Managed care is data management.

Managed Care Is Data Management

Accountability, the crux of the managed care movement, is the fundamental principle by which we will formulate new ways to manage patients and doctors better in response to demands for value. We must recognize, however, that as managed care decentralizes and delegates more care, and as patients' prerogatives increase, control and accountability will diminish. There are essentially two ways to restore some control:

* Data management, coupled with an advisory role for physicians.

* Restricting payment for services rendered.

The latter is called the adjudication process; it may cause difficulty at or near the point of service. The former recommends an accessible, organized medical database, one that actually recapitulates the medical logic used in comprehensive care. These vital data, however, have been unavailable, and making them available may create new problems for providers. The power of the medical database and concurrent monitoring/tracking systems can create new demands and threaten confidentiality. Thus, to manage care optimally, we must not only organize the care so that less work is required, but also be mindful of our primary purpose: to improve the health of our patients. This all requires expertise in patient communication, the careful handling of sensitive information, and active support of the advocacy role.

We have or should have the ability to supply service-based information about medical claims and clinical processes. We know how to take care of patients comprehensively, and we are expert at managing care. But we do not have the data we need. A more accessible medical database is needed at every point of patient contact. Nice, but not required, is the assistance of a computer in making this relevant information available.

Is there a better way to obtain valid and accurate medical data than by mandating it? There is, if we pay physicians only for what is appropriate (i.e., gaming is kept out of the system), all financial incentives that conflict with the advocacy role of physicians are removed, and the utility of the data is increased. The data will be more useful if they are available 24 hours a day; if physicians are taught to represent their logic in a summary list of medical problems, concerns, or preliminary assessments; and if effectiveness and efficiency are regularly evaluated from the data.

The data can be made more accessible by having well-organized, up-to-date lists of vital medical information, such as the problem list (an index and table of contents of the medical record) and a chronology of recent medical interactions, on-line and available from a centralized location. The confidentiality of this medical information is essential. It can be protected by making the managing entity responsible for the security of the data.

Finally, how will such a system be paid for? Insurers and others who have a vested interest in the "appropriateness decision" will be more than happy to pay for the opportunity to work with physicians to improve care. What better way to establish protocols and standards of diagnosis and treatment and influence the appropriateness decision at the point of contact and/or service? Also, having a computer-assisted database management program would facilitate the documentation of advice-giving education, the flagging of patient problems that warrant special attention and follow-up, and feedback procedures.

The best opportunity not only to advocate for patients but also to improve accountability lies at the point of patient contact. Although this is not the usual place to begin documentation, it is an important, underrecognized, and constant referent. Furthermore, it may be the only reliable point where patients will contact us, if for no other reason than for authorization or advice.

Present Medical Data

What touches all aspects of medicine is "information." Data are not yet information. Nevertheless, medical information and clinical logic are essential ingredients of effectiveness and efficiency studies. [1-3] Unfortunately, much of the analysis of medical care is cost driven, with the claims trail being its only reliable dataset. The fact that allocation of resources does not recapitulate the medical logic has not seemed to matter. In theory, health maintenance organizations and other managed care systems should excel at linking "claims" to the "clinical trail" (i.e., utilization monitoring in the pursuit of medical diagnoses and treatment). Nevertheless, no medical management system in America (and certainly not in Canada) asks physicians for their deductive reasoning. When logic is sequestered in our charting (even in the computerized medical record) or is absent entirely, or the analysis of medical practice is kept so distant from the care process itself, it is not surprising that it is no easier to control costs than it is to anticipate physicians' behavior.

Present data management systems fail to document what happens to the patient throughout the entire episode of care and thereby inadvertently contribute to discontinuity and depersonalization. Gaps in our documentation (knowledge) interfere with our efforts to improve the efficiency of health care. Furthermore, an incomplete database plays havoc with our ability to learn from the past, improve quality and effectiveness, and develop strategies for the future.

There is confusion about what is meant by efficient health care delivery systems. Central to our understanding of the concept kf efficiency are its components: the actual outcomes of clinical intervention, the efficacy of the treatment, and the desirability [1] of observed outcomes and productivity. We must be able to feed information back about the physician performance, if we are to modify medical practice. [2,3] Somewhere in the available dataset, there must be information that will allow us to evaluate performance. Not unexpectedly, this also happens to be the place where we will find support for appropriateness decisions.

Some of the vital information by which we should measure/evaluate production may be missing or ignored. A comprehensive dataset about efficiency and cost-effectiveness, for example, should include data about patient calls, inquiries, understanding, and the like. Our methods of getting this type of information, however, are onerous (e.g., computerization of the medical record) or unnecessarily labor-intensive (e.g., surveys and studies).

A Complete Picture

A complete picture of care in medically managed systems begins with any contact and ends at the outcome (intermediate or final). The continuity of medical care is the bridging of these two reference points. Presently, we do not have in place an effective method to track the patient from the point of contact. [4] If we wait for a point of service, we may delay tracking, or we may actually miss a few medical transactions that do not involve an encounter. We may even compromise our ability to assess accessibility.

Furthermore, the medical database is incomplete by our own admission. It is rare to find well-organized, comprehensible, readily available, and accessible "problem lists." The incomplete database is the proximate cause of our inability to control or audit what we do and to improve the practice of medicine. With an incomplete or inaccurate database, we can neither improve continuity nor prove discontinuity. We cannot fully evaluate the use of health care resources (e.g., when trying to provide cross-coverage). Finally, we cannot, at times, even properly identify patients or their medical histories!

Outcome measures, largely relegated to quality assurance discussions, evade practical appliction in utilization management. The unfortunate truth is that the results of our medical efforts may be unrelated to what we have done or will do on behalf of the patient. There are essentially two reasons for this state of affairs. When we have not sufficiently compensated or accounted for population, demographics, severity of illness, and local variation in practice patterns, what we claim to observe may, in fact, not be real. A "complete picture of care" and the ability to develop medical strategies requires a fairly complete database and sophisticated statistical understanding. The second reason is the contribution of unforeseen events, variable compliance, unstandardized practices, and a general inability to observe the complete process of care.

Tier 1 is cause and effect; it is the medical part of the process. Tier 2 is the quality and utilization review/management process. The complete process encompasses both tiers, and every step is critical.

Linking Trails

The clinical process of medicine is fairly straightforward. The pursuit of the diagnosis, or the clinical trail, is a scientific process of elimination. How, then, can the result of that pursuit be made credible or confirmed as reasonable or appropriate through retrospective utilization review of health care resource expenditures? Physicians will tell anyone that utilization control is not their overriding concern in patient management. Moreover, because the e ical outcome may or may not be related to a physician's skill, industry, or past performance, it is perplexing that management would judge physicians' performance on the basis of utilization standards alone. Indeed, retrospective auditing and trend studies do not anticipate utilization very well. It might be better to manage the care before or as it happens, rather than simply react to it.

Health promotion personnel have advocated a more preventive approach to medicine. However, they have not been able to convince the insurance and medical establishment of the efficacy of their method. Insurance systems cannot relate to known risk and, furthermore, do not want to remove any premiums from their existing revenue stream; i.e., excluding people who are only potentially sick.

The major challenge seems to be bringing to the forefront information about the "more manageable aspects" of health behavior, e.g., smoking, sedentary life-styles, stress reactions, and the like. How can we use these data to improve our advocacy role? How can the information help us decide what service is appropriate and what is not? The answer, it would appear, is the cooperative creation of an accessible profile of the patient's medical history--with past, present, and potential problems.

MOM*

The value of having a reliable chance to determine appropriateness or at least to monitor care should be obvious. The most reliable opportunity to not only perform this service, but also improve the utility of the medical database, [1-3] is the point of contact. (This may or may not be the point of service.) Medicine Optimally Managed is different from other database systems in that it interfaces an intelligent, caring, expect communictor who is knowledgeable about the patient and the policies and procedures of the managed care system. Although the specific items in the database are likely to vary between settings, they will always contain organized lists of data, generally prepared choronologically to reflect the process and the complete picture of care.

The theoretical advantages of such an information system includes its ability to advocate for patients, explain the rules, determine appropriateness, reduce medical/legal risk, monitor care from the point-of-contact to final disposition, ensure accountability, provide follow-up for patients, and even create the framework for risk prediction and medical strategy. Furthermore, it also presents an opportunity to insulate physicians from inapproprite, inefficient (nonreimbursable) use of their time.

In our conceptualization of a relational database management scheme, it is necessary to first tract the flow of information that a triage-type person would access and control (figure 1, page 22). Next, the triage algorithm is put forth (figure 2, page 22). The key to this method is that critical information can be put to good use from a centralized location (figure 3, above). The eventual disposition for the patient contact or transaction can be easily correlated with (or sought because of) contact information.

Dispositions may reflect way stations on the clinical trail or "immediate outcomes" (e.g., the patient was hospitalized), or when not "final," they may serve as a sort of "tickler file" in the interests of folow-through and continuity of medical care. In summary, it becomes exceedingly important to relate the clinical trail to the claims trail when assessing efficiency or cost-effectiveness. In addition, other dispositions--administrative decisions, precedents, feedback procedures, etc.--may be of relevance. It is noteworthy, in this context, that any such systems must have the ability to remember the decisions we make and the thought processes we went through. Thus, it must have updating capability and ways to recall the clinical pathways we have taken.

Conclusions

Data management requires that we (1) access all the relevant data, from point-of-contact to outcome (intermediate and final disposition), (2) provide a computer interface to improve our ability to interact with and then advocate for the patient, (3) control patient movement (or at least monitor it well), (4) use the same data to link the process of care with the outcome, and (5) learn from these data what works and what doesn't. [5]

Doctors need feedback about their patients and effcacy. What transpires and is recorded in the office and hospital chart is only of use during an encounter. Rather than rely on the vicissitudes of medical practice and variable documentation skills, systems should be designed to capture, correlate, and reconcile certain lists of medical care/service information. We must standardize the dataset. In so doing we may use it in the determination of effectiveness and efficiency. In preparation for the future, we must use such intelligent systems to learn from the past and our shared, collective experience.

A relational database program that takes mainframe data and brings them into a micro- or minicomputer also brings this kind of data access and management into the realm of possibility for most physicians working in patient care environments. The "appropriateness-decision" that we must make as part of our responsibility to insurers is also justifiction for their subsidy of the installment of such a program.

Studies of point-of-contact data management schemes have demonstrated the utility of the approach. [5] It is not hard to realize the benefits of such a system: repersonalization, continuity, documentation, tracking, the preparation of data to become information.

The major challenges include demonstrating the utility and security of the data management scheme to physicians, achieving cooperation from them to contribute their medical databases (even diagnoses related to claims will do at the outset), and setting up the programming that is required to allow a generic software system to be loaded into computer memory (RAM) and run with the data stored in mainframe computers. [6]

References

[1] Bohen, O., and Burbe, T. "New Directions in Effective Quality of Care: Patient Outcome Research." Federation of American Health Systems Review 21(5):50-3, Sept./Oct. 1988.

[2] Kaplan, J., and O'Connor,P. "Productivity Can Be Econometrically Considered in an HMO Setting: Its measurement and Implications for an Incentive Program." Presented at the Group Health Institute, Philadelphia, Pa., June 1984.

[3] Kramer, R., and Kaplan,J. "Physician Performance Evaluation and a Further Definition of Productivity." Presented at the Group Health Institute, Minneapolis, Minn., June 1986.

[4] Kaplan, J., and others. "An Argument for a

Point-of-Contact Management system." Medical Interface 2(2):23-8, Feb. 1989.

[5] Kaplan, J. "Accountability, Efficiency, Effectiveness." Medical Interface 3(5):13-16,40, May 1990.

[6] Kaplan, J. "Real Time Management Technology: Improving Accountability, Efficiency, and Effectiveness." Presentation at Group Health Institute, Los Angeles, Calif., June 12, 1990.

Jeffrey G. Kaplan, MD, MPS, is Executive Vice President for Medical Affairs, Health Services Medical Crop., Baldwinsville (Syracuse), New York. He is a member of the College's Forum on Computers and Information Technology and an associate member of its Society on Managed Health Care Organizations.