I have written a lot about primary care. I’ve covered its role in reducing ED visits, discussed the benefits of one of its most unique facets (coordination of care), and described the challenges of younger adults moving away from traditional models of primary care. It is hard to overstate just how important this field of medicine is to a high-functioning health care system.
Because of this, we increasingly turn to primary care to help meet the growing national demand for productivity, reduced costs, improved patient experiences, and better population health. These are all things that primary care providers can do. But in a time-short world, they are working in teams to help meet these goals. From a research and policy perspective, however, identifying team-based primary care is a major challenge.
What is Team-Based Primary Care?
Collaborations between providers are often referred to as team-based primary care. The goal is to engage more professionals in shared patient care. Doing this decreases the reliance on a single physician while maximizing the roles that other professionals play.
The Institute for Quality Improvement–an organization at the heart of the team-based care movement–notes that teams can take a number of forms. Physicians have often built teams with physician assistants (PAs) and nurse practitioners (NPs). They also regularly involve registered nurses (RNs), medical assistants (MAs), and non-clinical staff. Providers are also starting to collaborate more closely with pharmacists, physical therapists, behavioral health and social workers.
Team-Based Enhancements to Patient Care
It is common for a primary care physician to be responsible for handling all aspects of care. But for a patient with diabetes, does a physician need to be the one to conduct a foot exam? An MA could just as easily do this and record the results, just like taking blood pressure. Is the physician the right person to counsel patients on nutrition? A health coach might be better positioned to spend more time giving advice on nutrition, exercise, and other health behaviors. Regular meetings with a team pharmacist might also help improve how regularly patients take their medications.
Team-based care has its rewards. A 2016 white-paper [pdf] published by the Agency for Healthcare Research and Quality, attributes increases in access, comprehensiveness and coordination of care to team-based care. Patients received more health education, support for managing their own chronic diseases, and behavioral health. Providers were also more satisfied working in teams. An investigation of 23 high-performing practices also found that team-based care can result in greater joy in practice.
Identifying Team-Based Primary Care is a Major Challenge
It is not easy to define or measure what makes team-based care. Yet, we need to identify and distinguish among the variety of types of teams to answer important questions for both practice and policy. For example, what types of teams work best? What divisions of labor result in the most cost-effective practice? What degree of teamwork is needed to see better patient outcomes?
An article by Yong-Fang Kuo and colleagues published ahead of print in Medical Care explores these challenges. They examined whether Medicare billing data for 68 practices in Texas could be correlated with team-based care identified using a survey. The study explored only collaborations between MDs, PAs, and NPs.
Primary care office managers were surveyed about the degree of team-based care in the office. If the office manager said that two of the providers in a practice “tended to share patients” that was identified as a team-based care dyad. Out of the 63 practices surveyed, the team identified 361 team-based care dyads (MD-MD, MD-PA, or MD-NP).
The researchers also used Medicare claims data to identify similar provider dyads. Team-based care was defined as two providers billing Medicare for services to the same patient. Interestingly, out of the 361 dyads identified through survey data, the Medicare definition of team-based care revealed only 252 of these dyads. And among those dyads, they reflected a median of only about 10-11 shared patients. That is very little shared care!
Trade-Offs in Identifying Team-Based Primary Care
The authors also considered what number of shared-care patients could be used to best predict the actual team-based care reported via survey. And, of course, they found trade-offs. With a cut-off of 9 shared-care patients, the Medicare data had a sensitivity of 52% (relatively poor), a specificity of 73% (relatively good), and a combined positive-predictive value of 60%. Changing the cut-off to 30 shared-care patients resulted in poorer sensitivity (28%), much better specificity (92%), and a positive predictive value of 72% (not much better overall).
These trade-offs suggest that spotting team-based primary care through claims data is possible, but imperfect. The authors acknowledge this. They suggest that if methods could be refined to better identify team-based care using claims, it would open the door to deeper evaluations of health care use, quality and costs. And the use of claims data only begins to capture the range of possible team-based care. It doesn’t offer an easy look at collaborative roles for MAs, RNs, and other non-billing providers. Techniques will need to be developed to quantify those relationships.
Gregory Stevens