Correspondence
Reimbursement for Evaluation and Management Services
N Engl J Med 1999; 341:1619-1622November 18, 1999
- Article
To the Editor:
Lasker and Marquis (July 29 issue)1 offer a simple scheme to deal with the complexities of reimbursement for medical services, but it suffers from two major flaws, which the authors themselves point out. First, the system can easily be “gamed.” Because payment rates decline as more time is spent with a patient, income can be maximized by fragmenting care — treating two problems in two short visits rather than in one long one. More seriously, this payment scheme provides a very strong disincentive to treat difficult or time-consuming conditions, further aggravating the trend in managed care toward preferential treatment of patients with minor problems.
Stripped to its essence, the proposed scheme is to pay physicians on a piecemeal basis — so many dollars for so many visits per hour. It is a payment system that was tried by industry earlier in the century and largely abandoned because it was considered ineffective in a modern society. The system, which is still used to pay minimally skilled laborers in some Third World countries, rewards quantity over quality, speed over skill.
Lasker and Marquis's scheme is certainly a solution for the evaluation and management–coding morass, but the price will be further degradation in the quality of the medical care provided to our most seriously ill citizens.
1 ReferencesJean-Claude Bystryn, M.D.
New York University School of Medicine, New York, NY 100161
Lasker RD ,Marquis MS . The intensity of physicians' work in patient visits -- implications for the coding of patient evaluation and management services. N Engl J Med 1999;341:337-341
Full Text | Web of Science | Medline
To the Editor:
We have a strong sense of déjà vu. We analyzed the resource costs of evaluation and management services as a fundamental part of the resource-based relative-value scale (RBRVS) study. In 1988, we reported that regardless of the type of evaluation and management service, the site at which it is performed, or the specialty of the physician performing it, the level of work per unit of time varied only slightly.1 By 1992, we had surveyed more than 4000 physicians in 31 specialties on roughly 400 evaluation and management services, using vignettes that panels of clinicians had drafted. Multiple regression analyses showed that intraservice time (encounter time for office visits and consultations and time on the unit for hospital visits) predicted 90 percent of the variation in the level of work. Objective criteria, including the site of the service, whether the patient was new or established, and referral status, predicted a portion of the remaining 10 percent.2 We therefore recommended using time and these other criteria to bring Current Procedural Terminology (CPT) codes for evaluation and management services into closer accord with resource costs.1,2
Lasker and Marquis support our finding that intraservice time is a powerful predictor of physicians' perceptions of the work involved in evaluation and management services. Furthermore, whereas Lasker and Marquis surveyed 399 urologists, rheumatologists, and general internists about actual consultations and office and hospital visits and Iezzoni, in an accompanying editorial, noted that “calibrating their system would require new data on diverse specialties,”3 we had surveyed 10 times that number of physicians in every specialty and major subspecialty, using vignettes involving office, hospital, consultative, nursing home, critical care, and emergency services, according to established RBRVS study methods.
In its 1989 annual report to Congress, the Physician Payment Review Commission stated, “The coding system for evaluation and management services should be revised so that visits are classified on the basis of time as well as site of service, type of patient and referral status.”4 Indeed, the Health Care Financing Administration (HCFA) set the relative-value units for the work involved in evaluation and management services by selecting an intraservice time for each code and taking the level of work directly from our curves for work and time. Unfortunately, the American Medical Association's CPT Editorial Panel has defined evaluation and management services increasingly on the basis of content. Moreover, the resource costs of evaluation and management services performed according to current, content-based CPT documentation guidelines have, to our knowledge, not been studied.
The empirical evidence that time is an excellent predictor of physicians' work and the principle of avoiding needless administrative interference in how physicians practice lead us to urge again that intraservice time be used as a criterion for determining and documenting the CPT codes for evaluation and management services.
4 ReferencesPeter Braun, M.D.
Codman Group, Andover, MA 01810-1088Daniel L. Dunn, Ph.D.
Integrated Healthcare Information Systems, Lexington, MA 024201
Braun P ,Hsiao WC ,Becker ER ,DeNicola M . Evaluation and management services in the Resource-Based Relative Value Scale. JAMA 1988;260:2409-2417
CrossRef | Web of Science | Medline2
Braun P ,Dernburg J ,Dunn DL ,Cohen W . Predicting the work of evaluation and management services. Med Care 1992;30:Suppl:NS13-NS27
CrossRef | Web of Science | Medline3
Iezzoni LI . The demand for documentation for Medicare payment. N Engl J Med 1999;341:365-367
Full Text | Web of Science | Medline4
Annual report to Congress, 1989. Washington, D.C.: Physician Payment Review Commission, 1989.
To the Editor:
The most appropriate code for evaluation and management services is a dollar sign followed by the appropriate digits and decimal points. We now have sufficient information about what is involved in various medical services to establish the validity of such a code by having an experienced clinician evaluate the medical record. Everything from patients' complaints to statistical outliers could be used to trigger such inspections. Physicians found to have charged low fees could be given a gold star; in the case of ordinary fees, nothing would happen. Physicians whose fees were a little too high would receive a warning, those whose fees were very high would be thoroughly investigated and might have to refund money, and those whose fees were exorbitant or were for services that had not been performed would be charged with fraud and put in jail. The possibility of such an investigation would be just as effective as the possibility of an audit by the Internal Revenue Service in the case of an income tax return.
The coding industry has become a fraud itself. It is wasteful and abusive in the extreme.
John R. Dykers, Jr., M.D.
401 North Ivy Ave., Siler City, NC 27344To the Editor:
. . . If 10 percent of health care dollars are lost to fraud and abuse, that means that 90 percent of the claims are appropriate. Most health care providers do the right thing even with perverse incentives built into the system, so we should not be afraid of the flaws in Lasker and Marquis's proposal. As Iezzoni points out, “their general framework is . . . reasonable, straightforward, and consistent with clinical practice.” I believe we should adopt it now, while calibrating it to current data and working out the problems, rather than wait for it to be perfected. Their proposed system is no worse than the system we have today, it is much simpler to use, and I suspect it represents the way in which many physicians really do their coding in the privacy of their offices. Physicians should pressure HCFA to adopt a time-based system now and spend its money on evaluating and refining it rather than trying to fine-tune the current arcane and cumbersome system.
Leon D. Goldman, M.D.
Beth Israel Deaconess Medical Center, Boston, MA 02215To the Editor:
Reading the article by Lasker and Marquis, I was reminded of the Gary Larson cartoon in which, after covering a blackboard with sophisticated and arcane mathematical equations, Einstein discovers that time equals money.
A. Julianna Gulya, M.D.
George Washington University Health Sciences Center, Washington, DC 20037-3201To the Editor:
Lasker and Marquis's attempt at solving the problem of reimbursement for medical services is courageous and daring. Yet it does not take into account a vital part of care involving complex situations: the research and paperwork required for a large number of patients as part of their health care. I shall call it the after-office-hours work, which is often as important as the face-to-face encounter. This work includes but is not limited to telephone consultations with patients and other physicians, completion of insurance forms, and written requests for nonformulary medications. All other professionals who have a consulting role (accountants, lawyers, engineers, marketing experts, and others) charge a fee for time that is real but does not involve face-to-face encounters. In a primary care setting, a large portion of the day, and often the night, is spent reviewing charts, studying cases, reviewing recent literature, filling out forms, and making multiple calls about patients' problems. Is this time to be considered a free service when it comes to medical care?
Joseph Gutman, M.D.
Desert Samaritan Hospital, Mesa, AZ 85202To the Editor:
. . . Life would be simpler if all physicians were paid according to a fixed formula based on the worldwide standard for measuring labor: actual time spent working with the patient. The following formula appears reasonable: $50 for the first 20 minutes, $35 for the next 20 minutes, and $30 for each additional block of 20 minutes. To prevent the dehumanization of medical care, provision of low-quality care, and assembly-line economics, there should be no reimbursement for seeing more than three patients in an hour.
Minor adjustments will be necessary to account for differences in overhead and malpractice costs, length of subspecialty training, and differences in the cost of living. But in the long run, a fixed formula for paying physicians will prevent medicine from degenerating into a branch of science that is preoccupied with irrelevant documentation, bizarre codings, and artful billing.
Surendra Kelwala, M.D.
14555 Levan Rd., Suite E-307, Livonia, MI 48154To the Editor:
Primary care physicians face coding dilemmas every day. How do I code for the visit of a patient who presents for periodic follow-up of stable diabetes, hypertension, and hyperlipidemia but who also wants me to evaluate a new skin lesion and a newly sore knee and to suggest a treatment for periodic constipation? Not uncommonly, this same patient will have contracted an upper respiratory infection two days before the visit and will expect me to evaluate and treat it, too — all in a 15-minute visit. I face this type of scenario multiple times a day. The “oh, by the way's” that patients with stable chronic conditions save up for their regular visits to their doctors are a financial disaster for primary care physicians.
Lasker and Marquis's proposed coding scheme suffers from the same deficiency that the current CPT evaluation and management guidelines do. Both approaches assume that the physician is dealing with only one problem at a time or, with the CPT guidelines, up to three stable chronic problems. Until a coding scheme is developed that encompasses a mix of chronic, acute, and subacute problems, primary care physicians are left with four bad options for dealing with these visits. We can bend the guidelines and bill at the next higher level, bill for each service separately with the CPT code “-25” modifier (and usually not get paid for the extra services), bill for the service with the highest charge and give away the rest of the care, or require the patient to come back on a separate day for each problem. Obviously, all these options are less than satisfactory.
Ronald D. Reynolds, M.D.
1050 Old U.S. 52, New Richmond, OH 45157To the Editor:
Lasker and Marquis's proposal for simplifying coding for physicians' services is certainly “reasonable, straightforward, and consistent with clinical practice,” as Iezzoni notes in the accompanying editorial. It is also familiar.
Long before the current enthusiasm for increasingly complex documentation, the general internal medicine department in our large multispecialty group practice used simple guidelines to charge for outpatient care. They were based on time with fees weighted in favor of shorter visits, and distinctions among visits involving new, returning, and referred patients were made.
As far as I can make out, this arrangement was just what Lasker and Marquis now propose. And it worked well. Since current coding policies seem to satisfy no one, it may be time for HCFA to go “back to the future” and adopt a plan that is simple, fair, and cheap.
David R. Gutknecht, M.D.
Penn State Geisinger Medical Center, Danville, PA 17822Author/Editor Response
The authors reply:
To the Editor: We appreciate the support by Drs. Goldman, Kelwala, and Gutknecht for the simple reimbursement framework we proposed. Their comments are in accord with the feedback we have received directly from the medical community.
Dr. Bystryn is concerned that our scheme is a piece-rate system, which rewards quantity rather than quality. Although all fee-for-service reimbursement systems have incentives to provide more services (and none, to our knowledge, reward a good outcome or high quality), the system we propose is intended to achieve equity by paying physicians in proportion to their work. The scheme uses easily measured proxies for total work — face-to-face encounter time, new or established patient, new or ongoing problem, and referral or nonconsultative care — thereby reducing the burden of documentation. It reflects the work involved in treating patients by providing higher payments for longer visits than for shorter visits and by paying more per unit of time for types of visits that are more complex or that require more intensive care. Yet the scheme also rewards physicians who are efficient in delivering care — that is, those who deliver care that is similar to the care provided by other physicians but in a shorter period of time.
We agree with Dr. Gutman that “after-office-hours” work is as important as the face-to-face encounter. Indeed, the system we propose explicitly recognizes, and pays for, this effort. Although the coding in our scheme is based on blocks of encounter time, physicians would be paid according to the total amount of their work (i.e., the work they perform before, during, and after the face-to-face encounter with the patient). One reason that our scheme does not increase the payment in proportion to the amount of time spent in a face-to-face encounter is that shorter visits were shown to involve more time before and after the encounter than were longer visits.
Our finding that face-to-face encounter time is correlated with physicians' work in a study of actual visits substantiates the findings of Braun and his colleagues, who studied hypothetical visits. Moreover, the actual visits we studied included the full range of problems that physicians address during a single visit. Thus, contrary to Dr. Reynolds's concern, our scheme does not assume that the physician is dealing with only one problem, but instead reflects the mixture of problems in actual practice. Despite this mixture, we find that much of the variation in total work can be accounted for by a few easily measured and documented characteristics, and it does not require detailed and complex coding of the content of the visit.
Roz D. Lasker, M.D.
New York Academy of Medicine, New York, NY 10029-5293M. Susan Marquis, Ph.D.
RAND, Washington, DC 20005Author/Editor Response
In their totality, the letters confirm impressions of recent years: that debates about the coding of evaluation and management services engender strong emotions; that considerable energy is being expended on this topic, diverting attention from more pressing matters related to patient care; and that consensus about an optimal coding scheme is impossible to achieve. These letters largely reinforce my plea that HCFA move expeditiously to a straightforward approach for coding evaluation and management services that is consistent with clinical practice. In an increasingly electronic environment for medical-records documentation, it is unreasonable to expect the coding system to be the bulwark against Medicare fraud. Time — that vanishing commodity — should be spent on patient care, not on needless documentation and complex decision algorithms for quantifying that care.
Lisa I. Iezzoni, M.D.
Harvard Medical School, Boston, MA 02115
