NR 506NP Week 4 NP Outcomes Evidence Review Paper Example

Outcomes Evidence on Preventable Hospitalizations Under Nurse Practitioner and Physician Primary Care: One Randomized Trial, a VA Reassignment Study, Two Claims Analyses, and Why Self-Selection Decides What Each Can Support

Student Name

College of Nursing, Chamberlain University

NR 506NP: Healthcare Policy and Leadership

Instructor Name

Month Day, Year

Outcomes Evidence on Preventable Hospitalizations Under Nurse Practitioner and Physician Primary Care: One Randomized Trial, a VA Reassignment Study, Two Claims Analyses, and Why Self-Selection Decides What Each Can Support

One Outcome, One Comparison

Arguments over AB 890 cite "NP outcomes" as if there were one body of evidence. There are many outcomes, and studies of one do not answer questions about another. This review holds to one outcome, hospitalizations and acute care use that good primary care should prevent, and one comparison: patients whose primary care provider is a nurse practitioner against patients whose provider is a physician. It groups studies by what their design can support, because the other side in this debate has read the same studies and knows the weaknesses of each.

Hospitalizations were chosen because they matter to patients and payers alike and because both sides in California cite them. A preventable admission for uncontrolled diabetes or heart failure is a direct sign that outpatient care failed, and it is recorded in claims data in the same way whoever the primary care provider was, which makes it one of the few outcomes that can be compared across provider types without relying on chart review.

Studies by Data Source and Adjustment

Table 1 lists the four studies reviewed, each with its data source and how it dealt with differences between patient groups.

Table 1

Four Studies of Acute Care Use by Primary Care Provider Type

Note. OR = odds ratio. PA = physician assistant. VA = Veterans Affairs.

Self-Selection as the Main Problem

Every observational comparison of NP and physician patients faces the same threat: the patients are not the same people. Patients who choose an NP, or who are steered to one, may be younger, healthier or less complex, and if they are, they will be hospitalized less often regardless of who treats them. Kuo et al. (2015) state this directly: patients receiving all their care from NPs differed from physicians' patients by age, sex, race and ethnicity, socioeconomic status, residence and previous visits. Their propensity score matching substantially reduced those differences, but their instrumental variable approaches, designed to handle differences the data could not see, did not satisfactorily balance the groups. The odds ratio of 0.90 is therefore an association after adjustment for what Medicare claims record, and no more.

The designs in the table differ mainly in how they handle this problem. Adjustment for recorded characteristics, as in the Morgan and Kuo studies, handles only what is recorded. The Liu study used a natural experiment: patients whose physician left the VA were reassigned to a new provider, NP or physician, largely according to who had openings rather than patient choice. That reduces self-selection considerably, though it does not remove the possibility that managers assigned sicker patients to physicians. The Mundinger trial removed self-selection by randomizing.

What Each Supports Against What It Is Quoted For

Mundinger et al. (2000) supports the claim that, for adults without a regular source of care in one New York academic system in the 1990s, patients randomized to NPs used services no differently from those randomized to physicians. It is often quoted as proof that NPs are equal to physicians everywhere. It cannot support that: it was one setting, most patients were Hispanic women in their forties, and follow-up for service use was one year.

Liu et al. (2020) supports the claim that, in the VA, patients reassigned to NPs had fewer hospitalizations at similar cost and similar diabetes, cholesterol and blood pressure control. It is the strongest observational evidence in the set. It is quoted for private practice, where NPs are not embedded in an integrated system with shared records and specialists; the VA context may be part of the result.

Kuo et al. (2015) supports a modest association in Medicare patients with diabetes. It is quoted as showing that NPs reduce hospitalizations, which its own analysis does not claim.

Morgan et al. (2019) supports lower costs for complex VA patients with diabetes cared for by NPs and PAs together. It is often quoted for NPs alone, although it pooled NPs and PAs in its headline finding.

Where the Other Side Is Right

Opponents are right that none of these studies tests what AB 890 creates: an NP practicing alone, outside any system, after the minimum required years. The VA studies involve NPs working in teams inside a single health system. The trial involved NPs with the same authority as physicians in the same academic center. Evidence about independent practice in private settings, with patients who choose their provider, is thin. They are also right that other outcomes tell a different story in places: in Medicare claims, advanced practice clinicians ordered imaging more often after office visits than primary care physicians did (Hughes et al., 2015). A review limited to hospitalizations should not be quoted as if it covered every measure of quality or cost.

The Study That Would Settle It

The question California now faces could be answered with a study the state is well placed to run. Starting in 2023, some NPs began practicing under the new sections while others continued under standardized procedures. A study linking Medi-Cal claims to the Board of Registered Nursing's certification records could compare preventable hospitalizations among patients of 103 and 104 NPs, standard NPs and physicians in the same counties, before and after each NP's certification date. Better still would be random assignment of new Medi-Cal enrollees who do not choose a provider, which managed care plans already do, to primary care panels led by NPs or physicians. That would give the evidence the other side asks for, and it would give it for the practice the law actually allows.

References

Hughes, D. R., Jiang, M., & Duszak, R., Jr. (2015). A comparison of diagnostic imaging ordering patterns between advanced practice clinicians and primary care physicians following office-based evaluation and management visits. JAMA Internal Medicine, 175(1), 101-107. https://doi.org/10.1001/jamainternmed.2014.6349

Kuo, Y.-F., Chen, N.-W., Baillargeon, J., Raji, M. A., & Goodwin, J. S. (2015). Potentially preventable hospitalizations in Medicare patients with diabetes: A comparison of primary care provided by nurse practitioners versus physicians. Medical Care, 53(9), 776-783. https://doi.org/10.1097/MLR.0000000000000406

Liu, C.-F., Hebert, P. L., Douglas, J. H., Neely, E. L., Sulc, C. A., Reddy, A., Sales, A. E., & Wong, E. S. (2020). Outcomes of primary care delivery by nurse practitioners: Utilization, cost, and quality of care. Health Services Research, 55(2), 178-189. https://doi.org/10.1111/1475-6773.13246

Morgan, P. A., Smith, V. A., Berkowitz, T. S. Z., Edelman, D., Van Houtven, C. H., Woolson, S. L., Hendrix, C. C., Everett, C. M., White, B. S., & Jackson, G. L. (2019). Impact of physicians, nurse practitioners, and physician assistants on utilization and costs for complex patients. Health Affairs, 38(6), 1028-1036. https://doi.org/10.1377/hlthaff.2019.00014

Mundinger, M. O., Kane, R. L., Lenz, E. R., Totten, A. M., Tsai, W.-Y., Cleary, P. D., Friedewald, W. T., Siu, A. L., & Shelanski, M. L. (2000). Primary care outcomes in patients treated by nurse practitioners or physicians: A randomized trial. JAMA, 283(1), 59-68. https://doi.org/10.1001/jama.283.1.59