Jackson CA, Brown JA, Relles DA, Lee PP. Work Time Estimates for Ophthalmic Diagnoses and ProceduresResults From the Eye Care Workforce Study. Arch Ophthalmol. 1998;116(7):922–928. doi:10.1001/archopht.116.7.922
To provide estimates of patient-level annual ophthalmologist work times for the care of common ophthalmic conditions and patient-level global surgical care time for common or important ophthalmic procedures.
A random sample of the domestic membership (excluding members-in-training) of the American Academy of Ophthalmology, stratified by self-designated practice concentration, was surveyed in 1994 to provide estimates of work times for common ophthalmic services.
Comprehensive and subspecialty-specific results were obtained for ophthalmic diagnoses, services, and surgical procedures. For ophthalmic diagnoses and services, initial and follow-up visit work times are reported for comprehensive and subspecialty ophthalmologists separately. For common surgical procedures, aggregate results based on comprehensive and subspecialist survey responses are reported.
These ophthalmology-specific survey results can be used for a variety of purposes, including practice management, "benchmarking," health plan administration, and national workforce planning. Such surveys should be repeated as techniques and practice patterns change over time.
THE WORK time required to provide medical services and to perform procedures has become an increasingly prominent consideration in health care provision. On a national policy level, work time estimates are critical to the estimation of workforce requirements for health care provision for the US population.1- 3 On a health plan level, estimates of the time required to provide health care services and to perform procedures are important not only for physician staffing purposes, but also for management of support staff, facilities, and other resources. For individual physicians, practice patterns and work times are critical to practice management and cost-based accounting for practice development and management.4 Thus, an accurate and up-to-date assessment of work times for common and important procedures and services is central to planning the provision of health care for individual ophthalmic care providers as well as national health policy makers.
This article gives the results of a survey of a random sample of the membership of the American Academy of Ophthalmology (AAO) of the work times associated with the most common and important medical services and procedures provided in eye care. The survey conducted by RAND, Santa Monica, Calif, for the AAO formed the basis for the work time estimates that underlie the Eye Care Workforce Study.2,3 While time estimates for some ocular services and procedures are available from a study conducted during the late 1970s,5,6 the results we report reflect current practices within ophthalmology.
As part of the Eye Care Workforce Study, surveys were mailed to 2500 members of the AAO. The surveys were allocated so that at least 200 surveys (except as noted for the low vision and uveitis subspecialties) were mailed to members of every subspecialty within ophthalmology, and 500 surveys mailed to comprehensive or general ophthalmologists. The diagnoses and procedures selected were the most common as derived from analyses of national data sets, such as the National Ambulatory Medical Care Survey,7 as identified by the subspecialty societies within ophthalmology as being the most important for their respective memberships or as mentioned in trade publication surveys of ophthalmic practice. Additional details and descriptions can be found in the full report of the Eye Care Workforce Study.2
For all subspecialties except neuro-ophthalmology, low vision, and uveitis, domestic fellows or active members of the AAO who are included in the organization's membership database were sorted by their primary area of practice concentration. For neuro-ophthalmology, low vision, and uveitis, ophthalmologists with a secondary area of concentration were added to the subspecialty eligibility list. Even so, fewer than 200 ophthalmologists were eligible to be surveyed for the uveitis and low vision concentrations. Members without a self-identified primary area of concentration were included in the comprehensive ophthalmologist category. No individual was included on more than 1 eligibility list. Two hundred members from each subspecialty eligibility list (except 31 for low vision and 76 for uveitis where all eligible ophthalmologists were selected) were randomly selected to receive a survey that focused on services and procedures within that subspecialty.
Each of the surveys followed the same format. First, demographic information was requested about the physician and the nature of the physician's practice. Second, the duration (in minutes) and frequency of visits in a 1-year period for a set of comprehensive and subspecialty-specific common diagnoses were requested. The comprehensive ophthalmology survey, as noted, was constructed to include the subspecialty-oriented services that were also commonly performed by comprehensive ophthalmologists. Assignment of services and procedures to specific subspecialty surveys or to the comprehensive ophthalmology survey was done with the input of the professional staff, advisory board, and survey results published in the non–peer-reviewed literature. For example, laser trabeculoplasty and laser panretinal photocoagulation were included on the comprehensive ophthalmology survey and the respective subspecialty survey, since these services are likely to have been performed by comprehensive and subspecialty ophthalmologists.
Third, the preoperative, intraoperative, and postoperative surgical times and the time for postoperative visits for common procedures were elicited. The choice of these procedures was made with the input of the advisory board and the subspecialty societies. As with the medical services portion of the survey, the procedures likely to be performed by comprehensive ophthalmologists, as well as by subspecialists, were included in the comprehensive ophthalmologist survey.
The survey was mailed to the 2500 randomly selected members of the AAO. The survey package included a self-addressed, prepaid, return envelope, and a strongly supportive cover letter from the AAO and RAND. Ophthalmologists were encouraged to call for a different survey package if the subspecialty designation of the original survey did not reflect their current practice; fewer than 1% (0.5%) requested a different survey. A reminder letter was mailed 3 weeks after the initial mailing. Five weeks after the initial mailing, a second package with the same survey form was mailed to those who had not responded.
The overall response rate to the mail survey was 40%. Data analysis was conducted to determine whether differences occurred in demographic, practice, or result variables between early and late responders. Similarly, demographic analyses were conducted between responders and nonresponders to identify statistically significant differences between the 2 groups.
The survey responses were keypunched and entered into a database. The reliability of the data entry process was enhanced by using duplicate key-entry procedures and verified through a review of a sample of surveys.
The survey responses were analyzed across comprehensive and subspecialty respondents for each diagnosis and procedure. When multiple subspecialties were surveyed, results were analyzed separately by each subspecialty and then in the aggregate. We report the median rather than mean work times to avoid undue influence of any extreme (outlier) responses. As noted in the Eye Care Workforce Study, the effects of sampling variation were explored and incorporated into the modeling using the bootstrap method, a statistical nonparametric approach to estimating confidence intervals.2
Table 1 and Table 2 give the median times for the major disease categories surveyed for initial visit times (Table 1) and total (number of follow-up visits times the duration of each visit) follow-up care in a 1-year period (Table 2). When subspecialists and comprehensive ophthalmologists were surveyed or when multiple subspecialties were surveyed, the disaggregated subspecialty-specific times are provided in Table 1 and Table 2.
The work time estimates reported in this study reflect the self-reported time in minutes spent by the responding ophthalmologist to provide care to the typical patient receiving the service or undergoing the procedure in his or her practice. While the survey also asked for time estimates for clinical assistants or associates such as technicians or optometrists, the results we report are based on only the time spent by ophthalmologists.
For office-based services, the times reflect the effort needed to provide patient care regardless of the type of third-party payer or insurance plan. In other words, the work times reported include the time needed to review medical charts or letters, speak with the patient or family, document the findings, or speak to another physician about the medical care of a patient. What is not included is the additional time required to complete extra paperwork or to obtain special authorization required by certain payers. Thus, the study results are robust across different payment schemes precisely because they do not include the time-consuming demands of insurance plan–specific, noncare-related paperwork.
Table 3 gives a list of the major procedures surveyed in the study, the length of the preoperative visit in minutes, the intraoperative time in minutes, the same-day surgery time (discussion with family members and dictation, for example), and the total 90-day postoperative follow-up care time (the product of the number and median length of follow-up visits during the 90 days after surgery).
Comparison of survey responses between early and late responders showed no differences in the work times reported. More important, comparisons between responders and nonresponders revealed no significant differences in those demographic characteristics reported in the AAO membership database. For example, 90% of respondents were US born, while 88% of the full membership of the AAO is US born. The mean age of respondents was 45.0 years, while that of the AAO membership was 49.4 years. The slight age difference may be due to the AAO's membership file including retired members who were ineligible to participate in the study. Respondents were distributed proportionally across the 4 main census regions, with 19.5% from the North, 25.7% from the central region, 30.6% from the South, and 24.2% from the West, compared with 24.5% in the North, 21.0% in the central region, 32.6% in the South, and 21.8% in the West in the full AAO membership. Ninety-five percent of the respondents were board-certified, a percentage similar to nonresponders. Practice settings were also similar, although there was a slight excess of university-type practices among responders compared with nonresponders.
The results we report from the Eye Care Workforce Study survey provide the most complete and up-to-date results from a nationally representative sample of the membership of the AAO for the performance of common or important ophthalmic services and procedures. As such, they provide useful information for individual practitioners, health plan and eye care network managers, and policy analysts interested in workforce and health care structure decision making. For example, comparing the payment and time required to produce different services or procedures may guide individual physicians in making decisions about the services or procedures they prefer to provide in a specific market situation. For health care managers, multiplying the time required to provide care for a patient with a specific diagnosis by the number of patients with that condition may permit more precise estimates of staffing needs, especially compared with the typical practice of staffing by population ratios for eye care services.4,9 In addition, as suggested in earlier research,10 these work times may come to be used as initial benchmark work times for providers of care to patients with specific conditions or for the performance of specified procedures.
This last possibility raises the need to understand clearly the limitations of the data in the present study. First, the survey asked providers to estimate the time required for their average patient. We did not ask providers to keep a work time diary, nor did we conduct a time-motion study; such methods may provide more accurate information.5,6 Validation of these survey estimates using different data collection methods in future research would be important. However, the survey method we used has been used elsewhere and has been accepted in workforce and reimbursement rate-setting studies (as in the Health Care Finance Administration's development methods for the resource-based relative value scale [RBRVS]).
Second, the survey results are not specifically adjusted for case-mix differences; survey respondents were asked to provide mean times and visit frequencies across all their patients. However, the survey results do reflect the subspecialty sampling stratification and the slight overrepresentation of university-based ophthalmologists. These sample characteristics are likely to be biased toward longer work time estimates if such ophthalmologists treat more complex and time-consuming cases than comprehensive ophthalmologists. However, comparison between subspecialty and comprehensive ophthalmologists for the diagnoses and procedures that are most likely to be shared do not show a significant time difference. It would be speculation to ponder why differences do not exist. It may be that comprehensive ophthalmologists take more time to provide care for patients than do subspecialty ophthalmologists caring for similar patients, but further study is needed. What is important is that the time estimates are stable across types of ophthalmologist providers and that underestimates of work time and, thus, the number of eye care providers required are extremely unlikely.
Third, we received a 40% response rate to our survey. While American Medical Association surveys reach response rates in the 60% range, we believe that the results reported here are stable estimates. We found no difference between early and late responders. There were no demographic differences between responders and nonresponders except for a slightly higher response rate by university-based ophthalmologists. Furthermore, we report results using median times as opposed to mean times to reduce the effects of any outlier responses.
A fourth limitation may arise from the misidentification of provider subspecialty or the subspecialty assignment of procedures to particular surveys. As mentioned, few respondents indicated that they received the wrong type of survey. During the survey development phase, the AAO and the specialty societies provided input for diagnosis and procedure selection and assignment. Thus, we expended considerable effort to avoid this type of data collection problem.
Despite the limitations of the survey, the results provide an informative initial assessment of the work times for eye care services and procedures. When these results are compared with the RBRVS survey results of Hsiao et al11 and during the restudy phase of the ophthalmology codes that was funded by the AAO,12 we find that our survey-generated work time estimates are generally higher, but with some exceptions.
Of note, the times for cataract surgery reported in our survey are significantly less than those of the RBRVS survey results.10,11 This most likely represents the more widespread use of phacoemulsification techniques and the resulting time efficiencies of such techniques. Indeed, our survey results are more consistent with the recent decision by the Health Care Financing Administration to reduce payments for cataract surgery beyond the amount projected in the initial Medicare reductions resulting from application of the RBRVS system. Whether a time-based rationale for reimbursement of physician services is desirable clearly remains a very different matter.
The use of time-based efficiency measures and monitoring will only increase as our health care system continues to change. Already, anecdotal reports indicate that multispecialty group practices are limiting cataract surgery to those ophthalmologists who are able to perform cataract surgery in a prescribed length of time, generally half an hour or less. With the ongoing industrialization of medicine, such changes will only accelerate, in keeping with business principles of using the most efficient providers to provide care.
Thus, there are obvious implications not only for time efficiency, but also for patient outcomes and quality of care. Addressing vital issues, such as the outcomes of care given by providers who have particularly short procedure times, whether the appropriate process steps are undertaken, and the satisfaction of patients with such services, is critical to understand as we create an integrated system of health care. We look forward to future investigations of these issues in different settings around the country.
Accepted for publication March 3, 1998.
Supported by the American Academy of Ophthalmology, San Francisco, Calif.
Reprints: Catherine A. Jackson, PhD, RAND, Health Sciences Program, PO Box 2138, 1700 Main St, Santa Monica, CA 90407-2138 (e-mail: email@example.com).