et al., 1988). Only about one-quarter of high-risk persons have ever received pneumococcal vaccine.

Physician reminders have been shown to be successful in increasing rates of immunization; physicians who received computer-generated reminders vaccinated their eligible high-risk patients twice as often as they vaccinated patients in a control group (McDonald, 1992). A simple reminder sheet completed by the clinician and detailing vaccine eligibility, patient status, and reasons for refusal of the vaccine was successful in significantly increasing the rate of influenza immunization for high-risk outpatients (Merkel, 1994).

An increasing fraction of physicians’ offices in the United States are becoming computerized, and most physicians’ offices are now capable of creating an electronic database for the patients they treat. These systems were originally introduced largely to improve office administration and billing practices, and their role in enhancing the delivery of preventive services has not been well developed.

Such an approach has been tested and has been shown to enhance the delivery of influenza vaccine to elderly people (Bennett et al., 1994). From 1988 to 1991 internists and family practitioners in private practices in Monroe County, New York, participated in a series of demonstration studies to determine whether a target-based approach could increase the rate of influenza immunizations among elderly people. These studies indicate that the rate of immunization can be increased substantially after physicians are made aware of target groups within their practices and are given a simple means of monitoring their rate of coverage (Buffington et al., 1991).

One important logistic hurdle in the comprehensive delivery of influenza vaccine in the office setting is that this vaccine is given during a 3-month period each fall, and eligible patients may not be scheduled to see their physicians during that time. To achieve high immunization rates, physicians must develop initiatives to immunize all eligible patients, not just those who have a visit scheduled during the period when the influenza vaccine is being given.

Hospitals are also important sites for immunization. One of the most effective interventions appears to be the implementation of standing orders for vaccination. This consists of an institutional policy stating that everyone eligible for vaccination is to be vaccinated. Under this protocol, nurses can initiate immunizations without specific orders. Perhaps the best-documented multifactorial hospital-based interventions are those described by Nichol and coworkers (1990) at the Minneapolis Veterans Affairs Medical Center. Programs that ensure that hospitalized patients are immunized with influenza and pneumococcal vaccines are particularly important because two-thirds of the patients hospitalized for pneumococcal infections had been hospitalized within the previous 5 years, and 25% or more of elderly patients admitted for influenza-associated respiratory conditions had been discharged during the immunization season immediately preceding the outbreak period (Fedson, 1987).