A Qualitative Comparative Analysis of Combined State Health Policies Related to Human Papillomavirus Vaccine Uptake in the United States

Abstract

Objectives. To examine how combinations of state policies, rather than single policies, are related to uptake of human papillomavirus (HPV) vaccine.

Methods. Using publicly available records and the literature, we characterized policies for each US state and Washington, DC, in 2015 (n = 51), including (1) Medicaid expansion, (2) policies permitting HPV vaccination in pharmacies, (3) school-entry requirements, (4) classroom sex education mandates, and (5) parental education mandates. Using qualitative comparative analysis, we identified which existing combinations of these policies were necessary and sufficient for high HPV vaccine initiation among adolescents, with National Immunization Survey-Teen data.

Results. No single policy was necessary or sufficient for high HPV vaccine uptake; however, 1 set of policies had consistently high HPV vaccine uptake: adoption of all policies except parental education mandates (girls: consistency = 1.00, coverage = 0.07; boys: consistency = 0.99, coverage = 0.08).

Conclusions. We identified a set of polices related to high HPV vaccine uptake. Future studies should examine how these policies and others, individually and in combination, are associated with HPV vaccine uptake.

Public Health Implications. This study provides insight into what sets of policies are consistently related to high HPV vaccine uptake.

Human papillomavirus (HPV) is the most common sexually transmitted infection in the United States.¹ Each year, persistent HPV infection leads to more than 26 000 cases of HPV-associated cancers (including cervical, oropharyngeal, anal, vulvar, vaginal, and penile).¹ The United States faces more than $8 billion annually in direct medical costs as a result of HPV,² demonstrating considerable health and financial burden.

Much of the burden of HPV infections and HPV-associated cancers is avoidable through prophylactic vaccination.^3,4 The Advisory Committee on Immunization Practices recommends routine administration of HPV vaccine to 11- and 12-year-old girls (since 2006) and boys (since 2011), with catchup vaccination recommended through adolescence and early adulthood.⁵ Despite these recommendations, vaccine uptake (i.e., receipt of at least 1 dose) is suboptimal: as of 2015, only 63% of girls and 50% of boys (adolescents aged 13–17 years) had received at least 1 dose of the 2- to 3-dose series.⁶ As of 2015, uptake of HPV vaccine varied considerably across states, ranging from 48% to 88% for girls and 35% to 81% for boys.⁶

Annual increases in HPV vaccination appear to be stagnating, and only 4 states experienced statistically significant increases in initiation among girls from 2014 to 2015.⁶ Thus, more interventions are needed to address suboptimal HPV vaccination uptake. Interventions such as policy changes can be more effective at changing health outcomes than individualized public health programs⁷; state-level policies may therefore be particularly useful for changing health care processes and outcomes. Because policies are not implemented in a vacuum, it is important to understand how different combinations of policies are associated with HPV vaccination. The lack of research in this area limits our ability to design effective policy interventions that promote uptake of HPV vaccine. To our knowledge, no studies have examined how combinations of state policies, rather than single policies, are related to uptake of HPV vaccine.

Qualitative comparative analysis (QCA) is a formalized qualitative analytic approach that can be leveraged to determine which sets of state policies may be necessary or sufficient for high state-level HPV vaccination uptake. States have enacted multiple, often overlapping policies that may influence HPV vaccine uptake; QCA is well suited for characterizing these complex relationships. Unlike statistical approaches that seek to isolate the relationship between a single variable and outcome, QCA can examine the relationship between multiple policies and an outcome. We used QCA to examine which sets of state policies are related to high levels of HPV vaccination uptake among adolescent boys and girls.

METHODS

To accomplish our objective, (1) we characterized the policy environment relevant to HPV vaccine uptake among adolescents for each of the 50 US states and Washington, DC (hereafter collectively referred to as “states”), and (2) we used QCA to identify sets of state policies that are necessary and sufficient for higher uptake of HPV vaccine initiation (i.e., receipt of ≥ 1 dose) among adolescents.

Policymakers have used various policy approaches to address HPV uptake. These approaches have focused on increasing access to the vaccine through insurance coverage, expanding points of care, and mandates or requirements. We examined the following 6 state-level policies that are related to these approaches and have been associated with HPV vaccine uptake in previous empirical studies.

1. Medicaid expansion through the Affordable Care Act. Continuous health insurance coverage has been associated with higher odds of HPV vaccination among adolescents,⁸ and Medicaid coverage significantly increased the use of most preventive services in Oregon.^8,9

2. Policies permitting HPV vaccination in pharmacies. Pharmacies may be an acceptable and convenient location for HPV vaccination,^10,11 and they may increase uptake by expanding access to vaccination services.¹²

3. School-entry requirements for HPV vaccination. Although vaccination rates have increased dramatically after implementation of school-entry requirements for other adolescent vaccines,¹³ only 3 states currently have school-entry requirements for HPV vaccination.¹⁴

4. Nonmedical exemption from school-entry HPV vaccination requirements. Such policies may weaken the effects of school-entry requirements for vaccination.¹⁵

5. Sex education mandates. Sex education curricula that include information about the prevention of sexually transmitted infections may raise awareness and improve acceptability regarding HPV vaccination among both adolescents and their parents.¹⁴

6. Parental education mandates. There is no strong evidence that state mandates to provide HPV vaccine educational materials to parents is effective in increasing uptake. Cross-sectional analyses have demonstrated that these types of laws are not correlated with higher uptake¹⁶; however, education mandates may have an effect when combined with other policies.

Of note, we found that policies allowing nonmedical exemptions from school-entry requirements were collinear with other policies, and there was little state-to-state variation in this variable (only 2 states had these exemption policies). We therefore dropped this variable from our analysis.

Data Sources

We gathered data on the state health policies from publicly available state legislation records and reports (e.g., the National Conference of State Legislatures)^17–20 and the literature.²¹ We obtained additional information on Medicaid expansion,²² pharmacy policies,²³ and sex education^24,25 from online reports. We dichotomized these policies as “has any regulation” versus “has no regulation” as of 2015. We also obtained 2015 state rates of HPV vaccination uptake among adolescents aged 13 to 17 years (combined, and separately for boys and girls) from published reports from the National Immunization Survey-Teen,⁶ a population-based survey of parents of adolescents that verifies vaccination status against health care provider records.

Analysis

We employed fuzzy-set QCA,^26,27 a set-theoretical methodology, to identify combinations of state policies that are related to HPV vaccine uptake. QCA is a formalized qualitative analytic approach that allows researchers to identify combinations of characteristics associated with an outcome without relying on statistical power to draw inferences. It is therefore ideal for the small sample sizes often found in state-level studies. Our data set meets the minimum recommended sample size of 3 times the number of state policies (i.e., 5 state policies, or 3 × 5 = 15 units).²⁷

QCA differs from regression analysis in that it does not assume a linear, symmetrical relationship among variables. Instead, QCA determines which set(s) of policies may be necessary or sufficient for a specific outcome (i.e., high HPV vaccination rates) to occur. Policies considered necessary are those that must be present in a combination to have the expected outcome (i.e., high vaccine uptake). A policy that is considered sufficient would, on its own, be associated with the expected outcome.

High HPV vaccine uptake.

For our outcome measure, high HPV vaccine uptake, we considered 2015 HPV uptake rates and goals set by Healthy People 2020. Currently, no states have achieved the Healthy People 2020 goal for HPV vaccination (80% of adolescents aged 13 to 17 years vaccinated with 3 doses of HPV vaccine).²⁸ We therefore implemented a comparative performance approach and anchored the set to the bottom and top quartiles of state vaccine uptake, with the 50th percentile serving as our crossover point. In other words, we assigned states full membership (i.e., assigned value of 1) to the set of high HPV vaccine uptake states if they fell within the top quartile of observed HPV vaccine initiation. We considered states to be fully out of the set of states with high vaccine uptake (assigned value of 0) if they were within the bottom quartile of observed HPV vaccine initiation. For states that had observed HPV vaccine initiation levels between the 25th and 75th percentiles, we assigned values between 0 and 1 based on their comparative uptake (i.e., states with uptake in the 26th percentile would be assigned a value near 0, whereas a state with uptake in the 74th percentile would receive a value close to 1). This allowed us to incorporate the range of HPV vaccine uptake, rather than dichotomizing uptake as high or low; thus, the assigned values are continuous between 0 and 1.

Configurations of state policies.

For our configurational measures (state policies), we assigned membership or nonmembership into each set (k); thus, we included states in the set if they had enacted the policy (coded 1) and did not include them in the set if they had not enacted the policy (coded 0).

Measuring consistency.

Consistency refers to the degree to which states sharing the same sets of policies exhibit the same outcome (i.e., high or not high vaccination rates). Consistency is measured as a proportion, such that if 4 of 5 states with a particular combination of policies have a particular outcome, consistency is 4/5 = 0.8. We used a validated measure of consistency (Xi ≤ Yi) = ∑[min(Xi , Yi)]/ ∑(Xi), where X is the set membership (i.e., set of policies) and Y is the set membership in the outcome set (i.e., high rate of HPV vaccination).²⁷ We included configurations with a consistency threshold of 0.8 or greater, which is higher than the minimum recommendation published elsewhere.²⁷ We included in the analysis only policy configurations that had both at least 1 observation and a consistency above 0.8.

Measuring coverage.

We used a Boolean algebra-based algorithm (Quine–McCluskey algorithm) to reduce the number of logically redundant configurations. For example, if policies A and B and C have an outcome of high HPV vaccination, and policies A and B and no policy C have the same outcome, then we can logically reduce the policy sets to policy A and B. Once we had logically reduced and simplified our combinations, we measured coverage, or the degree to which a condition (a policy) accounts for the outcome (high HPV vaccination rates): (Xi ≤ Yi) = ∑[min(Xi, Yi)]/∑(Yi), where X is the set membership (i.e., policies) and Y is the outcome of set membership (i.e., high rate of HPV vaccination or not).²⁷ Conceptually, coverage is similar to an R² value in regression analyses, indicating whether most of the configurations of policies explaining the outcome were included in our analyses. The stronger the relationship between the set of policies and the outcome, the higher coverage will be (i.e., closer to 1). Unlike consistency, there is no standard threshold for measuring adequate coverage.

We conducted all analyses using Stata version 14 (StataCorp LP, College Station, TX), which has a program developed for QCA.²⁹

RESULTS

The highest rate of HPV vaccine initiation for both adolescent girls and boys was in Rhode Island (uptake for adolescent girls: 88%; uptake for adolescent boys: 81%; Table 1). The most common policy to be adopted across states was allowing HPV vaccination in pharmacies (47 states), followed by Medicaid expansion (32 states) and policies related to sex education mandates (23 states). School-entry requirements for the vaccine were the least common policy (3 states). Overall, there was state variation in what combinations of state policies had been adopted (Table 2). The most common combination of state policies included adoption of (1) Medicaid expansion, (2) policies permitting HPV vaccination in pharmacies, and (3) classroom sex education mandates; and nonadoption of (4) parental education mandates and (5) school-entry requirements for the vaccine (n = 11 states; Table 2).

TABLE 1—

Policy Adoption by State and by State Vaccine Rates: United States, 2015

State	Medicaid Expansion Through the Affordable Care Act	Policies Permitting HPV Vaccination in Pharmacies	School-Entry Requirements for Adolescent Vaccination	Classroom Sex Education Mandates	Parental Education Mandates	Vaccine Rates Among Adolescent Girls, %	Vaccine Rates Among Adolescent Boys, %
RI	X	X	X	X		87.9	80.6
DC	X	X	X	X		76.5	73.0
NH	X					74.2	69.8
MA	X	X				73.5	63.0
NV	X	X		X		72.0	44.5
HI	X	X				71.3	62.5
CT	X	X				70.9	65.3
ND	X	X		X	X	70.5	62.3
OR	X	X		X		70.0	58.6
NJ	X	X		X	X	69.0	50.9
VT	X	X		X		68.7	66.1
AZ	X	X				68.3	51.3
MI	X	X		X	X	67.6	52.3
DE	X	X		X		67.6	62.9
NE		X				67.3	54.3
IA	X	X		X	X	66.7	48.0
CA	X	X		X		66.7	58.5
NM	X	X		X		66.7	54.3
MD	X	X		X		66.0	55.0
ME		X		X		66.0	65.8
WA	X	X			X	65.8	46.8
NC		X		X	X	65.7	48.0
MN	X	X		X		65.5	57.1
CO	X	X			X	65.3	63.2
AR	X	X				63.5	44.2
FL						62.5	45.3
NY	X				X	62.3	60.3
PA	X	X				62.2	55.9
IL	X	X				62.0	44.3
WV	X			X		62.0	45.3
VA		X	X			61.2	40.1
OH	X	X		X		61.0	43.7
WI		X				60.5	46.4
LA	X	X			X	60.3	49.5
TX		X		X	X	60.1	41.4
TN		X		X		59.7	38.2
MO		X		X		59.3	44.7
OK		X				58.1	52.9
AL		X				57.7	39.4
KY	X	X		X		57.4	34.8
ID		X				57.3	44.2
AK	X	X				57.0	41.6
MT	X	X		X		55.0	46.0
GA		X		X		54.4	51.0
IN	X	X		X	X	53.7	43.2
SC		X		X	X	53.7	35.1
SD		X				53.2	39.2
MS		X			X	52.4	38.9
KS		X				50.9	36.0
UT		X		X	X	47.8	40.9
WY		X				47.7	37.1

Note. HPV = human papillomavirus. X = the state has adopted the policy.

TABLE 2—

Average Rates of Vaccine Uptake Among Adolescent Boys and Girls, by Set of State Policies: United States, 2015

		Vaccine Rates Among Adolescent Girls, %				Vaccine Rates Among Adolescent Boys, %
Policy Sets	No. of States	Mean	Median	Min	Max	Mean	Median	Min	Max
MPsCe	11	65.1	66.7	55.0	72.0	52.9	55.0	34.8	66.1
MPsce	8	66.1	65.9	57.0	73.5	53.5	53.6	41.6	65.3
mPsce	8	56.6	57.5	47.7	67.3	43.7	41.8	36.0	54.3
MPsCE	5	65.5	67.6	53.7	70.5	51.3	50.9	43.2	62.3
mPsCE	4	56.8	56.9	47.8	65.7	41.4	41.2	35.1	48.0
mPsCe	4	59.9	59.5	54.4	66.0	49.9	47.9	38.2	65.8
MPscE	3	63.8	65.3	60.3	65.8	53.2	49.5	46.8	63.2
MPSCe	2	82.2	82.2	76.5	87.9	76.8	76.8	73.0	80.6
MpsCe	1	62.0	. . .	. . .	. . .	45.3	. . .	. . .	. . .
MpscE	1	62.3	. . .	. . .	. . .	60.3	. . .	. . .	. . .
Mpsce	1	74.2	. . .	. . .	. . .	69.8	. . .	. . .	. . .
mPSce	1	61.2	. . .	. . .	. . .	40.1	. . .	. . .	. . .
mPscE	1	52.4	. . .	. . .	. . .	38.9	. . .	. . .	. . .
mpsce	1	62.5	. . .	. . .	. . .	45.3	. . .	. . .	. . .

Note. M = Medicaid expansion through the Affordable Care Act; P = policies permitting HPV vaccination in pharmacies; S = school-entry requirements for adolescent vaccination; C = classroom sex education mandates; E = parental education mandates. Capitalization indicates that the state has the policy; lower case indicates that the state does not have the policy.

We found no single policy that was necessary and sufficient for high HPV vaccine uptake. We found 1 combination of policies that had consistently high uptake of HPV vaccine for both boys and girls: adoption of (1) Medicaid expansion, (2) policies permitting HPV vaccination in pharmacies, (3) school-entry requirements for the vaccine, and (4) classroom sex education mandates and (5) absence of parental education mandates (girls: consistency = 1.00, coverage = 0.07; boys: consistency = 0.99, coverage = 0.08). Washington, DC, and Rhode Island were the only 2 states with these sets of policies.

DISCUSSION

Many factors can influence HPV vaccine uptake, including adolescent and parental preferences, provider recommendations, financial costs, and a host of social factors.^{1,9,10,30–32} State-level policies may influence these factors by acting as antecedents, or they may moderate their effect (e.g., Medicare expansion can mitigate financial cost barriers). Although states can enact a variety of policies to influence health outcomes, little is known about the combined impact of different sets of policies on health outcomes. Whereas most studies on state-level policies and HPV vaccine uptake have sought to describe the effect of an individual policy, we sought to identify combinations of state policies that were associated with high uptake of HPV vaccine.

We found that Washington, DC, and Rhode Island had a configuration of policies that was related to high HPV vaccine uptake: Medicaid expansion through the Affordable Care Act, polices allowing pharmacies to provide HPV vaccination, implementation of school-entry requirements for the vaccine, requirement of classroom sexual education, and the absence of parental education mandates for HPV vaccination.

By using a configurational approach, we were able to illustrate that a combination of these policies may have led to a particular outcome. For example, consider school-entry requirements for HPV vaccine. Currently, only 3 states (Rhode Island; Washington, DC; and Virginia) have enacted these policies. In our study, Rhode Island and Washington, DC, both had high vaccine uptake (i.e., above 75th percentile), whereas Virginia was not considered to have high uptake. Rhode Island and Washington, DC, had a different combination of policies than Virginia: Virginia had no classroom sex education mandates and did not expand Medicaid. Virginia (as well as Washington, DC) also offers an extremely broad exemption policy for school-entry requirements for HPV vaccination (exemptions for any reason), making the requirement less effective. The opposite is true for Rhode Island, which has the highest HPV vaccination rate and a narrower exemption policy for HPV vaccine school requirements (medical and religious reasons only). A recent study of the impact of the Virginia school-entry HPV vaccine requirement showed that there was no statistically significant increase in HPV vaccine uptake in Virginia.³³ On its own, the policy has yet to achieve the desired effects. Our configurational analysis aligns with these previous findings, as we found that a school-entry requirement may not be sufficient on its own for high HPV uptake. However, given a specific combination of other policies, even broad school-entry mandates can be effective in achieving high HPV vaccine uptake, as seen in Rhode Island and Washington, DC. Thus, public health professionals, policymakers, and researchers should consider the broader policy environment when advocating for or evaluating the effectiveness of a given policy.

Interestingly, we found that not having parental education mandates included in the state policy set was associated with high HPV vaccine uptake. It is important to note that these findings do not suggest that parental education mandates are not effective at increasing HPV vaccine uptake or that they decrease HPV vaccine uptake. First, no state or district had all 5 state policies in place, and we cannot make inferences about sets of policies that were not observed (i.e., counterfactuals), making this a limitation of our analysis. Second, parental education policies may not be necessary or sufficient for high HPV vaccine uptake, as evidence suggests that parental knowledge of HPV-associated diseases does not by itself predict vaccination of adolescents³¹; however, research suggests that brief educational interventions—such as those using informational pamphlets—can effectively change HPV vaccine attitudes and intentions.³⁰ These psychosocial factors are associated with HPV vaccine uptake when there are limited external constraints (e.g., no health insurance coverage).³² Therefore, education policies may work best in concert with other policies that minimize disparities in access, such as Medicaid expansion. For this reason, policymakers should continue to consider the value of parental educational interventions for improving HPV vaccine uptake in combination with their other policies.

Currently, research on the impact of Medicaid expansion on HPV vaccine uptake is limited, and the effect of Medicaid expansion on vaccine uptake is unclear. In our study, we found Medicaid expansion to be a necessary but not sufficient policy related to high uptake of the HPV vaccine. Medicaid expansion may increase adolescents’ access to care, including preventive vaccines; alternatively, Medicaid expansion may have little relevance for this population and outcome, as adolescents who were uninsured or underinsured may already have been eligible for vaccines through programs such as Vaccines for Children, which is associated with increased HPV vaccine uptake.³⁴ Thus, other factors that influence uptake—including attitudes, knowledge, and physician recommendation⁹—should be targeted in HPV vaccination interventions.

Limitations

Although this study examines sets of state policies that are related to high HPV vaccine uptake, it has limitations. First, state policies occur within a broader context of systemic-, interpersonal-, and personal-level factors.³⁵ Thus, we cannot parse out individual effects of policies given the underlying policy environments, nor can we establish causality between combinations of policies and high HPV vaccine uptake. Statistical approaches that are complementary to QCA will be important to fully understand HPV vaccine uptake. Second, not all possible combinations of state policies were represented in our data set. We cannot make inferences about unobserved sets of policies; instead, our findings describe how the observed combinations of policies were related to high HPV vaccine uptake, which provides important information about current policy approaches for vaccination outcomes. Third, we included 5 current policies considered to directly or indirectly affect HPV vaccine uptake. There is considerable variation in state and federal policies and programs that may influence HPV vaccine uptake, which we did not consider in our analysis. Thus, the findings are not prescriptive but rather illustrate the value of considering the interrelatedness of policies and their possible impact on public health outcomes. Future research may consider or incorporate other policies or programs in combinations to explore the configurational effects of state policies on HPV vaccine uptake. Finally, we were unable to account for the length of time states had implemented each policy, which may affect the relationship between the sets of policies and HPV vaccine uptake; future studies using statistical methods may examine this effect directly.

Conclusions

Overall, we found that states (including the District of Columbia) that implemented Medicaid expansion, allowed pharmacies to provide HPV vaccines, implemented school-entry requirements for the vaccine, and required classroom sexual education, in the absence of parental education mandates for HPV vaccination, had higher rates of HPV vaccination. In combination, these policies may be helpful for improving HPV vaccine uptake. Future studies should examine how these policies and others, individually and in combination, are associated with HPV vaccine uptake.