To evaluate the varicella immunization practices of physicians in Rochester, NY, and to identify factors that predict whether physicians administer varicella vaccine to children.
Evaluation of a 40-item survey addressing varicella immunization practices and opinions about varicella immunization that was sent to 241 pediatricians and family physicians.
A total of 172 physicians (71.4%) completed the survey. Sixty-three percent administer the vaccine to some or all varicella-susceptible children aged 1 through 5 years, while 57% administer it to children aged 6 through 11 years, and 74% administer it to adolescents 12 years and older. Physicians who did not offer the immunization were more likely to believe that (1) the vaccine should not be given to preadolescent children because "varicella is a normal part of childhood"; (2) by giving the vaccine, varicella may shift from being primarily a childhood illness to being primarily an adult illness; (3) children get enough immunizations already and should not be given additional immunizations to prevent varicella; (4) it may be preferable to only immunize immunocompromised individuals and their close contacts; and (5) it would cost too much to immunize all American children who have not had varicella. Physicians most likely to offer the vaccine were pediatricians and those whose patients were covered primarily by private insurance plans. After becoming aware of morbidity and mortality rates for varicella-related complications, many physicians who did not administer the vaccine, or administered it only at the request of a parent, indicated that they would be more likely to offer it.
Most physicians in the Rochester area administer varicella vaccine, especially to adolescents 12 years and older. Specialty, predominant insurance type billed, and various opinions characterized those who did not offer the immunization. Providing these physicians with information about varicella-related complications may make them more likely to immunize.
IT IS ESTIMATED that more than 3.9 million cases of chickenpox occur in the United States annually, resulting in approximately 10,000 hospitalizations and 90 to 100 deaths. Most cases occur in previously healthy children.1 A 1985 study2 estimated annual US health care costs resulting from chickenpox to be $399 million. In addition, a typical case of chickenpox results in the loss of 8.7 days of school for the affected child, while the parent or guardian loses an average of 0.5 to 1.8 days of work outside the home while caring for the child during the illness.3,4
Live attenuated varicella vaccine was first developed in Japan and Korea, more than 2 million doses of the in the 1970s. Since its initial licensure for routine use in Japan and Korea, more than 2 million doses of the vaccine have been administered in those nations. Subsequently, the vaccine was also licensed in several European countries, where its administration is limited to immunocompromised children. After many years of testing, varicella vaccine was licensed in the United States in 1995 for the vaccination of healthy children and adolescents who had not previously had chickenpox. Soon thereafter, the American Academy of Pediatrics recommended that most children who had not previously had chickenpox be given the varicella vaccine,1 a view that was subsequently endorsed by the Advisory Committee on Immunization Practices of the Centers for Disease Control and Prevention.5 Since that time, 11.7 million doses have been distributed in the United States (Jane Seward, MBBS, MPH, Centers for Disease Control and Prevention, oral communication, July 23, 1998).
Routine vaccination of children with varicella vaccine remains controversial. Concern has been expressed by physicians6- 9 that widespread varicella vaccination may be unnecessary because chickenpox is generally seen as a mild, self-limited disease; others1,6,7,10,11 are concerned about the length of time that immunity afforded by varicella vaccine may last. Some physicians10,11 have concerns about the cost of widespread varicella vaccination, while others10 believe it is unethical to immunize children to prevent an illness that primarily affects adults. Finally, concern has been expressed1,6,7,10 that widespread varicella vaccine administration may cause chickenpox to shift from being a childhood illness to being an adult illness, with the associated potential for more severe adverse effects. This survey was developed to elucidate the varicella vaccine immunization practices of a representative group of physicians in 1 community, to evaluate their opinions about common concerns related to widespread varicella vaccine administration, to identify factors that characterize those who do or do not administer varicella vaccine, and to determine if providing physicians with information about morbidity and mortality rates associated with varicella may affect their likelihood to immunize.
A 40-item, anonymous, self-administered survey addressing varicella immunization practices and opinions about varicella vaccination, and including demographic information, was sent in the summer and fall of 1996 to all 241 attending pediatricians, family physicians, and internists with admitting privileges at the Children's Hospital at Strong, in Rochester, NY, who routinely provided well-child care to children or adolescents. This included the vast majority of physicians who provide health care services to children and adolescents in the community and was therefore considered to be a representative group. Being an anonymous survey, written consent for participation was not required from survey subjects.
Participating physicians were asked to return separately the completed anonymous survey and a postcard listing their names so as to acknowledge that they had sent back the survey. All surveyed physicians who did not return the postcard after the first mailing received a second mailing of the survey; if they had previously completed the survey, they were asked to not complete it again. The survey included 8 items about the physician's varicella immunization practices, 25 opinion items related to varicella immunization, 2 knowledge questions, and 5 demographic items. Opinion items on the survey were categorical and used a 5-point Likert scale (1 indicating strongly agree; 2, agree; 3, uncertain; 4, disagree; 5, strongly disagree).
Results of the survey were analyzed using descriptive statistics and χ2 analysis. Specialty and sex of surveyed physicians and respondents were compared to evaluate how representative were respondents of the physician population. Responses to opinion items were condensed into "agree," "uncertain," and "disagree" categories for univariate and bivariate analysis. Variables with significance levels of P≤.05 in bivariate analyses were included in multivariate analyses performed using backward, stepwise logistic regression. SPSS statistical software was used for all analyses.12
One hundred seventy-two (71.4%) of the 241 surveyed physicians completed the survey. Seventy-four percent of respondents specialized in pediatrics, 25% specialized in family medicine, and 1% specialized in internal medicine (but primarily provided care to adolescents and young adults). Sixty-three percent of both the surveyed physicians and the respondents were male. Pediatricians were more likely to be represented among responding physicians than among the group of surveyed physicians as a whole (74% vs 69%, P=.005).
Table 1 depicts the stated varicella immunization practices of the surveyed physicians toward 3 age groups of children: preschool children ages 1 through 5 years, school-age children ages 6 through 11 years, and adolescents 12 years and older. The table illustrates the proportion of physicians who routinely offered varicella immunization compared with the proportions who either immunized children by parental request only or who did not immunize at all. The age group that physicians were most likely to routinely offer varicella immunization to were adolescents 12 years and older (74%).
Table 2 lists the percentages of responding physicians who agreed, disagreed, or either indicated that they were uncertain or did not respond to each of several opinion statements related to varicella immunization. Notably, most responding physicians disagreed with the statement that varicella vaccine should not be given to preadolescent children because chickenpox is a normal part of childhood. Most also agreed that the potential loss of parental work time due to chickenpox should be a consideration in deciding whether to immunize.
Table 3 compares the percentages of physicians who offer varicella immunization to children in the aforementioned age groups based on whether they agreed or disagreed with each opinion statement. Physicians who believed that chickenpox is a normal part of childhood, who believed that young children get enough vaccinations already, who were concerned that universal administration of varicella vaccine to susceptible individuals may cause chickenpox to shift from primarily being a childhood illness to primarily being an adult illness, who preferred a strategy of only giving varicella vaccine to immunocompromised individuals and their close contacts, who did not think that universal immunization is necessary to induce herd immunity, or who did not agree that the potential loss of parental work time due to chickenpox should be a consideration in deciding whether to immunize were much less likely to offer varicella vaccine to children ages 1 through 11 years than were responding physicians who held contrasting views.
When responding physicians who did not routinely recommend varicella immunization were then evaluated separately, they were found to have particularly strong beliefs in 3 areas. Their stated chief concern was that universal administration of varicella vaccine may cause chickenpox to shift from being primarily a childhood illness to being an adult illness followed, in turn, by a concern that adults would be unlikely to receive booster shots of varicella vaccine if immunity wanes over time and a belief that the potential loss of parental work time should not be a consideration in deciding whether or not to immunize.
Table 4 compares percentages of physicians who report offering varicella vaccine to their patients according to demographic characteristics of the physicians and their practices. As noted in the table, specialty and predominant insurance type billed were highly significant predictors of whether the physician routinely offered varicella immunization to children in each of the 3 age groups. When pediatricians and family physicians were compared within each insurance type predominately billed, pediatricians who predominately bill private insurance plans were significantly more likely to offer varicella immunization to children in each of the 3 age groups (72% vs 37%, P=.01 for children ages 1 through 5 years; 72% vs 40%, P=.003 for children ages 6 through 11 years; and 83% vs 48%, P<.001 for adolescents 12 years and older) than were family physicians who also predominately bill private insurance plans. Among pediatricians and family physicians who predominantly bill Medicaid plans, the percentages who offer varicella vaccine did not differ significantly. However, these percentages were lower than those found among physicians within the same specialities who primarily bill private insurance plans.
Table 5 lists the variables from Tables 2 and 3 that were found by multivariate logistic regression to be independently associated with the offer of varicella vaccination for children in each age group. Few of the variables that were significant predictors of the offer of immunization in the bivariate analyses were also significant in multivariate analysis. In fact, physician's specialty was the only variable independently associated with the offer of varicella immunization to all 3 age groups. Physicians who specialized in pediatrics were much more likely to offer varicella immunization to their patients in each age group evaluated than were physicians from other specialties. Predominant insurance type billed was independently associated with the offer of varicella immunization to children younger than 12 years. However, the New York State Vaccine for Children program did not provide varicella vaccines for eligible children until shortly after completion of the survey.
Sixty-three percent of responding physicians reported being aware that, in the United States each year, approximately 100 previously healthy individuals (mostly adolescents and children) die of complications related to chickenpox. However, only 45% of responding physicians reported being aware that, in the United States each year, approximately 10,000 individuals are hospitalized as a result of complications related to chickenpox. Physicians who were aware of these statistics were significantly more likely than those who were not to offer varicella vaccine to children in each of the 3 age groups (Table 6). In addition, most physicians who were not previously aware of these morbidity and mortality figures (59% and 62%, respectively) reported that, as a result of knowing this information, they are more likely or significantly more likely to consider immunizing their patients who have previously not had chickenpox.
Although varicella vaccine has been used extensively for several years in a number of countries, in 1995, the United States became the first country to recommend administration of varicella vaccine to almost all varicella-susceptible individuals older than 1 year.13 The licensure of the vaccine for routine use in the United States was followed in quick succession by recommendations for near-universal immunization of susceptible children by the appropriate policy-setting committees of the American Academy of Pediatrics1 and the Advisory Committee on Immunization Practices.5 Considerable controversy resulted regarding the advisability of immunizing nearly all varicella-susceptible individuals. Several arguments have been brought forth challenging this recommendation, including beliefs that chickenpox is a benign childhood disease with few serious sequelae for which no preventive immunization is needed6,7,10; that the cost of extensive immunization is prohibitive10; that waning immunity may result over time after immunization, which may cause chickenpox to become a common, serious adult illness,10 with the likely result that adults who had not been immunized as children would then have a much more severe illness course when they contract chickenpox1,6,7,11; and that it is unethical to immunize children to prevent an illness for which the primary morbidity occurs in adults.10 Other concerns relate to the cost of obtaining a deep freezer to store the vaccine,11 insurance reimbursement for the cost of the vaccine and its administration,14 and the potential for significant adverse effects should the vaccine be given to a pregnant woman or to another individual who shares a household with a pregnant woman.8
Results of the present survey, conducted among physicians in a community in which immunization awareness is generally high, confirm many of the aforementioned concerns about universal varicella immunization. Also evident are differences by specialty between pediatricians and family physicians regarding the offer of varicella vaccine. Other studies9,15,16 that have compared rates of immunizations administered by pediatricians and family physicians have shown a similar pattern of significantly higher immunization rates attributed to pediatricians than to family physicians. This variation between specialties is probably largely due to differences in the characteristics of family physicians' and pediatricians' practices. Individual family physicians generally provide medical care to far fewer children than do pediatricians and, hence, may be less likely than pediatricians to be aware of and to implement new pediatric immunization recommendations.
Insurance reimbursement for the purchase of varicella vaccine varied considerably at the time of the survey. Private and employer-sponsored insurance plans reimbursed physicians for the cost of purchasing and administering varicella vaccine, but the Vaccine for Children program, which provided vaccines for children insured by Medicaid and Medicaid managed care plans, did not distribute varicella vaccine. Not surprisingly, as a result, the primary insurance type billed was noted in this survey to be an independent predictor of whether physicians offered varicella vaccine to patients 1 through 11 years of age. It is notable, however, that although the Vaccine for Children program did not provide varicella vaccine, a high proportion (60%) of physicians who reported primarily seeing Medicaid patients nevertheless offered it to adolescents, who generally constitute a higher-risk group for complications due to varicella than younger children.
Many individuals, both in the general public and in the medical community, see chickenpox as a mild illness with few serious sequelae. As a result, chickenpox infection is considered a virtual rite of passage for children in the United States.6,8 Indeed, this survey demonstrates that many, if not most, physicians are unaware of the extent of the morbidity and mortality associated with chickenpox (approximately 10,000 annual varicella-related hospitalizations and 100 annual varicella-related deaths in the United States). The study also demonstrates that physicians who are aware of these morbidity and mortality figures are significantly more likely to offer varicella vaccine to their patients than are physicians who are unaware of these data. Therefore, it is quite conceivable that providing physicians with information about the potential adverse effects of varicella could result in increased immunization levels. Presumably believing this to be the case, the company that produces and distributes varicella vaccine in the United States has pursued a vigorous marketing campaign centered on informing physicians of the serious potential complications of varicella.
A number of changes have been made to the pediatric immunization schedule in recent years, primarily reflecting the introduction of new vaccines. At the same time, nonimmunization-related recommendations have also been substantially revised. When the Advisory Committee on Immunization Practices first recommended routine hepatitis B immunization of children in 1991, there was initially a great deal of opposition among physicians.15,16 Similarly, when the American Academy of Pediatrics first adopted its recommendation that infants be put to sleep on their backs, many physicians were opposed.17 Nevertheless, with the passage of time and increasing evidence of the benefit of these recommendations, a large majority of physicians now routinely both immunize children against hepatitis B16,18 and advise parents of newborns to place their infants in the supine sleeping position.17,19,20 It is possible that, over time, the recommendations for near-universal immunization to prevent varicella will similarly become accepted by increasing numbers of physicians and parents. It is also possible that the major reason for the current high rates of hepatitis B immunization is the proliferation of state laws requiring completion of the hepatitis B series for school entry. Although no states require varicella immunization or proof of disease for school entry as of yet, the District of Columbia instituted such a requirement in April 1998 and broadened it to include children older than 1 year entering day care. Several states have enacted and are in the process of instituting similar requirements. Certainly, incorporation of proof of disease or of varicella vaccination into the requirements for school entry would dramatically increase vaccination rates. In the meantime, the strength of the recommendation of their child's physician is likely to be among the most important criteria that parents use in determining whether to have their children immunized with varicella vaccine.
Certain limitations may have affected this survey. It is possible that nonrespondents had opinions and practices that were substantially different than those of respondents. In addition, since the survey relied on self-report, actual physician practices may differ significantly from those noted in physician responses to survey items. Finally, physician opinions and practices may have changed somewhat since the survey was conducted.
While this survey was performed in a single community, its results are likely to be widely generalizable. Whether to recommend varicella immunization and how strongly to do so are questions that confront every primary care physician in the United States who provides services to children and adolescents. The controversy about varicella vaccine has been widespread and heated and reflects substantive concerns that must be addressed. In that regard, this study's findings should be of benefit in assessing the reasons why many physicians do not currently immunize with varicella vaccine, in developing responses to address the concerns of physicians who do not routinely recommend the vaccine, and in promoting effective strategies for educating physicians and the public regarding both chickenpox and the vaccine used to prevent it.
Accepted for publication August 18, 1998.
Presented in part at the 1997 Annual Meeting of the Pediatric Academic Societies, Washington, DC, May 3, 1997.
We acknowledge the many pediatricians and family physicians in the Rochester, NY, area who took time from their busy practices to complete this survey.
Corresponding author: Stanley J. Schaffer, MD, MS, Division of General Pediatrics, Box 777, University of Rochester Medical Center, 601 Elmwood Ave, Rochester, NY 14642 (e-mail: Stanley_Schaffer@urmc.rochester.edu).
Schaffer SJ, Bruno S. Varicella Immunization Practices and the Factors That Influence Them. Arch Pediatr Adolesc Med. 1999;153(4):357-362. doi:10.1001/archpedi.153.4.357