Abstract
In response to the COVID-19 pandemic, a temporary change in policy was implemented in 2020. Breast screening services in England were advised to change from timed appointments to an open invitation for invitees to contact the service and arrange an appointment. This change to invitation methodology had potential benefits and risks including impacting inequalities in uptake. Qualitative data were collected by online questionnaire from 23 service providers and routinely collected quantitative uptake data were analysed to investigate the impact of open invitations on the National Programme in the South of England. Office for National Statistics and general practitioner (GP) practice profile data enabled the modelling of sociodemographic characteristics of breast screening invitees at each GP practice. Most services changed to open invitations (17/23), 82% of which altered administrative capacity and/or procedures to accommodate this change. Logistic benefits were reported including a more consistent flow of participants, fewer long gaps and fewer wasted slots. The change to open invitations was associated with a 7.2% reduction in the percentage of participants screened, accounting for participant sociodemographics and historical screening provider uptake. The inequality in screening uptake experienced by participants of minority ethnic background was exacerbated by the change to open invitations. Open invitations, whilst affording logistic benefits in an unprecedented pandemic era, were associated with reduced overall uptake and exacerbation of existing health inequality experienced by women of minority ethnic background. The broader impact on services highlighted the need for sustainability of measures taken to accommodate such operational changes.
Introduction
The National Breast Screening Programme in England invites anyone registered with a general practitioner (GP) as female, and aged 50 < 71 years old, to attend every 3 years. Screening services are locally commissioned in line with a nationally agreed specification. This evaluation formed part of the quality assurance of the 23 South of England breast screening services to minimise unwarranted variation, and provide intelligence to inform future policy, guidance and standards.
The South of England includes two regions (South East and West), covering 16 counties and 8217 lower super output areas (LSOA). The population has on average a less diverse ethnic background, and is less deprived than the country as a whole,1,2 and includes an eligible breast screening population of 1.7 million. 3
Before the COVID-19 pandemic all initial breast screening invitations were to a timed appointment, followed by a reminder of either an alternative timed appointment or an ‘open’ invitation for women to schedule their own appointment. Due to the temporary suspension of the programme in March 2020, increased infection control procedures and therefore longer time between appointments, there was restricted clinic capacity and a backlog of participants when the programme restarted. Consequently, a temporary policy change advised breast screening services to change initial invitations to an open format. This aimed to optimise available capacity by screening those who wished to attend, reducing ‘did not attend’ (DNA) rates.
Those known to have greater challenges in making contact to arrange an appointment (e.g. people with learning disabilities) were to continue to receive timed invitations. Timed appointments have been shown to significantly increase uptake for those who did not attend their first appointment.4,5 It was noted that there was the potential to adversely impact those whose disability was not known by the service, whose first language is not English, those with different cultural understanding of cancer/screening and others who may find setting up appointments difficult. Additionally, there was potential to adversely impact those who are screen-hesitant, as the use of timed invitations is known to lower barriers to attendance, removing a participant-led intermediary step. 6
The aim of this evaluation was to explore whether changing screening invitation method influenced uptake, and how this impacted health inequalities in the South of England.
Methods
Qualitative data
Regional commissioning teams provided local knowledge to coproduce a Microsoft Forms questionnaire completed by all 23 breast screening services. Information regarding which services changed invitation method, and when, informed the selection of time periods for quantitative data comparison.
Quantitative data
The percentage of eligible population screened was available by GP practice and mapped to breast screening services. Two three-month periods before and after the change to open invitations (July–September 2019 and July–September 2021) were selected.
Estimates of ethnicity and Index of Multiple Deprivation (IMD) decile of the female population aged 50 to <71 years at each practice were calculated by applying a scaling factor according to the proportion of the GP practice population that resided in each LSOA.1,7,8
Statistical analysis
Data were extracted from Microsoft Forms, NHS England internal databases and Office for National Statistics published data, cleaned and managed in Microsoft Excel, and analysed using STATA13. Difference in percentage screened in 2019 compared to 2021 by change to invitation method was analysed using a two-sided non-parametric test of matched pairs, making no assumption about variance or direction of change.
Data were not available at a participant level so linear regression was used to model the impact of change to open invitations on the percentage of participants screened in 2021. Additional variables were added in a stepwise manner. To estimate the effect on health inequalities, separate regression models were constructed according to invitation method change.
Differences in incidence rates of COVID-19 and local pandemic restrictions over the study period were not explored.
Results
All services responded to the survey between 28/09/2022 and 19/10/2022 (median completion time 11:11 [IQR 03:55–24:01]). Most services changed to open invitations (17/23; 73.9%) doing so between July and September 2020. Six services continued to use timed invitations.
Reported benefits of changing to open invitations were themed around improved logistics; knowledge of who and how many participants to expect, fewer DNA/wasted slots, more consistent participant flow and fewer long wait times. It was reported that open invitations were preferred by mammographers and helped service recovery.
Administrative capacity or procedures were altered by 82% of services that implemented open invitations. Fourteen services used the booking call for more than scheduling an appointment; four included elements of the appointment previously delivered on site (e.g. pre-screen questionnaire), and 10 gave the client an opportunity to ask questions on the call.
Over half (53%) of services had the overall impression that attendance had reduced following the change to open invitations. Five services (29%) believed attendance had improved, however corresponding quantitative data showed a reduction in uptake. Three services (18%) reported the impression that open invites had no effect on attendance, however in two of these three services uptake was found to have reduced. This was further explored, asking services for their overall impression of client attendance from specific population groups. Notably a substantial proportion of services were unable to give an overall impression of how the change had affected attendance by people of minority ethnic (41%) or deprived (35%) background, and those whose first language is not English (29%), responding ‘do not know’ when asked to estimate the effect of open invites on these population groups.
As shown in Figure 1, practices screened by services that changed to open invitations showed a significant reduction in percentage screened between 2019 and 2021 (p < 0.0001) that was not seen in practices continuing with timed invitations (p = 0.4594).
Change in percentage of participants screened between July–September 2019 and July–September 2021 by GP practice. GP practices mapped to a breast screening service that did not change the invitation method in 2020: Unchanged 2019 and Unchanged 2021; those that changed to open invitations in 2020: Changed 2019 and Changed 2021. Box and whisker plots detail the median, interquartile range, minimum and maximum values. A non-parametric test of matched pairs was used to assess statistical differences within groups.
The change to open invitations was associated with a 7.29% reduction in percentage screened (95% confidence interval (CI) −9.11 to −5.46) accounting for ethnicity, IMD decile and percentage screened in 2019. Those living in more deprived areas (1.42 [95%CI 1.00–1.85]; p < 0.001) and participants of minority ethnic background (0.46 [95%CI 0.36–0.57]; p < 0.001) had lower uptake regardless of invitation method.
Before changes to invitation method (July–September 2019) there was evidence of health inequalities; for every 1% increase in the proportion of patients of white ethnicity there was a 0.48% increase in the percent screened (95% CI 0.37–0.59; p < 0.001) accounting for deprivation decile. For each decile increase in IMD there was an associated 1.80% increase in percent screened (1.34–2.25; p < 0.001) accounting for ethnicity (adjusted r2 = 0.1105; n = 1329).
Whether changing to open invitations exacerbated these pre-existing health inequalities was then explored. Health inequality by ethnicity and deprivation remained whether or not invitation method was changed (Table 1). Of note is that the change to open invitations increased the slope of inequality experienced by women of minority ethnic groups. When invitation method did not change a 1% increase in the proportion of patients of white ethnicity was associated with a 0.22% increase in percent screened. This association rose to 0.82% amongst practices where the breast screening service changed to open invitations (Table 1).
Multivariable regression model of data after changes to invitation method (screened in July–September 2021) split by whether or not the breast screening service to which each GP was mapped made changes to the invitation method.
N: the number of GP practices; CI: confidence interval.
Services were asked for their preferred method of invitation going forward. Twelve of 23 services wished to remain with/return to timed appointments. Only three expressed a preference for open appointments alone, the rest a hybrid approach.
Discussion
The change to open invitations was recommended to support the recovery of breast screening services, by maximising slot availability for those wishing to attend. Some services remained with timed invitations resulting in a natural experiment, utilised here to evaluate the impact of open invitations on service delivery and percentage screened.
Whilst our findings regarding participant ethnicity and deprivation are similar to those described by others, 9 we uniquely provide an estimate of the sociodemographics of all participants invited to screening, including those who chose not to attend. Our findings provide intelligence to policy makers regarding the potential impact of operational processes on health inequalities.
Approximately half of breast screening services in the South of England had an impression of the impact of open invitations on attendance that was not supported by the quantitative data on uptake. This could have been influenced by the effect of the recovery of breast screening services following the COVID-19 pandemic. Furthermore, the difficulty some breast screening services found in estimating the effect of the change to open invitations on health inequalities could indicate the need for increasing awareness of health inequalities experienced within breast screening, and potential ways to reduce such inequality. We report that health inequality experienced by women of minority ethnic groups was exacerbated by the change to open invitations. This indicates a need for increased understanding of how invitation methodology affects different population groups, to ensure that operational changes reduce, rather than increase, inequalities in uptake.
The intended logistic benefits were reported by services; however the sustainability of administrative changes made to accommodate open invitations would need to be assured if this invitation method was to continue. Additionally, some services reported abuse from callers, reduced staff morale and resignations.
The booking phone call required as part of the open invitation method represented an opportunity to answer client queries and/or include parts of the breast screening process over the phone prior to the screening appointment. Such additional interventions are likely to have influenced client attendance and warrant further exploration.
The biggest challenge when evaluating the impact of open invitations on inequalities of uptake was that sociodemographic data are not collected at participant level by the Breast Screening Programme. A weighted estimate of ethnicity and IMD decile introduced inaccuracy into the analysis, resulting in minorities within an LSOA failing to be represented. Whilst this method has limitations it was preferable to not exploring the issue. Of note are the sociodemographic characteristics unable to be estimated based on residential location, such as social communication difficulties and English as a second language. Therefore, the effect on uptake amongst women in these population groups, and others likely to be impacted by invitation method change, were not explored and represent an important area for further research.
Other potential confounding variables between services that were not accounted for within this study include local incidence of COVID-19 infection and resulting pandemic restrictions over time, and ‘self-referrals’ – women who have passed the original invitation window who subsequently attend screening at a later date.
Conclusion
Open invitations, whilst affording important logistic benefits in an unprecedented pandemic era, were associated with reduced overall uptake and exacerbation of existing health inequality experienced by women of minority ethnic background. The reported additional impact on service delivery highlighted the need to consider sustainability of measures taken to accommodate such operational changes. Further research into the impact of operational changes on service delivery and participant attendance, in particular health inequalities, is warranted. Findings of such work will have relevance to the future design and implementation of several national screening programmes.
Footnotes
Acknowledgements
We thank the following colleagues within the South East and South West NHSE Regional public health commissioning teams for sharing local operational context and for support with questionnaire construction and distribution: Amber Codd, Heather Gilpin, Dave Harris, Gemma Harris, Lainey Hillier, Elizabeth Luckett, Lianne Straus, Sarah Tomkinson and Hayley Ware.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
