The INTEGRATE-D intervention is designed to support primary care practices with implementing ADA recommendations for integrating medical and psychosocial care for patients with DMII. Primary care practices will be the recipients of this support. For practices, the intervention will be 15 months long. The first three months will be a "Readiness Phase" where the study team assesses practices' capacity at baseline and a facilitator works with intervention practices to address capacity issues (e.g., help the practice produce a quality report) that can be barriers during active implementation. Then, intervention practices receive 12 months of "Active Intervention" with the same facilitator to align care processes with ADA recommendations. Pilot study design: The study team will use a randomized, mixed-method, pre-post design for this pilot study. The goal of this randomized pilot design is to evaluate the feasibility of recruitment, randomization, retention, assessment procedures, implementation of the INTEGRATE-D intervention, and provide insights into the effect of this pilot intervention on practice-level outcomes. Practice Sample and eligibility criteria: The sample for this pilot is four primary care practices. Practices with more than 100 adult patients (≥18 years) with a diagnosis of DM (Type II; not gestational) are eligible. Practice Recruitment: Practices will be recruited from the ORPRN practice-based research network. ORPRN works with 279 primary care practices in the state of Oregon, including 51% that are rural and 42% clinician-owned practices. ORPRN recruiters will identify 50 potential practices and conduct a faxback to assess interest, capacity, and eligibility. Of interested practices, ORPRN and the study team will select 4 practices (2 rural and 2 non-rural) that vary on whether practices employ a behavioral health clinician. The study team will purposefully vary practices on these characteristics based on previous work that demonstrated they influenced intervention uptake. Patient Sample: The study team will have three different patient samples. For chart audits, there will be two samples. (1) Patients with a diagnosis of DMII who have been seen at least once at the practice in the 15 months prior to the intervention start date and at least once after the intervention start date will be included in the chart audit. The study team will randomly select 50 patients meeting these criteria. (2) Patients meeting the above criteria and screening positive for depression (i.e., those patients who have minimally exposed to the INTEGRATE-D intervention will also be randomly selected for chart audit. The study team will audit charts until there is a sample of 30 patients meeting these criteria. (3) From this latter group, the study team will select 5 patients at each of the two intervention practices to participate in an interview. Randomization Procedure: Practices will be randomized in a 1:1 ratio through a covariate-constrained randomization procedure. This is accomplished as follows. Prior to randomization, the study team will collect baseline data on practice-level factors (e.g., practice size, ownership, rurality, patient panel demographics, etc.) that could influence process and outcome measures. These data will be collected as part of the recruitment process. The study team will construct all possible combinations of eligible practices into two groups. For each selected possible randomization, the study team will compare the balance between study arms in the important baseline data described above through a balance criterion. The study team will randomly select one randomization combination (from the 2 combinations with the best balance) which will yield 2 practices in each study arm with good balance in important covariates. Data Collection: For this pilot study, the study team will collect mixed methods (survey, observation, interview, and chart audit) data. For control practices, data collection will include the collection of survey and chart audit data. These data will be collected at the same time as data collected from intervention practices. Intervention practices, in addition to completing surveys and participating in the chart audit, will participate in practice site visits, which include the study team observing the practice and how it delivers care and interviewing a varied group of practice members, including those that participated in implementing the intervention. These visits will happen before and after the intervention. In addition, during the implementation of INTEGRATE-D, the study team will conduct monthly telephone calls with a key informant at the intervention practices. Measures: The study team will measure practice capacity for quality improvement using the Change Process Capability Questionnaire (survey). The study team will measure changes in care processes, as follows: for depression - screening for depression using the Patient Health Questionnaire (PHQ)-9, rates of referral to behavioral health when patients screen positive on the PHQ-9; for diabetes - the study team will use a composite measure of DMII process of clinical care (e.g., screening for HbA1c, eye, and foot exams) and assess discussion of and engagement in self-management activities (e.g., healthy eating, physical activity). The study team will measure outcome change as a change in PHQ-9 and HbA1c. All process and outcome measures will all be collected via chart audit. The study team will measure feasibility and acceptability, perceived and actual, of INTEGRATE-D using observation, semi-structured interviews, survey, and facilitator tracking data. Surveys will use validated acceptability and feasibility measures. Statistical power: As this is a pilot study, it is not meant to have the sample size needed to test for statistical significance. Rather, this pilot is designed to show the promise of the INTEGRATE-D intervention and provide the data need to do accurate power calculations for the larger, subsequent pragmatic trial. Data Analysis. Qualitative analysis will follow a rigorous process whereby a team of experienced qualitative researchers will identify emerging patterns or findings, and then analyze data again to solidify these patterns into a coherent set of findings. Quantitative analyses will use standard descriptive statistical methods and the study team will utilize Generalized Linear Mixed Models to compare the effect of INTEGRATE-D intervention on process outcome measures in the context of a cluster-randomized design.
