Feasibility of a Manualized Occupation-Based Diabetes Management Intervention

Design

We modeled our feasibility assessment after Tickle-Degnen’s (2013) exemplar for how to adapt the framework presented in Thabane et al. (2010) to occupational therapy research. Thabane et al. proposed a framework of four categories of questions to evaluate in feasibility studies. First, process assessment includes an analysis of the appropriateness of eligibility criteria, assessment procedures, and participants’ intervention adherence. Second, resource assessment examines the time, equipment, and space needed to carry out the study procedures. Third, management assessment evaluates the capacity of research staff to complete tasks as assigned in the study protocol. Finally, scientific assessment evaluates the dose, duration, treatment effect, and participant burden associated with the intervention. We evaluated these dimensions of feasibility using a mixed-methods approach that incorporated both process data collected in the course of conducting the study (e.g., recruitment and retention rates, participant safety, intervention adherence) and poststudy interviews with participants, the intervener, and research staff.

Recruitment

Within the process assessment, we monitored our ability to recruit and retain participants throughout the study duration. Eligibility criteria for participants were as follows: (1) age 18–30 yr, (2) diagnosis of type 1 or type 2 diabetes for a minimum of 1 yr, (3) hemoglobin A1c (HbA1c; measure of average blood glucose) of ≥8% (indicating poor diabetes control), and (4) not pregnant or planning to become pregnant. Participants were recruited using two strategies: (1) A graduate student worker recruited participants in person at a young adult diabetes clinic held weekly at a large urban county hospital, and (2) potentially eligible participants were identified from a medical chart review at the young adult clinic and invited to participate by telephone. The study was approved by the University of Southern California institutional review board (IRB), and all participants provided informed consent at the time of study enrollment.

Measures

Participants completed a demographic questionnaire, and their medical charts were reviewed to ascertain their medical history, current diabetes treatment regimen, and relevant comorbid conditions. An assessment battery was administered at enrollment and again immediately after completion of the intervention. This included HbA1c testing, using an Afinion point-of-care machine (Alere, Waltham, MA) which provides immediate results using a finger-prick blood test, as well as the following behavioral and psychosocial self-report measures:

As part of the process assessment, we monitored the completeness of assessments throughout data collection to evaluate the practicality and burden of data collection procedures and usability of outcome measures. In addition, participants were interviewed to provide feedback about their experiences at the conclusion of the study.

Intervention

The REAL intervention framework and process of intervention development are described elsewhere (Pyatak et al., 2015). Three staff members were involved in the implementation of the study: The principal investigator (PI; Elizabeth A. Pyatak) oversaw the study and performed administrative tasks, a bachelor’s-level student recruited participants and conducted testing, and a licensed occupational therapist pursuing a clinical doctoral degree provided the intervention.

In brief, the occupational therapist and participant established diabetes-related goals in accordance with each participant’s values and needs. The occupational therapist then provided customized, activity-based sessions based on four core intervention principles and seven modules of manualized intervention content. Before intervention delivery, the occupational therapist completed advanced training including 8 hr of diabetes-related continuing education and 12 hr of motivational interviewing training.

As part of the scientific assessment, the occupational therapist recorded detailed treatment notes and general notes regarding feasibility and acceptability of the intervention after each session using REDCap, a secure online data management system (Harris et al., 2009). To maintain intervention fidelity and problem-solve emergent challenges, the occupational therapist had weekly meetings with the PI and had access to a social worker and endocrinologist for consultation as needed. The occupational therapist was interviewed at the end of the study to collect suggestions for improving the intervention and overall study procedures.

Process Assessment

Our partnering hospital clinics reported that they followed approximately 90 patients who met eligibility criteria. Target recruitment was reached in 7 wk. Eight participants (age 22.1 ± 1.2 yr; 87.5% Hispanic–Latino; 75.0% with type 1 diabetes) completed baseline testing, and 7 completed follow-up testing. Testing sessions averaged 120 min, which we found to be too long to maintain the attention span of both tester and participant. Postintervention focus groups also proved challenging to implement; after several attempts to schedule focus group sessions, only 2 participants attended. Therefore, the remainder of participants completed individual interviews in their homes.

Two changes were incorporated during the recruitment process. First, a comorbidity exclusion criterion was added, because we found it necessary to exclude otherwise-eligible participants with conditions that the intervention was not sufficiently equipped to address (e.g., significant intellectual disabilities). Second, we learned that many of the patients with type 2 diabetes seen at our partnering clinic had been lost to follow-up. Thus, we submitted an IRB amendment to identify prospective participants with type 2 diabetes through medical chart review and recruit them by telephone.

In terms of retention, 1 participant was lost to follow-up after one intervention session and was unable to be reached for follow-up testing. The 7 remaining participants attended a median of 8 weekly occupational therapy sessions each and completed follow-up testing. Treatment sessions ranged from 45 min to 120 min each, with initial evaluations averaging 89 min and subsequent sessions averaging 66 min.

Resources Assessment

Participants’ study involvement lasted an average of 3 mo, including 8 days to complete baseline testing after participant enrollment and informed consent, 8 days to schedule the initial intervention session after baseline testing, 60 days of intervention, and 16 days to complete follow-up testing after being discharged from the intervention. The occupational therapist and tester used office and cellular telephones to contact participants; text messaging emerged as participants’ preferred communication modality.

The study took place in two settings: (1) the PI's office space to conduct treatment planning, documentation, and store and analyze data and (2) community settings to conduct assessments and deliver the intervention. No issues in either location affected completion of the respective tasks. We experienced one issue with the Afinion HbA1c testing machine: Because of a delay in conducting follow-up testing with one participant, the test cartridges had expired, resulting in missing HbA1c data for this participant.

The study was conducted within an occupational therapy department at a research university, which provided access to technological and administrative support. Software and computer programs available for data management, qualitative and quantitative data analysis, and data storage compliant with the Health Insurance Portability and Accountability Act of 1996 (Pub. L. 104-191) were adequate. We also partnered with a local hospital clinic whose physicians fulfilled their commitments to assist with recruitment. Finally, the doctoral resident, study supplies, and PI's time and salary were supported through departmental financial support and grant funding from the university’s Clinical and Translational Science Institute.

Management Assessment

The occupational therapist delivering the intervention experienced burden related to participants’ frequent requests to cancel or reschedule appointments, often at the last minute. Because the intervention had to be delivered within a short time frame, the therapist needed to remain flexible to accommodate participants’ preferences, regardless of practical considerations such as driving time. The therapist further reported that the diabetes continuing education course was invaluable and felt that additional training surrounding psychosocial challenges specific to the target population would be beneficial.

All data were collected using paper forms and entered into REDCap for data management. Two separate databases were used, one for participants’ contact information and therapist treatment notes and the other for deidentified medical information and survey responses. The online database had convenient and secure accessibility in both the office and the field. The therapist requested an additional narrative text box to document events not captured in previously designated response options.

Study procedures included a suicide assessment protocol that was used with one participant who revealed suicidal ideation to the occupational therapist during an intervention session and to the tester at follow-up testing. The therapist consulted with the social worker to develop a suicide risk assessment and continually monitor the client during the intervention. The tester appropriately followed the suicide assessment protocol during follow-up testing.

Scientific Assessment

There was a general consensus among participants that the intervention made a meaningful contribution to improving their diabetes management. There were no indications that the disclosure of suicidal ideation mentioned previously was related to the intervention, and no other adverse events were reported. Although the number of sessions was seen as adequate to address participants’ self-identified diabetes management goals, both the participants and the therapist indicated that the intervention would benefit from delivery over a longer period of time and with greater flexibility regarding the timing and duration of sessions.

The primary outcome, HbA1c, was feasible to measure and meaningful to most participants. However, because participants received the intervention over an average of 60 days we could not detect changes in this outcome given that HbA1c measures average blood glucose levels over a 3-mo period. With respect to other outcomes, although the purpose of a feasibility study does not include evaluation of intervention effects, we found that participants’ changes in behavioral and psychosocial outcomes were in the expected direction, with effect sizes (Cohen’s d correcting for dependence among means) ranging from 0.20 (small effect) to 0.80 (large effect).