Background Achieving high human papillomavirus (HPV) vaccine coverage may reduce inequalities in cervical cancer prevention by mitigating the inequalities seen in the cervical screening programme. This paper assesses whether the same sociodemographic factors are associated with both cervical screening and HPV vaccination.
Methods Girls’ HPV vaccination records were linked by address to cervical screening records for their mothers in the North West of England. Index of Multiple Deprivation scores (2010) and census ethnicity data (2001) were used to investigate the association between deprivation and ethnic composition of area of residence with HPV vaccination and cervical screening uptake, along with potential differences between Primary Care Trusts (PCTs), which were responsible for vaccine delivery.
Results Deprivation was not associated with routine (12–13-year-olds) vaccination initiation, but girls living in the most deprived quintile were significantly less likely to complete the three vaccine doses (OR 0.75; 95% CI 0.63 to 0.88). Mother–daughter pairs failing to engage in either screening or vaccination were also more likely to live in deprived areas (routine vaccination OR for most deprived quintile: 2.35; 95% CI 2.00 to 2.77). There were differences between PCTs after controlling for demographic effects (OR 1.35; 95% CI 1.23 to 1.52).
Conclusions Ensuring completion of the vaccine schedule is critical for organisations responsible for vaccine delivery in order to reduce cancer risk among girls living in deprived areas. There remains a small minority of mothers and daughters from disadvantaged backgrounds who do not participate in either cervical screening or HPV vaccination.
- CANCER: CERVIX
Statistics from Altmetric.com
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.
Studies in a number of countries have demonstrated that the risk of developing cervical cancer is greater among low socioeconomic groups1–3 and Asian and Black women over 65 years.4 Sociodemographic variance in women's use of, and access to, cervical screening contributes to this disparity.1–3 5–7 The introduction of the National Health Service (NHS) Cervical Screening Programme in England has halved cervical cancer incidence8 and reduced inequalities in the delivery of the screening programme.1 This is mainly attributed to the call–recall system that sends invitations for a free cervical screening test every 3 years to women 25–49 years, and every 5 years to women 50–64 years.9 National cervical screening coverage has, nevertheless, been declining since 1995,10 falling below the target of 80 to 78.6% (2011/2012).11 The reasons for this are complex and range from patient choice, growing distrust of health services and lower perceived risk in older women.11 Variation in service accessibility and structure, in addition to poor knowledge and understanding, is thought to contribute to low uptake in deprived areas with further inequalities by ethnicity.1 ,5 ,6 ,10 ,12–14
The North West is one of the most deprived regions in England,15 with an average screening uptake (<5 years since last adequate test) of 78.4%.16 In contrast, the region achieved a human papillomavirus (HPV) vaccine uptake of 84% compared with only 76.4% for the England average (2009/2010).17 Routine HPV vaccination of 12—13-year-old girls was introduced in 2008, using the bivalent Cervarix vaccine. This provides protection against high-risk HPV types 16 and 18, which account for ∼70% of cervical cancers18 and with additional cross-protection against types 45 and 31, protecting against 80% of cervical cancers.19 Low HPV vaccine coverage is likely to perpetuate social inequities,7 but it is not yet clear whether high vaccination coverage will reduce disparities in cervical cancer incidence that currently relate to inequalities in screening coverage.20
A mother's engagement and participation in the cervical screening programme is associated with her daughter's HPV vaccination status.21–25 By linking mothers’ cervical screening records to daughters’ HPV vaccination records in the North West of England, a recent study23 reported that daughters were more likely to have initiated the HPV vaccination programme if their mothers had previously attended for screening, with the association strongest when mothers had attended within the last 5 years (adjusted OR (AOR) 3.5; 3.1 to 4.0). This was thought to relate to the role of the mother in consenting for the vaccination programme on behalf of her daughter. As such it is possible that the same sociodemographic groups will remain at risk of cervical cancer as a result of poor participation in both programmes.
Building on this earlier work,23 the purpose of this analysis was to investigate whether the same sociodemographic factors predicted cervical screening participation by mothers, and HPV vaccination by their daughters, in the North West of England, and the extent to which uptake of these interventions varied by commissioning/delivery organisation (Primary Care Trust (PCT)). A unique aspect of this research is that these factors were assessed in the same sample of the same population and therefore allow an examination of the factors associated with missing both interventions.
The study was conducted in the NHS North West region covered by 24 PCTs, which, at the time of the study, were responsible for the commissioning of screening and delivery of immunisation services. The region had a population of 1.78 million women eligible for screening (2009/2010)16 and ∼42 000 girls aged 12 who were eligible for HPV vaccine.17 The routine vaccination of girls 12–13 years for HPV was included as part of the childhood immunisation programme from September 2008 with a time-limited catch-up programme delivered within schools and general practices (GPs) to girls 14–18 years born on, or after, 1 September 1990.26 The North West region has the third largest proportion of non-white ethnic groups (8% Census 2001).27 There is also a North–South divide in England in terms of social class, northern regions having higher proportions in lower social classes among most ethnic groups.28 Large variations exist in the ethnic distribution between PCTs in the North West.29
The selection of the North West region was greater than that indicated by statistical necessity, allowing for representation of a wide range of population characteristics.
The project was considered by the North West 12 Research Ethics Committee to be service evaluation and, as such, ethical approval was not required. As patient-identifiable data were required from PCTs, Caldicott approvals were obtained in order to release data to Lancashire and South Cumbria Agency (LaSCA), an NHS organisation that operates as a partnership agreement between PCTs in the North West to deliver administrative support services.
The combined mother–daughter dataset developed in the previous study23 was used, which successfully obtained linked mothers’ and daughters’ records by home address for 20 PCTs. This incorporated HPV vaccination data from PCT Child Health Systems (CHS) and cervical screening records from the National Health Authority Information System (NHAIS). A mother was designated as an adult female of eligible screening age (25–64 years) living at the same address as girls scheduled for HPV vaccine in 2008 and 2009. This included girls eligible for vaccination with the routine programme, aged 12–13 years on September 2008 (date of birth (DOB) 1/9/1995 to 31/8/1996) and September 2009 (DOB 1/9/1996 to 31/8/1997). Data were also obtained for girls aged 14–16 years (DOB 1/9/1992 to 31/8/1995) included within the catch-up programme. Analysis was restricted to addresses where only one adult female and one vaccine-eligible girl within a particular vaccination cohort were living at an address. Further details regarding this selection process can be found in Spencer et al.23
In order to assess the role of sociodemographic factors on cervical screening uptake in the wider population, anonymised cervical screening history for a random sample of women not linked to a girl of vaccination age was also extracted from the NHAIS system with otherwise identical inclusion criteria.
The following sociodemographic variables were considered:
PCT: This was based on area of residence and was common to mothers and daughters.
Deprivation: The Index of Multiple Deprivation (IMD) 2010 associated with the lower super output areas (LSOA) derived from the address postcode.15 The indices of deprivation are derived from aggregated measures of income, employment, health, disability, education, crime, barriers to housing and services and living environment from the UK 2010 census. Deprivation is presented as quintiles based on the distribution in each cohort. Quintiles were computed using the whole sample and this quintile used as a covariate in the analyses.
Ethnicity: Ethnicity is not recorded within the NHAIS or CHS systems; therefore, an area-based ethnicity density variable was used to consider the proportion of each ethnic group in the LSOA of residence. This was derived from the UK 2001 census data.29 Aggregated proportions of White, Mixed, Black, Asian and Other (“Other” includes Chinese) ethnic groups were used based on the UK census definitions.29
The outcomes measured were
Cervical screening attendance in the last 5 years. Although younger women are invited every 3 years, a 5-year screening interval was chosen as all women of eligible screening age should have attended within this time period. Five years is also the measure used to monitor the overall uptake of the NHS cervical screening programme.16 For linked women, this was the 5 years prior to their daughters’ vaccination (prior to September 2008 or 2009) and for non-linked women, prior to 2009.
HPV vaccination initiation, receipt of at least one vaccine dose.
HPV vaccination completion, receipt of all three doses (retention in the programme).
Failure to participate in either intervention, mother not having attended screening in the last 5 years, as well as daughter not having initiated HPV vaccination.
Logistic regression was used to quantify the association between IMD quintile, ethnicity (OR are expressed per 10% of each minority in the LSOA) and PCT on the cervical cancer prevention outcomes, with 95% CIs. Significance levels were adjusted for all other variables in the model. The effect of PCT is summarised as a median OR (MOR) derived from refitting the model with PCT represented as a random effect. The MOR is an estimate of the median of the OR between the higher and lower of any two randomly selected PCTs30 and therefore represents a ‘typical’ difference between two organisations.
Analysis was performed using IBM SPSS V.20.0 and Stata V.12.31
The demographic characteristics of the three groups of women are shown in table 1. The sample was predominantly white (∼93%), and the ethnic distribution was similar across all groups. The sample of women and girls covers 99% of the LSOA areas in the North West.
Overall screening uptake was 83.6% for mothers in the routine group, 84.5% for mothers in the catch-up group and 76.3% for non-linked women. 82.4% of girls in the routine group initiated vaccination and 76.1% in the catch-up group. The proportion of those who initiated the vaccination programme who went on to complete all three doses was 95.6% and 87.0% for the routine and catch-up groups, respectively. The proportion of mother–daughter pairs in which mothers had not attended for screening in the last 5 years and daughters had not initiated the vaccination programme was 4.0% for the routine group and 5.1% for the catch-up group.
Cervical screening uptake was strongly associated with deprivation across all groups with ORs of ∼0.4 for the most deprived compared with the least, with similar ORs in linked mothers and unlinked women despite the small difference in overall uptake (table 2). Weaker associations were seen for mothers living in areas with high ethnic minority density, where effects were largely confined to the mixed and other categories and negligible for mothers linked to girls in the routine vaccination cohort. There were only very small differences between PCTs (MOR 1.1).
There were weak associations between vaccination initiation and deprivation in the routine group, although the associations were a little stronger in the catch-up group (OR 0.73; 95% CI 0.68 to 0.79 in the most, compared with the least, deprived). It is of interest to note that there is a small but statistically significant effect (OR 1.09; 95% CI 1.00 to 1.17), whereby uptake in the routine group is higher in the most deprived compared with the least. There were more substantial differences in uptake in areas with high ethnic minority density. Girls living in areas with high proportions of Asians in the routine group and Black and Other ethnic groups in both routine and catch-up vaccination groups were all significantly less likely to initiate vaccination. The PCT of residence had a strong association with vaccine uptake (MOR 1.47, 95% CI 1.32 to 1.69 in the routine group; and MOR 1.67, 95% CI 1.45 to 2.01 in the catch-up group) after adjustment for deprivation and ethnicity (table 3).
For those who initiated the programme, there was a clear association between non-completion and deprivation (OR 0.75, 95% CI 0.63 to 0.88 between the most and least deprived in the routine; and OR 0.64, 95% CI 0.57 to 0.71 in the catch-up group) (table 3). Ethnic differences were not apparent in the routine group, but mixed and other ethnic groups did show evidence of reduced completion rates in the catch-up group (OR 0.51, 95% CI 0.33 to 0.79; and OR 0.72, 95% CI 0.52 to 1.00). There were strong differences between PCTs in completion rates (MOR 1.51, 95% CI 1.34 to 1.78; and MOR 2.19, 95% CI 1.77 to 2.91 for the routine and catch-up groups, respectively).
Failure of mother–daughter pairs to participate in either intervention
There was a very strong association between deprivation and the combined failure of both mother and daughter to participate in cervical cancer prevention, with OR 2.35 and 95% CI 2.00 to 2.77 between the most and least deprived in the routine group and OR 2.77 and 95% CI 2.40 to 3.19 in the catch-up group (table 4).
Across all minority ethnic groups there was an association with living in an area of high minority density and greater risk of missing both interventions, particularly in the mixed and other groups. This joint failure was strongly associated with PCT of residence (MOR 1.4).
We found that dual non-engagement of mothers and daughters in cervical cancer prevention programmes was twice as common in the most socially deprived areas and affected both routine and catch-up vaccination cohorts. Even in the routine cohort, girls from the most deprived backgrounds who started the course were less likely to complete the three doses. Differences between PCTs in vaccination uptake after controlling for sociodemographic effects suggest organisational factors also contribute to non-engagement.
For girls in the routine vaccination group, vaccination initiation was actually highest in the two most deprived IMD quintiles, although the effects are very small (OR 1.1). Earlier acceptability and feasibility studies found that more affluent parents were least likely to consent to the vaccine32 ,33 but that completion was lower in deprived groups.32 One explanation for higher initiation among deprived girls could be school-based delivery of the vaccine, including the efforts made by school nurses to obtain parental consent34 ,35 and the non-requirement for parents to organise the first appointment. Girls vaccinated routinely at school were more likely to complete the vaccine course than girls in the catch-up programme, but school delivery did not entirely prevent deprivation-associated drop-out. Vaccination completion may be impacted by irregular school attendance, which is higher in socially deprived groups.36 An aggregated analysis of PCT data for the first year of the HPV vaccination programme (2008) by Kumar and Whynes37 also found that relative deprivation as defined by IMD was associated with lower two-dose uptake of girls vaccinated in the catch-up programme (aged 17–18 years) compared with girls in the routine programme (aged 12–13 years). As the catch-up programme was more likely to be delivered using a variety of methods via school, GP-based programmes or both,17 ,38 this may explain why both vaccine initiation and completion in older girls remained associated with deprivation. However, school-based delivery was still the primary method used by the catch-up age group in the present study (91% girls compared with 94% in the routine group).17 ,38
Although there were only small differences between PCTs in cervical screening uptake, much stronger associations were seen in vaccination initiation and completion, as well as in the joint non-compliance of mother and daughter pairs. This suggests that additional organisational factors affect mothers’ and daughters’ engagement with cervical cancer prevention. Details of PCT delivery methods were not collected as part of the current study, but it was known that some PCTs used social marketing campaigns and reminder letters to promote dose completion. A study of three PCTs in South West England39 also reported relatively weak associations between vaccination uptake and social deprivation but found uptake to vary by area ethnic composition and with PCT. Higher uptake was reported in the PCT that funded an additional member of staff to encourage participation and follow-up missed vaccine doses in the community. The positive role of school nurses in approaching non-attenders and vulnerable groups illustrates how the delivery strategy of the vaccination programme can increase uptake and reduce inequalities.35
Asian and black women have previously been identified as at greater risk of developing cervical cancer4 and, with mixed ethnicity, more susceptible to lower screening coverage.6 ,10 A less clear association was found with cervical screening in this study. However, in common with another study,40 we found that girls living in areas with high concentrations of Asian, black and other ethnic minorities were significantly less likely to initiate the vaccination programme. Significantly, in these areas mother and daughter pairs failed to participate in either preventive programme. Possible reasons for this could relate to religious and cultural concerns about sex before marriage or relatively low levels of awareness within certain minority groups.40
Limitations and strengths
We used area-based, not individual-level, measures for both deprivation and ethnicity as individual ethnicity was not available from either data source. We adjusted by PCT to allow for the potential for PCT performance to affect the uptake of both interventions. Deprivation and ethnicity are clearly correlated with PCT, which could potentially lead to underestimates of these effects. However, the removal of PCT from analysis did not substantially change the effect estimates (data not shown). Although 95% of the girls were matched to a putative mother, a further 10.3% had to be excluded due to multiple matches.23 The use of linked mother and daughter pairs controlled for a range of unmeasured and unmeasurable factors potentially confounding comparisons of the two interventions as well as allowing the assessment of the potential for jointly missing both interventions. In addition, by considering screening uptake in both mothers and unlinked women, we were able to demonstrate the sociodemographic factors influencing screening uptake in the general population and among mothers of girls of vaccination age.
Despite the existence of an organised call–recall system within the cervical screening programme, inequalities in participation persist. The lack of association between HPV vaccine initiation and social deprivation in the routine HPV vaccine cohort indicates that delivery of the HPV vaccination programme within schools has the potential to reduce inequalities in cervical cancer prevention, provided three-dose completion remains a priority. Local delivery and organisational factors are central to achieving this because there remains a small minority of mothers and daughters from disadvantaged groups who are resistant to participation in both cervical cancer prevention programmes. A joint organisational focus on both the vaccination and screening programmes is required, without which such disadvantaged groups will continue to be at greatest risk of cervical cancer.
What is already known on this subject?
Women who are socially deprived have a higher cervical cancer risk.
Cervical screening coverage overall is declining.
Mothers participation in cervical screening has been shown to be associated with her daughter's HPV vaccination status.
What this study adds?
High human papillomavirus (HPV) three-dose vaccination coverage, including girls from disadvantaged backgrounds has the potential to reduce inequalities in cervical cancer prevention provided focus remains on three dose completion.
Failure to accept either cervical screening or HPV vaccination (4% of the sample) was highest among socially deprived compared with affluent groups, irrespective of whether the vaccine was offered routinely or in the catch-up programme.
Such resistant mothers and their daughters may remain at high risk of cervical cancer.
We would like to acknowledge Suzanne Milward, LaSCA, for her role in creating the database and Dr Peter Elton, NHS Bury, for his advice and support.
Contributors LB and SAR conceived and designed the study. AMS conducted the investigations, created the data set, undertook the analysis and drafted the paper. All authors contributed to reviewing and producing the final manuscript.
Funding The study was funded by NHS Cancer Screening Programme. AMS is funded by NHS Bury. LB is funded by the Max Elstein Trust. Researchers at the University of Manchester receive support from the Manchester Academic Health Science Centre and the Central Manchester University Hospital NHS Trust.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement NA. Further analysis is continuing to be undertaken by the research group.