All job advertisements in the journals Ophthalmology, American Journal of Ophthalmology, and Archives of Ophthalmology during the 318-month period from January 1980 through June 2006 were reviewed. Each job advertisement was disaggregated into the number of individual positions advertised, and each position was coded. Coding was done along three axes: practice type, geography, and subspecialty.
Advertisements were coded into one of three practice types: academic, private, and government. For an advertisement to be coded as an academic position, the originating institution had to be an academic center or an academic center affiliate, or the position had to be advertised as either an academic appointment or a professorship. Positions at Veterans Affairs hospitals were also coded as academic positions, because advertisements for ophthalmologists from Veterans Affairs hospitals invariably mentioned academic affiliations. Positions were coded as government jobs if they involved working for governmental agencies other than the Department of Veterans Affairs. All other positions were designated as private practice positions.
Geographical coding within the United States was done according to a scheme derived from Forman and associates41
that is summarized in . Positions outside the United States were coded as either Canadian or International.
GEOGRAPHIC CODING SCHEME FOR HELP WANTED INDEX
Advertisements were divided among 15 groups: comprehensive ophthalmology, retina, cornea (and/or anterior segment), glaucoma, uveitis, pediatrics, oculoplastics, neuro-ophthalmology, research, pathology, cataracts, low vision, oncology, education, and leadership. Jobs that did not fit into any of the preceding categories were coded as “other.” Comprehensive ophthalmology was used to code advertisements for comprehensive ophthalmologists, positions without subspecialty designations, and positions mentioning more than two specialties. In cases where two subspecialties were mentioned for a single position, the position was coded as 0.5 of each of the mentioned specialty positions. Advertisements for directors of residency programs were categorized as education positions. Advertisements for leaders of ophthalmology departments or leaders of ophthalmology departments in satellite locations were coded as leadership positions. Positions that involved research for more than 50% of the time were coded as research positions.
The HWI generated in this study includes all job advertisements for ophthalmologists printed in three academic journals, namely, Ophthalmology, American Journal of Ophthalmology, and Archives of Ophthalmology. This represents a subset of all ophthalmology job advertisements and does not capture other sources of job information, such as the Internet (notably the American Academy of Ophthalmology’s Academy Jobs Web site), nonacademic ophthalmology journals, and headhunter firms. Therefore, when appropriate, sensitivity analyses are carried out to further assess the conclusions drawn from the study.
Coded positions were aggregated on a monthly basis to generate the HWI. To determine the economic factors that drive demand for ophthalmologists, multiple regression analysis was used to discern relationships between subgroups of the annualized HWI and various economic variables. Only real (inflation-adjusted) variables were used for the analysis. If only nominal (non-inflation-adjusted) data were available from the various sources, the nominal data were adjusted by dividing the variable by the ratio of the consumer price index (CPI) deflator values for the year in question and the year preceding it. Symbolically, real Xn
= nominal Xn
), for variable X in year n. CPI data was obtained from the US Department of Labor’s Bureau of Labor Statistics.46
To assess the historical impacts of recession on the marketplace, demand for ophthalmologists during and after the recessions from July 1990 to March 1991 and March 2001 to November 2001 were compared. Similar analysis was done with published data on the marketplace demand for radiologists.40–45
Multiple regression analysis was also used to establish a relationship between the marketplace demand for radiologists and economic drivers. Statistical analysis mostly consisted of a comparison of proportions or means, and statistical significance was determined by means of Student’s t
test. For multivariate correlations an analysis of variance test was done and statistical significance was determined using the F
test. Resultant P
values were deemed significant if less than .050 The variables and HWI subcomponents that are used in this thesis are described below:
refers to an aggregate of private, federal, state, and local government funding dedicated to research. These numbers were obtained from the Centers for Medicare and Medicaid Services of the US Department of Health and Human Services.47
is the annual percentage appreciation or depreciation of the Standard and Poor’s 500 stock market index. S&P was calculated by dividing the adjusted closing price of the S&P 500 on the last trading day of a particular year with the price on the first trading day of that year. Symbolically S&Pn
= (pricelast trading day
/pricefirst trading day
). Data for the S&P 500 was obtained from Yahoo! Finance using the ticker symbol: ^GSPC.48
refers to the gross domestic product, which is the sum of all goods and services produced in a country in a year. The GDP data used for this analysis was obtained from the Bureau of Economic Analysis of the US Department of Commerce.49
CMS refers to the annual expenditures of the Centers for Medicare and Medicaid Services.
refers to the annual sums spent on durable medical equipment. These numbers were obtained from the Centers for Medicare and Medicaid Services of the US Department of Health and Human Services.47
refers to out-of-pocket expenditures on “services provided in establishments operated by health practitioners other than physicians and dentists. These professional services include those provided by private-duty nurses, chiropractors, podiatrists, optometrists, and physical, occupational, and speech therapists, among others.”47
is the annual number of ophthalmology fellowship positions offered by training institutions. This information is published in the Fellowship Match Report of the Association of University Professors of Ophthalmology.50
is the annual number of applicants to ophthalmology fellowship programs. This number is the sum of applicants to the various ophthalmology fellowship programs. These numbers are published in the Fellowship Match Report of the Association of University Professors of Ophthalmology.50
is the number of applicants who register for the San Francisco Matching Program. This information was obtained from the Ophthalmology Match Report of the Association of University Professors of Ophthalmology.51
Physician compensation variables
refer to the physician compensation data from the Physician Compensation and Production Survey series of the Medical Group Management Association.52
The specific compensation categories used in this thesis are listed below:
- Median academic compensation is the median total compensation of academic ophthalmologists.
- Mean academic base compensation is the mean base salary of academic ophthalmologists.
- Mean retina compensation is the mean total compensation of retina specialists.
- Median academic retina compensation is the median compensation of academic retina specialists.
Academic HWI is the annualized sum of all advertisements for academic ophthalmologists. Academic HWI is a proxy for the marketplace demand for academic ophthalmologists.
Private practice HWI is the annualized sum of all advertisements for private practice ophthalmologists. Private practice HWI is a proxy for the marketplace demand for private practice ophthalmologists.
Private subspecialist HWI is the annualized sum of all advertisements for nongeneralist, private practice ophthalmologists. This series was obtained by subtracting general practice positions from the private HWI. Private subspecialist HWI is a proxy for the marketplace demand for private practice subspecialists.
Academic retina HWI is the annualized sum of all advertisements for academic retinal specialists. Academic retina HWI is a proxy for the marketplace demand for academic retinal specialists.
Radiology HWI is the annualized sum of all advertisements for radiologists. Radiology HWI is a proxy for the market demand for radiologists.