The ideal time for vaccination against human papillomavirus (HPV) strains associated with cervical cancer is before onset of sexual activity. Human papillomavirus vaccination is recommended for females aged 11–12 years.1 “Catch-up vaccination” is vaccination outside of the recommended schedule, between 13 and 26 years of age for HPV. Of 13- to 17-year-old girls in the 2011 Teen National Immunization Survey, 38.4% had received the three-dose series,2 indicating that many young women remain eligible for catch-up vaccination.
Completion rates for females initiating the series have ranged from 28% to 70%.3–8 To date, racial and economic disparities have been associated with series noncompletion.2,9 Furthermore, HPV infection is more prevalent in less-educated women and those of lower socioeconomic status,9 indicating that public health benefit could be significant if vaccine coverage gaps in these subgroups were narrowed.
Catch-up vaccination occurs during a time of increased likelihood of sexual activity.10 Because vaccination is not recommended during pregnancy, pregnancy may theoretically interrupt the series.11
Similar to HPV vaccination completion, racial and economic disparities are seen for adolescent unintended pregnancy. Female adolescents of color and those living in poverty are more likely to have unintended pregnancies than white adolescents and those above the poverty level, respectively.12 Barriers to vaccination may overlap with barriers to prevention of unintended pregnancy. Among females initiating catch-up vaccination, we aimed to assess whether the occurrence of pregnancy was associated with series noncompletion and to identify demographic characteristics of noncompleters. We hypothesized that pregnancy during the vaccine series would be associated with noncompletion.
MATERIALS AND METHODS
Baseline data for this study were first collected for another related study,13 but follow-up data for vaccine completion were collected later to meet the aims of this retrospective cohort study. We abstracted data from the electronic medical records of females aged 13–26 years who presented to the University of Illinois Hospital & Health Sciences System for their first HPV vaccination between June 2006 and April 2010 (n=507). University of Illinois Hospital & Health Sciences System clinical sites at which these vaccine visits occurred were those providing primary care including gynecology, family medicine, and adolescent medicine (a pediatric subspecialty) as well as a procedural site, the colposcopy clinic, which provides evaluation and treatment of abnormal cervical cytology.
Participants were identified through records of dispensation of the quadrivalent HPV vaccine at the University of Illinois Hospital & Health Sciences System outpatient pharmacy at which patients fill their prescriptions for the HPV vaccine according to clinical protocol (patients receive a prescription for the vaccine, go to the University of Illinois Hospital & Health Sciences System pharmacy to fill the prescription, and return to the clinic for vaccine administration with a nurse). The date of initiation was defined as the date of vaccine dispensation from the pharmacy. During the study period, the quadrivalent vaccine was the only HPV vaccine available at our institution.
The study was approved by the University of Illinois at Chicago institutional review board. Inclusion criteria were female gender, age 13–26 years, having filled a prescription for HPV vaccine at a University of Illinois Hospital & Health Sciences System pharmacy, and adequate electronic medical record for data collection. Adequacy was defined as presence of a clinical note that could be linked to prescription of the vaccine (the “vaccine visit”). Exclusion criteria included male gender and inadequate medical records as well as individuals younger than 13 years and older than 26 years at the time of vaccine initiation. Thirteen was chosen as the lower age limit because this age is generally considered to be the start of early adolescence14 and we were interested in the population receiving catch-up vaccination, ie, vaccination outside of the recommended timeframe of 11–12 years of age.1 We chose 26 years as our upper age limit because this was the upper age recommended by the manufacturer.
After identifying the patient's first immunization, we retrieved the clinical encounter note from the prescribing health care provider and notes from clinical encounters subsequent to vaccine initiation. A single physician (R.P.) reviewed the electronic medical records and identified clinical notes appropriate for data extraction. Data were extracted from the clinical notes into the database by one of three individuals (R.P., M.C.Y., or Ms. Eden Pappo) and entry was verified by R.P. for all patients. Data cleaning, using wild-code checking and consistency checking, was performed by R.P. and B.H. To allow for sufficient follow-up time for vaccine completion, patients were included if at least 1 year had elapsed between the first dose and the end of the study period. In addition, although the vaccine was available on the market starting in 2006, only one patient had their first dose at University of Illinois Hospital & Health Sciences System in 2006 and was excluded from the study. Therefore, the current study includes patients who received their first HPV vaccination between January 2007 and April 2009 (n=376) with follow-up through April 2010 for vaccine series completion.
The main exposure collected from the electronic medical record was a pregnancy occurring during the vaccination series (defined as occurring within 1 year of the first vaccination unless the series had already been completed). Evidence of pregnancy was recorded if the patient's electronic medical record had a clinical note documenting pregnancy or a laboratory entry indicating a positive human chorionic gonadotropin test. Pregnancy status could not be ascertained for 66 patients without any clinical notes or laboratory results during the follow-up time period. These patients were coded as missing for the pregnancy variable and excluded from the study for a final analytic sample of 310 (sampling flowchart shown in Fig. 1).
Demographic variables collected from the electronic medical records included age at first HPV vaccination, race or ethnicity, and insurance used to pay for vaccine (private, public, or self-pay). During the time period of analysis, the out-of-pocket cost for vaccination was approximately $120 per dose ($360 for series).
Clinical variables collected included clinical setting of HPV vaccine provision (primary care or gynecology clinic or colposcopy clinic) and health care provider type (resident physician, attending physician, or certified nurse midwife). As a result of the overlap of clinical setting and health care provider type, we created a composite variable with the following categories to describe the site of vaccination provision: resident clinic, attending clinic, midwife clinic, or colposcopy clinic.
The primary outcome was completion of the third vaccination. Although the recommended timeframe for vaccine completion is ideally 6 months (doses at 0, 2, and 6 months), we allowed a minimum 1-year of follow-up before participants were considered to have failed to complete the series. Because we reviewed charts over an approximately 3-year period, those patients who received their first vaccine early in the study period were allowed a greater follow-up time than those who received the vaccine later in the study period. Therefore, the maximum follow-up period was 3 years 3 months for a patient who had received the vaccine in January 2007, and the minimum follow-up period was 1 year for a patient who had received the vaccine in April 2009.We used descriptive and χ2 statistics and log binomial regression modeling for multivariable analysis. Log binomial regression was chosen over logistic regression because odds ratios overestimate the relative risk (RR) with a relatively common outcome such as vaccine completion. Relative risks and 95% confidence intervals (CIs) were therefore directly estimated from log binomial regression models.15 Only patients with complete data on the key exposure variable (whether pregnancy had occurred) were included in the analysis (n=310). A sensitivity analysis was also performed using the entire sample (n=376), assuming everyone with missing data for pregnancy had not become pregnant. Data were analyzed using SAS 9.2. Significance level was set at α of 0.05. A post hoc power analysis revealed that, given our sample size and the observed rate of HPV vaccine completion of 59.9% among the nonpregnant, we had 80% power to detect an RR of 0.57.
The characteristics of participants are in Table 1. The mean age at receipt of first vaccination was 22 years. The majority received vaccination through a primary care clinic, and the rest received vaccination through colposcopy clinics. The largest proportion was provided vaccination by resident physicians. The patients were racially and ethnically diverse. Most vaccinations were covered through public insurance.
Thirty-one pregnancies occurred within 1 year of the first vaccination (10%). Among those who became pregnant and were using contraception at initiation of HPV vaccination, the methods were as follows: five were using oral contraceptive pills, three were using condoms, two were using depot-medroxyprogesterone, two were using intrauterine devices, one was using the contraceptive ring, and one was abstinent. Among those who were not using contraception, four were provided a method (one each contraceptive ring, depot medroxyprogesterone, intrauterine device, and condoms), and three were not. One was missing information on contraception at initiation of the HPV vaccination. We did not collect follow-up contraceptive data to assess for method changes.
Of the 310 patients included in the analysis, 77.1% received the second of three doses, and 56.1% completed the series. Mean follow-up time was 2.1 years (standard deviation 0.57).
In the unadjusted results, females who experienced pregnancy were less likely to complete the series than those who did not have a pregnancy (22.6% compared with 59.9%, respectively; RR 0.38, 95% CI 0.20–0.73). Black females were less likely to complete the vaccine series than were white females (48.0% compared with 72.1%, respectively; RR 0.66, 95% CI 0.53–0.83), and patients between the ages of 19 and 21 years were less likely to complete than 22–26 year olds (44.1% compared with 60.6%, respectively; RR 0.73, 95% CI 0.53–0.99). Compared with private insurance, those with public insurance were less likely to complete the series (47.4% compared with 71.8%, respectively; RR 0.66, 95% CI 0.52–0.84). Likelihood of series completion was not related to the type of clinic in which vaccination occurred (Table 2).
When the factors were subjected to multivariable analysis, pregnancy remained significantly negatively associated with vaccine series completion (adjusted RR 0.42, 95% CI 0.21–0.81) (Table 3). Black race also remained negatively associated with vaccine completion (adjusted RR 0.71, 95% CI 0.54–0.91). Clinic type and age were not associated with vaccine completion in multivariable analysis. Insurance status was highly correlated with race (87.6% of black patients had public insurance compared with 11.4% of white patients) and was not significantly associated with vaccine completion in multivariable analysis, so it was dropped from the model to avoid issues with multicollinearity.
In a sensitivity analysis using the entire sample (n=376), which assumed that all patients with missing pregnancy status (n=66) did not become pregnant within 1 year of vaccine initiation, results were slightly attenuated (RR 0.52, 95% CI 0.27–1.00). Those excluded from the sample as a result of missing pregnancy status were not significantly different on measured characteristics from those included in the sample, except that they were more likely to be in the 19- to 21-year age category (33% compared with 19%, respectively, P=.06).
We found a low rate of HPV vaccination completion (56.1%) among adolescents and young women initiating vaccination on a catch-up schedule. The overall rate of HPV vaccination completion in our study is similar to rates reported in other retrospective reviews. Pregnancy during the series was associated with vaccination noncompletion with those who became pregnant having a 68% lower likelihood of completion. Black patients were less likely than white patients to complete the series, consistent with other studies describing racial and socioeconomic disparities in HPV vaccination.5–8,16
Pregnancy during vaccination was uncommon, but those who experienced pregnancy were less likely to complete the series. This suggests that preventing unintended pregnancies among females receiving the vaccine on a catch-up schedule, who are at risk for both pregnancy and HPV infection, may increase completion rates.
It has been recommended that all patients receiving HPV vaccination be counseled on contraception and provided a reliable method to prevent unintended pregnancy if at risk.17 Completion may also be increased by incorporating HPV vaccination into postpartum visits, or into inpatient postpartum care, which Wright has shown to be feasible and acceptable to patients.18
Females who become pregnant during the vaccination time frame are ineligible to complete the series during pregnancy at our institution, consistent with the American College of Obstetricians and Gynecologists' Committee Opinion.11 Our finding of series “noncompletion” may therefore be series delay as a result of the pregnancy itself, at least among those with shorter follow-up times. However, follow-up time for the study ranged from 1 to 3 years with an average of approximately 2 years, suggesting that patients experiencing pregnancy are less likely to return at all for vaccine completion after conclusion of the pregnancy.
It is not in our clinical protocol to contact vaccination noncompleters, but reminders such as text messages that have been shown to be useful in other reproductive health areas19–21 should be studied as tools to increase completion. All health care providers at our institution had access to vaccination status in the electronic medical record and could assess for incomplete vaccination at postpartum visit; however, determining this would have added time to a visit. Systems-based interventions should therefore be considered such as noting incomplete HPV vaccination on a problem list or including it on postpartum discharge summaries. Because the series does not need to be restarted regardless of the time elapsed between doses,11 females may continue the series at any point after pregnancy resolution, even if several years have elapsed. Because many individual clinicians counseled patients, we were unable to ascertain whether females received counseling on the extended time frame allowable for the HPV vaccination.
This study has several additional limitations. The rate of pregnancy may be underestimated as a result of patients who may have sought prenatal care or termination of pregnancy elsewhere. The proportion completing the vaccine may also be artificially low because some patients may have completed the series at another clinic. We did not control for visit indication at the time of vaccine initiation; therefore, a differential follow-up rate may exist for patients presenting for different indications, and this analysis would not uncover those differences. We were unable to consider change in insurance as a potential confounder for vaccine completion because insurance status was only collected at baseline. The mean age of females in this analysis may be older than the average female receiving catch-up HPV vaccination and representative of an age cohort that has a higher pregnancy rate than younger teens. Therefore, our results may not be generalizable to populations of primarily younger adolescents. The follow-up period was limited to approximately 3 years and varied across patients. We may have found an increased rate of completion had we used a longer minimal follow-up period.
Our sensitivity analysis, imputing values of “not pregnant” for all of those with missing data for pregnancy occurrence, resulted in an attenuation of results. This attenuation may be attributable in part to the introduction of some exposure misclassification with the assumption that all females with missing data did not become pregnant when some of them likely did. Regardless, the overall conclusion that pregnancy interrupts the HPV vaccination series remains unchanged.
In summary, this study reports significantly lower HPV vaccine completion rates for those who become pregnant during the vaccine series, suggesting that HPV vaccination completion on a catch-up schedule may be improved by efforts to prevent pregnancy among those receiving the vaccine and increase postpartum HPV vaccination counseling. It also adds to the growing body of literature on the disparity between black and white HPV vaccine completers.
1. Markowitz LE, Dunne EF, Saraiya M, Lawson HW, Chesson H, Unger ER, et al.. Quadrivalent human papillomavirus vaccine: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 2007;56:1–24.
2. Centers for Disease Control and Prevention (CDC). National and state vaccination coverage among adolescents aged 13–17 years—United States, 2011 [published erratum appears in MMWR Morb Mortal Wkly Rep 2012;61:844]. MMWR Morb Mortal Wkly Rep 2012;61:671–7.
3. Dorell CG, Yankey D, Santibanez TA, Markowitz LE. Human papillomavirus vaccination series initiation and completion, 2008–2009. Pediatrics 2011;128:830–9.
4. Jemal A, Simard EP, Dorell C, Noone AM, Markowitz LE, Kohler B, et al.. Annual report to the nation on the status of cancer, 1975–2009, featuring the burden and trends in human papillomavirus (HPV)-associated cancers and HPV vaccination coverage levels. J Natl Cancer Inst 2013;105:175–201.
5. Widdice LE, Bernstein DI, Leonard AC, Marsolo KA, Kahn JA. Adherence to the HPV vaccine dosing intervals and factors associated with completion of 3 doses. Pediatrics 2011;127:77–84.
6. Niccolai LM, Mehta NR, Hadler JL. Racial/ethnic and poverty disparities in human papillomavirus vaccination completion. Am J Prev Med 2011;41:428–33.
7. Cook RL, Zhang J, Mullins J, Kauf T, Brumback B, Steingraber H, et al.. Factors associated with initiation and completion of human papillomavirus vaccine series among young women enrolled in Medicaid. J Adolesc Health 2010;47:596–9.
8. Chou B, Krill LS, Horton BB, Barat CE, Trimble CL. Disparities in human papillomavirus vaccine completion among vaccine initiators. Obstet Gynecol 2011;118:14–20.
9. Dunne EF, Unger ER, Sternberg M, McQuillan G, Swan DC, Patel SS, et al.. Prevalence of HPV infection among females in the United States. JAMA 2007;297:813–9.
10. Finer LB, Philbin JM. Sexual initiation, contraceptive use, and pregnancy among young adolescents. Pediatrics 2013;131:886–91.
11. Human papillomavirus vaccination. Committee Opinion No. 467. American College of Obstetricians and Gynecologists. Obstet Gynecol 2010;116:800–3.
12. Chandra A, Martinez GM, Mosher WD, Abma JC, Jones J. Fertility, family planning, and reproductive health of U.S. women: data from the 2002 National Survey of Family Growth. Vital Health Stat 23 2005;25:1–160.
13. Perry R, Yu MC, Harwood B. Contraceptive need, counseling, and provision at the time of human papillomavirus vaccination on a catch-up schedule. J Pediatr Adolesc Gynecol 2014;27:–6.
14. Sanfilippo JS, Lara-Torre E. Adolescent gynecology. Obstet Gynecol 2009;113:935–47.
15. Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005;162:199–200.
16. Schluterman NH, Terplan M, Lydecker AD, Tracy JK. Human papillomavirus (HPV) vaccine uptake and completion at an urban hospital. Vaccine 2011;29:3767–72.
17. Smith-McCune K, Sawaya GF. Update on quadrivalent human papillomavirus vaccination and pregnancy outcomes: is contraception advisable? Obstet Gynecol 2009;114:1168–9.
18. Wright JD, Govindappagari S, Pawar N, Cleary K, Burke WM, Devine PC, et al.. Acceptance and compliance with postpartum human papillomavirus vaccination. Obstet Gynecol 2012;120:771–82.
19. Castaño PM, Bynum JY, Andrés R, Lara M, Westhoff C. Effect of daily text messages on oral contraceptive continuation: a randomized controlled trial. Obstet Gynecol 2012;119:14–20.
20. Hou MY, Hurwitz S, Kavanagh E, Fortin J, Goldberg AB. Using daily text-message reminders to improve adherence with oral contraceptives: a randomized controlled trial. Obstet Gynecol 2010;116:633–40.
21. Moniz MH, Hasley S, Meyn LA, Beigi RH. Improving influenza vaccination rates in pregnancy through text messaging: a randomized controlled trial. Obstet Gynecol 2013;121:734–40.