Journal of Public Health Advance Access published online on March 13, 2008
Journal of Public Health, doi:10.1093/pubmed/fdn017
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Coverage in the National Health Service Breast Screening Programme, 1996–2005: correcting for the first invitation of women between 50 and 52 years
R. L. Bennett, Research Fellow
R. G. Blanks, Senior Staff Scientist
S. M. Moss, Associate Director
Cancer Screening Evaluation Unit, Institute of Cancer Research, 15 Cotswold Road, Sutton, Surrey SM2 5NG, UK
Address correspondence to Rachel Bennett, E-mail: rachel.bennett{at}icr.ac.uk
Background Coverage measures the ability of the National Health Service Breast Screening Programme (NHSBSP) to reach the eligible population and has a target of 70%.
Objective To estimate coverage accurately for women aged 50–64.
Methods Routine data from the KC63 return were used to calculate coverage for women aged 50–64 using an adjusted method that allows for the fact that women receive a first invitation to screening between 50 and 52.9 years.
Results The adjusted average coverage between 1996 and 2005, for women aged 50–64 was 74.3% and the standard unadjusted average measure for the same period was 68.3%. Therefore, previous measures of coverage for this age group have underestimated coverage by 
9% and the adjusted figure is actually well above the target.
Conclusion In terms of coverage the programme has been performing better than previously reported. It is important to monitor the effect of an increasing workload on the programmes ability to re-invite women within three years of their last screen as maintaining coverage is an important factor in ensuring that the NHSBSP is effective in reducing mortality from breast cancer.
Keywords: cancer, screening
 |
Introduction |
|---|
 TOP Introduction MethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
In England, women aged between 50 and 70 (previously 64) are invited to breast screening at three-yearly intervals. The success of the National Health Service Breast Screening Programme (NHSBSP) in reducing mortality from breast cancer is dependent on a number of factors including that a high proportion of women from the target population are screened.1 The ability of the NHSBSP to screen its target population is measured by coverage and relates to the screening activity over the previous three years, for example, coverage for 1996 relates to women screened between 1 April 1993 and 31 March 1996, who were resident in an area on the 31 March 1996.
The percentage of invited women who are screened by the programme is measured by uptake and has a minimum standard of 70% or more. If we assume that a three-yearly screening round is maintained, and that all eligible women are invited to screening, this target can also be applied to coverage. Since the programme began both uptake and coverage have been relatively stable, and the target has always been met for uptake but generally not coverage.2–5 Coverage has been calculated for women aged between 50 and 64. As women are first invited for screening between 50 and 52.9 years of age rather than their 50th birthday, approximately only half of women aged 50–52 would be expected to have been screened within the past three years. This has previously not been taken into account although it has been highlighted by Shah and colleagues.6
This paper presents coverage, for the 10-year period from 1996 to 2005 for the 50–64 age group, using a correction factor that allows for the fact that women are invited to screening between 50 and 52.9 years of age.
|
Methods |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
Coverage in the NHSBSP is defined as the proportion of eligible women who have been screened by the programme in the previous three years. Women may be ineligible for screening because they are terminally ill, have had a bilateral mastectomy or have withdrawn their consent to be invited. Coverage can be calculated from the KC63 statutory return, which is generated annually by primary care organizations from their population databases, and summarizes call and recall activity. The return presents aggregated data for the age groups 50–52, 53–54 (previously combined as 50–54), 55–59 and 60–64.
Using routine data from the KC63 return, we calculated coverage in the 50–64 age group for each year from 1996 to 2005. However, women are first invited for screening between 50 and 52.9 years of age i.e. at a mean age of 51.5 years. The true eligible population aged 50–54 (i.e. those eligible and due for invitation in the three-year period) is not known. Changes to the data collected on the KC63 return since 2002 now allow women aged between 50 and 52.9 to be excluded from the calculation of coverage as all eligible women aged 53–54 will be due for invitation.
We estimated adjusted coverage for the 50–64 age group by calculating the ratio of the invited population to those eligible for invitation for the age group 55–64 and applying this to the eligible population aged 50–54, in order to give an estimate of the true eligible population aged 50–54 and hence the total eligible population aged 50–64 (method outlined below).
Step1: Percentage of eligible women aged 55–64 years invited to screening in <3 years
|
|
We calculated the number of women invited as a proportion of the eligible population for the 55–64 age group.
Step2: Estimating the true eligible population aged 50–54 years
By assuming that the pattern of invitation for the 55–64 age group to be the same as the 50–54 age group then,
|
|
This formula can be re-arranged, and substituted in the standard formula for coverage as shown below.
Step3: Coverage with adjustment for the 50–54 age group
|
|
The estimates of adjusted coverage were validated, using primary care level data, by calculating adjusted coverage for the 50–64 age group after 2002 and comparing it with that for the age-group 53–64. The agreement between the two estimates was tested using a paired t-test with a null hypothesis that the difference between the two was zero.
|
Results |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
Table 1 shows coverage by age group for the 10-year period from 1996 to 2005, including the 53–64 age group for 2002 onwards. Average coverage in the 50–64 age group during the 10-year period was 68.3%, increasing from 65.4% in 1996, to a peak of 70.2% in 2001 and decreasing slightly in more recent years. Adjusted coverage for the same period was 74.3%, and was more than 70% in each of the 10 years.
|
Average coverage between 2002 and 2005 was 75.4% in women aged 53–64; 75.2% in the 53–54 age group and 75.4% in the 55–64 age group. In women aged 50–52, average coverage was lower (46.9%) and for the same time period it was 69.4% for the age group 50–64. In each of the age groups, coverage showed a similar trend over the four-year period.
Comparison of unadjusted and adjusted coverage for the 50–64 age group suggested that crude estimates of coverage have underestimated true coverage by 9%, while comparison of the 50–64 and 53–64 age groups suggested that estimated coverage in the 50–64 age group underestimated coverage in the 53–64 age group by 8%. Our validation of the estimates of adjusted coverage (for the 50–64 age group) suggested that they underestimated coverage in the 53–64 age group by <1%, which therefore suggests that calculations of adjusted coverage prior to 2002 can be used alongside more recent calculations of coverage for the 53–64 age group to permit examination of longitudinal patterns in coverage at both a national and more local level [e.g. by primary care trusts (PCT)].
|
Discussion |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
Main finding of this study
Coverage in the NHSBSP has always been significantly lower than both uptake and the target of 70%, mainly because it could only be crudely calculated for the 50–64 age group and therefore has underestimated the true coverage. Changes to routinely collected data have allowed us to show that, by measuring coverage in the 53–64 age group, and by adjusting earlier estimates of coverage, the programme is and has been meeting the target.
What is already known on this topic
Coverage is lowest in urban/inner city areas, as these areas have low uptake. Reasons for low uptake may include the ethnic mix of the population, the mobility of the population and socio-economic status.7 A mobile population may result in incorrect addresses, the presence of ‘ghosts’ on PCT lists no longer resident and conversely the existence of women in the area not registered with a general practitioner.8 Low coverage is also evident in a small number of other areas with good uptake but poor round length (defined as the percentage of eligible women whose first offered appointment is within 36 months of their previous screen) and has been reported to be as a result of difficulties in recruiting radiologists and radiographers.9
In 2001, units began to implement the expansion of the programme outlined in the NHS Cancer Plan for England.10 This increase in workload has the potential to increase round length, and thus to lower coverage. However, the results of an observational study to estimate the effect of moving to two views at the incident screen did not suggest that the increase in workload did increase round length.11 This may be as a result of the programme expansion attracting additional resources.
What this study adds
The programme has always met the target for uptake of 70%.3–5 We have now also shown that the programme has also been meeting the target for coverage. Furthermore, the technique used in this paper can be modified to take into account the fact that, as a result of the recent announcement of the extension of screening to women aged 47–73, in the future women will be first invited to screening between 47 and 49.9 years of age.12
Careful monitoring of coverage in future years should continue, particularly in areas with difficulties in recruiting staff. This will help ensure that the increase in workload resulting from further expansion of the programme as well changing demographics does not impact on factors which influence coverage such as round length, as maintaining coverage is an important factor in ensuring that the NHSBSP is effective in reducing mortality from breast cancer.
|
Limitations of this study |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
Our analyses were limited by the use of routinely collected aggregated data, and by the fact that these data changed during the 10-year period. The detailed data now collected on coverage will allow it to be calculated more accurately in future.
|
Funding |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
This work was funded by the NHS Cancer Screening Programme. The Cancer Screening Evaluation Unit also receives funding from the Department of Health Policy Research Programme.
|
Acknowledgements |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
The views expressed in this publication are those of the authors and not necessarily those of the Department of Health or the NHS Cancer Screening Programme.
|
References |
|---|
|
TOPIntroductionMethodsResultsDiscussionLimitations of this studyFundingAcknowledgementsReferences |
|---|
- Day NE, Williams DRR, Khaw KT. Breast cancer screening programmes: the development of a monitoring and evaluation system. Br J Cancer (1989) 59:954–8.[Web of Science][Medline]
- The Information Centre. Breast Screening Programme, England: 2005–06. (2007) Leeds: The Information Centre.
- Moss SM, Michel M, Patnick J, et al. Results from the NHS breast screening programme 1990–1993. J Med Screening (1995) 2:186–90.[Medline]
- Blanks RG, Moss SM, Patnick J. Results from the UK NHS breast screening programme 1994–1999. J Med Screening (2000) 7:195–8.
[Abstract/Free Full Text] - Bennett RL, Blanks RG, Patnick J, et al. Results from the UK NHS breast screening programme 2000–05. J Med Screen (2007) 14:200–4.[CrossRef][Web of Science][Medline]
- Shah S, Roche T, Henderson G. NHS performance tables for breast screening. Lancet (1998) 351:529.[Web of Science][Medline]
- The All Party Parliamentary Group on Breast Cancer. Early Detection Saves Lives: Tackling inequalities in breast screening. In: London (2004).
- Chamberlain J, Moss SM, Kirkpatrick AE, et al. National Health Service breast screening programme results for 1991–2. BMJ (1993) 307:353–6.
[Abstract/Free Full Text] - Eaton L. Breast cancer detection rates increase but coverage varies. BMJ (2005) 330:500.
[Free Full Text] - Department of Health. The NHS Cancer Plan. A Plan For Investment. A Plan For Reform (2000) London: Department of Health.
- Blanks RG, Bennett RL, Patnick J, et al. The effect of changing from one to two views at incident (subsequent) screens in the NHS breast screening programme in England: impact on cancer detection and recall rates. Clin Radiol (2005) 60:674–80.[CrossRef][Web of Science][Medline]
- Department of Health. Cancer Reform Strategy (2007) London: Department of Health.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||