Measuring Patient-Based Outcomes
By Martha Bayliss, MSc, and John E. Ware Jr., PhD
Asthma’s serious and substantial influences are reaching more patients and causing more harm. In 1995, asthma was estimated to affect 14 million people in the United States, and its prevalence is expected to continue to increase worldwide.1,2
Meanwhile, the cost of health care continues to rise, compelling patients and payers to become more involved in medical decisions. Data documenting what works in patient care urgently is needed so they can make informed choices.3,4
As such, the goals of asthma treatment have extended beyond the traditional clinical markers of disease severity to include the long-term benefits of treatment and patient satisfaction. Physicians, patients, payers, employers and government must understand how different approaches to asthma management accomplish the “fundamental objectives of prolonging life, relieving distress, restoring function and preventing disability.”5
For clinicians, the ability to track asthma patients’ outcomes holds promise on many fronts: individualizing patient treatments, establishing optimum regimens and best practices, and negotiating contracts with managed care companies on the basis of demonstrated results.
A comprehensive understanding of asthma should combine a biomedical measurement model (etiology, pathology, biology and clinical outcomes) with a social science paradigm based on assessment of many dimensions of health-related quality of life (HQL).6,7 Underlying this measurement strategy is a theoretical model linking clinical variables with overall HQL. (See Figure 1.)
Until recently, clinicians did not rely on HQL surveys in their practices or research. The surveys’ length made it impractical for patients to complete them in most clinical settings, and clinicians couldn’t easily interpret their results.
In a busy practice setting, the length of the health survey must reflect the amount of time the patient has to complete the questionnaire. Results from experience and experiments suggest that an ideal survey for clinical use can be completed in 10 minutes or less.8,9 For the average patient, this means a survey of between 40 to 60 questions.
During the past several years, researchers have developed short-form surveys, both for general health and for asthma in particular. They examined the trade-offs involved in using fewer vs. more questions to assess each health concept. In general, while shorter surveys are more practical, user-friendly and economical to administer, they provide information that is not as comprehensive, accurate or reliable than that yielded by longer forms.
Administering standardized assessments via the Internet is one approach that achieves the goals of standardizing content and scoring of HQL surveys, while minimizing costs and burden to respondents and researchers alike.
Simple, reliable and inexpensive methods for measuring patients’ health-related quality of life have implications for clinical practice, research, health care reform and other objectives of medicine. Standardization of item content, response choices and administration guidelines is necessary to achieve reproducible results that can be interpreted and compared meaningfully.
While many features of HQL scales affect the usefulness of data gathered, four features should be considered prerequisites to their use:
* Reliability–This indicates how confident we can be that an observed score is the “true” score. It’s based on the concept of obtaining consistent results with repeated administration of the same questions. We’re confident in a score if we obtain the same score again and again upon repeated assessment of an unchanging patient, such as when repeat sphygmomanometric measurements are averaged to determine blood pressure.
* Range of measurement–Health states range from very poor, including disability and dysfunction, to very good, including high levels of behavioral functioning and well-being. The wider the range of health status captured by survey questions, the broader the range of measurement. Restrictions in the range represented across questions in a health survey will set limits on the ability of scores to make distinctions among patients and to detect changes over time. On any given scale, patients scoring at the “top” of the scale cannot show improvement, and those at the “bottom” of the scale cannot show decline.
* Number of levels measured–The number of levels or score values into which a HQL measure can classify patients also influences the scale’s ability to distinguish differences and changes over time. The more levels available, the greater the precision that a scale will have.
* Confidence intervals–As with all measurement tools, the interpretation of a score must take into account the amount of “noise” in the score. This noise level can be quantified and displayed visually as a confidence interval (CI) around a score. The size of the CI is a function of the reliability and standard deviation of the scale and the sample size. Because reliability affects the size of the confidence interval when individual scale scores are interpreted, measures used with individual patients require use of a higher standard of reliability in comparison with scores for large groups of patients.10
Measurement of both generic and disease-specific HQL in patient-based outcomes assessment provide complementary information. 2, 11,12,13 Specific measures often are more useful in determining whether treatments have their intended specific effects.
Specific measures that are valuable in asthma studies include the questionnaires developed by Juniper and colleagues,14 Marks and colleagues,15,16 the Living with Asthma questionnaire,17 St. George’s Respiratory Questionnaire,18 the ITG Asthma Short Form,19 and the comprehensive Asthma Outcomes Monitoring System (AOMS).20
Generic tools can be used to compare disease burden and treatment effects. Examples applicable to asthma outcomes assessment include the Nottingham Health Profile,21 the SF-36 Health Survey,22 the Quality of Well-Being Scale,23 and the Sickness Impact Profile.24
However, simply using these tools to measure patient populations will not influence changes in health status, cost or utilization. These generic and asthma-specific tools are meant to serve as components of well-designed efforts at tracking patient outcomes and evaluating the costs and benefits of therapies or interventions.
As important as the choice of an HQL tool is the design of the study or program within which it’s applied. Applications of HQL tools in asthma include:
* Population monitoring, to track trends in health levels, risk factors and utilization
* Clinical and epidemiological studies, to identify determinants or triggers of asthma, to study the course of disease, and to determine the efficacy of treatments
* Clinical practice, to select treatments and monitor outcomes
* Program evaluation and policy analysis, to set priorities, evaluate policies and programs, and allocate resources.25
Assessing outcomes in health care policy is a high priority because the U.S. health care system is being restructured to contain rising health-care expenditures. Those implementing cost-containment measures have given less attention to the health effects of these strategies than to the cost savings. Accountability demands that we consider patients’ point-of view when evaluating treatment programs and policies.
Dr. Bayliss is senior director of clinical applications at QualityMetric Inc. in Lincoln, R.I. Dr. Ware is president and CEO of QualityMetric Inc. and executive director of the Health Assessment Lab at the Health Institute, New England Medical Center, Boston. He is also research professor in the department of psychiatry at Tufts University School of Medicine and adjunct professor in the School of Public Health at Harvard University, both located in Boston.
For a list of references, please call Tracy Schmierer at (800) 355-5627, ext. 350.