BACKGROUND

Engaged patients have been referred to as “the blockbuster drugs of the 21st century.” Under the Affordable Care Act, accountable care organizations (ACOs) are required to engage patients in their own health care decisions and treatments. ACOs are incentivized to improve patients' engagement in their own health and health care because health plan contracts involve global payment and substantial financial risks and rewards. As a result, adult primary care practices affiliated with ACOs are beginning to adopt and learn from a range of patient activation and engagement (PAE) strategies, including motivational interviewing, shared decision-making, shared medical appointments, health risk assessments, and the inclusion of patients in quality improvement and clinic governance. Health systems and ACOs are emphasizing the use of PAE for adult patients with diabetes and/or cardiovascular disease (CVD) because engagement with care, self-management, treatment adherence, depression, and social function may affect disease progression and prognosis. We specifically address the following question: Do adult patients with diabetes and/or CVD who are receiving care from ACO-affiliated practices with high adoption and implementation of PAE activities achieve better patient-reported outcomes (PROs), report better experiences of care, and achieve better clinical outcomes compared with patients receiving care from ACO-affiliated practices with low adoption and implementation of PAE?

OBJECTIVES

Using a 39-item survey of lead clinicians, we set out to do the following: 1. Collect information on PAE activities, which included assessments of disease prevention and health promotion activities, care team patient communication, shared decision-making, self-management support, advanced serious illness care, and patient involvement in care redesign in 16 practices of 2 ACOs during 2 consecutive years. 2. Assess the differences in PROs, patients' experiences of care, and selected clinical outcomes between adult patients with diabetes and/or CVD receiving care in practices with high vs low adoption and implementation of PAE. 3. Explore the extent to which practice-level variation in culture, leadership, teamwork, relational coordination, and patient-centeredness account for differences in PRO levels between practices with high vs low PAE.

METHODS

We assessed the naturally occurring variation in the adoption and implementation of the 39 PAE activities in 16 practices of 2 ACOs treating adult patients with diabetes and/or CVD using mixed methods, including a combination of organizational surveys, patient surveys, patient-level clinical data, and key informant interviews. A random sample of adult patients from each ACO completed a survey instrument that included selected PROs, the Patient Activation Measure [PAM-13], and patient experience measures in 2015 (time 1; T1) and then again 1 year later in 2016 (time 2; T2). We surveyed ACO practice members in 2015 and again 1 year later in 2016 to assess changes over time in organizational culture, leadership, team effectiveness, relational coordination, and patient-centeredness using previously validated instruments. We then supplemented these assessments using 103 informant interviews with key themes coded and analyzed using Atlas.ti. Multilevel regression analyses estimated the association of practice site PAE adoption level (high vs low PAE) on PROs, controlling for patient characteristics.

PATIENT OUTCOMES

PROs included PROMIS® participation in social roles and activities (Short Form-8a); a 12-item short form survey measuring physical functioning (Short Form-12a); a 4-item measure of anxiety and depression (PHQ-4 [Patient Health Questionnaire-4]); the PAM-13, and the Patient Assessment of Chronic Illness Care (PACIC-11). Clinical outcomes included systolic and diastolic blood pressure, glycated hemoglobin (HbA1c), and low-density lipoprotein cholesterol (LDL-C) control.

RESULTS

Overall, patients surveyed had stable PRO scores in T1 and T2. We found no statistically significant differences between high- and low-PAE practices at T1 or T2 for any of the PROs or in LDL-C levels, HbA1c levels, or diastolic blood pressure. We found a significant mediating effect of patient activation on the relationship between better patient experience and all 3 PRO scores at both T1 and T2, highlighting the important role of patient activation in translating practices' efforts to engage patients in diabetes and/or CVD care management. Patients who had 1 or more comorbid mental health conditions and high activation scores had 4 times the odds of having high social functioning compared with similar patients with low activation. Relational coordination of team members, innovation culture, and patient-centered culture scores were better for high-PAE practices compared with low-PAE practices in T2, while no differences existed at T1. Qualitative analyses of key informant interviews highlighted challenges of adoption and implementation of PAE strategies and elucidated reasons for a lack of relationship between high-PAE adoption among practices and PROs in quantitative results. Key informants recognized PAE as mutual goal setting, motivational interviewing, and shared decision-making; some felt that they used these approaches, but they had limited understanding of what the approaches entailed and did not use them routinely in caring for their patients with diabetes and/or CVD.

CONCLUSIONS

Patients with diabetes and/or CVD who received care from adult primary care practices that were in the top quartile on a 39-item index of PAE activities did not report better experience of care or better emotional, physical, or social outcomes of care compared with patients of practices in the bottom quartile of PAE activities. Some patients are highly activated and some are not, irrespective of practice efforts to engage them. Based largely on our qualitative interviews, we found that another reason for the lack of association is that practice teams had a superficial and sometimes incorrect understanding of PAE and strategies (eg, motivational interviewing and shared decision-making) and inconsistent and uneven application in daily practice. They also were deterred from implementing PAE strategies because of the greater attention given to having patients achieve clinical targets compared with improving PROs. To improve the impact of practice adoption and implementation of PAE on PROs, the use of evidence-based audit, feedback, and facilitation strategies should be part of future dissemination work.

LIMITATIONS

Only 1 individual per practice (generally the practice leader) completed the PAE survey. The practice leader was considered the most knowledgeable informant about whether the practice was using a specific PAE activity; however, it is possible that others within the practice might have responded differently. The relatively small degree of change between T1 and T2 in the PRO measures, the PAM-13 measure, and in relational coordination and team participation measures limited the ability to draw causal inferences based on naturally occurring differences between the 2 time periods. Examining changes in PROs over a longer time period and providing more structured and intensive dissemination of PRO feedback to practices might more effectively improve them. Given the limited variation in core measures in this sample of primary care practices of 2 ACOs, additional research using a larger sample of practices across a greater number of ACOs could enhance the generalizability of studies examining organizational and team influences on PROs.