AUCTORES
Globalize your Research
Research Article | DOI: https://doi.org/10.31579/2642-973X/028
1Department of Mathematics and Statistics, University of Arkansas in Little Rock, Little Rock, AR 72205
2Department of Neurology, University of Arkansas for Medical Sciences, Little Rock, AR 72205
3Department of Interventional Radiology, University of Arkansas for Medical Sciences, Little Rock, AR
4The Kidney Clinic, Snellville, GA 30078
*Corresponding Author: Aliza Brown, PhD, FAHA, Associate Professor, Department of Neurology, University of Arkansas for Medical Sciences, Little Rock, AR, 72205.
Citation: Xixi Wang, Wei Zhang, Sen Sheng, Rohan Sharma, Mudassar Kamran, et al. (2022). Residency Attrition and Associated Characteristics, a 10-Year Cross Specialty Comparative Study. Brain and Neurological Disorders. 5(4); DOI: 10.31579/2642-973X/028
Copyright: © 2022 Aliza Brown, this is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Received: 19 April 2022 | Accepted: 10 June 2022 | Published: 02 September 2022
Keywords: residency attrition; graduate medical education (gme); medicare funding
Background/Aim: With the annual cost of training a single resident estimated at $141,240, the implication of resident attrition imposed on the public could far exceed that dollar amount. However, not all specialties face the same challenge. A study of the trend and dispersion of attrition rates across different specialties would detect physician shortages and misallocations from the outset.
Materials and Methods: Data of 20 major specialties from academic year 2010-2011 to 2019-2020 was collected from the ACGME data resource book. Annual attrition rate was calculated and its spread was visualized via box-plot. Median and inter-quartile range (IQR) of annual attrition rate were calculated to draw comparison among specialties. Attrition rates’ association with time was analyzed to identify temporal trends. The Kruskal-Wallis test was performed to identify any significant difference among attrition rates of 20 major specialties. Pairwise comparison was followed to differentiate high- and low-attrition specialties.
Results: Dermatology consistently had the lowest attrition rate (Medium, 0.46%; IQR, 0.32% - 0.70%) while Psychiatry had the highest (Medium, 7.53%; IQR, 6.74% - 8.60%). Urology had the fastest decline in attrition rate (r= -0.93691; p<0.0001), followed by Internal medicine (r=-0.92173; p=0.0001). Primary care specialties including family medicine, obstetrics and gynecology and pediatrics have had more difficulty retaining their residents. A lower percentage of US medical school graduates (p<0.0001) and a higher percentage of female residents (p<0.0001) were found in high-attrition specialties.
Conclusion: Attrition rate remained vastly different among specialties over the past decade, necessitating inter- specialty dialogue to effectively tackle this issue. Left simply to workforce supply and demand, physician shortage and maldistribution could further expose the more vulnerable of our society to disastrous consequences.
The novel coronavirus pandemic of 2020 forced us to confront many hardships, many of which embedded in the inequality of our society, disproportionately affecting the ethical minority, the elderly and the poor [1-3]. Among many other things, this pandemic also highlighted the significance of physician shortages in the United States, with current projections anticipating a national shortage of up to 122,000physicians by 2032 [4,5]. However, not everyspecialty in medicineis facing the same shortageand not everyshortage has the same dire consequences. Studieshave shown thatFamily medicine physicians play a vital role in caring for vulnerable populations [6] and yet the number of primary care physicians has grown at a fraction of the rate of specialized physicians [7]. Also emerging during the pandemic is the mental health crisis brought on by lockdown and isolation [8]. Demand for psychiatrists will increase in already strainedemergency and mental health systems [9].
However, while the United States may face a future shortage of physicians, it does not presently have a shortageof doctors [10]. Studiesfound the graduatemedical education had become the primary bottleneck in the physician pipe line with limited residency training positioned constrained by funding availability4. Yet, some specialties report as high as 17-26% attrition rate [11-13]. Premature departureof a resident from training programis disruptive, and with annual cost of training a single resident estimated at $141,240it has financial implications for the traininginstitution [7,11] and poseseven greater opportunity cost in aggregate [10].
Multiplestudies have examinedresident attrition [11,14], with a vast majority of them investigating it for individual specialties [15-19]. A detailed analysis of resident attrition across various specialties is lacking. Given different data sources and varied methodologies, consistent evaluation of attrition rate even within a particular specialty may be challenging [18]. Moreover, there has been insufficient analysis of change in attrition rates over time. A thorough understanding of the attrition rate for variousspecialties driven from a comprehensive database, its comparison across other specialties, and its evolution over a periodmay help institutions better address this issue with educational and financial implications.
With that in mind, we began by examining the attrition rates for 20 major specialties and their evolution over a period of 10-years, employing a uniform Accreditation Council for Graduate Medical Education (ACGME) database.
Data collection
All data utilized in this study was collected from the ACGME data resource book (available at acgme.org). Data on attrition, and programmatic characteristics by specialty were gathered from academicyear (AY) 2010-2011 to 2019-2020. Twenty major specialties, defined as those with the greatest number of active residents by the end of AY 2019-2020, were analyzed. Cost and benefit analysis of GME programs were extracted from prior studies.
Annual attrition rate
Annual attrition rate was calculated as dividing ‘the number of residents leavingprior to completion of their trainingduring an academicyear’ by ‘the number of active residents at the end ofthe same academicyear’. It has been calculated for each of the 20 specialties yearlyfor the past 10 years.
A box-plot was devised to visually reflect the location and spread of these attrition rates by specialty. Since it’s easily observed there is significant variation in variances among attrition rates of different specialties, the Kruskal-Wallis test was carried out to compare their differences followed by post-hoc pairwise comparison using Dwass, Steel, and Critchlow-Fligner Method. Medianand inter-quartile range (IQR)of attrition rates were calculated for eachspecialty.
A correlation analysis was then performed across specialties to identify any changes in attrition rates over time.
Characteristics of high-attrition and low-attrition specialties
Setting the specialtywith the lowest sum of Wilcoxon scoresas control and based on the result of pairwise comparison by Dwass, Steel, Critchlow-Fligner Method, specialties were separated into high- and low-attrition groups. The Wilcoxon Two-Sample Test was performed to exam the difference in characteristics includingmean number of residents per program, percentage of female residents and percentage of US medical school graduates between specialties in high- and low- attrition groups.
The data analysisfor this paper was generated using SAS software.
Annual attrition rate variation by specialty
The level as well as the dispersion of annual attritionrate vary significantly by specialty over the past 10 years (Figure). Psychiatry has consistently had the highestattrition rate (Medium,7.53%), and the widest variation(IQR, 6.74% - 8.60%). Last year, 417 out of its total of 6,618 (6.30%) residentsleft their programprior to graduation. Following it, Surgery(Medium, 3.37%; IQR, 3.10% - 3.70%), Pathology-anatomic and clinical (Medium,2.91%; IQR, 2.55% - 3.29%), Neurological surgery (Medium, 2.43%; IQR, 2.06% - 2.97%),Family medicine (Medium,2.21%; IQR, 1.94% - 2.69%) and Obstetrics and gynecology (Medium,1.94%; IQR, 1.39%- 2.49%) also had relatively high attrition rate. On the other end of the spectrum, Dermatology for most of the time had the lowest attrition rate (Medium, 0.46%) and one of the smallest variations (IQR, 0.32% - 0.70%). In AY 2019-2020, only six out of 1,594 (0.38%) Dermatology residents left prematurely. Ophthalmology (Medium, 0.82%, IQR, 0.56% - 1.14%),Emergency medicine (Medium,0.83%; IQR, 0.72% - 1.00%), Otolaryngology (Medium, 0.84%; IQR, 0.76% - 1.06%) and Urology (Medium, 0.93%; IQR, 0.75% -1.24%) were also among low-attrition rate specialties.
Figure: The annual attrition rate of residents varies significantly by specialty. Over the past 10 years psychiatry has consistently had the highest attrition rate.
Recent trend in attrition rate byspecialty
Attrition rate is declining with time as there is a significantly negative correlation betweenattrition rate and year (r=-0.16809; p=0.0173). As for individual specialty, Urology (r= -0.93691;p<0 r=-0.92173; p=0.0001) r=-0.90114; p=0.0004) r=-0.89831; p=0.0004) and Pediatrics r=-0.89330; p=0.0005) (Table).>
| Specialty | AY 2010-2011 - AY 2019-2020 Attrition Rate | AY 2019-2020 | |||||
| Median | Lower Quartile | Upper Quartile | Correlation with Time** | No. of Residents not Graduating | Number of Programs | Number of Active Residents | |
| Anesthesiology | 1.55% | 1.34% | 1.84% | -0.89831 0.0004 | 76 | 160 | 6698 |
| Dermatology* | 0.46% | 0.32% | 0.70% | 0.14942 0.6803 | 6 | 144 | 1594 |
| Emergency medicine* | 0.83% | 0.72% | 1.00% | -0.69837 0.0247 | 45 | 265 | 8293 |
| Family medicine | 2.21% | 1.94% | 2.69% | -0.82447 0.0033 | 246 | 701 | 13725 |
| Internal medicine | 1.20% | 0.95% | 1.59% | -0.92173 0.0001 | 251 | 569 | 29243 |
| Internal medicine/Pediatrics* | 1.11% | 0.89% | 1.84% | -0.81233 0.0043 | 11 | 79 | 1511 |
| Neurological surgery | 2.43% | 2.06% | 2.97% | -0.59617 0.0689 | 37 | 118 | 1515 |
| Neurology | 1.68% | 1.49% | 2.16% | -0.69858 0.0246 | 29 | 160 | 3062 |
| Obstetrics and gynecology | 1.94% | 1.39% | 2.49% | -0.87110 0.0010 | 73 | 285 | 5677 |
| Ophthalmology* | 0.82% | 0.56% | 1.14% | -0.37851 0.2808 | 8 | 124 | 1512 |
| Orthopaedic surgery* | 0.78% | 0.64% | 1.00% | -0.60780 0.0623 | 26 | 197 | 4342 |
| Otolaryngology | 0.84% | 0.76% | 1.06% | -0.46913 0.1714 | 12 | 124 | 1689 |
| Pathology-anatomic and clinical | 2.91% | 2.55% | 3.29% | -0.84538 0.0021 | 60 | 142 | 2324 |
| Pediatrics | 1.45% | 1.19% | 2.10% | -0.89330 0.0005 | 98 | 211 | 9323 |
| Physical medicine and rehabilitation | 1.19% | 0.78% | 1.43% | -0.55579 0.0953 | 15 | 94 | 1453 |
| Plastic surgery - integrated* | 1.00% | 0.70% | 1.66% | -0.04909 0.8929 | 5 | 82 | 961 |
| Psychiatry | 7.53% | 6.74% | 8.60% | -0.90114 0.0004 | 417 | 269 | 6618 |
| Radiology-diagnostic | 1.83% | 1.66% | 2.05% | 0.78353 0.0073 | 126 | 197 | 4551 |
| Surgery | 3.37% | 3.10% | 3.70% | -0.65429 0.0401 | 273 | 330 | 8809 |
| Urology* | 0.93% | 0.75% | 1.24% | -0.93691 <.0001 | 10 | 145 | 1734 |
* Represent specialties in low-attrition group. **In the 5th column of correlation with time, correlation coefficient is presented on top, corresponding p-value at bottom. Specialties are listed in alphabetic order. | |||||||
Table: Comparison of Attrition Rate by Specialty.
Division between high- and low-attrition specialties
In a more concrete statistics analysis, the Kruskal-Wallis test shows that there is significant difference among the annual attrition rate of various specialties (p<0 p=0.0783), p=0.2605), p=0.4120) p=0.7571)>
Characteristics comparisons betweenhigh- and low-attrition specialties
When comparing characteristics between high- and low-attrition groups, primary care specialtiesincluding Family medicine (Medium 2.21%; IQR 1.94% - 2.69%), Obstetrics and gynecology (Medium 1.94%; IQR 1.39% - 2.49%) and Pediatrics (Medium 1.45%; IQR 1.19% - 2.10%) fall within the high-attrition rate group. No significance difference in mean number of residents per program has been identified. A higher percentage of female residents (p<0>
This study identified a significant variationin annual attritionrates over the past 10 years, rangingfrom 0.26% to 10.05%, among 20 major specialties. The huge variation in attrition rate amplifies the specialty maldistributions in the U.S. directly at the level of graduate medical education [4]. Reform has been calledupon to addressthe discrepancies betweenthe type of health care available and those in demand by patients and health care facilities [19]. However, the current incentive structure with which Medicare supports residency training makes inpatient care more lucrative than focusing on community healthand outpatient care [12]. One study that investigates the costs and benefits of operating graduate medical education (GME) programs found internal medicine and family medicine faculty practice plans are estimated to operate at a loss, whereas the other specialties are estimated to operate at a profit, with the highest profit per resident estimated for Urologyand the lowest profit estimated for cardiology and general surgery [7].
As our study indicated, primarycare specialties like Family medicine(Medium, 2.21%; IQR, 1.94%2.69%), Internal medicine(Medium, 1.20%; IQR, 0.95% - 1.59%), Obstetricsand gynecology (Medium, 1.94%; IQR, 1.39% - 2.49%) and Pediatrics (Medium, 1.45%; IQR, 1.19% - 2.10%) have all fallenunder high-attrition group.Also alarming is that Psychiatry (Medium, 7.53%; IQR, 6.74% - 8.60%), facing a national shortage in the millennial generation [19], constantly experienced several times the attritionrate of any other specialty. This leads to speculation that specialties with a lower or negative profit margin per resident under the current Medicare funding structure might be less incentivized or effective at retaining their residents. Conversely, specialties like Urology (Medium, 0.93%; IQR, 0.75% - 1.24%), Orthopedic surgery (Medium, 0.78%; IQR, 0.64% - 1.00%) and Plastic surgery – integrated (Medium, 1.00%; IQR, 0.70% - 1.66%), known to be procedure-heavy and lucrative, fall straight into the low-attrition group. This attrition rate disparity between outpatient-focused specialties and procedure-heavy specialties should raise concerns for public healthofficials when addressing the publics access to healthcare, especially in the pandemicera when inequality has been exacerbated [22,23].
Another important result in our study pointsto the fact that the percentage of International medicalschool graduates (IMGs) is higher in the high-attrition group of specialties than that in the low- attritiongroup. This may be explained by their high concentration in primary care specialties while facing more obstacles to complete training. For example, a prior study found that family practice is becoming increasingly reliant on IMGs as they accounted for an increasing percentage of familypractice residency positions filled despite a drop in total positionsfilled [24]. A studyhas also shown IMG serves an important role in fostering diversity, equity and inclusion in its local communities throughtheir language and culture connection to minority populations [25]. However, IMGs also self- reported considerable bias and prejudice, ranging from difficulty with getting externships and interviews to the critical view of their USMG counterparts.
Also, it is notable from our studythat female residents present a higherpercentage in high-attrition specialties, coinciding with prior studiesthat identified women being more susceptible to generational priorities and family issues [26, 27]. A recent study also shows that female PCPs generated 10.9% less revenue from office visits than their male counterparts and yet spent more time in directpatient care per visit, per day, and per year [28]. It’s been argued that formal maternity policies, a shift in culture and ongoing discussion, are needed to retain femaleresidents [29], which is especially urgentas the increasing need forprimary care physicians accelerates during the pandemic.
This study is limited by insufficient transparent data on the cost and benefit of residency training by specialty, partiallydue to the complicated incentiveformula for fundingfrom Medicare and the difficulty with monetizing the full benefit of having a resident. Having more quantitative data on the financial engagement of residency training would enable us to derive more concrete results and analysis.
By differentiating factorsinfluential to the resident attritionrate of various specialties, policymakers would be better informed in drawing up policies that are accountable to medical educators and receiversof medical care. The higherpercentage of IMGs and female doctorsin the high-attrition group of specialties, which includes most of the primary care specialties, signals an opportunity for health care officials to tackle maldistribution in physicians by implementing targeted measures that address specific challenges faced by these two groups of residents. If left unattended to, resident attrition could exacerbate difficulty with accessing healthcare for the vulnerable populations and worsening inequalities highlighted by the coronavirus pandemic.
Acknowledgement: None
Clearly Auctoresonline and particularly Psychology and Mental Health Care Journal is dedicated to improving health care services for individuals and populations. The editorial boards' ability to efficiently recognize and share the global importance of health literacy with a variety of stakeholders. Auctoresonline publishing platform can be used to facilitate of optimal client-based services and should be added to health care professionals' repertoire of evidence-based health care resources.
