Curbside Consult with Dr. Jayne 4/4/22

I spent some time this weekend at a non-healthcare, non-technology conference. It was nice to get away for a few days and spend time learning ways to improve my skills for one of my hobbies.

I’ve been attending this particular conference since 2018, and many of the attendees know about my past life in the emergency department. There were quite a few questions about COVID and whether I think it’s really over. I typically respond that I don’t think it will ever be over, but we’re learning how to cope with it in the US. Because our lives are back to normal, at least in part, many people have forgotten that there are other nations where people still haven’t had adequate opportunities to receive vaccines.

The Our World in Data website is one of my favorites. It shows that a high percentage of people in Africa have yet to receive even one dose of vaccine. It makes you think twice about living in a country where a large number of people still believe that COVID isn’t real and vaccines aren’t safe, despite there having been more than 11 billion doses administered worldwide.

I had some time to kill at the airport, so I participated in an online research study from Harvard University. The study was designed to evaluate strategies to influence vaccine-hesitant individuals to become up to date with the COVID vaccine schedule. Participants were educated on several strategies to try to persuade people to receive vaccines and then were asked to create narrative statements that they felt might work. Messages were to be in response to a patient who was concerned that the vaccine was rushed, that mRNA technology is too new, that fetal cells were used in vaccine development, and that vaccines cause death. The researchers plan to use a natural language processing algorithm to evaluate the messages, and which are best at demonstrating receptiveness. They also gathered data on the respondents’ perception of the concerned patient and whether they would be willing to interact with that person again, which I thought was interesting. I’ll have to keep my eye out for the results of the research in the future.

I also had time to read a study that was recently published and has been regarded as somewhat controversial. The Journal of the Mississippi State Medical Association published the study, “Targeting Value-Based Care with Physician-Led Care Teams” in its January issue. It details findings from Hattiesburg Clinic’s value-based care journey with its Accountable Care Organization. When cost of care was examined, the study revealed that care delivered by non-physician providers who were practicing independently was more expensive than care delivered by physicians. The findings led the Clinic to redesign its care model as well as to publish its findings. Multiple news outlets and physician organizations picked up on the article, leading to headlines about how midlevel practitioners just might not be the answer to the primary care physician shortage at all.

Looking at the organization’s journey, in 2005 it employed a combined total of 26 APPs (advanced practice providers), including nurse practitioners and physician assistants. Today it employs 118. Over the last 15 years, Hattiesburg Clinic had made decisions to expand care teams by allowing these providers to manage primary care patient panels on a largely independent basis. The Clinic has more than 33,000 Medicare beneficiaries and an associated Accountable Care Organization, so it was monitoring its outcomes carefully. The study found that by allowing APPs to operate independently, the organization “failed to meet our goals in the primary care setting of providing patients with an equivalent value-based experience.”

The authors looked at 2017-2019 CMS cost data on Medicare patients who did not have end-stage renal disease and who were not in a nursing home. The data showed that per member, per month spending was $43 higher for patients who had a non-physician in charge of their primary care needs. When applying risk adjustment factors for patient complexity, the difference was $119 per member, per month. Originally, the analysis was to help the organization identify high-cost providers so they could intervene. They didn’t expect the results they identified, including increased testing utilization, more specialist referrals, and more emergency department utilization for patients who were under non-physician care.

They also found that physicians performed better on nine of 10 quality measures, with notable differences in vaccination rates for influenza and pneumococcal disease. Physicians also had higher patient satisfaction scores across multiple domains measured via Press Ganey. Although they concluded that non-physician providers are valuable members of the care team, the organization determined that independent practice was not in the organization’s best interest. They then embarked on a year-long transition that would allow APPs to inform their patients that they would start seeing the supervising physician as well, and that the physician would become the primary care provider of record. Additionally, APPs in specialty areas were restricted from seeing new patient consultation visits except in emergencies or when approved by the referring physician.

There are some interesting factors to note with regard to the findings. First, the Hattiesburg Clinic is focused on value-based care. Their experiences may not translate to organizations that are still operating under a predominantly fee-for-service model. Under the value-based care model, excess testing and referrals cut into the organization’s bottom line, so there’s an inherent level of buy-in for operational changes. In a fee-for-service model, the organization can benefit from certain kinds of overutilization, which doesn’t encourage restricting services. Also interesting is the finding that the patients who had the best quality were those who had alternating visits with both the physician and the APP.

There are also some weaknesses in the study itself, including controlling for years of experience of the APPs compared to years of experience of the physicians, and any variation in the organization’s onboarding and training of different types of providers. Having worked with new and experienced nurse practitioners, physician assistants, and physicians, I’ve seen across the board that inexperience is directly related to the propensity to order increased testing and referrals. When you’ve seen a given clinical presentation hundreds or thousands of times, you’re likely to be more confident in your ability to manage the patient on your own and are also experienced enough to refine testing to the minimum necessary. The published writeup also doesn’t include enough information on the analysis to determine whether some of the differences were statistically significant.

It will be interesting to see if the authors submit their work for the additional scrutiny of one of the national journals and what the findings look like when they are subjected to additional statistical analysis. Although the findings seem dramatic, they underscore the need for critical reading and to determine whether findings are likely to be similar to other situations. There are hundreds of organizations across the country who have the same types of data as Hattiesburg Clinic, and it would be interesting to see whether they reach the same conclusions. We’ve entered an era where there is more healthcare quality and cost data at our fingertips than we’ve ever had, and it’s time to really start using it.

What does your organization think about Hattiesburg Clinic’s findings? Have you looked at this issue yourselves? Leave a comment or email me.

Email Dr. Jayne.