Facilitating the transition of young people with long-term conditions through health services from childhood to adulthood: the Transition research programme

Developments in health-care transition during the period of the research programme

Since the start of the programme in 2012, there have been developments in the field of transition of which we have taken account.

Results of published evaluative studies remain difficult to generalise from because they are usually about an intervention in one clinical setting, for one condition and using a locally designed intervention. The Cochrane review36 of interventions to improve transition, published in 2016, confirmed the lack of good evidence. Studies with more rigorous protocols have since been conducted. 37–40 A systematic review41 in 2017 reported the limitations of methods used in comparative studies of transition interventions. It found, for example, that most were small, single-condition studies, used unvalidated questionnaires, and did not report the age of transfer or the age at and time of questionnaire completion. In designing our research, we had anticipated such methodological limitations and addressed them in the methods we used for the longitudinal study [work packages (WPs) 2.1, 2.2 and 2.3].

Two Delphi studies proposed a suite of outcomes42 and process measures43 for evaluating transition. Our choice of outcomes in the longitudinal study is in line with these international studies.

A synthesis of qualitative studies of young people’s views about transition44 confirmed that there is much literature about the issues and difficulties encountered, and supported our decision not to undertake further exploratory work with young people.

As anticipated, arrangements for commissioning NHS services have continued to change during the research programme. However, the commissioning focus of our research concerns ‘what’ and ‘how’ to commission. We intended that our findings would be relevant to any commissioning structure, including whether or not commissioning and service provision are brought together in the same organisation.

Aims and objectives of the research programme

The overall aim of the programme was to promote the subjective well-being and health of young people with long-term conditions by generating evidence to enable NHS commissioners and trusts to facilitate the successful transition of young people from children’s to adults’ health care, thereby improving health and social outcomes.

To address this aim, the programme had three objectives:

1. to work with young people with long-term conditions to determine what successful transition means to them and what is important in their transitional health care

2. to identify the features of transitional health care that are effective and efficient

3. to determine how transitional health care should be organised, provided and commissioned.

Three separate WPs addressed each of these objectives. We present a summary of each WP and then bring together the results of all nine strands to inform our conclusions and their implications. As an overview, and to aid orientation, the WPs are summarised in Box 1, followed by Figure 1, which shows how the WPs influenced each other.

FIGURE 1.

How the WPs influenced each other. DCE, discrete choice experiment; UP, United Progression.

At the end of the account of each WP, we indicate its links to the other WPs and how each WP contributed to the seven main implications of our work.

In Appendix 1 is a summary of work we completed before applying for a Programme Grant.

In Appendix 2 are the management arrangements for the programme and a diagram showing how key groups interacted.

Aims

• To form and maintain UP.

• To consult with UP throughout the duration of the programme.

Method

Key findings

• UP influenced many of the WPs of the programme, and was involved in a wide range of dissemination and advocacy activities.

• As the members of UP grew in confidence, their contributions became increasingly useful to the programme. The support and training offered to UP were very important, although they were resource intensive.

• Members of UP reported that they enjoyed the challenges, the relationships they formed with other young people and their greater understanding of research.

Inter-relationships with other parts of the programme

UP influenced many aspects of the programme through their input to several WPs. The only two WPs they did not influence directly were those (WPs 3.2 and 3.3) on commissioning for transition.

Further analysis and discussion of the role of UP is presented in Involvement of patients, the voluntary sector and the public. This section also summarises the internal evaluation of the work UP undertook and points to our published paper. 45

Although not unique to our programme, when professionals, managers and commissioners attend meetings where young people contribute and have been closely involved in the work presented, such young-person involvement strengthens the assurance that the audience attaches to the work.

UP thoughts and opinions directly influenced implication 2 on DAH and the toolkit we developed to assist with introduction of such care across an organisation. By advising on questionnaire content, wording and feasibility, UP directly influenced both what research participants were asked to complete and the key implications of our work, namely implication 4 concerning joint planning, implication 5 concerning young people’s approaches to transition, implication 6 concerning features associated with better outcomes and implication 7 concerning features likely to optimise service uptake.

Aim

• To explore the usefulness of health passports in a project co-led by UP.

Methods

UP examined a range of passports designed to help young people navigate health services and concluded that there was no clear definition of what a ‘health passport’ was. As there were many health passports already available, UP decided that it would not be useful to develop another one. However, as no robust analysis of health passport use has been undertaken, UP aimed to examine the use and usefulness of some health passports already used by young people with long-term conditions.

UP contacted health-care professionals from a children’s hospital, a physical disabilities team, an adolescent rheumatology team and a voluntary organisation. UP asked professionals about their experience of health passports and this helped to identify questions and issues to explore with young people, such as:

• What kind of passports are used and what are their essential components?

• Who uses health passports and who benefits from them?

• How are they used and how are young people supported to use them?

UP then worked with the patient satisfaction team in the lead trust to devise a questionnaire to assess the views of young people about health passports. Then the professionals in the above organisations helped UP link to young people who used health passports. However, this proved difficult as ethics and research governance permissions were required; it was impossible to achieve this across several organisations within the available time.

Analysis

Young people from one service returned questionnaires (n = 13) and, although the response rate was low (20%), interesting findings emerged. None of the 13 participants said that they completed the health passport alone; the majority said that they thought passports were ‘useful’, but half of them never took them to health-care appointments. The passports were considered more useful to health professionals than to young people themselves.

Strengths and limitations

This work identified some qualitative findings about health passports. However, the work did not generate, as intended, enough quantitative data from which to draw definitive conclusions. It was difficult to obtain all of the research governance permissions in a timely manner to enable UP to approach as many young people as they had intended. The approaches also depended on the willingness and time of local health professionals to approach young people. For other parts of the programme, all such preliminary contacts and undertakings had been formalised before the programme started. For this work, co-generated by UP, planning could not begin until UP had worked through its ideas, which was at least 1 year into the programme.

Such challenges have been reported before. A Dutch group undertook participatory research with young people with chronic illness48 and concluded that, although the activity benefited the young people, it was an inefficient and unreliable way to gather research data. Similar benefits to young people were reported in a study in Vancouver. 49

Key findings

• There is a lack of conceptual clarity and no clear definition about what is meant by ‘health passport’.

• The work raises questions about whether or not a health passport is a tool truly held by young people. Who should complete it and when? Do young people really want it to ‘travel’ with them to appointments and between services?

• Even studies by enthusiasts for health passports found difficulties. Participants used their document only occasionally and its perceived utility was modest;50 only one-quarter completed a survey, and, of those who then agreed to a telephone discussion, only one-third had used their passport. 46

Inter-relationship with other parts of the programme

This WP contributed directly to the longitudinal study WP 2.1. A health passport was discussed as a candidate PBF of transitional health care. However, the initial work that UP undertook indicated that ‘health passport’ meant different things to different people, and although it was considered useful by some professionals, young people appeared to value it much less. We therefore decided not to include health passport as a PBF. This decision was vindicated by the subsequent survey work that UP undertook.

The confidence that members of UP gained by carrying out their research and their greater appreciation of research methodology enhanced the quality of all of their contributions to the programme, which were set out in the earlier account of WP 1.1. It also increased their self-confidence, and this is discussed in Involvement of patients, the voluntary sector and the public and in our associated published paper. 45

UP presented the health passport work at the Department of Health and Social Care’s ‘Takeover Day’, where interest was expressed in taking the passport work forward nationally.

UP presented their work as a poster at several meetings, including the final dissemination meeting at The King’s Fund, London, in October 2017. UP’s poster presentation is in Appendix 4.

Aims

• To explore the views of young people about transition.

• To identify and describe these views using Q-methodology.

Methods

Q-methodology combines quantitative and qualitative methods to investigate the range of possible views on a subject. It involves the rank ordering (Q-sort) by participants of a set of statements (Q-set) about a topic, after which a by-person factor analysis of the Q-sorts is used to reveal shared perspectives on the topic. Further details are in our published paper. 52

The Q-set of statements relevant to young people in transition was generated using quotations from two reviews of qualitative studies44,51 and our own mapping review of service models. 53 These were coded into a set of emergent themes, ‘planning’, ‘staff related’, ‘maturity’, ‘parent-related’ and ‘other’, which were then recategorised into several subthemes. One representative statement from each subtheme was selected. Following consultation with UP, the wording was adjusted and each final statement was prepared for sorting by participants onto a triangular grid. The statements are set out in table 1 of the published article. 52

Q-methodology requires that individuals in the sample are likely to hold different views. Therefore, maximum variation sampling was used, by age, gender, health condition and whether or not the young person had transferred. We sought 45 young people, aged 14–22 years, with a long-term condition that would soon require or had recently required them to transfer from children’s to adults’ services. Those with intellectual disability who could not complete the Q-sort were excluded. The number of participants is usually, but not necessarily, smaller than the Q-set. 54 The aim is to have four or five persons defining each anticipated viewpoint; there are often two to four, and rarely more than six. Accordingly, the breadth and diversity of viewpoints is probably best achieved when a participant group contains between 40 and 60 participants. 55

Clinicians from 10 adults’ and children’s specialties in a NHS Trust with tertiary and secondary care responsibilities invited young people with long-term conditions to participate. Q-sorts were administered face to face by the researcher.

Analysis

The degree to which an individual’s Q-sort corresponds to each factor derived from the factor analysis is given by their ‘factor loading’. This is a correlation coefficient between +1 and –1; the closer to 1, the more similar an individual’s Q-sort is to the factor. Individuals are ‘exemplars’ for a factor if they have a significant factor loading on that factor alone (p ≤ 0.01). We used PQMethod software (version 2.11; Peter Schmolck, University of the Bundeswehr, Munich, Germany) with centroid factor analysis followed by varimax rotation. 56 Outputs include the number of exemplars per factor, eigenvalues and factor variance, which provide information on the proportion of variance for the entire study explained by each factor. These are used alongside the postsort qualitative information to determine the ‘factor solution’: the final number of factors identified.

For each factor, an idealised ‘composite’ Q-sort is computed, illustrating how a person with a factor loading of 1 would have laid out their statement cards. Attention is paid to statements that characterise each factor, for example those placed in the +3 and –3 position on the grid, and those statements that distinguish between factors.

We found that a four-factor solution was optimal from a statistical viewpoint and most meaningful in clinical terms.

Strengths and limitations

Owing to the purposive sampling, claims cannot be made about how many individuals hold each point of view. Nor can inferences be made about subgroup sizes based on gender, age or condition.

We do not know if the four styles are stable over time. The views of some young people might relate to underlying personality characteristics and, therefore, might not change substantially. Alternatively, preferences may change as young people develop or as their health or health care changes.

Key findings

• Reviews of qualitative studies44,51 of transition have not revealed how individuals approach the dilemmas they face. Our results add a new dimension.

• Four factors or interaction styles of approaching transition were identified: ‘laid back’, ‘anxious’, ‘autonomy-seeking’ and ‘socially oriented’.

• These four diverse styles show that there is not one view on transition. Thus, a ‘one size fits all’ approach is inappropriate and tailoring care to each to each young person is crucial.

Inter-relationship with other parts of the programme

The recognition of these four interaction styles for approaching transition is a new finding and yields one of the seven key implications of the programme: implication 5.

The Q-sort study contributed to the design of the discrete choice experiment (DCE) (WP 2.3), which identified aspects of transitional health care that young people are more likely to take up. This in turn informed the economic analyses (WP 2.3).

Aims

To examine whether or not the PBFs of services are associated with better outcomes for young people with type 1 diabetes mellitus, cerebral palsy, or ASD and an associated mental health problem.

Hypothesis

In combination, and separately for type 1 diabetes mellitus, cerebral palsy and ASD, access to PBFs during the transition from children’s to adults’ health care predicts better health and social outcomes.

Methods

Analysis

Satisfaction with services

Satisfaction with services was generally good, as the ‘gap’ scores (the difference between the respondent’s ideal and current care) were small. Parents’ satisfaction was significantly lower than their children’s. Over the four visits, satisfaction with services did not change for those with diabetes mellitus, but for those with cerebral palsy or ASD it steadily worsened (Figure 2).

FIGURE 2.

Mind the Gap median score, young person reported: change over time by condition. CP, cerebral palsy; DM, diabetes mellitus.

Well-being

The median baseline well-being scores of those with diabetes mellitus or cerebral palsy were similar to each other and to those in the general population, and significantly higher (p < 0.001) than the scores of those with ASD. In the general population of 13- to 16-year-olds, the median score on the WEMWBS was 48.8. 77 In those aged > 16 years in the general population in England, the median was 50.7. 82

Over the four visits, the well-being scores of those with diabetes mellitus, cerebral palsy or ASD were steady. Those with ASD continued to have significantly lower well-being scores (p < 0.001) (Figure 3).

FIGURE 3.

The WEMWBS: change over time by condition group. CP, cerebral palsy; DM, diabetes mellitus.

Rotterdam Transition Profile

On every domain of the Rotterdam Transition Profile, except ‘education and employment’, significant differences were found between the three conditions (most p-values were < 0.01). A larger proportion of young people with diabetes mellitus were in a more independent phase of participation than those with ASD or cerebral palsy. Over the four visits, young people steadily moved to a more independent phase of transition on every domain. Those with diabetes mellitus remained further ahead than those with ASD or cerebral palsy (Figure 4). As the phase for the domain ‘using specialist services for condition’ was determined entirely by the structure of the services and could not be influenced by the young person, we dropped it from this and further analysis.

FIGURE 4.

Rotterdam Transition Profile. Percentages of participants per phase at (a) baseline and (b) visit 4, by condition. CP, cerebral palsy; DM, diabetes mellitus.

Independence in Appointments

The Independence in Appointments questionnaire was added from visit 2 onwards. At visit 2, the average score was 8. Those with diabetes mellitus had higher scores than those with ASD or cerebral palsy, and these differences persisted even though all scores increased, indicating more independence (Figure 5).

FIGURE 5.

Independence in Appointments scale: change over time, by condition. CP, cerebral palsy; DM, diabetes mellitus.

Cerebral palsy condition-specific measure of unmet need

As reported by parents, the proportion of young people with cerebral palsy with some unmet needs increased during the follow-up on the ‘daily living health care’ and ‘medical care’ domains. 64 As reported by the young person, an increase in unmet needs was seen only in the ‘daily living health care’ domain (Figure 6). When considering only those with some unmet needs, mean scores decreased slightly for the ‘daily living health care’ domain and remained the same for the ‘medical care’ domain, as reported by both young people and parents. These results were for the 64 young people who provided data at every visit. The results were similar when we examined data from each visit, whether or not data were available from other visits.

FIGURE 6.

Needs reported by (a) young person; and (b) parent.

These results draw attention to the need for more co-ordinated care in adults’ services for those with cerebral palsy. We discuss this further in our published paper64 and compare UK practice with that of other countries.

Autism spectrum disorder condition-specific measure of anxiety and depression

The proportion of young people with ASD and an associated mental health problem with an ‘abnormal’ score on the ‘anxiety’ and ‘depression’ domains (as measured by the HADS) remained similar between baseline and visit 4 (‘abnormal’ means above the recommended threshold score for a likely disorder) (Figure 7). The results were similar if analysed for the participants with data at baseline or visit 4, or if analysed for only those participants with data at baseline and visit 4.

FIGURE 7.

Proportions of normal, borderline and abnormal scores from the HADS at baseline and visit 4.

Further analysis of those with ASD is provided in Appendix 5. Having an additional diagnosis of attention deficit hyperactivity disorder or taking medication were predictors of a transfer to adult mental health services. Overall, the young people in our sample appeared to be doing reasonably well, with 67.5% in full-time education (either school or college) at the end of the study. Under one-third were in neither education nor employment. Analysis of HADS trajectories found that some young people were doing well, managing their mental health and able to engage successfully with services. However, the majority were experiencing episodes or continued levels of high anxiety. Qualitative data revealed that many young people struggled when faced with more challenging academic and social environments.

Proposed beneficial features

Some material in this section is from Colver et al. 63 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

The proposed beneficial features that services reported that they provided over the first year (i.e. before visit 2)

The following paragraph is covered in more detail in our published paper. 63 Overall, the nine PBFs of services for transitional health care were often not offered to young people in our study by NHS services. Fewer than half of the 35 services stated that they provided an ‘age-banded clinic’, ‘written transition plan’, ‘transition manager for clinical team’, ‘protocol for promotion of health self-efficacy’ or ‘holistic life-skills training’. Features were offered by services most often for young people with diabetes mellitus and least often for those with cerebral palsy (Figure 8).

FIGURE 8.

Percentage of ‘yes’ responses from each service by condition, stating whether or not the PBF was offered. CP, cerebral palsy; DM, diabetes mellitus.

The proposed beneficial features that young people reported that they experienced over the first year (i.e. before visit 2)

The content of the following paragraph is covered in more detail in our published paper,63 which integrates the findings from this WP and the qualitative work of WP 2.2. At visit 2, young people often reported that they had not experienced the features that services said they provided. Agreement between the young person’s report and what the service said it offered was 30% for ‘written transition plan’, 43% for ‘holistic life-skills training’ and 49% for ‘key worker’. Agreement was better for ‘appropriate parent involvement’, at 77%; for ’age-banded clinic’, at 77%; for ‘promotion of health self-efficacy‘, at 80%; and for ‘co-ordinated team’, at 69%.

Associations of proposed beneficial features with outcomes

Two approaches were adopted to assessing the associations of PBFs with outcomes. In both, we used the young person’s account of whether or not they had experienced a PBF, rather than what the service reported that it provided or aspired to provide.

In the first approach, it was established at each research visit whether or not the young person had experienced each PBF at least once in the previous year, and this was recorded as yes or no. A separate analysis was then conducted for each PBF (as the independent variable) against the outcomes. This was undertaken for each annual period. Thus, this approach is a series of analyses that examine associations over 1 year, directly related to what the young person is experiencing at the time.

In the second approach, a consolidated indicator was compiled for each PBF for the duration of the study. The criteria for being optimal or suboptimal varied by PBF and was calculated as outlined in Methods. The proportions of young people experiencing optimal or suboptimal consolidated indicators are shown in Table 3. In the analyses, the independent variable was whether the consolidated indicator was optimal or suboptimal for each PBF; the dependent variables were the outcomes captured at the ‘final’ visit.

TABLE 3 - Consolidated indicator of PBFs at final visit, by condition group

N/A, not applicable.

PBF	Consolidated indicator	All, n (%)	Diabetes, n (%)	Cerebral palsy, n (%)	ASD, n (%)	p-value
Age-banded clinic	Optimal	145 (53)	109 (97)	16 (22)	20 (23)	< 0.001
	Suboptimal	111 (40)	2 (2)	54 (73)	55 (62)
	Missing	18 (7)	1 (1)	4 (5)	13 (15)
Meet adult team before transfer	Optimal	111 (40)	73 (65)	16 (22)	22 (25)	< 0.001
	Suboptimal	133 (49)	31 (28)	54 (73)	48 (55)
	Missing	30 (11)	8 (7)	4 (5)	18 (20)
Promotion of health self-efficacy	Optimal	116 (42)	76 (68)	18 (24)	22 (25)	< 0.001
	Suboptimal	151 (55)	29 (26)	56 (76)	66 (75)
	Missing	7 (3)	7 (6)
Written transition plan	Optimal	48 (17)	32 (29)	11 (15)	5 (6)	< 0.001
	Suboptimal	185 (68)	62 (55)	59 (80)	64 (73)
	Missing	41 (15)	18 (16)	4 (5)	19 (21)
	N/A	2 (1)		1 (1)	1 (1)
Both young person and parent happy with parent involvement	Optimal	93 (34)	36 (32)	28 (38)	29 (33)	0.444
	Suboptimal	141 (51)	55 (49)	33 (45)	53 (60)
	Missing	38 (14)	21 (19)	12 (16)	5 (6)
	N/A	2 (1)		1 (1)	1 (1)
Key worker	Optimal	79 (29)	56 (50)	3 (4)	20 (23)	< 0.001
	Suboptimal	170 (62)	47 (42)	68 (92)	55 (62)
	Missing	25 (9)	9 (8)	3 (4)	13 (15)
Co-ordinated team	Optimal	167 (61)	104 (93)	25 (34)	38 (43)	< 0.001
	Suboptimal	66 (24)	2 (2)	40 (54)	24 (27)
	Missing	25 (9)	6 (5)	4 (5)	15 (17)
	N/A	16 (6)		5 (7)	11 (13)
Holistic life-skills training	Optimal	132 (48)	74 (66)	18 (24)	40 (45)	< 0.001
	Suboptimal	117 (43)	28 (25)	52 (70)	37 (42)
	Missing	25 (9)	10 (9)	4 (6)	11 (13)
Transition manager for clinical team	Optimal	60 (22)	27 (24)	14 (19)	19 (21)	0.947
	Suboptimal	143 (52)	67 (60)	34 (46)	42 (48)
	Missing	71 (26)	18 (16)	26 (35)	27 (31)
Total (n)		274	112	74	88

All models were adjusted for age, sex, condition and potential for clustering by site. Significant associations (p < 0.05) from these models were further adjusted for transfer status, time since transfer to final visit (if applicable) and time to first adult appointment (if applicable).

Examining the Rotterdam Transition Profile, we performed the logistic regression by comparing phases 1 and 2 with phase 3.

Table 4 sets out the effect sizes for all associations that reached significance of p ≤ 0.05.

TABLE 4 - Associations of PBFs with outcome measures

MTG, Mind the Gap; RTP, Rotterdam Transition Profile.

Odds ratio.

Notes

Coefficients and odds ratios ‘year-by-year’ and for the consolidated PBF indicator have suboptimal PBF delivery as the reference group.

A higher Mind the Gap score means less satisfaction with services than a lower score.

Numbers in square brackets indicate a negative association, i.e. satisfactory PBF was associated with worse outcome.

Bold indicates p ≤ 0.01.

PBF	PBFs by ‘year-by-year’ visits					Consolidated PBF indicator at final visit
	Period	Outcome	Whether regression b-coefficient or odds ratio	95% confidence interval	p-value	Outcome	Whether regression b-coefficient or odds ratio	95% confidence interval	p-value
Appropriate parent involvement	1	MTG: total	−0.48	−0.76 to −0.19	0.001	MTG: total	−0.67	−0.97 to −0.37	< 0.001
	1	MTG: environment	−0.57	−0.92 to −0.21	0.001	MTG: environment	−0.71	−1.17 to −0.26	0.006
	1	MTG: provider	−0.52	−0.82 to −0.22	0.001	MTG: provider	−0.75	−0.99 to −0.52	< 0.001
	1	RTP: domestic	[0.14a]	0.03 to 0.67	0.01	RTP: finances	[0.6a]	0.92 to 0.39	0.02
	1	RTP: health care	[0.35a]	0.17 to 0.72	0.004	WEMWBS	2.18	0.21 to 4.15	0.03
	1	RTP: services and aids	[0.42a]	0.18 to 0.97	0.04
	2	MTG: total	−0.60	−1.00 to −0.21	0.003
	2	MTG: environment	−0.65	−1.08 to −0.21	0.004
	2	MTG: provider	−0.48	−0.90 to −0.06	0.03
	2	MTG: process	−0.82	−1.32 to −0.31	0.002
	3	WEMWBS	4.47	1.96 to 6.97	0.001
	3	MTG: total	−0.87	0.45 to 1.29	0.001
	3	MTG: environment	−0.91	−1.44 to −0.37	< 0.001
	3	MTG: provider	−0.95	−1.37 to −0.52	< 0.001
	3	MTG: process	−0.63	−1.18 to −0.07	0.03
	3	WEMWBS	3.45	0.92 to 5.99	0.008
Promotion of health self-efficacy	1	MTG: total	−0.51	−0.87 to −0.15	0.006	MTG: total	−0.32	−0.62 to −0.03	0.04
	1	MTG: environment	−0.63	−1.06 to −0.18	0.005	MTG: environment	−0.57	−1.03 to −0.11	0.02
	1	MTG: provider	−0.47	−0.85 to −0.09	0.02	MTG: process	−0.37	−0.70 to −0.04	0.03
	1	MTG: process	−0.46	−0.91 to −0.01	0.04
	1	RTP: finances	[0.26a]	0.07 to 0.93	0.04
	1	RTP: domestic	[0.04a]	0.01 to 0.21	0.001
	2	MTG: total	−0.51	−0.91 to −0.11	0.01
	2	MTG: provider	−0.49	−0.90 to −0.08	0.02
	2	MTG: process	−0.70	−1.21 to −0.19	0.007
	2	MTG: total	−0.60	−1.01 to −0.20	0.004
	3	MTG: environment	−0.90	−1.40 to −0.40	< 0.001
	3	MTG: provider	−0.52	−0.95 to −0.11	0.01
Meet adult team before transfer	1	RTP: domestic	6.29a	1.60 to 24.80	0.009	RTP: education/employment	2.33a	1.21 to 4.55	0.01
	1	RTP: health care	2.71a	1.24 to 5.90	0.01	RTP: finances	2.78a	1.10 to 7.14	0.03
	1	RTP: services and aids	5.15a	2.08 to 12.78	< 0.001	RTP: services and aids	2.50a	1.06 to 5.88	0.04
	1	RTP: transport	2.01a	1.06 to 3.79	0.03	Autonomy in appointments	1.60	0.32 to 2.87	0.02
	1	RTP: education/employment	3.24a	1.09 to 9.65	0.04
Meet adult team before transfer (continued)	1	Autonomy in appointments	1.69	0.80 to 2.58	< 0.001
	2	RTP: finances	2.64a	0.92 to 6.62	0.02
	2	RTP: transport	2.08a	1.11 to 3.90	0.02
	2	Autonomy in appointments	1.00	0.01 to 2.0	0.05
	3	RTP: health care	2.11a	1.03 to 4.34	0.004
	3	Autonomy in appointments	1.46	0.34 to 2.59	0.01
Key worker	1	RTP: leisure	[0.56a]	0.32 to 0.96	0.04
	2	MTG: provider	−0.66	−1.04 to −0.28	0.001
	2	MTG: process	−0.69	−1.17 to −0.21	0.005
	3	RTP: education/employment	[0.40a]	0.20 to 0.95	0.02
Holistic life-skills training	1	RTP: domestic	[0.19a]	0.04 to 0.93	0.04
	1	RTP: services and aids	[0.34a]	0.12 to 0.94	0.04
	2	MTG: provider	[0.43]	0.03 to 0.84	0.04
	3	MTG: total	−0.46	−0.87 to −0.05	0.03
	3	MTG: provider	−0.57	−0.99 to −0.14	0.009
	3	RTP: domestic	2.47a	1.10 to 5.58	0.003
	3	RTP: romantic relationships	[0.52a]	0.26 to 0.98	0.04
Written transition plan	1	RTP: romantic relationships	[0.43a]	0.19 to 0.96	0.04
	2	MTG: total	−0.72	−1.39 to −0.04	0.04
	2	MTG: process	−1.19	−1.39 to −0.04	0.007
Co-ordinated team	1	RTP: domestic	[0.19a]	0.04 to 0.93	0.04	MTG: provider	−0.67	−1.25 to −0.09	0.03
	3	RTP: health care	[0.17a]	0.04 to 0.82	0.03	RTP: education/employment	[0.31a]	0.11 to 0.82	0.02
	3	RTP: services and aids	[0.22a]	0.06 to 0.81	0.02	RTP: domestic	[0.41a]	0.19 to 0.91	0.03
Transition manager for clinical team		No associations				RTP: domestic	2.63a	1.16 to 5.88	0.02
						RTP: services and aids	[0.41a]	0.20 to 0.86	0.02
						RTP: romantic relationships	[0.52a]	0.31 to 0.89	0.02
Age-banded clinic	1	WEMWBS	3.08	0.18 to 5.98	0.04
	1	Autonomy in appointments	1.44	0.48 to 2.4	0.003
	2	RTP: education/employment	5.22a	1.21 to 22.53	0.03

Interpretation of the associations of proposed beneficial features with outcomes

This is presented in more detail in our published paper. 66 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Three PBFs of transitional health care had significant (p ≤ 0.01) positive associations with better outcomes, namely ‘appropriate parent involvement’, ‘promotion of health self-efficacy’ and ‘meeting the adult team before transfer’. The b-coefficients indicated changes of approximately 0.5 SDs with respect to the satisfaction with services scale (SD 1.5 in our population), well-being (SD 7.0 in our population) and autonomy in appointments (SD 3.0 in our population). The odds ratios indicated increased likelihoods of being in a more independent phase of transition.

The other six PBFs had few statistically significant positive associations (p ≤ 0.01) with better outcomes in the year-by-year analysis, had a number of negative associations and had no positive associations with the consolidated indicator of exposure to PBFs.

‘Appropriate parent involvement’ and ‘promotion of health self-efficacy’ were reported to have been experienced satisfactorily by fewer than half of participants across transition across the three conditions. However, they were experienced by more young people with diabetes mellitus than those with cerebral palsy or ASD. For ‘meeting the adult team before transfer’, around two-thirds of young people with diabetes mellitus reported that they had met a member of the adult team, but fewer than one-quarter of those with cerebral palsy or ASD had done so. Thus, we found a different quality of experience of transitional health care between young people with a long-term illness (diabetes mellitus) and those with a long-term disability.

Strengths and limitations

Looking first at the strengths, our study drew 374 young people from across the UK and from many NHS Trusts. This is the largest sample for research in the UK in terms of examining transition longitudinally and collecting hypothesis-driven data from young people at home visits (compared with data-linkage studies from administrative data sets). Although we cannot be certain of representativeness and, therefore, generalisability across UK regions and NHS Trusts, we studied individuals from Greater London; north, south-east and south-west England; and Northern Ireland. Furthermore, we drew individuals from 27 NHS Trusts, and the trusts varied greatly in the number and variety of the PBFs they offered (see Figure 8).

We recruited from all young people with cerebral palsy in two population registers and all young people with diabetes mellitus or ASD and an associated mental health problem in nine NHS Trusts widely distributed across England (all such young people are seen in secondary care). Thus, complete populations of individuals with one of the three conditions were sampled from (e.g. rather than those attending particular schools, specialised tertiary health-care services or voluntary support groups).

The three conditions chosen were exemplars of chronic illness (diabetes mellitus), complex physical impairment (cerebral palsy) and neurodevelopmental impairment (ASD); they were deliberately chosen because they varied in terms of health needs, psychosocial complexity and availability of adults’ services. Thus, we consider that generalisability to other conditions is likely.

The 374 participants did not differ significantly from those declining to take part, other than for a small effect of deprivation. The distribution of severity of the three conditions was similar to that in national samples. 62 Those with the conditions had a wide range of severity; for example, the young people with cerebral palsy ranged from wheelchair users to those with independent ambulation. Attrition did not appear to create bias, as there were no significant differences between those remaining and those not remaining in the study with respect to sex, age, condition, diabetes site or ASD site. In Northern Ireland, there was a small effect of deprivation on the attrition of those with cerebral palsy. Thus, we think that, after non-participation and attrition, our study group continued to be representative of those we aimed to recruit.

Turning to the limitations, our study did not include young people with a learning disability because in the UK there is already a lifespan service for such individuals, and we would not have been able to base our quantitative studies on self-completed questionnaires, nor our qualitative work on interviews with young people. However, the exclusion of such individuals means that our results cannot be generalised to them.

Recruitment, especially of those with cerebral palsy, was more difficult than we expected. A possible explanation may be the method of recruitment, which was letters of invitation, rather than the young people’s clinician approaching them directly, as was undertaken for those with diabetes mellitus or ASD. Overall retention at 70% was acceptable in comparison with other longitudinal studies of adolescent populations. 32

It might be desirable to follow up young people for longer in adults’ services to assess later outcomes of transitional health care, but this is unlikely to be feasible. The fact that some young people were still in children’s services does not invalidate analyses; our study was about transition, not specifically health-care transfer. Table 1 shows the number still in children’s services.

Because of their personal circumstances, there was variability in precisely when the young people were seen by researchers. This is one reason why we chose the ‘final visit’ to be visit 4 or visit 3 (if the young person did not have a visit 4). Some of the visit 3s were considerably delayed and too close to a fourth visit to justify further data collection during the study. We examined the association of exposure to a PBF over the previous year in relation to outcomes. However, what constitutes optimal exposure over 3 years when, for instance, exposure occurs in one year but not the others? We made a clinical judgement about this for each feature for the consolidated longitudinal analysis. The decisions were agreed among our group, but others might have made different judgements.

Some PBFs, such as ‘written transition plan’, were infrequently provided and therefore we had less statistical power to detect associations with outcomes.

Key findings

• There were striking mismatches between the PBFs of transitional health care that a service said it provided and the features that young people reported to have experienced.

• PBFs were both provided and experienced less often in services for those with cerebral palsy and ASD than for those with type 1 diabetes mellitus.

• Average satisfaction with services was reasonable at baseline (rating of ideal service only slightly higher than rating of actual service). It remained steady for those with type 1 diabetes mellitus but worsened significantly over the study period for those with ASD or cerebral palsy.

• The well-being scores for each group remained similar over time. Those with type 1 diabetes mellitus or cerebral palsy had an average subjective well-being at baseline similar to that of the general population. Those with ASD and an associated mental health problem reported significantly lower well-being than those in the other two groups, and this difference persisted over the course of the study.

• For participation in life activities, there were significant differences between the three conditions, with young people with type 1 diabetes mellitus in more independent phases of transition than those with ASD or cerebral palsy. All made progress over time, but those with type 1 diabetes mellitus remained more independent.

• Condition-specific outcomes worsened during the study for those with type 1 diabetes mellitus or cerebral palsy. The proportion of young people with ASD with so-called ‘abnormal’ anxiety and depression scores remained the same from baseline to visit 4, with over one-third having abnormal levels of anxiety.

• The features of services for transitional health care associated with better outcomes were ‘appropriate parent involvement’, ‘promotion of health self-efficacy’ and ‘meeting the adult team before transfer’.

Inter-relationship with other parts of the programme

We were guided by advice from UP (WPs 1.1 and 1.2) in determining the wording of some questionnaires, which PBFs to include, how to approach young people and how to interpret the differences between what PBFs services said they provided and what young people experienced.

A subsample of the cohort in this WP were recruited to the qualitative longitudinal sample for WP 2.2.

Work package 2.3 constructed the DCE that was administered at visit 3.

This WP addressed the main hypothesis of the programme. In drawing conclusions about the clinical significance and validity of the findings (see Conclusions), we were further influenced by discussions with UP (WPs 1.1 and 1.2); we triangulated the findings with the results of the qualitative work in WP 2.2 and the health economic analysis in WP 2.3.

The findings of this WP informed the content of the DAH toolkit in WP 3.1b.

The findings of this WP contributed directly to implication 6 of the programme.

Aim

The aim was to investigate how young people with long-term conditions experience services for transitional health care.

Methods

More information about methods is in Appendix 6, which includes the interview guide.

A longitudinal multimethod qualitative study was conducted with young people recruited from the longitudinal quantitative study (WP 2.1). Young people were initially approached by a member of the WP 2.1 study team and, if they wanted more information, they were contacted by telephone or e-mail by WP 2.2 researchers previously unknown to them.

Data collection took place between June 2013 and June 2016. Recruitment was driven by criterion sampling (i.e. diagnosis, age, service) and theoretical sampling (i.e. experience of service, comorbidities, social network). 85 Young people were aged between 14 and 18 years at the first interview. Each young person identified a ‘significant other’ person in their life and a health professional significant to their health care whom they wanted to include in this qualitative study, and to nominate one consultation that the researcher could observe. Thirteen young people took part in this study. Family members who took part were 10 mothers, three fathers or stepfathers and one brother. Nineteen health professionals were also identified: seven paediatric consultants, five paediatric nurses/allied health professionals, three adult consultants and four adult nurses/allied health professionals. We interviewed 13 of these: nine from children’s services and four from adults’ services.

Thirteen young people received a first interview, of whom four were interviewed again at least 18 months later. They were interviewed on their own or with a relative, if they wished, at their home or at a clinic. Fourteen family members received a first interview, of whom eight were interviewed again at least 18 months later. In total, we undertook 40 first interviews (Table 5). In 30 of these, the young person was pre transfer to adults’ services and in 12 they were post transfer.

Observations were recorded in contemporaneous field notes. Formal interviews (face to face or by telephone) lasting between 30 and 120 minutes were audio-recorded, transcribed and edited to ensure respondents’ anonymity and then analysed alongside anonymised field notes. All analysis was conducted according to the standard procedures of qualitative analysis. 86 We used procedures from first-generation grounded theory (coding, constant comparison, memoing87), and from analytic induction (deviant case analysis88). Data collection and analysis occurred concurrently, so that issues raised in earlier phases of fieldwork were explored in subsequent ones, including in follow-up interviews. We undertook independent coding and cross-checking, team data sessions and data triangulation (different times and persons) and between-method triangulation (interview, observation).

Analysis

The full report is in Appendix 6. The report focuses on the observational data and interviews, with accompanying commentary and quotations in relation to the nine PBFs. The work is summarised below under three headings:

1. disrupting relationships of trust

2. supporting successful transfer

3. enabling in(ter)dependence.

Strengths and limitations

Conducting serial and triangulated interviews enabled us to confirm stories and capture evolving perspectives. Observing consultations enabled us to compare what people do against what they said they did in these contexts. Although both the form and the content of health care received may differ among the three exemplar groups, the organisation and experiences of the transition and transfer process showed many more similarities than differences.

The findings have face validity, as they relate well to the findings in other WPs and other studies in the literature, especially how transition and transfer can be disorientating for young people and their families. 44,51,83 This study strongly echoes work on how young people build relations of trust with health professionals and services,44,51 especially where those relations are consistent and supportive,58 as well as work on how the family, especially the mother, have built relations of trust with health professionals and services. 83

In its nature, qualitative work is in-depth but small-scale, a consequent strength being the richness of data and a weakness being the small number of case examples. We aimed to recruit 15 families and achieved 13. We had difficulties in recruiting young people, with many seeing the breadth and depth of data collection as an additional burden while they were already taking part in WP 2.1. The retention of young people for follow-up and/or post-transfer interviews was poor, so learning about the adult context was limited. We were unable to contact some young people, and some of those we did contact declined follow-up interviews; they either withdrew or suggested that we speak to their parents or health professionals. Thus, our findings are valid in relation to the data we collected; however, it is possible that we did not identify key messages that would have been relevant after the transfer to adults’ services.

Key findings

• Transition and transfer are disorientating for young people and their families. They have practical, cognitive and emotional impacts. Importantly, they disrupt often hard-earned knowledge of how best to navigate the children’s health-care system effectively.

• Health professionals need to help young people and their families adjust to and mitigate these disruptions. The key practices that minimise this disruption are being clear in discussions and documents about processes, discussing how adults’ and children’s practices differ, undertaking joint clinics and meetings with members of the adult team that the young person will actually be seeing and including family members in the process.

• Transition is about enabling young people to become responsible for their own health care. However, family members remain crucial. Such involvement is normal, not dysfunctional, and for people in other age groups is largely accepted. Family members can be helped to adjust their roles to have more partnership and sharing of care and consultation with their son or daughter.

Inter-relationship with other parts of the programme

This WP, 2.2, was informed by WP 1.1 in which the UP group commented on the draft interview schedule and the practicalities of the planned visits to young people and clinic consultations.

The sample for WP 2.2 was a subsample of that in WP 2.1 and so the two WPs worked closely together.

Our published article,63 on intended provision of PBFs and whether or not they are experienced by young people, integrated quantitative findings from WP 2.1 with qualitative findings from this WP 2.2. Our article directly influenced our decision to use the young person’s experience of each PBF as the exposure in our regression analyses, rather than what the service said it aimed to provide.

The findings of this WP influenced interpretation of the three PBFs associated with improved outcomes in the conclusions section. This discussion informed one of the seven key implications of the programme, implication 6.

Discrete choice experiment

Discrete choice experiments are regularly used in health economics to elicit preferences for health-care products and programmes, and in the valuation of preferences for health states. 90–92 Further details are in Appendix 7.

Aim

The aim was to quantify the strength of young people’s preferences for aspects of health services for transition using a DCE.

Methods

A DCE describes a service in terms of a number of characteristics, or ‘attributes’ (e.g. the flexibility of appointments, parental involvement). The extent to which an individual values an intervention is expected to depend on the ‘level’ these attributes take (e.g. can appointments be made outside office hours or not?). Thus, a DCE explores relative preferences about how services can be organised, by choosing between services with differing levels of their attributes. 93

The design and conduct of the DCE involved four steps.

Analysis

In total, 247 participants completed at least one DCE question (fully completed, n = 223; partially completed, n = 24); 101 participants did not complete the DCE as they had withdrawn or were lost to follow-up. Young people with ASD were significantly less likely to respond than those with diabetes mellitus [16/87 vs. 7/112 (χ² = 10.50, degrees of freedom = 2; p ≤ 0.001)].

Appendix 7 presents detailed results. In summary, 43 respondents always chose current care and 60 never chose current care. Those not currently receiving any service were significantly less likely to choose current care (p < 0.001). Each attribute and level was often preferred by many young people, as was current care. Table 8 shows that in a service where none of the attributes was present, the uptake of the service would be 78%. Adding a new service attribute, such as flexible appointments, predicted an uptake of 81%.

When demographic characteristics (gender, age at data collection and condition) were taken into account, all attributes, except flexibility of appointments, were preferred to a service with none of the attributes (see Appendix 7, Table 23).

Preference for current care was stronger among young men than among young women. Preference for current care was also stronger among those who had not transferred from children’s to adults’ services than among those who had. Young men had a stronger preference for out-of-hours clinics than young women did, and a weaker preference for making their own decisions about treatment. Young people with ASD had weaker preferences for parental involvement, and for being the ones to make decisions about their treatment, than those with cerebral palsy.

Strengths and limitations

A DCE can be demanding to complete. Although most young people were able to respond, proportionately fewer with ASD did so. It is possible that these young people have different preferences from those who did complete the DCE. Furthermore, it is possible that the DCE failed to accurately capture the preferences of those young people who completed it. On this latter issue, some reassurance was provided by the broad similarities between the DCE findings and the ranking exercise. The advantage of the DCE over the ranking exercise was that the DCE provided an indication of the relative importance of each attribute.

Discrete choice experiments are widely accepted as a methodological approach preferred by many organisations. For example, they form the basis of eliciting EQ-5D-5L (EuroQol-5 Dimensions, five-level version) population tariffs throughout the world95 and the Center for Devices and Radiological Health has stated that DCEs are a suitable method of eliciting patient preferences. 96 The approach we adopted was consistent with best practice. The aids to completion of the DCE might have infringed strict DCE methodology, but they enabled young people, with a wide range of confidence, intellectual ability and flexibility of thinking, to complete it. Indeed, the use of aids to completion has been advocated. 97–99

A fixed-effects modelling assumed that respondents were consistent in their choices over time. Although such an approach was not incorrect, more sophisticated econometric modelling could be used to explore heterogeneity and whether or not participants considered only a subset of attributes when choosing among alternatives. This might help explain why an appreciable proportion of respondents always chose current care.

Key findings

• Young people with long-term conditions, including those with ASD, could complete a DCE.

• In total, attributes were preferred to them not being present, except for flexibility of appointments.

• Preference for current care was strong and more pronounced among young men than among young women. This preference was stronger among those who had not transferred.

• Most young people (and especially women) least valued being able to have appointments outside office hours.

• Young people valued services in which information was passed to the right person, parental involvement was welcomed, staff offered choice and allowed young people to make decisions about their care, the same staff were seen at each clinic appointment and extra support was available for preparation for everyday life. Young people valued seeing the same staff at each clinic appointment more than having a key worker.

Inter-relationship with other parts of the programme

From WP 1.1, UP contributed much to the design of the DCE questionnaire. DCEs have not been undertaken often with young people. The understanding of the choice options was therefore discussed and piloted with UP.

The content of the DCE was, in part, informed by the Q-sort study of WP 1.3.

The DCE was administered at visit 3 in the longitudinal study WP 2.1.

This WP, 2.3.1, informed WP 2.3.2 on economic analysis.

The results of this WP led directly to one of the implications of the programme, implication 7, concerning elements of service that would make it more likely that transition services were taken up by young people.

Aims

• To determine the resource use and costs of the PBFs.

• To assess the relative efficiency of illustrative models of transition.

Methods

Using the same decision rule as in WP 2.1, a trajectory for each PBF, deemed satisfactory or unsatisfactory, was assigned to each participant based on the data captured from the health notes.

Costs to the NHS covered the costs of consultations, medication and investigations. Data needed for costing were obtained by the research assistants from the health records, with additional information on service use gathered at each annual visit after the baseline visit. To aid recall, young people had kept diaries of contacts with health and personal social services. A unit cost was then applied to each service accessed using routine sources. 100,101 Average pre- and post-transfer costs were then calculated.

Costs were then analysed to estimate the annual cost for a service where there was a PBF and for a service where there was no PBF. These costs were further split according to whether the young person was pre or post transfer, and, if post transfer, whether to primary or secondary care.

Young people completed the EQ-5D-Y at each of the four annual visits. This instrument captures HRQoL for use in economic analyses. 102 Data from the EQ-5D-Y were converted to utility scores using the value set for adults. 103 Utility scores range from 0 to 1, with 1 indicating best health. A simple scoring system was also adopted for the EQ-5D-Y. Each of its five domains was assigned equal weight, and each level of each domain was scored 1, 2 or 3, depending on severity. Thus, each domain scored from 1 to 3, with 3 being worst health. Summing the scores for each domain generated a ‘sumscore’ that ranged from 5 to 15, with 15 being worst health. A similar approach to analysing these data was adopted for cost.

The data on costs and HRQoL were incorporated into a simple Markov model (Figure 9) in which all young people start in children’s services, aged 14 years, and are followed for 10 years, the time over which a young person would be expected to be undergoing transition. The model considers a hypothetical cohort of 1000 young people. Each year, they have a probability of transferring from children’s services to primary care or moving into adults’ services. If, in any year, they do not make the transfer, then they remain in children’s services, indicated by the ellipse connected to the children’s services state. The chance of being in a given state in a given year (the ‘rate of transfer’; see Appendix 7) was calculated by considering the number of people who transferred from children’s services each year, and the ratio of young people transferred to adults’ services to those transferred to primary care.

FIGURE 9.

Structure of the economic model.

Those remaining in a given year in the children’s services state are represented by the ellipse connected to the children’s services state.

For each state in Figure 9, an annual cost, utility and sumscore value was assigned as described above. Cumulative costs, quality-adjusted life-years (QALYs) and sumscores were estimated for the 10-year time horizon, and all costs and effects were discounted at the UK recommended rate of 3.5% per annum.

The results of the model and the DCE were combined with findings from other areas of the research programme in a balance sheet analysis. In this analysis, we present those outcomes that could favour a hypothetical intervention compared with a situation in which there was no support during the transition from children’s to adults’ services.

Analysis

Strengths and limitations

A concern was whether or not the HRQoL instrument was sufficiently sensitive to capture an influence of PBFs. The study was powered not on the EQ-5D-Y but rather on the outcomes used in WP 2.1. We did have sufficient power to identify clinically significant changes in the EQ-5D-Y but, for analysis by condition, sample size meant that some real effects of difference may not have been detectable. In relation to the PBFs, power was limited if a feature was often absent.

Alternatively, the HRQoL instrument might have been sufficiently sensitive, but follow-up was too short to capture any influence of PBFs.

The DCE suggested that experiencing some of the PBFs might increase young people’s engagement with health services, which, in turn, might alleviate longer-term harms (and costs) due to inadequate health care. The economic evaluation was exploratory and compared only services with individual PBFs with a service with no PBFs. Therefore, no probabilistic sensitivity analysis was conducted, as is normally recommended for health technology assessment economic evaluations; a quantified estimate of the effect associated with various combinations of PBFs was not possible. Therefore, the economic conclusions should be interpreted with care.

It was necessary to annuitise the cost data for participants to allow a comparison with the EQ-5D-Y data. Thus, certain costs for participants for items mentioned occasionally, but likely to apply over the course of the study, were calculated as a daily rate following the aggregation of the visit data. For example, as insulin prescriptions were not consistently reported in the data, these were costed separately according to existing literature and then inflated to the current price year. We think that this made best use of the available data and is unlikely to have distorted the results.

The data collected on outcomes, costs and health service use were complex to interpret because there was loss to follow-up and differing health-care transfer arrangements, and the intervals between the annual visits by the research assistants were not always exactly 1 year. Strict econometric modelling, despite considerable efforts, proved to difficult to undertake. Therefore, a less robust exploratory analysis was adopted, which drew out key implications and generated a balance sheet. Conclusions were not wholly in concordance with some conclusions from other parts of the programme, but the triangulation of findings allowed key consistent conclusions and implications to be identified and reported.

Key findings

• There were significant differences in both costs and HRQoL between conditions.

• The average utility and sumscores were similar in those pre and post transfer.

• The median annual cost (covering the costs of consultations, medication and investigations) was £7003 for an individual with diabetes mellitus, £2934 for an individual with cerebral palsy and £2350 for an individual with ASD.

• Both costs and HRQoL declined over time for all conditions.

• A service involving a PBF is unlikely to reduce NHS costs over a 10-year time horizon but might increase them.

• We have examined the potential consequences of a service with flexible appointments and that encourages ‘appropriate parental involvement’ and ‘promoting health self-efficacy’. A judgement is needed about whether or not such a service would be worth the extra cost, given that it may increase engagement with health services and prevent deterioration in health in the long term.

Inter-relationship with other parts of the programme

Data collected during WP 2.1 (the quantitative longitudinal study) about costs of consultations and exposure to PBFs were essential for the economics analysis.

The results from the DCE (WP 2.3a) contributed to interpretation of the economic analysis.

The results of this WP led directly to one of the implications of the programme, implication 7, concerning value for money.

Work package 3.1.1

There is a need to provide health care that meets the needs of young people. Youth-friendly health services104,105 is one concept discussed in the literature. However, we need to move from the ad hoc provision of youth-friendly services to youth-responsive health systems106 that respond to young people’s changing developmental needs. 25 DAH recognises the changing biopsychosocial developmental needs of young people and the need to empower young people by embedding health education and health promotion in consultations. In this WP we developed a working definition of DAH:

DAH recognises the changing biopsychosocial developmental needs of young people and the need to empower young people by embedding health education and health promotion in consultations. In operational terms, DAH focuses on health-care professionals’ approach to and engagement with each young person and their carers alongside the structure of the organisations in which care takes place.

DAH offers a foundation for good practice in the health care of young people and so it should underpin transitional health care.

Work package 3.1.2

Aims

• To identify the structures, processes and relationships between commissioning entities in the NHS and other agencies relevant to transitional health care.

• To identify the facilitators of and barriers to commissioning for services for transitional health care.

• To identify how services for transitional health care could be better commissioned.

Methods

This work was overseen by an invited Steering Group, which met on four occasions.

The work was undertaken sequentially in six stages.

• Stage 1: literature review. A systematic review was undertaken to identify published peer-reviewed evidence about commissioning for transition.

• Stage 2: qualitative interviews with stakeholders. Twenty-nine interviews were conducted by telephone (n = 3) or in person (n = 26) with individuals from Clinical Commissioning Groups (CCGs), Health and Well-being Boards, local authorities, general practices and the third sector. Most of the interviews were conducted with people in north-east England. A second site in the Midlands was selected because the local Health and Well-being Board had transition as a priority, and there was a Commissioning Support Unit. Further details of the methods are in Appendix 9, which includes the interview schedule, and in our published paper. 112

• Stage 3: case studies. Using the findings of the literature review and the interviews, criteria were developed for seeking sites at which transition was being formally addressed. Three sites, covered by three CCGs, were visited in the north-west, the south-west and the West Midlands of England. Across the three sites, 11 semistructured interviews were conducted. The details of the methods are in Appendix 9.

• Stage 4: seminar in London on commissioning for transition. This brought together additional expertise and experience from public health, social policy, commissioning and research into commissioning. Thirty invited delegates attended.

• Stage 5: interviews with commissioners for adults’ services. We realised that we had talked with few commissioners for adults’ services. Therefore, we arranged face-to-face interviews (n = 5) with commissioners for adults’ services in two CCGs in north-east England (n = 3), a CCG in south-east England and a regional commissioner for specialised adults’ services for NHS England in the Midlands.

• Stage 6: synthesis and conclusions.

Analysis

Strengths and limitations

Despite increased awareness about transition since the publication of the NICE transition guideline14 and quality standard117 in 2016, commissioning for transition is a new field of enquiry. This means that our research conclusions are primarily based on the evidence generated during the research programme through research interviews, site visits and discussions with commissioning groups.

Given the fast pace of policy change, and with over 200 CCGs across England, one of the challenges of this WP was to capture a wide range of contemporaneous views and perspectives. We think that the combination of the range of methods employed (interview, seminar, site visits), the relatively broad geographical distribution across the UK, and engagement with CCGs and specialised commissioning at NHS England gave us a reasonably comprehensive understanding.

Key findings

• Commissioners and providers of adults’ services regard transition as the responsibility of commissioners and providers of children’s services to deal with and improve.

• Whether contract specifications include transition very broadly or in detail, implementation of transition is not usually then monitored by commissioners – partly because there are no nationally agreed outcome and quality indicators to use.

• Although a CQUIN for transition can raise the awareness of service managers and clinical teams about transition, the short-term requirement for indicators to be documented can make the introduction of a CQUIN a box-ticking exercise, rather than a commitment to organisational change that may in turn lead to quality improvement.

• A Transition Steering Committee and trust-wide transition co-ordinator in a NHS Trust are effective in discharging an organisation-wide commitment to improving transition.

Inter-relationship with other parts of the programme

Our understanding of commissioning processes developed from this WP allowed us to ensure, as far as possible, that we were well informed when we visited commissioners in WP 3.3. We were able to approach commissioners with specific questions and specific suggestions and then test these out. As will be shown in WP 3.3, we then formulated definitive conclusions with specific implications for commissioners. Our final conclusions are relevant, possible to implement and likely to be adopted by commissioners because of this sequential process of identifying issues and then testing them out with commissioners.

The findings of this WP informed the content of the DAH toolkit in WP 3.1b.

Through their effect on WP 3.3, the results of WP 3.2 directly determined what we included and how we framed implications 1, 2 and 6 of the programme.

Aims

• To synthesise learning from the programme on what and how to commission.

• To learn about the most useful way to provide research-based evidence to inform commissioners, through local and national meetings and visits.

Methods

Using results from WP 3.2 on commissioning, from WP 3.1 on DAH and from WPs 2.1 and 2.2 on features of services for transitional health care associated with better outcomes, we developed two brief documents setting out for commissioners the preliminary key research findings and implications of the programme. We wanted to learn from commissioners whether or not we were presenting material that they thought was relevant and useful, and in an appropriate manner. We presented these documents to the following groups and individuals, modifying their contents as we proceeded:

• an academic at a university in south-west England who had undertaken research into evidence-based commissioning and what influences commissioners’ decision-making

• a children’s health commissioner (also general practitioner) in a CCG in north-east England

• chief executives of a CCG consortium in north-east England

• chief executives of a CCG consortium in south-east England

• the Clinical Leadership Committee for specialist services at NHS England

• three telephone interviews with members of the above Clinical Leadership Committee

• a commissioner of specialist services for mental health in north England

• a commissioner of specialist services for long-term conditions in adults in the Midlands.

Thus, we talked with seven commissioners individually and with three committees at which 21 commissioners were present.

During the consultation meetings, we presented the key findings about the outcomes and lessons from the programme, focusing on the commissioning pathway and guidance. We encouraged, through discussion with various stakeholders, critical reflection to ensure the relevance of the emergent findings, identify key lessons learned, and explore the implications of these lessons for developing a nationally consistent and comprehensive approach to commissioning for transition. The discussions were central to challenging, modifying and enhancing our guidance for commissioning for transitional health care for young people with long-term conditions.

We kept notes of the content of all the meetings.

Analysis

Below are summaries of our discussions based on the notes of the meetings. We used an iterative process, based on the advice and suggestions we received, to modify and revise the summary documents and the way we presented our proposals to commissioners.

Strengths and limitations

This is, as far as we are aware, the first work to study commissioning for transition. We were able to identify some major issues for commissioning, and gain an understanding of the context in which commissioners work and the parameters they use to direct health-care provider organisations in commissioning transition.

As most commissioners we talked with agreed that transition was important and that now was an opportune time to introduce change, it would have been useful to meet with commissioners who were more sceptical about the importance of transition.

Our presentations to commissioners focused on a small number of implications for commissioners; these were refined as discussed above. This inevitably meant that we could not share the detail and complexities of all of the findings of the research programme.

We have published two papers on commissioning112,113 and we are preparing a third.

Key findings

Box 2 shows the final version of the document, after all of its modifications, that we discussed with commissioners. It is also the final version of the document for commissioners that we presented at the dissemination meeting at The King’s Fund in October 2017.

Inter-relationship with other parts of the programme

The summary of findings for commissioners in Box 2 includes proposals about DAH that were explored in WP 3.1 and about the PBFs that were explored in the longitudinal study WPs 2.1 and 2.2.

What we learned about the commissioning process in WPs 3.2 and 3.3 has directly informed our conclusions about how commissioning might be improved, and directly informed how best to present this evidence to those responsible for commissioning health care in different NHS settings.

This WP directly determined what we included and how we framed implications 1, 2 and 6 of the programme.

Integration of work packages

We achieved our aim to have a set of nine WPs that were closely integrated. Our conclusions and their implications were informed by the input of young people, the longitudinal study and the work on implementation in trusts and with commissioners.

Patient and public involvement

Our PPI was of high quality. We formed this view based on the interest NIHR showed in highlighting the work of UP on its website, the invitation from the Department of Health and Social Care to UP to attend the ‘take over’ day, the views of the External Advisory Board and the internal evaluation undertaken by Gail Dovey-Pearce (reported in Involvement of patients, the voluntary sector and the public and in our published paper45).

Small number of feasible implications

We succeeded in our intention to conclude with a small number of feasible implications that will have utility for NHS commissioners and for those responsible for the delivery of clinical services. During consultations, we were advised that documents containing large numbers of recommendations and aspirational statements are difficult for commissioners and chief executives to use.

Administrative challenges

External evaluation of real world rather than internal evaluation of a locally developed intervention

Most groups undertaking research into transition evaluate their own interventions; essentially, they undertake an audit or clinical evaluation of their service. However, we have successfully completed a hypothesis-driven, longitudinal, observational study of features being provided, or not, across 27 NHS Trusts (35 services). We had a large cohort, and this was the first study of its kind.

Prospective data collection during transition

We captured data from young people while they were in transition and across their point of transfer. Therefore, we have information about young people’s experiences and outcomes at the time they were receiving services, rather than on what they thought afterwards about what they had or should have received. This has rarely been done before and gives us confidence that our conclusions will be helpful to young people who are in the midst of transition.

Seeking findings relevant to a broad range of long-term conditions

We sought to identify the features of good transitional health care that were appropriate to those with any long-term condition. Although we selected specific conditions for the longitudinal study, they were deliberately chosen as conditions that give rise to a wide range of different health needs, psychosocial complexity and availability of adults’ services.

Engagement with commissioners

Our examination of commissioning is a novel aspect of the research programme. Arrangements for commissioning NHS services have continued to change over the period of the research programme and will continue to do so. However, the commissioning focus of our research concerns ‘what’ and ‘how’ to commission. Thus, our findings should be relevant to any commissioning structure, including whether or not commissioning and service provision are brought together in the same organisation.

Outcomes and processes measures (indicators) of transition

The range of outcomes we chose was informed by the International Classification of Functioning, Disability and Health: Children and Youth Version73 and, after discussion with international transition researchers, conformed to many of the recommendations of later international Delphi surveys,42,43 as set out in Table 14.

Subjective well-being and participation in life (e.g. social, educational, workplace) are generic outcomes (i.e. they are not specific to a particular condition). Well-being has been emphasised as an important and often neglected outcome in the evaluation of transitional health care. 128

Few studies have addressed such generic, distal outcomes. Assessing such outcomes brings some attendant difficulties:

• Poor adult outcomes are not necessarily due to a failure of transition. For example, in complex physical problems such as cerebral palsy, participation outcomes for adults are particularly poor for those with more severe impairment,129 yet participation is already much reduced in 8- to 12-year-old children with cerebral palsy. 125

• Many chronic illnesses such as diabetes mellitus, renal failure and cystic fibrosis generate more medical complications the longer the individual has the condition. Thus, even when process measures of optimal care may be satisfactory, health outcomes may worsen.

Process measures of transitional health care usually assume that providing services that young people like and maintaining a young person’s contact with health services will yield better outcomes. However, although Betz and Smith130 recognised that such measures may be relatively easy to collect, we do not know if they lead to an improvement in more distal outcomes.

Non-attendance at appointments might seem an easily captured and interpretable process measure, but we found that it is not. Many of the reasons for non-attendance are logical and valid, rather than indicating disengagement with services. In addition, rating by number or percentage of missed appointments is difficult to interpret. The number of expected appointments per year varies by condition (individuals with diabetes mellitus may expect at least four per year) and by disease activity (e.g. is the appointment for a routine follow-up or because of a worrying deterioration of disease control?).

In investigating relationships between process measures and outcomes, Cramm et al. 131 found that satisfaction with transitional care was associated with better social and emotional quality of life 1 year later in those with diabetes mellitus or neuromuscular disorders. Sattoe et al. 132 found that process measures such as ‘patient not lost to follow-up’, ‘attending scheduled visits’ and ‘satisfaction with care’ were not related to social participation. In their study, ‘continuing attention to self-management’ was the only process measure associated with better HRQoL.

Quality of delivery of proposed beneficial features

The nine PBFs that we studied were carefully defined and captured but we could not assess the quality with which each feature was delivered. For example, ‘meeting the adult team before transfer’ might have been tokenistic in some trusts and carefully planned in others. However, we examined exposure to each PBF across many trusts and on three occasions, and we think that much of any variation in quality will have been ironed out by the statistical method we adopted.

Longitudinal representation of proposed beneficial features

It was relatively straightforward to examine the association of exposure to each PBF over the previous year with generic and condition-specific outcomes. However, it was more difficult to decide how to represent each PBF over the duration of the longitudinal study. What constitutes optimal exposure over 3 years when, for instance, exposure occurs in one year but not in the others? We made a pragmatic, clinically informed judgement about this for each feature. The decision tree (WP 2.1) was agreed by the team of the Transition programme, but others might have come to different judgements.

Social participation

After 1 year, we realised that the Rotterdam Transition profile was not sufficient to capture some aspects of social participation. We added a social participation questionnaire, but this could be administered only at the third and fourth visits, thereby limiting our longitudinal analysis of this outcome.

Self-reported questionnaires

Most outcome data are from self-reported questionnaires. We recognise that this may be influenced by individual reporting styles, which make valid comparisons more difficult. However, we used questionnaires that were well validated; we had a large sample size, which tends to minimise the impact of variation due to reporting style; the questionnaire completion was manualised and repeated each year so that any reporting styles would be consistent for each participant; we used a number of different instruments with score ranges suitable for group statistical analysis; and, finally, the constructs about which we were largely concerned were about young people’s perceptions of their health, well-being and social participation.

Statistical power

In line with our original calculations, we had sufficient power to identify clinically significant changes in the outcomes. However, for analysis by condition, the sample size meant that some real effects of difference might not have been detectable. In relation to the PBFs, power was limited if a feature was often absent.

Economic modelling

The data collected on outcomes, costs and health service use were complex to interpret because there was loss to follow-up and differing health-care transfer arrangements, and the interval between the annual visits by the research assistants was not always 1 year (young people have a lot going on in their lives and visits often had to be rescheduled). Strict econometric modelling, despite considerable efforts, proved more difficult to undertake as a result of the nature of the data. A less robust exploratory analysis was adopted. This was able to draw out some key implications and generated a balance sheet. This was not wholly in concordance with some conclusions from other parts of the programme, but the triangulation of findings allowed key consistent conclusions and implications to be identified and reported.

With respect to young person involvement

Our work represents the views of young people because we consulted over 5 years with the UP group; UP disseminated findings to young-person groups across the UK; in the longitudinal study, young people with long-term conditions completed the questionnaires; and in the qualitative research work, young people were interviewed.

With respect to the longitudinal study

This is the largest sample for research in the UK in terms of examining transition longitudinally and collecting hypothesis-driven data from young people at home visits (compared with data-linkage studies from administrative data sets). Although we cannot be certain of representativeness and, therefore, generalisability across UK regions and NHS Trusts, we studied individuals from Greater London, north, south-east and south-west England, and Northern Ireland. Furthermore, we drew individuals from 27 NHS Trusts, and these trusts varied greatly in the number and variety of the PBFs they offered.

We recruited from all young people with cerebral palsy in two population registers and all young people with diabetes mellitus or ASD and an associated mental health problem in nine NHS Trusts widely distributed across England (all such young people are seen in secondary care). Thus, complete populations of individuals with one of the three conditions were sampled from (e.g. rather than those attending particular schools, specialised tertiary health-care services or voluntary support groups).

The three conditions chosen were exemplars of chronic illness (diabetes mellitus), complex physical impairment (cerebral palsy) and neurodevelopmental impairment (ASD); these were chosen deliberately to provide variation in health needs, psychosocial complexity and availability of adults’ services. Thus, we consider that generalisability to other conditions is likely. This is supported by the fact that studies of young people with specific conditions usually conclude that the important issues are generic rather than about the specific disease (e.g. diabetes mellitus,133 inflammatory bowel disease,134 cystic fibrosis,135 attention deficit hyperactivity disorder,136 urology137 and disabilities138).

However, as intended in our protocol, we did not include individuals with a significant learning disability, and our findings cannot generalise to this group. Nor did we include those with degenerative conditions. If progression is slow, as for instance with cystic fibrosis, transition is similar to that in other conditions and a formal transfer of health-care needs to take place. If progression is rapid, as for instance with some brain tumours, then the wider social and educational aspects of transition take place very differently. Furthermore, it may not be appropriate for transfer of health care to occur because it would be unnecessarily disruptive, with the young person likely to become incapacitated and/or die fairly soon after transfer; rather, advance care planning should be in place, as recommended by Together for Short Lives. 139

The 374 participants did not differ significantly from those declining to take part, other than for a small effect of deprivation. The distribution of severity of the three conditions was similar to that in national samples. 62 Those with the conditions had a wide range of severity; for example, the young people with cerebral palsy ranged from wheelchair users to those with independent ambulation. Attrition did not appear to create a bias, as there were no significant differences between those remaining and those not remaining in the study with respect to sex, age, condition, diabetes site or ASD site. In Northern Ireland, there was a small effect of deprivation on the attrition of those with cerebral palsy. Thus, we think that, after non-participation and attrition, our study group continued to be representative of those we aimed to recruit.

With respect to commissioning

Our work is novel. We interviewed, carried out case studies and consulted widely with commissioners and senior managers, CCGs and specialised commissioning, and regional and national specialised commissioning, and in many locations across the UK. Thus, we think that our findings are generalisable to the UK and to other countries with well-developed health-care provision.

With respect to the economic analysis

As we state in our account of WP 2.3, strict econometric modelling, despite considerable efforts, proved too difficult to undertake because of the nature of the data, and a less robust exploratory analysis was adopted. This drew out some key implications and generated a balance sheet. This was not wholly in concordance with some conclusions from other parts of the programme, but the triangulation of findings allowed key consistent conclusions and implications to be identified and reported.

Policy context

In Setting the scene at the start of the Synopsis, we summarised the policy context that led to our application for a programme of research. Since then, there have been a number of key developments in England. NICE guidance on transition was published in 2016. 14 A national CQUIN for transition of those with mental health problems35 was introduced by NHS England in 2018. Guidance on commissioning for transition for individuals with special educational needs was published by NHS England in 2018. 140 Following the death of an individual with ASD just before the transfer of their care, in July 2018 the Healthcare Safety Investigation Branch published a report that included recommendations for improving the services for young people with mental health problems in transition. 141

In 2018, the Chancellor of the Exchequer indicated that there would be considerable financial investment in the NHS over the following 10 years. We were asked, based on our research findings, to contribute a proposal for improving transitional health-care practice. Proposals about transitional care were subsequently mentioned in five places in the 2019 NHS Forward Plan. 142

Commissioning for transition

Our literature review (see Appendix 9) found no published or grey literature about commissioning for transition. Our discussions with commissioners showed that transition has been regarded as the responsibility of children’s service commissioners and children’s service providers. This is inappropriate as transition extends to approximately the age of 24 years. Commissioners also tend to consider the numbers of young people in transition to be small; this seems to be because commissioners may be actively engaged with only the small number of very complex cases of young people who require bespoke packages of care integrated across health care, social care and education. Commissioners also reported that a lack of evidence made it difficult for them to be prescriptive about transitional health care. We hope that our research programme will help in this respect. Similarly, commissioners found it difficult to be prescriptive because they lacked indicators with which to assess compliance. Development of indicators is needed, and a workshop on indicators for transition was led by NHS England at the meeting in London in October 2017 at which we disseminated our complete findings for the first time.

Engaging with clinical services: ensuring continuity

Our qualitative work in WP 2.2 found that the point of transfer (whatever the adult provision might be) is intrinsically disruptive for families and young people. Families and young people will have been familiar, often for many years, with how to navigate children’s services, who to talk with if worried, where to get medication from and so on. Transfer disrupts their knowledge and relationships, and thus their confidence and trust, not because the adults’ services are inherently better or worse but because of the change that, at first, families and young people do not fully understand or feel confident to engage with. Thus, there is a need to move away somewhat from asking how young people are helped to mature and adjust, to asking how systems and processes for young people and their families might enable them to regain their knowledge, confidence and trust as quickly as possible. These processes might include ensuring that parents remain closely involved if the young person wants this (as they provide continuity); and meeting with the adult team before transfer (which may remove some uncertainty, especially if this meeting introduces the professional whom the young person will be seeing in future). These are two of the PBFs that were associated with better outcomes in WP 2.1. The above interpretation reinforces the conclusion of a major study by Allen et al. ,58,145 albeit that it focused solely on those with diabetes mellitus.

Improving transition across a NHS organisation

Several of the trusts we visited, or whose managers we talked with in WP 3.2, had a Transition Steering Committee, chaired or convened by a trust-wide transition co-ordinator. These arrangements were effective in bringing about change across adults’ and children’s services and across specialties. They took advantage of the skills and enthusiasm of those already providing good practice and they organised trust-wide training that brought together those from different specialties and from adults’ and children’s services.

Young people’s approach to transition

In WP 1.3, the Q-sort study identified four interaction styles that young people adopted when approaching transition: ‘laid back’, ‘anxious’, ‘autonomy-seeking’ and ‘socially oriented’ (welcoming support from, and frequent discussions with, family, friends and health-care professionals). Discussing and understanding young people’s views and preferences about and during transition would be likely to help clinicians and young people develop personalised planning for transition, and effective engagement with adults’ care.

In the discussions led by UP and the Council for Disabled Children with other young person groups (see Involvement of patients, the voluntary sector and the public), this finding was validated. Young people readily identified with one of the styles. However, some also said that their style had changed; for instance, when they were younger they might have been ‘anxious’ but some years on they were now ‘seeking autonomy and being in control’.

The literature on asking young people about their transition is large (including systematic reviews44,51) and we did not want to repeat such studies. However, ‘Q-sort’ methodology has not been used before in transition and, in identifying these four interaction styles, we have added new findings to the literature.

Changes in outcomes in the longitudinal study (work package 2.1)

Average satisfaction with services was reasonable at baseline (rating of ideal service only slightly higher than rating of actual service). It remained steady for those with diabetes mellitus but worsened significantly for those with ASD or cerebral palsy across the study period. The importance of transition has been recognised by diabetes services for over a decade, and the Diabetes Tariff146 includes transition. It is probable that, in services for ASD and cerebral palsy, the transition process is less streamlined, with consequent reductions in young people’s and parents’ satisfaction with services.

For those with diabetes mellitus or cerebral palsy, their average subjective well-being at baseline was similar to that of the general population and remained so during the study. Those with ASD and an associated mental health problem reported significantly lower well-being than the other two groups and this difference persisted over the course of the study – almost certainly due to the impact of their mental health problem.

For participation in life activities, there were significant differences between the three conditions, with young people with diabetes mellitus in more independent phases of transition than those with ASD or cerebral palsy. All made progress over time but those with diabetes remained more independent.

Condition-specific measures worsened over the course of the study for those with diabetes mellitus or cerebral palsy. The proportions of young people with ASD with ‘abnormal’ anxiety and depression scores remained the same from baseline to the ‘final’ visit.

Proposed beneficial features

We found that the provision of our nine PBFs in 2013/14 was patchy. Fewer than half of the health services stated that they provided ‘age-banded clinics’, ‘transition plans’, ‘transition manager for clinical teams’, ‘protocols for promoting health self-efficacy’ or ‘holistic life-skills training’ (all but ‘transition manager for clinical team’ are recommended by NICE14). Features were provided less often in services for those with cerebral palsy or ASD than for those with diabetes mellitus. Thus, there is much room for improvement across all of the services attended by the young people recruited to this UK-wide study.

Three PBFs were consistently associated with better outcomes in the longitudinal study WP 2.1: ‘appropriate parent involvement’, ‘promotion of health self-efficacy’ and ‘meeting the adult team before transfer’. We advise that strong effort be directed to maintaining or introducing them to a high standard.

In WP 2.1, ‘appropriate parent involvement’ was found to be associated with outcomes concerning satisfaction with services, well-being, some aspects of participation and meeting the health needs of those with cerebral palsy. Interviews in WP 2.2 revealed how important parents were to the process of transition. The involvement of parents as young people think appropriate is a crucial element of DAH. In WP 2.3.1, the DCE showed that ‘appropriate parent involvement’ was the second most highly rated choice. WP 2.3.2 (economic modelling) concluded that ‘appropriate parent involvement’ would be likely to maximise service uptake and may represent good value for money.

Allen et al. ’s147 study of diabetes and transition emphasised the importance of parents during transition. Heath et al. ’s83 systematic review of studies of parents’ perceptions of their role in transition concluded:

Parents can be key facilitators of their child’s healthcare transition, supporting them to become experts in their own condition and care.

Heath et al. 83

Two other recent reports148,149 investigated the parent/young person dyad and reached similar conclusions to ours: that the parent and young person need to share care but that this dynamic will continuously change. NICE14 emphasises ‘appropriate parent involvement’ throughout its report; it was an overarching principle in section 1.1.1, and sections 1.2.19–1.2.22 were devoted to the involvement of parents.

In WP 2.1, ‘promotion of health self-efficacy’ was found to be associated with outcomes concerning satisfaction with services and meeting health needs of those with cerebral palsy. Interviews in WP 2.2 revealed that both parents and young people thought that this was a central element of transition. The interviews also suggested that health self-efficacy may be not only about managing one’s condition but also about how to manage one’s health services. ‘Promotion of health self-efficacy’ is a central component of DAH. WP 2.3.2 (economic modelling) concluded that ‘promotion of self-efficacy’ would be likely to represent good value for money.

Klaasen et al. 150 developed a health self-management scale. Sattoe et al. 151 found that ‘continuing attention to self-management’ was associated with better HRQoL. There is conflicting evidence about whether a structured approach, including motivational techniques for raising health self-efficacy in diabetes mellitus, influences glycaemic control. 152,153 Mackie et al. 154 showed the benefit of a 1-hour, nurse-led intervention to promote knowledge and confidence about one’s condition, in this case congenital heart disease. NICE14 (in section 1.2.17) recommends ‘promotion of health self-efficacy’.

In WP 2.1, ‘meeting the adult team before transfer’ was found to be associated with better outcomes in participation and in diabetes management. There was a negative association between having met with an adult team and one of the Mind the Gap subdomains. We infer that lower satisfaction was due to the young person’s lack of familiarity with the adult team. Interviews in WP 2.2 found ‘meeting the adult team’ to be important, although sometimes the arrangements rendered it ineffective because the young person could not attend if the opportunity was only offered once per year, or the young person met someone who was subsequently not responsible for their care. If undertaken well, the meeting really helped with the potentially disrupting and disorientating consequences of transfer described in WP 2.2. WP 2.3.2 (economic modelling) concluded that evidence of the value for money of ‘meeting the adult team’ was ambiguous.

Our definition of meeting the adult team before transfer included having clinics where adult and paediatric physicians consulted jointly. In other studies, such joint clinics have shown to lead to improvements in process indicators in renal services,155 diabetes services,156 urology services137 and rheumatology services. 157 NICE14 (in sections 1.3.5 and 1.3.6) recommends ‘meeting the adult team before transfer’.

One PBF, ‘having a key worker’, was significantly associated with greater service satisfaction only during the second study period (between visits 2 and 3); it had no associations with longitudinal analysis and some negative associations with participation.

Sloper et al. 158 found strong evidence for the value of key workers for families with children with special educational needs. NICE14 (in sections 1.2.5–1.2.10) calls for a ‘named worker’.

The difficulty with ‘key worker’ may be the practicality of operationalising this rather than the principle of having one. Staff changes as a result of leaving post, restrictions in job plans, service restructuring, or sickness or maternity leave make it difficult to provide a key worker for all young people with long-term conditions, even though it may work well for the few for whom there is continuity.

Five PBFs had few significant associations with better outcomes: ‘holistic life-skills training’, ‘transition plan’, ‘age-banded clinic’, ‘having a transition manager for clinical team’ and ‘co-ordinated team’. WP 2.3.2 (economic modelling) did not find any benefit in terms of costs or HRQoL. Nor were the last three recommended by NICE. 14

Regarding ‘having access to holistic life-skills training’, our qualitative work in WP 2.2 found that all parents and young people thought that signposting to educational and social services and to charities was essential. Some services were more ambivalent, and were not even sure if it was their responsibility to probe these wider areas of a young person’s life.

‘Having a written transition plan’ is recommended by NICE14 (in section 1.3.4) but it does not specifically mention that the plan should be a written document. Qualitative work in WP 2.2 found conflicting views about a ‘written’ plan: such plans take time to draw up, they become out of date and they get lost, and who should hold them? Some professionals said that personal interaction was far more important. On the other hand, the lack of a formal plan left many families disorientated and wondering if the service did not have the resources to provide for care after transfer.

These last six PBFs, for which we did not find evidence of benefit, were included in our programme because there had been a number of small studies, typically internal evaluations of local interventions, that suggested benefit (e.g. benefit of a transition manager for clinical team159). We acknowledge that the absence of evidence is not the same as lack of benefit. However, our study was hypothesis driven and had clear, pre-identified outcomes. Even if these five PBFs should not be rejected, our findings indicate that resources (which include training) would be best directed first to embedding into service provision the three beneficial features for which we found convincing evidence of benefit.

Health economics

The short-term cost of introducing the three key PBFs that we found to be associated with better outcomes will depend on precisely how the feature is delivered. If a NHS organisation intends to adopt a trust-wide approach to transition, including ensuring that services are developmentally appropriate, then it could be advantageous for them to consider allocating sessions to a co-ordinator. Such a role could involve chairing a Transition Steering Committee and assisting directorates and teams of health-care professionals and providing instruction in DAH. The toolkit we have developed could assist this process (see Appendix 8).

Arranging for every young person and their parents to meet an adult team member before transfer will require organisation and clinic time. Where there are already clinics at which adults’ and children’s clinicians consult together, such meetings happen automatically. However, where care is being transferred to primary care, a meeting attended by a member of the children’s services team may enable this feature and support the young person’s confidence to make further appointments.

The DCE (WP 2.3.1) suggested that, over a 10-year time horizon, a service allowing appropriate parental involvement, where the same staff are seen at each clinic, and where young people have the opportunity to make decisions about care, would maximise uptake. A service involving ‘appropriate parent involvement’ and ‘protocol for promotion of young person’s confidence in managing their health condition’ may represent good value for money, but a service involving a ‘transition manager for clinical team’ or ‘age-banded clinic’ might represent less value for money. However, the results are tentative as the economic evaluation was exploratory. The balance sheet highlighted that adoption of the new approaches to transition, outlined above, needs to balance the extra cost against the increased engagement with health services and prevention of deterioration in health in the long term.