Moral hazard in influenza vaccine receipt patterns among the Medicare elderly

Research Objective:

Influenza vaccination coverage level among the non-institutionalized elderly appeared to reach a “plateau” of about 70% during the late 1990s. To achieve the Healthy People 2010 goal of 90% vaccination coverage among the elderly population, a new approach is needed to explore individuals' incentive to seek vaccination, e.g., moral hazard. Moral hazard is exemplified by postponing or missing vaccination after observing a lower epidemic level. A past study indicated the presence of such moral hazard during consecutive influenza seasons, i.e., individuals became less likely to seek influenza vaccination if the immediate past season's epidemic level was lower. Our study hypothesized that such moral hazard in vaccination behavior was observed even during one influenza season.

Methods: Cross-sectional survival analyses during each of six influenza seasons from 1999-2000 to 2004-05 seasons, using the Medicare Current Beneficiary Survey (MCBS) and its associated claims data. The study population was the community-dwelling Medicare elderly. The outcome variable was weekly vaccine receipt. The influenza epidemic level was measured by the weekly mortality rate due to pneumonia and influenza at the nine census region level. Our main analysis's key independent variable was the change rate in this weekly mortality rate, between the immediate past two weeks. Other covariates included the weekly vaccine supply at the national level and individual factors such as health status and socio-economic characteristics.

Principal Findings:

Out of our preliminary analyses on six influenza seasons, results of five seasons implied the presence of moral hazard (p<.05). For instance, during the 2004-05 season, an individual was 35% (p<.001) less likely to receive an influenza vaccine during a week when the influenza epidemic level decreased by 10% in the past two weeks. On the other hand, an individual was at most 58% (p<.001) more likely to be vaccinated subsequent to the weeks experiencing the epidemic increase. The magnitude of such moral hazard tended to be greater during a season with a longer period, e.g., the 2001-02 season, or with the delayed vaccine supply like the 2004-05 season.

Conclusions:

During one flu season, individuals' vaccination patterns were strongly associated with the changes in the ongoing epidemic levels. When vaccination incentives declined due to moral hazard, it would be difficult to achieve herd immunity. As a result, the secondary upsurge of influenza epidemic could occur afterward during the same season. Therefore, the constant efforts to vaccinate throughout an entire influenza season would be needed.