Cognitive behavioural therapy compared with standardised medical care for adults with dissociative non-epileptic seizures: the CODES RCT

Screening phase

Participants were initially identified in neurology/specialist epilepsy outpatient clinics. Neurologists would make and explain the DS diagnosis (see Neurologists’ delivery of standardised medical care) and give the patient a booklet with further information about their condition (downloadable from www.codestrial.org/information-booklets/4579871164; accessed 10 January 2021). A protocol was developed for this process, as follows. If patients met the eligibility criteria and were interested in participating, the patient consented to the neurologist forwarding their contact details to the CODES team. A research worker would then contact the participant, explain the trial in more depth and cover the material in the participant information sheet (see Report Supplementary Material 1). If the patient was interested in proceeding, the research worker confirmed eligibility and obtained informed consent. This was carried out mostly in person but occasionally by post. Basic demographic information and a very brief medical history pertaining to the patient’s condition were obtained, and the patient was instructed how to keep seizure diaries. Patients were referred to the designated liaison/neuropsychiatry service by the neurologist. During this intervening period, a research worker contacted the participant fortnightly by telephone, text message or e-mail (based on participants’ preferences) to obtain seizure diary data, comprising the number of seizures experienced per week and how many were severe.

Intervention phase

A liaison or neuropsychiatrist assessed the patient approximately 3 months after their initial neurology diagnosis. This appointment was used in part to undertake further screening for eligibility for the RCT. The appointments included a reiteration of diagnostic points, provision of a more in-depth booklet about DSs (downloadable from www.codestrial.org/information-booklets/4579871164; accessed 10 January 2021) and a detailed clinical psychiatric assessment. A guide was written for the psychiatrist to facilitate consistency and good communication. If the patient met the eligibility criteria and was interested in participating in the RCT, with the patient’s consent, the psychiatrist informed the research worker, who then contacted the participant to explain the RCT further. If the participant still wished to participate, the research worker met with them and covered the material in the participant information sheet for the RCT (see Report Supplementary Material 2). The research worker confirmed eligibility, obtained informed consent and completed the baseline assessments with the patient. Participants were then randomised to one of the two treatment conditions. Enrolled patients were asked to consent to the research team contacting a carer/informant who could provide their own perspective on the participant’s DS frequency at the follow-up stages. If they agreed to this, carers/informants received a participant information sheet (see Report Supplementary Material 3) and, if they agreed to participate, they completed a consent form either at a face-to-face meeting or by post.

Inclusion criteria (screening phase)

• Adults aged at least 18 years old who had experienced DSs within the previous 8-week period and whose diagnosis was corroborated by either video EEG or if not available, clinical consensus (see Neurologists’ delivery of standardised medical care).

• No recorded history of intellectual disability.

• Had the ability to keep seizure diaries and fill out questionnaires.

• Showed readiness to keep seizure diaries on a regular basis and attend a psychiatric assessment 3 months following receipt of their DS diagnosis in the study.

• Were able to provide informed consent.

Exclusion criteria (screening phase)

• A diagnosis of currently occurring epileptic seizures in addition to DSs (‘current’ is characterised as an epileptic seizure occurring in the prior year).

• Lacking the ability to independently maintain seizure records or fill out questionnaires.

• Met criteria for current alcohol or drug dependency in line with the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)94 criteria since this might, among other things, make attendance and symptom reporting less reliable.

• Insufficient fluency in English to complete questionnaires or later undergo CBT without an interpreter.

• Currently undergoing CBT for another diagnosis, if this intervention would still be ongoing by the time the psychiatry assessment takes place.

• Having previously had a CBT-based treatment for DSs at one of the trial participating centres.

Inclusion criteria for randomised controlled trial (intervention phase)

• Adults aged at least 18 years old who had been recruited into the study in the screening phase following their diagnosis.

• Indicated willingness to continue filling out seizure diaries and complete questionnaires.

• Had given the research team data about their seizure occurrence on a regular basis since receiving their diagnosis of DSs in the screening phase.

• Indicated that if they were allocated to CBT, they would be willing to attend weekly or biweekly therapy sessions.

• Both the participant and their clinician believed that randomisation was acceptable.

• Ability to provide written informed consent.

Exclusion criteria for randomised controlled trial (intervention phase)

• Was currently experiencing epileptic seizures in addition to DSs.

• DS had not been experienced during the 8-week period leading up to the psychiatry assessment.

• Had previously had a CBT-based intervention for DSs at one of the centres taking part in the RCT.

• Was currently undergoing a CBT intervention for another condition.

• Was experiencing active psychosis.

• Met criteria for current alcohol or drug dependence according to DSM-IV criteria since, in addition to making symptom recording and session attendance less reliable, it might be used to reduce anxiety and would reduce the impact of exposure during CBT, and have possible impact on patients’ memory for sessions.

• Evidence of current use of benzodiazepines that exceeded the equivalent dose of 10 mg of diazepam per day, for reasons similar to those for alcohol or drug dependence.

• Was at high risk of imminent self-harm, following the psychiatry assessment or according to the results of the structured psychiatric assessment [the Mini-International Neuropsychiatric Interview (M.I.N.I.)] administered by the research worker, subsequently followed up by a discussion with the relevant psychiatrist.

• Had received a diagnosis of factitious disorder.

Control intervention

Across the UK, medical and psychological care for DSs is variable, with different specialties contributing in specific ways. 71 To produce a broadly consistent treatment environment across the trial, key approaches were employed to create what we termed ‘standardised medical care’ for patients with DSs. This included providing briefing sessions to the clinicians (e.g. at site initiation visits), a detailed leaflet about how they might explain the diagnosis to patients, crib sheets containing the essential information that they should provide to patients during sessions (see, for example, Appendix 2 for the crib sheet for psychiatrists) and sets of frequently asked questions for clinicians providing SMC (see Report Supplementary Material 4–8 for the remainder of these materials). These are techniques that have been found to be acceptable in other studies. 93 It was intended that SMC would be provided by both neurologists and psychiatrists, and would start from the initial meeting with the neurologist at diagnosis. There was no mandatory number of SMC sessions; however, after the initial neurology and psychiatry assessment, we estimated that there would be up to two SMC sessions with the neurologist and three to four sessions with the psychiatrist. However, owing to local service procedures and clinical need, we could not be prescriptive about this.

One important component of SMC was the provision of information. We created two information booklets about DSs to be given at different stages of SMC to supplement the information given to patients by their medical clinicians. These booklets were devised by the clinical members of the project team, with input from SUs with DSs and a hospital information officer. The two booklets were:

1. Dissociative seizures factsheet (neurology). This was to be given to patients by their neurologists when the diagnosis of DSs was first communicated. It was also downloadable from www.codestrial.org/information-booklets/4579871164 (accessed 10 January 2021).

2. Dissociative seizures factsheet (psychiatry). This included content from the neurology leaflet but also provided more extensive information that could help provide further details relevant to psychiatric assessment and treatment. This was to be given to patients when they attended their psychiatric assessment. It has subsequently been made available at www.codestrial.org/information-booklets/4579871164 (accessed 10 January 2021).

When research workers contacted patients about potential participation in the study phases, they asked whether or not these factsheets had been provided; if not, they ensured that participants received the relevant factsheet.

Cognitive–behavioural therapy for dissociative seizures

Cognitive–behavioural interventions have generally now been shown to have positive benefits for a range of medically unexplained symptoms, as reported in a number of systematic reviews;78,95–99 however, these are not specifically studies of DSs and have not adopted specific models to underpin treatment of DSs. Although mechanisms of change have been studied in adults with medically unexplained symptoms,100 this work has not included adults with DSs. Recent conceptualisations of factors giving rise to DSs have, nonetheless, been developed and include an integrative cognitive model that seeks to explain DSs as automatic activations of a central representation of seizures (referred to as a seizure scaffold) in the context of dysfunction of inhibitory processing. 101 The seizure representation may be shaped by different factors that are relevant to seizures in the person’s life and yet further factors, such as chronic stress and arousal, may compromise adequate inhibitory processing. This and other conceptualisations of DSs (e.g. ‘panic without panic’43) offer the possibility of employing cognitive–behavioural interventions to address DS occurrence.

However, although acknowledging the richness of relatively recent models,101 the development of our CBT approach, which predates such models, stems from a single case study81 and was further refined in an open-label study82 and then in our pilot RCT. 1 A further description of the approach is given elsewhere. 83 A finding of increased symptoms of autonomic arousal as relevant to the concept of ‘panic without panic’43 is not in conflict with the approach taken in these studies.

Our cognitive–behavioural model incorporated the fear escape–avoidance model. 102,103 We considered DSs to represent a dissociative response to heightened arousal accompanying cognitive/emotional/physiological or environmental cues that may or may not be associated with previous/current distressing or life-threatening experiences. Alternatively, DSs may have occurred after events, such as panic attacks or syncope. All of these events may previously have led to unbearable feelings of distress and/or fear.

The treatment broadly consisted of engagement and rationale giving; helping the patient develop and use seizure control techniques; helping the person reduce avoidance behaviours via exposure; helping the person tackle maladaptive cognitions associated with seizure occurrence and facilitate emotional processing; dealing with trauma; and planning for relapse prevention. The key elements of the conceptualisation of the disorder and the treatment elements incorporated in the approach used here are illustrated in Figure 1.

FIGURE 1.

Model showing responses associated with DSs and the CBT techniques used to target these in our DS-specific CBT. ABC, antecedents, behaviours, consequences.

Cognitive–behavioural therapy was intended to be delivered in 12 sessions (plus one booster session) informed by a treatment manual. This number of sessions was based on our previous work1,82 that we were seeking to extend here; in our 2010 pilot RCT1 that provided preliminary evidence for efficacy, not all patients attended both booster sessions so, for easier implementation for the therapists, we limited this to one booster session. The DS-specific CBT was designed to help facilitate the patient to:

• develop an understanding of their seizure disorder

• develop an understanding of how cognitive, emotional, physiological and behavioural aspects of their DSs are related

• understand factors that led to the persistence of their DSs

• learn how to prevent DSs by interrupting behaviours, cognitions or physiological responses occurring before or at the start of their DSs

• improve their lifestyle by undertaking previously avoided activities

• address thought patterns and attributions about their disorder that act to maintain their DSs

• address and deal with previous traumas, poor mood, anxiety or reduced self-esteem, where relevant

• increase their levels of independence and help them to comprehend the contribution of significant others to their disorder.

The CBT sessions included usual CBT components, such as session agendas and planning and reviewing homework activities, but also required patients to complete seizure diaries and undertake activities designed to help with seizure control. Although treatment was manualised, there was room for flexibility, allowing individual formulations. Participants receiving CBT were given a booklet (the ‘Manual for Patients Attending CBT’) that contained written material to supplement the therapy sessions and included pages for making notes. The titles of the individual topics covered in this manual were Introduction to cognitive behaviour therapy and dissociative seizures; A guide for other people; Distraction and re-focusing techniques; Progressive muscle relaxation exercises; Breathing exercises; Graded exposure; Trauma in the context of dissociative seizures; Identifying negative automatic thoughts; Alternatives to negative thoughts; Preparing for the future; and Discharge plan.

Clinical and demographic information

We collected the following clinical and demographic information: date of birth, gender, ethnicity, living arrangements, marital status, dependants, attained qualifications, employment status, receipt of disability benefit, previous epilepsy diagnosis, prescription of AEDs, age at first seizure and whether or not the person had previously sought medical help for a mental health concern. Postcodes were collected from all participants to derive an Index of Multiple Deprivation (IMD), which provides a score that is indicative of the level of deprivation in a specific area. The separate databases for England (IMD), Scotland (Scottish IMD) and Wales (Welsh IMD) were published in different years and were based on slightly differing numbers of domains used to derive the scores. We used the versions in use at the time of the first recruitment into the study108–110 and, for consistency, we used the same versions throughout the study. We allocated IMD scores to quintiles ordered across the three databases so that the lowest quintile reflected the least deprivation.

At recruitment to the screening phase, participants rated how strongly they believed that they had been given the correct diagnosis (0 = not at all, 10 = extremely strongly). At recruitment into the intervention phase, self-report information was collected on other medical conditions with which participants were currently diagnosed, along with their treatment preference for CBT + SMC or SMC alone or whether they had no preference. In addition, expectation of the outcome of treatment was assessed by questions on how logical treatment seemed and how confident they were that the treatment would help them when considering CBT, being seen by a neurologist and being seen by a psychiatrist (see Appendix 4).

As well as demographic data, at recruitment into the intervention stage psychiatric comorbidities were assessed using a structured screening instrument (the Mini-International Neuropsychiatric Interview; M.I.N.I. v6.0)111 and a screening measure of personality [the Standardised Assessment of Personality Abbreviated Scale, Self-Report (SAPAS-SR)]. 112 We included a measure of personality in the light of accounts of personality disorder or personality clusters in people with DSs49,57 and to allow potential future examination of the effect of personality in moderating outcome.

The M.I.N.I. is a commonly used, structured, psychiatric diagnostic interview and is divided into modules that correspond to diagnostic categories including major depressive episode; suicidality; manic and hypomanic episodes; panic disorder; agoraphobia; social phobia; obsessive–compulsive disorder; PTSD; alcohol dependence/abuse; substance dependence/abuse; psychotic disorders and mood disorder with psychotic features; anorexia nervosa; bulimia nervosa; generalised anxiety disorder; and antisocial personality disorder. Research staff who administered the M.I.N.I. attended a 1-day training event that was led by one of the psychiatrists in the project team (Nick Medford). In addition to didactic teaching, role plays were held. Follow-up consultations with Nick Medford were arranged to address administration/scoring difficulties.

The eight-item SAPAS-SR poses questions about how the person sees themselves. ‘Yes’ responses are scored as 1 and ‘no’ responses are scored as 0, giving a final score between 0 and 8 with a cut-off score of 4 for the presence of personality disorder. It has high test–retest reliability; the cut-off score of 4 demonstrates a sensitivity of 0.83 and specificity of 0.8112 when identifying the presence of personality disorder. As we did not then undertake formal assessment of personality disorder for those with scores of ≥ 4, we interpret these scores as indicative of maladaptive personality traits rather than frank personality disorder.

Primary outcome measure

The primary outcome measure used in the trial was self-reported monthly DS frequency at 12 months post randomisation. This enabled us to include all participants’ outcomes irrespective of their improvement or otherwise during the trial. DSs were recorded by patients in a seizure diary that was collected by research workers every 2 weeks, using the format most acceptable to the participant (e.g. via paper diary, text or e-mail). In addition, participants were asked about seizure occurrence at baseline, 6 months and 12 months by a single-item measure for which they were asked to write down how many seizures they had experienced in the previous 4 weeks. The gold standard measure for the primary outcome was to calculate it as the sum of four seizure diary entries for weeks 49–52 (post randomisation). Where these data were not available, full details about constructing the primary outcome can be found in Appendix 5.

Secondary outcome measures

A number of measures were used to assess secondary outcomes. These are grouped together with measures of the same general psychological or other process in the following sections.

Health economics

These measures are described in Chapter 4.

Generalisability of the trial sample

To assess whether or not our trial sample of participants (n = 368) was similar to those who were eligible for the trial but did not go on to be randomised (n = 58), we compared key characteristics measured after consent to screening. Variables were chosen for this comparison if they might predict outcome in patients with DSs, that is to check that our randomised sample of 368 participants did not differ from the 58 participants in factors that might later be relevant to the outcome in the RCT. Factors that may predict outcome in patients with DSs more generally include symptom duration, receipt of social security benefits, having previous psychiatric diagnoses, employment status, educational achievement and gender (see Chapter 1). Although we tested a wider range of variables to encompass age, gender, social relationships, diagnostic method, predominant DS semiology, previous diagnosis of epilepsy, current prescription of AEDs, belief in diagnosis of DS and previous medical help-seeking for a mental health problem, to compare these subgroups more widely we did not compare the two subgroups on all of our demographic variables to avoid excessive testing. We employed Wilcoxon’s rank-sum test and Fisher’s exact tests for continuous and categorical variables, respectively. The variables incorporated in such comparisons are indicated in Chapter 3, Tables 6–8.

Descriptive statistics

In Chapter 3, descriptive statistics for all baseline, primary and secondary outcomes are reported by trial arm and overall for all of the time points at which data were collected (baseline, 6 months and 12 months); a formal analysis of a treatment difference was conducted for outcomes at 12 months only. Medians [interquartile range (IQR)] are used to describe count outcomes (seizure frequency or seizure freedom) owing to potential skewness, means (SD) are used to describe all other continuous outcomes, and frequencies (%) are used to report categorical and binary outcomes. All totals (n) are reported. One of the secondary outcomes (informants’ rating of patient’s seizures at 12 months post randomisation) could not be formally analysed because it was completed by only a small proportion (7.3%) of participants.

Predictors of loss to follow-up and the multivariate imputation by chained equations procedure

Prior to unblinding the trial statisticians, a binary variable was created to indicate whether or not participants provided complete 12-month follow-up data for the primary outcome (1 = provided primary outcome data, 0 = did not provide primary outcome data). This binary variable was sent to an independent King’s CTU-affiliated statistician along with the CODES therapy database; the independent statistician then created a second binary variable to indicate treatment compliance. As per the SAP, this was defined as attending at least nine sessions of CBT if allocated to the CBT + SMC arm (1 = nine or more CBT sessions attended, 0 = fewer than nine CBT sessions attended). This number of sessions was chosen to be consistent with our pilot RCT. 1 The independent statistician then ran a chi-squared test to assess whether or not treatment compliance within the intervention arm was predictive of missing 12-month primary outcome data. The chi-squared test confirmed this association to be statistically significant (p < 0.001): 94% (131/140) of participants in the intervention arm who were compliant with CBT provided primary outcome data at 12 months, compared with 54% (25/46) of those who were non-compliant. Therefore, as described in the TSC-agreed and published SAP,92 multivariate imputation via chained equations (MICE) was deemed appropriate and necessary for the main analysis to produce inferences valid under the detected missing-at-random (MAR) data-generating process. This meant that missing outcome variables at all time points were imputed instead of using the mixed-modelling approach originally suggested in the brief (pre SAP) analysis section of the protocol. 84 Multiple imputation (MI) consists of an imputation step and an analysis step. First, missing values in specified variables are multiply imputed. Second, an analysis model is fitted to each imputed data set and the analysis results are combined using Rubin’s rules. 131 This two-stage procedure provides inferences that are valid under a MAR missing data mechanism. Importantly, as defined by the imputation model, the observed variables are allowed to drive missingness at 12 months.

To inform the imputation model, logistic regression methods were used to detect which baseline variables were associated with missing follow-up at 12 months in each randomisation stratum (sites). Almost all baseline variables were included in this process, with just a few exceptions for the following reasons: (1) if measured only in small subgroups (IMD quintile Wales, type of dependant, type of carer and status of previous epilepsy diagnosis); (2) if small subgroups were merged to reduce the chance of perfect prediction (ethnicity was grouped into white, black or other; relationship status was dichotomised into married or living with partner vs. single or other; binary variables were used for ‘any current M.I.N.I. diagnosis’ and ‘any previous M.I.N.I. diagnosis’); or (3) if measured more than once (continuous score of SAPAS-SR was used, so the binary version was not included).

First, unadjusted logistic regression was implemented, with the binary follow-up variable as the outcome and the baseline measures were each tested separately. Seven variables were found to be significantly associated at the α < 0.05 level: modified PHQ-15 score (p = 0.002), at least one current M.I.N.I. diagnosis (p = 0.003), number of days seizure free in the last 6 months (p = 0.005), having a carer (p = 0.007), relationship status (p = 0.023), previously sought help for a mental health problem (p = 0.035) and SF-12v2 Mental Component Summary score (p = 0.047).

Second, manual forward stepwise regression was used, with a liberal inclusion threshold of α < 0.1. The process was as follows: a logistic regression on binary follow-up was run fixing psychiatrist site as a covariate because randomisation stratified on this; each of the seven variables above, found to be univariately associated with missing primary outcome data, was then included one by one in a step-wise fashion; all of the p-values were compared and the variable with the lowest p-value (if p < 0.1) was added. This process was repeated until no further variables could be added (i.e. p > 0.1).

The variables that were found to be associated with loss to follow-up (within sites) were number of somatic symptoms on the modified PHQ-15 – those with fewer symptoms were more likely to provide primary outcome data (17 vs. 19.7 median symptoms; p = 0.0030); carer – those with a carer were more likely than those without a carer to provide primary outcome data (91.3% vs. 80.8%; p = 0.0021); relationship status – those who were single were more likely than those married/living with partner to provide primary outcome data (89.6% vs. 81.0%; p = 0.0078); number of days seizure free in the last 6 months in the study – those with longer periods of time without a seizure were more likely to provide primary outcome data (9.5 vs. 7 median days; p = 0.0078); and previously sought help for a mental health problem – those who had never sought help were more likely than those who had to provide primary outcome data (90.6% vs. 82.2%; p = 0.037).

Once the baseline predictors of missingness had been established, the MICE procedure132 to impute missing values for each outcome could continue. We generated 100 imputed data sets for each outcome and combined the analysis results according to Rubin’s rules,131 where each imputed data set was analysed according to the corresponding analysis model.

The primary outcome imputation model included dummy variables for treatment compliance in the CBT + SMC arm and included the five baseline predictors of dropout, as above. All dummy variables were coded 0 and 1. The imputation model also included all variables of the analysis model because MI theory requires all variables of the analysis model to be included in the imputation model. Any previous (or ‘auxiliary’) measures of the primary outcome (e.g. baseline and 6 months) were included to ensure that the observed values of the same outcome measured at other time points could contribute to the prediction of missing values in the MICE procedure. Unless complete, all auxiliary variables were imputed. For the primary outcome, incomplete auxiliary variables were 6-month seizure frequency, longest period of time seizure free in the last 6 months and baseline modified PHQ-15 score.

The three count variables (baseline and 6-month seizure frequency, and days of seizure freedom) were log-transformed prior to the MI step (plus 1 to avoid logging zero). The reason for modelling the longest period of seizure freedom (days) in the last 6 months on the log-scale was because of its inverse relation to the primary outcome and, although it refers to a 6-month period, there is no maximum value owing to its self-reported nature. Baseline and 6-month seizure frequency were also log-transformed because the relationships between them and 12-month seizure frequency are multiplicative, translating into an additive effect of log-counts on the linear prediction scale. Predictive mean matching was used to impute both incomplete count variables because of zero-inflated distributions.

Continuous outcomes were imputed using linear regression. Psychiatry sites were included as random intercepts in the analysis model; therefore, we included dummy variables for fixed effects of psychiatry site in the imputation model. This was acceptable because a fixed-effects model is more general than a random-effects model. To avoid perfect prediction and overparameterisation, some small categories were merged within dummy variables (e.g. one site had only one participant and another had three participants).

Both the imputation and the analysis model assumed a negative binomial distribution because monthly seizure frequency (count) was very overdispersed (i.e. the variance was much greater than the mean). Otherwise, a Poisson model may have been appropriate. Count outcomes are constructed using the number of events (count) over an exposure period (time): for us, this corresponded to the number of seizures over number of days (7, 14, 21 or 28 days from weekly diaries).

For the primary outcome, the analysis model hence included trial arm and baseline monthly seizure frequency as the independent variables for the following reasons, respectively: to formally assess treatment arm difference and because baseline values of the outcome variable are known to be predictive of the post-randomisation outcome. Psychiatry site needed to be conditioned on because it was the randomisation stratification factor, and we chose to model the effects by site-varying random intercepts rather than by fixed effects. This approach allowed us to estimate the treatment effect in the population from which the trial sites are a sample, which provides better generalisability than if we had used a fixed-effects approach in this multicentre trial.

We also considered the random effects for SMC doctors and CBT therapists (in the CBT + SMC arm). SMC doctors (generally psychiatrists) were nested within psychiatry sites, which meant that there was commonly only one or two doctors per site. Therefore, SMC doctor effects were not distinguishable from site effects and were not included. Therapist effects were assessed empirically and, again, no evidence for their existence was found. This was evaluated using complete cases (CCs): a likelihood ratio (LR) test was run for each outcome that compared the log-likelihoods of the models including therapist effects (in the intervention arm) with the models that did not include therapist effects. None of the LR tests was significant at the α < 0.1 level and half of the tests reported a p-value equal to 1.0000 [LR χ²(1) = 0]. For some outcomes, the LR test could not even be run [i.e. p-value = not applicable (N/A)] because the model with the therapist effects could not converge; our interpretation here was that therapist effects were not helping to explain the variability in the outcome. Most therapists were also nested within psychiatry sites (in the CBT + SMC arm only), so this may have influenced the LR tests.

Given the random effects for site, the primary outcome analysis model was a mixed-effects negative binomial model that was fitted using the Stata command ‘menbreg’. The dependent variable (monthly seizure frequency at 12 months) was constructed as a count and exposure period, and the two independent variables were constructed as one dummy variable (CBT + SMC vs. SMC) and one log-transformed continuous variable (baseline monthly seizure frequency). As explained above, random intercepts for psychiatry site were included to account for common site experiences. The inbuilt Stata option ‘cmdok’ was used to allow estimation of the mixed-effects negative binomial model with multiply imputed data. Incidence rate ratios (IRRs) were reported.

All secondary outcomes were analysed following the same approach: using MI to allow for the detected MAR process. For all outcomes, the imputation model included the five predictors of missingness at 12 months; a treatment compliance dummy variable in the CBT + SMC arm; a trial arm dummy variable; psychiatry site dummy variables, and any previous measures of the same variable. A few secondary outcomes were strongly correlated and had different missing patterns, so were imputed simultaneously (i.e. imputations for both outcome variables were generated from a single, larger, imputation model). This was the case for seizure severity and bothersomeness, and self-report and doctor-rated CGI scale score change.

For the one other count outcome, the longest period of time (consecutive days) seizure free in the last 6 months, a mixed-effects negative binomial model (with site-varying random intercepts) was also used. For the secondary outcomes that were treated as continuous, mixed-effects linear regression (with site-varying random intercepts) was used: seizure severity, seizure bothersomeness, Physical Component Summary score (SF-12v2), Mental Component Summary score (SF-12v2), health today (EQ-5D-5L VAS), impact on functioning (WSAS), anxiety (GAD-7), depression (PHQ-9), distress (CORE-10), other somatic symptoms (modified PHQ-15), CGI scale change self-report, CGI scale change doctor rated and satisfaction with treatment. Finally, mixed-effects logistic regression (with site-varying random intercepts) was used for the secondary binary outcomes: seizure freedom in the last 3 months of the study and > 50% reduction in seizure frequency.

Owing to the different scales of the secondary outcomes, all estimated treatment effects were standardised to aid comparisons. For outcomes with a baseline measure, this was calculated by dividing the estimated difference between arms on the original scale by the baseline SD or by the pooled SD of the outcome. None of the p-values was adjusted for multiple testing; therefore, interpretation of statistically significant (p < 0.05) secondary outcomes should be undertaken with caution.

Checking of regression assumptions and multiply imputed data

As written in the SAP,92 regression assumptions were checked for all 17 outcomes to ensure that imputed values looked sensible. First, to check the multiply imputed data, the 100 MI data sets were saved for each outcome and summary statistics were used to compare the imputed values with the observed data; namely, the minimum, mean and maximum values of m = 1 to m = 100 were compared with these summary statistics from the observed sample to check that they were reasonably similar.

Second, a subsample of MI data sets were used to perform MI diagnostics of all count or continuous outcomes: this consisted of every 10th MI data set (m = 10, 20, . . ., 90, 100). Kernel density estimates were plotted for these 10 MI data sets against the observed and completed data sets. This meant that the distributions of the observed, imputed and completed values could be compared graphically. Kernel density plots are similar to histograms in that the y-axis depicts the density function but, compared with a histogram that uses bins (or bars), they plot the distribution using smooth lines [similar to a Locally Weighted Scatterplot Smoothing (LOWESS) curve].

For the primary outcome and other count variables, the reason for using negative binomial models instead of Poisson models was the violation of the assumption that the variance is equal to the mean; for example, for our primary outcome data the mean was 37.6 and the variance was 8611.7. It was not possible to test whether or not each seizure was independent of each other, but it was reasonable to make this assumption.

For all continuous variables, the following regression assumptions were checked:

1. Homoscedasticity of residuals – the constant variance of residuals across all data points on regression line. This is checked by inspecting scatterplots of fitted values versus standardised residuals.

2. Linearity of independent variables – the linear relationship between dependent and independent variables. For those outcomes with a corresponding baseline variable that was adjusted for in the model, this was checked by inspecting scatterplots using Stata’s inbuilt ‘avplot’ command.

3. Normality of residuals – error terms display a normal distribution. This was checked by inspecting box plots of the residuals.

Agreement between auxiliary measures and sensitivity analysis

Given that there were two different methods of recording the primary outcome measure, the diary (which allowed a 1- to 4-week exposure period) and the single measure (which specifically asked about a 4-week period), we tested whether or not there was sufficient agreement between the two variable types at the primary outcome time point (12 months post randomisation). Both versions were log-transformed (after adding 1 to avoid logging 0) so that the agreement could be measured using log-frequencies (i.e. on the log-scale). An ICC was calculated by treating the log-transformed seizure diary (pro rata 4-weekly seizure frequency) as one ‘rater’ and the log-transformed single measure (seizure freedom question) as a second ‘rater’. The Stata command ‘kappaetc’ was used to measure inter-rater reliability.

In addition to calculating the ICC, a sensitivity analysis was run to check whether the size, direction or level of significance of treatment effect was ‘sensitive’ to how the primary outcome was constructed. To do this, the gold standard measure was treated as the only accepted measure for monthly seizure frequency: all participants who had not provided seizure diary data at 12 months but had provided the single measure were treated as missing follow-up. The same MICE procedure was then followed.

Complete-case and complier-average causal effect analysis

As well as the main ITT analysis and the sensitivity analysis, each outcome was analysed with just the CCs, that is without any MI and without adjusting for baseline predictors of missingness. The same analyses models as for the MI were used. This was carried out as a ‘reality check’ to see whether or not the results differed in comparison with those derived from the imputed data; if a difference was found in terms of the statistical significance, direction of effect or size of effect, then it may highlight the bias that has been corrected for by the MICE procedure.

Similarly, to assess the efficacy of the CODES intervention (CBT) in the presence of non-compliance, we ran a complier-average causal effect (CACE) analysis for the primary outcome, as stipulated by the SAP. This analysis models the effect of receiving the intervention (at least nine sessions of CBT) and aims to estimate the efficacy of the trial, rather than the effectiveness, which is estimated by ITT analysis. Treatment receipt regardless of trial arm allocation was used as the explanatory variable of interest. The same imputation step was used as for the main ITT analysis, but this time the log-transformed primary outcome was modelled and an instrumental-variables regression was run using the two-stage least squares estimator.

Introduction

Participants were initially identified and then recruited from neurology/specialist epilepsy services in 27 NHS trusts. Of the 698 initially recruited patients, 568 patients attended psychiatry assessments at a neuropsychiatry/liaison psychiatry service in 1 of 17 NHS trusts around 3 months later, and 368 patients were subsequently randomised into the trial. CBT was delivered in 1 of 18 NHS trusts linked to the psychiatry services for the purpose of the study (see Chapter 2 for the list of sites contributing to the different aspects of the study).

Participant flow through the study

Recruitment

Recruitment into the screening phase of the study (i.e. from neurology/specialist epilepsy clinics) ran from October 2014 to February 2017. Randomisation from this pool of patients into the RCT ran from January 2015 to May 2017. As a result of initial concerns regarding loss to follow-up rates, these timelines included an extension of recruitment to both phases of the study by 3 months.

Neurologists recruiting patients for the study were asked to provide information on all patients with DSs attending their clinics. Although this may be an underestimate of the number of DS patients attending neurology/epilepsy clinics, 901 patients were identified whose eligibility details were provided [the Consolidated Standards of Reporting Trials (CONSORT) flow diagram for CODES, part 1; Figure 2]. Recruitment in the liaison/neuropsychiatry settings involved psychiatrists further assessing patients’ eligibility for the RCT. This was undertaken for the 568 patients attending their psychiatric assessment session.

FIGURE 2.

The CONSORT flow diagram: part 1.

Enrolment

Of the 901 patients identified in neurology settings, 698 were consented to the screening phase of the study (i.e. consented for psychiatric assessment) (see Figure 2). This was the upper limit of the extended recruitment target for this phase of the study and was possible because of a surge in recruitment in the last month of recruitment. In addition to not meeting eligibility criteria (n = 56), as seen in Figure 2, the majority (n = 85) of the 147 patients not enrolled did not want to take part in the study, and 61 patients did not respond to attempts to make contact.

Recruitment into the screening stage occurred across all sites in parallel (Table 3). As anticipated, sites varied considerably in the number of patients recruited; this was, in part, influenced by the number of clinics held at each site.

The 27 neurology/specialist epilepsy sites (NHS trusts) have been listed in order of the date that the first participant consented rather than by name to help anonymise data from participants from sites with small numbers of participants (fewer than participants). In alphabetical order, the neurology/specialist epilepsy services were at the following NHS trusts: Barts Health NHS Trust; Birmingham and Solihull Mental Health NHS Foundation Trust; Brighton and Sussex University Hospitals NHS Trust; Cambridge University Hospitals NHS Foundation Trust; Cardiff and Vale University Health Board; Chesterfield Royal Hospital NHS Foundation Trust; Croydon Health Services NHS Trust; Dartford and Gravesham NHS Trust; East Kent Hospitals University NHS Foundation Trust; East Sussex Healthcare NHS Trust; Guy’s and St Thomas’ NHS Foundation Trust; Imperial College Healthcare NHS Trust; King’s College Hospital NHS Foundation Trust; Leeds Teaching Hospitals NHS Trust; Lewisham and Greenwich NHS Trust; Maidstone and Tunbridge Wells NHS Trust; Medway NHS Foundation Trust; Newcastle upon Tyne Hospitals NHS Foundation Trust; NHS Lothian; Royal Berkshire NHS Foundation Trust; Royal Free London NHS Foundation Trust; Sheffield Teaching Hospitals NHS Foundation Trust; St George’s University Hospitals NHS Foundation Trust; University College London Hospitals NHS Foundation Trust; University Hospital Southampton NHS Foundation Trust; University Hospitals Birmingham NHS Foundation Trust; and Western Sussex Hospitals NHS Foundation Trust.

Once patients consented to the screening phase of the study in the neurology/specialist epilepsy clinics, research workers maintained fortnightly contact with the patients wherever possible to obtain seizure diary data and remind patients about their psychiatry assessment. A total of 130 people did not attend their psychiatric assessment appointment, so could not be considered for the RCT. Of the 568 people who were assessed by a psychiatrist, 426 were considered eligible for the RCT and 368 were consented to the RCT (Figure 3). The number consented to the RCT slightly exceeded our extended target (n = 356); however, we could not ethically refrain from recruiting some rather than other eligible participants. Our recruitment graphs for both phases of the study are shown in Figure 4.

FIGURE 3.

The CONSORT flow diagram: part 2. a, One participant allocated to SMC received CODES CBT by mistake. b, Treatment receipt of CBT (compliance) was defined as receiving at least nine sessions.

FIGURE 4.

Cumulative target vs. achieved recruitment and randomisation. Note that recruitment targets were subsequently revised from November 2016 to a total of 698 in phase 1 with an extended recruitment period to February 2017 and a revised RCT total of 356 (phase 2), similarly extending recruitment by 3 months from February to May 2017.

Enrolment into the RCT took place across all sites in parallel and was completed at the end of May 2017 (Table 4).

The 17 psychiatry sites (NHS trusts) have been listed in order of the date that the first participant was randomised rather than by name to help anonymise data of participants from sites with small numbers of participants (fewer than five participants). In alphabetical order, the psychiatry sites are as follows: Berkshire Healthcare NHS Foundation Trust; Birmingham and Solihull Mental Health Foundation Trust; Cambridgeshire and Peterborough NHS Foundation Trust; Cardiff and Vale University Health Board; Derbyshire Healthcare NHS Foundation Trust; East London NHS Foundation Trust; Kent and Medway NHS and Social Care Partnership Trust; Leeds and York Partnership NHS Foundation Trust; NHS Lothian; Northumberland Tyne and Wear NHS Foundation Trust; Sheffield Health and Social Care NHS Foundation Trust; South London and Maudsley NHS Foundation Trust; Southwest London and St George’s Mental Health NHS Trust; Sussex Partnership NHS Foundation Trust; University College London Hospitals NHS Foundation Trust; University Hospital Southampton NHS Foundation Trust; and West London NHS Trust.

Unblinding

Table 5 reports the data from the study research worker (RW) treatment guess forms to assess whether or not blinding of treatment allocation was successful. Forms were completed at 12 months post randomisation or at participant withdrawal. Forms were available for 342 out of 368 participants. The data in Table 5 imply that blinding was generally successful, with only eight randomised participants being reported to have unblinded RW(s) in some way. Most guesses were completely at random.

To formally test whether or not trial arm predicted type of guess, ‘Strongly think he/she was allocated CBT’ was combined with ‘Think he/she was allocated CBT’, and ‘Strongly think he/she was allocated SMC only’ was combined with ‘Think he/she was allocated SMC only’. A Fisher’s exact test was then run to compare guesses between trial arms and indicated that there was an association between trial arm and RW guess (p < 0.001). For participants allocated to the SMC-alone arm, RWs guessed that the majority were allocated to SMC alone (81.3%), whereas this percentage fell to 49.7% for the CBT + SMC arm. This implies that the RWs may have been able to guess when participants were having CBT to some extent.

Baseline characteristics

We gained clinical and other demographic data at initial recruitment of the 698 participants into the study from neurology clinics, and again prior to randomisation into the RCT for 368 participants. Some data were obtained only once, whereas some data were gained at both time points. Tables 6–8 present clinical and demographic data for those consented to the screening phase (n = 698), those subsequently eligible for the RCT (having attended the psychiatric assessment, n = 426) and those randomly allocated to the two treatment arms (n = 368). Although there were some fluctuations in terms of values across study stages, Tables 6–8 indicate that those randomised into the RCT were very similar to the initially recruited sample (n = 698), and characteristics were similar for those randomised to the two treatment arms.

As described in Chapter 2, key characteristics that were measured after consent to the screening phase were compared using statistical tests between participants in our trial sample (n = 368) and participants who were eligible for the trial but did not go on to be randomised (n = 58). None of the comparisons was significant at the p-value < 0.05 level.

Our participants had a median age in the mid-30s, and the proportion of women in the randomised subsample, at around three-quarters, was in line with expectations and was very similar to the proportion of women in the overall pool of 698 participants from which the subsample was drawn (see Table 6). We have described the total sample of 698 participants in detail elsewhere133 and reported that, although overall there were more women than men, women were more likely than men to develop DSs at a younger age, whereas men were equally likely to develop DSs across the age span.

The IMD scores, reflecting levels of deprivation based on participants’ postcodes, indicated that, when examining quintiles, > 50% of participants from England, Scotland and Wales fell in the two quintiles indicating areas of the highest levels of deprivation.

The majority of the participants were white, and > 50% had attained secondary-level education or vocational qualifications. Across all subgroups, approximately two-thirds of participants were not employed or in education. We defined ‘unemployed and not in education’ as including those who were unemployed, employed full-time/part-time but off sick, students whose studies were interrupted because of illness, retired owing to age or ill-health, or househusbands/housewives. Those categorised as being employed (approximately one-third of cases) were employed full- or part-time (and working), students or self-employed. The clear majority of those aged < 65 years and not working were receiving disability benefits, in contrast to those of working age and working. Most participants lived with other people, although roughly equal proportions were either married/cohabiting or single. Most (> 60%) reported having no dependants; around one-third (> 35%) reported having a carer, with this most often being a partner.

For the most part, at each stage of the study just over 50% of the patients had received their DS diagnosis based on video EEG (see Table 7). The median age at onset of DSs in our initial sample and those subsequently randomised was late 20s (not all participants were able to say when their DSs had started). The median duration of experiencing DSs until the delivery of the diagnosis in this study was mostly 3 years. Around two-thirds of patients were reported to have predominantly hyperkinetic DSs. The percentage of patients self-reporting a previous diagnosis of epilepsy was generally slightly higher than the percentage of patients thought by clinicians to have had a previous epilepsy diagnosis. Clinicians indicated that an appreciable number of patients were thought to have been previously misdiagnosed with epilepsy, and indicated that for between one-quarter and half of those patients it was not possible to verify the previous diagnosis. Although around one-third (30.3%) of the initially recruited (n = 698) sample self-reported that they were taking AEDs (30.9%), this percentage was nearer 20% in those entering the RCT.

The majority of patients indicated a strong belief in their diagnosis (scores between 8 and 10), with a median score of 8 (out of a maximum of 10) across all subgroups. A negligible percentage of participants had previously received CBT for their DSs. Around two-thirds had previously sought help for a mental health problem and > 70% of those entering the RCT reported suffering from another medical problem.

Psychological comorbidities and other characteristics measured pre randomisation

Healthcare practitioners involved in delivering CODES

A wide range of neurologists, psychiatrists and therapists participated in the CODES trial; SMC was delivered via local permutations of neurology and psychiatry appointments. Of the 63 CBT therapists who were trained to deliver CODES CBT, 39 were subsequently allocated patients. Reasons for non-allocation of patients to trained therapists included therapists changing job between training and patient randomisation, therapists’ change of job role at the site and parental leave.

The demographics and clinical experience of the three types of HCPs are summarised in Appendix 6.

Numbers of sessions attended for each intervention

Table 11 presents the number of SMC appointments that participants were offered and received across both trial arms. SMC consisted of neurology and/or psychiatry appointments. All participants allocated to CBT + SMC were offered up to 13 CBT sessions, whereas the number of SMC appointments that were offered varied.

Table 11 indicates that, as anticipated, participants in both trial arms were offered more psychiatry SMC appointments than neurology SMC appointments. The median number of neurology and psychiatry SMC appointments attended was one and three, respectively, in line with expectations. There were some participants who were offered and attended > 10 SMC appointments because of clinical need; however, overall, the number of SMC appointments offered and attended was similar across trial arms.

The median time between randomisation and the first CBT session was 38.5 days (IQR 26–59 days). Over half (55.9%) of participants who were randomised to receive CBT attended all 13 sessions; only eight (4.3%) participants did not attend any, and the median number of attended CBT sessions was the full course of 13. Three-quarters (75.3%) of participants attended at least nine CBT sessions, which meant that the majority were defined as treatment compliant. One participant received three extra sessions beyond the 13 offered in the trial design because they were considered to be at high risk by their therapist. For this participant, there was also no interval between session 12 and the booster session. Figure 5 illustrates the distribution of the number of CBT sessions attended in the CBT + SMC arm (n = 186). One participant who was randomised to the SMC arm was mistakenly offered the CODES CBT and subsequently attended all 13 sessions.

FIGURE 5.

Number of CBT sessions attended in the CBT + SMC arm.

Over 700 CBT sessions were scheduled and missed; around 30–40 participants missed different sessions. Although most missed appointments were cancelled in advance, over one-quarter (28.4%) were non-attendances without notification. Of 410 appointments cancelled by patients, feeling unwell (43%) and other (work and family) commitments (22%) were the most common reasons for participant non-attendance; however, nearly 12% of cancelled sessions were because of patients reporting being unable to travel alone and 7% were cancelled because of seizures. Nearly all sessions (95.8%) were attended face to face, and most participants (78.6%) attended alone. Only 10 (5.4%) participants of the total allocated to the CBT + SMC arm informed their therapist that they wished to withdraw from therapy, which included three participants who had not attended any sessions and seven participants who had attended more than one CBT session (range 2–12 sessions).

For the second CBT session onwards, therapists were asked to complete a series of questions about participant adherence to treatment. Of these responses (n = 1725), 84.4% were reported to have implemented therapy techniques well from the previous session (moderately well, very well or completely) and 80.9% were perceived by the therapists to have completed at least half of their homework. The mean duration of the therapy sessions (therapist reported) was 60.5 minutes; we did not collect data concerning the time of day that the sessions were held.

Information collected on participants who experienced a seizure during therapy indicated that only 42 (2.2%) of the attended CBT sessions were significantly disrupted by a person having a seizure; this corresponded to 18 participants in total. Most of these seizures were between 3 minutes and 10 minutes in duration, and only four resulted in the participant suffering physical injury.

Therapy protocol deviations

We observed only two deviations from the therapy protocol, one in each treatment arm. In the SMC arm, one participant received CODES CBT in error. In the CBT + SMC arm, one participant received more than 13 sessions of CBT; the participant had an additional three sessions after the standard 12th session and booster session because of clinical need.

Treatment fidelity

As described in Chapter 2, two independent CBT therapists undertook blind ratings of one session from each therapist, with half of the ratings being undertaken for session 3 and half for session 7. In total, 18 tapes for each session were rated using the rating scale described in Chapter 2, Evaluation of treatment fidelity (see Appendix 3). As indicated, scores were converted to standardised scores out of 100. For the single-item subscales (i.e. Overall Therapist Adherence, Therapeutic Alliance and Overall CBT Delivery), scores were calculated by dividing the score by 7 and then multiplying by 100. For the specific DS skills scale (i.e. DS skills items 3–6 in Appendix 3), scores were totalled and divided by [7 × number of relevant (yes-rated) items scored] × 100 to generate standardised scores. The median values are shown in Table 12.

Overall, the median ratings for the therapy being delivered in accordance with the CODES therapy manual, the therapeutic alliance and whether or not the therapy being delivered was CBT all fell in the upper end of the scales (between considerably and extensively for adherence to the manual; just below excellent for therapeutic alliance; and between considerably and extensively for delivery of CBT). More mid-range median scores were obtained for DS-specific techniques, but it was possible that, although DS control techniques were expected to take place in session 3, by session 7 therapists may have adopted some flexibility in session content (see Wilkinson et al. 135) and teaching seizure control techniques may have become less relevant by then.

Retention of participants

As seen in the CONSORT flow diagram (see Figure 3), we obtained primary outcome data for 313 (85%) of the randomised participants. The follow-up rate did not differ in the two arms [84% (156/186) in the CBT + SMC arm; 86% (157/182) in the SMC-alone arm]. Fifty-five participants withdrew or were lost to follow-up (CBT + SMC, n = 30; SMC alone, n = 25) but in no case was withdrawal accounted for by death or other AEs (Table 13). In only one instance (in the SMC arm) was the withdrawal decision initiated by the clinician rather than the participant, but no further information was provided for this withdrawal.

Primary outcome: seizure frequency

Our primary outcome measure for the RCT was participants’ monthly DS frequency at 12 months post randomisation. As described in Chapter 2, these data were also collected at baseline and 6 months post randomisation, and these were used as auxiliary variables in the MICE procedure. The outcome was defined as DS occurrence over the previous 4 weeks and data were collected using seizure diaries. Participants were also asked how many seizures they had experienced in the past 4 weeks as a single-item measure to impute diary seizure frequency for those who had not provided it, and to help estimate the reliability of the diary data.

Both trial arms appeared to show a decrease in monthly DSs from baseline to 12 months post randomisation. An illustration of the observed change over time by trial arm is shown in Figure 6a. Geometric means have been used because they are arithmetically similar to the median and allow 95% CIs to be constructed. The between-group comparison of monthly DS frequency did not reach statistical significance at 12 months (IRR = 0.78, 95% CI 0.56 to 1.09; p = 0.144), although the inferential statistics estimated a 22% advantage for the CBT + SMC trial arm (see Table 16). When comparing the MI and CC results (see Table 16), the size, direction and significance of the treatment effect are very similar (IRR = 0.80, 95% CI 0.58 to 1.09; p = 0.156); indeed, all CC models in Table 16 lead to the same substantive conclusions as the ITT analysis models.

FIGURE 6.

Plot over time of the primary outcome and a secondary seizure experience outcome. (a) Monthly seizure frequency; and (b) longest period of seizure freedom in the last 6 months. T0, baseline; T1, 6-month follow-up; T2, 12-month follow-up. Adapted with permission from Goldstein et al. 134 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Adapted with permission from Goldstein et al. 134 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Furthermore, the CACE effect estimate of the primary outcome, as described in Chapter 2, leads to the same effect size as the ITT analysis (IRR = 0.78), with a similar p-value (p = 0.217). This shows that the ‘efficacy’ estimate is the same as the ‘effectiveness’ estimate. Finally, the ICC between the two methods of collecting the primary outcome (seizure diary and single measure) was 0.95, which implies very high agreement; the results of a second sensitivity analysis for the primary outcome (for which the seizure diary was treated as the only acceptable measure, resulting in 21 participants being treated as having missing outcome data) yielded a similar result again (IRR = 0.74, 95% CI 0.53 to 1.04; p = 0.086). We can, therefore, conclude that the primary outcome analysis was not sensitive to how we constructed monthly seizure frequency.

Secondary outcomes

Health-related quality of life

For neither the SF-12v2 Physical Component Summary score nor the Mental Component Summary score was the between-group difference at 12 months statistically significant. On the EQ-5D-5L VAS at the 12-month follow-up, the CBT + SMC arm reported higher overall health scores than the SMC-alone arm (standardised treatment effect 0.27, 95% CI 0.06 to 0.47; p = 0.010). An illustration of the observed difference between arms over time is shown in Figure 7a.

FIGURE 7.

Plot over time of HRQoL and psychosocial functioning secondary outcomes by trial arm. (a) EQ-5D-5L VAS; and (b) WSAS. T0, baseline; T1, 6-month follow-up; T2, 12-month follow-up; EQ-5D-5L VAS: 0 = worst health, 100 = best health; WSAS: 0 = minimum impact, 40 = maximum impact with higher scores reflecting worse functioning. Adapted with permission from Goldstein et al. 134 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Adapted with permission from Goldstein et al. 134 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Effect sizes

To be able to compare sizes of estimated treatment effects between outcomes on different scales, we standardised all effect sizes for continuous outcomes (standardised treatment effects are shown in Table 16). Where the outcome was measured at baseline, we divided the estimated difference between arms by the SD of the baseline score; for other measures, the difference was divided by the pooled SD of the outcome.

Figure 8 displays the standardised treatment effects with 95% CIs for all continuous secondary outcomes. For outcomes in Table 16 that are interpreted as showing better outcomes from the CBT + SMC arm if the standardised treatment effect is negative (< 0), the estimated treatment effect has been reversed (multiplied by –1) in Figure 8 so that a positive standardised effect size (> 0) is in favour of CBT + SMC compared with SMC. Figure 8 shows that the estimated treatment effects for all 13 outcomes are positive (above the dashed line at d = 0), which illustrates a trend that favours the CBT + SMC arm.

FIGURE 8.

Forest plot of standardised treatment effects and 95% CIs of CBT + SMC compared with SMC alone for all outcomes analysed on a continuous scale (standardised treatment effect of > 0 favours CBT + SMC). Adapted with permission from Goldstein et al. 134 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Adapted with permission from Goldstein et al. 134 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

All p-values in Table 16 are prior to any adjustment for multiple testing, which is consistent with the formally agreed and published SAP. However, if conservative adjustments were to be made to the level of significance (α = 0.05), for example a Bonferroni adjustment (α/17), five of the secondary outcomes still strongly suggest that offering CBT plus SMC provided significant benefit compared with offering SMC only (i.e. at p < 0.003). Specifically, participants in the CBT + SMC arm experienced larger numbers of consecutive seizure-free days, better psychosocial functioning, better self-reported and doctor-rated CGI and greater treatment satisfaction than those in the SMC-alone arm.

Adverse events and other indices of harm

In total, 176 AEs were reported during the RCT by 110 participants (Table 17). In the CBT + SMC arm, 97 AEs were reported by 57 participants, and in the SMC-alone arm 79 AEs were reported by 53 participants; the incidence of AEs was, therefore, similar across trial arms. In the RCT, 49 out of 368 (13.3%) randomised participants reported SAEs: 31 SAEs were reported by 25 participants in the CBT + SMC arm and 27 SAEs were reported by 24 participants in the SMC-alone arm. None of the AEs or SAEs recorded in the RCT in the CBT + SMC arm was deemed likely to be related to the study intervention [likelihood of relatedness: remote, n = 33 (34%); none, n = 64 (66%)]. An example of a SAE for which the relatedness was judged as ‘remote’ was sickness and inability to eat or drink anything for several days after a seizure and resulting in the patient's admission to hospital. No AEs were considered to be SARs.

Only one patient reported a decrease of 20 points on the SF-12v2 Physical Component Summary score between baseline and both the 6-month and the 12-month follow-up assessments (i.e. a change in scores of 2 SDs).

Regarding self-reported scores of ‘much worse’ or ‘very much worse’ on the participant-rated CGI scale change score at the end of the study, 25 (13.7%) participants in the SMC-alone arm and 13 (6.9%) in the CBT + SMC arm self-reported being ‘much worse’ or ‘very much worse’ on the CGI scale at the 12-month follow-up.

Summary of results

In the CODES trial, participants with DSs who met other eligibility criteria were first recruited to a screening phase; eligible and consenting patients were subsequently randomised to receive SMC alone or CBT + SMC, with a 12-month post-randomisation follow-up. We randomised 368 people in total: 182 to the SMC-alone arm and 186 to the CBT + SMC arm. At baseline, the median 4-weekly DS frequency overall was 15 DSs (IQR 4–47 DSs), indicating often frequent DSs. There was also evidence of a high level of comorbidity and functional impairment. The study demonstrates the feasibility of delivering a CBT intervention for this group of patients across multiple centres and by therapists of different levels of seniority and experience. Patients demonstrated a high rate of belief in the diagnosis and the treatment offered. Our SMC, given the guidance we provided to clinicians in terms of its content and delivery and the information booklets provided to patients, might be suitably considered to be standardised and specialist medical care rather than standardised medical care.

The characteristics of the trial arms were generally well balanced at baseline. Our overall primary outcome follow-up rate at 12 months was 85% (313/368), balanced by treatment arm. Our analysis adopted the ITT approach with MI and followed our published SAP,92 whereby we evaluated outcomes at 12 months post randomisation only.

Regarding our primary outcome, DS frequency at 12 months post randomisation, the inferential statistics (using the MICE procedure and statistical modelling) estimated a 22% advantage for the CBT + SMC arm; however, this was not statistically significant (p = 0.144). Raw data suggested that both trial arms appeared to show an overall reduction in DS occurrence at 12 months.

All secondary outcomes for which analysis was possible were in the same direction and suggested an overall benefit of CBT + SMC over SMC alone. In 9 of the 16 comparisons this was accompanied by statistical significance (p < 0.05) and, of these, five outcomes were significantly better in the CBT + SMC arm at a p-value of ≤ 0.001. Thus, compared with the SMC-alone arm, the participants in the CBT + SMC arm were found to be less functionally impaired by their DSs, as measured by the WSAS (p < 0.001); they reported longer periods of DS-free days in the last 6 months of the study (p = 0.001), with inferential statistics estimating a 64% advantage in the CBT + SMC arm; their CGI scale scores were better when self-rated (p = 0.001) and clinician rated (p < 0.001); and their satisfaction with treatment was greater (p < 0.001). Effect sizes were moderate and, given that there is no other information on our outcomes from studies that have asked patients with DSs what might represent a clinically meaningful change for them, we cannot say whether or not these effect sizes were clinically meaningful. 136 It is the case that the CBT + SMC arm’s scores on the WSAS, for example, were in the range associated with significant functional impairment but less severe clinical symptomatology than at baseline. 117

Further considerations of the data

Raw data suggested that in the CBT + SMC arm an improvement in DS frequency was observable at the 6-month follow-up point, although our SAP did not allow formal testing at that time point. Although we cannot evaluate which component(s) of the CBT intervention offered particular benefit, our DS-specific CBT approach did include seizure control techniques, which were generally viewed positively by patients and therapists. 135,137

In a related manner, raw data suggest that the SMC-alone arm experienced a reduction in DS frequency at a later follow-up point than the CBT + SMC arm. It is not clear what led to this apparently later improvement in the SMC-alone arm and whether or not this could be attributed to discussion (but not practice) of distraction techniques during SMC sessions or whether direction to self-help websites containing descriptions of potential seizure control approaches may have resulted in the use of techniques to attempt to reduce DSs, but at a slower rate than in the CBT + SMC arm.

Considering the apparent reduction in both trial arms’ DS frequency, it is important to note that we do not know whether or not this overall pattern simply represents the natural progression of the disorder over a comparable period either in general or for people with DSs specifically involved in a RCT. Although other studies have included long-term follow-up after diagnosis and information provision,55 such data are not comparable to those for people in a RCT with frequent research team contact and in a different cultural context. We are aware that SMC might itself be considered to be an active intervention, although we had not conceived of it as such when designing the study. Thus, the material in the information booklets given to participants at both study stages may have led to a better understanding of their disorder that they could share with families (see Chapter 5). Furthermore, the frequent contact by research workers and the interest shown in the participants as a result, and the requirement to self-monitor DSs throughout the study, may have served as an unintended intervention. We know from our qualitative work that many participants also found the contact with the SMC psychiatrists to be supportive, which may have helped reduce arousal levels and DS occurrence.

We did not seek to quantitatively investigate participants’ own perceptions of any advantages deriving from more rapid seizure reduction; however, for recent-onset cases in particular, this might be hypothesised to facilitate their return to work or study and may prevent progression to chronicity, although this requires detailed empirical study. Although CBT + SMC arm participants also seemed to show an earlier decrease in how bothersome they considered their DSs to be (see Table 14), and in their WSAS scores (see Table 15 and Figure 7b), further studies could usefully consider the real-life implications of benefit of CBT + SMC on DS frequency and WSAS scores and whether these are accompanied by improved actual work status or whether any ongoing seizures continue to pose a barrier. Again, although we cannot know which components of the CBT approach were selectively effective, our DS-specific CBT did address avoidance behaviour and this may potentially relate to the improvements seen on the WSAS.

Relation to other studies

Our findings support and extend those in our pilot RCT,1 yielding a similar trajectory for DS frequency over the course of the study. Thus, this pattern of improvement in DS occurrence is not limited to a single centre or set of therapists. The current findings, from the first adequately powered trial in this area, however, indicate a wider range of benefits from CBT + SMC than from SMC alone than we were able to demonstrate previously, even if they were not reflected in our primary outcome. The only other relevant pilot RCT68 that suggested the value of what was termed CBT-informed psychotherapy for seizure reduction was conducted on a much smaller scale, was not powered to permit a comparison between groups and did not evaluate findings at 12 months post randomisation; in addition, that study excluded patients with an epilepsy diagnosis.

Our study suggests that it is possible to engage patients with DSs in psychotherapy at an acceptable level and for the most part in SMC. Our guidance to neurologists in terms of what they should tell patients about the treatment pathway included asking them to communicate to patients some of the advantages of seeing a psychiatrist. It is possible that the guidelines given to neurologists about diagnosis delivery and the information they should convey to patients communicated a level of interest in their condition by professionals that encouraged them to participate in therapy.

Strengths and limitations

We will consider the strengths and limitations of the study overall in Chapter 6, along with recommendations for future research.

Perspective and resource use

The primary economic evaluation adopted a health and social care perspective. This approach is in line with recommendations from the National Institute for Health and Care Excellence (NICE) and is a convention followed in most health technology assessments conducted in the UK. However, it is often the case that interventions have effects outside the health and social care system, so we approach a societal perspective in the secondary analyses. To do this, we include costs associated with lost employment and costs of care provided by family and friends alongside the health and social care costs. We do not include patient time costs, costs associated with lost leisure/recreational activities or costs of any health impacts on carers; therefore, this is not a complete societal perspective.

The number of therapist contacts was recorded as part of the study. Other health and social care service use, receipt of care from family members and time off work were measured at baseline and at 6 and 12 months’ follow-up using a bespoke version of the Client Service Receipt Inventory (CSRI). 141 The CSRI is a widely used questionnaire that is usually adapted for each specific study and allows for the collection of retrospective data on health and social care use. The CSRI listed key services and to measure contacts with individual professionals the participant was asked whether or not this had occurred, how many times in the previous 6 months and what the typical duration of the contact was. For inpatient care, the CSRI asked for information on the specialty and number of days in hospital. For outpatient and accident and emergency (A&E) care it collected information on the number of visits and those visits made by an ambulance. Medication received (for all reasons) was also recorded. If an individual had a missing number of service contacts, then the median of others with these data was used.

The questions on care from family or friends specified specific areas: help in the home, help outside the home, personal care and help with medical problems. The participant was asked to state how many hours of care per week were provided in these areas because of their health problems. Participants were asked to state how many days they had lost from work during the previous 6 months because of health problems. Some services were measured with the CSRI but not costed owing to lack of appropriate values. This included self-help groups and online resources.

In most studies, including CODES, the CSRI relies on participant self-reporting and so may be prone to recall accuracy problems. To address this, we also used data from the HES as an alternative for assessing use and estimating the hospital inpatient, outpatient and emergency care costs. Data were requested from NHS Digital, Electronic Data Research and Innovation Service (NHS Scotland) and NHS Wales Informatics Service, and permission was granted for the use of these data. Ethics approval was granted for us to apply for HES data for participants who had not formally withdrawn from the study; we supplied these organisations with the participants’ personal identifier numbers and NHS (or in Scotland Community Health Index) numbers, and we were sent the data by personal identifier number but without other identifiers. Information was recorded differently for England, Scotland and Wales, although it was still possible to combine the data.

Cost calculations

Intervention costs were derived from information on salaries, overheads, training and supervision. The salaries were based on self-reported bands of the therapists taking part in the study. Costs of other health and social care contacts were calculated by combining the service use data with recognised national unit costs. 142,143 Average wage rates144 were used to value lost work and care provided by family/friends, assuming that time spent providing this could be spent on work or leisure. The use of wage rates in principle reflects the opportunity cost of this time. Medication costs were calculated using information in the British National Formulary. 145 Nationally applicable unit costs were also applied to the HES data that consisted of inpatient days, A&E visits and outpatient contacts. Costs are reported in 2017/18 Great British pounds. The list of unit costs used here is given in Appendix 8.

Quality-adjusted life-years

The main outcome measure in the economic evaluation was QALYs derived from EQ-5D-5L. 116 QALYs are a composite measure combining HRQoL and time, and allow comparisons to be made across different conditions. The EQ-5D-5L is a short scale that is used to measure HRQoL considering five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. For each domain, participants rate their current health using a five-point scale to describe the extent to which they can perform activities (ranging from ‘No problems/symptoms’ to ‘Unable to/extreme problems’). This gives rise to 3125 unique health states ranging from 11111 to 55555. The scale also includes a VAS that has been described in Chapter 2. Participants completed the EQ-5D-5L at all study points. Weights that are assumed to be proxies for ‘utilities’ were applied to the health states measured by the EQ-5D-5L. These weights are anchored by 1 (reflecting full health) and 0 (reflecting death), and were obtained from work conducted by Devlin et al. 146 QALYs were generated from the utility scores using area under the curve methods with an assumption of linear change between adjacent time points.

Analysis

Analyses were conducted using Stata software version 15. The use and cost of individual services were described but statistical tests were not conducted to analyse the group differences. Total health and social care costs and societal costs were compared between the two arms over the follow-up period. This was carried out using an ordinary least-squares linear-regression model with follow-up cost as the dependent variable and the baseline costs and group identifier included as independent variables. Cost data usually follow a skewed distribution and, therefore, bootstrapping methods were used to produce 95% CIs around the cost differences. Comparisons of QALYs over the follow-up period were also made using a regression model, controlling for baseline utility scores. 147

To assess cost-effectiveness, costs were combined with QALYs and the primary clinical outcome measure. Missing costs and QALYs were imputed using a best-subset regression procedure in Stata. Incremental costs and outcomes were estimated using regression models and 1000 differences obtained from bootstrapped resamples were plotted on a cost-effectiveness plane to investigate uncertainty around the incremental cost-effectiveness ratio obtained from point estimates. The intervention is considered ‘dominant’ if it is less expensive and more effective than SMC. If it is more expensive and more effective, incremental cost-effectiveness ratios were constructed to indicate the extra cost incurred to achieve a one-unit reduction in DS frequency or one extra QALY. To explore uncertainty around cost-effectiveness estimates (based on QALYs), we plotted cost-effectiveness planes (based on incremental cost–outcome pairs from 1000 bootstrapped resamples) and plotted cost-effectiveness acceptability curves (CEACs) (that were derived using the net benefit approach). We did not produce cost-effectiveness planes or CEACs for the analyses based on the difference in seizure frequency. CEACs reveal to decision-makers the probability of the intervention being cost-effective compared with the alternative, given different (implicit monetary) values placed on incremental improvements in the outcome measurement.

The HES data were analysed by examining the numbers of participants receiving inpatient care, receiving outpatient care or visiting A&E departments at baseline and in the 6 months prior to the 12-month follow-up, and making comparisons between the two arms. Numbers of contacts/days were also compared, as were mean costs. Owing to restrictions agreed when applying for the data, we did not link HES data to any variables other than participant ID and group idenifier. For this reason, these could not be directly used in the cost-effectiveness analysis.

Sensitivity analyses

In the sensitivity analyses, the Short Form questionnaire-6 Dimensions (i.e. utility score) (SF-6D), generated from the SF-12v2, was used to derive QALYs. 148 The SF-12v2 contains a wider range of items than the EQ-5D-5L and, thus, could reflect the impact of DSs more appropriately. A comparison of HRQoL has demonstrated that QALYs derived from the SF-6D performed well for people with epilepsy in discriminating between those with and those without seizures over a 2-year follow-up period,149 although there was no relevant follow-up data for DS patients. Further sensitivity analyses were conducted by increasing and decreasing the intervention costs by 25% and 50%. We examined the impact on societal costs of using alternative unit costs for informal care and lost employment. Specifically, we used the healthcare worker unit cost (£26 per hour) to derive informal care costs and a minimum wage of £7.50 for informal care and lost employment.

Service use

Service use during the 6-month period prior to baseline and at the 6- and 12-month follow-up period is described in Table 18. At baseline, > 80% of the participants in each arm had contact with GPs. Other community services were used by far fewer participants, with the next most commonly used being practice nurses. The mean number of contacts with community services was largest for GPs, followed by home helps. There were no obvious differences in the use of community services between the arms. During the 12-month follow-up period, around three-quarters of each trial arm had contacts with GPs. Other community services were again used by relatively few participants, although there were some differences between arms with the CBT + SMC arm participants more likely to have contact with practice nurses and the SMC-alone arm more likely to have contact with physiotherapists. It is also evident that a small number of participants were reporting receipt of CBT at baseline in each of the arms. At follow-up, this occurred for the CBT + SMC arm only. It may be that the seven participants who reported such contacts were referring to the actual trial intervention, or it may be that they were in fact receiving non-intervention CBT.

Turning to hospital use, Table 18 shows that around one-third of each arm had inpatient stays during the baseline period, with a similar average number of inpatient days. Outpatient appointments also occurred for about one-third of each arm. Use of an ambulance was slightly more likely in the SMC-alone arm, as were visits to A&E departments. Contacts with psychiatrists occurred for about three-quarters of each arm and contacts with neurologists for > 80%. During the entire 12-month follow-up period, there was less use of inpatient care, especially in the CBT + SMC arm. Outpatient use remained similar to baseline and there was reduction for both trial arms in visits to A&E departments. There were also reductions for both trial arms in contacts with psychiatrists and neurologists.

Receipt of informal care was very high in both trial arms, with nearly half of the participants at baseline reporting having received help with personal care and over half reporting having received help inside and outside the home. Over one-third reported that informal carers spent some time ‘on-call’ (i.e. had to remain with the participant in the home even if no care was being directly provided). When this occurred, it accounted for the most hours of care received per week. During the follow-up period, levels of informal care remained high and with no clear differences between the arms.

Service costs

The mean number of CBT sessions attended by participants in the CBT + SMC arm was 10. Combined with the unit cost of a clinical psychologist (used because this is the largest professional grouping of the current therapists) produced associated mean costs of £1064. One participant in the SMC-alone arm had access to the intervention and the mean cost for the arm was estimated at £7.

Mean costs of community services during the baseline period were highest for GP contacts and home help providing both personal care and support for household tasks (Table 19). The latter was more expensive for the SMC-alone arm. At follow-up, intervention costs were the highest healthcare cost for the CBT + SMC arm. At both the 6-month and the 12-month follow-up, GPs and home help service items again accounted for relatively high costs. Medication costs at baseline were similar between the arms, whereas at both follow-up points they were somewhat higher for the SMC-alone arm. Inpatient costs were similar at baseline but higher at follow-up for the CBT + SMC arm. However, the difference at 12 months of £332 is equivalent to < 1 extra inpatient day. Other hospital costs were relatively similar between trial arms at both baseline and follow-up, as were the total hospital costs.

Informal care costs were substantial at all time periods and for both trial arms. No major differences between the arms were apparent. Lost employment costs were higher for the SMC-alone arm at both baseline and follow-up.

The total mean health and social care costs at baseline were £4092 for the SMC-alone arm and £3798 for the CBT + SMC arm. From baseline to the 6-month follow-up, the costs were £2560 and £2922, respectively, whereas at the 12-month follow-up they were £2683 and £2915, respectively. The difference in health and social care costs over the entire follow-up period, adjusted for baseline, was £1834 (bootstrapped 95% CI £478 to £3475), which is around £840 more than the extra cost accounted for by the intervention itself.

The mean societal costs at baseline were £33,261 for the SMC-alone arm and £29,066 for the CBT + SMC arm. At 6 months, the societal costs were £22,828 for the SMC-alone arm and £26,183 for the CBT + SMC arm. During the combined 12-month follow-up period, the societal costs were £55,503 and £52,933, respectively. The imputed difference, adjusted for baseline costs, was £6566 (bootstrapped 95% CI –£5909 to £18,919).

Quality-adjusted life-years

Table 20 reports the utility scores and QALYs derived from both the EQ-5D-5L and the SF-12v2 measures. At baseline and at the 6- and 12-month follow-ups, the CBT + SMC arm had a slightly higher utility score. The CC analysis shows a QALY difference of 0.0416 in favour of the CBT + SMC arm. After imputation, the difference is similar at 0.0457. However, the QALY difference after controlling for baseline is less: 0.0152 for imputed cases (bootstrapped 95% CI –0.0106 to 0.0392).

When the SF-12v2 was used to generate the SF-6D, it was shown from the CC analysis that CBT + SMC accrued 0.0293 more QALYs than SMC alone. After imputation, the difference was 0.0233 QALYs. When baseline utility was controlled for, the differences became 0.0231 and 0.0157 QALYs, respectively. We also used the crosswalk method150 to convert the EQ-5D-5L scores to tariffs derived for the EuroQol-5 Dimensions, three-level version (EQ-5D-3L). This resulted in lower utility scores for both trial arms and an incremental QALY gain for CBT + SMC of 0.0089 QALYs.

Cost-effectiveness results

After imputation and controlling for baseline costs and utility, the CBT + SMC arm was shown to have incremental costs of £1834 and incremental QALYs based on the EQ-5D-5L of 0.0152 (Table 21). This gives an incremental cost per QALY of £120,658 compared with SMC alone. The incremental QALYs based on the SF-6D were 0.0157 QALYs, resulting in an incremental cost per QALY of £116,815.

Uncertainty around the cost-effectiveness results based on the EQ-5D-5L is depicted in Figure 9. Most of the bootstrapped replications fall in the top-right quadrant in which costs and QALYs are higher for the CBT + SMC arm. The percentage of replications falling below the red line indicating that the intervention is cost-effective based on a threshold of £20,000 per QALY is 27%. The corresponding CEAC (Figure 10) indicates that the probability of cost-effectiveness increases as the threshold value on a QALY rises, but is relatively low. Similar results apply to the cost-effectiveness results derived from the SF-6D (Figures 11 and 12).

FIGURE 9.

Cost-effectiveness plane using the EQ-5D-5L and costs measured from the health and social care perspective.

FIGURE 10.

Cost-effectiveness acceptability curve using the EQ-5D-5L and costs measured from the health and social care perspective.

FIGURE 11.

Cost-effectiveness plane using the SF-6D measure derived from the SF-12v2 and costs measured from the health and social care perspective.

FIGURE 12.

Cost-effectiveness acceptability curve using the SF-6D derived from the SF-12v2 and costs measured from the health and social care perspective.

With the crosswalk method used for the utility scores, the incremental cost-effectiveness ratio (ICER) for CBT + SMC increased to £206,067 per QALY. When the intervention costs were reduced by 25%, the ICER fell to £103,224 per QALY (Table 22) and further to £85,724 when the costs were reduced by 50%. By definition, the ICERs rise with higher intervention costs to £138,158 and £155,592 for 25% and 50% increases in intervention costs, respectively.

We have shown that societal costs are higher by £6566 for the CBT + SMC arm. With our main analysis showing a QALY difference of 0.0152, this results in an ICER of £431,974 per QALY.

Applying the unit cost of a healthcare worker resulted in increased informal care costs for both trial arms (£66,239 for the SMC-alone arm and £65,486 for the CBT + SMC arm). The difference between the arms (£2359) was not statistically significant (95% CI –£19,583 to £24,280). Applying the minimum wage of £7.50 to informal care would reduce the informal care cost to £19,209 and £18,990, respectively. Lost employment costs would be reduced from £2303 to £1174 for the SMC-alone arm and from £1721 to £903 for the CBT + SMC arm using the minimum wage.

Cost-effectiveness based on the number of seizures

The difference in median seizure rate between the two arms at 12 months is three seizures (see Table 14). The difference in costs is £1251 for the CC analysis and £1834 after imputation, producing ICERs of £41,740 and £611, respectively. The results imply that the cost for an avoided seizure over the 4 weeks prior to the 12-month follow-up is £611. There is, however, no accepted (cost-effectiveness) value of a seizure avoided and, therefore, it is simply a value judgement as to whether or not this represents value for money.

Hospital Episode Statistics analysis

The findings from the analysis of the HES are shown in Table 23. Between baseline and 6 months, > 80% of each arm had outpatient contacts, which fell substantially for the follow-up period. The number of outpatient contacts was very similar between arms and over time. Direct comparison with the CSRI data is difficult when we look at all outpatient contacts because we separated out psychiatrist and neurologist contacts with the CSRI data. At baseline, the CSRI data report 76% of patients having psychiatrist contacts in the CBT + SMC arm and 78% in the SMC-alone arm. The figures for neurologist contacts are 83% and 84%, respectively. The figure for other hospital care is 13% for both trial arms. Although all participants should have had neurologist and psychiatrist contacts, both the CSRI and the HES data indicate similar levels of use, and this is < 100%. Between one-fifth and one-quarter of patients in both trial arms at both time points have missed appointments. This information is not provided in the CSRI.

If we combine the cost of outpatient care from Table 23 with the psychiatrist and neurologist costs (these categories should be mutually exclusive), we find similarities with the costs in Table 19. (Other hospital costs are not included here.) At baseline, this cost is £497 for SMC alone and £491 for CBT + SMC (whereas the HES data show costs of £498 and £487, respectively). In the 6 months prior to the 12-month follow-up, the costs are £384 for SMC alone and £359 for CBT + SMC. The HES-based costs are £350 and £436, respectively. Therefore, there is reasonably good agreement between the CSRI and the HES data for outpatient care, except for CBT + SMC for which HES gives higher costs. At follow-up, the CBT + SMC arm does make more use of this than the SMC-alone arm, but the difference is not large.

The HES data also show very similar rates of A&E department contacts between trial arms, with participants of both arms less likely to attend in the follow-up period. The rate of use at baseline for the CBT + SMC arm is very similar when comparing HES (46%) with CSRI (47%) data; there is more of a difference for the SMC-alone arm (45% and 57%, respectively). The HES-based costs of A&E care are rather different at baseline from the CSRI-based costs (with higher costs for the CSRI), particularly for the CBT + SMC arm. However, at the 12-month follow-up there are very similar costs (£119 and £138 for the SMC-alone arm and £116 and £131 for the CBT + SMC arm, respectively).

At baseline, HES report that 31% of the SMC-alone arm and 33% of the CBT + SMC arm use inpatient care. The CSRI data are 36% for each arm. However, the costs of care are noticeably different when using the CSRI or the HES data. Costs based on the former are higher at baseline for both trial arms and for the CBT + SMC arm at follow-up. The implication is that although participants report a similar rate of inpatient use to HES, they also report more days in hospital. The difference at baseline between methods is comparable for the two arms, whereas at follow-up it seems to result in costs that are potentially overestimated by £341 for the CBT + SMC arm. If we were to reduce the overall incremental costs of £1834 for the CBT + SMC arm compared with the SMC-alone arm by this amount, the ICER is reduced to £95,096 per QALY (based on the EQ-5D-5L).

Main findings

To our knowledge, this is the first economic evaluation of CBT for people with DSs participating in an RCT evaluating a psychotherapeutic intervention. The main clinical results of the study have shown that although CBT + SMC was superior to SMC alone on most of the secondary outcomes, the primary outcome was not significantly different between the trial arms (see Chapter 3). In this chapter, we have found that CBT + SMC produces slightly more QALYs than SMC alone, but that the difference is not substantial. This is the case whether the EQ-5D-5L or the SF-6D was used to generate QALYs. The costs for the CBT + SMC arm were higher than those for the SMC-alone arm, with most of the extra cost accounted for by the intervention itself. It was apparent that the intervention did not seem to offset costs in other sectors. It is not unusual for interventions such as this to result in higher health costs than usual care, and this is not in itself evidence against cost-effectiveness. It is recognised that to achieve improved outcomes it is often necessary to have higher costs of care and this is where the incremental cost per QALY is relevant. Interventions that are assessed by NICE are usually considered to be cost-effective if they achieve a cost per QALY below £20,000–30,000. In this study, the cost per QALY for CBT + SMC was £120,658 compared with SMC alone. The figure was similar when QALYs were generated from the SF-6D. For this reason, the addition of CBT to SMC would be unlikely to be considered cost-effective. Costs were not markedly different between the trial arms when informal care and lost employment costs were included.

It was of interest that costs in both trial arms fell for some services. This is not unusual in trials, given that recruitment often requires various investigations, and health improvements (often experienced by receiving care in the control groups) should lead to reductions in use. The study was not, however, designed to explore changes in costs. There were not obvious cost differences between the two arms (other than for the therapy itself). However, medication costs were higher for SMC alone. This may be because of compensation for the lack of CBT in that arm.

Limitations

The analyses were conducted in a way that is consistent with similar NIHR-funded economic evaluations. However, a number of limitations need to be considered when interpreting these results. First, we relied on self-report data for the evaluation. The CSRI asked participants to recall what services they had received in the previous 6 months, which may have led to inaccuracies. There is, however, no reason to suppose that potential inaccuracy would differ between arms, and for most services that we wanted to collect data on it was the only option. The possibility of errors did, however, lead us to examine HES data in addition to those collected with the CSRI. For outpatient use, the findings from the CSRI and from HES were remarkably similar. For A&E visits, the costs were somewhat different at baseline but at follow-up were very similar using the different methods. The methods produced the largest discrepancies for inpatient care, with higher costs indicated by the CSRI. When patients are admitted they are likely to be particularly unwell and recollection of exact numbers of days may be exaggerated. When we produced an ICER taking this into account, it fell to £95,096.

What is clear from the analyses is that both the CSRI and the HES suggest that fewer than 100% of participants had psychiatrist and neurologist contacts during the baseline period. The trial procedures were that such contacts should take place for recruitment to occur. The CSRI data could quite easily be underestimates of real use, but HES data seem to validate the CSRI. Although HES should capture all contacts, it is still reliant on contacts being recorded and reported, which may not always have occurred. Some of the missed appointments reported in HES will also have been for neurologists and psychiatrists. One clinician in one site provided both neurological and psychiatric input (see Appendix 6), but it is unlikely that would be the reason for the figures reported here.

Second, although CSRI data covered baseline and the whole follow-up period, the HES data covered the baseline period and the 6 months prior to the 12-month follow-up. This means that we are not able to make full comparisons between the two methods. However, the comparisons we have made are still informative and largely validate the CSRI method.

Third, the outcome measure used in the economic evaluation was QALYs generated with the EQ-5D-5L. This is the approach recommended by NICE and used by most UK health economists. However, it is unclear whether or not the EQ-5D-5L is a sensitive measure in this patient group, and given that most secondary outcomes favoured CBT + SMC we might have expected larger differences. Use of the SF-6D-based QALYs revealed smaller differences between groups.

Fourth, the tariffs that were attached to the health states were taken from work conducted by Devlin et al. 146 Since we embarked on this study, concerns have been raised about this value set,151 with some arguing for conversion to tariffs derived for the EQ-5D-3L. 151

Fifth, the time horizon may be too limited to fully assess cost-effectiveness; it may be that the effects of CBT continue long term. Although a 1-year follow-up is reasonable for a trial, we might wish to conduct modelling work to determine the effects beyond this period. For CBT + SMC to be cost-effective, there would need to be ongoing differences in HRQoL compared with SMC alone. If ‘top-ups’ were required, then this would have cost implications. Continued health improvements would hopefully lead to future reductions in health service use that would also improve cost-effectiveness. We addressed this in sensitivity analyses and the ICERs rose substantially when using the crosswalk method.

Finally, there was one person in the SMC-alone arm who received the intervention. There were also participants in both trial arms who reported other CBT and talk therapy. However, the mean costs were low (see Table 19) and this would not substantially affect the results. In addition, we were not able to verify the nature of this additional therapy owing to the risk of unblinding the research workers. We cannot therefore be sure as to its nature or extent and had no analogous HES data against which to compare it.

Methods

Results

Summary

The CODES psychiatrists and CBT therapists felt that the patients who were referred to them in the study on the whole had a better understanding of the DS diagnosis than patients who might normally be seen outside the trial, owing to the more detailed delivery and explanation of the diagnosis from neurologists; this suggested that this early stage in the care pathway was important to the subsequent engagement of patients. There was consensus among trial participants, CBT therapists and psychiatrists that the dissociative seizures factsheets helped in understanding the diagnosis and were a valuable resource, providing clear information that could be understood by family and friends and other HCPs. Many of the participants randomised to CBT reflected very positively on their experience of treatment and felt that they had developed useful CBT skills to help with seizure management; however, treatment was not without its challenges for both the participants and the therapists and not all techniques were equally well experienced or implemented in all cases, highlighting the need to have a range of strategies to offer to patients. SMC was also positively received by most patients, and the psychiatrists noted that it could by itself lead to improvement. For the psychiatrists, keeping CBT techniques out of SMC appointments could be difficult, especially because all the psychiatrists interviewed had some therapy training.

Introduction

Within the CODES trial, neurologists were required to assume an important role in the treatment pathway. The aim of this process study was to explore neurologists’ experiences of their often more-than-usual in-depth role within the treatment of DSs and their thoughts on working with patients recruited to the CODES trial. More specifically, we investigated neurologists’ knowledge of DSs and their clinical practice for the condition before and after their involvement in the trial, as well as their opinions about CODES-related components. As we could not undertake a representative sample of qualitative interviews because 91 neurologists/epilepsy specialists had participated in the study, clinicians were asked to complete an online survey comprising 43 questions yielding qualitative and quantitative data.

Methods

Results

Participants’ clinical practice when working with patients with dissociative seizures

Thirty per cent of the neurologists reported that participating in the CODES trial had changed their practice regarding referring participants to psychological interventions. This included making earlier and more direct referrals for DSs, and the development of a DS pathway within their organisation as a consequence of the study. Of those who had not changed their practice, some already had local and well-developed specialised psychotherapy services in place for DSs. Others were still unable to refer to psychotherapy for DSs following the trial. During the trial, there was an improved ability to refer patients who lived outside the practice area to psychological services.

Around half (53.5%) of the respondents reported being able to refer patients to DS-specific psychological intervention before the CODES trial. This figure did not change after the conclusion of the trial. The availability of services appeared an important determinant. Doctors who had access to DS-specific psychological intervention prior to the trial referred around 78% of patients to these services. The comparable estimate from those who could not directly refer patients was 12%.

Sixteen per cent of neurologists reported that their involvement in the CODES trial changed their referral practice to psychiatrists/neuropsychiatrists for DSs. Both before and after their involvement in the CODES trial, the majority of neurologists agreed that psychiatry services should be involved in the care of all patients with DSs (before: median 4, IQR 3–5; after: median = 4, IQR 4–5; p = 0.1). However, only 55% said that they would have referred patients with DSs to psychiatrists/neuropsychiatrists before their involvement in CODES; for several respondents, this was because of a lack of local experts and resources.

Unsurprisingly, before their participation in the CODES trial, neurologists reported many different referral practices (Figure 13). Most neurologists referred patients with DSs directly to psychological interventions or to another professional who might then make this referral.

FIGURE 13.

Referral practice prior to involvement in the CODES trial. (Note that multiple responses were allowed.)

Neurologists’ confidence in recommending psychological therapy for DSs did not change as a result of their involvement in the CODES trial (before: median 4, IQR 4–5; during: median 4, IQR 4–5; p = 0.16). Although doctors reported having a good level of knowledge before the CODES trial of how psychological intervention may benefit patients with DSs, their ability to describe the psychological therapy offered to help patients with DSs improved during the study (before: median 3, IQR 2–4; during: median 4, IQR 3–4; p = 0.001).

Clinicians were asked to indicate their usual practice with respect to following up patients with DSs before, during and after their participation in the CODES trial (Figure 14). Overall, participation in the trial did not appear to affect how patients were followed up. Nonetheless, 26% of doctors indicated that the study had an impact on their practice relating to patients with DSs in other ways. Frequently reported changes included having more detailed conversations with patients; being better able to signpost patients and provide them with information; having greater confidence in the exploration of aetiological factors with patients with DSs; employing the CODES factsheet; making direct referrals for psychological treatment; making a greater number of referrals to other professionals; and feeling more confident making and delivering the diagnosis without stigma.

FIGURE 14.

Neurologists’ usual follow-up practice for patients with DSs before, during and after their involvement in CODES. (Note that respondents could indicate more than one option.)

CODES trial-related components

Doctors were asked their opinion on the usefulness of the different elements of the CODES care pathway and materials provided (Figure 15). Most doctors found all elements ‘very’ or ‘extremely’ useful, with particular emphasis on CBT for those patients who received it (86.1% of doctors) and the dissociative seizures factsheet (81.4%).

FIGURE 15.

Ratings of the usefulness of CODES study elements.

In addition, most neurologists reported that patient satisfaction for the psychiatric care (69.7%) and CBT (79.1%) within CODES was ‘very good’ (Figure 16), with 90% reporting that they would like this pathway to continue within their service. Overall, 91% of doctors also said that they would like to continue using the factsheet, with one specifically deeming it ‘extremely useful’ and suggesting that it offered ‘therapeutic benefit’ in itself. In addition, 84% of respondents reported referring their patients to relevant websites ‘often’ or ‘very frequently’.

FIGURE 16.

Ratings of CODES study components.

Around 84% of the neurologists reported ease of recruitment of patients with DSs to the study, as well as a general satisfaction of patients with their participation (> 69%). Despite this, doctors expressed more uncertainty with regard to keeping patients engaged (around 35% neither agreed nor disagreed), alluding to a level of difficulty in maintaining some patients’ engagement in the trial.

When providing additional thoughts on the CODES trial, positive comments included reference to raising awareness and understanding of DSs and that the pathway allowed patients quicker access to assessment and treatment of the disorder, avoiding existing waiting lists. The study also facilitated multicentred collaboration that was considered as a ‘lasting legacy of the trial’.

Two doctors said that they would like additional clinical resources to meet the increased clinical demand resulting from the CODES trial; another mentioned the problems faced by patients who had to travel long distances to receive treatment during the study.

Published conference abstracts

Goldstein LH, Chalder T, Carson AJ, Landau S, McCrone P, Magill N, et al. COgnitive behavioural therapy vs standardised medical care for adults with Dissociative non-Epileptic seizures (CODES): an RCT protocol. J Neurol Neurosurg Psychiatry 2015;86(9). (Conference: British Neuropsychiatry Association, London, February 2015.)

Goldstein LH, Chalder T, Carson AJ, Landau S, McCrone P, Medford N, et al. COgnitive behavioural therapy vs standardised medical care for adults with Dissociative non-Epileptic Seizures (CODES): progress in a randomised controlled trial. Epilepsia 2016;57:23. (Conference: 12th European Congress on Epileptology, Prague, September 2016.)

Goldstein LH, Chalder T, Carson AJ, Landau S, McCrone P, Medford N, et al. COgnitive behavioural therapy vs standardised medical care for adults with Dissociative non-Epileptic seizures (CODES): update on a pragmatic randomised controlled trial. J Neurol Neurosurg Psychiatry 2017;88:A26. (Conference: British Neuropsychiatry Association, London, February 2017.)

Goldstein LH, Chalder T, Carson AJ, Landau S, McCrone P, Medford N, et al. COgnitive behavioural therapy vs. standardised medical care for adults with Dissociative non-Epileptic Seizures (CODES): baseline characteristics of participants in a randomised controlled trial. Epilepsia 2018;59:S239. (Conference: 13th European Congress on Epileptology, Vienna, August 2018.)

Protocol available

Goldstein LH, Mellers JDC, Landau, S, Stone J, Carson A, Medford N, et al. COgnitive behavioural therapy vs standardised medical care for adults with Dissociative non-Epileptic Seizures (CODES): a multicentre randomised controlled trial protocol. BMC Neurol 2015;15:98.

See also update (statistical analysis plan)

Robinson EJ, Goldstein LH, McCrone P, Perdue I, Chalder T, Mellers JDC, et al. COgnitive behavioural therapy versus standardised medical care for adults with Dissociative non-Epileptic seizures (CODES): statistical and economic analysis plan for a randomised controlled trial. Trials 2017;18:258.

Explaining the diagnosis

• Name the condition DISSOCIATIVE SEIZURES is preferable here.

• Real attacks – not ‘imagined’ or ‘put on’ – can be frightening and disabling.

• Tell them what they do not have and why – i.e. why tests etc. show this is not epilepsy.

• Common and recognised condition diagnosed on positive grounds.

What are dissociative seizures?

• Dissociation is the medical word for a ‘trance-like’ state or ‘switching off’.

• ‘Normal’ examples of dissociation (e.g. day-dreaming, divided attention, ‘shock’).

What causes dissociative seizures?

• Causes often not obvious: may be related to ‘stress’ but this is often not apparent.

• Discuss aetiological factors emerging from clinical history. Relevance of predisposing, precipitating and perpetuating factors.

• Provide a model for the attacks – e.g. brain becomes overloaded and shuts down.

Treatment

• Potentially reversible condition.

• Information available on the CODES leaflet (e.g. self-help websites).

• Anti-epileptic drugs are not effective and usually stopped.

• Treatment of psychiatric comorbidity.

Information about the CODES study at point of randomisation

• We do not know what treatment is best for dissociative seizures.

• Understanding and acceptance of the diagnosis is often very helpful in its own right.

• We know that many people get better when they are given information about the condition and receive support from a specialist doctor (known here as ‘Standardised Medical Care’).

• We are doing the CODES study to find out if CBT is any better than ‘Standardised Medical Care’ (SMC) as we don’t know that CBT is actually better than SMC alone.

• Explain randomisation.

Record if ELIGIBLE? and WILLING TO BE CONTACTED? by researcher

How are you defining dissociative (non-epileptic) seizures?

1. Episodes of apparent altered responsiveness or ‘transient loss of consciousness’ [if in doubt refer/discuss with trial team (for example a paroxysmal movement disorder with some impairment of responsiveness would count even if the patient can remember the event)] 2. Clinical findings that demonstrate incompatibility between the episodes and recognised neurological or general medical conditions (e.g. typical features of a prolonged motionless episodes or a prolonged generalised thrashing attack with eyes tightly closed and a normal EEG) 3. Not better explained by another medical or mental disorder 4. The symptom or deficit causes significant distress, psychosocial impairment, or warrants medical evaluation.

I think the patient needs to be seen by a community mental health team. What should I tell them about treating the seizures?

You should refer them in the normal way. The patient should receive normal care for problems such as self-harm or depression. The local team need to know that treatment of the seizures remains your responsibility as part of a trial which the patient has consented to, and should not form part of their work with the patient.

Can I give the patient medication for anxiety or depression?

There is no recognised drug treatment for non-epileptic seizures, so medication should not be prescribed for the seizures themselves. However, some patients may have co-morbid anxiety or depression that warrants prescription of medication and in such cases medication may be given, as per normal clinical practice.