Risky business: challenges in vaccine risk communication.

Abstract

Concern over vaccine safety has been a major determinant of immunization policy—from passage of the Biologics Control Act in 1902 to the recent American Academy of Pediatrics (AAP) and Advisory Committee on Immunization Practices (ACIP) recommendations for expanded use of inactivated polio (IPV) and acellular pertussis vaccines.1-4 With the availability of new vaccines and vaccination options comes the challenge of how best to communicate these options to patients and their parents. Lessons from the field of risk communication on predictors of risk acceptability and vaccination decision heuristics provide insight into why some parents resist vaccination. This article provides a historical perspective on vaccine adverse events and applies lessons from risk communication research to help physicians improve their ability to discuss vaccine risks.Immunizations have been described as the single most effective health intervention after clean water and sewage disposal,5 and have an extraordinary safety record. However, public policy has long reflected concern for the safety of vaccines and related biologics. In 1902, Congress passed the Biologics Control Act in response to the death of 13 children who had received injections of diphtheria antitoxin contaminated with tetanus toxin.6 This act required biologics to be manufactured in a manner that assured their safety, purity, and potency. Responsibility for these regulations was assigned originally to the Hygienic Laboratory of the Public Health Service, which evolved into the National Institutes of Health. The regulation of biologics including vaccines was transferred to the Food and Drug Administration (FDA) in 1972.The first published accounts of serious adverse events after whole cell pertussis vaccine occurred in 1933 with Madsen's7 report of two deaths within 48 hours of immunization, and in 1948 with the report in Pediatrics by Byers and Moll8 of encephalopathy after diphtheria and tetanus toxoids combined with whole-cell pertussis vaccine (DTwP). In 1955 the Cutter incident occurred, in which incompletely inactivated Salk vaccine was associated with over 200 cases of paralytic poliomyelitis.9 By the 1970s the incidence of many vaccine-preventable diseases had decreased significantly and concern over safety of whole cell pertussis vaccine surfaced in Japan and Europe.10 Pertussis immunization rates plummeted in several countries. Outbreaks of pertussis followed, leading to the development of an acellular pertussis vaccine (DTaP) in Japan.11The pertussis debate raging in Europe and Japan reached the United States in 1982, when the Emmy Award-winning television program,DPT: Vaccine Roulette, was aired with allegations of serious neurologic sequelae after DTwP.12 This was followed by the publication of Coulter and Fisher's book, DPT: A Shot in the Dark, in 1986.13 A dramatic increase in diphtheria, tetanus, and pertussis vaccine (DTP) litigation led to the withdrawal of several manufacturers from the marketplace. The Centers for Disease Control and Prevention (CDC) began stockpiling DTP vaccine as a shortage loomed. With the nation's health at stake, the National Childhood Vaccine Injury Act of 1986 (NCVIA) was passed,14after a collaborative effort involving the AAP, consumer advocacy groups, and others. This act included several provisions relevant to vaccine safety, creating the National Vaccine Injury Compensation Program (VICP) and the Vaccine Adverse Event Reporting System and mandating comprehensive reviews of vaccine-related adverse events by the Institute of Medicine (IOM) and the development of vaccine information materials.The expanded role of DTaP and IPV vaccines in childhood immunizations is in direct response to vaccine safety considerations. Concerns about local and systemic reactions prompted the recent revisions to the ACIP and AAP acellular pertussis vaccine recommendations, despite the lack of data establishing conclusively a causal link between DTwP and serious sequelae.15 After over a decade of research to develop less reactogenic pertussis vaccines for infants, the FDA licensed acellular pertussis vaccines for the fourth and fifth doses in 1992 and for the primary series in 1996.16 DTaP is now the preferred product, with DTwP an acceptable alternative during the transition period.24Since 1980, an average of 8 to 9 cases of paralytic polio per year have been associated with the live, trivalent oral polio vaccine (OPV) in the United States, either through direct receipt or contact with a recipient.3 In 1994, the Western Hemisphere was certified by the World Health Organization to be free of indigenous wild polio.17 The greatly diminished risk of acquired paralytic polio from wild poliovirus, either indigenous or imported, prompted a reevaluation of the risks of OPV-associated paralytic polio. After much debate, the AAP and ACIP have moved toward expanded use of IPV.13 With the availability of new options comes the challenge of how to communicate these options to our patients. Moreover, pediatricians are occasionally faced with parents who question the need for vaccinations altogether.Why is vaccine risk communication so challenging? Perhaps the most important factor may be the lack of disease awareness. The dramatic decline of vaccine-preventable diseases has inevitably decreased public awareness of these illnesses, likely prompting greater reluctance to accept adverse reactions after vaccination. Another factor is the power of temporal association—ie, post hoc, ergo propter hoc—or what follows immunization must be caused by it. Neurologic sequelae after DTwP are one example of this logic. Although paralytic polio is a demonstrated risk after vaccination with OPV, many serious adverse events temporally associated with immunization lack a clear cause-and-effect relationship. In addition, vaccine risk communication is hampered by a lack of data. In 199118 and 199419 the IOM undertook extensive reviews of adverse events associated with childhood vaccines, and concluded that there was either no evidence or insufficient evidence to establish a causal relationship for two thirds of the conditions it studied. Moreover, experts often disagree about the interpretation of existing data, further confusing a public looking to science for answers. The rarity of an adverse event, the lack of defined clinical syndromes, and the absence of pathophysiological understanding limit investigation into adverse events. For example, no specific clinical syndrome or neuropathology has been defined for DTwP encephalopathy, despite more than five decades of DTwP use.Underlying this issue is the inherent tension between protecting public health and allowing individual autonomy. Some consumer advocacy groups formed from the DTwP controversy of the 1980s maintain the importance of parental choice in vaccinations.13 However, the enactment of state laws mandating immunizations for school entry, in large part responsible for wide vaccine coverage and the dramatic drop in vaccine-preventable diseases, is in direct collision with free parental choice. Some health care practitioners also support the concept of choice in vaccinations. For example, a survey of American chiropractors found that 81% of respondents believe immunizations should be voluntary.20 Currently, most states permit religious exemptions to vaccination requirements and 15 states permit philosophical exemptions (Joel Kuritsky, CDC, National Immunization Program, unpublished data). Reconciling patient empowerment with the goals of public health remains problematic.The media and other sources of public information play a role in vaccine risk perception.21 In 1994, media reports incorrectly attributed deafness in Heather Whitestone, the former Miss America, to the DTwP vaccine, although her condition was later confirmed by her pediatrician to be the result of Haemophilus influenzae type b meningitis.22 The irony that her condition was caused by an illness now preventable by vaccination may have been lost on the American public because of the media fanfare surrounding the initial report. This episode underscores the fact that health professionals are not the only source of vaccine information, which can come from family members, neighbors, and an array of media outlets such as newspapers, magazines, and television. More recently the Internet, with its home pages and electronic bulletin boards, has emerged with vast potential for information dissemination but without any editorial control, much less peer review.23 Even when presented with accurate information, parents and physicians may differ in how they interpret data and make decisions on risk. What then are the determinants of risk perception, and what risk communication approaches are available to the pediatrician?By and large, pediatricians are viewed as a credible source of vaccine recommendations, a notion supported by a recent study of private pediatric practices that indicated provider behavior may be the most important determinant of immunization rates.24The majority of parents follow their pediatrician's recommendation regarding immunizations, and may not engage in an independent decision-making process. However, a minority of parents question vaccination recommendations. For them, short explanations of risks and benefits may not suffice.Over the past two decades the field of risk communication has developed from the need to find more effective ways to communicate health risks. Researchers in this field recognized the discrepancy between how scientists explained health risks and what the public believed. Risk communication research has drawn from diverse disciplines such as cognitive and social psychology, behavioral decision theory, and risk assessment and management, and has concentrated on environmental hazards including nuclear reactors,25 chemical plants,26 and radon.27 A broad review of risk communication was conducted by the National Research Council in 1989.28 Theories developed in this field have only recently been applied to vaccine risk communication.29Of paramount importance is that individuals perceive risk differently. Although physicians may focus on the statistics regarding general vaccine effectiveness and known risks of vaccine-preventable diseases, parents making vaccination decisions may perceive risks in a broader religious, cultural, and personal context. Individual characteristics affect decisions to vaccinate; data from the CDC tell us that immunization rates vary by race, education, socioeconomic status, and other factors.30 For example, mistrust of the medical system by some African-Americans has been identified as a barrier to optimal health care and participation in clinical trials,31and was highlighted by the Clinton Administration's recent decision to issue a formal apology for the Tuskegee Syphilis Study.32Confronting reasons for undervaccination such as lack of access and missed opportunities may be easier than addressing issues of trust and cultural perspective.Other attributes of risk also affect risk acceptability and may complicate parental risk decisions.33 Voluntary, controllable risks are more acceptable than involuntary risks.29 Consumer groups advocating greater parental choice in vaccination contend that state laws mandating vaccination for school entry render vaccination an involuntary risk, and thus less acceptable.29 Researchers have uncovered another way in which the perception of control influences parental vaccination decisions. Meszaros et al34 demonstrated that parents refusing vaccination were more likely than parents who vaccinated their children to agree with the statement, “[I]f there were no vaccination given to my child, I could prevent the disease.” Thus, some parents may not accept vaccination recommendations because they perceive control over events in ways not recognized by the pediatrician.Additional factors may influence the acceptability of risks. Some risk communication researchers suggest that natural risks are generally more acceptable than man-made risks.35 At an IOM forum on polio vaccine policy, an advocate for alternative therapy argued that immunity acquired after natural infection is often preferable to vaccine-induced.36 This logic neglects the often severe morbidity and mortality associated with vaccine-preventable diseases, as well as the fact that some vaccine-preventable diseases induce no significant immunity (eg, tetanus). In addition, the acceptability of a risk is determined by whether it induces fear or dread and whether it is memorable. Any event adversely affecting the well-being of a child is dreaded, and risks that are memorable are more aversive than those that are not. For example, parents' recall of the fever and fussiness after DTwP in their infant may influence their acceptance of further vaccinations.Finally, risk communication researchers have demonstrated that some parents are unlikely to undertake a risk control measure unless they perceive both a serious threat and some control over the risk.37 When a perceived threat is low, individuals are unlikely to accept a health intervention regardless of the efficacy of that intervention. Thus, a decision to vaccinate is most likely to be made by a parent who recognizes the threat of vaccine-preventable illness and perceives vaccination as an appropriate resource to control that risk.How risks are perceived depends, in part, on how the message isframed. Studies have shown that an option described in terms of benefits may result in a different decision than the equivalent option described in terms of risks.3839 Applying this to vaccines, risk communication that emphasizes lives saved by vaccination may be more effective in promoting vaccination than communication that mentions the lives lost despite vaccination. In addition, the extent to which an individual perceives a health intervention as risky may determine how a message is processed and what action is taken.38 Research in this area suggests that for parents who view vaccines as generally safe, emphasizing the benefits of vaccines may be the most successful approach to encourage vaccination. For those parents who question vaccine safety, a more effective approach may be to frame the decision in terms of the risks of illnesses preventable by vaccination.Because most risk decisions involve more information than can be processed readily, individuals tend to simplify. An intervention such as immunization is often categorized as safe or unsafe without acknowledging the spectrum in between. Heuristics are cognitive shortcuts that people use to simplify complex decisions and judgments. Parents may use heuristics to quantify risk. In some cases, heuristics may result in a decision not to vaccinate.The heuristics individuals use to make risky decisions includecompression, or overestimating the frequency of rare risks and underestimating the frequency of common risks.40Concerns about DTP encephalopathy and vaccine-associated paralytic polio, which are rare, may be increased by use of this heuristic. Moreover, the availability of an event (one that is accessible or easily remembered) can lead to overestimation of its frequency; witness the effectiveness of sensationalized media reports alleging vaccine injury.Several studies have elucidated an omission bias, which explains the reluctance for certain parents to vaccinate their children.344142 That is, they withhold vaccinations because of the perception that actions (commissions) are more harmful than inactions (omissions). A study by Ritov and Baron41using a hypothetical vaccine scenario found that some individuals would feel more responsible if their child died after a vaccination than after a vaccine-preventable disease. This study also demonstrated the preference of some individuals to eliminate risk, with 23% of individuals in one experiment indicating that they would vaccinate their child only if the risk of vaccination was zero. Moreover, when given a choice, individuals tend to avoid ambiguity.43 For example, a risk from a known disease may be more acceptable than an equivalent or smaller risk that is perceived as more ambiguous (eg, from a new vaccine). This heuristic may be operative in some parents' avoidance of the varicella vaccine. Some parents who withhold vaccinations may use freeloadinglogic, relying on high vaccination rates and herd immunity to protect their unvaccinated child.29Some heuristics may work in favor of vaccinations.Bandwagoning is the tendency for parents to vaccinate if “everyone else is doing it,” without fully evaluating the options themselves.34 Altruistic individuals are willing to accept personal risk if society as a whole will benefit (ie, herd immunity).Effective risk communication has four components. First, it communicates existing knowledge, taking into account what individuals already know. Second, successful risk communication recognizes factors influencing parental risk perception and addresses vaccine decision heuristics. Third, it acknowledges potential risk communication pitfalls. Finally, effective risk communication engages parents appropriately, which for active, concerned parents means a decision-making partnership with their physician.Because belief systems are difficult to change once established, it is appropriate to provide parents with understandable, reliable information on vaccines at the outset so their first impression is correct. What vaccine risk communication tools exist today? The NCVIA mandated that vaccine administrators provide parents with written information materials detailing the benefits and risks of childhood vaccinations. These initially took the form of Vaccine Information Pamphlets, developed by the CDC and implemented in April 1992. The AAP, among others, argued that these pamphlets were unnecessarily long and confusing to parents,44 and an amendment to the NCVIA led to the shorter, simplified Vaccine Information Statements (VIS) adopted in October 1994. Vaccine administrators are required by law to provide VIS for vaccines covered by the VICP. Of note, the VIS should not be construed as informed consent, as informed consent requirements are determined by state law. State laws dictate whether informed consent is required before vaccination, whether informed consent must be oral or in writing, and whether additional vaccine information must be provided (Kevin Malone, CDC, Office of General Counsel; personal communication). Although the VIS have been hailed as an improvement, they have been criticized for requiring too high a reading level for some parents, and not providing enough information for others. Clearly, one size does not fit all.To overcome these shortcomings, experts in risk communication recommend layering information according to individual needs. An option for future vaccine information materials includes providing multiple levels of information that can be targeted to various audiences. Most importantly, the VIS and other vaccine information materials cannot replace the dialogue between pediatrician and parent but can be viewed as a starting point.If a pediatrician does not possess the knowledge to answer a parent's particular question, referral to appropriate resources should be made. Technical information on vaccines and vaccine-preventable diseases can be found in the AAP Committee on Infectious Diseases' Red Book, the ACIP's Reports and Recommendations of the Morbidity and Mortality Weekly Report, and the manufacturers' package inserts. The CDC has developed a pamphlet directed at providers for addressing common misconceptions on vaccines such as the existence of DTP hot lots or that the immune system can be overloaded by simultaneous immunization with multiple vaccines.45 In addition, information on vaccines can be found on CDC's National Immunization Program Internet home page athttp://www.cdc.gov/nip/home.htm.Second, effective risk communication acknowledges factors influencing risk perception and risk acceptability, and addresses heuristics. Parents expressing a reluctance to vaccinate their children require more than a quantitative analysis of the risks and benefits; physicians should seek to understand the cultural, religious, and other personal factors influencing vaccination decisions. Vaccination decisions are likely to be improved if physicians educate parents about the risks of diseases and encourage parents to participate in and control vaccination decisions. Message framing can influence parental vaccination decisions, but as discussed above, its effects are likely to depend on parents' prior beliefs and perceptions of risks and benefits.Successful risk communication also addresses vaccine decision heuristics. Providing parents with accurate risk data is the first step toward correcting compression and other biases in estimating risks. One approach suggested by risk communication researchers to overcome the omission bias is to reframe the vaccination decision from the viewpoint of the child. Baron46 found that individuals opposed to vaccination could be persuaded to vaccinate if they placed themselves in the child's position and then asked themselves whether they preferred a greater or lesser chance of death, and whether it mattered if the outcome occurred as a result of someone's act or omission. Parents can be educated on the limits of freeloading by discussing the incidence of vaccine-preventable diseases in unvaccinated populations. Because risks which are easily accessible to the imagination are more compelling, examples given in the context of a personal story can be persuasive. For example, parents reluctant to vaccinate against pertussis can be told of pediatrician's personal experience treating children hospitalized with pertussis.In communicating risk, there are potential pitfalls. Physicians must recognize their own use of heuristics, as professional training does not preclude biases and errors in judgment.4748Researchers describe the tendency for experts to extrapolate from limited data and fit equivocal data into preconceived patterns.49 In addition, some providers withhold vaccine risk information for fear of sending a mixed message to parents. However, pediatricians are a trusted source of information and are well-situated to address safety concerns that parents are likely to pick up elsewhere. Other potential pitfalls include the use of risk comparisons, comparing, for example, the risk of a serious vaccine reaction with the risk of being struck by lightning. Risk communication researchers argue that risk comparisons often backfire39; these comparisons may be confusing and may appear to belittle a parent's concerns. Finally, some physicians adopt the paternalistic “Me Doctor Me God”50 approach to dictate vaccine choices to their patients. In addition to the stylistic deficiencies of this approach, one study demonstrated a higher litigation rate in primary care physicians who deemphasized patient education and provider-patient communication.51Finally, effective risk communication strategies entail a decision-making partnership between the pediatrician and parent.28 Although some parents may want to follow their pediatrician's advice, others may want to make their own decisions, using information provided by their pediatrician. Recent AAP and ACIP polio vaccine guidelines recommend informing parents or caregivers of the three acceptable polio vaccine regimens, allowing greater latitude for parent-provider choice. Increasingly, parents are concerned with process as well as outcome, and they want to be part of the decision-making process. At the heart of this partnership is the understanding by parents that they retain some control, as well as ultimate responsibility, for their child's health.Ultimately, whose responsibility is risk communication? Public health clinics, which administer >50% of childhood immunizations in the United States, are often not structured to handle such discussions. Vaccine manufacturers develop package inserts and promotional brochures but lack the means for engaging in a direct dialogue with the public. Pediatricians, while ideally situated to communicate the benefits and risks of vaccinations, are faced with ever-increasing demands on their time, and lengthy discussions about risk are unlikely to be covered by capitation payments or third-party payers. In the final analysis, however, the responsibility for vaccine risk communication rests on the pediatrician or primary care provider who recommends vaccination.Additional challenges remain. Licensure of new vaccine products will further crowd an already full immunization schedule and may complicate risk communication efforts. Innovative technologies in the pipeline such as DNA vaccines may streamline vaccine combinations and diminish the need for booster doses, but present the theoretical safety questions of integration into host genome and the production of anti-DNA antibodies.52 Recent media reports alleging injury after vaccination53 and scientific studies linking vaccination with chronic disease54 attest to the need for further research on vaccine adverse events. Pediatricians must assimilate ever-increasing data on vaccines and apply principles of risk communication in order to translate AAP and ACIP vaccine recommendations into actual practice.Vaccine risks, whether rare or common, scientifically documented or not, can affect public attitudes toward immunization and influence public policy. Individual resistance to vaccination arises from many factors, including declining disease rates, scientific uncertainty, and the desire for individual autonomy. Explanations of vaccine benefits and risks are necessary but not sufficient to ensure effective risk communication. Risk communication research describes predictors of risk acceptability and vaccine decision heuristics that can help pediatricians understand the reluctance of some parents to vaccinate their children. Effective risk communication acknowledges these factors, avoids potential pitfalls, and engages parents appropriately in vaccination decisions.We thank Edgar Marcuse, MD, MPH, Bascom Anthony, MD, and Robert Ball, MD, MPH, for their invaluable comments on the manuscript.