Hot off the press: SGEM#308: Taking care of patients every day with physician assistants and nurse practitioners

The use of advanced practice providers (APPs) in U.S. emergency departments (EDs) has expanded in recent years, with about double the ratio of patients seen by APPs in 2016 than in 2008.1-3 APP scope of practice has also expanded to include more high-acuity patients.4 Little data exist regarding ED-based APPs when it comes to productivity, safety, and flow. The studies that have been published either are single center or use national databases.5 This study is set in multiple hospitals in a large ED group and compares APP productivity, safety, and flow metrics to those of ED physicians at the same sites. This study is a planned secondary analysis of a preexisting database from a national emergency medicine group and includes data from 94 hospitals in 19 states and includes approximately 13 million ED visits. For scheduling, 1 hour of physician coverage was equivalent to 2 hours of APP coverage. Outcomes included productivity (patients/hour, RVUs/hour, RVUs/visit, RVUs/relative salary), flow (length of stay—LOS, proportion of patients who left without completion of treatment—LWOT). In addition, exploratory outcomes included incident reports and Press-Ganey percentile rank. Over 13 million patient visits were included, and physicians had higher patients/hour and RVUs/hour than APPs, with small decreases in these associated with increasing APP hours. No effects were seen on safety and flow outcomes. This study is strengthened by the large number of ED visits included from a number of hospitals in numerous states. The data are robust, and the outcomes are specific, of interest, and applicable to EM providers. One potential weakness of the study is external generalizability. These hospitals are all part of the same system, and there is likely to be training, onboarding, and ongoing education that may differ from other systems. In addition, these 94 hospitals consist of a variety of settings, including trauma centers, community hospitals, and academic and nonacademic institutions. If analyzed on a smaller level, differences may arise in outcomes that are not seen when examining all the data from all of the hospitals pooled together. Finally, some social media commentary (see below) pointed out that comparing RVUs and patients/hour between physicians and APPs may not get to the root of where APPs can be helpful—lower patients/hour may result because they are “offloading” minor procedures that can be time-consuming, allowing physicians to continue evaluating more higher-acuity patients. This effect may be difficult to quantify. Over 13 million visits were included. Physicians independently evaluated 74.6% of patients, while physician assistants (PAs) independently evaluated 18.6% and nurse practitioners (NPs) evaluated 5.4%. A mean (±SD) of 3.7 (±1.5) RVUs/visit were generated by physicians, with 2.8 (±1.1) by NPs and 2.7 (±1.1) by PAs. PAs and NPs both evaluated 1.1 patients/hour, while emergency physicians (EPs) evaluated 2.2, generating 8.5 RVUs/hour compared to 3.1 by NPs and 3.0 by PAs. Increasing APP coverage by 10% reduced overall patients/hour by 0.12 and RVUs/hour by 0.4. There was no significant effect seen on length of stay, left without completing treatment, or 72-hour returns. The results show that in this hospital system, physicians generated more RVUs/patient and saw a higher number of patients/hour than APPs. APPs can be used to offload lower-acuity patients and those needing minor procedures. Comparing productivity numbers alone may not fully reveal the benefits of utilizing APPs in the ED setting. David Hiram: A lot of interesting data here, but I don't think the take away is “APPs don't contribute to RVU production” and certainly not “APPs see less patients per hour.” They say APPS only see 1.1 per hour, but that is only patients not seen by both APP and physician. This could even be tolerated, but they had several policies that forced APPs to involve the physician. The physician gets full half credit for that RVU, but did they do half the work? It is not uncommon for them to briefly review the case and just say hi to a patient. Did that patient really need physician involvement? Do onerous requirements of physician involvement (they stated abnormal vital signs forced physician involvement, which is incredibly common) deter APPs from taking more complex patients since it takes more time to present a patient to an attending, slowing ED flow? If I were paid by RVU or had a patient per hour expectation, and I also was instructed to present complex patients to another clinician, I would be incentivized to take less complex patients that I did not have to present. At many institutions, it is the expectation that APPs see the lower acuity and the higher acuity goes to residents for learning purposes (totally appropriate). Does this contribute to lower complexity seen? Was there any local expectation that APPs see the less complex? The less complex does not mean less time either. Lacerations and reductions can take far more time than a STEMI. Many places also require that every admitted patient is seen by a physician as well, taking away from the “seen independently” patient per hour of APPs. What best demonstrates RVU production was "hidden," for lack of better phrase, is the critical procedures where physicians did 0.7 percent alone and it was 5.3 percent for APP and Physician together. The majority was done by APP and physician together, but we all know only one person can do a procedure in reality. Most likely the APP did the procedure and was likely remotely supervised by the attending. Yet, none of these RVUs were attributed to the APP from what I can see since it wasn't "independently." This all goes to how PAs are invisible when it comes to billing. When CMS [Centers for Medicare & Medicaid Services] is billed for joint work by the PA/Physician team, it goes under the physician NPI [National Provider ID] and the PA becomes invisible. This is why there is a paucity of data on our RVU production. NPs can bill CMS directly with their NPI number and I am unsure how it plays into this. Last point, but certainly not the last flaw, the APPs had several other duties such as working triage and reviewing patient results, which do not count towards RVUs. To clarify, I appreciate their conclusions that APPs can be effectively integrated without negative effect on patient safety. I particularly appreciate their point of "[APPs] independently evaluating critically ill ED patients suggests the potential for enhanced use of [APPs] in EDs." I just feel the attention grabbing point here is that APPs see 1.1 per hour and generate half the revenue, which feels misleading. Thank you David for your comment. The conclusion was not that APPs do not contribute to RVU production, but that their observed patient per hour was approximately half that of physicians. I was also surprised by this finding, but based on our data we are certain that it represents the reality across this very large sample of 13 million ED visits. However, based upon how APPs are used within the study EDs, you are correct that this likely underestimates the contribution of APPs because we did not attribute RVUs to APPs who performed supportive ED-level functions, such as provider-in-triage (PIT). APPs can be utilized to “offload” lower-acuity cases, while allowing physicians to care for higher-acuity patients. Physicians overall had higher levels of productivity, as measured both by patients/hour and by RVUs/hour. The authors have no relevant financial information or potential conflicts to disclose.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.