"Measuring the Outcome of Care - Are Patients Really Better?"

Michael J. Goldberg, MD

The clinicians' goal is to manage their patients' care in a way that achieves the desired outcomes while not producing any adverse consequences. In order to accomplish this, it is important to have outcome studies. Outcome studies are of four types: technical, functional health status, patient satisfaction, and resource utilization. Technical outcomes assess physical parameters that are measurable with tools such as radiographs, blood tests and goniometers. A change in range of motion of a joint, or in the magnitude of scoliosis are examples. This is quite different from functional health assessment which evaluates the tasks that people perform, their roles, and their quality of life. Patient satisfaction is the patient's opinion of the processes of care and the consequences of care. Most patient satisfaction studies emphasize the process of care and measure such things as access to the doctor, parking and other amenities. Resource utilization deals with utilization of services and utilization of dollars.

That which is measured gets interest and investment. Thus, there is great interest at this time in the cost of care and cost effectiveness. It is important to keep in mind that cost-effectiveness analysis may set health care priorities in a way that violates peoples' values. (Ubel, PA et al. Cost Effectiveness Analysis and the Setting of Budget Constraints. New England Journal of

Medicine 334:1174-77,1996.) Not all benefits are measurable. Indeed, the measurable may drive out the important. Nevertheless, the measurement tool may determine the results. Callahan (Callahan, JJ. Assessing the Results of Hip Replacement. A Comparison of Five Different Rating Systems. Journal of Bone and Joint Surgery 72-B: 1008-1009, 1990) show that the same patients had variable results following total hip replacement depending on what hip assessment instrument was used to evaluate their outcomes. In general, questionnaires constructed by or administered by those who have a vested interest in the success of the treatment are suspect.

There are two types of functional health outcomes assessment instruments. One is a "generic health measure", and the other is a health status. The best-known is the SF36. It is very good for comparing across conditions. However, it is population-based and it is not causal. It measures the functional health of the patient at a point in time, but improvement in the patient's status can not be attributed to a particular medical intervention and may be related to other social circumstances. Condition or disease-specific measures, on the other hand, are specific for a condition. They are able to measure small changes, but one may need many. The SF36 is suitable for adults only. It measures the physical domains of physical function, physical role, and bodily pain, and the mental health

domains of general health perceptions, vitality, social functioning, and emotional role. In pediatrics however, in addition to mental and physical domains, one needs to measure the concepts of family functioning and impact of disease on family.

Questionnaires filled out by a proxy have limitations. Proxies include physicians, other clinicians such as physical therapists, as well as parents. Patients and their parents report functional performance differently.

Nevertheless, patient-based assessment instruments are reliable and as sensitive as traditional measures. They measure functional status and they measure things that patients feel are important.

Self-administered tests have the advantage of ease of administration and low cost; but patients may not answer all the questions because of language or educational difficulties and tend not to answer questions about economic or sexual matters. Patient- based instruments are not substitutes for traditional physical exams.

A good functional health instrument must be reproducible (reliable). When the patient is stable, it measures the same results. It also needs to be valid (specific). It has to measure what we want it to do. Lastly, it needs to be responsive (sensitive). When a patient changes clinically, the instrument

detects it. Generic instruments have been used in clinical settings for specific diseases. (ET Shapiro, JC Richmond, S Rockett. The use of a Generic Patient-Based Health Assessment (SF 36) for Evaluation of Patients with Anterior Cruciate Ligament Injuries. American Journal of Sports Medicine 24 (#2) : 196-200, 1996). Theoretically functional health instruments might be used to reduce unnecessary referrals and reduce unnecessary surgery. Most importantly, functional health data allows patients to make good decisions. Basing surgery on technical parameters only may yield neither improved functional status nor patient satisfaction.

There is a difference between outcomes research and outcomes assessment. Outcomes research is hypothesis-based, prospective and statistically significant, whereas outcomes assessment is often informal and practice-based. Outcomes assessment in some ways, is developing a thermometer that one might insert into a practice to see how they are doing compared to other groups managing the same disease. Outcomes research is basic science.

Managing patients from an outcomes perspective involves three steps: There is outcomes measurement where we assess the patient at a point in time. Then outcomes monitoring where we assess patients over a period of time. Finally, outcomes management where we change treatment to effect the outcomes. Physicians and other clinicians often find participating in outcomes and sharing data both stressful and threatening. It is vitally important however, for all clinicians to participate.

Address correspondence to:

Michael J. Goldberg, MD New England Medical Centre

The following was a summary of the Hector Kay Memorial Lecture from the recent ACPOC meeting in Atlanta