Vaccine-Associated Liability Risk and Provider Immunization Practices FREE

THE POTENTIAL for malpractice litigation has become a part of life for physicians in the United States. A segment of the health care delivery system associated frequently with claims of adverse reactions and outcomes is the manufacture and administration of pediatric vaccines.

Although vaccines have greatly reduced the incidence of many infectious diseases in our society,¹ they also have been the focus of a notable amount of litigation against their manufacturers and those responsible for their administration.² Physicians and nurses are among those who have been defendants in cases alleging permanent damage to children from vaccines.³

The quantity of suits filed and damages claimed reached a level such that in 1986 the federal government acted to limit the liability exposure of vaccine manufacturers and providers. In that year, Congress passed the National Childhood Vaccine Injury Act after reports appeared that described some vaccine manufacturers ceasing to produce vaccines and some providers unwilling to administer vaccines secondary to the risk of litigation.⁴ This act established a federal compensation program for those who claimed injury from vaccines, to be funded by excise taxes imposed on vaccine manufacturers.⁵

Although the act was well received by manufacturers and providers of vaccines, concerns regarding the extent of protection it affords to these parties have been expressed.^{3 ,6 - 7} Practitioners are still at risk for civil suits even if parents pursue claims through the National Childhood Vaccine Injury Act.⁶

There are numerous reports of physicians not providing immunizations to children with minor illnesses.^{6 ,8 - 9} Despite recent reports from the Institute of Medicine regarding the safety of childhood vaccines¹⁰ and the Standards for Pediatric Immunization Practices¹¹ promulgated by the Centers for Disease Control and Prevention, physicians continue to use such inappropriate contraindications to justify the delay of immunizations.^{12 - 13} It is unclear what role the fear of adverse reactions and the potential for subsequent litigation play in a physician's reluctance to give immunizations to children with minor illnesses; it is also unknown whether differences in this regard exist among pediatricians and family physicians. We are unaware of any other studies assessing this issue directly. Other reasons for immunization refusal during minor illness may include parental preferences, the belief by physicians that their patients always will return for an immunization following resolution of their illness, the belief by physicians that children may have a poor immunologic response, or time constraints of sick-child visits.

This study examines the perceptions of pediatricians and family physicians regarding their liability exposure for immunization provision and determines if that perception is associated with a physician's immunization practices.

As part of a larger study assessing the methods of immunization recommendation dissemination,¹⁴ questionnaires were mailed in late 1993 to physicians selected from the American Medical Association's Masterfile, a database of all physicians in the United States, including nonmembers of the American Medical Association. A total of 3014 physicians (1165 pediatricians and 1849 family physicians) in selected metropolitan and nonmetropolitan areas (as defined by the Office of Management and Budget) of 9 states were surveyed. States were chosen to provide variability in region of the country, population dispersion (metropolitan vs nonmetropolitan), patterns of organization of care (fee-for-service vs managed care), state vaccine distribution systems, and immunization rates. The states selected were California, Texas, Wisconsin, Colorado, Massachusetts, Tennessee, Pennsylvania, Georgia, and Hawaii. Family physicians were oversampled relative to pediatricians because our previous studies of immunization practices indicated that approximately 20% of family physicians limit their care to adults or do not administer pediatric immunizations.^{15 - 16}

Questionnaires were accompanied by hand-signed, personalized cover letters. Three reminder letters and questionnaires were sent to all nonrespondents during the 4 months following the initial mailing. The questionnaire was pretested with a sample of practicing pediatricians and family physicians in South Carolina to ensure clarity of interpretation and ease of completion. The pretest data are not included in our analysis. Additional descriptive information regarding the questionnaire has been published elsewhere.¹⁴

The study variables were constructed from the survey data and from demographic data available from the American Medical Association's Masterfile. Outcome variables of primary interest included physicians' self-reported perceptions of their legal and financial risks for providing specific childhood immunizations, their perceptions of the protection against immunization-related lawsuits afforded by state and federal programs, and their immunization practices (ie, documentation of immunization provision and diphtheria, tetanus, and pertussis [DTP] vaccine use in patients with an acute or chronic illness). Variables hypothesized to predict physician attitudes and behavior included physician characteristics (specialty and specialty society membership), practice type (solo, group, or public clinic), practice location (metropolitan vs nonmetropolitan), and the proportion of patients enrolled in managed care plans and in Medicaid.

An initial analysis involved frequency distributions and a univariate analysis of all pertinent items from the survey and the American Medical Association's Masterfile. Likert scales¹⁷ were collapsed to form dichotomous variables, and cross tabulations and χ² analyses were used to determine the significance of the association of each predictor variable with the outcome variables. Next, comparisons were made between the attitudes and practices of pediatricians and family physicians using χ² analysis. Predictors identified as significant in these models are those at the P<.01 level. We believe that this rigorous criterion is appropriate given the many variables in each model.

Finally, multiple regression analysis was used to determine the independent effect of each predictor variable on the outcome variables of interest. The variables used in the regression equations are found in Table 1. The outcome variables for the equations were (1) physician perceptions of the safety of immunizations in children with a mild respiratory tract illness without a fever, (2) physician perceptions of the safety of DTP vaccine use in children with a family history of seizures, (3) physician perceptions of the safety of DTP vaccine use in children with a local reaction and a fever (temperature <40.5°C) after a previous DTP vaccine, and (4) whether physicians reported giving immunizations at acute care visits if not specifically contraindicated.

All analyses were adjusted (weighted) for the probability of selection and differential response among those surveyed. Factors associated with differential response included physician specialty, state of residence, and board certification. The initial weights were rescaled to allow for generalization to the population from which the sample was drawn without inflating the sample size. Estimates obtained using rescaled weights are the same as those obtained using original weights; rescaling was performed primarily to avoid exaggerated statistical significance due to the large sample size.

After removing from the sample those who had retired, died, or moved out of state, responses were received from 1863 of 2814 eligible subjects, for an overall response rate of 66% (72% of pediatricians and 63% of family physicians). The 442 respondents who did not give immunizations and who did not provide well-child care were excluded, leaving a sample of 1421 for subsequent analysis (742 pediatricians and 679 family physicians). The response rates did not vary by physician sex, age, country of training (United States vs international), graduation year, discipline (allopath vs osteopath), or location (metropolitan vs nonmetropolitan). The application of rescaled weights yielded effective sample sizes of 704 pediatricians and 723 family physicians. The demographic characteristics of the respondents have been described elsewhere¹⁸ and are given in Table 2.

Overall, less than 30% of all respondents believed that federal and state programs protected them against vaccine-related lawsuits. There were significant differences between pediatricians and family physicians regarding these beliefs. Pediatricians were more likely to believe that current programs protect them against lawsuits (32% vs 21%, P<.001). Yet, only 13% of family physicians and 12% of pediatricians believed that providing pediatric immunizations significantly increased their chances of a malpractice suit. Pediatricians were also more likely to believe that the DTP vaccine could be administered safely to children with a family history of seizures, minor respiratory tract illnesses, or a previous local reaction to a DTP injection (Table 3).

Regarding immunization practices in the acute care setting, no significant differences were seen between the 2 specialties. Most pediatricians (82%) and family physicians (79%) report that they routinely review the immunization needs of children at acute care visits and provide vaccines at those visits if there are no contraindications (pediatricians, 87%; family physicians, 85%).

More than 80% of physicians in both specialties reported systems to document immunization provision (Table 4). However, differences between the 2 specialties in manner of documentation exist. Pediatricians were more likely than family physicians to record immunizations on a special page in the patient's medical record and to record immunizations in some type of computer-based system.

The results of the logistic regression analysis showed significant differences in the immunization practices of family physicians compared with pediatricians (Table 5). Table 5 includes only those variables from Table 1 that achieved statistical significance in the regression equations. Family physicians were more likely to defer immunization because of a mild illness or reactions to previous doses of vaccine, as were physicians older than 40 years in both specialties. Liability issues were not significantly associated with any of the outcome variables, except that those who believed that the whole-cell DTP vaccine increased their risk for lawsuits were less likely to believe that the DTP vaccine was safe for children with a family history of seizures.

Additional regression models (not in the table) demonstrated that family physicians were less likely to believe that state and federal laws protected them against immunization malpractice claims (OR, 0.59; P<.001). No relationship was observed between this outcome and the percentage of a physician's patients with Medicaid insurance or in a managed care group.

Data from this study support the notion that most pediatricians and family physicians do not believe that state and federal programs protect them from immunization-related lawsuits.

Although the National Vaccine Injury Compensation Program, created by the National Childhood Vaccine Injury Act, had a notable positive effect on the vaccine liability crisis of the 1980s by decreasing the risk for manufacturers (and the number of suits filed), practitioners were not totally protected.⁶ As written, the program left practitioners still at risk for civil suits and penalties theoretically beyond their control.¹⁹ Although vaccine manufacturers are expressly protected under the statute, the act does not specifically confer similar immunity on physicians.^{3 ,20}

However, most respondents did not believe that administering vaccines placed them at increased risk of litigation even though parents still could choose to pursue civil claims rather than use the National Vaccine Injury Compensation Program. Loopholes in the plan led to assertions of successful "double dipping" by some claimants (pursuing claims through the National Vaccine Injury Compensation Program and through the civil courts simultaneously).⁶ The legality of this practice was confirmed by the First Circuit Court of Appeals in Boston, Mass, which ruled in 1994 that families of children injured by vaccines can sue for damages even after the child has been compensated by the special fund for vaccine victims (Schafer v American Cyanamid Co, No. 93-1422, 1994).²¹ Yet, physicians in our study reported that they do not believe that they are at risk for immunization-related lawsuits.

Documentation of office immunization practices has been shown to be highly variable, yet in this study physicians have self-reported that their practices maintain accurate and complete records. It is possible that the act has had an effect on the maintenance of permanent vaccine records.²² Providers are required by law to record the date a vaccine was administered, the manufacturer and lot number of the vaccine, and the name, address, and title of the person administering the vaccine.^{23 - 24} Data from this study demonstrate little or no association in the perceptions of liability among physicians and their self-reported record-keeping practices.

The most striking finding of the study is the continued practice by so many physicians to withhold immunizations inappropriately for mild illnesses, previous mild reactions to vaccines, or a family history of seizures.⁹ That only two thirds of family physicians would provide a DTP vaccine to a child with a family history of seizures, and that only slightly more than half would give the same vaccine to a child experiencing a fever (temperature <40.5°C) following a previous dose of the vaccine, is a cause for concern. Although fewer pediatricians reported practicing inappropriate contraindications, many did as well.

The liability-related variables were not found to be independently associated with most of the vaccine-related behaviors examined. This raises the question as to whether physicians may use liability concerns as a reason for not immunizing ill children when their concerns are actually otherwise. These data suggest that physicians are not comfortable immunizing ill children, regardless of their concerns about liability. There were 2 exceptions, the notable associations between (1) the belief that the use of the whole-cell DTP vaccine increases the risk for lawsuits and the lack of agreement with the statement that the DTP vaccine is safe for children with a family history of seizures and (2) the belief that laws protect physicians against vaccine-related liability and the belief that it is safe to immunize children with a mild respiratory tract illness. This final association should be appreciated within the context that the data also demonstrated that less than one third of all respondents believed they were protected by such laws.

To the extent that the link between malpractice fears and inappropriately withholding immunizations emanates from physicians themselves, it is possible that physicians are not fully accurate in their understanding of the factors affecting their own immunization behaviors. Although it is commonly believed that people know why they do the things they do, evidence suggests that our causal attributions for our own behavior (and that of others) are frequently faulty.²⁵ Two of the most common biases in people's explanations for their own behaviors are the tendency to see them as due to external forces beyond control, called actor-observer bias, and the tendency to deny responsibility for negative behaviors, called self-serving bias.^{26 - 28} These biases alone may explain why physicians believe that malpractice fears affect their immunization practices, yet no statistical link between their malpractice fears and immunization practices can be found. A similar inaccuracy has been found in family physicians' explanations for the role of malpractice fears and insurance costs in their decision not to provide obstetrical services.^{29 - 30}

There are some limitations to this study. First, these are self-reported data and we did not independently verify actual clinical practice. Second, there is the possibility that response bias played a role in these results as those with a particular perspective of these issues may have been more or less likely to respond to the survey.

Private physicians are an essential component of the multidimensional vaccine provision system in the United States.³¹ These data suggest that the decision to offer immunizations to ill children may not be related to physicians' perceptions regarding their potential liability for vaccine-related injury but rather to other concerns. Such concerns may include the potential harm or suffering caused to their patients, parental preferences, and even nursing preferences. Although these data confirm physicians' skepticism regarding protection from liability provided by current programs, it is likely that educational efforts focused on that issue alone will have little effect on the practice of inappropriately postponing immunizations. Additional efforts are needed, especially among family physicians, to understand the basis behind the perceptions of physicians regarding the safety of vaccine administration to children with mild illnesses, a family history of seizures, and mild reactions to previously received DTP vaccines.

Accepted for publication October 24, 1997.

This study was supported by grant AHCPR R-01 HS 07286-0 from the Agency for Health Care Policy and Research, Rockville, Md.

We thank Monica A. Price for administrative assistance; and Sarah J. Clark, MPH, for her critical review of the manuscript.

Editor's Note: It looks as though eliminating lawyers would not increase immunizations; any other ideas...about the immunizations, that is.—Catherine D. DeAngelis, MD

Reprints: Gary L. Freed, MD, MPH, Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, CB 7590, Chapel Hill, NC 27599-7590.