Objective To evaluate productivity of mid-level providers (MLPs) compared with emergency medicine (EM) resident physicians in an emergency department (ED) low acuity area, and to compare patient satisfaction when cared for by MLPs versus EM residents.
Methods This was a retrospective review of EM resident physicians and MLPs in an ED low acuity area. The number of patients seen and relative value units (RVUs) generated per clinical hour worked were evaluated. A t test was used to compare resident and MLP productivity. Additionally, patients were prospectively surveyed to assess satisfaction, using survey items based on the Press-Ganey survey. Non-parametric statistics were used to analyse patient satisfaction scores.
Results MLPs treated 2.21 patients per hour (CI ±0.09), while resident physicians treated 1.53 patients per hour (CI ±0.08). MLPs generated 4.01 RVUs per hour (CI ±0.18) while resident physicians generated 3.14 RVUs per hour (CI ±0.18). Resident physicians generated 2.07 RVUs per patient (CI ±0.08) while MLPs generated 1.82 RVUs per patient (CI ±0.03; p<0.001). Of the 201 completed satisfaction surveys, 126 patients were seen by MLPs and 75 were seen by residents. Overall patients were highly satisfied with their ED visit. There were no differences in any survey responses based on provider type or resident level of training.
Conclusion In a low acuity area of the ED, MLPs treated more patients per hour and generated more RVUs per hour than EM resident physicians. However, resident physicians generated more RVUs per patient. Patient satisfaction did not differ.
Statistics from Altmetric.com
Emergency department (ED) patient visits have been steadily increasing, and despite significant growth in the number of board certified emergency physicians, physician workforce shortages continue. Over the past 11 years, the National Hospital Ambulatory Medical Care Survey estimated that ED visits have grown from 94.9 million in 1997 to 123.8 million in 2008.1 ,2 Many EDs utilise physician extenders, or mid-level providers (MLPs), to help augment the emergency physician workforce. The proportion of EDs reporting use of MLPs has increased from 28.3% in 1997 to 77.2% in 2006, and is likely even higher in academic EDs.3 ,4 The number of ED patients seen by MLPs has increased dramatically, from 5.5% in 1997 to 12.7% in 2006.3
Utilising MLPs has allowed EDs to better manage increasing patient volumes, and helps to offset the need for more emergency physicians.4 ,5 Nationally, 53% of patients are seen in the lowest triage acuity groups and this is the typical case mix seen by MLPs. There is no contemporaneous physician involvement documented in 45% of these cases.3 There is limited conflicting information on the quality of care provided by MLPs in the ED setting.4–7
Currently, there are about 200 EDs hosting Accreditation Council for Graduate Medical Education (ACGME) or American Osteopathic Association (AOA) emergency medicine (EM) residency programmes in the USA. Although utilisation of EM resident physicians will typically allow an ED to better manage increasing patient volumes, patient service responsibilities are often affected by other important priorities, such as education and research. Many academic EDs also utilise MLPs, particularly in a fast track setting.
In general, patients who have been seen by MLPs tend to be very satisfied with their experience. However, ED patients expect to see a physician when they present for care, and up to 20% would prefer to wait longer to see a physician rather than see an MLP.8 In the current culture that focuses highly on customer satisfaction and often ties reimbursement to this measure, it is important to consider the level of patient satisfaction regarding MLP care, as well as resident care.
In addition to patient satisfaction, there is growing popularity linking reimbursement and ED compensation to the relative value unit (RVU). RVUs are derived from a complex formula that provides a metric tool to gauge provider productivity and reimbursement by Medicare. Simply put, an RVU is generated for each CPT code by taking into account the work provided by the clinician, the practice expense of the services provided and the malpractice cost for the service with a geographic adjustment factor. The total RVU is then multiplied by a conversion factor (what Medicare pays for 1 RVU generated) to determine the Medicare payment.9 RVUs are not only used for reimbursement but also for compensation. ED productivity is determined by RVUs generated per hour, which is commonly used to guide ED provider compensation. Therefore, in a world focused on maximising productivity, it is important to determine where MLPs would be most productive when considering staffing in an EM training programme.
In this study, we compare the productivity of MLPs compared with EM residents in a fast track setting. Additionally, we compare the level of patient satisfaction when being cared for by MLPs versus EM residents.
This was a retrospective cohort study of resident and MLP productivity and a prospective cohort study of patient satisfaction in an ED low acuity fast track area. The study was reviewed and approved by the hospital's institutional review board.
Study setting and population
The study was performed in the low acuity fast track area at a tertiary care community ED with an annual census of 70 000; 19% of the annual ED census is seen in the fast track area. The study population included MLPs (five physician assistants and one nurse practitioner), EM residents (allopathic and osteopathic residents of all levels) assigned to the fast track and patients seen in the fast track (triage ESI 4 and 5). EM residents assigned to the remainder of the ED were excluded. Off service resident rotators and attending physicians were also excluded.
For the retrospective portion of this study, research associates collected census and productivity data through query of the Verinet coding system. The Verinet system records individual provider shift data regarding total number of patients seen, total number of RVUs generated and mean RVUs generated per patient. In the event of patients being signed over from shift to shift, the transfer of care to the next provider is recorded on the electronic medical record but the system credits the original provider with care of the patient. Shift hours were recorded from the resident and MLP work schedules. Resident shifts were almost entirely 10 h shifts from 10:00 to 20:00, although on conference days they occasionally worked 13:00 to 21:00 or 16:00 to midnight. MLP shifts were predominantly 8:00–18:00 or 13:00–23:00. Patients seen per hour (pt/h) and RVUs generated per hour (RVU/h) were calculated.
For the prospective portion of this study, a convenience sample of patients seen in the fast track was surveyed. After their visit, patients verbally consented to complete an anonymous four item satisfaction survey. The items were as follows: “The healthcare provider cared about me as a person,” “The healthcare provider explained my problem and follow-up to me,” “The healthcare provider kept me aware of tests and studies” and “I am very satisfied with my experience here today.” Survey items were based on Press-Ganey surveys to best represent parameters by which healthcare providers are evaluated. Each question was rated on a 10 point scale, with a score of 10 corresponding to ‘strongly agree’ and a score of 1 corresponding to ‘strongly disagree’.
A power calculation determined that 10 subjects per group were required for a power of 0.8 to determine a 25% difference in productivity among residents of varying levels of training and MLPs (α=0.05). Data for productivity measures, including pt/h, RVU/h and RVU/pt were analysed using descriptive statistics, two tailed t test and analysis of variance. Comparison was made between MLPs and all residents as well as with subgroups of residents with different levels of training. Regression analysis was used to determine whether ED census contributed to differences in productivity. Data for satisfaction were analysed with descriptive statistics. Non-parametric tests (Mann–Whitney and Kruskal–Wallis) were used to compare patient satisfaction scores for residents of various levels of training and MLPs, as well as satisfaction scores based on shift time of day, as these data were not normally distributed.
Ninety resident (15 PGY1, 33 PGY2 and 42 PGY3) and 208 MLP fast track shifts were included for productivity analysis. During the study period (June to October 2009), MLPs saw 2.21 pt/h (95% CI ±0.09) while residents saw 1.53 pt/h (95% CI ±0.08; p<0.001). There were no differences in productivity based on resident level of training. PGY1s saw 1.5 pt/h (95% CI ±0.17) versus PGY2s, who saw 1.49 pt/h (95% CI ±0.12), versus PGY3s, who saw 1.58 pt/h (95% CI ±0.14).
MLPs generated 4.01 RVU/h (95% CI ±0.18) while residents generated 3.14 RVU/h (95% CI ±0.18; p<0.001). There were no differences in RVU/h based on resident level of training.
Residents, however, generated 2.07 RVU/pt (95% CI ±0.08) while MLPs generated 1.82 RVU/pt (95% CI ±0.03; p<0.001). There were no differences in RVU/pt based on resident level of training.
The mean ED census during the study period was 195 visits/day (SD=20). Regression analysis to determine the relationship of census to residents' productivity yielded r2 values of 0.024 (pt/h) and 0.059 (RVU/h); r2 values for MLPs were 0.275 (pt/h) and 0.272 (RVU/h).
A total of 201 patients completed the satisfaction surveys; 126 patients were seen by MLPs and 75 were seen by residents (22 by PGY1s, 17 by PGY2s and 36 by PGY3s). The majority of patients were highly satisfied with their ED visits. For the first survey item (“The healthcare provider cared about me as a person”), median/mean values were 10/7.0 for MLPs and 10/7.4 for residents (p=0.4, Mann–Whitney). For the second item (“The healthcare provider explained my problem and follow-up to me”), median/mean values were 10/7.1 for MLPs and 10/7.4 for residents (p=0.70, Mann–Whitney). For the third item (“The healthcare provider kept me aware of tests and studies”), median/mean values were 10/7 for MLPs and 10/7.3 for residents (p=0.42, Mann–Whitney). For the fourth item (“I am very satisfied with my experience here today”), median/mean values were 10/7.1 for MLPs and 10/7.1 for residents (p=0.95, Mann–Whitney). There were no differences in any survey responses based on resident level of training (p=0.52 for item 1, p=0.54 for item 2, p=0.60 for item 3 and p=0.61 for item 4, Kruskal–Wallis). There were also no differences found in patient satisfaction based on shift time of day.
In an ED fast track setting for one community teaching hospital, MLPs treated more patients per hour and generated more RVUs per hour than EM resident physicians. These productivity differences should be considered when determining staffing needs in an ED fast track setting with an EM residency training programme.
There was no significant difference in productivity among EM residents by postgraduate year. This finding is discordant with multiple other studies which have demonstrated increased productivity over the course of training.10–15 These studies, however, were done in a mixed or high acuity setting. It is possible that level of training is correlated with number of patients seen in higher acuity settings because of the increasing complexity of patients. It seems intuitive that residents of a higher level of training would be more proficient at managing higher acuity patients then their intern counterparts who may need more time and supervision with such complex cases.
The ED patient census had little impact on the productivity of either group. Concordant with the current study, in their study from an academic urban emergency department, Jeanmonod et al found that ED volume had a weak correlation with resident productivity while working in a high acuity environment.10 This may be because providers are ‘maxed out’ in terms of the number of patients they are seeing and therefore cannot increase their productivity in response to volume. It may also be because of unmeasured time limiting factors, such as bed turnover and laboratory turnaround. Resource limitations, such as nursing staffing or trauma alerts during shifts, also likely play a role.
At the study ED, fully licensed physician extenders working in a low acuity area evaluate the great majority of patients without attending physician involvement. On the other hand, resident physicians with restricted or unrestricted licenses must present to an attending physician for all of the patients they evaluate due to hospital restrictions. We expect that the time spent locating an available attending physician, presenting and discussing the case, as well as teaching time between the resident and attending physician may be significant. This time may be responsible for much of the difference in productivity between physician extenders and resident physicians. In their study of direct observation of resident/attending interactions, Chisholm et al demonstrated that approximately 18% of resident time spent working in non-critical areas of the ED was spent interacting with the attending staff.16
Resident physicians generated more RVUs per patient. We expect this is due to improved documentation practices, perhaps affected by the fact that the MLPs are accustomed to working mainly in the fast track area, while the resident physicians are accustomed to mainly treating higher acuity patients in the ED throughout their residency. This may result in more thorough documentation by residents who are taught necessary billing and coding practices for high acuity patients compared with MLPs, who are urged to emphasise turnaround times. A review of the medical literature revealed no other studies comparing revenue generation of MLPs versus residents working in an ED. Residents in our ED receive periodic didactic teaching covering appropriate documentation and billing, which might not be available to the MLPs. Previous studies have demonstrated that educational initiatives covering documentation and billing can increase resident RVU production.17 ,18 Although MLPs are reimbursed by Medicare at 85% of the physician fee schedule, the RVU data included in this study are at the billed rate as opposed to reimbursed rate and so is not responsible for the decreased RVU generation per patient seen by the MLPs.
In the current study, there were no differences in patient satisfaction on four recorded measures. Previous studies have demonstrated that patient's overall satisfaction with their ED visits decreases with decreasing level of acuity, making the low acuity environment a challenging place to practice.19–22 A British study comparing patient satisfaction with their encounters in an ED setting found no difference in patient satisfaction when seen by a general practitioner, a specialty registrar (the equivalent of an EM attending) or an emergency nurse practitioner, but did find that patients were slightly less satisfied with the adequacy of information provided by senior house officers as well as being slightly less satisfied with their overall visit when seen by a senior house officer.23 A second study from the UK found that patients randomised to be seen by nurse practitioners versus a house officer had increased overall visit satisfaction that was reflected in higher satisfaction scores in multiple measures of communication.24 Although not studied in the ED setting, Hooker et al found that patient satisfaction in the Kaiser Permanente system did not differ based on care by physician, physician assistant or nurse practitioner.25 In their study of the use of a nurse practitioner in the ED under direct supervision of an attending physician, Rhee, et al found there was no difference in the satisfaction of patients seen by the MLP when compared with controls of medical students and house staff.26 Finally, in their systematic review, Carter et al reviewed the impact of nurse practitioners in EDs and found that overall, patients were either as satisfied or more satisfied with the care provided by nurse practitioners.27
Our study was limited to one academic ED staffed by six different MLPs and 40 individual EM residents. The productivity of this limited number of MLPs and residents may not be reflective of the average productivity of physician extenders or residents at other EDs.
This study was non-randomised, potentially resulting in some bias in patient selection by providers. When evaluating patients seen per hour, some variability between MLPs and residents may be due to provider's selection of certain types of patients because of interest or lack of comfort with a particular presentation. Part of this variability is mediated by department policy that the provider picks up the patient who has been waiting the longest for care. There may have also been some variability in patient seen per hour between residents and MLPs because MLPs staff more fast track cases than residents, potentially making them more efficient at seeing these types of cases.
Some of the increase in RVU per patient experienced by residents may be due to the fact that attending physicians are required to see and document their involvement in patient care when seeing patients with residents but not necessarily for MLPs. This produces both an increased reimbursement for every EM code as well as giving the attending the opportunity to review the chart for maximal reimbursement.
Lastly, when conducting the patient satisfaction surveys, the total number of patients asked to complete the survey was not recorded, only the total number of patients who actually completed the survey was noted. Therefore, a response rate was not determined. Perhaps if a response rate was calculated it would have further legitimised the results of our survey population.
In an ED fast track setting at a community teaching hospital, physician extenders treated more patients per hour and generated more RVUs per hour than EM resident physicians. Resident physicians, however, generated more RVUs per patient, likely due to improved documentation practices. The ED patient census had little impact on productivity of either group. Additionally, patient satisfaction in a sample of patients seen in a low acuity area of the ED did not differ as a factor of provider type, provider level of training or the shift time of day.
Competing interests None.
Ethics approval The study was approved by St Luke's institutional review board.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.