Klatt, Timothy E. MD; Hopp, Elizabeth BS
Influenza causes between 3,300 and 49,000 deaths annually within the United States.1 When compared with similar-aged women, pregnant women who contract influenza have an increased risk of both needing medical care and requiring hospitalization. There was increased morbidity and mortality among pregnant women during the 1918–1919, 1957–1958, and 2009–2010 influenza pandemics. Severe infections among postpartum women, defined as those who delivered within the previous 2 weeks, were also observed in the 2009–2010 pandemic.2
In response to these risks, the American College of Obstetricians and Gynecologists recently stated that “preventing influenza during pregnancy is an essential element of prenatal care, and the most effective strategy for preventing influenza is annual immunization.”3 Vaccination against influenza is considered safe throughout pregnancy.
Maternal vaccination is also very important to the infant. Infants younger than 6 months of age who contract influenza have a high rate of hospitalization.4 Childhood deaths associated with influenza are most frequent in this age group.4,5 The vaccine is not licensed for use before 6 months of age and may be less effective in children younger than 2 years of age. The infant acquires protection through antibodies acquired both transplacentally and during breastfeeding. Maternal vaccination has been shown to reduce proven influenza illness in infants and to avert approximately one third of all febrile respiratory illnesses in young infants.4
Starting in 2004, both the American College of Obstetricians and Gynecologists and the Advisory Committee on Immunization Practices recommended that health care providers offer influenza vaccination to all pregnant women, in any trimester, during the influenza season. Despite these recommendations, most pregnant women (76% during the 2007–2008 influenza season) still go unvaccinated.2 Previous investigators reported that electronic reminders in their medical records have led to both nonsignificant and significant improvements in influenza vaccination rates among populations of children with asthma and patients with rheumatology receiving immunosuppressants, respectively.6,7 In an effort to increase our vaccination rate in the Medical College of Wisconsin Obstetrics and Gynecology Clinic, we implemented a “best-practice alert” within our electronic medical record.
MATERIALS AND METHODS
This study of the effectiveness of our quality improvement project was approved by the Medical College of Wisconsin/Froedtert Hospital institutional review board. We have used an electronic chart from Epic Systems, based in Verona, Wisconsin, in the Medical College of Wisconsin Obstetrics and Gynecology Clinic since 2005. Our clinic is a single site. During the 2007–2008 and 2008–2009 influenza seasons, eight faculty and nine resident obstetrician–gynecologists saw obstetric patients there. During the 2007–2008 season, two certified nurse midwives and three nurse practitioners also provided prenatal care there. One certified nurse midwife left before the 2008–2009 season. All of the health care providers' patients were included. At the beginning of both influenza seasons when health care providers were informed that the vaccine was available, they were also reminded of the importance of vaccinating every pregnant woman, without a contraindication, at her next prenatal care visit regardless of gestational age. We administered an inactivated, preservative-free vaccine during both influenza seasons. Vaccination charges, billing, and insurance coverage of vaccination were the same for both seasons.
Despite the vaccine shortage during the 2007–2008 season, we did not perform rationing and did not stagger the delivery of vaccinations. We operated with the understanding that our system would shift its stores to the highest risk populations, including pregnant women, and expected access to as many as we could use. Despite this, we received very few vaccinations after exhausting our initial supply. Centers for Disease Control and Prevention guidelines for the vaccination of pregnant women were the same for both influenza seasons.
Beginning on October 1, 2008, we included a best-practice alert, an electronic reminder, within our electronic prenatal record (Figs. 1 and 2). The best-practice alert alerted the health care provider during each prenatal visit if the medical record did not yet contain documentation of an influenza vaccination during the 2008–2009 influenza season. This alert, once satisfied, no longer appeared at subsequent visits. The alert could be satisfied by ordering the vaccine, documenting that a vaccine had been given elsewhere, or by selecting “declined” and providing a reason. If the health care provider simply selected “declined” without providing a reason, however, the best-practice alert would continue to fire at future visits until satisfied.
Health care provider and patient education materials were identical in both years. No other educational efforts or systemic changes were instituted during the time this study was conducted. A Nursing Informatics Specialist spent a total of 6 hours building and testing the best-practice alert and instructing health care providers in its use. The total cost to the Medical College of Wisconsin, including supplies, salary, and benefits, was approximately $321.
We then compared our 2008–2009 vaccination rate with our 2007–2008 rate. In 2007–2008, we began vaccination on October 1 and, because of a shortage of the vaccine, exhausted our supply of the influenza vaccine on the 31st business day. We therefore compared the influenza vaccination rates for the first 30 business days of the 2008–2009 and 2007–2008 influenza seasons. If a patient was not vaccinated, we reviewed her chart to determine whether a discussion regarding vaccination was held and, if so, the reason she went unvaccinated.
Our power calculation was based on a two-sample two-sided χ2 test. We assumed a baseline vaccination rate of 36.6% based on a review of deidentified clinic billing records from October 1, 2007, through January 31, 2008. One hundred sixteen patients from each influenza season were necessary, at an α level of 0.05, to provide 80% power to discern a 50% difference in vaccination rates, that is, a difference between a rate of 36.6% and 54.9%.
In both 2007–2008 and 2008–2009, approximately 640 unique women attended a prenatal visit. Between the 2007–2008 and 2008–2009 influenza seasons, only one health care provider changed within our clinic. One of our nurse midwives left before the start of the 2008–2009 influenza season. Analysis of her 2007–2008 performance showed that her rate of documented discussions and influenza vaccination did not differ from those of the remaining health care providers. Her patients were included within the study population. During the 2007–2008 influenza season, the number of vaccinations administered per day did not taper as we reached the end of our supply.
In 2008–2009, after implementation of the best-practice alert, more women were vaccinated both in our clinic and in total. The total figure includes vaccines administered at locations other than our clinic. Our 2008–2009 vaccination rate was significantly higher (P<.001 for all comparisons; Table 1). In 2007–2008 and 2008–2009, 9.4% and 10.9%, respectively, of the total number of vaccinations occurred outside of our clinic. Health care providers documented significantly more discussions regarding influenza vaccination in 2008–2009 compared with 2007–2008, 577 (89.5%) compared with 316 (49.5%), respectively (P<.001, 95% confidence interval for the difference in proportions 0.35–0.45).
The reasons why pregnant women went unvaccinated in 2007–2008 and 2008–2009 are shown in Table 2. In 2007–2008, most of the unvaccinated women had no discussion documented in their medical record. In 2008–2009, the most common reason for going without vaccination was an informed refusal of vaccination. In 2008–2009, 68.1% of the women whose charts documented a discussion accepted vaccination.
In 2008–2009, after activation of the best-practice alert within our electronic prenatal chart, we vaccinated 61% of our pregnant patients. This was a significant improvement over our relatively high 2007–2008 rate of 42%. For comparison, during the 2007–2008 influenza season, only 24% of pregnant women in the United States, without additional high-risk conditions, reported vaccination.2 Our improved rates were not the result of public, health care provider, or both awareness of the need for vaccination resulting from the recent H1N1 pandemic because our data collection was completed before its onset. In fact, both our baseline and improved rate exceeded that of 38% reported by the Centers for Disease Control and Prevention for October to December 2009, the height of the recent H1N1 epidemic.2 Use of the best-practice alert enabled us to reach the Healthy People 2010 objective of 60% vaccination of adults with high-risk conditions aged 18–49 years.8
The continuing poor rate of influenza vaccination of pregnant women in the United States is vexing. Numerous medical articles have reported the serious risks that influenza poses to pregnant women. Both the American College of Obstetrics and Gynecology and the Advisory Committee on Immunization Practices, among others, have stressed the importance of vaccinating pregnant women. Previous efforts to increase the rate of influenza vaccination among pregnant women have been reported. At the University of New Mexico, a standing order was implemented within a paper prenatal record. This enabled their obstetrics and gynecology clinic nurses to screen for and administer the influenza vaccine without first consulting with a health care provider. Their rate increased from 4% to 37%.5 At the Kelsey-Seybold multispecialty clinic in Houston, a number of strategies were implemented over six influenza seasons to increase their rate of vaccination of pregnant women. These strategies included assessing baseline immunization rates for obstetric health care providers, direct encouragement and behavior modeling, implementing standing orders for influenza vaccination, and offering vaccination training to obstetricians and nurses. They also simultaneously implemented additional programs to increase their rate of health care worker vaccination. Their rate of vaccination of pregnant women against influenza increased from 2.5% (2003–2004) to 37.4% (2008–2009).9 By implementing the best-practice alert, we were able to significantly improve on a baseline rate that exceeded both of these end points.
Implementation of the best-practice alert also resulted in a significantly higher rate of documented health care provider–patient discussions regarding vaccination during the 2008–2009 influenza season when compared with 2007–2008. This likely resulted from the best-practice alert's functioning as both a timely reminder and an option for documentation. A previous study reported that 71% of pregnant women who were offered the vaccine accepted it.10 Our 2008 acceptance-after-discussion rate, 68%, is consistent with this figure.
The 2007–2008 and 2008–2009 influenza seasons were similar with only baseline activity reported during October and November and peak activity occurring in early March.11,12 We are aware of no other new programs to increase the 2008–2009 influenza immunization rate among pregnant women either within the Froedtert Hospital and Medical College of Wisconsin community or within the greater Milwaukee community.
One concern about the use of best-practice alerts is the risk of alert fatigue. If health care providers encounter what they feel are too many alerts, they may become frustrated, overwhelmed, or desensitized and begin ignoring or overriding these alerts. Strategies for minimizing this risk include limiting the number of active alerts, only using alerts to flag very important issues, and instituting “hard stops.” Hard stops, unlike “soft stops,” cannot be ignored or overridden because they force the health care provider to address the alert before the health care provider can proceed. Our best-practice alert was the only active alert during the study period. It was also a hard stop. Although a hard stop guarantees a response, health care providers can find them bothersome. In an attempt to minimize their frustration, we approached our health care providers as a group, in the design phase, to capture their input on the structure of the best-practice alert they would be required to satisfy before they could close the chart.
Limitations of this study include that we studied a single site that was already performing well, compared with published data, before the intervention. Our health care providers accepted the hard stop best-practice alert quite well, possibly, in part, because they were convinced of the importance of vaccination against influenza. Because contemporary Centers for Disease Control and Prevention guidelines recommended vaccination of every pregnant woman without a contraindication, we did not collect demographic data related to the patients. These issues may limit the generalizability of our findings. Our health care providers requested that the best-practice alert disrupt their day as little as possible. Because of this, we did not collect detailed data specifying the reasons patients voiced for declining the vaccine. We, therefore, cannot recommend specific patient populations or concerns to target for future improvements. Another limitation is that we only studied the first 30 days of both influenza seasons. Our findings may not reflect the vaccination rates for the complete influenza seasons.
In both influenza seasons, relatively few women received their vaccine outside of our clinic. This clearly indicates that the increased vaccination rate was not the result of better ascertainment of vaccinations performed outside of our clinic. This also suggests that future interventions to increase the rate of influenza vaccination should be designed for settings in which prenatal care is delivered. We did not collect data regarding individual health care provider's 2008–2009 vaccination rates. Although nearly 90% of patient records included a documented discussion regarding vaccination, the discussion was not standardized. Future evaluation of both high and low performers may help develop recommendations for effectively conducting this discussion.
We recommend that all users of electronic medical records consider adding a best-practice alert or its equivalent to the records of their pregnant patients to improve their rates of offering influenza vaccination. Its benefits likely derived from its function as a timely reminder for our health care providers. During the course of their busy days, both health care providers and patients may simply forget to discuss the influenza vaccine. Use of a best-practice alert could be expanded to include all other high-risk groups. In fact, nearly every patient may benefit from a best-practice alert because current Centers for Disease Control and Prevention guidelines recommend annual vaccination of all individuals older than 6 months who do not have specific contraindications to the vaccine.2 This relatively simple, inexpensive intervention proved highly effective in our clinic.
1. Centers for Disease Control and Prevention (CDC). Estimates of deaths associated with seasonal influenza—United States, 1976–2007. MMWR Morb Mortal Wkly Rep 2010;27:1057–62.
2. Fiore AE, Uyeki TM, Broder K, Finelli L, Euler GL, Singleton JA, et al.. Prevention and control of seasonal influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices (ACIP), 2010. MMWR Recomm Rep 2010;59:1–62.
3. Influenza vaccination during pregnancy. ACOG Committee Opinion No. 468. American College of Obstetricians and Gynecologists. Obstet Gynecol 2010;116:1006–7.
4. Zaman K, Roy E, Arifeen SE, Rahman M, Raqib R, Wilson E, et al.. Effectiveness of maternal influenza immunization in mothers and infants. N Engl J Med 2008;359:1555–64.
5. Ogburn T, Espey EL, Contreras V, Arroyo P. Impact of clinic interventions on the rate of influenza vaccination in pregnant women. J Reprod Med 2007;52:753–6.
6. Fiks AG, Hunter KF, Localio AR, Grundmeier RW, Bryant-Stephens T, Luberti AA, et al.. Impact of electronic health record-based alerts on influenza vaccination for children with asthma. Pediatrics 2009;124:159–69.
7. Ledwich LJ, Harrington TM, Ayoub WT, Sartorius JA, Newman ED. Improved influenza and pneumococcal vaccination in rheumatology patients taking immunosuppressants using an electronic health record best practice alert. Arthritis Rheum 2009;61:1505–10.
8. Centers for Disease Control and Prevention. Immunization and infectious diseases. In: U.S. Department of Health and Human Services. Healthy people 2010: volume 1. 2nd ed. Washington, DC: U.S. Government Printing Office; 2000. p. 14-1–49.
9. Mouzoon ME, Munoz FM, Greisinger AJ, Brehm BJ, Wehmanen OA, Smith FA, et al.. Improving influenza immunization in pregnant women and healthcare workers. Am J Manag Care 2010;16:209–16.
10. Yeager DP, Toy EC, Baker B 3rd. Influenza vaccination in pregnancy. Am J Perinatol 1999;16:283–6.
© 2012 by The American College of Obstetricians and Gynecologists.