Pediatricians’ intention to administer human papillomavirus vaccine: the role of practice characteristics, knowledge, and attitudes

Journal of Adolescent Health 37 (2005) 502–510

Original article

Pediatricians’ intention to administer human papillomavirus vaccine: the role of practice characteristics, knowledge, and attitudes Jessica A. Kahn, M.D. M.P.H.a,d,*, Gregory D. Zimet, Ph.D.b, David I. Bernstein, M.D.c,d, Jeremy M. Riedesel, M.S.d, Dongmei Lan, M.S.e, Bin Huang, Ph.D.a,d,e, and Susan L. Rosenthal, Ph.D.f a

Division of Adolescent Medicine, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio b Division of Adolescent Medicine, Indiana University, Indianapolis, Indiana c Division of Infectious Diseases, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio d University of Cincinnati College of Medicine, Cincinnati, Ohio e Center for Epidemiology and Biostatistics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio f Division of Adolescent and Behavioral Health, University of Texas, Medical Branch at Galveston, Galveston, Texas Manuscript received June 23, 2005; manuscript accepted July 24, 2005

Abstract

Purpose: The objective of this study was to examine pediatrician characteristics and attitudes associated with intention to recommend two hypothetical human papillomavirus (HPV) vaccines. Methods: A survey instrument mailed to a random sample of 1000 pediatricians assessed provider characteristics, HPV knowledge, and attitudes about HPV vaccination. Intention to administer each of two HPV vaccines types (a cervical cancer/genital wart vaccine and a cervical cancer vaccine) to girls and boys of three different ages (11, 14, and 17 years) was assessed. Linear mixed modeling for repeated measures and multivariable linear regression models were performed to identify variables associated with intention to recommend vaccination. Results: The mean age of participants (n ⫽ 513) was 42 years and 57% were female. Participants were more likely to recommend vaccination to girls vs. boys and older vs. younger children, and were more likely to recommend a cervical cancer/genital wart vaccine than a cervical cancer vaccine (p ⬍ .0001). Variables independently associated with intention to recommend a cervical cancer/genital wart vaccine were: higher estimate of the percentage of sexually active adolescents in one’s practice (␤ .084, p ⫽ .002), number of young adolescents seen weekly (␤ 1.300, p ⫽ .015), higher HPV knowledge (␤ 1.079, p ⫽ .015), likelihood of following the recommendations of important individuals and organizations regarding immunization (␤ .834, p ⫽ .001), and fewer perceived barriers to immunization (␤ ⫺.203, p ⫽ .001). Conclusions: Vaccination initiatives directed toward pediatricians that focus on modifiable predictors of intention to vaccinate, such as HPV knowledge and attitudes about vaccination, may facilitate adherence to emerging national immunization guidelines. © 2005 Society for Adolescent Medicine. All rights reserved.

Keywords:

Pediatrician; Intention; Human papillomavirus; Vaccine

*Address correspondence to: Dr. Jessica A. Kahn, Division of Adolescent Medicine, MLC 4000, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45229. E-mail address: [email protected]

Human papillomavirus (HPV) infection is common among sexually experienced youth: prevalence rates up to 82% in adolescent and young adult women have been reported [1,2]. Sexually transmitted HPV genotypes are categorized into low-risk types, which may cause genital warts, and high-risk types, which may cause cervical cancer [3,4]. Safe

1054-139X/05/$ – see front matter © 2005 Society for Adolescent Medicine. All rights reserved. doi:10.1016/j.jadohealth.2005.07.014

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

and clinically effective vaccines to prevent HPV infection are being developed [5,6]. Vaccination will likely be recommended for preadolescents and early adolescents because once adolescents initiate sexual intercourse, they are at high risk for HPV infection. Almost half of the 15–19year-old young women who participated in a recent study were HPV-positive within 3 years of sexual initiation [7]. Vaccines designed to prevent cervical cancer, genital warts, or both will be available soon for clinical use. The effectiveness of vaccine delivery will depend largely upon whether providers recommend the vaccine to patients, those at risk for infection and its sequelae agree to be immunized, and effective public health strategies for immunization are developed and implemented [8,9]. Pediatricians will play a critical role in HPV vaccine delivery because early adolescents are more likely to visit a pediatrician than any other provider [10], and because a pediatrician’s recommendation is likely to influence strongly a parent’s or adolescent’s decision to receive the vaccine [9]. Therefore, an important first step in developing effective vaccine delivery programs is to understand pediatricians’ attitudes about HPV immunization and factors associated with their intention to recommend immunization. Previous work has demonstrated that factors associated with physician immunization practices include provider characteristics, knowledge, and attitudes about vaccination [11–17]. Physician attitudes about recommending HPV vaccines may differ from their attitudes about routine childhood vaccines. For example, providers may anticipate specific barriers to immunizing children against sexually transmitted infections (STIs), including personal reluctance to discuss STI vaccines with preadolescents or parental concerns about STI vaccines. In addition, provider attitudes may be linked to the disease targeted (cervical cancer vs. genital warts) or patient characteristics [18,19]. The aims of this study were to examine pediatricians’ attitudes about HPV immunization, to assess intention to recommend two hypothetical HPV vaccines (one that would prevent both cervical cancer and genital warts, and one that would prevent cervical cancer only) to male and female patients of varying age, and to identify provider characteristics and attitudes about immunization associated with intention to recommend each HPV vaccine type. We hypothesized that more clinical experience with adolescent patients, knowledge about HPV, and positive attitudes about HPV vaccine would be associated with intention to recommend vaccination. Methods The survey instrument developed for this study was based on a validated survey of physicians’ immunization knowledge, attitudes, and practices [20] and a survey previously developed by the investigators [21]. It was pretested in a sample of 20 pediatricians and family physicians, and pilot tested in a sample of American Academy of Family

503

Physicians (AAFP) members [22]. The revised, self-administered survey instrument was then mailed twice over a two-month period in 2004 to a random sample of 1000 American Academy of Pediatrics (AAP) members. The sample size was determined using a power calculation based on a return rate of 40% (n ⫽ 400), and 80% power to detect differences in intention to recommend HPV vaccine between physicians with varying characteristics. The first survey contained $10 compensation. The research protocol was approved by the Cincinnati Children’s Hospital Medical Center Institutional Review Board. The design of a number of survey constructs was based on the Theory of Planned Behavior [23]. This theory has been used to predict physician behaviors including delivery of preventive heath services [24] and immunization practices [20]. The Theory of Planned Behavior proposes that the most important determinant of a behavior is an individual’s intention to perform that behavior. A substantial body of literature confirms the strength of the association between behavioral intention and actual behavior [25,26]. Determinants of behavioral intention include the individual’s attitudes about performing the behavior, perceptions about the attitudes of others important to that individual, and perceived control over performing the behavior (including barriers to performing the behavior). Because the HPV vaccine is not yet available, we modified our assessment of the determinants of behavioral intention (e.g., we assessed anticipated beliefs about the vaccine) and focused on intention to recommend the vaccine as the primary outcome variable. The survey instrument included sections assessing: 1) provider characteristics (3 items); 2) practice characteristics, including number of 10 –15-year-old children seen weekly in one’s practice and percentage of those visits that are for routine health maintenance (5 items); 3) attitudes about adolescent patients, including comfort discussing issues of sexuality with this population, frequency of these discussions during health maintenance visits, and estimate of the percentage of 15-year-old male and female patients in one’s practice who are sexually active (4 items); 4) knowledge about HPV (5 items, correct items summed for a possible knowledge score of 0 –5), attitudes about HPV vaccination, and intention to administer HPV vaccine. The following attitudes were assessed: ideal age for vaccination (single item), importance of HPV vaccine compared with other vaccines (6-item index), beliefs about outcomes of vaccination (8-item scale, Cronbach alpha .68), strategies for vaccine delivery (3-item scale, alpha .74), likelihood of following the recommendations of important individuals and organizations regarding immunization (6-item scale, alpha .65), and barriers to recommending immunization (14-item scale, alpha .86). Responses to most items assessing attitudes used four- or five-point Likert-type scales and the sum score of each scale was used in univariate and multivariable analyses. Intention to administer two hypothetical HPV vaccines was assessed separately for male and

504

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

Table 1 Practice and provider characteristics Characteristic Age Gender Decade completed residency

Location of practice

Number of children 10–15 years of age seen per week

Percentage of 10–15-year-old patient visits that are for routine health maintenance

Frequency with which providers discuss sexual activity with adolescent patients

Comfort addressing adolescent sexual activity Estimated percentage of 15-year-old patients in providers’ practice who are sexually active

Category

n (%)

Male Female Before 1980 1980–1989 1990–1999 2000 or later Urban Suburban Rural Other 0–10 11–25 26–49 50 or more 0–24 25–49 50–74 75–100 Almost all of the time (⬎ 90% of visits) Most of the time (50–90% of visits) Some of the time (⬎ 25% but ⬍ 50% of visits) Occasionally/never (⬎ 0% but ⬍ 25% of visits/0% of visits Somewhat/very comfortable Neither/somewhat/very uncomfortable Male Female All

220 (43.3) 288 (56.7) 73 (14.5) 107 (21.2) 140 (27.8) 184 (36.5) 264 (52.6) 184 (36.7) 44 (8.8) 10 (2.0) 171 (33.7) 186 (36.7) 104 (20.5) 46 (9.1) 218 (44.5) 171 (34.9) 84 (17.1) 17 (3.5) 264 (53.7) 138 (28.0) 43 (8.7) 47 (9.6) 428 (84.4) 79 (15.6)

Mean (SD) 41.7 (10.9)

female patients of three different ages and for two different types of vaccines. The items assessing intention were: “If an HPV vaccine were available that was effective (e.g., at least 70% effective) against cervical cancer [against genital warts and cervical cancer], what is the likelihood you would recommend the vaccine to . . . an 11-year-old girl, 11-yearold boy, 14-year-old girl, 14-year-old boy, 17-year-old girl, or 17-year-old boy in your practice?” The two primary outcome variables were intention to administer a cervical cancer vaccine and a cervical cancer/genital wart vaccine. These variables were defined as the mean value for all six responses regarding intention to recommend each of the two vaccines. There were five possible responses to these items (5 ⫽ extremely likely and 1 ⫽ extremely unlikely) so the possible score range was 1–5. The Cronbach alpha was .85 for responses to the six items measuring intention to vaccinate against cervical cancer, and .88 for the six items measuring intention to vaccinate against cervical cancer/genital warts. Univariate analyses were performed to examine associations between predictor variables (provider characteristics, practice characteristics, knowledge, and attitudes) and intention to vaccinate, using Wilcoxon rank-sum tests (for categorical predictor variables) and Spearman correlation coefficients (for continuous predictor variables). Those variables associated with intention to vaccinate at p ⱕ .10 were

37.0 (22.6) 35.8 (20.9) 36.3 (21.2)

entered into separate multivariable linear regression models. A forward selection technique confirmed with a backward elimination procedure was used to determine those variables independently associated with intention to vaccinate against each of the two vaccine types. To examine the effects of patient gender, patient age, and vaccine type on intention to recommend vaccination, a linear mixed model for repeated measures was used for all 12 items assessing intention to vaccinate boys or girls who were 11, 14, or 17 years of age against cervical cancer/genital warts or cervical cancer. Statistical analyses were performed using SPSS version 12.0 (SPSS Inc., Chicago, Illinois) and SAS version 8.2 (SAS Institute Inc., Cary, North Carolina). Results Of the 1000 surveys sent to providers, 603 (60%) were returned and of those, 33 reported that they did not care directly for patients. Of the remaining 570 who were eligible to participate and returned a survey, 513 (90% of those eligible, 51% of total n) completed the survey. Those who completed the survey, compared with those who did not, were more likely to be female (p ⫽ .018) and were younger (p ⬍ .0001). Participant characteristics and HPV knowledge are shown in Table 1. Most participants regularly cared for children 10 –15 years of age, and most of those visits were

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

505

Table 2 Attitudes and beliefs of providers concerning an HPV vaccine Attitude

n (%)

Importance of specific HPV vaccine characteristics Vaccine does not cause adverse side effects Vaccination leads to long lasting immunity Condom use in patients would not decrease if vaccinated Vaccine protects against condylomata (genital warts) as well as cervical cancer Patients could receive other vaccines at the same visit Vaccination provides providers with an opportunity to discuss sexuality issues with patients The likelihood of patients having sex would not increase if vaccinated Vaccination would eliminate the need for annual Pap tests Likelihood of following the vaccination recommendations of the following organizations/individuals American Academy of Pediatrics/Redbook Centers for Disease Control and Prevention Advisory Committee on Immunization Practices Pediatric Infectious Disease Society Colleagues Pharmaceutical representative Barriers to vaccinating 10–15-year-old patients Parental refusal/reluctance because parent concerned about safety of the vaccine Parental reluctance to have children immunized against a sexually transmitted infection Parental refusal/reluctance because parent thinks child already receives too many vaccines Provider’s ability to get older children/adolescents to show up for well visits/immunizations Parental reluctance to discuss issues of sexuality and sexually transmitted infections Parental concern that immunization may lead to an increase in risky behavior Tendency for parents to pay less attention to immunization status once child is over 5 years old Parental refusal/reluctance because parent doesn’t believe in vaccines Parental belief that their child is being singled out as one who would be at risk for a sexually transmitted infection Acquisition and administration cost of another vaccine to provider’s practice Time pressure during a well visit Recommendations for immunizations change too often Provider reluctance to talk with adolescents about issues of sexuality and sexually transmitted infections Provider reluctance to administer multiple vaccines at one time

Extremely/very important 486 (96.0) 482 (96.0) 376 (74.6) 348 (68.6) 339 (67.0) 322 (63.4) 243 (48.5) 230 (45.9) Extremely/somewhat likely 509 (99.8) 497 (98.0) 482 (94.5) 398 (78.5) 372 (73.1) 47 (9.2) Extremely/Somewhat likely 356 (70.8) 251 (50.1) 237 (47.0) 216 (43.0) 216 (43.0) 211 (41.9) 181 (36.2) 176 (35.0) 132 (26.3) 126 (25.1) 93 (18.6) 51 (10.2) 32 (6.4) 23 (4.6)

for well-child care. Most participants reported that they frequently discussed sexual activity with adolescent patients and were comfortable addressing the issue. Knowledge about HPV was fair: 45% correctly identified those subtypes associated with cervical cancer, 42% correctly identified subtypes associated with genital warts, 41% knew approximately what percentage of cervical cancer was caused by HPV infection, 32% knew the approximate prevalence of HPV infection in sexually active adolescents, and 28% knew the approximate prevalence of genital warts in sexually active adolescents. The mean HPV knowledge score was 1.86 (SD 1.29) out of a possible 5.0. Provider attitudes about HPV vaccine are shown in Table 2. Each attitude is dichotomized in the table to make the descriptive statistics easier to comprehend, but attitudinal scales were analyzed as continuous variables in univariate and multivariable analyses. Pediatricians believed that vaccine safety and long-lasting immunity would be the most important vaccine characteristics. Almost all participants reported that they would be likely to follow the immunization recommendations of the AAP/ Red Book, the Centers for Disease Control and Prevention (CDC), and Advisory Committee on Immunization Practices (ACIP). Responses to items assessing barriers

to vaccination were dominated by beliefs about parental barriers such as parental concerns about vaccine safety or reluctance to have children immunized against an STI. For example, at least 40% of pediatricians believed that the following issues would create barriers to HPV vaccination: parental concern about vaccine safety (71%), parental reluctance to allow their child to be immunized against an STI (50%), parental belief that their child receives too many vaccines already (47%), parental reluctance to discuss sexuality and sexually transmitted infections with their child (43%), and parental concern that immunization may lead to riskier adolescent behaviors (42%). Participants believed that HPV vaccine was more important than/as important as vaccines against the following diseases, respectively: anthrax (62%/21%), Chlamydia (28%/59%), tuberculosis (25%/43%), genital herpes (23%/65%), influenza (15%/41%) and hepatitis B (5%/55%). Participants reported that the following strategies would be extremely or very important for successful delivery of HPV vaccine to children and adolescents: coverage by most insurance policies (98%), availability at a reasonable cost (96%), affordability for those without insurance or whose policy would not cover the vaccine

506

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

Table 3 Pediatricians’ intention to recommend two different HPV vaccine types to patients of varying age and gender

Table 4 Pediatricians’ intention to recommend HPV vaccines as a function of vaccine type, patient age, and patient gendera

Patient age (years)

Patient gender

Cervical cancer vaccine n (%)a

Cervical cancer and genital wart vaccine n (%)a

Vaccine type

11

Female Male Female Male Female Male

313 (61.6) 188 (37.3) 433 (85.6) 271 (54.0) 483 (95.5) 322 (64.1)

341 (67.8) 305 (60.8) 463 (91.7) 418 (82.9) 498 (98.4) 462 (91.5)

Cervical Cervical Cervical Cervical Cervical Cervical Cervical Cervical Cervical Cervical Cervical Cervical

14 17 a

n (%) represents the number and percentage of those who reported they were extremely or somewhat likely to recommend the vaccine.

(96%), and mandated for school enrollment (58%). Most participants (60%) thought that the ideal age to administer an HPV vaccine would be between 9 and 13 years of age, whereas smaller proportions thought that the ideal age would be younger than 9 years of age (21%), between 14 and 16 years of age (16%), or older than 16 years of age (3%). Intention to immunize children of different ages against the two vaccine types is shown in Table 3 and Figure 1. The results of linear mixed modeling for repeated measures indicated that there were significant effects of the following factors on intention to recommend vaccines: patient gender (F value 286.4, p ⬍ .0001), patient age (F value 194.4, p ⬍ .0001), type of vaccine (F value 274.5, p ⬍ .0001), the interaction of age and gender (F value 11.7, p ⬍ .0001), the interaction of age and vaccine type (F value 3.6, p ⫽ .028), and the interaction of gender and vaccine type (F value 206.0, p ⬍ .0001). At each age, participants were most likely to recommend a cervical cancer/genital wart vaccine to girls, followed by a cervical cancer vaccine to girls,

Fig. 1. Pediatricians’ intention to recommend HPV vaccines by vaccine type, patient age, and patient gender. Intention was measured using the mean value of the five possible responses (5 ⫽ extremely likely to 1 ⫽ extremely unlikely).

cancer/genital cancer cancer/genital cancer cancer/genital cancer cancer/genital cancer cancer/genital cancer cancer/genital cancer

wart wart wart wart wart wart

Age

Gender

Mean intention scoreb

Standard deviation

17 17 17 17 14 14 14 14 11 11 11 11

Girl Girl Boy Boy Girl Girl Boy Boy Girl Girl Boy Boy

4.74 4.65 4.52 3.62 4.42 4.23 4.16 3.27 3.77 3.58 3.57 2.80

.02 .03 .04 .07 .04 .04 .05 .07 .06 .06 .06 .06

a Within each of the three age groups (17, 14, and 11 years of age), differences in intention to vaccinate were significant (p ⬍ .001) with the exception of the following: the difference between intention to recommend a cervical cancer vaccine to girls and intention to recommend a cervical cancer/genital wart vaccine to boys did not differ significantly at age 11 and age 14. b Mean value of the five responses (5 ⫽ extremely likely to 1 ⫽ extremely unlikely) to questions assessing intention to recommend HPV vaccine.

followed by a genital wart/cervical cancer vaccine to boys, followed by a cervical cancer vaccine to boys (Table 4). Within each of the three age groups, all of these differences in intention were highly significant (p ⬍ .001) except for the difference between intention to recommend a cervical cancer vaccine to girls and intention to recommend a cervical cancer/genital wart vaccine to boys, which did not differ significantly at age 11 or age 14. Although most differences in intention were statistically significant, actual differences in intention were not substantial with the exception of the difference between intention to vaccinate boys against a cervical cancer vaccine and all other categories. Thus, the main finding was that providers were particularly reluctant to recommend a cervical cancer vaccine to boys. In addition, participants were significantly more likely to recommend vaccines to older vs. younger patients for both genders and vaccine types. That is, for a given gender and vaccine type, they were significantly more likely to recommend a vaccine to a 14-year-old than an 11-year-old, to a 17-year-old than a 14-year-old, and to a 17-year-old than an 11-year-old (p ⬍ .0001 for all comparisons). Similarly, participants were significantly more likely to recommend vaccines to girls vs. boys for each age group and vaccine type (p ⬍ .0001 for all comparisons). Results of the multivariable analyses are shown in Table 5. Provider characteristics, practice characteristics, and attitudes about HPV vaccine were independently associated with intention to recommend a cervical cancer vaccine. These included age, male gender, percentage of 10 –15-

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

507

Table 5 Provider characteristics, practice characteristics, and attitudes independently associated with intention to administer HPV vaccines

␤ coefficient

Intention to recommend HPV vaccine

Independent variable

Vaccine effective against cervical cancer

Provider age Provider gender (male) Percentage of visits by 10–15-year-old patients that are for well-child care Estimated percentage of 15-year-olds in practice who are sexually active Likelihood of following the vaccination recommendations of influential organizations and individuals Number of children 10–15 years of age seen weekly in practice Knowledge about HPV Estimated percentage of 15-year-olds in practice who are sexually active Likelihood of following the vaccination recommendations of influential organizations and individuals Perceived barriers to immunization

Vaccine effective against cervical cancer and genital warts

year-old patients seen for routine health maintenance visits, estimate of the percentage of 15-year-old patients in one’s practice who are sexually active, and likelihood of following the vaccination recommendations of influential organizations or individuals (normative beliefs). Male gender and provider age were correlated significantly in univariate analyses; however, the interaction between gender and age was not significant. Practice characteristics, knowledge, and provider attitudes about HPV vaccine were independently associated with intention to recommend a cervical cancer/genital wart vaccine. These included number of 10 –15-year-old patients seen in one’s practice per week, knowledge about HPV, estimate of the percentage of 15-year-old patients in one’s practice who are sexually active, likelihood of following the vaccination recommendations of influential organizations or individuals, and fewer barriers to recommending immunization. Discussion In this study, we examined pediatricians’ attitudes about HPV vaccination and intention to recommend two hypothetical HPV vaccines to children and adolescents. Intention to recommend vaccination differed according to patient gender, patient age, and vaccine type. Pediatricians’ intention to recommend both HPV vaccines was markedly higher for older vs. younger patients, as demonstrated in recent studies examining attitudes about STI vaccination among gynecologists, nurse practitioners, and family physicians [18,19,22]. Personal reluctance to recommend vaccination to young adolescents and their parents may be linked to a number of clinician beliefs and attitudes. These clinicians may perceive young adolescents to be at low risk for HPV infection or may be hesitant to discuss issues related to sexuality with pread-

Standard error

p Value

.014 .258 .127

.004 .092 .051

.001 .006 .013

.008

.002

⬍.0001

.058

.018

.001

1.300

.531

.015

1.079 .084

.442 .027

.015 .002

.834

.240

.001

⫺.203

.063

.001

Model R2 .09

.10

olescents [27], particularly if they do not include these issues as a component of routine anticipatory guidance. They may be concerned that adolescents who receive STI vaccines will perceive sexual intercourse to be safer and perhaps practice riskier behaviors. Clinicians may also be concerned about negative parental reactions to a discussion of STIs with their child or adolescent. Clinician disinclination to vaccinate preadolescents and early adolescents could represent a significant barrier to effective HPV vaccine delivery, because prophylactic HPV vaccination should occur before sexual initiation. Approximately one-third of adolescents initiate sexual intercourse by ninth grade [28], and this group will be vulnerable to HPV infection and its sequelae if providers are hesitant to immunize early adolescents. Pediatricians were also more likely to intend to recommend vaccination to girls vs. boys at all ages. Family physicians also reported higher intention to recommend an HPV vaccine to girls than to boys [22], whereas nurse practitioners did not report higher intention to recommend STI vaccines to girls. It is possible that the attitudes of nurse practitioners participating in the latter study, who were recruited in two Midwestern cities during professional conferences, varied less than those of a national sample of physicians. Clinicians who are more likely to vaccinate girls than boys may believe that HPV vaccination will have a more significant impact on women’s health. However, these beliefs are inconsistent with evidence demonstrating that vaccinating both genders will be more effective in reducing the burden of HPV disease than vaccinating women only [29,30]. The findings also suggest that future research exploring further the reasons underlying the relative reluctance to vaccinate younger children and boys would be valuable, particularly for the

508

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

design of educational interventions targeted to providers that focus on STI vaccine delivery. Participants were more likely to recommend an HPV vaccine against cervical cancer and warts compared with a vaccine against cervical cancer only, especially for male patients. These attitudes are consistent with those reported by gynecologists and family physicians [19,22]. Physicians may consider a vaccine effective against both cervical cancer and genital warts as having a greater impact on morbidity and mortality in both men and women. Providers may also believe that parents of male patients or male adolescents would be more likely to accept an HPV vaccine that protects against genital warts in addition to cervical cancer because it provides a direct benefit to boys. Identification of modifiable factors associated with intention to recommend HPV vaccination may be helpful in the design of effective vaccine delivery programs. Theories of health behavior such as the Theory of Planned Behavior assume that specific factors, such as attitudes about and perceived control over a behavior, may be modifiable and also are causally associated with behaviors. Thus, interventions that aim to change salient attitudes may lead to behavioral change. Although some factors identified in our analyses are not modifiable (such as provider age, gender, and previous experience with adolescents), others are potentially modifiable and thus could be targeted in interventions to increase intention to recommend HPV vaccines or actual vaccine recommendations. For example, normative beliefs were associated with intention to recommend HPV immunization. Previous studies similarly have demonstrated that decisions to recommend STI immunization among nurse practitioners, obstetricians and gynecologists, and family physicians would be influenced by the recommendations of their professional organizations [18,19,22]. In fact, normative beliefs have been associated with a variety of physician behaviors, including recommendations for childhood vaccines [20,24]. In addition, knowledge about HPV and beliefs about patients’ sexual experience were associated with intention to recommend vaccination. Pediatric residents may receive little training about HPV, and practicing pediatricians may have less experience with HPV-related disease than family physicians or obstetricians and gynecologists. Those with less knowledge about HPV may not recognize the value of an HPV vaccine for adolescent health. Similarly, pediatricians who underestimate the percentage of their patients who are sexually active may believe their patients to be at relatively low risk for HPV infection and thus may be less likely to advocate for HPV immunization. Collectively, these findings suggest that effective communication about guidelines for immunization from trusted sources such as the AAP and CDC and about adolescent

risk for HPV infection could be powerful determinants of pediatricians’ intention to recommend HPV vaccines. Perceived barriers to immunization were also associated with intention to recommend HPV immunization, and included parental concerns about vaccines in general (e.g., concerns about vaccine safety and belief that children already receive too many vaccines) and STI vaccines specifically (e.g., immunization may lead to riskier sexual behaviors or reluctance to discuss issues of sexuality). Provider perceptions are supported by recent research demonstrating that although overall vaccine refusal rates are fairly low, parental concerns about immunization (particularly related to safety) have been associated with actual receipt of childhood vaccines [31– 33] as well as intended receipt of an HPV vaccine [34]. Nevertheless, research focusing on parental attitudes about STI vaccines suggests that most parents would approve of vaccination for their children, regardless of parent or child sociodemographic characteristics [35,36]. Furthermore, parents participating in one study were equally accepting of STI and non-STI vaccines [36]. Given that physician recommendation is one of the most important influences on a parent’s decision as to whether their child should be immunized against an STI [34], and that effective strategies for communication between providers and parents about the risks and benefits of non-STI and STI vaccines are being developed [37,38], it will be important to provide pediatricians with information about parental beliefs and effective educational tools about STI vaccines that can easily be incorporated into a busy practice. One limitation of this study is possible nonresponse bias: the attitudes of those who responded to the survey may differ from those who did not respond. Compared with all pediatricians in the United States, those who responded to this survey were slightly more likely to be female and younger on average (www.aap.org/workforce/ Peddemograph2003.ppt). However, the response rate was reasonable for a national survey of physicians and comparable to other studies using similar methodology [39 – 41]. Furthermore, physicians tend to be relatively homogeneous with respect to attitudes and behaviors, which may limit nonresponse bias [42]. A second limitation was that perceived barriers were self-reported: an examination of the association between actual barriers and vaccination behaviors should be performed after the vaccine is available for clinical use. Similarly, the outcome variable was intention to recommend an HPV vaccine rather than actual recommendation, which was not possible to assess because the vaccine is not yet available for clinical use. However, behavioral intention is a consistent predictor of actual behavior, including physician immunization behavior [20,23,24]. Furthermore, pediatricians demonstrated relatively low knowledge about HPV infection; higher knowledge may change their attitudes about vaccination. Finally, the R2 for both multi-

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

variable models was relatively low. Possible explanations include the following. First, there may be unmeasured factors influencing intention to recommend HPV vaccines, such as additional barriers to vaccination. Further study is underway to identify these factors. Second, there is a relative imbalance of both dispersion and variation in the independent variables that may decrease the magnitude of the R2. Third, as noted by Freedman, the magnitude of R2 depends on the number of parameters as well as the sample size [43]. When the number of explanatory variables is large relative to the sample size, and the variables are not well correlated, the beta coefficients and R2 may be highly significant, even if the independent and dependent variables are not associated. Using a method that Freedman proposed [43], we would expect an R2 of .005 in our study if the association between the independent variables and the outcome was due to chance. In contrast, the R2 of one model was .09 and of the other model was .10, and the p values for overall model fit were p ⬍ .001. In conclusion, vaccination initiatives focusing on modifiable predictors of intention to vaccinate may help to facilitate early acceptance of STI vaccines and adherence to emerging national immunization guidelines among pediatricians. Communication strategies should include education about HPV infection and its consequences, information about the safety and efficacy of vaccines, and recommendations for immunization by national organizations (particularly the AAP, CDC, and ACIP). Although guidelines are not yet available, it is likely that preadolescents and early adolescents will be targeted for vaccination, and possible that vaccination will be recommended for boys as well as girls [29,44]. If so, communication with providers should also include information about the importance of administering the vaccine to both boys and girls before sexual initiation to achieve the greatest public health impact.

Acknowledgments Dr. Kahn was supported by Grant #K23 AI50923-01 from the National Institute of Allergy and Infectious Diseases (NIAID). Mr. Riedesel was supported by Grant #T35 DK60444-02 from the National Institute of Diabetes and Digestive and Kidney Diseases. Dr. Zimet was supported by Grant #R01 AI49644 from NIAID. Dr. Rosenthal was supported by Grant #HD 40151 from the National Institute of Child Health and Human Development.

References [1] Tarkowski TA, Koumans EH, Sawyer M, et al. Epidemiology of human papillomavirus infection and abnormal cytologic test results in an urban adolescent population. J Infect Dis 2004;189:46 –50.

509

[2] Brown DR, Shew ML, Qadadri B, et al. A longitudinal study of genital human papillomavirus infection in a cohort of closely followed adolescent women. J Infect Dis 2005;191:182–92. [3] Beutner KR, Reitano MV, Richwald GA, Wiley DJ. External genital warts: report of the American Medical Association Consensus Conference. AMA Expert Panel on External Genital Warts. Clin Infect Dis 1998;27:796 – 806. [4] Bosch FX, Lorincz A, Munoz N, et al. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 2002;55: 244 – 65. [5] Koutsky LA, Ault KA, Wheeler CM, et al. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 2002;347: 1645–51. [6] Harper DM, Franco EL, Wheeler C, et al. Efficacy of a bivalent L1 virus-like particle vaccine in prevention of infection with human papillomavirus types 16 and 18 in young women: a randomised controlled trial. Lancet 2004;364:1757– 65. [7] Woodman CB, Collins S, Winter H, et al. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet 2001;357:1831– 6. [8] Hinman AR, Orenstein WA, Mortimer EA. When, where, and how do immunizations fail? Ann Epidemiol 1992;2:805–12. [9] Zimet GD, Mays RM, Fortenberry JD. Vaccines against sexually transmitted infections: promise and problems of the magic bullets for prevention and control. Sex Transm Dis 2000;27:49 –52. [10] Ziv A, Boulet JR, Slap GB. Utilization of physician offices by adolescents in the United States. Pediatrics 1999;104:35– 42. [11] Freed GL, Bordley WC, Clark SJ, Konrad TR. Universal hepatitis B immunization of infants: reactions of pediatricians and family physicians over time. Pediatrics 1994;93:747–51. [12] Loewenson PR, White KE, Osterholm MT, MacDonald KL. Physician attitudes and practices regarding universal infant vaccination against hepatitis B infection in Minnesota: implications for public health policy. Pediatr Infect Dis J 1994;13:373– 8. [13] Freed GL, Freeman VA, Clark SJ, et al. Pediatrician and family physician agreement with and adoption of universal hepatitis B immunization. J Fam Pract 1996;42:587–92. [14] Kahane SM, Watt JP, Newell K, et al. Immunization levels and risk factors for low immunization coverage among private practices. Pediatrics 2000;105:E73. [15] Koepke CP, Vogel CA, Kohrt AE. Provider characteristics and behaviors as predictors of immunization coverage. Am J Prev Med 2001;21:250 –5. [16] Prislin R, Sawyer MH, De Guire M, et al. Missed opportunities to immunize: psychosocial and practice correlates. Am J Prev Med 2002;22:165–9. [17] Guajardo AD, Middleman AB, Sansaricq KM. School nurses identify barriers and solutions to implementing a school-based hepatitis B immunization program. J Sch Health 2002;72:128 –30. [18] Mays RM, Zimet GD. Recommending STI vaccination to parents of adolescents: the attitudes of nurse practitioners. Sex Transm Dis 2004;31:428 –32. [19] Raley JC, Followwill KA, Zimet GD, Ault KA. Gynecologists’ attitudes regarding human papillomavirus vaccination: a survey of fellows of the American College of Obstetricians and Gynecologists. Infect Dis Obstet Gynecol 2004;12:127–33. [20] Prislin R, Nader PR, De Guire M, et al. Physician’s immunization knowledge, attitudes and practices. A valid and internally consistent measurement tool. Am J Prev Med 1999;17:151–2. [21] Kahn JA, Rosenthal SL, Hamann T, Bernstein DI. Attitudes about human papillomavirus vaccine in young women. Int J STD AIDS 2003;14:300 – 6. [22] Riedesel JM, Rosenthal SL, Zimet GD, et al. Family physicians’ attitudes about HPV vaccines. J Pediatr Adolesc Gynecol 2005; In press.

510

J.A. Kahn et al. / Journal of Adolescent Health 37 (2005) 502–510

[23] Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process 1991;50:179 –211. [24] Millstein SG. Utility of the theories of reasoned action and planned behavior for predicting physician behavior: a prospective analysis. Health Psychol 1996;15:398 – 402. [25] Godin G, Kok G. The theory of planned behavior: a review of its applications to health-related behaviors. Am J Health Promot 1996; 11:87–98. [26] Sheeran P, Orbell S. Do intentions predict condom use? Metaanalysis and examination of six moderator variables. Br J Soc Psychol 1998;37:231–50. [27] Schuster MA, Bell RM, Petersen LP, Kanouse DE. Communication between adolescents and physicians about sexual behavior and risk prevention. Arch Pediatr Adolesc Med 1996;150:906 –13. [28] U.S. Department of Health and Human Services. Trends in the WellBeing of America’s Children and Youth. Washington, DC: U.S. Department of Health and Human Services, 2003. [29] Hughes JP, Garnett GP, Koutsky L. The theoretical population-level impact of a prophylactic human papilloma virus vaccine. Epidemiology 2002;13:631–9. [30] Garnett GP, Dubin G, Slaoui M, Darcis T. The potential epidemiological impact of a genital herpes vaccine for women. Sex Transm Infect 2004;80:24 –9. [31] Bardenheier B, Yusuf H, Schwartz B, et al. Are parental vaccine safety concerns associated with receipt of measles-mumps-rubella, diphtheria and tetanus toxoids with acellular pertussis, or hepatitis B vaccines by children? Arch Pediatr Adolesc Med 2004;158:569 –75. [32] Gust DA, Strine TW, Maurice E, et al. Underimmunization among children: effects of vaccine safety concerns on immunization status. Pediatrics 2004;114:e16 –22. [33] Freed GL, Clark SJ, Hibbs BF, Santoli JM. Parental vaccine safety concerns. The experiences of pediatricians and family physicians. Am J Prev Med 2004;26:11– 4.

[34] Davis K, Dickman ED, Ferris D, Dias JK. Human papillomavirus vaccine acceptability among parents of 10- to 15-year-old adolescents. Obstet Gynecol Surv 2004;59:820 –2. [35] Mays RM, Sturm LA, Zimet GD. Parental perspectives on vaccinating children against sexually transmitted infections. Soc Sci Med 2004;58:1405–13. [36] Zimet GD, Mays RM, Sturm LA, et al. Parental attitudes about sexually transmitted infection vaccination for their adolescent children. Arch Pediatr Adolesc Med 2005;159:132–7. [37] Davis TC, Fredrickson DD, Bocchini C, et al. Improving vaccine risk/benefit ommunication with an immunization education package: a pilot study. Ambul Pediatr 2002;2:193–200. [38] Davis TC, Fredrickson DD, Kennen EM, et al. Childhood vaccine risk/benefit communication among public health clinics: a timemotion study. Public Health Nurs 2004;21:228 –36. [39] Cummings SM, Savitz LA, Konrad TR. Reported response rates to mailed physician questionnaires. Health Serv Res 2001;35:1347–55. [40] Balk SJ, O’Connor KG, Saraiya M. Counseling parents and children on sun protection: a national survey of pediatricians. Pediatrics 2004; 114:1056 – 64. [41] Sices L, Feudtner C, McLaughlin J, et al. How do primary care physicians manage children with possible developmental delays? A national survey with an experimental design. Pediatrics 2004;113: 274 – 82. [42] Kellerman SE, Herold J. Physician response to surveys. A review of the literature. Am J Prev Med 2001;20:61–7. [43] Freedman DA. A note on screening regression equations. Am Stat 1983;37:152–5. [44] Goldie SJ, Kohli M, Grima D, et al. Projected clinical benefits and cost-effectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst 2004;96:604 –15.