Contribution of Communication Inequalities to Disparities in Human Papillomavirus Vaccine Awareness and Knowledge

Abstract

Objectives. We examined the association of Internet-related communication inequalities on human papillomavirus (HPV) vaccine awareness and infection knowledge.

Methods. We drew data from National Cancer Institute’s 2007 Health Information National Trends Survey (n = 7674). We estimated multivariable logistic regression models to assess Internet use and Internet health information seeking on HPV vaccine awareness and infection knowledge.

Results. Non–Internet users, compared with general Internet users, had significantly lower odds of being aware of the HPV vaccine (odds ratio [OR] = 0.42; 95% confidence interval [CI] = 0.34, 0.51) and knowing that HPV causes cervical cancer (OR = 0.70; 95% CI = 0.52, 0.95). Among general health information seekers, non–Internet seekers compared with Internet information seekers exhibit significantly lower odds of HPV vaccine awareness (OR = 0.59; 95% CI = 0.46, 0.75), and of knowing about the link between HPV infection and cervical cancer (OR = 0.79; 95% CI = 0.63, 0.99) and the sexual transmission of HPV (OR = 0.71; 95% CI = 0.57, 0.89). Among cancer information seekers, there were no differences in outcomes between Internet seekers and non–Internet seekers.

Conclusions. Use of a communication channel, such as the Internet, whose use is already socially and racially patterned, may widen observed disparities in vaccine completion rates.

In 2006, the Food and Drug Administration approved the first vaccine for the 4 major types of human papillomavirus (HPV), 2 of which cause approximately 70% of cervical cancer cases in the United States.¹ Many health officials and researchers have indicated that appropriate uptake of the vaccine could serve as a way to eliminate disparities in cervical cancer rates seen throughout the United States.² However, according to the Centers for Disease Control and Prevention’s (CDC’s) most recent National Immunization Survey–Teen, only 37% of US female adolescents aged 13 to 17 years initiated the HPV vaccine series (≥ 1 dose) in 2008 and a mere 18% of teens completed the necessary 3-dose series.³

More important than the overall low series completion rates nationally are the noted disparities across racial and social groups. Several large-scale studies, including the 2008 National Immunization Survey, have reported that, although African American and Hispanic girls are more likely to initiate the series, they are less likely to complete the 3-dose series compared with their White counterparts.^4–6 In addition, reports indicate that girls living in poorer states and in neighborhoods with low education levels, and those who rely on public insurance are also less likely to complete the vaccine series compared with girls living in wealthier states and in neighborhoods with higher aggregate education levels, and those who are covered by private insurance.^4,6,7 If these racial and social disparities in completion rates continue, they may further exacerbate what is already an unequal burden of cervical cancer among these groups of women.

We suggest that the discrepancy between initiation and completion rates may in part be attributable to the well-documented knowledge gaps among lower–socioeconomic status (SES) and minority parents, caregivers, and other influential adults about what HPV is, how it is spread, and how it can be prevented across these vulnerable groups.^8–10 The initiation shot typically occurs during a well-child visit along with a host of other vaccinations, raising the issue as to whether patients and guardians are truly informed of what vaccinations they are receiving, but completion shots need to be separately scheduled. Knowledge and awareness therefore may play a larger role in the decision-making process as to whether a child or adolescent will complete the series because they will occur outside the normal doctor’s visit.

What is less known, and what needs to be examined to reduce the disparities in completion rates, are the driving factors behind these knowledge gaps. We argue that communication is one thread that could potentially connect the different levels of social determinants that could ultimately explain individual health status and population-level health because of the different functions it serves.^11–13 If so, communication inequalities—differences among social groups in accessing, seeking, processing, and using health information—could play an important role in shaping HPV-related knowledge gaps, which may, in turn, influence vaccine completion rates.¹¹ In the case of HPV, there are a number of communication inequalities that warrant examination. For example, disparities in completion rates may in part be influenced by a breakdown in patient–provider communication. However, there is a substantial body of evidence indicating the ever-growing reliance on the Internet, over physicians, as an important source of health information.¹¹ Therefore, we offer that HPV-related knowledge gaps and communication inequalities may primarily be a product of (1) the vaccine manufacturer’s reliance on Internet-based direct-to-consumer advertising (DTCA) as the main information delivery system for both the infection and the vaccine, combined with (2) differential access and use of the Internet among those of low SES and minority populations.

Merck’s Gardasil is one of the first cancer vaccines for prophylactic use, and as such, they created a lucrative market with revenue estimates close to $3.2 billion by the year 2012.¹⁴ Because Merck has a large financial stake in the success of the vaccine, industry analysts report that an overwhelming majority of the messages and information that the public receives about HPV and the vaccine are somehow linked to the pharmaceutical developer.^14,15 If one considers how limited public health funds are, there is a perceived benefit of having a large pharmaceutical company with a substantial marketing budget take on the burden of educating the public about the harmful effects of the disease while advertising its product.

The goal of print and television DTCA, however, is to make consumers aware of the brand name of a drug and to drive them to perform external searching or, rather, information seeking, which is the purposive and deliberate seeking of information on a topic because of its salience and need.¹⁶ External information seeking is recognized in the health marketing industry as the way consumers can expand their knowledge of the drugs represented.¹⁷ Moreover, research shows that consumers are increasingly turning to the Internet, over other channels, to search for more information about a drug they saw on television.^18,19 With increased advertising spending, concerns have been raised that print and TV DTCA may mislead consumers with unbalanced and incomplete information making the reliance on external searching, specifically Internet information seeking, even more crucial.¹⁹

Unfortunately, reliance on Internet-based external searches to provide needed information for informed decision-making on HPV vaccination could only exacerbate existing knowledge gaps. Research has shown that education is strongly associated with consumer information seeking as well as with time spent looking for the information.²⁰ Research also shows that DTCA is not appropriate for low-literacy audiences as the information provided is too complex and written at a high reading grade level.^21–23 There is the additional consideration that many low-SES and minority groups lack Internet access as well as the needed computer literacy skills to adequately search the Internet for health-related information.²⁴ Recent analyses of the National Cancer Institute’s Health Information National Trends Survey (HINTS) 2007 offer further evidence of digital disparities. Although overall penetration of Internet access in the United States has increased to 70%, the digital divide remains with nearly half of Hispanics and non-Hispanic Blacks lacking Internet access even after controlling for other important variables. The same is true across socioeconomic strata with adults with lower education levels and income having significantly lower odds of Internet access compared with their higher-SES peers.²⁵ Additional research on the 2007 HINTS data also highlight disparities in online health information seeking with lower odds of using the Internet as a first source of health information among those with lower incomes and education levels, and for Blacks compared with Whites.²⁶ Moreover, limited computer literacy skills could potentially interact with limited health literacy skills among low-SES and minority groups to not only inhibit the initiation of Internet-based health information seeking but also result in procuring misinformation because of improper Web searching. This misinformation could negatively impact HPV and vaccine knowledge as well vaccination uptake and completion rates.

Our goal was to examine the association between Internet-based communication and knowledge of the HPV vaccine with the hope of identifying more effective approaches to improve completion rates among underserved groups. In light of the Internet-based communication strategy for the HPV vaccine emphasized by Merck, we hypothesized that those individuals and groups that do not rely on the Internet for health information are likely to have lower levels of awareness and knowledge of both HPV and the HPV vaccine.

METHODS

We drew the data for this study from the National Cancer Institute’s 2007 HINTS. HINTS is focused on reporting the cancer communication behaviors and trends of American adults based on a survey of nationally representative, noninstitutionalized adults, and has been fielded every 2 to 3 years beginning in 2003. HINTS 2007 data were collected from January 2008 through May 2008. The HINTS is a public use data set and the questionnaire can be found at: http://hints.cancer.gov. HINTS 2007 consisted of 2 sample frames. One sample was drawn as a random digit dial (RDD) telephone survey, which used a computer-assisted telephone interview (CATI) format. Data were collected from 4092 respondents via CATI. The second national random sample was selected from a list of addresses from the US Postal Service administrative records. Because of the tendency for mail respondents to represent nonminority groups and to have characteristics associated with higher SES,²⁷ the mail survey included a stratified sample selected from a list of addresses that oversampled for minorities. Data were collected from 3582 respondents via mail. The National Cancer Institute chose the dual frame based on research by Link et al.,²⁸ which suggested that use of a mail survey, with appropriate follow-up, can achieve a higher response rate than RDD alone; this design was also adopted by the CDC’s Behavioral Risk Factor Surveillance Survey. The response rates for HINTS 2007 were 31% for the mail sample and 24% for the RDD sample.²⁹

Measures

Research has shown that purposive, active seeking of health information on a given topic as well as incidental exposure to health information through one’s routine use of media are correlated with health knowledge.^13,16 Therefore, we measured the impact of both general Internet use as well as Internet health information seeking on our outcomes. We measured general Internet use by the question “Do you ever go on-line to access the Internet or World Wide Web, or to send and receive email?” We then measured Internet health information seeking through 2 dimensions: general and cancer-specific information seeking on the Internet. We measured general health Internet information seeking by the questions “Have you ever looked for information about health or medical topics from any source?” and “The most recent time you looked for information about health or medical topics, where did you go first?” We measured cancer-specific Internet seeking by the questions “Have you ever looked for information about cancer from any source” and “The most recent time you looked for information about cancer, where did you go first?” For both general and cancer-specific seeking, we classified respondents into 3 categories (1) Internet seekers, (2) non–Internet seekers, and (3) nonseekers. Internet seekers reported “yes” to either seeking question and identified the Internet as their source of information. Non–Internet seekers reported “yes” to either seeking question but identified another source for their information such as books, magazine, family or friend, or doctor. Finally, respondents who reported “no” to either seeking question were classified as nonseekers.

To assess HPV knowledge and HPV vaccine awareness, we included 3 dichotomous outcome variables in all analyses. We measured HPV vaccine awareness by the question

A vaccine to prevent HPV infection is available and is called the cervical cancer vaccine or HPV shot. Before today, have you ever heard of the cervical cancer vaccine or HPV shot?

We measured HPV knowledge by 2 questions, “Do you think HPV can cause cervical cancer?” and “Do you think you can get HPV through sexual contact?” Respondents who reported not knowing about HPV were not asked the knowledge questions.

We included several sociodemographic variables and covariates identified in other studies as being influential on HPV knowledge as well as Internet use and health seeking behaviors in all analyses. These variables included age, household income, education, race/ethnicity, employment status, sex, HPV status, having a daughter, immigrant status, insurance status, having a regular physician, and general health status.^8,30–32

Data Analysis

We analyzed data from both survey frames (mail and RDD) and included the entire adult population in our analyses (n = 7674). We did not limit our analyses to parents or guardians of girls younger than 18 years for 2 reasons. First, studies have shown that knowledge, attitudes, and beliefs of family and friends are influential on one’s own understanding as well as on actual behaviors.^33–36 Therefore, HPV awareness and knowledge held by grandparents, aunts, uncles, and other relatives and family friends can play an important role in vaccination rates and should be measured. Second, the HPV vaccine is now being marketed to a larger age spectrum among females as well as to young males warranting a broader assessment of knowledge among the US adult population.

We conducted analyses with SAS version 9.2 (SAS Institute, Cary, NC) to properly calculate standard errors to accommodate the multistage sampling design. We excluded observations with missing data for any of the key variables from analysis, except for observations with missing income data. Because there was a substantial number of observations with missing income data (n = 500) we created a separate category for these respondents and included it in all analyses. We re-estimated the sampling weights as recommended by the National Cancer Institute and included survey mode as a covariate in each regression model to account for mode effects. Detailed descriptions of how the sample and replicate weights were calculated can be found in the HINTS 2007 Final Report.²⁹

We first examined descriptive statistics of the 3 outcomes of interest: HPV vaccine awareness, HPV as a cause of cervical cancer, and HPV contracted through sexual contact among the entire survey population, and by our key predictor variables: general Internet user, general health Internet information seeker, and cancer-specific Internet seeker. We then estimated the unadjusted associations (χ²) between the predictors of interest and our knowledge outcomes. Finally, we re-estimated these associations by using multivariable logistic regression techniques with control for spurious relationships by including all listed confounders and covariates into the regression models.

RESULTS

Fifty-one percent of the HINTS sample was female and 49% was male. Thirty-one percent of the survey population was aged 18 to 34 years, 30% was aged 35 to 49 years, and 39% was aged 50 years and older. Non-Hispanic Whites represented 65% of the sample, 11% were non-Hispanic Black, 12% were Hispanic, and 12% were of other racial and ethnic background. Thirty percent of the sample reported yearly household incomes of less than $35 000, 28% earned $35 000 to $74 999, 25% earned $75 000 or more, and 17% of the sample did not report their income. More than 40% of HINTS respondents reported a high-school or less education, 35% obtained some college education, and 25% reported a college degree.

Descriptive and Bivariate Results

In relation to the main outcomes of interest for this study, nearly 70% of HINTS respondents had heard of the HPV vaccine and were knowledgeable that HPV can be spread via sexual contact. Three fourths of the population (75%) knew of the link between HPV infection and cervical cancer (Table 1). However, there were noted disparities in awareness and knowledge across Internet usage and Internet information seeking as well as sociodemographic characteristics.

TABLE 1—

Weighted Frequencies and Bivariate Analyses Results Regarding Human Papillomavirus Awareness and Knowledge: National Cancer Institute’s 2007 Health Information National Trends Survey

	Aware of HPV Vaccine		Knowledgeable That HPV Is a Cause of Cervical Cancer		Knowledgeable That HPV Is Transmitted Through Sexual Contact
	%	P	%	P	%	P
Total	68		75		68
Internet use		<.001		<.001		<.001
Nonuser	42		66		58
User	75		81		70
General health information seeking		<.001		<.001		<.001
Internet seeker	80		84		74
Non–Internet seeker	61		79		68
Nonseeker	46		64		50
Cancer-specific health information seeking		<.001		<.001		<.001
Internet seeker	82		84		75
Non–Internet seeker	71		80		73
Nonseeker	57		75		61
Household income, $		<.001		<.001		<.001
< 20 000	50		81		62
20 000–34 999	58		73		65
35 000–49 999	67		80		68
50 000–74 999	72		81		65
≥ 75 000	75		82		75
Did not report	59		67		61
Education		<.001		<.001		<.001
< high school	41		71		50
High school graduate	57		73		57
Some college	72		79		70
≥ college degree	76		85		72
Race/ethnicity		<.001		.02		.01
Non-Hispanic White	72		80		69
Non-Hispanic Black	55		72		57
Hispanic	45		81		63
Other	50		71		68
Sex		<.001		<.001		.14
Male	51		75		65
Female	77		81		69
HPV status		<.001		<.001		<.001
Ever	95		94		88
Never	64		78		67
Children		.04		.05		.34
None	63		77		68
≥ 1 daughter	69		81		65
No daughter	65		82		68

Note. HPV = human papillomavirus. P values determined by χ² test.

Only 42% of non–Internet users reported awareness of the HPV vaccine compared with 75% of Internet users. Two thirds (66%) of non–Internet users compared with 81% of Internet users were knowledgeable that HPV can cause cervical cancer. A little more than half (58%) of non–Internet users knew that HPV could be transmitted via sexual contact compared with 70% of Internet users (Table 1).

Adults who did not seek out general health information (nonseekers) reported the lowest levels of vaccine awareness and infection knowledge. Only about half had heard of the HPV vaccine (46%) and knew that the infection could be transmitted via sexual contact (50%) and 64% knew that HPV could cause cervical cancer. Among general health information seekers, Internet seekers reported higher levels of awareness and knowledge compared with non–Internet seekers. Eighty percent of Internet seekers had heard of the HPV vaccine, 84% knew that HPV could cause cervical cancer, and 74% knew that the infection could be contracted through sexual contact (Table 1). On the other hand, only 61% of non–Internet seekers were aware of the vaccine, 79% knew of the link between HPV and cervical cancer, and only 68% were aware of the infection’s transmission (Table 1).

We saw a similar trend among cancer-specific information seekers with nonseekers reporting the lowest levels of knowledge and awareness. Only 57% cancer-specific nonseekers had heard of the vaccine, 75% knew of the link between HPV and cervical cancer, and 61% knew that the infection could be transmitted via sexual contact. Among cancer-specific seekers, Internet seekers again reported the highest levels of awareness and knowledge. Eighty-two percent of cancer-specific Internet seekers had heard of the HPV vaccine, 84% knew that HPV causes cervical cancer, and 75% knew that the infection could spread through sexual contact. By comparison, only 71% of non-Internet cancer-specific seekers were aware of the HPV vaccine, 80% knew of the link between HPV and cervical cancer, and 73% knew of the infection’s transmission (Table 1).

Increased levels of vaccine awareness and infection knowledge were reported among adults with higher levels of education and income (Table 1). For example, only half of adults whose household income was less than $20 000 had heard of the HPV vaccine compared with three fourths of those whose household income was $75 000 or greater. Less than half (41%) of adults with less than a high-school degree were aware of the vaccine whereas 76% of college-educated adults knew of the vaccine. Non-Hispanic Whites reported the highest levels of vaccine awareness (72%) and knowledge (80% link with cervical cancer; 69% sexually transmitted) whereas non-Hispanic Blacks reported the lowest levels of knowledge (72% link with cervical cancer; 57% sexually transmitted). Hispanics had the lowest level of vaccine awareness (45%) but high levels of infection knowledge (81% link with cervical cancer; 63% sexually transmitted).

More women than men were aware of the vaccine and knew that HPV could cause cervical cancer; however, there were no differences by sex in knowing that HPV can be transmitted through sexual contact. As would be expected, respondents with a history of HPV reported very high levels of awareness and knowledge across all outcome variables (Table 1). In addition, respondents who reported having at least 1 female child younger than 18 years living in their household were also more likely to be aware of the vaccine and know that HPV can cause cervical cancer; yet, they were not any more likely to know how the infection is transmitted (Table 1).

Multivariable Logistic Regression Results

Non–Internet users, compared with users, had nearly half the odds (odds ratio [OR] = 0.42; 95% CI = 0.34, 0.51) of having heard about the HPV vaccine after we controlled for important sociodemographic characteristics including age, household income, education, race/ethnicity, employment status, sex, HPV status, having a daughter, immigrant status, insurance status, having a regular physician, general health status, and survey mode. Non–Internet users also had 30% lower odds (OR = 0.70; 95% CI = 0.52, 0.95) of knowing that HPV is a cause of cervical cancer compared with Internet users. There was no significant difference between Internet use and nonuse in reported knowledge of HPV being contracted through sexual contact (Table 2).

TABLE 2—

Multivariable Logistic Regression of the Association Between Internet Use and HPV Vaccine Awareness and HPV Knowledge: National Cancer Institute’s 2007 Health Information National Trends Survey

	Aware of HPV Vaccine (n = 6853)		Knowledgeable That HPV Is a Cause of Cervical Cancer (n = 4483)		Knowledgeable That HPV Is Transmitted Through Sexual Contact (n = 4497)
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Internet use		<.001		.02		.64
Nonuser	0.42 (0.34, 0.51)		0.70 (0.52, 0.95)		1.07 (0.77, 1.45)
User (Ref)	1.00		1.00		1.00
Household income, $		<.001		.45		.3
< 20 000	0.70 (0.49, 1.01)		1.24 (0.78, 1.98)		1.14 (0.75, 1.72)
20 000–34 999	0.85 (0.58, 1.24)		0.89 (0.58, 1.35)		1.06 (0.72, 1.55)
35 000–49 999 (Ref)	1.00		1.00		1.00
50 000–74 999	1.16 (0.82, 1.64)		1.26 (0.81, 1.96)		1.04 (0.73, 1.49)
≥ 75 000	1.01 (0.74, 1.38)		1.12 (0.76, 1.65)		1.38 (0.95, 2.00)
Did not report	0.61 (0.44, 0.85)		0.97 (0.61, 1.56)		1.05 (0.73, 1.52)
Education		<.001		<.001		<.001
< high school	0.80 (0.55, 1.15)		0.81 (0.47, 1.41)		0.64 (0.41, 0.99)
High school graduate (Ref)	1.00		1.00		1.00
Some college	1.36 (1.11, 1.68)		0.93 (0.68, 1.26)		1.46 (1.08, 1.96)
≥ college degree	1.51 (1.20, 1.91)		1.69 (1.21, 2.36)		2.26 (1.70, 3.00)
Race/ethnicity		<.001		<.001		.05
Non-Hispanic White (Ref)	1.00		1.00		1.00
Non-Hispanic Black	0.64 (0.49, 0.83)		0.54 (0.35, 0.83)		0.61 (0.42, 0.88)
Hispanic	0.56 (0.37, 0.84)		1.26 (0.63, 2.51)		1.12 (0.67, 1.84)
Other	0.37 (0.25, 0.54)		0.54 (0.34, 0.87)		0.88 (0.52, 1.49)
Sex		<.001		.03		.56
Male (Ref)	1.00		1.00		1.00
Female	3.99 (3.33, 4.78)		1.31 (1.03, 1.68)		1.07 (0.85, 1.36)
HPV status		.34		<.001		<.001
Ever	2.12 (0.46, 9.75)		6.91 (1.82, 26.23)		5.44 (2.18, 13.58)
Never (Ref)	1.00		1.00		1.00
Children		.08		.65		.56
None	0.82 (0.64, 1.06)		0.85 (0.61, 1.20)		1.07 (0.82, 1.40)
≥ 1 daughter (Ref)	1.00		1.00		1.00
No daughter	0.66 (0.46, 0.95)		0.88 (0.56, 1.39)		1.21 (0.86, 1.71)

Note. CI = confidence interval; HPV = human papillomavirus; OR = odds ratio. Also in model: age; employment; immigrant, insurance, and health status; having a regular physician; and survey mode. P values determined by adjusted Wald test.

Non-Internet general health information seeking and nonseeking were associated with lower odds of HPV vaccine awareness and HPV knowledge (Table 3). Those respondents who actively looked for health information but reported using a non-Internet source such as a newspaper, magazine, or interpersonal source, had 41% (95% CI = 0.46, 0.75) lower odds of having heard of the HPV vaccine, 21% (95% CI = 0.63, 0.99) lower odds of knowing that HPV is a cause of cervical cancer, and 29% (95% CI = 0.57, 0.89) lower odds of knowing that HPV can be transmitted via sexual contact compared with Internet seekers, when we controlled for the sociodemographic characteristics listed previously. Nonseeking was associated with the lowest odds of knowledge with nonseekers reporting 66% (95% CI = 0.27, 0.44) lower odds of having heard of the HPV vaccine, 45% (95% CI = 0.39, 0.76) lower odds of knowing that HPV is a cause of cervical cancer, and 54% (95% CI = 0.34, 0.63) lower odds of knowing that the infection can be transmitted via sexual contact, compared with Internet seekers.

TABLE 3—

Multivariable Logistic Regression of the Association Between General Health Information Seeking Via the Internet and HPV Vaccine Awareness and HPV Knowledge: National Cancer Institute’s 2007 Health Information National Trends Survey

	Aware of HPV Vaccine (n = 6820)		Knowledgeable That HPV Is a Cause of Cervical Cancer (n = 4458)		Knowledgeable That HPV Is Transmitted Through Sexual Contact (n = 4471)
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
General health information seeking		<.001		<.001		<.001
Internet seeker (Ref)	1.00		1.00		1.00
Non–Internet seeker	0.59 (0.46, 0.75)		0.79 (0.63, 0.99)		0.71 (0.57, 0.89)
Nonseeker	0.34 (0.27, 0.44)		0.55 (0.39, 0.76)		0.46 (0.34, 0.63)
Household income, $		<.001		.57		.5
< 20 000	0.67 (0.47, 0.95)		1.19 (0.75, 1.92)		1.22 (0.80, 1.86)
20 000–34 999	0.86 (0.59, 1.24)		0.85 (0.55, 1.32)		1.08 (0.73, 1.59)
35 000–49 999 (Ref)	1.00		1.00		1.00
50 000–74 999	1.20 (0.85, 1.71)		1.20 (0.77, 1.89)		0.98 (0.69, 1.41)
≥ 75 000	1.04 (0.76, 1.42)		1.08 (0.73, 1.59)		1.30 (0.90, 1.86)
Did not report	0.63 (0.45, 0.87)		0.98 (0.60, 1.59)		1.07 (0.74, 1.55)
Education		<.001		<.001		<.001
< high school	0.77 (0.54, 1.11)		0.80 (0.47, 1.38)		0.70 (0.45, 1.09)
High school graduate (Ref)	1.00		1.00		1.00
Some college	1.31 (1.05, 1.64)		0.88 (0.63, 1.22)		1.27 (0.93, 1.73)
≥ college degree	1.44 (1.13, 1.83)		1.54 (1.10, 2.16)		1.82 (1.37, 2.42)
Race/ethnicity		<.001		<.001		.11
Non-Hispanic White (Ref)	1.00		1.00		1.00
Non-Hispanic Black	0.63 (0.47, 0.84)		0.54 (0.34, 0.85)		0.65 (0.45, 0.94)
Hispanic	0.53 (0.36, 0.80)		1.27 (0.64, 2.52)		1.17 (0.73, 1.88)
Other	0.37 (0.25, 0.55)		0.53 (0.33, 0.83)		0.88 (0.54, 1.44)
Sex		<.001		.06		.91
Male (Ref)	1.00		1.00		1.00
Female	3.80 (3.17, 4.56)		1.27 (0.99, 1.63)		1.02 (0.80, 1.29)
HPV status		.4		.01		<.001
Ever	1.84 (0.44, 7.74)		6.48 (1.69, 24.77		4.68 (1.89, 11.61)
Never (Ref)	1.00		1.00		1.00
Children		.11		.67		.5
None	0.83 (0.64, 1.07)		0.85 (0.60, 1.22)		1.09 (0.84, 1.43)
≥ 1 daughter (Ref)	1.00		1.00		1.00
No daughter	0.67 (0.46, 0.98)		0.88 (0.55, 1.40)		1.23 (0.87, 1.73)

Note. CI = confidence interval; HPV = human papillomavirus; OR = odds ratio. Also in model: age; employment; immigrant, insurance, and health status; having a regular physician; and survey mode. P values determined by adjusted Wald test.

There were no significant differences in HPV vaccine awareness or HPV knowledge between cancer-specific information Internet seekers and non–Internet seekers after we controlled for sociodemographic characteristics (Table 4). However, being a cancer-specific nonseeker (respondents who reported never having looked for information on cancer) was associated with lower odds of both vaccine awareness and infection knowledge after we controlled for important covariates. Cancer-specific nonseekers had 51% lower odds (95% CI = 0.37, 0.63) of having heard about the HPV vaccine, 36% (95% CI = 0.48, 0.86) lower odds of knowing that HPV is a cause of cervical cancer, and 39% (95% CI = 0.46, 0.80) lower odds of knowing that HPV can be transmitted via sexual contact compared with cancer-specific Internet seekers.

TABLE 4—

Multivariable Logistic Regression of the Association Between Cancer-Specific Health Information Seeking Via the Internet and HPV Vaccine Awareness and HPV Knowledge: National Cancer Institute’s 2007 Health Information National Trends Survey

	Aware of HPV Vaccine (n = 6777)		Knowledgeable That HPV Is a Cause of Cervical Cancer (n = 4427)		Knowledgeable That HPV Is Transmitted Through Sexual Contact (n = 4440)
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Cancer-specific health information seeking		<.001		<.001		<.001
Internet seeker (Ref)	1.00		1.00		1.00
Non–Internet seeker	0.90 (0.68, 1.18)		1.02 (0.73, 1.43)		0.99 (0.70, 1.40)
Nonseeker	0.49 (0.37, 0.63)		0.64 (0.48, 0.86)		0.61 (0.46, 0.80)
Household income, $		<.001		.48		.47
< 20 000	0.64 (0.44, 0.92)		1.18 (0.75, 1.85)		1.18 (0.78, 1.77)
20 000–34 999	0.80 (0.54, 1.17)		0.88 (0.58, 1.35)		1.10 (0.75, 1.61)
35 000–49 999 (Ref)	1.00		1.00		1.00
50 000–74 999	1.26 (0.88, 1.81)		1.27 (0.82, 1.98)		1.04 (0.73, 1.50)
≥ 75 000	1.11 (0.80, 1.54)		1.15 (0.78, 1.70)		1.36 (0.94, 1.97)
Did not report	0.62 (0.44, 0.87)		0.99 (0.61, 1.60)		1.08 (0.75, 1.55)
Education		<.001		<.001		<.001
< high school	0.73 (0.51, 1.05)		0.82 (0.48, 1.41)		0.69 (0.45, 1.06)
High school graduate (Ref)	1.00		1.00		1.00
Some college	1.44 (1.17, 1.78)		0.91 (0.66, 1.26)		1.41 (1.05, 1.89)
≥ college degree	1.67 (1.34, 2.07)		1.64 (1.16, 2.31)		2.10 (1.59, 2.78)
Race/ethnicity		<.001		<.001		.11
Non-Hispanic White (Ref)	1.00		1.00		1.00
Non-Hispanic Black	0.61 (0.46, 0.82)		0.54 (0.35, 0.84)		0.65 (0.45, 0.94)
Hispanic	0.50 (0.33, 0.75)		1.20 (0.59, 2.44)		1.12 (0.68, 1.86)
Other	0.36 (0.24, 0.53)		0.51 (0.33, 0.80)		0.85 (0.52, 1.38)
Sex		<.001		.03		.78
Male (Ref)	1.00		1.00		1.00
Female	3.92 (3.23, 4.74)		1.32 (1.03, 1.69)		1.04 (0.82, 1.31)
HPV status		.41		.01		<.001
Ever	1.93 (0.40, 9.22)		6.27 (1.68, 23.35)		4.86 (1.92, 12.30)
Never (Ref)	1.00		1.00		1.00
Children		.09		.7		.46
None	0.81 (0.63, 1.05)		0.86 (0.61, 1.22)		1.07 (0.82, 1.39)
≥ 1 daughter (Ref)	1.00		1.00		1.00
No daughter	0.67 (0.46, 0.97)		0.89 (0.56, 1.41)		1.24 (0.88, 1.75)

Note. CI = confidence interval; HPV = human papillomavirus; OR = odds ratio. Also in model: age; employment; immigrant, insurance, and health status; having a regular physician; and survey mode. P values determined by adjusted Wald test.

DISCUSSION

On average, US adults that do not use the Internet as well as those who did not cite the Internet as their first source the most recent time they searched for health information have significantly lower odds of being aware that there is a vaccine to protect against contracting HPV, the infection’s link with cervical cancer, and how the infection is transmitted compared with those adults that access or use the Internet or who rely on the Internet as a first source for health information. However, our results do not indicate any knowledge gaps among adults who specifically seek out cancer-related information on the basis of their chosen source of information. Regardless of source, adults who sought out cancer information had nearly 50% higher odds of knowing about the vaccine and information about HPV infection than adults who never sought out cancer information (nonseekers). These differences are understandable in that cancer information seekers may exhibit a priori higher levels of awareness and knowledge in relation to cancer and HPV because they are specifically looking for cancer information; therefore, the source of information may not be an important determinant of their knowledge as much as their behavior of purposive seeking.

Overall, our results underscore the importance of communication inequalities in shaping knowledge gaps and echo the concern of the dependence on Internet-heavy DTCA as the main source of vaccine and infection information in the United States. Unfortunately, previous research has also shown that African Americans, Hispanics, and low-SES individuals are significantly less likely to seek out health information online compared with their White and higher-SES counterparts³⁷ increasing the concern that the observed communication inequalities are racially and socially patterned, which may offer one potential explanation for disparities in vaccine completion rates across the United States.

There are some limitations to our study. First, the low survey response rates may increase sampling error in our estimates; however, overall sampling coverage was enhanced through the dual survey frame design, which included cell-phone–only households as well as an oversampling of minorities. In addition, we were not able to account for either general media exposure or actual vaccination behavior in our analyses because these data were not collected in HINTS 2007. It could be argued that Internet users and Internet health information seekers reported higher levels of HPV awareness and knowledge based on the fact that they are large consumers of all media and thus may have been exposed to more television advertisements related to the HPV vaccine. However, we know from communication research that media consumption patterns are not consistent across channel and, therefore, it cannot be determined that Internet exposure translates equally to television, newspaper, or other media exposure.³⁸ Although this study offers an important first-step assessment of HPV-related knowledge gaps and presents Internet-based health information–seeking behaviors as a potential driving factor behind these gaps, future research should both directly measure DTCA vaccine exposure by channel as well as actual vaccination behavior to better address these concerns. Also, as this was a cross-sectional survey study, it is challenging to account for unmeasured confounding, which may offer alternative explanations to the ones presented. Additional empirical evidence is needed to offer further support to these initial and novel findings.

Researchers and public health practitioners have begun to investigate potential factors behind the observed racial and social disparities in HPV vaccine series completion rates, yet most of this research has focused on specific health care issues, such as access, type and status of insurance, distance from clinics, type of provider, and type of visit, with mixed results across study and across age groups.^4–6,39 Little attention has been paid to the potential role of communication and subsequent communication inequalities in influencing these disparities. We know from decades of health behavior research that knowledge plays an important role in helping to shape attitudes and beliefs, which then, in turn, can influence intention to vaccinate—this has been demonstrated in the case of HPV as well.^40,41 Research is starting to emerge that illustrates the importance of how information is delivered and subsequently whether it influences attitudes, beliefs, and intentions toward HPV vaccination.^42–44 Our research extends this line of examination by highlighting the importance of public communication as well as the sources used for information delivery in shaping the public’s knowledge about HPV and the HPV vaccine. Our data offer support that the strategy of using a channel such as the Internet whose use is already racially and socially patterned is likely to not only exacerbate gaps in knowledge among these vulnerable groups but also to potentially contribute to disparities in health outcomes as evident in current vaccine completion rates.

It is important to acknowledge, then, that these data also provide direction on how to bridge observed racial and social disparities in HPV vaccine series completion rates, short of mandating vaccination. One possible mechanism is to increase Internet access among those groups that currently lack access while at the same time improving computer literacy skills, which serve as a cornerstone for Internet-based health information seeking. In this regard, the US government recently issued the National Broadband Plan which directs the Federal Communications Commission to work toward universal broadband Internet access for all Americans and includes direction for the establishment of digital literacy centers.⁴⁵ Progress toward achieving these goals could significantly reduce observed communication inequalities surrounding not only HPV and the vaccine but also in other health-related domains.

Yet, concurrently, realizing the limitations of Internet-based communication, public health agencies and educators should develop off-line communication strategies that promote HPV vaccine and infection knowledge. For example, as stated earlier, patient–provider communication should be examined to determine what if any role it plays in shaping disparities between HPV vaccine initiation and completion rates. In addition, school-based communication efforts and programs have garnered much success for not only prevention of other sexually transmitted infections but also for other health behaviors that warrant parental involvement such as nutrition and physical activity interventions.^46,47 These alternative communication approaches could prove vital in improving current HPV vaccination completion rates and ultimately in reducing the increased burden of cervical cancer among racial/ethnic minorities and women of lower SES.