Nothing About Us Without Us: A Scoping Review and Priority-Setting Partnership in Type 1 Diabetes and Exercise

Objectives: Examine the characteristics of patient engagement (PE) practices in exercise-based randomized trials in type 1 diabetes (T1D), and facilitate T1D stakeholders in determining the top ten list of priorities for exercise research. Design: Two methodological approaches were employed: a scoping review and a modified James Lind Alliance priority-setting partnership. Methods: Published (Medline, Embase, CINAHL, and Central databases) and grey literature (www.clinicaltrials.gov) were searched to identify randomized controlled trials of exercise interventions lasting minimum four weeks and available in English. We extracted information on PE and patient-reported outcomes (PROMs) to identify if patient perspectives had been implemented. Based on results, we set out to determine exercise research priorities as a first step towards a patient-engaged research agenda. An online survey was distributed across Canada to collect research questions from patients, caregivers and healthcare providers. We qualitatively analyzed submitted questions and compiled a long-list that a twelve-person stakeholder steering committee used to identify the top ten priority research questions. Results: Of 9,962 identified sources, 19 published trials and 4 trial registrations fulfilled inclusion criteria. No evidence of PE existed in any included study. Most commonly measured PROMs were frequency of hypoglycemia (n=7) and quality of life (n=4). The priority-setting survey yielded 194 submitted research questions. Steering committee rankings identified 10 priorities focused on lifestyle factors and exercise modifications to maintain short-term glycemic control. Conclusion: Recent exercise-based randomized trials in T1D have not included PE and PROMs. Patient priorities for exercise research have yet to be addressed with adequately designed clinical trials.


Introduction
Exercise provides numerous health benefits for individuals with type 1 diabetes (T1D) 1 and is an important component of diabetes self-management 2 . Despite vast health benefits, only one third of people with T1D meet minimum recommendations for regular exercise to achieve health benefits 3 . The unique barriers to exercise for people with T1D 4,5 are severe, particularly loss of glycemic control and hypoglycemia. With few evidence-based strategies available to overcome these barriers, novel approaches are needed to improve the efficacy of future exercise trials to address patient-relevant concerns.
Including patients in designing and delivering research studies can help address patientrelevant gaps in clinical research 6,7 such as understanding barriers to uptake of exercise among people with T1D. Patient-oriented research, being "a continuum of research that engages patients as partners, focusses on patient-identified priorities and improves patient outcomes" 8 , is becoming a priority within clinical trials, but has had little traction in exercise and T1D science 9 . Individuals with T1D have previously been involved in a range of patient engagement [10][11][12] (PE) or prioritysetting activities 13 to optimize blood glucose self-management and overall health. Notably, these studies have not centred on exercise research. The current status of PE in setting priorities for and conducting research within exercise science for individuals with T1D remains relatively unknown.
This study addresses these gaps in the literature. First, we conducted a scoping review of exercise training randomized trials for people with T1D to map patient engagement within recent trials. Informed by these results, we then engaged people with T1D, caregivers and healthcare providers in conducting a modified James Lind Alliance (JLA) model 14 of research prioritization to identify the most important questions about exercise and health.

STUDY DESIGNS AND RESEARCH METHODS Study 1: Scoping Review of Physical Activity/Exercise Randomized Trials and Type 1 Diabetes
We conducted a scoping review of published and grey literature available from the past twenty years to determine in a single narrative analysis: (1) the characteristics of exercise training interventions delivered to people with T1D and (2) the extent patient partners or patient-reported outcomes (PROMs) were involved in study development. The primary question guiding this review was "Is there evidence of patient perspectives being implemented in developing or implementing long-term exercise training trials for individuals with T1D?" 4 This review was conducted following the five-stage Arksey and O'Malley framework 15 and formatted in accordance with PRISMA Scoping Review reporting guidelines 16 . A review protocol was not published prior to its conduct. Our team conceptualized PE as per the Canadian Institutes of Health Research definition as being "meaningful and active collaboration in governance, prioritysetting, conducting research and knowledge translation" 8 .

Data Sources
Information for this review was collected from published and grey literature. Additionally, the Clinical Trials online registry (www.clinicaltrials.gov) was searched to identify ongoing trials (grey literature).
Included trial registrations satisfied the same inclusion criteria as published literature according to information provided. Registrations with related publications were added to the published literature analysis; otherwise, detailed aims and protocols were recorded.

Inclusion/Exclusion Screening
We included randomized controlled trials of exercise training for individuals with T1D, limiting to interventions lasting four weeks or longer to exclude lab-based acute exercise studies, which are common in exercise science. Interventions providing education to support behaviour change without directly implementing an exercise program were excluded. Full-text sources had to be available in the English language.
Screening of published literature occurred in duplicate. Two reviewers independently screened titles and abstracts from the initial (NK and AM) and updated searches (NK and NB).
Conflicts following independent screening were resolved through discussions between reviewers.
Full-text versions of potentially eligible articles were searched and uploaded to Rayyan software.
Full-text screening was undertaken by both reviewers concurrently. The principal investigator (JM) was consulted throughout screening where disagreement remained after reviewer discussions.

Data Extraction
. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org /10.1101/19006452 doi: medRxiv preprint Publications were randomly divided between two co-authors to independently extract data (NK and AM). Where further information was required 18,19 , corresponding authors were contacted electronically. A data extraction form was developed in Microsoft Excel for all published and grey literature data to identify: publication information, participant characteristics, intervention details (frequency, intensity, type, time and intervention duration) and measured outcomes (extracted as per reporting within each study). Reviewers noted evidence of patient engagement if authors declared involvement of people with T1D in research question selection, study design, recruitment, data collection, data analysis and interpretation, or manuscript preparation. Finally, reviewers recorded whether PROMs were measured, as a proxy for reflecting patient-relevant research interests. PROMs included quality of life, diabetes distress, perceived competence, problem areas, self-management behaviours, frequency of glycemic symptoms and several core outcomes identified by the Irish D1 Now Study 12 .

Study 2: A Priority-Setting Partnership for Research in Type 1 Diabetes and Exercise
Following the scoping review, our team conducted a priority-setting partnership with patients, caregivers and healthcare providers living or working with T1D. We adapted the JLA model of priority-setting, which is supported by the Cochrane Collaboration 20 . This study was approved by the University of Manitoba Health Research Ethics Board (H2018:187) and is reported in accordance with GRIPP2 reporting standards for patient and public involvement in research 21 .
The JLA approach to priority-setting is a multi-stage, mixed-methods research design 14,22 , which we modified to identify priorities for exercise and T1D.

Phase 1:
We created an online survey using REDCap Surveys server hosted at the University of Manitoba 23,24 to collect responses to the item, "What questions about physical activity and type 1 diabetes would you like to see answered by research?" Four open-text response boxes were available for respondent submissions. Demographic (age, province of residence and relationship to diabetes) and related patient information (current age, age of diagnosis, gender and ethnic identity) were also collected. The survey was distributed across Canada for six months through partnerships developed with diabetes advocacy organizations (JDRF, Diabetes Canada and Diabetes Action Canada), social media advertising, and posters in diabetes clinics or wellness centres in several urban centres. Concurrently, 12 individuals (eight patients, three caregivers and four healthcare providers) were recruited via maximum variation sampling methods 25 to participate in a steering committee. This steering committee was responsible for prioritizing submitted questions. Each committee member reviewed the list and ranked their top 10 questions in order of one (most important) to ten (tenth most important). Rankings were returned after the two-week review period by email in word documents encrypted with personalized passwords for each member. Rankings were collated through an inverted points-based system whereby top-ranked questions of each member were denoted ten points, and each successive ranking received one less point. Total points for each question were summed and each question receiving ten or more points (in keeping with the JLA T1D partnership) was short-listed for further prioritization. Committee members were asked to individually select, in no particular order, their top and bottom three questions from the short-list.
The steering committee was then divided into three groups, each with equal proportions of representative stakeholders. Printed cards for short-listed questions were presented to committee members in categories anonymously reflecting their independent selections (ie. consensus important, less important, mixed ratings or no ratings). Each group ranked all short-listed questions in order of importance. A collated ranking from all groups was presented to members, and the full steering committee was divided into three different groups. These new groups discussed the list and re-ordered printed cards, if necessary.
The final top ten list was presented to the full steering committee. Each member anonymously rated their level of agreement from one to ten. Consensus on the final top ten list was not reached after the first two rounds of group discussions, and a third round was necessary.
. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006452 doi: medRxiv preprint

Patient and Public Involvement
As per the IAP2 Spectrum of Public Participation 27 , patients and public stakeholders were consulted to prioritize research questions throughout three of four priority-setting project phases.
All research question data were developed and prioritized by patients, caregivers and healthcare providers of individuals with T1D. These research questions will inform our future research agenda, for which we plan to collaborate with stakeholder partners.

Study 1: Scoping Review
The however, four were excluded as relevant articles were already included in the published literature analysis. One registration provided a full-text publication, and was thus added to the published literature. Therefore after screening, 19 published articles and four registered trials were included for analysis.
Intervention summaries are provided in Tables 1 and 2. The majority of trials compared aerobic or combined aerobic and resistance training to a non-exercise control group. Interventions were delivered under supervised conditions by a kinesiologist or equivalent, for a median of 60 minutes/session (range: 10-90 minutes), three times/week (range: 1-5 times) for 20 weeks (range: 6 weeks-4 years). Twenty-three outcomes were reported across the 19 trials, of which nearly all focused on physiological factors, with glucose control and predictors of cardiovascular health being most common.
There was no evidence that any trials conducted to date engaged individuals with lived experience of T1D. Number of hypoglycemic events was the most commonly discussed PROM . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
No trial registration described partnerships with stakeholders in developing or implementing the study. Quality of life is the only PROM explicitly disclosed in one trial 38 .

Study #2 -Priority-Setting Exercise
The online survey was available and advertised to the public between July 2018 and January 2019. We collected responses from 115 individuals across nine Canadian provinces. Respondents were a mean age of 40.9 (±15.1) years, and the majority (73.9%) identified as a patient with T1D.
The remainder identified as caregivers (15.7%), friends (7.0%) or healthcare providers (12.2%), with some respondents identifying as more than one category (8.7%). More females completed the survey (63.4%) than males, and no participants identified as a non-binary gender. Most patients identified as being of Canadian ethnic origin (74%), with the next largest samples being European Canadian (15%), Métis (4%) and Caribbean Canadian (3%).
Of the 115 respondents, 100 submitted at least one research question, producing a total of 194 submissions. After qualitative analysis (Figure 2), 38 research questions were included in the phase 3 long-list and distributed to steering committee members for review. We received 100% of our steering committee rankings between February 21 st and March 8 th , 2019. Twenty-four questions were short-listed after receiving ten or more points in collated rankings.
Eleven of twelve steering committee members attended the Phase 4 workshop on April 6 th , 2019. The workshop lasted approximately six hours. Three rounds of small group discussions occurred. Following the second round, three members did not sufficiently agree with the aggregated list. The third round resulted in slightly lower agreement for two of those members. Therefore, we conducted a post-workshop analysis using small group rankings from the third round to supplement . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006452 doi: medRxiv preprint the aggregated list from the second round. This analysis resulted in the final top ten list of research priorities for T1D and exercise (Table 4).

DISCUSSION
This study examined the characteristics of randomized trials and patient priorities for exercise science research for people with T1D. In our scoping review, we determined that patient engagement methods and PROMs have not been historically used to inform exercise-based interventions. Guided by these results, we facilitated a priority-setting project with T1D stakeholders to identify priority research questions pertaining to exercise and health. We identified that patients and caregivers are interested in modalities and strategies to exercise safely and maintain glucose control. Collectively, these findings provide a novel patient-centred rationale for designing future randomized trials of exercise interventions for people with T1D.

Previous Literature
This is the first scoping review of exercise randomized trials for individuals with T1D designed to determine if patient engagement exists in exercise and T1D literature. This topic was not addressed in recent systematic reviews of exercise training and health outcomes in people with T1D [39][40][41] . We found that exercise randomized trials published or being delivered for individuals with T1D did not focus on stakeholder engagement. This gap is not exclusive to trials of T1D and exercise. A scoping review of priority-setting practices in all health research found only 27studies engaged patients in identifying research topics, and 12 in identifying specific research questions 42 .
Many studies simply inferred stakeholder priorities from qualitative data. Additionally, most trials engaging stakeholders do not integrate multiple stakeholders' perspectives (ie. patients, clinicians, caregivers etc) in the prioritization process. This is an important consideration when engaging stakeholders in research, as stakeholders with different experiences of a health condition may have different priorities for research topics or outcomes 43 . Engaging T1D stakeholders is a significant gap in exercise science literature and should be considered within future randomized trials.
Patient engagement and priority-setting projects identifying important research topics from stakeholders' perspectives are becoming more common within clinical research 42,44,45 . This project revealed that stakeholders are largely concerned with short-term outcomes, strategies to prevent hypoglycemia and stabilize short-term glucose control. This contrasts the JLA T1D treatments project 13 , where prioritized questions focused on long-term outcomes including adverse effects of various insulin analogues or potential cognitive impacts of living with T1D. This difference may . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint .
1 0 indicate some uncertainty felt by stakeholders regarding the safety of exercise given their individual situation. Fear of short-term health complications is a common barrier to regular exercise among people with T1D 5 . This fear itself has a range of health implications including reduced physical activity 46 , increased glycemic variability 46 , poorer sleep patterns 47

Strengths and Weaknesses
This study was strengthened by using two complimentary methods to identify gaps in exercise science for individuals with T1D. Our priority-setting project followed a modified JLA approach to priority-setting. We believe the recognition and support for this model 20 is a strength for our study. We were also supported by a highly engaged steering committee (100% phase 3 participation rate). Despite these strengths, several limitations should be addressed. The scoping review was limited to trials published recently and in the English language, therefore we may have missed some trials. Additionally, we were not able to achieve consensus in person at the workshop.
Although the JLA Guidebook 14 mentions a majority vote can be obtained in cases where consensus is not achieved, this process had to be conducted in a post-workshop analysis since several members had other commitments.

Impact of Patient Engagement
A quantitative evaluation of patient engagement was not conducted. Consulting end users as participants was integral to this study. The developed list of research questions was based on submissions collected directly from and prioritized by end users of research. Without input from patients and caregivers, prioritized research questions would likely have been very different.

CONCLUSION
We have outlined the current status of patient engagement in exercise research for individuals with T1D and engaged stakeholders in developing a list of priorities in T1D and exercise research. This list of priorities will be used to guide our future research agenda, and we . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint .

1
recognize the need to continue working with stakeholders in designing future research. It will also be critical to re-evaluate priorities as new information becomes available.

Summary Box
• We found no evidence of patient engagement or involvement in exercise science research for patients living with type 1 diabetes over the past twenty years • We worked with a Canadian sample of patients, caregivers and healthcare providers living or working with type 1 diabetes to develop a top ten list of priorities for future exercise science research • Current researchers interested in type 1 diabetes and exercise now have a framework of priorities set by stakeholders, which can be used to guide future research programs and result in more clinically relevant exercise science findings . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

FIGURE 1: STUDY FLOW DIAGRAM
. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint .    T  H  E  M  E  :  G  L  U  C  O  S  E  C  O  N  T  R  O  L  T  H  E  M  E  :  A  N  T  H  R  O  P  O  M  E  T  R  I  C  S  T  H  E  M  E  :  F  I  T  N  E  S  S  A  N  D  E  X  E  R  C  I  S  E  C  A  P  A  C  I  T  Y  T  H  E  M  E  :  B  L  O  O  D  M  A  R  K  E  R  S  O  F  H  E  A  L  T  H  T  H  E  M  E  :  C  O  M  P  L  I  C  A  T  I  O  N  S  T  H  E  M  E  :  S  U  B  J  E  C  T  I  V  E  M  E  A  S  U  R  E  S . CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the

(which was not peer-reviewed)
The copyright holder for this preprint  What explains the variation in responses that the same person can experience doing the same exercise between different days?

2
Which is the best for maintaining glycemic stability and glucose tolerance: aerobic training, strength training, or a combination of both? If a combination, does the order matter? 3 What modes of exercise (ie activity types, such as walking, cycling, weightlifting, rock climbing etc) produce the best health benefits while maintaining tight glycemic control?

4
What dietary plans can safely and effectively be followed for an active lifestyle in type 1 diabetes without compromising pre-and post-exercise glycemic control?

5
What is the optimal time of day and exercise prescription (example: how often, what type, how intense) in order to maintain ideal glycemic control and insulin sensitivity?

6
What is the best method of preventing post-exercise hypo-or hyperglycemia?

7
Will certain glycemic ranges before starting exercise consistently result in hypoor hyperglycemia? 8 What effect can various levels of hydration have on blood sugar levels during and after exercise?

9
How does hypo-or hyperglycemia affect muscle growth and strength training progress, or vice versa?

10
What is the effect of climate/temperature on blood sugar control during exercise and what causes this effect?
. CC-BY-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006452 doi: medRxiv preprint