectives on the consultation as providing a potential opportunity to intervene. Practitioners often viewed the encounter as unlikely to lead to changes, whereas patients generally saw practitioners as influential in supporting lifestyle management.

Strategies for communicating CV risk information varied widely between clinicians. In general, clinicians were more likely to discuss biomedical risk factors, that is, factors such as hypertension or hyperlipidaemia, most of which are managed pharmacologically. Disappointingly, the percentage of consultations in which clinicians engaged in discussions about behavioural risk factors (physical activity levels, diet, alcohol consumption or smoking) ranged between 27% and 40%. Even fewer of these consultations included specific discussions with patients or interventions to support patients to reduce risk factors.

Capacity of services to reduce cardiovascular risk factors for patients with psoriasis

Even if health-care professionals seeing psoriasis patients had a broad understanding of the CVD risk factors associated with the disease, IMPACT programme research has identified two crucial factors that mitigate against any likely improvement in psoriasis-associated CV risk. The first is a clear deficit in the knowledge and skill sets of both primary and secondary care clinicians in behaviour change techniques. These include a range of evidence-based techniques that are most likely to engage individuals in improved health behaviours. Across all IMPACT workstreams, but most noticeably in workstreams 2 and 4, practitioners indicated that they had received little exposure to behaviour change interventions, which were also absent from the curricula used for both undergraduate and postgraduate training of GPs, nurses and dermatologists. Additional in-depth interviews with practitioners indicated that they held narrow biomedical ‘personal models’ of psoriasis in which management strategies were restrictive and ‘skin focused’. They rarely identified a role for contributory lifestyle behaviours or saw the need to address these.

The second factor is the low level of understanding of the links between lifestyle behaviour, CVD risk and psoriasis outcomes in patients with psoriasis. At one of our earliest patient engagement meetings, participants let it be known that they did not understand why our research focus was on lifestyle health behaviours. Few had been aware of psoriasis-associated comorbidities. A small number had been told to avoid alcohol because of potential interactions with medication (usually methotrexate), but none had been aware of the links between psoriasis severity and weight, for example. The observation study undertaken as part of workstream 4 demonstrated that there were no psoriasis-specific materials for patients on lifestyle behaviours and how they might contribute both to psoriasis severity and to CV risk. Furthermore, even generic LBC materials were either badly displayed or not available in health settings most likely to be used by psoriasis patients.

Thus, IMPACT programme researchers identified correspondingly limited ‘models’ or understanding of psoriasis held by both practitioners and patients. The challenge where patients were concerned would be how to broaden patients’ understanding of psoriasis without raising levels of poor mood or adding to feelings of helplessness and hopelessness, particularly in patients with high (albeit largely unrecognised) levels of illness burden.

Psychological well-being

As anticipated, both the CPRD study and the screening study identified higher levels of low mood than those in matched populations. In the CPRD, just over 4% of those with psoriasis had a concomitant Read code for depression. In the CV screening study, a much larger number of patients reported high HADS scores, indicating likelihood of anxiety and/or depression (data available from programme team). These were also the psoriasis patients most likely to engage in unhealthy behaviours. These findings reinforced reported experiences of our patient RUG members and the in-depth qualitative findings of the workstream 3 study into coping with psoriasis.

Workstream 3 provided new insights into the experiences of people with psoriasis. Although there is extensive research evidence demonstrating that psoriasis is a stigmatising condition that has a major impact on quality of life and well-being, participants in workstream 3 confirmed that this is rarely, if ever, discussed in primary care consultations. The absence of discussions that acknowledged the demands of living with psoriasis or the significant self-management challenges associated with treatment adherence contributed markedly to patients’ feelings of low self-worth. Patients described withdrawing from primary care services and their low expectations of good-quality long-term care. Findings from this workstream also raised questions about how many people have been missed who would fulfil criteria for referral to secondary care because they had disengaged with services or how many remained unaware of the significant developments in psoriasis treatment that have occurred over the last decade.

New paradigm for psoriasis management

Consideration and discussion of IMPACT programme findings, including those undertaken as part of the synthesis phase, consolidated growing awareness of a need for a different approach to psoriasis management in both primary and secondary care. The paradigm shift would be about moving people’s ‘working model’ of psoriasis management from an approach that was ‘skin focused’ to one that recognised psoriasis as a complex long-term condition. This would mean a focus on the relapsing–remitting nature of psoriasis, the management (or even prevention of) physical and psychological comorbidities and, finally, a new recognition of the role of health behaviours in determining the severity and course of psoriasis and comorbidity risks. The team recognised that this is what occurred in the field of diabetes management > 20 years ago, when significant energy was subsequently directed to the transformation of diabetes services.

Following the planned evidence synthesis process undertaken by the team, two broad areas for intervention were identified. The first target was to improve the capacity of the health-care workforce to address the needs of people with psoriasis and the second was to help empower people with psoriasis by giving them a broader understanding of psoriasis including potential comorbidities and how to minimise them. These included an ‘understanding’ or knowledge component, an attitudinal component and a skills component:

• Knowledge component – the target was to broaden both patients’ and health-care practitioners’ understanding of psoriasis as a long-term condition, with associated comorbidities, and as one in which health behaviours contribute to both onset and severity.

• Attitudinal component – for health-care practitioners, the target was to raise the profile of psoriasis as more than a ‘simple skin condition’ and as an area worthy of their skilled attention; the target for patients was to transform attitudes of passive acceptance (or even hopelessness) to ones of assertive partnership with their health-care practitioners.

• Skills component – for practitioners, the skills improvement target was an increased capacity to perform integrated clinical management combining physical, psychological and behavioural components to improve outcomes while recognising the limited time constraints in which they are working; for patients, the skills component was focused on self-management, getting more out of health-care consultations, improved medication adherence and the ability to engage with relevant healthy lifestyle behaviours.

The outcomes of both the patient- and practitioner-focused components of workstream 5 have been very promising. The team identified means of improving understanding, attitudes and skills in both groups as well as providing some solid preliminary evidence for likely mechanisms of action and stakeholder acceptability. Importantly, the lessons learned from this work are likely to be relevant to other long-term conditions and their management. Finally, a practical contribution has been made in terms of the production and evaluation of very high-quality practitioner training materials and patient support materials that are now available for dissemination and further evaluation.

Qualitative research in dermatology

The IMPACT programme has resulted in a number of methodological innovations, the most notable being the application of qualitative and mixed-methods approaches to dermatology research for the first time. Qualitative research methods had rarely been used in dermatology research and the IMPACT programme is viewed as pioneering in this respect. Qualitative and mixed-methodology studies of the experiences of both patients and practitioners led to a number of new findings and insights that were subsequently utilised in designing the (workstream 5) interventions.

As well as standard semistructured interview and focus group approaches, qualitative data elicitation techniques included tape-assisted recall in workstream 2 and structured discussions with the multidisciplinary team of IMPACT programme experts to develop the models for the DES in workstream 1. Mixed methodological techniques included a broad range of techniques including completion of mixed open and closed question response sheets by professional ‘simulated’ patients, actors trained to respond as if they were a patient and stimulus response sheets containing a mixture of open and closed items completed by real patients in response to receiving the patient materials as part of workstream 5. The team also recognised that it may be challenging to ask patients to articulate why they found particular aspects of the interventions useful and so developed a system that allowed patients to simply identify sections of materials that they believed had changed their understanding of their condition using symbols. A structured observational study was undertaken to identify the quality and relevance of existing patient materials for individuals with psoriasis. As part of the evidence synthesis phase, stimulus materials were used to trigger structured discussions, which were recorded on specially designed tables; data from these were then collated and used as a focus for the final IMPACT programme workstream.

The visibility, impact and high quality of qualitative and mixed-methodology research outputs led to the IMPACT programme lead qualitative researcher, Pauline A Nelson, being appointed as the first qualitative research section editor of the British Journal of Dermatology. This appointment was heralded by the editor in the most recent British Journal of Dermatology position statement207 and the need for qualitative research in dermatology highlighted in one of Nelson’s early editorials:208

These findings often go beyond ‘static’ description to a dynamic explanation of the data . . . We owe it to patients . . . and . . . professionals to push methodological boundaries in the field, not only making qualitative methods an established component of dermatology research but also moving beyond the more conventionally employed interview and focus group.

Nelson208

Methods summary

The range of methods used enabled the team to apply key steps and processes and issues identified in the MRC framework,156 which was designed to facilitate the development of complex health interventions. These also included testing the use of existing or adapted quantitative measures and the development of new measures in the context of psoriasis. This pioneering work has value to both this team and others undertaking dermatology-related research.

Key messages and implications for practice

Epidemiology and mechanisms

From the epidemiological findings, mechanisms by which inflammatory arthritis drives CVD risk are still unknown. Similarly, although we identified links between low mood and risk, the specific direction of the relationship remains unclear. One limitation to the conclusions that could be drawn from the epidemiological work was the need to restrict inclusion criteria to those individuals for whom psoriasis developed in early adulthood (at minimum age of 20 years). There may be a different risk profile (and/or different mechanisms) for those who developed psoriasis in childhood, in whom onset following infection is more common.

The PSO WELL interventions

From the intervention stage (workstream 5), outstanding questions include whether or not (1) practitioners can implement and sustain the learning from the PSO WELL training in real-world clinical situations (and which types of practitioners are best placed to deliver this new way of working to psoriasis patients), (2) they can also incorporate effective use of the PSO WELL patient materials alongside their newly acquired knowledge and skills (and indeed whether or not and how this should be tested), (3) combining the PSO WELL clinician training and patient materials intervention would have a greater effect than using each separately and (4) the interventions are actually cost-effective when applied in practice.

Workstream 1

Parisi R, Symmons DP, Griffiths CE, Ashcroft DM, Identification and Management of Psoriasis and Associated ComorbidiTy (IMPACT) project team. Global epidemiology of psoriasis: a systematic review of incidence and prevalence. J Invest Dermatol 2013;133:377–85. https://doi.org/10.1038/jid.2012.339

Hamilton MP, Ntais D, Griffiths CE, Davies LM. Psoriasis treatment and management – a systematic review of full economic evaluations. Br J Dermatol 2015;172:574–83. https://doi.org/10.1111/bjd.13486

Parisi R, Rutter MK, Lunt M, Young HS, Symmons DPM, Griffiths CEM, et al. Psoriasis and the risk of major cardiovascular events: cohort study using the Clinical Practice Research Datalink. J Invest Dermatol 2015;135:2189–97. https://doi.org/10.1038/jid.2015.87

Workstream 2

Keyworth C, Nelson PA, Chew-Graham CA, Kane K, Pearce CJ, Griffiths CEM, et al. Communicating cardiovascular disease risk to people with psoriasis: what techniques do practitioners use? Int J Behav Med 2016;23:168–78. https://doi.org/10.1007/s12529-015-9517-8

Nelson PA, Kane K, Chisholm A, Pearce CJ, Keyworth C, Rutter MK, et al. ‘I should have taken that further’: missed opportunities during cardiovascular risk assessment in patients with psoriasis in UK primary care settings – a mixed-methods study. Health Expect 2016;19:1121–37. https://doi.org/10.1111/hex.12404

Rutter MK, Kane K, Lunt M, Cordingley L, Littlewood A, Young HS, et al. Primary care-based screening for cardiovascular risk factors in patients with psoriasis. Br J Dermatol 2016;175:348–56. https://doi.org/10.1111/bjd.14557

Keyworth C, Nelson PA, Bundy C, Pye SR, Griffiths CEM, Cordingley L. Does message framing affect changes in behavioural intentions in people with psoriasis? A randomized exploratory study examining health risk communication. Psychol Health Med 2018;23:763–78. https://doi.org/10.1080/13548506.2018.1427876

Workstream 3

Cordingley L, Nelson PA, Griffiths CE, Chew-Graham CA. Beyond skin: the need for a new approach to the management of psoriasis in primary care. Br J Gen Pract 2012;62:568–9. https://doi.org/10.3399/bjgp12X658133

Nelson PA, Barker Z, Griffiths CE, Cordingley L, Chew-Graham CA, IMPACT Team. ‘On the surface’: a qualitative study of GPs’ and patients’ perspectives on psoriasis. BMC Fam Pract 2013;14:158. https://doi.org/10.1186/1471-2296-14-158

Nelson PA, Chew-Graham CA, Griffiths CE, Cordingley L, IMPACT Team. Recognition of need in health care consultations: a qualitative study of people with psoriasis. Br J Dermatol 2013;168:354–61. https://doi.org/10.1111/j.1365-2133.2012.11217.x

Workstream 4

Keyworth C, Nelson PA, Chisholm A, Griffiths CE, Cordingley L, Bundy C, Identification and Management of Psoriasis-Associated Co-morbidiTy (IMPACT) team. Providing lifestyle behaviour change support for patients with psoriasis: an assessment of the existing training competencies across medical and nursing health professionals. Br J Dermatol 2014;171:602–8. https://doi.org/10.1111/bjd.13067

Nelson PA, Keyworth C, Chisholm A, Pearce CJ, Griffiths CE, Cordingley L, et al. ‘In someone’s clinic but not in mine’: clinicians’ views of supporting lifestyle behaviour change in patients with psoriasis – a qualitative interview study. Br J Dermatol 2014;171:1116–22. https://doi.org/10.1111/bjd.13231

Keyworth C, Nelson PA, Griffiths CEM, Cordingley L, Bundy C. Do English healthcare settings use ‘Choice Architecture’ principles in promoting healthy lifestyles for people with psoriasis? BMC Health Serv Res 2015;15:215.

Chisholm A, Nelson PA, Pearce CJ, Keyworth C, Griffiths CE, Cordingley L, Bundy C, IMPACT Team. The role of personal models in clinical management: exploring health care providers’ beliefs about psoriasis. Br J Health Psychol 2016;21:114–34. https://doi.org/10.1111/bjhp.12148

Workstream 5

Chisholm A, Nelson PA, Pearce CJ, Littlewood AJ, Kane K, Henry AL, et al. Motivational interviewing-based training enhances clinicians’ skills and knowledge in psoriasis: findings from the PSO WELL^® study. Brit J Dermatology 2017;176:677–86. https://doi.org/10.1111/bjd.14837

Nelson PA, Kane K, Pearce CJ, Bundy C, Chisholm A, Hilton R, et al. ‘New to me’: changing patient understanding of psoriasis and identifying mechanisms of change. The Pso Well^® patient materials mixed-methods feasibility study. Br J Dermatol 2017;177:758–70. https://doi.org/10.1111/bjd.15574

PhD thesis (associated with workstreams 2 and 4)

Keyworth C. Risk Communication and Lifestyle Behaviour Change in People with Psoriasis. PhD thesis. Manchester: University of Manchester; 2015. URL: www.escholar.manchester.ac.uk/api/datastream?publicationPid=uk-ac-man-scw:266982&datastreamId (accessed 28 January 2019).

PhD thesis abstract

People with psoriasis are known to engage in high levels of unhealthy lifestyle behaviours that may lead to poorer psoriasis outcomes and increase the risk of CVD. Thus, helping individuals with psoriasis understand the link between behaviours and health risks, that is health risk communication and direct support for LBC are important aspects in optimal management of psoriasis, a long-term inflammatory skin condition.

There are two aspects of the literature that remain unclear: first, whether or not adequate support is given to patients to enable them to understand that the links between lifestyle behaviours and health outcomes is part of psoriasis patient management strategies; and, second, whether or not there is agreement around effective health risk communication techniques. This programme of research aimed to examine these gaps in the literature using four related studies.

The first study used content analysis to examine general and dermatology-specific health-care professionals’ core training competencies for evidence of skills relating to LBC. An important finding was the lack of explicit skills relating to LBC and changing understanding of health risks. There was little or no reference to recognised LBC techniques that could be used to support and facilitate LBC with patients.

The second study used observational techniques to examine messages about the links between behaviour and health outcomes and LBC signposting (such as leaflets or posters about healthy living) for patients with psoriasis in primary and secondary care patient waiting areas. There was little evidence of psoriasis-specific information about healthy living. Generic information (not specifically about psoriasis) was often of poor quality and poorly displayed and did not conform to evidence-based recommendations for effective LBC signposting.

The third study combined observational and qualitative techniques to examine how health-care professionals communicate information about CVD risk to patients and the role of LBC in reducing risk in the context of primary care risk assessments with people with psoriasis. A key finding was that interpretation of risk information was not always linked to specific advice about how to modify each risk factor. Discussion was mostly instructional rather than a shared collaborative discussion about behaviour change and risk reduction.

The fourth study used experimental methods to examine the effects of message-framing theory as a health risk communication strategy on reported behavioural intentions in people with psoriasis. An important finding was that, for messages about psoriasis symptom reduction, gain-framed (positively framed) messages were more effective in increasing behavioural intentions for reducing alcohol consumption. Conversely, for messages about CVD risk reduction, loss-framed (negatively framed) messages were more effective for increasing behavioural intentions to reduce alcohol consumption.

The body of work presented in this thesis demonstrates that much needs to be done to increase the skill sets of health-care professionals to help people with psoriasis recognise the specific links between their own health behaviours and health outcomes. In addition, specific recommendations have been suggested as a way of improving risk communication strategies, such as using theory-based personally relevant health information for people with psoriasis.

Lay abstract

Psoriasis is a long-term skin condition that often appears as skin redness and ‘plaques’ on the surface of the skin. People with psoriasis are known to engage in an unhealthy lifestyle (increased alcohol intake, smoking, insufficient physical activity and poor diet). This worsens psoriasis and leads to the additional risk of CVD (e.g. heart attack). Telling people about these risks should be a key part of health-care management in patients with psoriasis. Yet there is limited evidence of the best way of doing this, in a way that leads people to live healthier lives and reduce long-term health risks.

The programme of research aimed to examine this using four research studies.

The first study examined whether or not the core training goals for primary and secondary health-care professionals (e.g. doctors, nurses, dermatologists) included information about the importance of healthy lifestyle. An important finding was the lack of reference to specific ways that health-care professionals could encourage and support patients to live healthier lifestyles (e.g. setting small achievable goals, setting an action plan).

The second study examined information available to patients about healthy living (such as leaflets or posters about healthy living) in primary and secondary care patient waiting areas (e.g. doctor’s surgeries). There was little evidence of psoriasis-specific information about healthy living. Generic information (not specifically about psoriasis) was sometimes available but the quality was often poor and not displayed in a way that patients could read and understand the information.

The third study examined how doctors and nurses talk about CVD to patients with psoriasis. A key finding was the lack of specific advice about how patients could have healthier lifestyles to reduce long-term health risks.

The fourth study examined the effect of differently worded health messages (positively worded or negatively worded) on whether or not people would decide to make healthy changes to their lifestyle (e.g. increasing physical activity). When patients were told that healthy living improves the visible appearance of psoriasis, positively worded messages were more effective than negative ones in encouraging people to consider making lifestyle changes. When patients were told that healthy living would reduce the likelihood of CVD (e.g. heart attack), negatively worded messages were more effective.

The research presented in this thesis suggests that more attention could focus on encouraging patients with psoriasis to live healthier lives. Specific ways of improving how patients with psoriasis are told about this involves appropriately worded, personally relevant health information.

Survey design: time attribute

A fifth attribute was added to the stated preference survey about the length of time it would take for patients to read and use the information, manage their psoriasis and make any lifestyle changes. Many stated preference surveys use cost as a quantitative attribute to assess the amount of money that respondents are willing to pay in exchange for their preferred type of intervention. In this survey, time is used rather than cost. On the basis of the work to develop the intervention and discussions with the IMPACT research team, it was felt that time better reflected the trade-offs relevant to users of psoriasis care, which is not directly charged for in the NHS. It was anticipated that, overall, participants would prefer to spend less rather than more time to understand and manage their psoriasis. To explore this in the analysis, the time attribute had six levels (30 minutes, 1 hour, 2 hours, 3 hours, 4 hours and 8 hours). The decision to use 30 minutes as the base level was based on the nature of psoriasis and the need to manage the condition on a daily basis.

Fractional factorial design

The five attributes, combined with the number of levels (four attributes with three levels and one attribute with six levels), gives a total number of 486 scenarios describing different combinations of the features of the PSO WELL patient material and clinician training intervention. It is not feasible to design a survey to assess preferences for all of these scenarios. To address this, a fractional factorial design was used to reduce the number of scenarios by selecting a sample of possible combinations that covers the combinations and effects of interest. 175 A published design catalogue (http://neilsloane.com/oadir/MA.18.3.6.6.1.txt; accessed 28 January 2019) and modulo arithmetic were used to generate an efficient, orthogonal fractional factorial design176 to estimate the main effects of each attribute. This gave a design with 18 choice sets. The efficiency of the design was assessed using online software (Discrete Choice Experiments). The design was estimated to be 94% efficient compared with an optimal two-choice design and 65% efficient compared with a six-choice design for a main-effects analysis.

Sample size estimate

There is no clear formula for working out sample sizes, but an initial estimate was derived based on the number of choice sets in the fractional factorial design211 for a main-effects design. Based on this approach, for each choice set it was assumed that the true probability of choosing option 1 is 50% and the probability of choosing option 2 is also 50%. It was also assumed that an acceptable error for this probability is ± 2.5% in the sampled population. With a survey of 18 choice sets, this equates to a minimum of 170 completed questionnaires in total (85 × 2 questionnaires of nine choice sets). Based on previous experience, we anticipated that approximately 30–40% of participants would complete all the choice sets in a questionnaire. This indicates that 213–283 participants are needed to get 85 complete responses per choice question. A higher number of responses will be required if there is significant heterogeneity in the sample of participants. Stated preference survey guidelines suggest that 500 participants is sufficient for a statistical analysis that accounts for heterogeneity. Accordingly, we initially aimed to recruit 500 participants.

Post hoc sample size estimates used the approach recommended by de Bekker-Grob et al. 177 to estimate the numbers needed for a main-effects analysis that can identify statistically significant preferences for each level of each attribute. The hypotheses for the sample size estimates were that participants would prefer:

• printed information and printed plus online information to no information

• personalised information and personalised information plus support for lifestyle changes to no information or support

• understanding of psoriasis, treatments and health to be the same as or better than before rather than worse than before

• ability to manage psoriasis and make lifestyle choices to be the same as or better than before rather than worse than before

• to spend less rather than more time to manage their psoriasis.

Sample sizes were estimated for time as a continuous variable. Inputs to the sample size calculations included the coefficients from the main analysis (used as effect size estimates), the study design, 5% statistical significance level and 80% statistical power. The calculations were done in R version 3.2.0, based on the code published by de Bekker-Grob et al. 177

Participant recruitment and survey administration

The stated preference survey was advertised via social media websites [e.g. Twitter and Facebook (Facebook, Inc., Menlo Park, CA, USA)], specifically targeting psoriasis pages or groups to recruit participants. The Psoriasis Association also advertised the survey on their website, newsletter and through social media channels. Psoriasis Association members and IMPACT programme participants (who had agreed to be contacted about further research) were e-mailed. The survey was developed for participants to complete online.

Participants who responded to the adverts followed a web link to see the survey instructions, a cover letter, a participant information sheet, survey instructions and the survey itself. The survey was then randomly selected for the participant to complete. Participants were not offered any incentives to complete the survey.

The survey software used was LimeSurvey (LimeSurvey GmbH, Hamburg, Germany), managed and hosted on a University of Manchester website.

Analysis

Descriptive statistics were used to summarise participant characteristics and the survey responses. Logistic regression was used to explore whether or not there were differences in respondent characteristics between the survey versions and the option chosen for each choice set.

Effects and dummy coding of the attributes give different absolute estimates of the coefficients or strength of preferences. However, the relative difference for each categorical attribute level compared with the lowest (reference) level will be the same for the two coding methods. 212 Accordingly, the primary, sensitivity and subgroup analyses used dummy coding of the attributes to explore how the coefficients (or preferences) changed for each level. Effects coding was used to retrieve the preference weight (coefficient) for the reference level to assess the functional form of the attributes. 175,213

The primary analysis of the choice questions included only participants with complete responses to these questions. The responses to the choice questions were analysed with a conditional logistic regression model using maximum likelihood estimation to estimate the preference weights (coefficients) for each of the attributes. This model is a simple fixed-effects model that takes into account the panel nature of the data and is widely applied to analyse discrete choice data in health care. 175,213,214

The coefficients for each attribute tell us how strongly users feel that it is an important part of the PSO WELL intervention. Random utility theory assumes that a responder chooses between two options by interpreting the information described as a set of characteristics and selects the one that provides the highest utility overall. Based on this assumption, the results of the regression analyses were used to calculate MRS between the different attributes. This allowed us to estimate the relative time participants were willing to trade for the different features and assess how important they are (e.g. how much does the provision of information materials matter to users compared with other aspects, such as their ability to make lifestyle changes?). The MRS was calculated in Stata 13 as the coefficient for each attribute level divided by the time coefficient using the non-linear combination of estimates command.

The conditional logit model assumes that, across all participants, all the choice questions/tasks measure utility equally well. If this is not the case then the coefficients from the regression cannot be used to assess the relative importance of different levels across attributes (scale heterogeneity). However, the overall aim of the analysis was to estimate the MRS or willingness to spend time to gain the preferred level of each attribute. This deals with the issue of scale heterogeneity by essentially converting each coefficient to the same scale.

The assumption of fixed effects means that the choice of options in a choice set is determined by the attributes and levels assigned to the option and assumes that there are no systematic, unobserved differences in preferences between respondents. The sensitivity analysis included a random parameters logistic regression model to explore whether or not relaxing this assumption affected the marginal rates of substitution.

The data were downloaded and prepared for analysis in IBM SPSS Statistics version 22 (IBM Corporation, Armonk, NY, USA) and analysed in Stata 13.

Participant characteristics and time to complete survey

The characteristics of participants are summarised in Table 15. The characteristics of participants completing surveys 1 and 2 are presented in Table 16 and characteristics of participants choosing option 1 or 2 are presented in Table 17. Participants who did not answer any choice questions (160/526; 30%) also did not complete the questions about their sociodemographic characteristics or treatment, which were at the end of the survey. Out of the 116 people who completed one or more but not all nine of the choice questions, 11 (5%) answered one or more of the sociodemographic questions. This lack of data means that it is not possible to assess whether or not there were any differences between those who did and those who did not answer any, or all, choice questions. As is typical with anonymous online surveys, it was not possible to collect comprehensive data about the number of people who saw the adverts or e-mail invitations or who then visited the website to look at the details about the survey.

Including all eligible respondents, the average time to complete the survey was 8 minutes (SD 22 minutes), although this ranged from < 1 minute to 416 minutes. Five participants had > 1 hour between the start and finish times (range 96–416 minutes). There were no differences in average time between those who started survey 1 (mean 7 minutes, SD 12 minutes; n = 251) and those who started survey 2 (mean 9 minutes, SD 28 minutes; n = 275) (p = 0.362). As expected, there were differences in the time taken between those who completed one or more choice questions (mean 10 minutes, SD 24 minutes; n = 366) and those who did not (mean 3 minutes, SD 14 minutes; n = 160). This difference was statistically significant (p < 0.001). Similarly, there were differences in the time taken between those who completed all choice questions (mean 12 minutes, SD 29 minutes; n = 250) and those who did not (mean 5 minutes, SD 9 minutes; n = 116; p = 0.005). For 240 (96%) of the 250 of those who completed all choice questions, the mean time taken was lower than or within the time range given in the participant instructions (10–20 minutes).

Analysis of choice data

Effects coding was used in a conditional logistic regression model to allow estimation of the coefficient for the base level and to plot the coefficients as shown in Figures 23 and 24. The full results of the primary analysis of preferences, using dummy-coded attributes, are shown in Table 18. For the information attribute, Figure 23 illustrates the similarity in preferences for levels 1 (no information) and 2 (printed information materials).

The results of the primary analysis of preferences using dummy-coded attributes are shown in Table 18. For the attributes about information, clinician visits, understanding of psoriasis and comorbidity and ability to make lifestyle changes, these analyses indicate that participants preferred higher levels (levels 2 or 3) of each attribute rather than the reference or base level, which is in line with what was expected. In addition, the strength of preference measured by the coefficients was statistically significant for both levels compared with the lowest or reference level for these attributes. The exception to this was the information attribute, for which only the coefficient for printed and online information materials (level 3) was statistically significantly different from the base level of no information.

When treated as a continuous linear variable, the coefficient for time was negative and statistically significant, as shown in Table 18. This is in line with expectations that, all other things being equal, participants would prefer to spend less rather than more time to manage their psoriasis and make lifestyle changes. However, when time was analysed as a categorical variable, the results indicate that participants preferred to spend up to 2 hours rather than the base level (30 minutes) to manage their psoriasis (Table 19 and see Figure 24). At and beyond 2 hours, participants preferred to spend less time (compared with the base level) to manage their psoriasis. In addition, the level of 3 hours was not statistically significantly different to the base level of 30 minutes (see Table 19).

The post hoc estimates of the sample size needed to identify a statistically significant coefficient for each attribute level are shown in Table 20. If time is treated as a continuous variable, the results show that the number of participants who completed all choice questions is sufficient to identify statistically significant main effects for all but one of the attribute levels. The exception is level 1 (printed information) of the information materials attribute, which would require a sample size of 9373. If time is treated as a categorical variable, the number of complete responses required is 9598 for level 1 of the information materials attribute and 744 for the 3-hour level of the time attribute.

Time willing to trade

The time that participants are willing to spend (or trade) to gain the level of each attribute level that they prefer is shown in Table 21. The results suggest that improvements in ability to manage psoriasis and make lifestyle changes, along with clinic visits that include personalised information and support to make lifestyle changes, were the most important components of the PSO WELL intervention. The information materials are the least important components. The sensitivity and subgroup analyses indicated that there were no differences in the relative importance of the different attribute levels according to the analysis methods or participants’ demographic and treatment characteristics (Tables 22–24). The 95% CIs overlapped, indicating no statistically significant differences compared with the primary analysis.

For the information attribute, the printed information level was included in the primary analysis to illustrate its relative importance, even though it was not statistically different from the base level of no information. The sensitivity analysis indicated that there were no differences in the overall results from combining levels 1 and 2 into one level. Similarly, the time level for 3 hours (which was not statistically significant) was included in the primary analysis. In the sensitivity analysis, this was combined with the level of 4 hours. Table 22 indicates that this did not affect the relative time participants were willing to trade to gain improvements in the different components of the intervention.

Treatment as usual

At the start of each simulated patient journey, the patient enters the model with a profile of sociodemographic and health characteristics. The sociodemographic characteristics include age, gender, ethnicity and Townsend deprivation score for the local area the patient lives in. The health characteristics include the severity of psoriasis, PsA, anxiety and depression, BMI, type 2 diabetes, MI and risk of future MI or stroke. Treatment history attributes include prior and current psoriasis treatments. Lifestyle behaviours included are alcohol consumption, extent of physical activity and smoking status. 147 It is assumed that the likelihood that patients are willing to change their behaviours is the same for each behaviour (reduce alcohol consumption, stop smoking, improve diet and physical activity or make no change to any of these behaviours).

Until either the model time horizon is reached or the patient dies, the patient uses health-care resources (e.g. interacting with clinicians at clinical consultations, receiving treatment). The number of health-care services used and the costs depend on each person’s characteristics and how they interact with the health-care system. At the same time, the patient may or may not make changes to their lifestyle that influence their health and need for health care. These events directly affect the costs and QALYs accumulated by the patient. Outcomes are recorded at the end of the simulation so that mean and total costs and QALYs across the whole simulated population can be calculated for each arm of the model.

The model focuses on interactions between the patient and their GP and between the patient and the dermatology service. This is based on the assumption that these are the key relationships that will influence psoriasis treatment. However, the model does also include the wider costs and health impacts of the comorbidities associated with psoriasis. Whenever a patient attends a clinical consultation at either a general practice or dermatology clinic, the model records an interaction between the patient and the clinician (clinical agent). For simplicity, the model assumes that clinical agents possess two key characteristics (attributes) that affect their interaction with the patient and influence patients’ health-care costs and health outcomes. The first is the clinician’s propensity to make treatment decisions that are in line with evidence-based guidelines. This depends on their knowledge of relevant guidelines and assessment of their relevance to the patient and treatment decision required. The second characteristic is the clinician’s propensity to use motivational interviewing skills during consultations. Again, this depends on the clinician’s knowledge of and ability to use motivational interviewing skills and their assessment of the relevance of these for the patient and treatment decision.

Treatment as usual plus PSO WELL

The PSO WELL treatment arm is structured in the same way as the usual care arm but with the addition of the PSO WELL interventions. The intervention arm of the model is, therefore, TAU plus the patient materials and clinician training, considered together as a single PSO WELL package. It is assumed that the distribution of patient materials and training of clinicians occur right at the start of the simulation. The key differences between the PSO WELL arm and the TAU arm is the assumption that (1) the patient materials could change the likelihood that patients make beneficial changes to their lifestyle and health behaviours and (2) training makes clinicians more likely to provide appropriate treatment and use motivational interviewing techniques. These in turn influence patient health, need for health care and health-care costs.

The PSO WELL patient education intervention potentially motivates a patient to seek a GP appointment. In the IMPACT feasibility study, 27.3% (15/55) of participants who received the patient education materials intended to seek more information about psoriasis within 4 weeks. In the model, this is the likelihood that a patient will seek a consultation following the intervention. In the usual care arm, no new help-seeking behaviour is assumed.

The clinician training intervention is more influential in the model, increasing the likelihood that clinicians make guideline-concordant decisions and use motivational interviewing skills. 164 GPs respond well to direct information about the management of psoriasis; 78% of referrals to secondary care were considered appropriate from GPs who received guidelines developed by dermatologists, compared with 59% from other GPs. 130 The model utilises these results as the probability that a clinician will act in a guideline-concordant manner in treatment and referral decisions or deploy motivational interviewing skills during a consultation. For clinicians on the PSO WELL model arm, this probability is 78% compared with 59% of TAU clinicians. This differential is assumed to last for 1 year, after which the likelihood is 59% for all clinicians.

Agent interactions at consultations

Patient–practitioner interactions at consultations, where diagnoses and treatment decisions are made, determine many events that happen during a patient’s journey through the model. Psoriasis severity determines what type of clinician a person sees routinely and what treatments are prescribed. All patients are assumed to have a GP interaction at the start of care. Non-severe psoriasis patients have their care managed by a GP, whereas severe psoriasis patients then have their psoriasis managed by a dermatologist. A GP can refer a patient to a dermatologist for management (who may accept or reject the referral). A dermatologist can refer a patient to a GP for non-psoriasis issues while maintaining management of the patient’s psoriasis care.

At a consultation, the clinician first judges the patient’s psoriasis severity. If the GP diagnoses mild psoriasis, a topical therapy is prescribed216 and a future check-up is scheduled. If the GP diagnoses severe psoriasis, the patient is referred to a dermatologist to potentially receive a more complex therapy: phototherapy, non-biologic systemic therapy or biologic systemic therapy. 216

A GP might misdiagnose a person with mild psoriasis as having a more severe condition. The probability of this happening may differ between TAU and PSO WELL clinician training. Similarly, a GP could classify severe psoriasis patient as having mild disease and prescribe a topical therapy. We assume that dermatologists will always correctly refer a person with mild psoriasis back to primary care.

Psoriasis treatment

People with mild psoriasis can progress through the sequence of topical therapies derived from current clinical guidelines:

1. topical corticosteroid once daily plus vitamin D once daily (NICE guidelines CG153)61

2. if there is insufficient control from (1), change treatment to vitamin D twice daily (NICE guidelines CG153)61

3. if there is insufficient control from (2), change treatment to corticosteroid twice daily (NICE guidelines CG153)61

4. if a patient exhausts all available topical therapies and they still have mild psoriasis, re-prescribe corticosteroid twice daily.

People with severe psoriasis who enter the model will start treatment with phototherapy if they have previously received only topical treatment. If the patient has previously received a complex therapy, the clinician will initiate the appropriate treatment in the sequence according to current guidelines. 61

A clinician who follows the clinical guidelines will stop prescribing ultraviolet B (UVB) phototherapy if the patient has received 10 courses in total or a maximum of 175 sessions (assumes a typical course consists of 2.5 sessions per week for 7 weeks). 188 Any complex treatment for severe psoriasis will be stopped if it does not lead to a treatment response (defined as a reduction in PASI score of ≥ 75% or a 50–75% reduction in PASI plus a reduction of at least 5 points on the DLQI measure). This is the guideline decision rule for biologic therapies61 and is assumed to apply for all systemic treatments for severe psoriasis.

A clinician who does not act in accordance with the guidelines will, in the model, prescribe phototherapy if required. However, in this case there is no clear information about which type of phototherapy will be prescribed. Accordingly, when clinicians do not act in accordance with the guidelines, the specific type of phototherapy is selected at random by the model. If a patient with severe psoriasis has previously received phototherapy, but the clinician does not adhere to the relevant clinical guidelines, it is assumed that a systemic therapy is prescribed. In this case, the specific type of systemic therapy will be selected at random by the model.

The model does not incorporate potential contraindications or side effects associated with severe psoriasis treatments.

Scheduling consultations

A patient’s time until next consultation depends on their current psoriasis treatment. Routine GP appointments occur every 6 months to review progress with topical therapy. An appointment is scheduled for 2 weeks after referral from a dermatologist, a psoriasis flare or patient help-seeking following the PSO WELL intervention. Routine dermatologist consultations occur every 10–16 weeks depending on the specific complex treatment regimen.

Lifestyle and comorbidity interventions

A clinician with a good knowledge of psoriasis and awareness of related health issues may identify the need for a formal lifestyle or comorbidity intervention. The patient will agree to a lifestyle intervention only if it is a behaviour they are intent on changing. To manage an identified comorbidity, patients will receive the next available intervention in the relevant treatment sequence derived from clinical guidelines. 182–187 If the end of a particular comorbidity treatment sequence is reached and further treatment is required, patients will continue to receive the last intervention given. Patients who have behavioural interventions will receive each behavioural treatment sequence only once.

Treatment effects

Psoriasis treatment effects (see Table 27) were identified from the literature review. Only studies that included all regimens within a particular type of therapy were considered. If multiple studies reported different effect sizes for equivalent treatments, the largest was selected. For simplicity, treatment effects occur immediately in the model.

The studies in the systematic review used a range of measures to quantify treatment effects. The two main measures used were, first, the DLQI, which assesses the impact of psoriasis on aspects of a patient’s life (symptoms and feelings, daily activities, leisure, personal relationships) and whether or not the psoriasis treatment is a problem. Whether or not the psoriasis treatment is a problem is particularly relevant for topical therapies. The second measure is the PASI, which measures severity of psoriasis symptoms. The model required a single measure of treatment response. The DLQI was considered to better reflect patients’ concerns about quality of life. Accordingly, published algorithms were used to generate a common measure from the data collected in the systematic review. The following algorithm217 was used to estimate the impact of topical therapies on DLQI:

For systemic therapies, the identified treatment effect was the proportion of patients who experienced a ≥ PASI75 (75% improvement in PASI score) response. Based on this, the following relationship was used to estimate the impact of systemic treatments on DLQI score (Table 33). 219

Treatment effects of behaviour change and comorbidity interventions were obtained from the sources used in developing clinical guidelines. 130,182–187 These effects are realised for a finite period of time before relapsing to prior values, with time to relapse based on the available evidence. Details of the effects and time to relapse are given in Table 28.

Psoriasis flares

Psoriasis flares occur in the model, potentially causing mild psoriasis to progress to severe psoriasis. It is estimated that 35.5% of psoriasis patients experience a flare by 4 weeks;221 from this, the model predicts a time to flare. A flare will cause an increase in DLQI score (estimated as the reported difference in DLQI scores of patients with frequent flares and infrequent flares). 246 If the DLQI score exceeds 10, the patient is classified as having severe psoriasis. It is assumed that even if patients with severe psoriasis have their symptoms effectively managed they cannot return to having mild or moderate psoriasis.

Quality-adjusted life-years

Quality-adjusted life-years were used as the measure of total health benefit experienced by simulated patients over the course of their journey through the model. A QALY measures the length of survival and adjusts the life-year to take into account any ill health during that time. This is done by estimating utility weights that reflect a person’s preferences for different health states. Utility weights are typically measured on a scale from 0 to 1. A person in full health for 1 year will have a utility value of 1 and a QALY of 1. A person who is in less than full health for 1 year may have a utility value of 0.25 and a QALY of 0.25. Death is assigned a value of 0. Some of the measures used to generate utility values also allow a person with very poor health to have utility values of < 0. The model allows a simulated patient’s health status to vary as events occur and their attributes change. If the simulated patient has clinical characteristics (e.g. severe psoriasis symptoms or diabetes) or experiences events (e.g. side effects or stroke) that indicate that they are in less than full health, they are assigned a reduction in utility (also termed a utility decrement). The patient’s overall utility for 1 year is estimated by subtracting any utility decrements from 1 (see Table 29). For example, if a patient has PsA, the associated decrement (–0.093)230 is incurred, giving a utility score of 0.907. To ensure that the cumulative impact of several utility decrements does not overestimate the total negative impact of characteristics and events occurring at once, only the largest decrement is applied in full at each utility update.

Mortality

Time to mortality is predicted based on all-cause mortality. 192 The only cause of early mortality included in the model is fatality due to MI (15.1% in females, 10.6% in males)193 and stroke (12.5%). 194 These were thought to be the most relevant to the higher CVD risk seen in people with psoriasis1,30,66 and the focus of the PSO WELL intervention.

Costs and resource use

Various health-care resources are required throughout the simulation. The PSO WELL interventions require two types: the resources needed to deliver the intervention (direct provision costs) and the resources used to develop the interventions. These are treated as sunk costs for conceptualisation and planning. Our primary analysis included only the delivery costs of the PSO WELL interventions, implying that all necessary development had taken place. Total delivery costs, obtained directly from the IMPACT programme records, were £24 per patient for the information materials and £11,515 per group of 10 clinicians trained (see Table 30). For the primary analysis, we conservatively assumed that each clinician trained would manage only one patient who benefited from the clinician training intervention. This was varied in threshold analysis. The average group size for clinician training was 10 participants.

Other health-care costs were sourced from UK sources (see Tables 31 and 32). The basic cost per consultation is £38 for a GP visit and £93–109234 for a dermatologist appointment. Topical and non-biologic systemic therapies were costed using the average of available generic treatments in the British National Formulary238 weighted by the total number of each pack prescribed in England in 2014. 239 The cost of phototherapy (£92) is incurred per session. 234 Biologics included in the model are all self-administered, with the exception of infliximab, which incurs an additional administration cost as it is delivered by intravenous infusion. The costs of providing interventions to support lifestyle changes and to manage comorbidities were also included and are detailed in Table 31. Clinical guidelines were used to determine typical treatment regimens. 182–187

Costs and QALYs were discounted at an annual rate of 3.5% to account for social time preferences as recommended in the UK. 195