An Assessment of Nurse Practitioner Low-Value Care Use in Primary Care

ABSTRACT

Objectives: To establish baseline prevalence rates associated with nurse practitioner (NP) use of 3 of the most commonly observed primary care low-value-care (LVC) services and to examine whether practice location and patient characteristics impact NP LVC use.

Study Design: Cross-sectional, secondary analysis.

Methods: Data for 14,579 adult beneficiaries in the 2021 Merative MarketScan Commercial and Medicare databases in Arizona, Nevada, and New Mexico were analyzed. Outpatient claims associated with NP care were used to examine the use of low-value lumbar x-ray, antibiotics for acute upper respiratory infection (aURI), and routine electrocardiogram (ECG) as described by the Choosing Wisely initiative. International Statistical Classification of Diseases, Tenth Revision and Current Procedural Terminology codes were used to apply inclusion and exclusion criteria. Relationships between LVC use and the state where a beneficiary received care, rural-urban practice location, and beneficiary sex and age were examined.

Results: Prevalence rates of NP use of low-value lumbar x-ray (13%), aURI antibiotic (42%), and ECG (6%) were lower or relatively similar to those found in other studies. Older beneficiary age was significantly associated with more low-value ECGs used (P < .001), but when adults 45 years and older were examined, age no longer remained significantly related. No significant relationships between NP LVC use and practice location or beneficiary sex were found.

Conclusions: NP LVC use in primary care was lower or relatively similar compared with the general clinician population. MarketScan may underrepresent rural care, and the relationship between NP LVC use and rural-urban location should be reexamined using an alternative classification system. To deimplement NP LVC use, other factors, such as NP characteristics, must be explored.

Am J Manag Care. 2025;31(10):In Press

Takeaway Points

Primary care nurse practitioners (NPs) use low-value care (LVC) at lower or relatively similar rates compared with the general clinician population.

• Findings suggest that urban practice location, female beneficiary sex, and younger beneficiary age are not associated with NPs’ decisions to use LVC.
• LVC antibiotic prescribing for acute upper respiratory infection remains high among NPs, indicating the need to examine this service in more detail.
• Additional variables should be investigated to create NP-specific LVC deimplementation interventions.

Low-value care (LVC)—defined as any health care service that is unsupported by current evidence, offers no net benefit in specific clinical situations, and is associated with harmful patient outcomes and wasteful spending—is a significant problem in health care.1,2 Health care costs associated with LVC have been estimated to be more than $100 billion annually.3 LVC has been linked to increased risks of unnecessary surgery, overtreatment, and misdiagnoses.4,5 Eliminating LVC is a clear priority to reduce health care spending and to potentially allocate health care services, a finite and limited resource, to those who truly can benefit. Evaluating LVC is one way to identify particular areas for minimizing health care waste, thus possibly improving patient outcomes while reducing overall health care spending.

Recent evidence has revealed that overall LVC use among all health care providers in outpatient care settings decreased from 2014 to 2018 (677.8 to 632.7 per 1000 individuals), but some LVC had actually increased during this time: imaging for acute, uncomplicated lumbar pain; antibiotic prescriptions for acute upper respiratory infection (aURI); and electrocardiogram (ECG) use in low-risk patients.2,6,7 Notably, these services are associated with some of the most common reasons for primary care visits.8 Thus, LVC use persists.

However, the degree to which nurse practitioners (NPs) contribute to LVC prevalence remains unclear because many studies omitted analyses by provider type2,9 or combined NPs and physician assistants (PAs) into a single group for analysis.10,11 There are approximately 280,000 NPs practicing in the US,12 with around 73% practicing in primary care,13 a number large enough to warrant analysis of NPs’ role in LVC prevalence separately from other providers.

Within the literature, evidence indicates that LVC is used more frequently in urban settings14-16 and among younger14,17,18 or female17 patients. Although these factors may influence physician and overall clinician LVC use, it is currently unknown whether or how they impact NPs’ decisions to use LVC. Nevertheless, evidence suggests they might. For example, one study found that urban NPs required low-value pelvic examinations more frequently than rural NPs.19

Much of the recent focus on LVC use has been on deimplementation, but there is a need to understand the baseline prevalence of the most commonly used primary care services among NPs and what may influence their decisions to use these services. It was the intent of this research study to describe primary care NPs’ LVC practice patterns in terms of prevalence in Arizona, Nevada, and New Mexico and to investigate whether NPs’ LVC use significantly differs by the state where a beneficiary received care, rural-urban practice location, or patient sex or age.

METHODS

A cross-sectional, secondary data analysis of the 2021 Merative MarketScan administrative claims data was performed. MarketScan contains a nationally representative sample of employer-based, privately insured individuals in the US.20 Because MarketScan contains deidentified beneficiary data, this study was classified as exempt by the University of New Mexico’s Health Science Human Research Review Committee.

The 2021 outpatient claims data were used to create a sample for each LVC service, as each service was analyzed individually. Data from the previous year, 2020, were used to obtain health data required to apply exclusion criteria. International Statistical Classification of Diseases, Tenth Revision (ICD-10) and Current Procedural Terminology (CPT) codes were used to identify past medical and procedural history and low-value services.

Three primary care low-value services that had the potential to be used most frequently in the adult primary care setting were selected for individual analysis.1,7 These were x-rays for acute, uncomplicated lumbar pain21,22; antibiotic prescribing for aURI21; and routine, annual ECG use.2 ECG screenings associated with a general adult medical examination without abnormal findings14 were considered routine because it was assumed individuals experiencing related symptoms would schedule an acute care appointment. A low-value service was defined as the presence of a CPT claim within 1 month of the respective diagnosis, with only the first service in that time frame included in the analysis. LVC was assumed to have not occurred if a procedural claim was absent during that time.

We included adult beneficiaries (≥18 years) who received treatment by an NP in Arizona, Nevada, or New Mexico; were continuously enrolled in a health plan for 2020 and 2021; and had a claim for outpatient lumbar pain, aURI, or a routine medical exam. These states were selected because they are states that the researchers have some knowledge of, are located in the southwest US, and offer NPs full-practice authority upon initial licensure, thus minimizing the potential for physician influence. Only the first 2021 diagnosis claim per beneficiary per service was selected for inclusion. Using this criteria, all qualified 2020 and 2021 MarketScan Commercial Database and Medicare Database (Medicare supplemented; eg, Medicare Advantage) data were obtained for analysis.

Low-value lumbar x-ray included beneficiaries without a history of chronic back pain or lumbar imaging 6 months prior to their diagnosis claim22,23; low-value antibiotic prescribing for aURI consisted of beneficiaries without a history of HIV or AIDS2,9,24 1 year prior to their aURI claim; and low-value ECGs included beneficiaries without a history of cardiac disease or risk factors 1 year prior to their first 2021 routine medical exam14 (see eAppendix [available at ajmc.com]).

We used SPSS Statistics 29.1 (IBM) for statistical analysis. Descriptive statistical tests were performed, and inferential statistics were used to test the relationship between LVC use and state, rurality, and beneficiary sex and age. When the effect size was too small for detecting a statistically significant difference, an equivalently sized, random undersampling of the majority group was selected for analysis. The threshold for statistical significance was set at P < .05.

This research study intended to identify low-value lumbar x-ray, antibiotic prescribing for aURI, and ECG prevalence among NPs, and it tested the hypothesis that urban practice location (defined as metropolitan statistical area [MSA]), female beneficiary sex, and younger beneficiary age would be associated with significantly more LVC use by NPs.

RESULTS

Sample distributions were similar for all samples, with most beneficiaries being female (62%-66%) and treated in urban areas (95%-96%). Mean beneficiary age was 39.32 to 44.64 years (Table 1). The majority of beneficiaries received care in Arizona (70%-73%); however, similar distributions were observed when data were analyzed by individual state.

Prevalence of NP use of low-value lumbar x-ray was 13% (n = 1820), antibiotic prescribing for aURI was 42% (n = 7610), and ECG was 6% (n = 5149). For low-value ECG use, when all beneficiaries were examined, as expected, older beneficiaries (mean [SD] age, 45.31 [12.1] years) were more likely than younger beneficiaries (38.63 [13.25] years) to receive a low-value ECG (t344.83 = –9.23; P < .001; Cohen d = .53). As older age groups were examined (beneficiaries ≥ 45 years), age no longer remained significantly related (P = .16). Age was not statistically significantly related to low-value lumbar x-ray use (P = .14). Age was statistically significantly related to low-value aURI antibiotic use (t6790.59 = –5.14; P < .001), yet the effect size was small (Cohen d = .12) and the mean difference (1.69 years) was not clinically meaningful (Table 2 and Table 3).

The state where a beneficiary received care was not significantly related to low-value lumbar x-ray (P = .14) or antibiotic for aURI (P = .6). MSA was not significantly related to low-value lumbar x-ray (P = .15) or antibiotic for aURI (P = .47). Beneficiary sex was not significantly related to low-value lumbar x-ray (P = .18) or ECG (P = .13). The minimum expected cell count for the χ2 test of independence was violated for all remaining relationships. Based on findings from services that did meet expected 20% minimum cell counts, there is no evidence to support that, if expected minimum counts were obtained, significant differences would be found.

DISCUSSION

Primary care NPs were found to use low-value lumbar x-rays at a rate of 13%, lower than the 19% prevalence found among NPs using 2013 Medicare data.22,23 Using the same NP data in both of these previous studies, no difference in low-value x-ray use was observed between NPs and physicians,22 but NPs were less likely than PAs to use x-rays.23 Using National Ambulatory Medical Care Survey (NAMCS) data, Mafi et al also found no significant differences in x-ray use among NPs, physicians, and PAs.21 On the contrary, 2 studies reported much lower low-value lumbar imaging use (x-ray, CT, and MRI) by all health care clinicians among Medicare beneficiaries—at 0.2% in 20182 and 2% in 2016.9 In both studies, the Milliman MedInsight Health Waste Calculator, an algorithm-based software program, was used to classify services as wasteful, likely wasteful, or not wasteful.2,9 Services were only considered low value when they were classified as wasteful, and excluding the likely wasteful category may explain why prevalence was much lower in these studies. Differences in prevalence may also exist due to different patient populations examined or methodologies used, or it could indicate decreasing prevalence among NPs. Additional studies examining low-value lumbar x-ray use are indicated.

Low-value antibiotic use for aURI was 42% among primary care NPs in this study. This was much lower than the 80% prevalence rate of low-value antibiotic prescribing for sinusitis estimated by the American Academy of Family Physicians (AAFP).6 When we isolated beneficiaries diagnosed with acute sinusitis from the total sample, prevalence increased to 64%, which is lower than the AAFP’s estimate but still a troubling figure. Even after accounting for the possibility of misclassified antibiotic prescriptions, low-value antibiotic use for aURI among NPs remains high. A study using 2006-2012 NAMCS and National Hospital Ambulatory Medical Care Survey data found low-value antibiotic prescribing prevalence to be 55%.24 Although that study reported a higher prevalence than the current study, it found no significant differences among NPs, physicians, and PAs, echoing results from the study by Mafi et al.21 Low-value antibiotic prescribing for aURI among NPs remains high but could be trending down, and more research is needed to confirm this.

Studies using the Health Waste Calculator reported a lower prevalence of low-value antibiotic prescribing, at 8%2 and 12%.9 It is unclear which diagnoses were used to create their samples because only acute sinusitis, URI, and viral respiratory illness were listed. For the current study, an exhaustive list of aURI diagnoses, including rhinitis (common cold), coryza, sinusitis, pharyngitis, tonsilitis, and laryngitis, was used.25 Furthermore, these 2 studies used concurrent infections as an exclusion criterion. Due to the considerable number of ICD-10 infection diagnoses, it would have been futile to attempt to capture all unique codes needed to apply this exclusion criterion. However, it is suspected that even if beneficiaries with concurrent infections were excluded, low-value antibiotic prescribing would not have decreased by 30%, leading to the conclusion that prevalence differences may be due to methodological differences (ie, Health Waste Calculator).

Primary care NPs were found to use low-value ECGs at a rate of 6% in this study. Prevalence was similar to that reported in another study, which found that 7% of Medicare patients getting a routine medical exam in 2014 received a low-value ECG.14 The 2 studies that used the Health Waste Calculator reported similar, yet slightly lower, prevalence rates at 2%2 and 4%.9 Rather than excluding individuals with cardiac disease and risk factors, individuals in these studies were retained in the analysis, and ECGs used in this group were classified as not wasteful, possibly artificially lowering prevalence rates.2,9 According to the AAFP, annual ECGs should not be used in the low-risk patient population,4 which is why, in the current study, high-risk individuals were excluded prior to the analysis, similar to the study by Ganguli et al.14 It is plausible that NPs use low-value ECGs at similar rates to other providers, but more evidence is needed to substantiate these findings.

Although sufficient evidence suggested that NPs would use more LVC in urban areas,9,14,19 this hypothesis was not supported using the MarketScan data. We suspect that the MSA classification system underrepresented rural NP care, which may have contributed to the nonsignificant findings. Another study, using an alternative rural and urban classification system such as the US Department of Agriculture’s Rural-Urban Continuum Codes, should reexamine this hypothesis.

It was hypothesized that female patients would be more likely than male patients to receive LVC from NPs. Despite living in rural areas, which had previously been associated with less LVC use, female Medicaid beneficiaries aged 18 to 64 years were more likely to receive low-value lumbar imaging.17 Our study, however, found no statistically significant relationship between beneficiary sex and LVC use. Based on the limited evidence in the literature and findings from this study, it is possible that the results from the study by Charlesworth et al17 were an isolated occurrence.

Younger beneficiary age was not found to be associated with more LVC used by NPs. It is possible that the role of younger age as a predictor of LVC use may only apply to certain populations, such as Medicare14 or Medicaid beneficiaries18; certain types of LVC, such as lumbar MRI18; or composite measures of multiple LVC services.14 Based on these findings, it may be more appropriate to test relationships between individual LVC services rather than creating composites of multiple services, as each LVC service is unique.

Limitations

It is possible that appropriate services were misclassified as low value. LVC may have also been underestimated because administrative claims were used as a proxy for NP prescribing; beneficiaries may have been prescribed a service but chose not to carry out that service. MarketScan does not include claims data for Medicaid, Medicare only, or individuals who are uninsured. However, this was a previously unexamined population (employer-based, commercially insured) and its benefits outweighed its drawbacks. It is plausible that some NP care was missed due to incident-to billing methods that allow an NP to bill under a physician. Additionally, although states offering full-practice authority upon initial licensure were selected, it is impossible to identify whether a physician was involved in decision-making related to LVC use. Lastly, results from this study may not be generalizable to regions outside the Southwest or to states that do not offer NPs full-practice authority upon initial licensure.

Implications for Future Research

Although relatively lower than in other studies, low-value antibiotic prescribing for aURI among NPs remains high, indicating the need to study this low-value service in more detail. Future studies must focus on identifying and testing other variables that may influence LVC use among NPs, such as NP characteristics (eg, the state where an NP received their education, years of registered nurse and NP experience) and other beneficiary characteristics (eg, ethnicity, mental health status). Qualitative inquiry should also be used to isolate previously unexamined variables that could lead to a better understanding of why NPs use or do not use LVC. This way, LVC deimplementation efforts can be specifically targeted to concerns in the NP population.

CONCLUSIONS

It was the intent of this research to establish baseline prevalence rates for some of the most common low-value primary care services among NPs and to examine whether certain geographical and beneficiary characteristics influenced NPs’ practice patterns. For all 3 services, NPs were found to use LVC at similar or less frequent rates compared with the general provider population or earlier studies examining NP practice. These results further demonstrate NPs’ ability to provide safe, cost-effective care. Equally important as significant findings, nonsignificant relationships must also be published because they reveal variables known to not influence NP LVC use. The 2021 MarketScan data did not support the association between urban practice location, female beneficiary sex, or younger beneficiary age with more NP LVC use. It is recommended that the influence of urban practice location be reexamined using an alternative rurality measure because the MSA classification system may have underrepresented rural NP care in this sample. Finally, results from this study can be used to track future trends in LVC use among primary care NPs.

Acknowledgments

Merative and MarketScan are trademarks of Merative Corporation in the United States, other countries, or both.

Author Affiliations: University of New Mexico College of Nursing (SBN, RPL, JL, BIHD), Albuquerque, NM; University of New Mexico Department of Economics (BPH), Albuquerque, NM.

Source of Funding: The University of New Mexico Graduate Studies Rogers Research Project Grant.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (SBN, RPL, BPH, BIHD); acquisition of data (SBN); analysis and interpretation of data (SBN, JL, BPH, BIHD); drafting of the manuscript (SBN, JL, BIHD); critical revision of the manuscript for important intellectual content (SBN, RPL, BPH, BIHD); statistical analysis (SBN, JL); provision of patients or study materials (SBN, RPL); obtaining funding (SBN); administrative, technical, or logistic support (BIHD); and supervision (RPL, BPH, BIHD).

Address Correspondence to: Sara B. Nugent, PhD, RN, University of New Mexico College of Nursing, 1650 University Blvd NE, Albuquerque, NM 87102. Email: snugent@salud.unm.edu.

REFERENCES

1. Choosing Wisely. ABIM Foundation. Accessed September 1, 2023. https://www.choosingwisely.org/

2. Mafi JN, Reid RO, Baseman LH, et al. Trends in low-value health service use and spending in the US Medicare fee-for-service program, 2014-2018. JAMA Netw Open. 2021;4(2):e2037328. doi:10.1001/jamanetworkopen.2020.37328

3. Shrank WH, Rogstad TL, Parekh N. Waste in the US health care system: estimated costs and potential for savings. JAMA. 2019;322(15):1501-1509. doi:10.1001/jama.2019.13978

4. Annual EKGs for low-risk patients. American Academy of Family Physicians. Accessed September 12, 2023. https://www.aafp.org/family-physician/patient-care/clinical-recommendations/all-clinical-recommendations/cw-ekg.html

5. Srinivas SV, Deyo RA, Berger ZD. Application of “less is more” to low back pain. Arch Intern Med. 2012;172(13):1016-1020. doi:10.1001/archinternmed.2012.1838

6. Antibiotics for sinusitis. American Academy of Family Physicians. Accessed September 12, 2023. https://www.aafp.org/family-physician/patient-care/clinical-recommendations/all-clinical-recommendations/cw-sinusitis.html

7. Imaging for low back pain. American Academy of Family Physicians. Accessed September 12, 2023. https://www.aafp.org/family-physician/patient-care/clinical-recommendations/all-clinical-recommendations/cw-back-pain.html

8. Finley CR, Chan DS, Garrison S, et al. What are the most common conditions in primary care? Can Fam Physician. 2018;64(11):832-840. https://www.cfp.ca/content/64/11/832/tab-article-info

9. Ganguli I, Morden NE, Yang CWW, Crawford M, Colla CH. Low-value care at the actionable level of individual health systems. JAMA Intern Med. 2021;181(11):1490-1500. doi:10.1001/jamainternmed.2021.5531

10. Degnan KO, Cluzet V, David MZ, et al; CDC Prevention Epicenters Program. Development and validation of antibiotic stewardship metrics for outpatient respiratory tract diagnoses and association of provider characteristics with inappropriate prescribing. Infect Control Hosp Epidemiol. 2022;43(1):56-63. doi:10.1017/ice.2021.44

11. Hwang S, Birken SA, Nielsen ME, Elston-Lafata J, Wheeler SB, Spees LP. Understanding the multilevel determinants of clinicians’ imaging decision-making: setting the stage for de-implementation of low-value imaging. BMC Health Serv Res. 2022;22(1):1232. doi:10.1186/s12913-022-08600-3

12. Nurse practitioners. US Bureau of Labor Statistics. May 2023. Accessed July 1, 2023. https://www.bls.gov/oes/2023/may/oes291171.htm

13. NP fact sheet. American Academy of Nurse Practitioners. Updated November 2022. Accessed July 1, 2023. https://web.archive.org/web/20230603022442/https://www.aanp.org/about/all-about-nps/np-fact-sheet

14. Ganguli I, Lupo C, Mainor AJ, et al. Assessment of prevalence and cost of care cascades after routine testing during the Medicare annual wellness visit. JAMA Netw Open. 2020;3(12):e2029891. doi:10.1001/jamanetworkopen.2020.29891

15. Oronce CIA, Fortuna RJ. Differences in rates of high-value and low-value care between community health centers and private practices. J Gen Intern Med. 2020;35(4):994-1000. doi:10.1007/s11606-019-05544-z

16. Scott A, Li J, Gravelle H, McGrail M. Physician competition and low-value health care. Am J Health Econ. 2022;8(2):252-274. doi:10.1086/718509

17. Charlesworth CJ, Meath TH, Schwartz AL, McConnell KJ. Comparison of low-value care in Medicaid vs commercially insured populations. JAMA Intern Med. 2016;176(7):998-1004. doi:10.1001/jamainternmed.2016.2086

18. Gidwani R, Sinnott P, Avoundjian T, Lo J, Asch SM, Barnett PG. Inappropriate ordering of lumbar spine magnetic resonance imaging: are providers Choosing Wisely? Am J Manag Care. 2016;22(2):e68-e76.

19. Henderson JT, Sawaya GF, Blum M, Stratton L, Harper CC. Pelvic examinations and access to oral hormonal contraception. Obstet Gynecol. 2010;116(6):1257-1264. doi:10.1097/AOG.0b013e3181fb540f

20. MarketScan for life sciences researchers. Merative. Accessed October 3, 2023. https://www.merative.com/documents/brief/marketscan-research-databases-for-life-sciences-researchers

21. Mafi JN, Wee CC, Davis RB, Landon BE. Comparing use of low-value health care services among U.S. advanced practice clinicians and physicians. Ann Intern Med. 2016;165(4):237-244. doi:10.7326/M15-2152

22. O’Reilly-Jacob M, Perloff J, Buerhaus P. Low-value back imaging in the care of Medicare beneficiaries: a comparison of nurse practitioners and physician assistants. Med Care Res Rev. 2021;78(3):197-207. doi:10.1177/1077558719877796

23. O’Reilly-Jacob M, Perloff J, Buerhaus P. Comparing the rates of low-value back images ordered by physicians and nurse practitioners for Medicare beneficiaries in primary care. Nurs Outlook. 2019;67(6):713-724. doi:10.1016/j.outlook.2019.05.005

24. Jiao S, Murimi IB, Stafford RS, Mojtabai R, Alexander GC. Quality of prescribing by physicians, nurse practitioners, and physician assistants in the United States. Pharmacotherapy. 2018;38(4):417-427. doi:10.1002/phar.2095

25. Thomas M, Bomar PA. Upper respiratory tract infection. In: StatPearls. StatPearls Publishing; 2023. Accessed October 31, 2023. http://www.ncbi.nlm.nih.gov/books/NBK532961/