|
What Does It Cost Physician Practices To Interact With Health Insurance Plans?
Physicians have long expressed dissatisfaction with the time they and their staffs spend interacting with health plans. However, little information exists about the extent of these interactions. We conducted a national survey on this subject of physicians and practice administrators. Physicians reported spending three hours weekly interacting with plans; nursing and clerical staff spent much larger amounts of time. When time is converted to dollars, we estimate that the national time cost to practices of interactions with plans is at least $23 billion to $31 billion each year.
ADMINISTRATIVE COSTS ARE HIGH IN U.S. HEALTH CARE, although there is disagreement over their causes and the benefits that may be provided by the activities that generate them.1 Costs incurred by physicians’ offices are an important contributor to overall administrative costs; interactions with health insurance plans appear to be responsible for a large proportion of physicians’ administrative costs.2 However, little information exists about the costs to physician practices of interactions with health plans. We present results from a national survey of physicians and medical group administrators inquiring about time spent interacting with health plans, specifically looking at prior authorization requirements, pharmaceutical formularies, claims, credentialing, contracting, and data on quality. In addition to generating a national estimate of the time—and the dollar value of the time—spent on these interactions, the survey data provide the first national information available on costs by the type of interaction and by the size and specialty type of physician practice.
Survey sample. Beginning with the 2006 American Medical Association (AMA) Physician Masterfile, our sample excluded physicians employed by the federal government, academic medical centers, or health maintenance organizations (HMOs) and physicians likely to spend most of their time in hospitals (such as emergency department physicians) or to have many self-pay patients (such as plastic surgeons). We then randomly selected 750 physicians from those identified by the Masterfile as working in solo or two-physician practices, and 560 physicians from those working in practices of three or more, for a total of 1,310 physicians. Selection was stratified by specialty type—primary care (including family physicians, general internists, and general pediatricians), medical specialists, and surgical specialists—for a total of 730 primary care physicians and 580 specialists. From the Medical Group Management Association (MGMA) Universe national file of 39,944 medical groups (the file includes information on both MGMA members and nonmembers), we selected the administrators for 629 groups, stratified by specialty.3 Survey instrument. We designed three survey instruments.4 The "physician survey" inquired about time spent on specific categories of interaction by the physician and by nursing staff working directly with that physician: registered nurses (RNs), medical assistants (MAs), or licensed practical nurses (LPNs). The "administrator survey" inquired about time spent by nursing and clerical staff providing practicewide functions; it was sent to administrators sampled from the MGMA group-practice database. Because physicians in solo and two-physician practices typically act as administrators, we developed a combined "physician-administrator survey," which was sent to 375 physicians randomly selected from the sample of 750 physicians in one-to-two-physician practices; the other 375 physicians in practices of this size received the physician survey. Survey questions asked about the minutes per typical day spent on each type of interaction.5 Surveys were designed based on review of previous research and on twenty-seven interviews conducted with physicians, practice administrators, and health plan executives. The surveys were then pilot-tested on fifteen physicians and administrators from different U.S. regions and were revised. Surveys were mailed in late July 2006 and again to nonrespondents in September and November.6 Statistical analysis. We categorized each physician and each administrator as belonging to one of nine size/specialty practice categories.7 We then calculated the mean time spent per day, week, and year for each type of interaction by physicians and the time spent per physician in the practice by each type of staff.8 We did so by adding the mean hours spent per physician calculated from the physician survey to the mean hours spent per physician calculated from the administrator survey. We converted time spent to dollars per year for physicians and for each type of staff, using external data on annual compensation, including benefits, and annual time worked.9 We conducted z-tests of differences between pairs of weighted means.10 Although we trimmed extreme outliers, the mean values in our data were increased by respondents who reported spending large amounts of time interacting with health plans. For the most important estimates, we therefore present median as well as mean values. The surveys and cover letter repeatedly and explicitly asked respondents to report time spent interacting with health plans; however, we were concerned that they would not be able to separate time spent on billing/claims interactions with health plans from billing/claims interactions with Medicare and Medicaid. Therefore, to be conservative, we reduced the claims/billing time estimates presented by 38.4 percent—the percentage of gross charges attributable to these payers—from the data reported.11 This adjustment was not made for other types of interactions, because they are rare or nonexistent for traditional Medicare and Medicaid (prior authorization, contracting), likely to involve small amounts of time (quality data), or require physicians to interact with private health plans (in Medicare Part D and Medicare and Medicaid managed care programs). The study was approved by institutional review boards at Cornell, the University of Chicago, and the University of Toronto.
Of the 1,939 physicians and administrators in the sample, 142 were ineligible because they were no longer practicing, were in an excluded category, or were no longer at their practice address. Of the remaining 1,797 people, 895 returned completed surveys, for a raw response rate of 49.8 percent. The adjusted response rate was 57.5 percent; this rate was 60.1 percent for physicians, 51.5 percent for physician-administrators, and 56.6 percent for administrators.12
On average, physicians reported spending forty-three minutes per workday—equivalent to three hours per week and nearly three weeks per year—on interactions with health plans (Exhibit 1
Of RN/MA/LPN time, 3.8 hours was spent per physician per day interacting with health plans (19.1 hours per physician per week). Note that this is not necessarily time spent by one person, but rather time spent per physician by all nursing staff in the practice. Median values were 1.8 hours per physician per day and 9.1 hours per week (median values not shown). Clerical staff spent 7.2 hours per physician per day—for a total of 35.9 hours per week. Median values were 5.9 hours per physician per day and 29.8 hours per week. Staff time did not vary systematically by specialty.
The average physician in a solo or two-person practice spent 3.5 hours weekly interacting with health plans (4.3 hours for primary care physicians, data not shown)—significantly more than for physicians in practices with ten or more physicians (2.6 hours; Exhibit 1
On average, physicians spent more time dealing with formularies than on any other interaction (Exhibit 2
When time is converted to dollars, practices spent an average of $68,274 per physician per year interacting with health plans (Exhibit 3  ). The median value was $51,043. Although per physician costs are lower in practices of ten or more physicians, there was not a statistically significant difference in costs by practice size. Primary care practices spent $64,859 annually per physician—nearly one-third of the income plus benefits of the average primary care physician (median spending for a primary care practice: $47,707). Practices comprising primarily medical specialists and surgeons spent $78,913 and $66,954 per physician, respectively. In comparing the amounts spent across specialties, one should keep in mind that the hourly compensation for specialists is much higher than that for primary care physicians.
The total cost of practices’ interactions with health plans cannot be precisely estimated using our data, because not all physician categories were included in the sample. An approximate calculation can be made by multiplying the number of physicians in U.S. office-based practice by the $68,274 per physician annual cost we found. The AMA lists 560,118 U.S. physicians in office-based practice in 2006, but this figure likely is approximately 19 percent higher than the actual number.14 Reducing the AMA number of physicians by 19 percent results in an estimate of 453,696 office-based physicians; multiplying this number by $68,274 results in an estimate of $31.0 billion as the total annual cost to physician practices of interacting with health plans ($23.2 billion if the median per physician value is used). This estimate does not include costs for interaction-related equipment, supplies, telephone, fax, or office space or for time spent by nurse practitioners (NPs) and physician assistants (PAs) interacting with health plans.
Surveys asked respondents to state, on a five-point scale ranging from "decreased a lot" to "increased a lot," the extent to which they believed their practice’s costs of interacting with health plans had changed over the prior two years. Forty-one percent of respondents stated that costs had increased a lot; 36.4 percent, that they had somewhat increased (Exhibit 4
Health plans claim to have taken steps to reduce the administrative burden they place on physicians; our data suggest that physicians continue to be discontented with this burden.15 The average physician reports spending nearly three weeks per year on these interactions (nearly two weeks if the median value is used); twenty-three weeks per physician per year are spent by nursing staff, and forty-four weeks, by clerical staff. A small fraction of this time—two hours per year by the average physician, and little more by their staff—is devoted to providing quality data to health plans or to reviewing quality data generated by plans about the physicians’ practice. Primary care physicians, especially those in small practices, spend larger amounts of time interacting with health plans than physicians in other specialties. Comparison with prior research. Three previous studies provided estimates of physicians’ administrative costs, but the first was conducted in 1995, the second relied heavily on the 1995 study and does not separate overall administrative costs from the costs of dealing with health plans, and the third relied heavily on a small sample of California practices.16 Our survey is recent; it is the first national survey to ask directly about time spent by nonphysician staff interacting with health plans and the first to provide data by type of interaction, type of staff, specialty, and practice size. The two prior studies that provide information on the time spent by physicians on interactions with health plans found results similar to our estimate of 3.0 hours per week. The 1995 survey found that physicians spent 3.0 hours a week on "insurance-related matters" comparable to those included in our study.17 Data from the small California-based study suggest that physicians spend 3.1 hours per week on these matters.18 A study conducted by Julie Sakowski and colleagues focused intensively on one very large multispecialty medical group and estimated that physicians spent thirty-five minutes per day (compared to our mean estimate of forty-three minutes and median estimate of twenty-eight minutes) on a set of activities similar to those included in our study.19 As in our study, physicians spent the most time dealing with formularies, and the next most time on authorizations. Study limitations. Our study has several limitations. First, although the physicians categories surveyed represent two-thirds of all physicians engaged in U.S. patient care, we did not include all categories of physicians. If physicians not included have lower-than-average costs of interacting with health plans, our per physician estimate will be too high. On the other hand, our total estimated cost may be too low: it did not include costs for the approximately 66,000 U.S. physicians classified as hospital staff, for the 97,000 U.S. residents and fellows, or for the staff working with these 163,000 physicians. Our estimate also did not include costs for health plan interaction–related equipment, supplies, telephone, fax, office space, or time spent by NPs and PAs. Second, our response rate was 58 percent. Recent work suggests that significant nonresponse bias is unlikely in physician surveys with response rates in this range.20 However, physicians and administrators who were more concerned with time spent interacting with health plans might have been more likely than others to respond to our survey. We cannot rule out this possibility, but we found no evidence for nonresponse bias in the check we were able to perform.21 Third, our estimates are based on respondents’ reports, not on direct observation. These reports varied considerably (in general, the standard deviations were approximately as large as the means); this could reflect inaccurate or exaggerated responses or true variation among physicians, whose practices have very different degrees of exposure to managed care. Some reassurance about the accuracy of the estimates may be gained from the fact that four different studies (including ours), using different methodologies, found quite similar estimates of the time spent by physicians; additionally, the patterns of responses across practices sizes, specialty types, and types of interaction are generally consistent. In the absence of a national time and motion study, which would be costly to conduct, respondents’ reports are the best source of data available. Fourth, our estimates include time spent by practices interacting with health plans in plans’ role as payers for individuals, employers, and the Medicare and Medicaid programs (which use health plans to administer managed care and formulary programs). They are not intended to include time spent dealing directly with Medicare or Medicaid (nearly all such time would be spent on billing) or general administrative time spent on overall physician practice operations. Our estimated mean $31 billion cost to physician practices of time spent on interactions with health plans is equal to 6.9 percent of all U.S. expenditures for physician and clinical services.22 It is six times the amount the federal government has spent annually on the Children’s Health Insurance Program (CHIP).23 The median cost estimated—$23.2 billion—is smaller but still substantial. Administrative spending produces benefits as well as costs. Nevertheless, although it is easy to portray administrative costs as "waste," it is important to emphasize that these costs can never be zero and that the interactions that generate these costs may produce benefits as well.24 Prior authorization and formulary requirements, for example, may reduce costs and improve the quality of care to the extent that they reduce inappropriate provision of services and promote the use of appropriate procedures and medications.25 Although it is expensive for physicians to deal with multiple health plans, each with its own requirements for authorizations, formularies, billing, and so on, it can be argued that having multiple plans generates the benefits that can flow from competition, including innovation and increased patient choice.26 It would be useful for further research to explore innovations in the policies and administrative procedures of health plans and physician offices that could reduce the costs of physician-plan interactions without also reducing benefits therefrom.27 Some efforts to achieve this goal are being made, with discussion occurring in such forums as the Healthcare Administrative Simplification Coalition. Created in 2006, the coalition includes, among other organizations, two large national health plans, the Centers for Medicare and Medicaid Services, Microsoft, the MGMA, the AMA, the American College of Physicians, the American College of Surgeons, and the American Academy of Family Physicians.
Larry Casalino (lac2021{at}med.cornell.edu) is chief of the Division of Outcomes and Effectiveness Research, Weill Cornell Medical College, in New York City. Sean Nicholson is an associate professor in the Department of Policy Analysis and Management at Cornell in Ithaca, New York. David Gans is vice president, Practice Management Resources, at the Medical Group Management Association (MGMA) in Englewood, Colorado; Terry Hammons is a senior fellow there. Dante Morra is an internist and Clinical Teaching Unit director at Toronto General Hospital, University of Toronto (Ontario). Theodore Karrison is a research associate professor in the Department of Health Studies, University of Chicago (Illinois). Wendy Levinson is the chair of the Department of Medicine, University of Toronto. Parts of this paper were presented at the AcademyHealth Annual Research Meeting in Washington, D.C., 9 June 2008. This study was funded by the Robert Wood Johnson Foundation and the Commonwealth Fund. David Gans and Terry Hammons are employees of the Medical Group Management Association, which is a founding member of the Healthcare Administrative Simplification Coalition. Gans coauthored (without compensation) a white paper to be published by the coalition. Lawrence Casalino is an uncompensated board member of the American Medical Group Foundation. The authors thank Alison Nair and Jessica Miotk for their assistance, and the American Academy of Family Physicians, American Academy of Pediatrics, American College of Physicians, and American College of Surgeons for providing letters of support for this study.
Related Article
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Study Data And Methods
Study Results
