country. Studies that evaluate underutilization of health care tend to be descriptive surveys of utilization patterns and unmet needs. These studies can provide data for assessing health service needs and for identifying areas of maldistribution or inequity in delivery of care for the elderly population (Rubenstein et al., 1989).
In medical care, underuse of services not related to direct access barriers may be classified as either underdiagnosis or undertreatment. Underdiagnosis (or lack of case finding) in the elderly has been studied for such conditions as depression, substance abuse, urinary incontinence, and confusional states. It may be attributed in some part to practitioners simply not identifying with medical problems they have not experienced or not treating certain categories of patients (the poor, women, racial and ethnic minorities) as thoroughly as others. Undertreatment includes, for example, lack of timely and appropriately vigorous medical therapies, follow-up, adequate nursing care, discharge planning, and home health visits. Malpractice suits frequently allege underdiagnosis and undertreatment (e.g., missed diagnosis, or lack of follow-up of abnormal x-ray or test).
Often underuse is only inferred. For instance, descriptive studies, surveys of utilization patterns, or controlled trials that demonstrate improved outcomes from a service that is not generally provided may support inferences about underuse. In addition, care may be found insufficient when individuals are hospitalized for complications of conditions that can (and should) be successfully managed in the outpatient setting or long-term-care facility or when they are hospitalized for the first time at excessively advanced stages of disease. Furthermore, care may be insufficient when patients receive services from primary care physicians that would be more appropriately given by specialists.
In cases discussed above, generalizing to the entire population of elderly might provide an estimate of the “room for improvement” in the overall level of care (often by improving access) rather than in the care provided by organizations or specific providers. However, where providers have an incentive to conserve resources, evidence of possible underprovision of services may need to be sought directly.
Methods to detect underuse related to access barriers are not the same as those used to detect underuse related to underdiagnosis or undertreatment. To measure underuse of services because of access barriers, population data are needed (e.g., all persons eligible for care, not just those using care). That is, when compared to population norms or to rates for other subgroups, rates of use in one group may suggest underuse. Conversely, underuse for patients already receiving services (underdiagnosis or undertreatment) may