Integration of quality assessment and physician incentives.

Integration of Quality Assessment and Physician Incentives

The measurement of quality is very elusive, and physicians have traditionally been reluctant to define it. Despite exhortations for self-examination from the physician community since early in this century, what we know and have done about quality assessment to date is remarkably little. Now, quality of care has become a public policy issue. Concerns about patient care are amplified by

well-performed studies that suggest that physicians do not consistently apply available knowledge in the care of patients and that uncertainty about the most effective diagnostic and therapeutic approaches is pervasive.

Government and others who pay for health care are impatient, as evidenced by accelerating efforts to fill the quality assessment vacuum. Crude and simplistic data on hospital mortality and other outcomes have cast the gauntlet before the physician community. Physicians must assume a central role in managing the quality assessment process. They should not abdicate their responsibility to create "standards kf care." The process of professional self-examination should not be an end in itself. Quality assessment should be a vehicle for improving the provision of care.

New measures will be necessary to judge the quality of entire health care systems, such as HMOs and PPOs. Employers are searching for new tools to make sophisticated judgments about the cost, quality, efficiency, and even efficacy of care. The reputation of the provider once served as a proxy for quality, but this is no longer enough. Employers will keep demanding cost containment, will keep seeking providers or systems that offer high quality with cost controls, and will demand hard evidence when they are asked to pay for it because it is "better."

As the Medicare prospective pricing system (PPS) progresses toward a single national payment rate, accounting for severity-of-illness variations becomes more important to providers and policymakers. Controlling for severity of illness and examining variables such as longer than average length of stay for specific conditions, higher than average readmissions, appropriateness of admissions, nosocomial infections, and deaths in anesthesia may allow valid cross-comparisons of hospitals and physicians.

Donabedian has set forth the formidable task facing clinical medicine. [1] Using Donabedian's classic categories of "Structure" (the settings in which care occurs), "Process" (that which is actually done in civing and receiving care), and "Outcome" (the effects of care on health status), inferences can be drawn about the quality of care. Good structure increases the likilihood of good process, and good process increases the likelihood of good outcome. Accrediting organizations are now developing methods of basing accreditation on quality assessment using measures of patient outcome. Ellwood has proposed a tool called "outcomes management," which uses epidemiological principles to evaluate the quality of care. [2] Unfortunately, it is still not possible to know for certain the extent to which an observed outcome is attributable to an antecedent process of care. It is most important that the measurement of quality of care is subordinate to improvement of the quality of care.

Properly managed, capitation is the most effective payment system for managed medical care. It rewards prevention and those interventions that promote healthy life-styles, such as weight reduction, smoking cessation, etc. The incentives in this system also reward low utilization, so there must be a strong emphasis on monitoring.

In capitated systems, there has been a great deal of discussion of individual incentives' being more perverse than group incentives. Any system that manage all aspects of medical care (including quality and cost) requires more effort. If we wish to promote a system that values good judgment in the management and provision of quality care, we should be certain that the physician who does this well by working harder receives better compensation than the physician who doesn't. Also, it is much easier to monitor the individual physician's decisions than to monitor the decisions of an entire group. A group's decision to limit care in a perverse fashion is much more difficult to track than such a decision by an individual physician.

U.S. Healthcare currently serves more than one million members in six Northeastern states from Delaware through Massachusetts. The company has a network of 13,000 private practice physicians and contracts with more than 300 hospitals. The majority of the population served lives in urban areas, the greatest concentration in Philadelphia and the surrounding area. US Healthcare contracts with private primary care physicians to provide care through compensation by capitation. We contract with specialty providers and other sources of health care in a modified fee-for-service arrangement. The amount of capitation and total compensation paid to primary physicians depends on their quality assessment rating and their ability to manage the cost of care effectively.

Since September 1, 1987, we have operated a system that uses quality criteria to categorize all primary care offices in our service areas. Prior to this time, physicians were paid on the basis of 80 percent of capitation with 20 percent "withheld" for possible return at the year's end. This withhold was a pooled fund and thus, under little control of the individual physician. In addition, at the end of the year additional distributions of funds were made based upon whether the office had been cost-effective and less had been spent than had had been allotted for the care of patients. Basing an incentive soley on the cost of care did not account for quality. Utilizing a pooled withhold did not allow for individual accountability. Because the distribution was paid 3-4 months after the end of the year, its lack of timeliness did not provide as meaningful an incentive.

Therefore, we decided to improve the system and make it more responsive to quality issues. This new categorization (figure 1, above), determines the rate or "tier of capitation" paid. It also determines the frequency of any distribution of dollars paid in excess of capitation earned by virtue of cost-effective care. In September 1987, the majority of offices in our system were categorized. This initial categorization did not affect their capitation. It served to demonstrate the model and to alter the frequency of distributions. In January 1988, the offices were categorized again based on the latest 6 months of data. This categorization affected both capitation and frequency of distributions.

The offices are now recategorized every 6 months from "1" to "5" (1 being the best). A "1" office is paid the highest tier of capitation. A "3" office is paid 80 percent of the highest tier of capittion. A "5" office is paid 60 percent of the highest tier of capitation. There is no longer any "withhold" of funds in our system. Category "1" and "2" offices receive distributions of earned incentive payments on a monthly basis and "3" offices receive distributions on an every other month basis. A category "4" or "5" office receives distributions on a yearly basis.

Approximately one half of an office's category is determined by valuing the utilization of services. Utilization is measured by looking at the cost of hospital care, specialist care, and emergency care. Offices are provided with a statement of their utilization in terms of hospital days per 1,000 members per year and specialty and emergency costs in dollars per member per month (table 1, page 16) and of their utilization performance in comparison to all similar primary offices (figure 2, page 15). These data are converted into a score that is incorporated into the calculation to determine the office category. The utilization data for the period involved (e.g., kna month or 2 months) determine the amount of distribution for that period. There is a cap on distribution of additional funds related to both high and low utilization of services.

Our measurement of quality has four aspects--chart audits, member surveys, transfer rates out of an office, and philosophy of managed health care (figure 1, page 14). Every office in our system has its charts audited for at least two medical care evaluations per year. These evaluations examine specific aspects of care according to designated and publicized standards. For example, we assess documentation of immunization status in children, measurement of blood pressure in hypertensive patients, or measurement of cholesterol in adults over 40. These data are compared bor all physicians in our system and to the designated clinical standards. The standards are developed and publicized by quality assurance committees composed of physicians in our network.

Whatever its strengths and limitations as a measure of quality, information about patient satisfaction should be an indispensable tool in the assessment of quality. We survey almost all our adult members each year for information about the quality of service in offices. We request information about physician availability, waiting time, office personnel, personal concern of the physician, etc. The data collected from each office are compared for all physicians in our system, and a detailed report is provided to each office (table 2, page 16). We use this information as a management tool for the office as well as a mechanism to compare and value offices.

We use detailed information about the transfer rate of patients out of every office to compare offices. We investigate an increased rate of transfer from an office, and, if there is an underlying quality concern, the office is valued less in this system than those that have low transfer rates. With few exceptions, an elevated transfer rate is a manifestation of poor quality of service and/or care.

We also have a measure of managed care philosophy, which combines all the aspects of working within the managed care system. Physicians who serve on our Executive advisory Committee, our Quality Assurance Committee, or our Membership Committee are valued highly in this regard. Offices that have been sanctioned because of quality concerns are valued much less in this model. Willingness to function within the managed care system is readily evaluated by virtue or our medical directors' being involved with and visiting each office at least once a year. Also, our professional service coordinators are in each office at least four times per year. This intense contact allows us to make judgments about the philosophy of an office. An office that uses our healthy outlook programs, participates in our committees, and attempts to help patients receive quality care in our system would score 70 to 100 out of a possible 100 points. The office that is unwilling to participate in our healthy outlook programs, has been resistant to changes in its behavior when significant deficiencies have been brought to its attention, and has been sanctioned for poor compliance with quality issues scores between 10 and 30 points.

Each of these quality elements is valued numerically and a score is calculated. The weighting was determined by our consideration of the importance of each aspect in delivering quality care to our subscribers. Each of the aspects carries approximately equal weighting, with the survey and the philosophy of managed care valued slightly more. The quality measures account for approximately one half of the categorization.

In addition, we have developed a mechanism to increase the distribution of dollars earned by multiplying the distribution by a percentage dependent upon the quality rating (figure 1, page 14). High-quality offices that function poorly as "managers" have their categories upgraded by their high-quality ratings, which balances their lack of efficiency. Offices whose tendency might be to give less than optimal care and service will tend to have poorer results on their surveys, poorer results on their audits, higher transfer rates, and poor evaluations of their philosophy. Thus, while their utilization scores may be high, their quality scores will be low and their categories will be lowered. In addition, their multipliers derived from poorer quality will diminish or even completely eliminate any additional distributions. We have handled the severity-of-illness variable by placing a limit on offices' fiscal responsibility and by equalizing the cost of catastrophic care for patients over the entire network.

We are beginning to accumulate evidence to support our contention tht this model has been able to motivate physicians to alter their style of practice in a very positive manner. Both quality of care and quality of service have improved, and there has been a positive financial impact on the physician. For example, we have seen a rate of compliance with out immunization standard rise from 62 percent to 84 percent. Many offices have documented a dramatic improvement on their patient surveys.

In 1988, Drs. Chassin, Kosecoff, and Dubois (authors of the RAND Corp. studies of appropriateness of care) evaluated our program. They summarized their findings by stating that as an IPA model, U.S. Healthcare has a more comprehensive set of tools to assess quality and a more definitive mechanism to improve quality than any other health maintenance organization about which they had information. Thus, a credible outside evaluator has confirmed that we do what we say we do and that we do it very well.

References

[1] Donabedian, A. "The Quality of Care: How can It Be assessed?" JAMA 260(12):1743-8, Sept. 23-30, 1988.

[2] Ellwood, P. "Outcomes Management: A Technology of Patient Experience (Shattuck Lecture)." New England Journal of Medicine 318(23):1549-56, June 9, 1988.

Neil Schlackman, MD, is Medical Director, U.S. Healthcare, Blue Bell, Pa.