To the Editor: The literature pertaining to the transition of care from paediatric to adult services in type 1 diabetes mellitus has only recently included attempts to establish high quality data. Indeed, a 2016 Cochrane review identified only four randomised controlled trials (n = 238) of interventions designed to improve the transition of care for adolescents with chronic medical conditions [1], including one study (n = 26) relating to the implementation of a transition coordinator [2].

We therefore read with interest the results of the pragmatic clinical trial by Butalia et al [3], which reported a reduction in absolute losses to follow-up in a cohort of young adults with type 1 diabetes who were exposed to a communication technology enhanced transition coordinator. The intervention was associated with a fourfold reduction in absolute disengagement from specialist medical services in the first 12 months post-transition, compared with a historical cohort [3]. This paper adds to the transition literature base, but a robust process of evaluation would allow understanding of the factors that would facilitate roll out of this intervention at scale. We need to specifically understand the uptake and impact of communication technology and web-based supports. Whilst communication preferences and numbers of communications are reported, the level of self-initiated interaction or engagement was not described.

Details on the most utilised aspects of the communication transition coordinator’s role would be useful for future planning. This is particularly important given that it is difficult to interpret the relative benefit of individual elements within multifaceted interventions. Combined medical and non-medical appointment attendance was included as an outcome measure in this study [3], but detail around the timing of these appointments (same day or otherwise) is not provided, and therefore it is unclear how attendance was truly affected by transition coordination. Considering that intervention participants are reported to have had a mean of 1.4 HbA1c measurements over 5.8 attendances in the 12 months following transition [3], it may be assumed that participants were seen by multiple members of the multidisciplinary team on each of 2 separate attendances for physician review. If this assumption is incorrect, detail around the nature of clinic attendances is vital. It is also interesting to note there was no increase in the absolute number of attended clinics between 12–18 months in either the control or intervention group [3]. Moreover, dismissal of randomised controlled trials as a source of recruitment bias should be cautioned, as this study design remains the gold standard for evaluation of novel interventions. Pre- and post-intervention studies introduce potential confounders such as centre changes in attitude or practice. While the authors are explicit in stating that no processes were altered within the paediatric setting during this trial, subtle changes post-transition may have been introduced as a consequence of this study, which cannot be measured or accounted for.

As the authors have noted, an increasing number of studies have reported outcomes related to transition coordinators in type 1 diabetes [2,3,4,5,6,7] (Table 1). Similarities across these studies include comparable primary outcome measures, such as absolute disengagement from specialist services, frequency of clinic attendance and change in HbA1c across the transition process. The role of the transition coordination has been noted to have limited, if any, impact on somatic outcomes [8], which, specific to type 1 diabetes, means that increased attendance as a result of transition coordination does not result in a reduction of suboptimal glycaemic control [2,3,4,5,6]. However, it should be noted that elevations in HbA1c commonly predate transition, with its origins in paediatric care. Therefore, targeted pre-transition interventions may more appropriately address this issue. There are notable points of difference in the findings across these studies. Rates of absolute disengagement or loss to follow-up in the first year post-transition appear to be higher in Canada [3, 4] than those noted in the Australian context, for example, where disengagement tends to occur in the second 12 months following transition [2, 5]. Potentially, the design of future intervention studies should be adapted to local context.

Table 1 Description and comparison of transition intervention studies in type 1 diabetes mellitus
Full size table

As a result of publication bias, typically only novel findings are published and highlighted, which limits meaningful comparison across similar studies. Transition to adult care is a field of medicine which is in its infancy in terms of prospective and randomised data. Authors should be encouraged and supported to acknowledge the positive aspects of previous studies, while maintaining constructive criticism of issues with study design. More so than in any other research domain, transition intervention trials are conducted in real-world clinic situations. Post-transition follow-up within a tightly controlled research domain is very difficult, and participants frequently move between specialist services [1, 2]. This inevitably invites criticism of study design, and yet transition of care is an issue where any attempt to manipulate the nature of post-transition clinical follow-up would not be sustainable or possible to replicate more widely. Whilst primary outcomes are of course important, relevant findings relating to secondary outcomes cannot and should not be completely ignored. Given the lack of standardisation in approach, definition of success or core outcome measures, rigorous comparison of data outcomes is crucial so that clinical practice and future research can move forward effectively.

This recent paper from Butalia et al [3] is a positive contribution to the growing evidence that coordination of care across the transition pathway for young adults with type 1 diabetes has beneficial effects in terms of reductions in absolute losses to follow-up in the short term. What remains untested is the optimal dosage and nature of exposure to this intervention (i.e. when should it start, how long should it be provided for, is it more appropriate to be driven from within the paediatric or adult setting?) and an understanding of how context impacts outcomes. Further, a robust method to identify and directly engage with those young people who are most likely to disengage from medical care remains intangible. Finally, the design of a consumer-driven intervention which can be provided in a manner that empowers and equips the young person for ongoing and lifelong healthcare transitions will ultimately embody the elusive gold-standard model of care for transition.