Abstract
Objectives
To examine current procedures for cancer screening invitation list (SIL) checking in primary care, and to make recommendations for the future use of these procedures.
Setting
Cancer screening departments/units and associated general practices across England.
Methods
1) An analysis of the outputs of screening programme SIL checking, and accompanying practice questionnaire, for cancers of the cervix (9 screening centres, 36 general practices), breast (6 centres, 76 practices) and bowel (pilot hub, 7 practices)-supplemented by an audit of calls to screening centres to identify inappropriate invitations; 2) a national postal questionnaire survey sent to all 80 breast screening departments across England and 320 associated general practices; 3) telephone interviews with 13 NHS screening staff to obtain detailed perspectives about SIL checking procedures.
Results
SIL checking in primary care is undertaken by a variety of clinical and non-clinical staff. It plays a useful role in cervical screening with tangible evidence of refinements to the SIL and support from both primary care and screening centre staff. Conversely, its role in breast and bowel screening is not supported by the results of this study. Overall, there is no significant evidence of adverse effects from inappropriate invitations.
Conclusions
SIL checking in primary care for cervical cancer should continue, but its use in breast and bowel cancer screening is not supported by our results. New ways of undertaking the SIL checking process to make it more accurate and less burdensome should be examined.
Introduction
Exclusion criteria for cancer screening programmes
Exclusion criteria for cancer screening programmes
Although an established feature of the cervical and breast screening programme, the role of primary care in the verification of screening invitation lists is poorly understood; recent surveys of cervical SIL checking processes identified a lack of consistency in removing women from SILs, and up to 40% inaccuracies in SIL checking.10, 11 Similarly, the thoroughness of primary care-based SIL checking procedures for breast screening has been questioned; time and resources appear to constrain activity.12, 13 In the present funding environment, general practice activities such as SIL checking are attracting increasing scrutiny; 14 it is not clear whether the time spent on the process is warranted.
This study examines the utility and feasibility of SIL checking in primary care for the NHS cervical, breast and bowel screening programme.
Methods
A mixed methods approach was used, comprising an analysis of SIL checking procedures (for all 3 programmes), and, for cervical screening, a questionnaire survey and qualitative interviews with screening staff.
Cervical
Nine cervical screening departments serving areas with a range of Sociodemographic characteristics were recruited to the study (Birmingham, Buckinghamshire, Croydon, London, Newcastle, Preston, South Leicestershire, West Sussex, West Yorkshire): 36 general practices within these areas (four from each screening department) were also invited to participate.
Breast
Six regional breast screening units were recruited, again on the basis of the diversity of their geographical location and urban/rural coverage (Somerset; Warwickshire, Solihull & Coventry; Greater Manchester; Northampton; Nottingham; South West London). Seventy-six GP practices within the jurisdiction of these six regional breast screening units were invited to participate.
Colorectal
Seven general practices in the Coventry and Nuneaton areas were identified through the English Bowel Cancer Screening Pilot, based in Rugby.
All general practices were invited to participate in the study by the screening department/unit. As per usual practice, details of patients in the practices to be invited for screening were obtained by the cervical screening departments, breast screening units and the bowel screening Pilot from the local NHS database. SILs were generated and sent to general practices between two and four weeks before the screening invitations were issued to patients registered at that practice. In general, the breast screening programme invites eligible women in a practice once every three years based on the availability of the mobile unit; cervical screening of eligible women in a practice takes place at set frequencies based on a woman's age and the time of her last test or non-response; screening in the bowel Pilot was on a practice by practice basis. General practices were asked to exclude people as appropriate, based on their current knowledge of their patients. SIL checking processes were evaluated through two key sources of data:
Analysis of data from the SILs and associated questionnaires
These were assessed against the national criteria for appropriate exclusions (listed in Table 1) by staff at the cervical screening departments, breast screening units and Bowel Screening Pilot. The number of patients excluded from SILs and reasons for the amendments were examined. SILs were accompanied by a questionnaire (sent to the practice manager) which examined process issues (date SIL was checked, which staff member had done so, how much time was spent on this task) and the number and nature of amendments.
Audit of all phone calls concerning inappropriate invitations to screening
Staff in the screening departments and units recorded phone calls or letters regarding invitations from the practices in the study and noted the reason an invitation was thought to be inappropriate on standard proformas. Local screening units were asked to provide a copy of local SIL checking guidance issued to their general practices.
National survey of current procedures in cervical screening
Two postal questionnaire surveys examining SIL checking processes were carried out in parallel: all 80 cervical screening departments in England were asked to complete a screening department questionnaire, and to identify four general practices in their catchment area that they considered to have a range of competencies in completing SILs and to forward the GP questionnaire to them.
Interviews with cervical screening programme staff
Semi-structured telephone interviews were carried out with thirteen NHS staff purposively selected from a range of disciplines and with differing levels of responsibility and engagement with the SIL checking process or the cervical cancer screening programme, including a senior project manager, quality assurance directors/assistant directors in regional screening departments, screening service managers, and general practice personnel. The interviews explored attitudes towards the current operation of the SIL system and suggestions for modification to SIL checking. The opportunity to raise other pertinent issues and concerns was provided.
Analysis
Quantitative data: Data were analysed using SPSS version 11.5. Differences between two groups (i.e. participating and non-participating practices) were analysed using the Mann-Whitney test. Spearman's correlations were used to investigate the relationship between different aspects of the SIL check. A significance level of 0.05 was used throughout.
Qualitative data: All interviews were recorded and transcribed; transcripts of interviews were read through several times to identify the main issues, problems and concerns raised by staff. A Grounded Theory approach to analysis was adopted.
Results
Analysis of SIL checking procedures
Analysis of SIL returns and associated questionnaires
Details of SIL checking for in the cervical, breast and bowel screening case studies
Details of SIL checking for in the cervical, breast and bowel screening case studies
The number and nature of amendments made to SILs varied by screening programme:
Cervical: 1550 women were listed on the SILs; 92 amendments were made (5.9%). Seven practices made no amendments, six practices (22%) had between 1 and 3, four practices (15%) had between 4 and 6 amendments and seventeen (63%) practices returned over 7 amendments each. Only one was queried by the screening department: the patient's record had been updated since the SIL had been sent to the practice. The majority of cervical SILs were checked by non-clinical staff; in no instance did a GP check the SIL. The time taken to check the SILs varied considerably, ranging from 10 minutes to 8 hours. However, the time required did correlate with the number of patients on the list (R 2 = 0.3882), which was itself related to both the practice size, and whether the screening department issued the SIL weekly, monthly, or quarterly.
Breast: 16,017 women were listed on the 31 returned and completed breast SILs: 42 amendments were made (0.3%). The number of patients excluded from the SIL for each practice ranged from 0-3. All amendments met the valid criteria for ceasing/adding to the SILs and were deemed appropriate by screening unit staff. Checking the breast SIL was also shared amongst staff, and the time taken to check the SILs varied from 20 minutes to 13 hours, and was again significantly associated with the number of patients listed on the SIL (Spearman's rho = 0.681, p = 0.01).
Bowel: 1430 patients were listed on the two completed and returned bowel SILs: eleven amendments were made for these two practices (0.8%). Of the 79 exclusions in the practices where SIL checking was incomplete, 69 were to note patients moving from the practice. Checking of bowel screening SILs again involved a variety of practice staff and took between 40 minutes and 12 hours.
Audit of inappropriate invitations
Reasons for reports of inappropriate invitations to screening to Screening Programmes
One general practice received a query about an inappropriate invitation to a male patient: it is unknown if this patient had always been a male or had had a sex change operation
These women had died between the SIL being produced for their practice and the date the screening invitation was issued
Currently in hospital = 1; Has lung cancer =1
Questionnaires sent to all cervical screening departments (and their selected general practices) further explored SIL checking processes. Fifty-three of 80 screening departments (66%), and 140 of 320 general practices (44%) responded. The number of women eligible for screening ranged from 162 to 8371 in participating practices. Consistent with our SIL checking analysis, in these practices both clinical and non-clinical staff checked SILs, and time taken ranged from an hour to ‘over a day’. SILs are produced at differing frequencies by screening departments: nearly one in five send these out weekly while the majority are issued monthly. The extent to which the task of inviting women for screening and its associated paperwork is shared by screening departments and general practices also varies considerably with women from 65 practices (46%) being contacted by both. Of 140 practices, 122 (87%) considered SIL checking for cervical screening to serve a useful function.
Qualitative survey of cervical screening programme staff
Three main themes emerged from the interviews with NHS staff.
Perceived value of SIL checking procedures
All staff recognized the value of SIL checking processes as part of the cervical screening programme; this was true across screening department, general practice, and management personnel. The refinement of screening invitation lists in general practice, based on local knowledge of patients’ circumstances, was widely regarded as a strength of the process.
Suggested areas for improvement
Participants suggested it would be beneficial if patient-specific details were shared between the screening department and general practices more regularly; for example, if practice staff have received diagnostic results that would lead to a change in a patient's medical records, this new information should be available to the screening department to enable them to update their records. The need for ongoing training of general practice staff was raised by all screening managers and QA personnel, and was considered vital for the ongoing efficiency of the screening invitation process. Such training was considered particularly important given the increasing numbers of non-clinical staff who check the SIL, and where turnover of practice staff may mean new and untrained administrative staff are expected to perform this task.
Use of ‘Electronic SILs’
Electronic templates for the screening invitation lists (eSILs) have been developed to improve the ease of competing and reduce time and expense of SILs. 17 eSILs are welcomed by both screening departments and GP practices. Some practice staff did express reservations that they might not necessarily reduce the amount of time spent checking SILs, as the same diligence would be required to check each patient's record. Several interviewees expressed the view that if eSILs are to be successful, staff must again be adequately trained in the new processes involved.
Discussion
Summary of results
This study examined the role of primary care in screening invitation list checking within the cervical and breast cancer screening programmes and the bowel cancer screening Pilot in England. We found wide variation in the time given to SIL checking, partly dependent upon the size of the invitation list and how frequently it is issued. Both clinical and non-clinical members of staff in general practice are involved in SIL checking procedures, and authorizing amendments. SIL checking is usually accurate, and consistent with NHS guidance. Despite the shift in responsibility for recruitment and follow-up away from individual practices to screening departments, the importance of the role of primary care in refining invitation lists for cervical cancer is recognized. Few amendments were made to the lists of eligible patients for breast and bowel screening (mostly relating to changes of address) despite the considerable time the process took. Few invitees contacted the breast screening department or bowel screening centre to report an inappropriate invitation being issued.
Study limitations
This study relied on self-reported data from participating screening centre and primary care personnel. General practices involved in the study were not randomly selected. While efforts were made to obtain a representative sample, they may not have been typical-we found, for example, that practices that contributed data on breast screening had above-average mammography uptake rates amongst their patients.
Participation in the national questionnaire survey of cervical screening processes was low. It is likely also that many inappropriate invitations are not detected though auditing calls to screening centres or general practices: many invitees will ignore them. Self-reported perceptions of workload from screening can also vary considerably.18,19
Comparison with existing evidence
Our findings are consistent with previous studies which have highlighted variation in practice, and concerns about the accuracy and thoroughness of SIL-checking;10,11,12,13 these issues seem particularly pertinent where screening invitees are mobile (such as young women in large urban centres).
Further research
Results from this study have led to SIL checking being discontinued in the breast and bowel screening programmes in England. In contrast, SIL checking is still considered an important component of quality assurance within the cervical screening programme and has been maintained. Reasons for these differences include:
Cervical screening is, uniquely, delivered in primary care and uptake is rewarded in the Quality and Outcomes Framework;
20
practices are motivated to use accurate screening invitation lists; Primary care has little involvement in breast and bowel cancer screening in the UK: while the advice of GPs and practice nurses may be sought on an individual basis, there is no direct involvement in the delivery of screening; Invitations to cervical screening are issued to women in the eligible age range at the appropriate frequency from her last test or non-response. Thus general practices will have patients being invited all the time to this programme, with SILs being produced at regular intervals and being relatively brief: practice staff become familiar with cervical SIL checking criteria. This contrasts with breast screening where one large SIL is produced every three years: lack of familiarity with exclusion criteria and turnover of staff since the last breast SIL check may make complete and accurate checking more onerous.
Future research should examine how SIL checking could be further refined; IT-based approaches should, ideally, replace manual checking of lists. 21 A broader analysis might look at the role of primary care in other elements of cervical screening recruitment, including measures which are known to improve uptake in specific groups of patients where participation in cervical screening is low. 22
Issuing inappropriate invitations for screening has a range of potentially adverse effects on programmes and invitees; a specific issue our analysis identified related to women with learning difficulties being removed from screening invitation lists-further research work has been undertaken regarding how to extend and develop access to and informed choice of screening services for all eligible patients, irrespective of physical or learning disabilities. 23 However, the decision to cease SIL checking in the breast screening programme, and not to include this step in the national implementation of bowel screening, also has the potential to cause distress to patients, for example women who have undergone a bilateral mastectomy (over 50% of complaints of inappropriate invitations in this study, Table 3) or individuals currently being treated for breast or bowel cancer. It may be sufficient for the screening invitation letter to patients to clearly state that for some patients the invitation may not be relevant, but it is essential that the breast and bowel cancer screening programmes have clear mechanisms in place to deal in a sensitive manner with concerns or complaints arising from inappropriate invitations.
Involvement of primary care in refinement of screening invitation lists
Our results suggest that primary care plays a worthwhile role in supporting the Cervical Screening Programme through refining screening invitation lists; in contrast, our results do not support this process for breast and bowel cancer screening, particularly as it is apparent that there is considerable variation in the approaches taken to this process.
Footnotes
Acknowledgements
We thank all cervical screening departments, the six regional breast screening centres, the English Bowel Cancer Screening Pilot, and the general practices who participated in the study. Thanks to Rosemary Porteous, University of Edinburgh, for transcribing the interviews. Thanks to anonymous reviewers for feedback on previous manuscripts.
Funding Body
NHS Cancer Screening Programmes
Ethics
Ethical approval was not required for the first two components of the study: they were considered service evaluations of ongoing activity conducted within the framework of the NHS Screening Programmes (data collected were not patient-identifiable). Conflicting advice was received regarding the need to obtain MREC approval regarding the qualitative interviews with NHS staff: an application was submitted, but the MREC Committee ruled that their authorization was not required.