Study: Patients More Satisfied with Specialists who Trained with Primary Care Physicians

Maximilian J. Pany, BA

Primary care physician and specialist co-training is associated with significantly improved patient experiences with specialists, according to findings from a quasi-experimental trial.

In new outcomes from a difference-in-differences analysis of electronic health record (EHR) data from Harvard Medical School, a team of Boston-based investigators observed better patient satisfaction per Press Ganey’s Medical Practice Surveys regarding their specialist care when specialists had co-trained with primary care physicians in the past.

The findings would suggest more strategies that encourage forming greater peer-to-peer relationships among physicians would continue to improve the patient experience.

Led by Maximilian J. Pany, BA, of the Department of Health Care Policy at Harvard Medical School, investigators sought to determine the role of peer relationships between primary care physicians and specialists formed during training. As they noted, physician behavior is a “primary driver of patient outcomes and health care spending,” with peer observation and approval potentially serving as a significant motivator in medicine.

“Accordingly, an audience of familiar peers may elevate physician performance, not only by subjecting it to informed scrutiny, but also by providing an opportunity to demonstrate commitment to what is valued by the profession,” investigators wrote. “If strong, the motivational effects of peer interaction could have profound implications for the organization of care delivery, including potential gains from models encouraging peer familiarity and visibility—gains that could accrue over many dimensions of care without requiring decision-specific interventions.”

Pany and colleagues used EHR data from 2016 – 2019 based on completed specialist referrals from primary care physicians to compare patient ratings of specialist care between those seen by specialists who trained with their primary care provider either in medical school or postgraduate medical programs, and those seen by a specialist that did not train with the same primary care provider.

The analysis included 8655 patients undergoing 9920 specialist visits; of those, 5562 resulted from directed referrals. Mean patient age was 57.4 years old and 62.9% were female. Patients had seen any of 502 specialists working across 13 different specialties. Only 3.1% (n = 306) of specialist visits involved a primary care physician-specialist dyad with overlap in training.

The team observed an 8.3 percentage point increase in mean composite survey rating of specialist care among patients to be treated by a specialist who co-trained with their primary care physician (95% CI, 4.9 – 11.8; P <.001). When adjusting for patient covariates, as well as for the year of treatment and physicians’ average survey ratings, the association increased to a 9.0 percentage point benefit (95% CI, 5.6 – 12.4; P <.001).

“This difference corresponds to an effect size of 1.31 SDs of the specialist-level distribution in composite ratings, analogous to an improvement from median performance among specialists to the 91st percentile,” investigators emphasized. “Item-specific analyses revealed consistently strong associations between co-training and patient experiences for 9 of 10 survey items.”

Investigators additionally observed similar estimated benefit from co-training familiarity among patients being treated by specialists on an undirected referral from their primary care physician.

What’s more, Pany and colleagues saw a 1.6 percentage point increase in medicine prescriptions for patients who were treated by specialists who co-trained with their primary care providers (95% CI, 0.3 – 2.9), as well as a 1.2 percentage point increase in images ordered during the specialist appointment (95% CI, -0.7 to 3.0).

The team believed these findings show that greater patient feedback was specific to the relationship between primary care physicians and specialists—as opposed to singular physician- or patient-specific factors. Such a conclusion is supported by the sustained improvement of scores among patients treated by a co-trained health care team who indirectly referred to specialists.

“The estimated effects of co-training on patient ratings of specialist care are much larger than the effects of other policies, interventions, or efforts to improve patient experiences, including public reporting, accountable care organizations, health plan effects, and hospital characteristics,” they wrote. “Accordingly, the present study’s findings, and extensions thereof, could have major implications for the organization of care delivery, as they suggest potentially large quality gains from models encouraging peer interactions—such as team-based care, digital consultations that foster direct communication, peer coaching, and multispecialty case discussions.”

Investigators concluded that peer relationships between collaborative care team members and physicians may broadly improve quality of care and even help to mitigate health care-related shortcomings such as treatment disparity.

“More generally, our study suggests that nonfinancial strategies to harness physician professionalism, including the pressure to live up to the expectations of colleagues, could generate gains over many dimensions of care without requiring decision-specific interventions that risk erosion of intrinsic motivation when deployed in aggregate,” they wrote.

The study, “Physician-Peer Relationships and Patient Experiences With Specialist Care,” was published online in JAMA Internal Medicine.