Scope of Practice and Autonomy of Physician Assistants in Rural Versus Urban Emergency Departments

This was a study of the scope of practice and autonomy of emergency medicine (EM) physician assistants (PAs) practicing in rural versus urban emergency departments (EDs). Using the American Academy of Physician Assistants (AAPA) Masterfile, a random sample of 200 U.S. EM PAs were surveyed, with oversampling of an additional 200 rural PAs. Location was classified by zip code–based rural–urban commuting area codes, and responses were compared about conditions managed, procedures performed, and physician supervision between rural versus urban groups. A total of 237 responses were received from PAs in 44 U.S. states, of which (201) were valid responses (105 rural, 96 urban) from PAs currently practicing in EDs (59.3% exclusion-adjusted response rate). Compared to urban PAs, rural PAs more frequently managed cardiac arrest (67% vs. 44%), stroke (86% vs. 72%), multisystem trauma (83% vs. 70%), active labor (44% vs. 23%), and critically ill children (82% vs. 65%) in the past year. They were more likely to have performed intubation (65% vs. 44%), needle thoracostomy (21% vs. 8%), and tube thoracostomy (46% vs. 26%). Rural PAs more often reported never having a physician present in the ED (38% vs. 0%) and less often reported always having a physician present (50% vs. 98%). Rural PAs were also less likely to report that a physician evaluates more than 75% of their patients (8% vs. 18%) and more likely that a physician never evaluates all of their patients (19% vs. 7%). Rural PAs reported a broader scope of practice, more autonomy, and less access to physician supervision than urban PAs. Estudio del ámbito de la práctica clínica y la autonomía de los asistentes médicos (AM) en Medicina de Urgencias y Emergencias (MUE) que ejercen en servicios de urgencias (SU) urbanos frente a los rurales. Se usó el fichero de la American Academy of Physician Assistants. Se encuestó a una muestra aleatoria de 200 AM de MUE en EEUU, con sobremuestreo de 200 AM rurales adicionales. La ubicación se clasificó por códigos de área de desplazamiento rural-urbana basados en códigos postales, y las respuestas sobre las enfermedades manejadas, los procedimientos realizados y la supervisión médica se compararon entre los grupos urbano y rural. Se recibieron 237 respuestas de AM de 44 estados de EEUU, de las cuales 201 fueron respuestas válidas (105 rurales, 96 urbanas) de AM que actualmente ejercen en un SU (proporción de respuesta ajustada, por exclusión de un 59,3%). En comparación con los AM urbanos, los AM rurales manejaron con mas frecuencia en el año previo el paro cardíaco (67% vs. 44%), el ictus (86% vs. 72%), el traumatismo multisistémico (83% vs. 70%), los partos naturales (44% vs. 23%) y los niños con patología critica (82% vs. 65%). Los AM rurales tuvieron mayor probabilidad de haber realizado intubación (65% vs. 44%), toracostomía con aguja (21% vs. 8%) y toracostomía con tubo (46% vs. 26%). Los AM rurales contestaron más a menudo no tener nunca un médico presente en el SU (38% vs. 0%), y con menos frecuencia contestaron tenerlo siempre presente (50% vs. 98%). Los AM rurales también fueron menos propensos a contestar que un médico evalúe más del 75% de sus pacientes (8% vs. 18%) y más probable que un médico nunca evaluase la totalidad de sus pacientes (19% vs 7%). Los AM rurales reportaron un ámbito más amplio de la práctica clínica, más autonomía y menos acceso a la supervisión médica que los AM urbanos. There are not enough emergency medicine (EM) board-certified emergency physicians (EPs) available to staff our emergency departments (EDs) in the United States,1-3 particularly in rural areas.4-6 Furthermore, the limited access to medical care in rural settings mandates that providers be comfortable with increased levels of care.7 To supplement these deficits and to improve the efficiency of EPs, physician assistants (PAs) are increasingly utilized in both rural and urban EDs.8-10 However, little is known about differences in the scope of practice and autonomy of EM PAs practicing in rural versus urban settings. A prior national survey of EM PAs had a low (21%) response rate, limited geographic variation, and no comparisons between rural and urban practice locations.11 As the employment of PAs in EDs continues to increase,12, 13 so does their level of autonomy and scope of practice.6, 11, 14 However, these trends are not necessarily accompanied by an increase in physician supervision. Illustrating the possible effects, recent data suggest that PAs who worked without direct physician supervision provided a lower quality of acute asthma care than did directly supervised PAs.14 To optimize the quality of emergency care, understanding how the scope of practice and autonomy compares between rural and urban PAs has implications for the training, supervision, and optimal utilization of PAs in a wide range of ED practice settings. We conducted a pilot, hypothesis-generating study to identify large differences in the scope of practice and autonomy of EM PAs practicing in rural versus urban EDs. We hypothesized that rural PAs would have a broader scope of practice and higher reported autonomy, while receiving less direct supervision. This study was a cross-sectional analysis of the 98,148 PAs in the 2011 American Academy of Physician Assistants (AAPA) Masterfile, followed by a survey of EM PAs working in both rural and urban locations. The Colorado Multiple Institutional Review Board approved the study as an exempt study with a waiver of written informed consent. We obtained a comprehensive list of practicing PAs who are both members and nonmembers of the AAPA from the AAPA Masterfile (www.aapa.org). The Masterfile provided data including demographic and contact information and type of primary practice site. We limited the survey to PAs who self-identified EM as their primary specialty (n = 6,080) and classified location (rural or urban) using zip code–based rural–urban commuting area codes (http://depts.washington.edu/uwruca/index.php). We developed our questionnaire based on a prior survey conducted by the Society of Emergency Medicine Physician Assistants with additional modifications in the attempt to identify major differences in scope of practice, autonomy, and physician supervision between rural and urban EM PAs.11 Seeking to determine if further and more granular study is warranted, we selected a subset of relevant items from the prior survey, revised these questions and response options, and added survey items. Based on the consensus of both authors, we created the initial survey draft that was then refined with feedback from two EP-scientists at the authors' institution with expertise in survey methodology, but without direct involvement in the study. Following these revisions, we pilot-tested the survey with four practicing EM PAs (none of whom were eligible for the actual study), followed by interviews with these respondents to determine ease of use, validate the interpretability of the questions and responses, and evaluate the ability of questions to accurately measure items of interest. We made additional minor revisions to improve clarity after this pilot testing to arrive at the final survey. Using a random number generator, we selected a sample of 200 (3%) from all 6,080 EM PAs in the 2011 AAPA Masterfile and an additional sample of 200 from the 581 rural PAs. This method intentionally oversampled rural PAs to ensure sufficient representation of this group. The full survey instrument is provided in Data Supplement S1 (available as supporting information in the online version of this paper). We asked all PAs about demographic characteristics, ED practice characteristics, physician supervision, frequency of procedures performed and conditions managed in the prior 12 months, methods of prior EM education, and preferences for future training. From August 2011 to October 2011, we mailed surveys up to three times to all possible participants. If no response was obtained from mailings, we attempted to reach the PA by phone during November and December 2011 before we determined nonresponse. Among complete respondents, missing values for individual items were excluded. Respondents who were not actively practicing in U.S. EDs or who had invalid addresses were excluded from the study. We selected 400 as our initial total sample size to obtain an estimated 60% response rate with a projected 1:1 ratio of rural to urban responses. We selected this target sample size to reliably detect two-sided differences of 15% with 95% confidence intervals (CIs) that were nonoverlapping with zero over a range of prevalence to evaluate our hypothesis that rural PAs have a higher scope of practice and greater autonomy while receiving less supervision. We believed that this level of precision was adequate to generate further hypotheses and inform future interventional studies. We performed analyses using Stata version 10.1 (StataCorp, College Station, TX). Data are presented with descriptive statistics. For between-group comparisons, we combined all greater than zero responses regarding procedures performed and conditions managed, determined a priori, to create a dichotomous comparison (performed vs. not performed). We compared responses from rural and urban EM PAs using differences in proportions with 95% CIs. Of the initial 400 PAs surveyed, we received 237 responses, of which 36 (15.2%) were excluded for the following reasons: 0% of clinical time spent working in the ED (n = 24), PA retired from practice (n = 3), reported not to be a PA (n = 1), active military duty (n = 1), and moved/undeliverable address (n = 7). The remaining 201 responses, representing PAs from 44 states, were included in the analysis (105 rural, 96 urban). Of the 163 nonresponders, we estimated that 25 (15.2%) would be excluded based on not meeting selection criteria. Thus we calculated our adjusted response rate to be 59.3% (201 valid responses of estimated 339 eligible), using the American Association for Public Opinion Research guidelines.15 Compared to the 201 valid respondents, the 163 nonresponders and 36 exclusions were similar in distribution of rural or urban classification, as well as, age, sex, EM board certification rate, and years postgraduation (data not shown). The comparison of demographic characteristics, scope of practice, and supervision of PAs in EDs, stratified by rural or urban classification, are presented in Table 1. Rural PAs were more likely to be older and male and have more years postgraduation. The clinical practice for rural PAs involved lower-volume EDs, fewer patients per hour, and higher proportion of clinical time outside the ED. Rural PAs more often reported never having a physician present in the ED and less often reported always having a physician present in the ED, compared to urban PAs. In addition, rural PAs were less likely to have a physician evaluate their patients or to have at least one supervising physician who was EM board-certified. The frequency of conditions managed and procedures performed by PAs in rural versus urban EDs are shown in Table 2. In the past 12 months, rural PAs more frequently managed cardiac arrest (67% vs. 44%), stroke (86% vs. 72%), multisystem trauma (83% vs. 70%), active labor (44% vs. 23%), and critically ill child (82% vs. 65%). While rural PAs were less likely to have performed bedside ultrasound (43% vs. 62%) and lumbar puncture (48% vs. 62%), they were more likely to have performed bag-valve-mask ventilation (79% vs. 52%), intubation (65% vs. 44%), needle thoracostomy (21% vs. 8%), and tube thoracostomy (46% vs. 26%) in the preceding 12 months. Laceration repair, central line placement, and procedural sedation were common procedures without large differences in rural versus urban EM PA practice. Table 3 details prior EM-specific training and desired training needs among PAs practicing in rural versus urban EDs. While prior EM-specific training was similar between the two groups, rural PAs were less likely to receive on-the-job training from physicians. The most common forms of prior EM-specific training included EM-specific rotations during PA training, brief certification programs (e.g., Advanced Cardiac Life Support, Advanced Trauma Life Support, Pediatric Advanced Life Support), and continuing medical education and were similar between rural and urban PAs. Regarding desired methods of future EM-specific training, rural and urban PAs similarly preferred continuing medical education training and procedure and skills lab training over increased physician supervision and guidance, while rural PAs reported higher interest in distance learning. As a pilot, hypothesis-generating study, our results indicate that EM PAs manage high-complexity conditions and perform complex procedures. PAs practicing in rural EDs appear to have an even greater scope of practice than urban PAs. We also found that rural PAs were more likely to work in EDs without onsite physician supervision, have less access to EM board-certified or any physician supervision if needed, and have fewer of their patients evaluated by physicians. The current trend of increasing PA use may be a consequence of the significant EP shortages nationwide and maldistribution of practitioners between rural and urban EDs.1-3, 12, 13 Lower patient volumes in rural areas make staffing EDs with multiple providers financially difficult, and increasing PA staffing may improve department cost-efficiency.7 As our data suggest, these issues have likely influenced the reported higher scopes of practice and less physician supervision among rural EM PAs. Prior studies have also identified that non–EM board-certified physicians provide a substantial proportion of rural EM care and that significant discrepancies regarding physician board certification and residency training exist between urban and rural ED physicians.1, 5, 16 Building on these data, our results indicate that rural EM PAs may also be less likely to receive supervision from EM board-certified physicians. Prior studies also suggest that non–EM board-certified physicians themselves lack confidence is providing certain types of emergency care.17 Further study is needed to quantify how variations in type and amount of PA supervision affect the quality and safety of patient care. Our results also imply that the complexity of conditions and procedures being performed by PAs in rural EDs were greater than those by urban PAs and similar to those of EM board-certified physicians; however, further studies are needed to confirm our findings. Compared to EM physician residency training, EM PA education provides limited exposure to EM-specific knowledge and skills training. Our findings indicate that few completed EM PA residency/certification programs (as few exist). Differences in complexity of conditions and procedures being performed between rural and urban PAs suggest two different scopes of practice; however, current training and education do not reflect these differences. Direct physician supervision that might mitigate these differences may be discordant and lacks standardization. Furthermore, few rural and urban PAs reported a desire for additional physician supervision, indicating a preference for enhanced autonomy. Emergency medicine specialty organizations largely support the use of EP-supervised PAs in the delivery of emergency care.2 Supporting this model, prior studies have demonstrated similar quality of care for patient from PAs with direct physician supervision compared to physicians.14 However, our findings suggest that in many rural EDs, this model may be undermined by limited availability of onsite physicians. Thus, establishing guidelines that standardize EM PA training, reassess competencies, and improve access to physician supervision may raise the quality of emergency care. While this study suggests relative differences between PAs practicing in rural versus urban EDs, exactly how such differences affect quality of care and patient safety remains unclear. Furthermore, as our data only describe EM PAs, our results cannot be generalized to other physician extenders, such as nurse practitioners. While our response rate was higher than that of previously reported data (59% vs. 21%),11 response bias remains possible. However, key characteristics between respondents and nonrespondents were similar. In addition, several key findings had large-magnitude differences, which would be unlikely to change direction with a higher response rate. Although our survey was not formally designed to adjust for amount of ED clinical time, rural PAs appeared to spend less clinical time in the ED (65% vs. 77%). Thus, adjusting for clinical time might, if anything, increase the relative differences observed. We chose the rural–urban commuting area codes to classify primary practice location and defined ED care settings by self-report. Several rural–urban classification systems exist, and using a different system may yield somewhat different results. Moreover, allowing individual PA respondent definitions of rural and urban would have yielded inconsistent classification, compared to our more objective measure. While the primary contact information in the AAPA Masterfile is supposed to yield work addresses, some zip codes might be from home addresses. This presents the potential for misrepresentation of PAs actual commuting status (living in urban location and commuting to a rural ED or vice versa). Finally, recall and measurement bias are possible as all responses were self-reported (including the definition of an ED and reported scope of practice) and were not verified in this pilot study. Thus, the responses are subject to imprecision, and we therefore focused the analysis on relative differences between rural and urban PAs, as potential sources of bias are unlikely to affect the rural and urban groups differently. This pilot study suggests large differences in the ED practice of rural and urban physician assistants; however, further study is needed and planned to validate these findings by chart review or prospective enrollment. Rural physician assistants reported a broader scope of practice, more autonomy, and less access to direct physician supervision than urban physician assistants. It remains unclear how these variables affect the quality and safety of emergency care in U.S. EDs. Defining and implementing training and supervision standards should be considered a high priority to optimize utilization of emergency physician assistants and the quality of emergency care, particularly in rural EDs. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.