Study the study before you cite the site

By Peter McMenamin posted 05-13-2014 15:14


Nurses do not pretend to be physicians

Nurse practitioners in New York will be able to operate more independently of doctors under a bill slated to become law as part of the state budget enacted April 1, 2014.  The Nurse Practitioners Modernization Act will remove the requirement of a written practice agreement between an experienced nurse practitioner and a doctor as a condition of practice. The law will take effect Jan. 1, 2015.  The Medical Society of the State of New York (MSSNY) strongly opposed the independent practice of a nurse practitioner without a collaborative agreement with a physician.  MSSNY referred to studies that showed that increasing the use of NPs does not lower costs as the patients of NPs tended to have higher rates of medical service utilization.  This argument was further pursued in a New York Times op-ed that appeared April 30, 2014 authored by Dr. Sandeep Jauhar.  In his article Dr. Jauhar cites a single article from 1999 that had concluded NP patients received more diagnostic tests, specialty visits, and hospital admissions than patients assigned to physicians. According to the study, those utilization differences “may offset or negate any cost savings achieved by hiring nurse practitioners in place of physicians.”  As it happens, there appears to be only a single published study in this regard, and that study is far from compelling.  As illustrated below, there were too many statistical and operational loose ends with respect to its execution and interpretation to place much reliance on the results. 

A major question involving choice between two alternative clinicians is whether each is competent for the job at hand.  Tamara Zurakowski, PhD, CRNP, RN, a faculty member from the University of Pennsylvania notes that a nuclear physicist has substantially more training and knows a great deal more about the production of electrical energy than a licensed electrician does.  But when wiring in a house needs fixing the first choice is rarely the physicist. The electrician's combination of knowledge and skill is a better match for the job at hand.  In this case, the physicist’s additional years of education would just amount to gilding the lily.

In this regard, a continuing staple of medical education is expressed in the notion, “When you hear hoof beats, think horses, not zebras.”  There are very complicated physical problems that a patient might present, but those tend to be relatively rare.  Only TV’s Dr. Gregory House found a zebra 100% of the time.  (In fact, with respect to cost control, the more recent question has been was there a House in the doctor?)  Regardless, every time the discussion turns to NPs and primary care, there will be at least one MD that feels compelled to tell a zebra story.  For example, Dr. Jauhar’s zebra invocation of goiter rather than simple hoarseness, a distinction that required “a doctor’s expertise.”

Dr. Jauhar also invokes a study published prior to the turn of the 21st century.  He concludes that that study found that NPs practice more expensive care than MDs.  Arguably randomized, the study was definitely not well controlled.  It appeared to be true that patients assigned to the “NP wing” of the study received more tests, although the limited information contained in the published study reported very few differences as statistically significant.  And, as Tay Kopanos from the American Association of Nurse Practitioners pointed out, the study did not track the identities of ordering clinicians so there was no specific evidence that NPs had actually ordered more tests, despite Dr. Jauhar’s assertion.  In fact, the authors mention that some patients were reassigned to different clinicians than their original assignment, but for study purposes those patients were deemed to continue in their original wing.  Mainly the language used in the study is precise, “utilization of patients assigned to nurse practitioners.”  But the authors repeatedly assert that “nurse practitioners had higher utilization rates.” 

Actually, there were all too many things that the study authors indicated that they did not do.  They considered multivariate analyses, but chose to proceed only with simple two way comparisons for the various measures, not controlling for covariates.  They did not appear to examine variations in utilization per patient where some patients used a service many times while other patients did not use the service at all.  That makes comparisons of average values in the different wings problematic.  They did not track the specialties of the clinicians that actually ordered lab and radiological tests, although tests and referrals at the Maryland VA study site could be ordered by any practitioner, not just the assigned primary care provider.  Despite the site being a Veteran’s Affairs clinic, there was no mention of lack of customary financial incentives that might otherwise have personally affected the ordering of services by clinicians or the lack of patient cost-sharing that might otherwise have affected clinician ordering and patient compliance.  The authors did note that it was important to emphasize that the study evaluated the practices of nurse practitioners who were supervised full-time by "dedicated attending physicians."  In this regard, the authors suggested that attending physicians might have been uncomfortable with the uncertainty inherent in joint management of patients with nurse practitioners and recommended ordering additional tests as a result. However, the authors failed entirely to take note of scope of practice restrictions that might have limited the nurse practitioners’ ability to directly care for their patients thus requiring additional referrals.   The authors separately hypothesized that increased specialist referrals might lead to increased use of lab and radiological tests “. . . i.e., the tests were ordered by specialists and not the primary care provider.”

Dr. Jauhar asserts that the most plausible reason NPs might have appeared to order more test is that they did so to compensate for a lack of clinical training.  He provides no additional documentation for this inference.  He glosses over that the ordering clinician was not recorded for any service so the genuine source or sources for arguably excessive ordering were unknown.  Dr. Jauhar omits any mention of the pedagogical inclinations of supervising physicians, attending physicians, pressures from VA medical directors for attendings to order fewer tests, or the possibility that NPs might appropriately refer more patients to specialist physicians compared to residents or attendings who may have perceived they were qualified to manage care for any patients.  There was no discussion in the original study about whether observed test ordering in any of the arms of the study was above or below existing standards of practice.  Was more better or worse than less?  We do not have a measure upon which to rely.

The study authors concluded, “. . . . our study suggests that the use of nurse practitioners in certain primary care settings may increase the utilization of health services, particularly specialty and inpatient care.” The authors seemed predisposed to believe that NPs in independent practice would order even more tests, in particular: “. . . . the presence of supervising physicians probably influenced their [NP] practice and brought it closer to physician norms. If indeed this was the case, then the differences that did exist may be magnified in an independent nurse practitioner practice . . . .”  However, they did not further conjecture whether bringing NP practice closer to physician norms involved increasing rather than decreasing patient utilization. 

There was no compelling case presented in the study that NPs increased health services utilization and there was no serious attempt to verify that inference.  Cost of care estimates of the various wings of the study were entirely absent despite the controlled (and single payer) environment in which the study took place.  Sadly, although the authors mentioned the need to study whether reported differences in utilization affected the quality of care, there was no attempt to examine whether any of the wings of the study adhered to existing standards of care.  Readers were given no guidance on whether the increased use of care by the patients in the NP wing brought their care closer to such standards or exceeded the standards.

With all the studies published over the last 15 years that Dr. Jauhar might have reviewed he chose one that is far from compelling.  In fact, there are no carefully controlled studies that provide information on the relative costs of NP provision or ordering of services to patients compared to the services provided or ordered for similar patient by physicians.  What we do know is that through the Balanced Budget Act of 1997 the Congress mandated a gender-associated pay differential of 15%; at that time 97% of registered nurses were women and more than 80% of physicians were men.  As then specified in Medicare’s section 1833(a)(1)O, for the identically billed clinical services (identified by specific CPT codes) NPs and CNSs were limited to 15% less than the approved amount to be paid to physicians.  The gender proportions have changed somewhat since that time but the differential remains.  It was not imposed because NPs were believed to order more tests. 

It is to physicians’ continuing shame that having accepted the RBRVS notions that any specific service has a single specific worth, that they continue to assert that NPs and CNSs should be paid less for the identical service.  There is added value of APRN services in that they do convey more patient information and education than MDs.  APRNs do offer services beyond traditional office hours, serving patients and families that might not otherwise be able to secure primary care services.  All of these observations were verified by an extensive two year study conducted by the Institute of Medicine and published in 2010.  There should be no question that NPs, in particular, are well qualified primary care providers.



02-04-2015 22:55

It is comforting to learn about the Nurse Practitioners Modernization Act (NPMA) effective on January 2015 in the state of New York. It is important to clarify that NPMA only applies to experience NPs with more than 3600 hours of practice or two years. The Nurse Practitioners (NPs) have demonstrated in numerous studies that they are excellent primary care providers. Moreover, the NPs greatly compensate for the shortage of primary care physicians in rural areas. As of today, there are one third of all states that allow NPs to function without a physician’s supervision. It is about time for the American Nurses Association (ANA), the American Association of Nurse Practitioners (AANP) to organize a national awareness campaign and start to work with the state members and especially the members of the United States Congress and their staff to create laws like the NPMA to be effective at the national level.

05-15-2014 15:56

Very compelling article! This trend is sure to start spreading