Background and methods The national immunisation records of over 220 000 girls offered vaccine in the routine or catch-up programme of the Human papillomavirus (HPV) programme in Scotland were analysed. Descriptive statistics and multilevel modelling were used to determine individual and organisational factors associated with uptake. Age, school year, school denomination, deprivation and, for school-leavers, mode of delivery were explored. Additional aggregate data were used to examine the effect of late uptake of missed doses in the routine vaccination programme.
Results School-based delivery initially achieved over 80% uptake of complete courses in routine and catch-up age groups. Within this context of generally high coverage, there was an association between individual level deprivation and lower uptake, and a decline in in-year course completion over time. However, later uptake of missed doses in the following year substantially decreased these effects. There was no influence on uptake of the type of school (non-denominational/denominational). Vaccination of school-leavers in the catch-up campaign had lower coverage, with 50% starting and 30% completing the course in-year. There was no clear advantage of vaccination through general practice or through Board-run clinics in reaching this group.
Conclusions School-based vaccination can achieve high and equitable uptake of a multidose vaccine in a routine immunisation programme. Sustained high coverage with HPV vaccine across Scotland provides a stable platform for planning future strategies for cervical screening and understanding the impact of the vaccination at a population level.
- Public Health
- Adolescents CG
- Cancer: Cervix
Statistics from Altmetric.com
Background and introduction
In September 2008, Scotland (population 5.2 million) began an HPV vaccination programme to protect girls against cervical cancer later in life. The policy decision to introduce the programme was supported by cost-effectiveness analyses and models that predicted the future impact of immunisation. These showed that the optimum benefit of introducing the vaccine, supplementary to an existing effective cervical screening programme, was predicated on achieving high routine vaccine coverage among 12-year-old to 13-year-old girls immunised with a three dose course.1 ,2
It was clear that the best outcomes for cervical cancer, a disease with a marked association with deprivation, required that vaccine be delivered equitably. Equity in health and healthcare has been defined as ensuring that resources are distributed fairly in relation to the health needs of different groups and areas. To help achieve this, work was performed during the planning stages to identify groups which may have poor access or other barriers to accessing HPV immunisation to support those delivering the programme.
The routine HPV vaccination programme in Scotland, for 12-year-old to 13-year-old girls, is delivered through an entirely school-based programme. This includes private and public secondary schools (considered together here), which are denominational (mainly Roman Catholic) or non-denominational.
A 3-year catch-up programme for older girls (aged 13–17 years) was delivered both at school, for those still in school, and at specific Health Board-run vaccination clinics or general practices (GPs) for school-leavers. The mix of provision for school-leavers was determined by local Health Boards, of which there are 14 in Scotland. The catch-up was phased, targeting the oldest age groups first, and ran between 1 September 2008 and 31 August 2011, with individual invitations and reminders being send to eligible girls during the first 2 years.
This paper explores to what extent the equitable uptake of the HPV vaccination programme (routine and catch-up) was achieved by socioeconomic indicators and by denomination and what individual and organisational factors were associated with this. Individual factors are age, timing of doses, socioeconomic status and school year. Organisational factors are school denomination, deprivation level of pupil population and, for school-leavers in the catch-up programme, vaccination setting. The study uses national disaggregate data to investigate the impact of these on completed courses of vaccination and additional aggregate data to examine late catch-up of missed doses.
Previous studies of HPV vaccine coverage have used data from a limited geographical area, covered a shorter time span and used aggregate figures or an uptake of incomplete courses.3–6 None have looked at a later catch-up of missed doses.
Datasets used to prepare national immunisation statistics were subject to descriptive and multivariable analyses. Three groups were examined: the routine cohort of 12-year-old to 13-year-old girls, that is, the age group that will continue to be offered the vaccine in future, and older girls immunised as part of a catch-up programme who were divided into those still at school and those who had already left school.
School-based vaccines are recorded on the Child Health System Programme-School, whereas those given out of school use the Scottish Immunisation Recall System. Both systems act as population registers for call and recall purposes. The anonymised data, provided by NHS National Services Scotland Information Services Division (ISD), were subject to rigorous, standardised quality assurance checks. Two analysis data files were created, one for the routine programme and one for the catch-up.
The routine programme analysis included three consecutive cohorts of 12-year-old to 13-year-old girls in S2, the second year of secondary school between September 2008 and August 2011. The denominator was based on the school roll. The numerator was the total number receiving one, two or three doses in mid-August, following each school year end, to allow for complete data recording.
The population for the catch-up campaign was girls aged 13–17 years inclusive between September 2008 and August 2011. It was divided into the following two groups: those in school years S3–S6 (the final school year) and those who had left school and were born on or after 1 September 1990. The denominator for the S3–S6 cohort was the school roll and, for school-leavers, was defined by date of birth. The numerator was the total of those who had received one, two or three doses in mid-February of the following year to capture final doses for school-leavers, for whom vaccination mostly began in January 2009, while the school-based vaccination started the preceding September.
There were no personal identifiers in the datasets as these had been removed prior to their release for analysis. Each immunisation record had been assigned to a Scottish Index of Multiple Deprivation (SIMD) quintile using postcode of residence (SIMD-1 being most deprived; SIMD-5 least).7 School-level characteristics using published education statistics were linked using school location code.8 These were: percentage of pupils eligible for free school meals (FSM) and school denomination. FSM was categorised into equal categories, to provide a graded measure of school-level deprivation. School-leavers were classified by the model of delivery used by their Health Board of residence. This was defined as GP delivery, a Board-run clinic or mixed (figure 1).
Further data were provided by ISD showing uptake in each of the routine 12-year-old to 13-year-old cohorts 1 year later, to show the additional effect of later uptake of missed doses.
Data management including linkage to school characteristics used Microsoft Access 2003. Multilevel modelling, using ‘school’ for the routine programme and ‘Health Board’ for the catch-up cohort as the highest nesting level, was used to investigate the association between individual and organisational factors and vaccine uptake. Adjusted ORs and corresponding 95% CIs were calculated in R V.126.96.36.199
Data quality and completeness
There were 86 769 records in the routine programme dataset. The catch-up dataset contained 139 742 records (90 553 school pupils and 49 189 school-leavers).
Individual-level data from one Board was not available for 117 and 110 girls in years 1 and 2, respectively, of the routine programme and 539 older girls, including 143 who had left school.
Data quality and completeness were high. SIMD was missing for only 380 (0.4%) routine and 556 (0.4%) catch-up records. School information could not be linked to 1523 (1.8%) routine and 1364 (1.5%) catch-up records. These records were excluded from the multivariable analysis. FSM data were missing for 4486 routine and 6958 older pupil records. Most of these (95% and 86%, respectively) were pupils attending independent/grant-aided schools. The category ‘FSM missing’ was retained in the multivariable analysis so as not to exclude these records.
Levels of uptake of completed courses of HPV vaccination in each of the three S2 cohorts are shown in table 1. Over 90% started and over 80% completed the course in-year in each year. Although remaining high overall, the uptake in years 2 and 3 was significantly lower than in year 1 for all doses (p<0.0001). In-year course completion decreased over time. By year 3, there was a 10.8% gap between starting and finishing the vaccine course compared with 4.3% in year 1. Table 2 shows the uptake for each cohort 1 year later. This shows that missed doses continued to be given in the following school year leading to overall coverage of more than 90% with all three doses in all years analysed. Over time, the contribution of late doses to completing a course grew with a 3.0% increase in course completion in 2008/2009 compared with 9.1% in 2010/2011.
Individual factors influencing uptake
After adjusting for programme year and organisational factors (table 3), there is no linear effect of individual level deprivation (SIMD) for uptake of dose 1 (p=0.115), but there were linear effects for doses 2 and 3. These both show lower uptake associated with increasing deprivation (p<0.001). The uptake of the first dose continued to have a high and even profile across SIMD quintiles in subsequent years. The uptake of complete vaccine courses was increasingly poor over time and by deprivation, although it still exceeded 80% in all but one of the data cells (table 2). The overall effect was a widening gap between starting and completing a course by deprivation. For example, in year 3 of the routine programme, 91.9% of girls in SIMD-1 started the course but only 76% completed it, while the equivalent figures for girls in SIMD-5 were 91.7% and 84.4%, respectively (table 2). The effect of providing opportunities for catching up on missed doses a year later was most marked in the most deprived quintile, increasing the rate of course completion by 4.9% in year 1 and 13.4% in year 3, and reducing the uptake differential by SIMD.
Organisational factors influencing uptake
There was no difference in uptake between denominational and non-denominational schools for all ages (tables 3 and 5; p=0.678 for uptake of the third dose in the S2 routine programme). There was some effect of deprivation measured at the school level through the percentage of pupils eligible for FSM. Those in the highest FSM quintile in the routine programme have a significantly lower uptake than those in schools with the lowest FSM quintile adjusted OR=0.66 (95% CI 0.56 to 0.78). Uptake for the ‘FSM unknown’ category is also significantly lower.
Patterns of uptake for the catch-up programme at school show a similar profile to the routine programme (table 4), although data on the effect of late doses were not available for these older pupils. Among school-leavers, uptake was much lower. For example, in year 1, 48.5% of 16-year-old to 17-year-old school-leavers started and 31.8% completed a course of vaccine, compared with 93% and 86.6% of school pupils of the same age.
Individual factors influencing uptake
As with the routine programme, the effects of deprivation on uptake for those in school were more evident for the third dose, OR=2.15 (with 95% CI 1.99 to 2.31) for uptake in the least deprived quintile compared with the most deprived quintile (table 5), although again this is within a generally high uptake, over 80%, of complete courses across all SIMD quintiles. In school-leavers, there are two effects of deprivation which are evident. First, that more than twice as many girls in the most deprived SIMD quintile had left school than in the least-deprived quintile (around 14 600 compared to 6600 across 2 years). Second, that vaccine uptake declines with the deprivation quintile, with a 5% decrease in coverage between SIMD-5 and SIMD-1. The adjusted OR for a school-leaver in SIMD-5 starting a course of vaccine was 1.45 (95% CI 1.37 to 1.54) compared with that in SIMD-1.
Five Boards arranged all school-leaver vaccination through Board-run clinics, seven used mixed Board and GP delivery and one used GP only delivery, accounting for 25.1%, 64.6% and 10.3%, respectively, of the school-leaver cohort. There was a strong interaction between the mode of delivery (measured at the Board level) and the year of programme (p<0.0001). Uptake in the first year was higher in Board-run clinics (52.3%) compared with mixed models (47.9%) and GP delivery (43.5%) for dose 1 and 36.8%, 31.4% and 21.9%, respectively, for completed courses. The pattern was reversed in year 2, with 40.1%, 51.6% and 56.7% starting and 22.4%, 30.5% and 33.7% finishing a course in Board-run, mixed and GP only models, respectively.
High, sustained and equitable uptake of HPV vaccine in the routine age group has been achieved through a school-based vaccination programme in Scotland. School-based delivery also effectively reached older girls in the catch-up programme. Recent experiences of adolescent vaccination in Scotland had led to an expectation that good coverage could be reached. These were the MenC catch-up in 2000 and a large-scale feasibility trial of three-dose Hepatitis B vaccine in 11-year-olds to 12-year-olds in 2004.10–12 The 80% uptake achieved in this latter exercise informed the modelling scenarios and cost-effectiveness work supporting the UK HPV programme introduction.1
Decreasing in-year uptake in the second and third years of the routine programme coincided with external factors and programmatic changes which are worth noting. In Autumn 2009 (year 2), cases of pandemic H1N1 influenza peaked between the scheduled first and second HPV dose. Vaccine uptake showed a temporary decline at this time, possibly related to sickness and absenteeism. In 2010, some Health Boards reduced the frequency of school vaccination days, from five to three, marking the end of enhanced vaccination activity for the catch-up programme. Hence, uptake of a complete course within the school year could not be achieved if any one dose was missed. Consequently, uptake of dose three in-year is comparatively low after 2010.
In our analysis, national datasets and large sample sizes have the statistical power to reveal significant variations in uptake over the factors considered. Using SIMD as an indicator of individual deprivation, there was a linear effect on in-year vaccine uptake in schools, particularly for the third dose. But this finding, while important, should not obscure the generally high uptake that was reached across all SIMD quintiles. Offering further vaccination opportunities in the next school-year for those who had missed doses had an important levelling effect on the equity of vaccine coverage.
In Scotland, more than 15% of pupils attend denominational schools. There was no association between school denomination and uptake. Elsewhere in the UK and, more markedly, elsewhere in Europe, religious belief has been identified or predicted as a negative influence.13–17 Our findings are more in line with the conclusions from a recent rapid evidence assessment of school-based HPV vaccination, which found that inequalities in uptake were mainly related to income and other social factors rather than ethnicity and/or religion.18
Despite the relatively low uptake in school-leavers, the catch-up vaccination coverage in Scotland is much higher than in most other countries with an HPV programme.19–21 Overall, 65.5% of 13-year-olds to17-year-olds completed a course and 71% had two doses of bivalent HPV vaccine, for which there is some evidence of a protective effect.22
There was a marked difference in the uptake of vaccine for those in school compared to those who had left. Overall, 50% of eligible school-leavers started and 31% completed the course within a year. The final uptake figures will be a little higher as catch-up was available to any of the eligible cohort until 31 August 2011 and remains available for all girls until their 18th birthday.
The arrangement made by each Health Board to run clinics, engage local GPs or deploy a mixed model depended on the geographical characteristics of the Board and local GP negotiations. At the national level, there was no clear advantage between GP-led delivery and Board-run clinics in reaching these teenagers. There was a 10% difference in uptake according to delivery model. Board-run clinics appeared to be more effective in year 1, but the converse was seen in year 2. It is not clear whether this difference is to do with a change in delivery, or differences in the age and socioeconomic profile of the group being targeted between the 2 years. In 2009/2010, eligible school-leavers were older (median age 17 years 3 months) and there were twice as many girls in SIMD-1 than SIMD-5. In 2010/2011, the smaller target cohort was predominantly new school-leavers, their median age was 16 years 7 months, and there were over 3.5 times as many girls in SIMD-1 than in SIMD-5. Previous vaccination of 16-year-old to 17-year-old school-leavers in the single-dose MenC catch-up10 not only had a generally better uptake, but also showed a difference between the older and younger school-leavers, 72.3% in 17-year-olds compared with 67.6% for those born a year later.
Study strengths and weaknesses
National, high-quality disaggregate data, collected in a standard way allows comprehensive analysis of uptake across the whole country over time. In the routine cohort, it allows the effect of late catch-up of missed doses to be measured. Some similar catch-up might be assumed for older girls, both in and out of school, but cannot be easily measured as the national vaccination data starts to lose its precision for older, more mobile age groups. Denominators and numerators are affected by in- migration and out-migration to and from Scotland.
Within these routine datasets, there are a limited number of variables to study. Although an overall indicator of deprivation, SIMD is an indirect measure that provides limited qualitative insights into behaviours that influence late and non-uptake of vaccine.
There was no individual level or other reliable school level measure of ethnicity to allow comparison with other UK studies which have shown an association with uptake.
Board-level data on the model of delivery to school-leavers do not allow clear conclusions on which, if any, is a better method to reach school-leavers. However, it may be that, in a country with such diverse geographies and very different social needs between Health Boards, a one-size answer will not emerge.
Vaccines given privately will not reliably be included in the uptake figures. Some will be recorded by chance, but the systems operate to manage and monitor the centrally funded vaccination programmes.
The target age for routine use of this vaccine aims to protect recipients before they are exposed to HPV. Careful follow-up is now important to assess the impact of the programme on levels of HPV infection and cervical precancers. In the early years, this will be complicated by the mix of unvaccinated and partially vaccinated individuals and the fact that vaccinated or not, some, particularly in the catch-up cohorts, will have already been exposed to high-risk HPV infection. Preimmunisation surveillance studies in an unvaccinated population of Scottish teenagers found that 6.5% of 15–18-year-old girls had evidence of HPV-16 and/or HPV-18 infection.23
In recent years, there has been an increasing focus on vaccinating adolescents to aid the control of measles, mumps, pertussis and meningococcal infections, among others. The issue of how to effectively deliver immunisations to this age group has become increasingly important.24 The results reported here reinforce the previously reported advantages of school-based programmes in achieving high uptake.25 However, for those who have left school, it is not clear if delivery in GP settings is any more successful than clinics specifically designed for vaccination.
In Scotland, incidence of cervical cancer remains twice as high for the most deprived compared with the least deprived,26 and achieving vaccine equity was important from the outset. This has been largely achieved particularly with the routine programme. It provides more support for the effectiveness of school-based programmes, not just in maximising uptake but in reducing the socioeconomic variability. The availability of repeated opportunities to catch up on missed doses at school contributes to this and is a reflection of the patterns seen earlier in childhood, where missed Measles, Mumps and Rubella (MMR) doses were caught up over time and late vaccination was related to deprivation.27
School-based vaccination can achieve high and equitable uptake of a multidose vaccination programme. Uptake has been resilient to external factors. Offering additional opportunities for missed doses is effective in minimising socioeconomic variation in the uptake of routine HPV immunisation in girls.
What is already known on this subject
Several studies have observed and predicted that in-year uptake of HPV vaccine is influenced by deprivation, ethnicity and religious belief. Of these, deprivation has consistently been shown in practice to be associated with differential uptake.
What this study adds
High uptake of a multidose HPV vaccine is sustained over time through a universal school-based programme. Late uptake of missed doses improves course completion and equity of coverage before adolescents are widely exposed to HPV infection.
Contributors KS led and coordinated this study; KK and CR provided expert statistical advice and undertook the statistical modelling; RG collated, prepared and provided the data; JL undertook additional data management and linkage; AP and MD contributed to the interpretation and all contributors helped to write the manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.