In our health system during the 2009-2010 influenza season, we found rates of vaccination of pregnant women of 54% for the seasonal influenza vaccine and 51% for the H1N1 vaccine. When restricted in a manner similar to a 10-state sample [19
] (births through March 12, 2010) our rates were approximately 10% higher: seasonal vaccine coverage was 61% versus 51% and H1N1 coverage was 59% versus 47%. This is notable since our population serves pregnant women who have been shown in previous studies to be less likely to receive the vaccine, those who are disadvantaged, that is, uninsured, nonwhite, less educated, and of low income [20
Other studies have reported on vaccination coverage of pregnant women in the 2009-2010 pandemic influenza season, many of which rely on self-report. In the United States, a nationally representative phone survey found coverage rates of 32% for the seasonal vaccine and 46% for the H1N1 vaccine [23
]. Another nationally representative phone survey found similar results, 42% for the H1N1 vaccine [24
]. Higher rates have been reported from samples in health care institutions. A study at a university hospital in Denver, CO, found higher coverage rates of 64% for the seasonal vaccine and 54% for the H1N1 vaccine [25
]. Unlike our study, the Denver study relied on self-report and demonstrated high rates of vaccination outside their health system, 15% for the seasonal vaccine and 29% for the H1N1 vaccine. We do not know the magnitude of the outside vaccination in our population. A high rate of H1N1 vaccine coverage (76%) was achieved in a public hospital prenatal clinic in Seattle, WA, using a system which included chart prompts and a vaccine registry [26
]. The highest rates were reported at the Massachusetts General Hospital with coverage rates of 88% and 86% for the seasonal and the H1N1 vaccines, respectively [27
]. Their sample was restricted to 3 months of the influenza season (January 2010 to March 2010) which could have raised these estimates as compared to other studies, since both vaccines were more readily available at that time. A time difference might in part explain the lower coverage rates of the H1N1 vaccine (26%) reported in one health system, Maimonides Medical Center in Brooklyn, NY, as the sample began shortly after the H1N1 vaccine was available (time frame of November 12 to December 19, 2009) [28
]. Coverage rates outside the United States have also been reported. In Canada, rates of seasonal coverage were reported to be 30% in Alberta [29
], while for H1N1 the rates have been reported to be between 63% and 76% [29
]. In France, rates of the H1N1 coverage of pregnant women were lower: 38% in a 3-hospital study in Paris [31
] and 22% using a national insurance database [32
]. Another insurance database from Australia reported even lower estimated coverage rates of 10% for the H1N1 vaccine in pregnant women from October 2009 to January 2010 [33
With one exception, all of the rates are lower than the 80% coverage rates targeted by the national Healthy People 2020 objectives for pregnant women [34
]. However, most represent large increases over the prior year rates [7
], which appear to be sustained with 49% pregnant women reporting being vaccinated during the 2010-2011 influenza season [20
]. In our health system we achieved rates higher than nationally reported rates; we attribute this to a comprehensive approach to influenza. For example, we undertook many of the prevention strategies cited by Goldfarb et al. [27
] in their successful program. Our actions included standing orders [35
] and awareness activities, [36
] both of which have been demonstrated to increase vaccination rates. Reasons for our rates being lower than the highest reported rates may have included not having chart prompts [37
] or not making use of a vaccine registry [26
]. Furthermore, our system of prenatal care includes clinics spread over a large geographic area, making messaging and implementing programs challenging. However, streamlining the order process as we have been able to do with the Tdap vaccine suggests that even higher rates can be achieved.
Our study has a number of strengths. It was based on chart review which indicates that we can be relatively certain of the lower limit of our vaccination rates. It is possible that actual coverage rates may be higher, especially in view of the high rates of vaccination outside of the health system reported by Fisher et al. [25
]. However, not knowing the magnitude of the outside coverage in our population, we would not want to speculate on how much higher the “real” coverage rates would have been. Another advantage of the chart review methodology is that many of the studies rely on self-report, which may have been an issue in the 2009 to 2010 season when 2 vaccines were available. Our own experience with our population suggests this may have occurred, with a frequent response of, “I already received the flu vaccine,” when they had received only one of the recommended two vaccines. Furthermore, many of the studies relying on self-report had substantial nonresponse rates of approximately 50%; this included studies with high reported coverage rates or those that were nationally representative [23
]. A limitation of our study is that it represents the cumulative efforts of one health system caring for the underserved, as such these results may not be generalizable to other populations or health systems. Finally, we do not know which of our interventions impacted vaccination rates the most.
Our study has a number of implications. For other health systems it would appear that improved vaccination rates as compared to the national rates are possible, even for health systems that are safety-net systems that provide medical care to large numbers of underserved patients. Our success with a population which usually experiences lower coverage rates—minorities, uninsured, and with limited English-speaking ability—could be a model for other systems. Taken with Goldfarb's study [27
], we also believe a comprehensive approach might be a rational strategy to achieve high coverage rates. Given initial reports of continued high rates in pregnant patients over the subsequent influenza season to this study, it will be interesting to see if high vaccination rates continue in this population. Finally, given that self-report might overestimate and chart review might underestimate coverage rates, it would be interesting to supplement chart review with self-report to further characterize this in our population. Moreover, self-report in such a study would assist in determining the upper limit achievable by segregating those who represent missed opportunities versus those who refuse the vaccine.
In summary, our public safety-net institution achieved influenza vaccine coverage rates, that were higher than those reported national samples by instituting a comprehensive strategy. Our rates however, were not as high as those in reported in “best practice” studies. This could be due to not capturing vaccinations done outside our health system, not including certain interventions in our vaccination strategy such as chart prompts, and not including an order for influenza vaccination on all postpartum orders.