Differences in Clinical Activity and Medicare Payments for Female vs Male Ophthalmologists

Key Points

Question  How do the earnings and clinical activity of men and women in ophthalmology compare?

Findings  This study of the Centers for Medicare & Medicaid Services database for payments to ophthalmologists in 2012 and 2013 revealed that the average female ophthalmologist collected $0.58 for every dollar collected by a male ophthalmologist; comparing the medians, women collected $0.56 for every dollar earned by men. Women also had lower levels of clinical activity.

Meaning  Women in ophthalmology collect less than men because their clinical activity is lower; an examination of root causes to eliminate obstacles to parity in work effort may be in the best interest of a growing number of practitioners, the patients we serve, and the specialty as a whole.

Importance  The number of women in ophthalmology is rising. Little is known about their clinical activity and collections.

Objective  To examine whether charges, as reflected in reimbursements from the Centers for Medicare & Medicaid Services (CMS) to ophthalmologists, differ by sex and how disparity relates to differences in clinical activity.

Design, Setting, and Participants  Retrospective review of the CMS database for payments to ophthalmologists from January 1, 2012, through December 31, 2013. The dates of the analysis were February 1 through May 30, 2016. After exclusion of J and Q codes, the total payments to and the number of charges by individual ophthalmologists were analyzed. The mean values were compared using a single t test, and the medians were compared by the nonparametric Wilcoxon rank sum test.

Main Outcomes and Measures  Primary outcome measures were the mean and median CMS payments to male and female ophthalmologists in outpatient, non–facility-based settings. Secondary outcome measures included the number of charges submitted by men and women and the types of charges most commonly submitted by men and women.

Results  This study included 16 111 ophthalmologists (3078 women [19.1%] and 13 033 men [80.9%]) in 2012 and 16 179 ophthalmologists (3206 women [19.8%] and 12 973 men [80.2%]) in 2013. In 2012, the average female ophthalmologist collected $0.58 (95% CI, $0.54-$0.62; P < .001) for every dollar collected by a male ophthalmologist; comparing the medians, women collected $0.56 (95% CI, $0.50-$0.61; P < .001) for every dollar earned by men. Mean and median collections were similar when comparing female vs male ophthalmologists in 2013 (P < .001). The mean payment per charge was the same for men and women, $66 in 2012 and $64 in 2013. There was a strong association between collections and work product, with female ophthalmologists submitting fewer charges to Medicare in 2012 (median, 1120 charges; difference −935; 95% CI, −1024 to −846; P < .001) and in 2013 (median, 1141 charges; difference −937; 95% CI, −1026 to −848; P < .001) than male ophthalmologists. When corrected by comparing men and women with similar clinical activity, renumeration was still lower for women. In both years, women were underrepresented among ophthalmologists with the highest collections.

Conclusions and Relevance  Remuneration from the CMS was disparate between male and female ophthalmologists in 2012 and 2013 because of the submission of fewer charges by women. Further studies are necessary to explore root causes for this difference, with equity in opportunity and parity in clinical activity standing to benefit the specialty.

While multiple publications address the topic of physician sex and salary in many medical specialties,^1-3 there is a paucity of published literature on the association between physician sex, clinical activity, and collections in ophthalmology.⁴ Previous reports have documented sex disparity in authored publications,⁵ federal research funding,⁶ and editorial advancement,⁷ but academic studies of financial differences have been limited by difficulty accessing meaningful and accurate data. In 2010, the Patient Protection and Affordable Care Act required the Centers for Medicare & Medicaid Services (CMS) to report physician demographics and reimbursements, and that information is publicly available for 2012 and 2013. Release of these data provides the opportunity to determine whether sex-based differences in payments exist and if any differences can be explained by specialization, remuneration, and clinical activity through analysis of Health Care Financing Administration Common Procedural Coding System (HCPCS) codes.

The main objective of this study was to describe the association between physician sex and payments from the CMS to ophthalmologists in the United States in outpatient settings. We also sought to determine if there was a sex-based difference in (1) clinical activity (defined as the number of charges submitted to the CMS), (2) the mean reimbursement per charge submitted, (3) percentage representation for the most common billing codes, (4) breadth of practice (as measured by the number of unique HCPCS code submissions), and (5) collections by level of clinical activity.

This study was a retrospective review of the CMS database for payments to ophthalmologists from January 1, 2012, through December 31, 2013. The dates of the analysis were February 1 through May 30, 2016.

Quiz Ref IDBecause the data analyzed in this study were publicly available and no human participants were involved, institutional review board approval and patient informed consent were not required by the Johns Hopkins University Wilmer Eye Institute. Data containing individual-level information on payments to all practicing ophthalmologists in the United States were downloaded from the CMS database⁸ for 2012 and 2013, filtering by provider type. Information within the database included the National Provider Identifier (NPI), physician name and sex, HCPCS code and description, the number of services provided, the number of distinct Medicare beneficiaries receiving service, the number of distinct Medicare beneficiaries per day of service, the mean charges the clinician submitted for the service, the mean charges for the service allowable by Medicare, and the mean Medicare payment. The CMS database identifies clinicians with a metadata field “gender” and values of M (male) or F (female). For this study, we report these variables as “sex.”

Per the CMS,⁸ the mean payment and charge variables reflect the total payments or charges for a given HCPCS code or place of service divided by the number of services provided. Therefore, to calculate the total payments or charges for individual physicians, the mean payment and charge variables for each physician were multiplied by the number of services provided but separated by the overall place of service. The 2 categories for place of service defined by the CMS were non–facility based (O [outpatient]) and facility based (F), with distinct reimbursement patterns. Non–facility-based locations (offices, schools, assisted-living facilities, and federally qualified health centers) constituted most charges and payments for ophthalmologists; therefore, this payment category was the focus of our analysis. Because CMS payments include reimbursement for physician-administered drugs, all J and Q code charges were excluded to improve accuracy in accounting for payment for services.

To determine the Medicare reimbursement for a given clinician, the mean Medicare payment amount for each HCPCS code was multiplied by the number of services provided (line_srvc_cnt, as defined by the CMS, which reflects the actual count of events for most physician services), and all charges for each HCPCS code per NPI were summed. The number of charges to the CMS was calculated by summing the number of services provided for each physician NPI. Submitted charges for each NPI were calculated by multiplying the number of services provided by the mean submitted charge amount. For each year, the codes used most frequently by each sex were compared in terms of frequency, charges, and payments. The number of unique HCPCS codes submitted by individual physicians was taken to be the best indicator of breadth or diversity of practice. The total number of unique codes submitted by men and women was counted per year. The means and medians were also compared for these outcomes. To compare payments by clinical activity, all ophthalmologists were categorized by the number of charges. To minimize differences between men and women that were revealed by statistical tests of the mean charges per group, each year was categorized into several groups that were best fit for the distributions to enable valid comparison, providing 8 (12.5 percentile) groups for 2012 and 7 (14.3 percentile) groups for 2013.

Statistical analyses were performed using 2 software programs (Microsoft Excel 2016; Microsoft Corporation and Stata 14.2; StataCorp LP). All data were tested for normality and skew. The Levene robust test statistic was used to test for equal variances of variables between the sexes. To meet the distribution criteria revealed by normality and variance tests, the nonparametric Wilcoxon rank sum test was used to compare the medians for each sex. The median differences and 95% CIs were estimated via quantile regression models. The mean values were compared using a single t test for unequal variances where appropriate. All P values were nominal.

Quiz Ref IDIn 2012, a total of 19.8% of ophthalmologists receiving CMS payments were women. Women accounted for 12.3% of the total charges to the CMS and received 12.0% of the total collections. Of the 45 844 855 charges to ophthalmologists, 90.9% were non–facility based (Table 1). The mean number of charges per ophthalmologist was 2811 for men and 1631 for women, for a difference of −1180 (95% CI, −1262 to −1097; P < .001). The mean collection by female ophthalmologists was $106 932. For male ophthalmologists, this value was $185 405, for a difference of −$78 473 (95% CI, −$84 192 to −$72 754; P < .001). A woman working in non–facility-based ophthalmology collected on average $0.58 for every dollar earned by a man, which was proportional to the level of clinical activity. Distributions of the total collections by men and women are shown in the Figure. A histogram of payments to men and women, including J and Q codes, is shown in eFigure 1 in the Supplement).

The mean payment per charge was $66 in 2012 for women and men (Table 1). The median number of charges per ophthalmologist was 2055 for men and 1120 for women (difference, −935; 95% CI, −1024 to −846; P < .001). For female ophthalmologists, the median collection received was $73 671 compared with $132 632 for their male counterparts (estimated median difference, −$58 929; 95% CI, −$64 656 to −$53 202; P < .001). Comparing the medians, women collected $0.56 (95% CI, $0.50-$0.61; P < .001) for every dollar earned by men.

There were 489 unique HCPCS codes submitted by ophthalmologists in 2012 (Table 1). The number of unique codes submitted was 466 for men and 225 for women. On average, men submitted 11.78 unique codes compared with 10.05 for women (mean difference, −1.73; 95% CI −1.93 to −1.54; P < .001).

The 10 most frequently submitted HCPCS codes in 2012 are listed in eTable 1 in the Supplement. The average male ophthalmologist submitted more charges than the average female ophthalmologist for all codes listed. The number of charges submitted was used to categorize male and female ophthalmologists based on clinical activity (Table 2). Women comprised less than 8% of ophthalmologists at the highest clinical activity level (>5010 charges); within that subset, there were differences in the number of charges submitted (95% CI, −922 to 94; P = .03). eFigure 1 in the Supplement shows this phenomenon: the highest collection by a woman was $1 721 993. A greater percentage of men had collections exceeding $1 000 000, with the highest (outlying) collection at $8 909 557.

Quiz Ref IDIn 2013, a total of 20.6% of ophthalmologists receiving CMS payments were women. Women accounted for 12.8% of the total charges to the CMS and received only 12.5% of the total collections. Of the 46 629 794 payments to ophthalmologists, 91.0% were non–facility based (Table 3 and Table 4). The mean number of charges per ophthalmologist was 2860 for men and 1660 for women, for a difference of −1200 (95% CI, −1283 to −1117; P < .001). The mean collection by female ophthalmologists was $106 846. For male ophthalmologists, this value was $184 310, for a difference of −$77 464 (95% CI, −$82 926 to −$72 001; P < .001). A woman working in non–facility-based ophthalmology collected on average $0.58 for every dollar collected by a man. Distributions of the total collections by men and women are shown in the Figure. A histogram of payments to men and women, including J and Q codes, is shown in eFigure 2 in the Supplement.

The mean payment per charge was $64 in 2013 for women and men (Table 3). The median number of charges per ophthalmologist was 2078 for men and 1141 for women (difference, −937; 95% CI, −1026 to −848; P < .001). For female ophthalmologists, the median collection received was $74 351 compared with $133 566 for their male counterparts (estimated median difference, $59 162; 95% CI, −$64 883 to −$53 441; P < .001). Comparing the medians, women collected $0.56 (95% CI, $0.50-$0.61; P < .001) for each dollar collected by men.

The number of unique HCPCS codes submitted was 465 for men and 225 for women (Table 3). On average, men submitted 11.86 unique codes compared with 10.08 for women (mean difference, −1.79; 95% CI, −1.98 to −1.59; P < .001).

The 10 most frequently submitted HCPCS codes in 2013 are listed in eTable 2 in the Supplement. The average male ophthalmologist submitted more charges than the average female ophthalmologist for all codes listed. The number of charges submitted was used to categorize male and female ophthalmologists based on clinical activity (Table 4). Women generally collected less at all levels of clinical activity, and these differences reached P < .05 in multiple groups. Women comprised less than 9% of ophthalmologists at the highest clinical activity level (>4736 charges); within that subset, there were differences in the number of charges submitted (95% CI, −1035 to −97; P < .001), the mean collections (95% CI, −$91 940 to −$33 132; P < .001), and the median collections (95% CI, −$58 479 to $4585; P = .004). eFigure 2 in the Supplement shows this phenomenon: the highest collection by a woman was $1 581 313. A greater percentage of men had collections exceeding $1 000 000, with the highest (outlying) collection at $7 576 210.

In 2007, Weeks and Wallace⁴ reported the influence of physician sex on the incomes of American ophthalmologists using American Medical Association survey data from 1992 through 2001. After adjusting for key variables, the authors concluded that the reduction in annual income independently associated with female sex was significant and substantial ($55 091 [20% lower], P = .005). Women reported 24% fewer patient visits but worked only 5% fewer hours than men. In fact, increased work time was associated with lower incomes in their regression model, suggesting that clinical productivity and time spent with patients were key determinants of compensation. Range of practice (ie, medical vs surgical interventions) and submitted charges were not analyzed. Therefore, while the authors concluded that there was strong evidence of inequity between the sexes, actionable information for those seeking to correct disparity was limited.

Quiz Ref IDWith the recent release of CMS data, we are able to more accurately describe the sex gap in pay in ophthalmology as well as the association between clinical activity and payments. Studies of physician sex and payments for clinical care often are limited by small sample size, reliance on surveys^4,9,10 (which are complicated by selection bias), and review of regional data,¹¹ making it difficult to draw broad conclusions. This study aimed to address these issues by including a large and diverse group of ophthalmologists and reviewing the most recently reported federal data. However, our analysis has limitations. First, payments from sources other than the CMS are not reported publicly and thus could not be included. Second, physician ages and specialty status are not uniformly available (from the American Academy of Ophthalmology [AAO]), so we could not account for seniority or subspecialization in this analysis. It may be that male ophthalmologists are on average older, better established, or subspecialized and thus busier or managing more diverse pathologic conditions than female ophthalmologists. However, male physicians are known to have higher starting salaries¹¹ than women even at the beginning of their careers. Third, this research aimed to assess whether sex disparity exists and how it relates to clinical activity. This study was not designed to explain why women submitted fewer charges to the CMS. Further research is necessary to determine whether women differ from men in the ratio of government to commercial payers (eg, payer mix), ratio of facility-based to non–facility-based patients, and other practice styles, which may better illuminate the reasons for clinical activity and financial differences.

Despite the acknowledged limitations, this work provides insight into the professional practices of male and female ophthalmologists and is further validated by recent reports with similar findings. The most recent ophthalmology compensation report by Medscape¹² revealed that self-employed female ophthalmologists made $85 000 less than self-employed male ophthalmologists. The 2016 report by Doximity¹³ report stated that a 21% sex pay gap existed across all medical specialties. In that analysis, ophthalmology had the largest pay gap between women and men (−36% [approximately −$95 000 annually], P < .05), with a gross earnings difference similar to what was observed in this study.

Why is the pay gap markedly larger in ophthalmology than in other medical specialties? Comparing the medians, women in ophthalmology collected $0.56 for every dollar earned by men through the CMS in 2012 and 2013, with no evidence to suggest that women were less competent or had more limited aspirations. As expected, there was a strong association between payments and work product, with female ophthalmologists submitting fewer charges to Medicare in 2012 and 2013 than male ophthalmologists.

Some component of disparity may be a natural consequence of personal choices in specialization and full-time employment made by women, who are more likely to elect flexible hours and maternity leave early in their careers than men. Although women represent half of all medical school graduates,¹⁴ female ophthalmologists may be less likely to obtain competitive fellowships, specifically in vitreoretinal disease and refractive surgery.¹⁵ Surveys of American ophthalmologists have not been performed, but female ophthalmologists earn less in other English-speaking countries,¹⁶ perceive that they do not have the same advancement opportunities as male ophthalmologists (P < .01),¹⁶ believe that they are unequally treated by peers (P < .001),¹⁶ and think that parenting slows progress in their career (P < .001).¹⁷

Men and women may also have different ideals for professional achievement, and the best current data on this subject come from public sources. In the most recent report by Medscape,¹² women in ophthalmology were less likely to report “making good money at a job I like” as the most rewarding aspect of their job (4% of women vs 7% of men) and more likely to report “making the world a better place” as their primary motivation (13% of women vs 8% of men). Thirty-eight percent of female ophthalmologists were satisfied with their income, which was slightly lower than the rate reported by their higher-earning male counterparts (41%).

If the effort required to address sex disparity in clinical activity and financial reimbursement seems immense, the potential cost of neglecting this issue is greater, as noted by the following 3 observations. First, a woman earning $50 000 less per year than a man in ophthalmology faces an individual relative loss of $1 500 000 during a 30-year career before interest and investment returns are even considered. Based on data collected by the AAO, there are approximately 4500 female ophthalmologists in the American workforce; therefore, correcting disparity would yield approximately $6 750 000 000 to current AAO members during their working lifetimes. For a collections difference of $95 000 per year, this number grows to $12 825 000 000. Second, efforts to attract and retain women strengthen the specialty. For example, female ophthalmologists are less likely to be sued than male ophthalmologists (−54%, P < .001).¹⁸ Failure to address inequity may discourage women from entering the field and make our specialty collectively weaker. Third, a nationwide shortage of ophthalmologists¹⁹ is the foundation of the argument for increased optometric scope of practice. As more women enter the specialty of ophthalmology, one way to address medical need expeditiously is to identify and eliminate obstacles to maximum productivity of new and established female ophthalmologists. When women are enabled and encouraged, the ophthalmology community is better able to respond to public need and support diversity.

It is anticipated that our findings will promote introspection and discussion among men and women. A critical question is whether a detailed understanding of the association between work product, lifetime earnings, collective worth, and patient care will motivate changes in practice patterns. A separate question is how ophthalmologists will define success in the future and what role high collections will have in that definition. Despite collecting less on average, women may achieve equivalent net financial gains in the long term by performing a greater share of tasks not directly remunerated by insurers, having lower litigation risk¹⁸ (and insurance premiums), and promoting greater patient satisfaction²⁰ (which may be tied to federal payments in coming years).

Approximately 20% of ophthalmologists registered in the United States are female, and this number is growing. Between 2005 and 2014, a total of 37% of candidates taking the American Board of Ophthalmology Written Qualifying Examination were women.²¹ While the present study was not designed to determine what obstacles women in ophthalmology face in the context of parallel disparity in authored publications,⁵ federal research funding,⁶ industry advisory roles,²² and editorial advancement,^7,23 there is face validity to the position that women have fewer opportunities to pursue the same economic prospects as men. This finding warrants formal attention. As sex roles have evolved, working fathers are as likely as working mothers to struggle balancing the responsibilities of work and family, so an analysis of issues facing working women may directly benefit both women and men in the future.²⁴ Recognizing the personal and macroeconomic consequences of reduced clinical activity and associated disparity is a necessary first step toward promoting the full enfranchisement and prosperity of all ophthalmologists.

Corresponding Author: Ashvini K. Reddy, MD, Wilmer Eye Institute, The Johns Hopkins University, 600 N Wolfe St, Room Woods 472, Baltimore, MD 21287 (areddy16@jhmi.edu).

Accepted for Publication: November 21, 2016.

Published Online: January 19, 2017. doi:10.1001/jamaophthalmol.2016.5399

Author Contributions: Dr Reddy and Mr Bounds had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Reddy, Bounds, Bakri, Berrocal, Thorne.

Acquisition, analysis, or interpretation of data: Reddy, Bounds, Gordon, Smith, Haller, Berrocal, Thorne.

Drafting of the manuscript: Reddy, Bounds, Bakri.

Critical revision of the manuscript for important intellectual content: Reddy, Bakri, Gordon, Smith, Haller, Berrocal, Thorne.

Statistical analysis: Reddy, Bounds.

Administrative, technical, or material support: Reddy, Bounds, Gordon, Thorne.

Study supervision: Reddy, Bakri, Haller, Berrocal, Thorne.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported.

Funding/Support: This study was supported in part by grant FT130101648 from the Australian Research Council (Dr Smith) and by an unrestricted grant from Research to Prevent Blindness (Dr Bakri).

Role of the Funder/Sponsor: Neither funding organization had any role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.