NICOLE WAGNER
PhD, Kaiser Permanente Colorado Institute for Health Research

Uptake of a new health intervention is dependent on patient acceptance and responsiveness. 1) Patient report is frequently used to assess patient responsiveness due to the low cost and limited labor requirements. 2) However, patient report is prone to errors, such as over-reporting and missing data. 2), 3) Incorporating measures of patient responsiveness from multiple sources presents an opportunity to sustain the low cost and labor requirements while increasing data accuracy and comprehensiveness. Measures from multiple sources may also provide an opportunity to identify strategies for adaptation. 4) To assess the value of a mixed methods approach, this study describes the use of electronic health record (EHR) data, patient report, and implementor logs for measuring patient responsiveness in a pragmatic intervention trial.

Just in Case (JIC) was a cluster randomized intervention trial, designed to increase the uptake of naloxone, an overdose antagonist medication, in patients on chronic opioid therapy. JIC was conducted between 2017 and 2019 at Denver Health Medical Center, a safety net health system serving the Denver metro area. Patients 18 years and older filling chronic opioid medications were eligible to receive naloxone co-dispensing. Eligible patients were recruited to complete surveys at baseline, 4 months, and 8 months (patient report). Pharmacy staff in the intervention arm recorded naloxone co-dispensing events and reasons for nonacceptance in a pharmacy fidelity log (implementor log).

Patient responsiveness measures included naloxone uptake and barriers to naloxone uptake. Naloxone uptake was captured in the EHR using naloxone dispensing data. Reasons for not accepting naloxone were coded for common themes in patient report surveys and implementor log.

Using EHR pharmacy records, 527 eligible patients were identified with 204 naloxone fills. Barriers to naloxone uptake from patient report included lack of knowledge on naloxone (40%), not thinking they needed it (88%), and fear of repercussions from pharmacists or doctors (20%). Barriers to naloxone uptake from implementors included patients already had naloxone (36%) and weren’t willing to pay for naloxone (28%).

A mixed methods approach using measures from the EHR, patients, and implementors provides a comprehensive assessment of patient responsiveness with increased accuracy. Each measure contributed unique data to inform potential opportunities for improvement and adaptation. EHR data contributed accurate counts of acceptance. Patient report identified a knowledge gap indicating a more robust education program may be needed. Implementors reported cost issues indicating a potential protocol modification could improve uptake.

AMY HUEBSCHMANN
Associate Professor, Clinician-Investigator, Division of General Internal Medicine, University of Colorado Center for Women’s Health Research

Medically complex patients with uncontrolled type 2 diabetes face diabetes self-management challenges, including managing blood glucose levels and lifestyle behaviors. Technology packages have improved clinical outcomes by allowing patients to share data with clinic teams on home glucose (Glooko©) and behavioral health risk data (My Own Health Report, MOHR). However, adoption of Glooko and MOHR remains low in primary care. In a pilot study to inform implementation efforts, we evaluated the acceptability of Glooko/MOHR among key stakeholders: patients and clinicians.

We recruited eligible patients with uncontrolled type 2 diabetes mellitus (Hemoglobin A1c >8%) and their treating clinicians from three academic primary care clinics.

Participants provided acceptability ratings after a demonstration of the process of sharing Glooko/MOHR data between patients and clinicians. We considered ratings of ≥ 70% in each of the 7 Technology Acceptance Model (TAM) domains as acceptable. All quantitative data are reported as mean ± SD. We considered survey ratings of 70-80% and >80% as moderately and highly acceptable, respectively.

Patients enrolled (n=12) were adults (age = 65.7 ± 12.8 years), 33% non-white, 58% female, and 50% reported use of internet to manage health issues. Clinicians (n=11) had 13.2 ± 9.9 years of practice experience. Patient acceptability for Glooko data sharing: Intention to use (91.5±12%), Perceived usefulness (89.5±8.1%) and Social influence (83±0%). No unacceptable ratings. Patient acceptability for MOHR data sharing: Perceived usefulness (85.5±8.1%), Self-efficacy (83.5±12%) and Social influence (83±0%). One TAM domain was rated as unacceptable: Resistance to change (58.5±12%), including 33% of patients agreeing that s/he did not want MOHR to change how s/he managed diabetes. Clinician (n=11) acceptability of sharing Glooko and MOHR data: Highest ratings were for Perceived usefulness (88.1 ± 4.2%), Facilitators (84.2 ± 8.6%) and Intention to use (82 ± 15.6%). The Subjective Norms/Others’ support was unacceptably rated (50.2 ± 16%), including anticipation of low perceived support among patients (27%), colleagues (55%), and health managers (55%).

Medically complex patients with diabetes and their clinicians expressed intention to use technology to share glucose and behavioral risk data between visits. However, to reach the promise of using remote technology and patient-reported data to address health challenges, clinics will need to identify and address factors leading to clinicians’ perceptions of limited support from others to use remote data monitoring, particularly patients, and also better discern why some patients are resistant to using MOHR as part of their diabetes management.

DENNIS GURFINKEL
Sr. PRA, ACCORDS

Diabetes group visits are historically challenging to implement in primary care. Pragmatic trials optimally use existing staff to deliver the intervention and allow flexibility in adherence and delivery.

1. To describe use of the Replicating Effective Programs for adapting and implementing an evidence-based intervention for use in real-world care settings.
2. To describe methods for establishing fidelity and adaptations to a study protocol to ensure rigor and feasibility of the conduct of a pragmatic trial.

The Invested in Diabetes study is an ongoing pragmatic cluster randomized comparative effectiveness trial testing two group visit models for delivering the Targeted Training in Illness Management (TTIM) curriculum for diabetes in primary care.1 In one model, TTIM is delivered by a health educator, with set topic order. In the other model, TTIM is delivered by a multidisciplinary care team, with topic order selected by patients. Practices are supported using the Replicating Effective Programs (REP) implementation framework plus intensive practice facilitation.2 Patient and practice stakeholder input was used to adapted TTIM curriculum and the study protocol and outcome measures. Dedicated research staff were used to help practices implement the project and collect outcomes data (patient-reported outcomes and Electronic Health Record data). Finally, the study team observes one session per practice/quarter to monitor fidelity to the TTIM curriculum and study core elements, documenting adaptations to content or delivery.

Study team members rated the pragmatic design of the study protocol according to the PRECIS-2 guidelines3 (Figure 1). Core elements of the study were identified and described to ensure fidelity. Stakeholder-led adaptations of the protocol outside of core elements were identified pre-implementation, including 6 two-hour sessions instead of 12 one-hour sessions and streamlining patient-reported outcomes to those with clinical utility and patient preference. Twelve in-person and virtual trainings have been conducted to date; trainings have progressively highlighted importance of skill building activities for SMA facilitators. Around 80 practice coaching sessions have been done to help practices start and sustain their group visits. Practices delivered test data extracts in summer 2019; data quality assessments revealed variability in accuracy and completeness. Thirteen months into the 24-month implementation period, 86 cohorts have gone live with 613 out of the goal of 1440 patients enrolled in group visits. Ongoing practice support is maintained through dedicated practice coaches to help troubleshoot issues and maintain fidelity to the study protocol.

To retain rigor in the study design, the REP framework allowed for adaptation to context while establishing core elements that must remain in place for hypothesis testing. The Invested in Diabetes study implementation processes help to ensure rigor of the design as well as feasibility of delivery in real-world primary care practices.