Scope of the COS

The COS has been designed for use in all trials and clinical research studies evaluating safety and effectiveness of all potential interventions for diabetes-related foot ulceration in adults aged ≥16 years with a diagnosis of diabetes.

Phase I: Generation of the Long List of Outcomes

Phase II: Delphi Process (May–December 2023)

Phase III: Steering Group Meeting (December 2023)

Given the expected heterogeneity of the patient population of interest, the steering group—composed of one trial methodologist (J.N.), four vascular surgeons (R.J.H., D.R., A.S., and C.A.), and one diabetologist (F.G.)—reviewed and rationalized the list of outcomes rated as either critical or without consensus to generate a list of relevant outcomes that would be feasibly discussed in the consensus meeting. The rationalization process included deduplication of outcomes, removal of measures that either were not considered to be true outcomes or were relevant only to a specific subgroup of patients with diabetes-related foot ulceration (such as revascularization-specific outcomes).

Phase IV: Consensus Meeting (January 2024)

The consensus meeting was held online on the Microsoft Teams platform and was attended by patients with lived experience of diabetes-related foot ulceration and international experts in the field of diabetes-related foot disease, including diabetologists, podiatrists, vascular, orthopedic plastic surgeons, movement scientists, and wound care specialists. The expert panel was selected purposefully from the members of the IWGDF and VSGBI diabetic foot specialist interest group to ensure balanced representation from the relevant stakeholder groups. All attendees provided written consent prior to attending the meeting.

All outcomes considered to be either critical or lacking consensus during the Delphi rounds were discussed in the consensus meeting. Each outcome was open for voting using the consensus criteria from the Delphi stage using anonymous Mentimeter software (Mentimeter; Stockholm, Sweden). This followed a detailed discussion for each outcome between all the stakeholders to help create the final feasible and practical COS.

Phase I: Generation of the Long List of Outcomes

Phase II: Delphi Process

The 103 unique outcomes identified through systematic review (n = 95) and patient interviews (n = 8) were taken forward to the Delphi survey. The two consecutive rounds were sequentially completed between May 2023 and November 2023. Following removal of 26 duplicate submissions, 336 responses were recorded (281 from health care professionals and 55 from patients) in the first round and 176 (152 from health care professionals and 24 patients) in the second round, corresponding to a 52% second round response rate (Table 1). Among the participating health care professionals, two-thirds practiced in the U.K., and the remaining third represented 33 other countries across six continents.

Table 1

Demographics of health care professionals participating in the Delphi survey

	Round 1	Round 2
Number of participants
Total	336	176
Health care professionals	281 (84)	152 (86)
Patients	55 (16)	24 (14)
Clinical role of health care professionals
Doctor	114 (41)	65 (43)
Podiatrist	133 (47)	69 (45)
Nurse	12 (4)	8 (5)
Other	22 (8)	10 (7)
Specialty of health care professionals
Vascular Surgery	72 (26)	43 (28)
Orthopedics	13 (5)	4 (3)
Diabetes and Endocrinology	124 (44)	74 (49)
Other	72 (26)	31 (20)
Seniority of participating doctors
Consultant	86 (75)	51 (79)
Staff grade	10 (9)	4 (6)
Registrar	18 (16)	10 (15)
Years of clinical experience
<5	27 (10)	10 (7)
5–10	52 (19)	27 (18)
11–20	82 (29)	51 (34)
>20	120 (43)	64 (42)

Open in a separate window

Values are provided as counts with proportion of the relevant group expressed in percentages in brackets. All health care professionals could express their specialty.

Of the 103 outcomes available for rating in the first round, 47 (46%) were ranked as critical and were automatically considered for inclusion in the final COS. The remaining 56 (54%) outcomes did not reach consensus and were taken forward to the second round. Of these, 16 (15%) were subsequently rerated as critical, giving a total of 63 critical outcomes, with 40 remaining without consensus (Table 2).

Phase III: Steering Group Meeting

Given the unexpectedly high number of outcomes being considered as critical (n = 63) or lacking consensus (n = 40), the steering group rationalized the list of outcomes to ensure that the COS was relevant, practical, and feasible. Following review of the critical outcomes, 41 out of 63 (65%) were excluded from discussion in the subsequent consensus meeting, as 12 were captured in other outcomes, 11 were not universally relevant to all patients, 7 were not specific enough for inclusion in the COS, 5 were not a true outcome measure, and the remaining 6 were treatments and not outcomes (Supplementary Table 1).

A similar process was undertaken for outcomes without consensus, with 25 of 40 outcomes (63%) being excluded (Supplementary Table 2).

COS in the Context of Previous Work

IWGDF and the European Wound Management Association have previously jointly published guidance on the core planning and reporting details for studies on prevention and management of diabetic foot ulceration (7). Nonetheless, this statement was a product of expert panel discussion rather than a formal methodological approach with active patient and public involvement. Despite these differences, six of the outcomes in the COS, including wound healing, time to healing, minor amputation, major amputation, health-related quality of life, and mortality, were already recommended as the key quality measures for studies on interventions for diabetes-related foot ulceration by IWGDF and European Wound Management Association.

Ratification of the Final COS

The consensus meeting, attended by patients and international experts in diabetes-related foot disease representing the key stakeholder groups, facilitated balanced debate on the key components of the COS. The overarching aim was to create a practical COS, of similar size to that of other disciplines (17–20), which can be applied to all patients with diabetes-related foot ulceration across various health care systems globally. Despite 22 outcomes being considered as critical in the Delphi process, the final 8 included in the COS captured most of these measures, respecting the development process. Most of the deliberation during the final consensus meeting related to issues surrounding the definition and subsequent reporting of new and recurrent ulceration, subtypes of infection requiring universal reporting, measurement of quality of life, physical activity and functional status, hospitalization, revascularization, and major adverse cardiovascular events.

The consensus group acknowledged that, ideally, new and recurrent ulceration should be reported separately as outcomes. It was felt that new ulceration could be regarded as a study safety measure to determine whether a specific treatment resulted in additional wounds during the course of treatment and before the initial ulcer was healed (e.g., a foot ulcer from inappropriate casting). Recurrent ulceration, on the other hand, would potentially imply that a particular intervention may have been ineffective, but would also require longer follow-up periods. However, based on experience from previous trials using diverse methodologies and a current lack of uniform definitions (21), the experts agreed that it may be challenging to differentiate between new and recurrent ulceration. Therefore, grouping new and/or recurrent ulceration together would allow for pragmatic outcome reporting of wounds guided by the needs of the individual research team.

Following debate, osteomyelitis, despite being seen as critical by 89% of Delphi participants, was not included in the final COS. Osteomyelitis was regarded as a specific subtype of infection and could be seen as a confounder of a poor clinical outcome. Including it in the COS would also potentially result in the need for additional diagnostic tests (e.g., imaging, bone biopsy) to facilitate the diagnosis in patients with diabetes-related foot ulceration, whose clinical presentation would not necessarily have supported these tests in routine practice. Furthermore, none of the current diagnostic tests are 100% specific or sensitive in the detection of osteomyelitis (22), suggesting that such an outcome may not be reliably reported. Infection was chosen as one of the core outcomes as an umbrella term to capture all infective processes.

Pain and ambulatory status were rated as critical in the Delphi process. However, both of these outcomes are often explicitly captured or are considered to be key determinants of health-related quality of life. For example, ambulatory status (defined as level of physical activity or walking ability) is captured in most health-related quality of life questionnaires. Furthermore, pain was considered not to be universally experienced by all patients, with previous studies suggesting that it affects 21% of patients with neuropathy (23). While tolerability was felt to be important to patients, its close links to adherence with a particular intervention and potential loss to follow-up would traditionally be captured in the study’s CONSORT flowchart. Similarly, falls were found to be specific to trials assessing offloading devices rather than all interventions.

Potential reporting of hospitalization for a diabetes-related foot ulcer was noted to be challenging given the inherent difficulties in comparison of such outcomes, because of variable admission thresholds across diverse health care systems and pathways of care. It has been recognized that physicians without access to designated multidisciplinary diabetes-related foot clinics are more likely to admit patients with an ulcer compared with those practicing in specialist centers (24). Current evidence suggests a global increase in hospitalizations for diabetes-related foot complications (25). Nevertheless, there has been a shift in paradigm, with various national strategies being implemented to promote admission avoidance, such as specific guidance from the Joint British Diabetes Societies for Inpatient Care in the U.K. (26). With the shift toward community-based interventions, including utilization of outpatient parenteral antimicrobial therapy or hospital at-home services (27), hospitalization rates may no longer be of key significance to the population of patients with diabetes-related foot ulceration. As most diabetic foot care is delivered in community or outpatient settings, hospital admission was ultimately considered not to be a critical outcome.

Revascularization was not included in the COS as it was considered to be a form of treatment aimed at improving overall patient outcome, such as wound healing or limb salvage, rather than an outcome itself. Even though 50% of patients with diabetes-related foot ulcers have underlying peripheral artery disease, fewer than 10% require revascularization (28), reducing its applicability to all studies.

While major adverse cardiovascular events were overall seen as critical, the consensus was that they would not be relevant to all studies, particularly those looking at effectiveness of specific dressings or wound care. As a result, acute kidney injury and cardiovascular or cerebrovascular events were not included in the final COS.

Limitations

Despite three reminders being sent to the second round nonresponders, the response rate in our Delphi survey was moderate at 52%, similar to that of other studies (16). Nonetheless, relatively low response rates in the consecutive rounds are a recognized limitation of the Delphi process, particularly in surveys with a high number of items available for rating (29). Given the moderate response rate and the fact that additional outcomes were rerated as critical but none were excluded after the second round, there was no role for further rounds, in keeping with COS development in other disciplines (17,20).

Despite promoting the study with support from patient charities and approaching all suitable patients under the care of two large multidisciplinary diabetic foot networks providing at least five diabetic foot clinics a week with 12 follow-up and four new appointments daily, patient participation in the Delphi survey was relatively low. However, patients were actively engaged in all stages of the COS development. The interview phase, which included 18 patients with diverse foot complications and exposure to a range of interventions, generated eight distinct patient-specific outcomes, which were subsequently incorporated into the Delphi survey. Furthermore, the consensus meeting was attended by five patient representatives who provided crucial input into discussions shaping the final COS.

Lastly, patients with diabetes-related foot ulceration form a very heterogenous group, with variable levels of underlying neuropathy, peripheral arterial disease, and infection, ultimately resulting in dissimilar personal experiences, disease progression, and treatment plans. It was therefore understandable why the Delphi process resulted in the unexpectedly high number of critical outcomes (n = 63), which, if included, would have resulted in an impractical and undeliverable COS. This issue necessitated deviation from the standard COS development protocol, with rationalization of critical and no consensus outcomes prior to discussion in the consensus meeting, to ensure valuable and structured discussion and formalization of a feasible COS. Nonetheless, the majority of outcomes excluded by the steering group prior to the consensus meeting either were not true outcomes or were too specific to a particular subgroup of patients, already precluding their incorporation in the final COS. Despite this deviation from standard protocol, both the list of outcomes discussed in the consensus meeting and the final COS included outcomes ranked as most critical during the Delphi process. Similar issues and other forms of protocol adjustment have been noted in other COS development studies (17).

Acknowledgments. The authors thank Dr. Jemima Dooley (Bristol Medical School, University of Bristol) for her help in designing the topic guide for patient interviews. The authors also express our gratitude to the patient representatives Linda Brodie, Janet Harford, Brenda Riley, Sarah Parsons, and Michael Vernon for their valuable contributions to discussion during the consensus meeting.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. G.D. and R.J.H. conceptualized the study. A.S. and G.D. completed the systematic review to facilitate the first phase of the COS development. A.S. and M.O.-G. interviewed patients and analyzed the transcripts. C.A. created patient descriptors for the outcomes in the long list. A.S. designed the Delphi survey on REDCap and provided oversight of data collection. A.S., A.J., and R.J.H. analyzed the data. A.S. drafted the manuscript. A.S., F.G., J.N., D.R., C.A., and R.J.H. formed the steering group. A.S., F.G., J.N., D.R., D.G.A., C.A., S.A.B., J.C., V.C., K.D., M.E., R.F., C.G., E.J.H., A.J., V.K., L.A.L., J.L.M., M.M.-S., E.J.G.P., J.S., J.v.N., D.K.W., and R.J.H. participated in the consensus meeting. All coauthors critically reviewed and approved the final version of the manuscript. R.J.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Elizabeth Selvin and Rodica Pop-Busui.