Provision of the progestogen-only pill by community pharmacies as bridging contraception for women receiving emergency contraception: the Bridge-it RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 15/113/01. The contractual start date was in April 2017. The draft report began editorial review in July 2020 and was accepted for publication in January 2021. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

© Queen’s Printer and Controller of HMSO 2021. This work was produced by Cameron et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2021 Queen’s Printer and Controller of HMSO

Background

Unintended pregnancy is a major public health problem. Despite having one of the highest rates of modern contraceptive use worldwide, the UK has among the highest abortion rates in Europe. 1 In 2018, almost 200,000 pregnancies ended in induced abortion. 2,3 Unintended pregnancy also ends in childbirth. Around 10% of UK births are unintended and 25% are mistimed. 4 Unintended pregnancy is costly to the NHS5 and distressing for women. Unintended pregnancies are over-represented in young women from deprived backgrounds. Unintended childbirth can have both socioeconomic consequences for women and their families, and mental health consequences. 6

Emergency contraception (EC) prevents pregnancy in individual women following unprotected sex or contraceptive accidents (e.g. burst condom). Approval of EC from pharmacies and making it free of charge to all women in Scotland and Wales, and free to many women in England, has increased use and EC is now largely obtained from pharmacies. 7 However, although trials have shown that facilitating access to EC increases use of EC, these trials have failed to show an effect on unintended pregnancy rates. 8

Emergency contraception [containing levonorgestrel (LNG)] is only effective if taken within 72 hours of unprotected sex. It does not prevent conceptions from subsequent sex. The risk of pregnancy is increased up to threefold among women who have further unprotected sex in the same menstrual cycle after using EC. 9 An effective method of contraception should therefore be started as soon as possible – known as ‘quick starting’. 10 However, the only contraceptives available from pharmacies without prescription are condoms, which have high failure rates. 11 To start an effective contraceptive (i.e. hormonal or intrauterine contraception) women must visit a general practitioner (GP) or attend a sexual and reproductive health (SRH) clinic. It may take time to get an appointment and many women may lose the motivation to seek contraception, leading to an unintended pregnancy. In addition, in some UK studies, fewer than half of pharmacists gave advice about ongoing contraception after EC. 12,13

There is a lack of evidence on interventions to increase uptake of effective contraception after EC. We conducted a literature search of PubMed English-language articles from 2000 to March 2020 using search terms (((bridging) OR bridge)) AND (emergency contracept*) and identified only two relevant articles. One was a cluster randomised trial14 from Jamaica, in which women seeking EC were offered a voucher that gave them a discount on the cost of contraceptive pills. This study reported no effect of the intervention on subsequent contraceptive use. The other article15 was a pilot study of 12 pharmacies in Edinburgh, with 168 women presenting for EC randomised to receive 1 month of a progestogen-only pill (POP), rapid access to a local SRH clinic or standard care. Participants were contacted by telephone 6–8 weeks later to determine which method of contraception they were using. In the POP arm, 35 out of 39 (90%) women used the pills provided and 9 out of 28 (32%) women in the rapid-access arm attended the SRH clinic. Compared with standard care, the proportion of women using effective contraception after EC was statistically significantly greater in both the POP arm (56% vs. 16%; p = 0.001) and the rapid-access arm (52% vs. 16%; p = 0.027). The study concluded that a 1-month supply of POP after EC or rapid access to a SRH service might increase short-term uptake of effective contraception following EC, but that a high-quality, adequately powered randomised trial was needed.

Rationale for study

Unintended pregnancy remains a public health problem in the UK. Guidance from the Faculty of Sexual and Reproductive Healthcare of the Royal College of the Obstetricians and Gynaecologists (London, UK) stresses the need to ‘quick-start’ ongoing contraception after EC. 10 In 2014, National Institute for Health and Care Excellence guidance on contraceptive services for young people endorsed this recommendation. 16 In some parts of the UK, pharmacies are offering women supplies of contraceptive pills after EC use,17 despite the lack of evidence that this is an effective intervention or is cost-effective. We must show that the intervention is effective before we can recommend adoption of this approach. A recent cost-effectiveness analysis of EC estimated that, in 2011, unintended pregnancies cost the NHS over £1B. 5 It is possible that even these costs are an underestimate of the ‘real costs’, as they did not include the cost of managing medical complications of pregnancy or account for the additional costs associated with teenage pregnancy. If a community pharmacy-based intervention could be designed that was effective in reducing unintended pregnancies and was shown to be cost-effective, then this would result in savings for the health systems, which could be invested elsewhere. Women who present for EC should be given the best chance to prevent an unintended pregnancy.

Long-acting reversible contraception (LARC), which includes the implant, intrauterine contraception and injectable contraception, has been shown to be the most effective reversible method of contraception for preventing pregnancy. 18 In recognition of the effectiveness of LARC, national initiatives in the UK to increase uptake have been encouraged. 16,18 Rapid referral of women using EC to SRH services capable of initiating LARC immediately (as in this study) could do much to improve LARC uptake among women at risk of unintended pregnancy.

We therefore conducted the ‘Bridge-It’ study: a robust trial to determine whether or not a pharmacy-based intervention designed to facilitate the uptake of effective contraception after EC increases use of effective contraceptive methods, compared with standard care alone. For the Bridge-it study, we used a composite intervention of a small ‘bridging’ supply of the POP and the offer of rapid access to a participating SRH clinic. This combined both temporary contraception (giving women time to arrange an appointment with their usual contraceptive provider) with facilitated access to a specialist contraceptive service (i.e. the SRH clinic) where all methods of contraception, including the most effective LARC methods, are provided.

Study design

The Bridge-it study was a pragmatic cluster randomised cohort crossover trial of community pharmacies in three regions of the UK [London (south and central), Lothian (Edinburgh and region) and Tayside (Dundee and region)]. 19 With this design, the order in which each community pharmacy provided intervention or control was randomised. There was an intervening period of at least 2 weeks (during which recruitment halted) before the pharmacy switched to the other recruitment phase. 19 The cluster design was chosen as our pilot study15 findings indicated that an individual randomised trial would not recruit sufficient numbers of women. The crossover was chosen for efficiency, with each cluster acting as its own control and fewer pharmacies required. The washout period between recruitment periods minimised any contamination effect of pharmacists between recruitment periods.

Pharmacies were chosen to be a mixture of large commercial chain stores and small independent stores. Each pharmacy had to have a sufficient volume of EC dispensing (at least 30 ECs per month) to participate.

Women receiving LNG EC from a study pharmacy were considered for study participation. The LNG EC was given in the appropriate dose (1.5 mg or 3 mg) for the woman’s weight. 10

The intervention was a composite intervention. Women received 3 months worth of the POP (75 µg of desogestrel/day) at no cost (as is the norm in the NHS) and the offer to attend a local participating SRH service to discuss and provide ongoing effective contraception, including the most effective LARC methods (i.e. intrauterine contraceptive methods and contraceptive implants). 18 The three packets of POP provide women with 3 months within which they could attend a contraceptive provider for ongoing contraception. The POP has very few absolute contraindications,20 making it safer for pharmacy provision than the combined oral contraceptive pill. The desogestrel POP was chosen as it is the market leader, the most effective POP (as it has high rates of ovulation inhibition) and is inexpensive (£9 for 3 months of the generic version). 21 Community pharmacists provided the POP using locally approved patient group directions (PGDs) (i.e. strict criteria to permit provision of specified medicines by non-prescribers). 21,22 Participating pharmacists were trained on the study protocol and use of the PGD for the POP, including information about medical contraindications to POP, potential drug interactions and ‘missed pill’ guidance. Pharmacists advised women to start the POP the day after the EC. Women in the intervention group also received a rapid-access card that on presentation to the local SRH clinic would help them to be seen as a walk-in patient to discuss contraception and obtain their preferred ongoing method. The card provided information about the location of the SRH clinic and the opening hours. The clinics (three in London, one in Lothian and two in Tayside) were located within 5 miles of the community pharmacies.

In the control group, women received standard care with the EC (i.e. usual advice about ongoing contraception). Mystery shopper visits with simulated patients12,13 were conducted in the pharmacies just before recruitment started in the control group to document the content of the standard care EC consultations (see Chapter 4).

Participants

Pharmacists assessed women’s eligibility for the study, invited them to participate, obtained written informed consent for the study (including access to SRH clinic records and data linkage with the national abortion registries) and provided a patient information sheet. Participants completed a baseline questionnaire that included demographic details, reproductive history and previous methods of contraception used, including ECs (see Appendix 3). To maximise retention in the study, participants received an online shopping voucher of £10 following recruitment. 23

Procedures

Participants in both groups of the study had a single follow-up at 4 months: either a telephone interview with a research nurse or a self-administered web-based questionnaire, according to participant preference (see Appendices 4 and 5). In addition to contraceptive use, women were asked about their interaction with the pharmacist and use of the rapid-access card (intervention group only). Participants reporting pregnancy also completed the London Measure of Unplanned Pregnancy, which is a validated questionnaire to measure the intendedness of the index pregnancy. 24,25 Serious adverse events since recruitment were also recorded.

The participating SRH clinics also searched their clinic record systems for data on if and for what reason participants in both arms of the study had used their service during the 4 months following recruitment.

Outcomes

The primary outcome was self-reported use of effective contraception (hormonal or intrauterine) at 4 months. Secondary outcomes were incidence of abortion in the 12 months following recruitment and an economic evaluation of the intervention. Both secondary outcomes will be reported later [as per agreement with the National Institute for Health Research (NIHR)].

A multimethod process evaluation was also conducted to assess implementation, mechanisms of change and context,19 and included qualitative interviews with participants, pharmacists and staff at SRH clinics. The process evaluation is reported separately in this report (see Chapter 5).

Sample size

An original power calculation for the study was based on two co-primary outcomes: (1) abortion rates at 12 months and (2) use of effective contraception at 4 months. For the abortion outcome, we required > 2000 women to be randomised in at least 26 pharmacies to have 90% power at 2.5% level of significance to detect a relative reduction of around 50. During the study, it became evident that recruiting this number of women was not feasible within the available time frame and resources. The independent oversight committees [the Trial Steering Committee (TSC) and Data Monitoring Committee (DMC)] and the funder (NIHR Health Technology Assessment programme) agreed to repower the study on a single primary outcome (i.e. use of effective contraception at 4 months). To have 90% power at a 5% level of significance and to demonstrate an increase from around 30% to 45% (absolute change 15%, relative increase 50%), the study required between 626 and 737 women, depending on the intraclass correlations (within period and between periods), which were not observable within the study at the time this change was made. 26 The sample size calculation assumed that we would not have the 4-month primary outcome data for 25% of participants and that there would be 25 participating pharmacies. For the intracluster correlations, the within-cluster, within-period correlation was 0.032 and the between-period, within-cluster correlation was also 0.032. The observed correlation between the two outcomes in the same cluster period compared with the two outcomes from different periods in the same cluster was effectively zero (< 0.00001).

Randomisation

The order of delivery of intervention or control for each pharmacy was generated using a computer software algorithm. This randomly allocated permuted blocks of size two, four and six, and blocking was used to ensure a balanced order. The randomisation file was prepared by the study statistician at the Centre for Healthcare Randomised Trials, University of Aberdeen (Aberdeen, UK), using SAS^® 6.4 for Windows (SAS Institute Inc., Cary, NC, USA). Pharmacists were informed of the randomisation allocation by the study trial manager. The study was not blinded.

Statistical analysis

Baseline characteristics of participants and 4-month follow-up data were summarised using mean [standard deviation (SD)] or median (interquartile range) for continuous variables. Discrete variables were summarised with numbers and percentages. The published study protocol19 specified a hierarchical mixed-effects logistic regression on individual data for the analysis of reported use of effective contraception at 4 months, and this represented our analysis intentions at the time of submission of the study protocol. However, we revised this primary analysis in the final statistical analysis plan (accepted 5 November 2019, before any unblinded data had been seen) to use a linear model on the unweighted proportion (expressed as a percentage) at site level, following methodological guidance from Morgan et al. ,27 and we used the hierarchical mixed-effects logistic regression as a sensitivity-type analysis. Although the linear model on percentages at site level makes less use of the available information, it can be more robust, with fewer assumptions, and expresses the treatment effect as a percentage difference in proportion rather than as an odds ratio.

Ideally, the design would generate equal contributions in each time period and each site would contribute equal numbers. However, we knew from the aggregated accruing data on the database, returned to date, that sites were quite different in size and also returning different numbers in each period. Therefore, and guided by the Morgan et al. 27 paper, which indicated that the simple approach of modelling the percentages by site had the best properties in terms of control of type 1 error, we took the unusual decision to prefer the simpler approach to the more complex hierarchical approach. This resulted in forgoing the potential increase in power traded against the possibility of model misspecification, as the model assumptions (regarding the several decompositions of within and between cluster correlation, the influence of missing data, and the influence of very different sizes by period and site) might have been violated. We felt that the hierarchical approach could still usefully be reported as a sensitivity-type analysis and, as it happened, it does seem to support the simpler analysis of proportions, giving similar estimates, but with the expected increase in precision.

For the primary outcome, the percentage of women reporting use of effective contraception in each cluster period was analysed using linear regression, adjusting for period, treatment arm and centre, and with centre as a fixed effect. 27 Prespecified baseline covariates (i.e. mean age, percentage of participants currently in a sexual relationship and percentage of participants who had previously used effective contraceptive methods) were included as an additional analysis. A two-sided p-value < 0.05 was taken as statistically significant.

For the primary outcome, prespecified subgroup analysis for LARC use18 was performed using a stricter level of statistical significance (two-sided 1% significance level) and 99% confidence intervals (CIs).

The sensitivities of treatment effect estimate to missing outcome data were undertaken using multiple imputation. The missing primary outcome (i.e. uptake of effective contraception) was imputed using all baseline characteristics as predictors, except for ‘previous ectopic pregnancy’ (because of small numbers). Stata^® version 16 (StataCorp LP, College Station, TX, USA) was used to impute 40 data sets.

Summary of changes to the protocol

Owing to resource and time constraints, the study was repowered on a single primary outcome (i.e. effective contraception use at 7 months). For the same reasons, the planned follow-up of participants at 12 months was removed. Abortion rates at 12 months will still be collected and these and the cost-effectiveness model will be reported separately (because of a later study end date than anticipated). A summary of changes to the protocol can be found in Box 1. The final and previous versions of the protocol can be found on the NIHR Journals Library [URL: www.journalslibrary.nihr.ac.uk/programmes/hta/1511301/#/ (accessed 15 February 2021)].

Ethics

Ethics approval was obtained from South East Scotland Research Ethics Committee in June 2017. Approvals were also obtained from NHS Research Scotland (Clydebank, UK) and Health Research Authority (London, UK).

Recruitment and participant flow

Fifty-six pharmacies were approached to participate and 32 agreed to take part in the trial (Figure 1). Twenty-nine out of the 32 pharmacies recruited participants (14 in London, 12 in Lothian and three in Tayside). Participants were recruited between 19 December 2017 and 26 June 2019. A total of 1252 participants were screened, of whom 762 were eligible and 636 consented and were randomised. Box 2 shows the inclusion and exclusion criteria.

FIGURE 1.

Consolidated Standards of Reporting Trials (CONSORT) flow chart.

Table 1 shows reasons for ineligibility and Figure 2 shows participant flow. The number of participants recruited per pharmacy ranged from 2–35 in the pharmacies’ first recruitment period and from 0–38 in the second period.

FIGURE 2.

Participant flow chart. n, number of women who returned a 4-month questionnaire; N, number of women recruited; P, number of pharmacies.

Table 2 summarises the baseline characteristics of participants by randomised group and recruitment period. The mean age of participants was 22.7 (SD 5.7) years and 22.6 (SD 5.1) years in the intervention and control groups, respectively. There were no statistically significant differences between groups.

Four-month follow-up

Follow-up data at 4 months were available for 198 out of 315 (62.9%) women and 208 out of 318 (65.4%) women in the intervention and control groups, respectively. Follow-up rates did not differ statistically between groups nor between those recruited in the first and second recruitment periods of pharmacies. There were no substantial differences in any baseline characteristics between responders and non-responders (Table 3). No significant adverse events were reported.

Primary outcome: effective contraception use at 4 months

The proportion of women using effective contraception (hormonal and intrauterine contraception) was 20.1% greater (95% CI 5.2% to 35.0%) in the intervention group (58.4%, SD 21.6) than in the control group (40.5%, SD 23.8) (adjusted for recruitment period, treatment arm and centre; p = 0.011) (Table 4).

Use of effective contraception remained statistically significantly higher in the intervention group when adjusted for recruitment period, treatment group, study centre, age, whether or not currently in a sexual relationship and history of past use of effective contraception, and was robust to the missing data (Tables 4–6) using a multiple imputation approach under an assumption of missing at random.

Contraceptive use and long-acting reversible contraception use at 4 months

The methods of contraception used at 4 months after EC are shown in Table 7. The most common methods used were oral contraceptive pills (POPs in the intervention group and the combined hormonal contraceptive pills in the control group). There were no sterilisations or women relying on vasectomy. Use of LARC methods (e.g. implant, intrauterine and injectable) did not differ between the intervention group (13/198, 6.6%) and control group (23/208, 11.1%) (95% CI –10.04% to 1.05%; p = 0.112).

The most common reason given by respondents for not using effective contraception at 4 months was that they were not currently sexually active (Table 8). Significantly fewer women in the intervention group than in the control group had used EC again since recruitment [20/198 (10.1%) vs. 37/208 (17.8%)] (95% CI –15.38% to –1.48%; p = 0.018) (see Table 8).

A significantly larger proportion of women in the intervention group reported that the pharmacist had advised them about starting ongoing contraception (98% vs. 75.5%; p < 0.001) (Table 9).

A total of 158 out of 198 (79.8%) respondents in the intervention group reported that they used some of the study POP that the pharmacist provided. Table 10 provides data on when women started the POP, the number of packets of POPs used and reasons for non-use or discontinuation of the POP.

Rapid-access card and sexual and reproductive health clinic utilisation

At the 4-month follow-up interview, 137 out of 198 (69.2%) respondents in the intervention group could recall receiving the rapid-access card, but only 31 respondents (15.6%) reported that they attended the SRH clinic. Only two respondents used the rapid-access card within 1 month of recruitment. Four out of 31 respondents (12.9%) received a LARC method at the SRH clinic (Table 11).

Data obtained from the SRH clinic databases showed that similar proportions of participants in both the intervention and control groups actually attended the SRH clinics within 4 months of recruitment [52/305 (17%) intervention group vs. 43/309 (13.9%) control group, 95% CI –2.60% to 8.87%; p = 0.284]. A total of 41 out of 95 (43.1%) attendees received contraception and the methods supplied are shown in Table 12. There was no statistically significant difference in LARC provision to women in each group.

A total of 75 women both attended the SRH clinic and provided data at the 4-month interview. For 34 of these women, the SRH clinic provided a contraceptive method. Thirty out of 34 (88%) women reported using the same method of contraception at 4 months as the method provided by the SRH clinic.

Pregnancies in the 4 months since emergency contraception

Nineteen respondents (4.7%) (intervention, n = 9; control, n = 10) reported that they had been pregnant (n = 17) or were currently pregnant (n = 2). The outcome of the pregnancies that had ended were abortion in 10 women (58.8%) and miscarriage in six women (35.3%); this information was unknown for one woman (5.9%). Based on the London Measure of Unintended Pregnancy24,25 scores from the questionnaire, 15 out of 19 pregnancies (78.9%) had been unintended (seven in the intervention group and eight in the control group).

Mystery shopper exercise

Pharmacists were advised that one or two mystery shopper visits would take place before the control group of the study started, but they were not told when these would occur. Visits took place between April 2018 and January 2019. Research nurses recruited female volunteers (aged ≥ 16 years) to be mystery shoppers and instructed them on the aims of the exercise, the scenario to be followed and the data to be collected. The members of the patient and public involvement (PPI) group for the study assisted in identifying potential mystery shoppers. The shoppers received £20 for each completed visit. They followed a scenario (a single episode of unprotected sex within 72 hours) relating to a request for EC. The scenario had been approved by the PPI group. Immediately after leaving the pharmacy, shoppers completed a data collection pro forma, recording the time of arriving and leaving the pharmacy, duration of the consultation and where the consultation about EC took place. They also noted any information provided by the pharmacist, including advice on ongoing contraception. Data from the proforma were entered into an Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA) database and descriptive statistics were conducted for each question.

Results

A total of 55 mystery shopper visits were conducted at the 30 study pharmacies participating in this part of the trial. The mean total time reported spent in the pharmacy was 12 (range 1–47) minutes. The median reported duration of the consultation with the pharmacist was 6 (range 1–18) minutes. Consultations took place in a private room with 34 women (62%) and the remaining consultations took place at the counter.

Eleven mystery shoppers (20%) were unable to get EC from the pharmacy. Seven mystery shoppers reported that they were told that no trained pharmacist was available to provide EC. Four mystery shoppers were told that EC was not in stock. Two of the latter women were aged 16 years and visited the same pharmacy on the same day and both were advised to return 90 minutes later. A further two mystery shoppers were advised that a copper intrauterine device would be more effective and were directed to a SRH clinic for insertion, but neither was offered oral EC as an interim measure. Eight mystery shoppers were told that they had to swallow the EC tablets in the pharmacy and seven mystery shoppers (in London) were told that the pharmacy could not offer free EC. Three mystery shoppers were asked to provide proof of identity and age to get free EC. Table 13 shows the results for information provided by the pharmacist around EC or ongoing contraception.

A total of 27 out of 55 (49%) mystery shoppers received advice (oral or written) about ongoing contraception. Forty-eight of the mystery shoppers made comments about their experience. Analysis of the free-text comments showed four common themes: (1) the manner of the pharmacist, (2) being requested to take the EC on the premises, (3) pharmacy not able to provide EC and (4) concerns about privacy. Examples are shown in Box 3.

Discussion

The mystery shopper exercise was undertaken to describe ‘standard care’ for the control group regarding the request for EC at a community pharmacy. The exercise showed that, although pharmacists were generally helpful, getting EC (for which all of the women were eligible) was not always easy (1/5 did not get EC). In addition, waiting times were sometimes long, consultations short and privacy was not always guaranteed. In a few cases, women were asked for proof of identity/age, which was not a requirement of the local PGD in place for supplying EC. In addition, fewer than half of pharmacists in the mystery shopper exercise gave any advice about ongoing contraception. These findings are similar to a previous mystery shopper study from Edinburgh, published 10 years ago. 12 Although the current mystery shopper exercise was conducted in only 30 UK pharmacies and may therefore not be fully representative of care across the UK, all participating pharmacists had recently undergone training for the Bridge-it study and so one might have expected that the performance of these pharmacies regarding advice around ongoing contraception would have been higher.

These findings suggest that opportunities to provide EC to women and to prevent unintended pregnancy are currently being missed in community pharmacies across the UK. Furthermore, opportunities for pharmacists to discuss ongoing contraception with women provided with EC are also being missed.

This adds to the debate of whether or not EC should be available on the general sales list (i.e. available without the need for a consultation and purchasable from a range of outlets). According to the Medicines and Healthcare products Regulatory Agency (London, UK), the general sales list is appropriate for medicines that can, with reasonable safety, be sold or supplied in ways other than by or under the supervision of a pharmacist. 29 The term ‘with reasonable safety’ has been defined as ‘where the hazard to health, the risk of misuse, or the need to take special precautions in handling is small and where wider sale would be a convenience to the purchaser’ (contains public sector information licensed under the Open Government Licence v3.0. URL: https://nationalarchives.gov.uk/doc/open-government-licence/version/3/). 29

It is now more than 20 years since EC was first approved as a pharmacy medicine in the UK. 7 There is abundant evidence to demonstrate that it does not present a hazard to health, nor is it widely misused, and does not need special precautions. If ECs were a general sales list medicine, then there would be more opportunities to obtain it, which would benefit women and, possibly, public health.

Methods: data sources and analysis

We employed a range of quantitative and qualitative methods, including qualitative interviews with providers (pharmacists and SRH staff) and participants, monitoring of pharmacy selection and training observations, analysis of relevant data from the 4-month follow-up questionnaire and monitoring of contemporaneous events (e.g. relevant media coverage). Further details on each data collection method and analysis approach are provided in Appendix 1, Table 16. Collection of such data allowed us to assess the role of our critical assumptions, mediators of change and intermediate outcomes that could impact on the effectiveness of the intervention.

All process data were analysed independently from the outcome data and documented before the outcomes were known. The process evaluation team (SP and LMcD) discussed the progress of data collection and analysis on a regular basis, allowing any issues that were encountered to be resolved (Table 14).

Qualitative interviews with pharmacists, sexual and reproductive health providers and participants

Qualitative interviews were conducted with those responsible for delivering the study (i.e. pharmacists and SRH providers) and those receiving the intervention (i.e. Bridge-it Study participants). The breakdown of recruitment by site is detailed in Figure 3. Specific details about sampling, recruitment, topics covered and analysis for each group is detailed below.

FIGURE 3.

Process evaluation flow chart.

Semistructured qualitative telephone interviews were conducted with pharmacists and SRH providers to explore the acceptability of the intervention and training received (pharmacists), the perceptions of barriers to and facilitators of implementation, and existing contextual challenges within their services (see Appendix 2). In total, 22 pharmacists were interviewed (12 from Lothian, three from Tayside and seven from London-based pharmacies). The aim had been to interview one pharmacist from each participating pharmacy in the study. The main pharmacies not represented (n = 7) are based in London (south), as it had not been possible to conduct interviews before the study was discontinued. Interviews were conducted between July 2018 and July 2019, with most interviews taking place once recruitment had ended in their pharmacy.

As well as interviewing pharmacists involved in the study, we interviewed five SRH providers in three out of four participating NHS sites (two in Lothian, two in Tayside and one in London) between May and October 2019. Our original aim had been to interview three or four staff members from each service; however, recruitment was challenging, particularly because of low Bridge-it study participant attendance at SRH clinics. Interviews were conducted 4–6 months after study recruitment had ended to allow time for SRH staff to have experience of participants attending their service.

Qualitative interviews were also conducted with a purposive sample of 36 intervention participants (24 participants in Lothian, nine participants in London and three participants in Tayside) to explore intervention acceptability, experiences of bridging from EC to effective contraception, facilitators of and barriers to continued uptake and the wider context of their contraceptive experiences. Unintended consequences were documented through the process evaluation, including asking participants whether or not they perceived any unintended negative outcomes resulting from participation in the trial. Research nurses asked participants for consent to be contacted by the process evaluation research assistant for a qualitative interview at the end of the 4-month follow-up questionnaire, and interviews were conducted between November 2018 and October 2019. Purposive sampling was employed to recruit a representative and diverse sample, with participants sampled by area, age, use of the study POP and attendance at a SRH clinic. The characteristics of the participants who were interviewed are shown in Table 15.

TABLE 15 - Intervention interview participant characteristics

Participants could attend both the GP and SRH clinic services.

Characteristic	Participants (n)
Age group (years)
18–19	9
20–24	14
25–29	5
30–37	8
Location
Lothian	24
Tayside	3
London	9
Contraceptive use at time of interview
POP	10
Combined pill	4
Intrauterine contraception	1
Implant	1
None	20
Contraceptive service attendeda
GP	14
SRH clinic	5
No service attended	13
Unknown	6
Previous effective contraceptive use
Yes	22
No	13
Missing	1

All provider and participant interview data were audio-recorded, transcribed verbatim and anonymised. Data management was assisted by ethnographic software (NVivo 10, QSR International, Warrington, UK). Transcription and analysis (proceeding case by case) began with the first interview and was ongoing during data collection, allowing emergent themes to be identified and explored in future interviews. Data analysis was undertaken using ‘framework analysis’, which is a method of proven validity and reliability in which data are coded, indexed and charted systematically, then organised using a matrix or framework. 30 Such a method helps to ensure systematic thematic analysis and facilitates the synthesis of key themes across the data set. Constant comparison was carried out to ensure that the analysis represents all perspectives and negative (‘deviant’) cases.

Synthesis of multiple data sources

After independent analysis of each data strand was completed, the findings from the multisource process evaluation were integrated to shed a light on implementation, mechanisms of change and the role of contextual factors on implementation and outcomes. To achieve this, key findings from each stage of data collection were synthesised in an analytical integration table. The table was organised to allow consideration of findings from each stage, with each strand being positioned in a single matrix for each PE component (e.g. implementation, mechanisms and context) (see Appendix 1). This method facilitated the systematic synthesis of the process data, drawing out synergistic interpretations to reveal a broader holistic perspective on how the intervention worked in practice.

Implementation

Acceptability of taking part in a research study

Figure 4 shows the pharmacist support and training that was provided for delivering the Bridge-it study. During training sessions, the majority of pharmacists seemed enthusiastic and motivated to take part in the study and often reflected on this in interviews, expressing a desire to take part in more research:

I’m kind of like, I’d love to be involved in more things so I’m always welcome . . . you know, I’m welcome to do more kind of projects like this.

Pharmacist 20, London

FIGURE 4.

A flow chart of pharmacist training and support for delivering the Bridge-it study.

In particular, pharmacists who had been involved in the earlier pilot study15 or who had previous experience of pharmacy-based interventions were particularly enthused and seemed to be more accepting of the pharmacist role in the study, potentially because of familiarity with the kinds of processes and documentation required for trials. It is important to note that pharmacists who were particularly enthusiastic about the benefits of bridging through community pharmacies for patients, and about taking part in the study, were those who tended to recruit the most participants, perhaps highlighting the importance of ‘buy-in’ to the premise of the intervention.

Although most pharmacists reflected positively on the premise of the intervention and on taking part in research to expand contraceptive services, some concerns were raised during training. Pharmacists’ concerns typically focused on worries relating to managing the additional time and workload pressures of taking part in a research study (rather than of delivering POP per se), scepticism that additional time would be set aside to accommodate this, and concerns about how the study would fit with existing practices and changing guidelines relating to EC. These concerns are detailed in The impact of context on the Bridge-it study implementation and outcomes and focus on facilitators of and barriers to delivery and implementation in practice.

Pharmacist and sexual and reproductive health provider awareness of the study and preparedness to deliver

Process evaluation summary and key implications

Key learning and implications

Interpretation

The Bridge-it study demonstrated that this community pharmacist-delivered intervention of provision of a small bridging supply of oral contraception along with EC and the invitation to a SRH clinic resulted in a significantly larger proportion of women using effective contraception 4 months later than standard care alone. The difference in uptake of effective contraception between the groups was large. Furthermore, significantly fewer women in the intervention group than in the control group sought EC again during the study. In addition, consistent with the known safety of the POP,20 no serious adverse events were reported. These findings are clinically important. Given that use of effective contraception prevents unintended pregnancies, the public health benefit of this simple and safe intervention is potentially considerable.

Generalisability and limitations

The Bridge-it study was a well-designed trial, with a sufficiently large sample size to show the expected effect with sufficient power. The primary outcome findings also remained consistent, even when adjusted for a range of factors and the missingness of data. The study was conducted at three different sites across the UK and the trends were consistent at each site. The Bridge-it study findings are also reassuringly consistent with the findings of our earlier pilot study. 15

The pharmacies in the Bridge-it study included a mix of small independent pharmacies and large chain pharmacies, and so should reflect EC service provision across the UK. In addition, the women who participated in the Bridge-it study were of similar age and shared similar characteristics to EC users in previous studies in the UK. 32 The findings should therefore be applicable across the UK.

The intervention tested was a simple one (i.e. a 3-month supply of POP provided by PGDs). In addition, the intervention should not be costly, as POP is an inexpensive drug. However, this will be determined by our cost-effectiveness analysis.

This was a difficult study to undertake. Pharmacists are not usually familiar with undertaking research. The pharmacy is a commercial setting. Requiring pharmacists to recruit participants, provide study information and complete additional paperwork as part of a research study was an added burden to EC provision. Although cluster randomised trials are generally susceptible to bias as allocation is known to participants at consent, our pilot study informed us that a study design of individual randomisation in the community pharmacy setting would not work. 15 The study design meant that different groups of women participated in the two periods. This mix of factors contributed to varying recruitment rates across pharmacies and between periods, resulting in different lengths of required recruitment and not all pharmacies successfully recruiting in both periods. Although this added to the complexity of the study and its management, it produces more robust and generalisable findings.

The main limitation of this study (as for many contraceptive studies) is that the outcome was the uptake of effective contraception rather than the number of unintended pregnancies. We had originally intended to examine abortion rates in each group as a co-primary outcome, but during the course of the study it became evident that we could not recruit sufficient numbers of participants within a realistic time frame, and so we chose to focus solely on uptake of effective contraception, which required smaller sample size. Use of effective contraception should prevent unintended pregnancy; however, the association is more complex, particularly as the use of a method of contraception does not always equate to correct and consistent use. 33 Thousands of participants would have been required for this study to have the power to demonstrate an effect of the intervention on unintended pregnancies. Although participants gave consent to allow subsequent data linkage with abortion registries, the sample size may be too small to show any difference between the groups.

Another consideration is that contraceptive use was self-reported. However, there is some evidence that women’s self-reporting of contraceptive method is reliable. 34 The findings from the limited validation exercise among those women who both obtained contraception from the SRH clinic and responded to follow-up showed that 88% of women were provided with the same method as they had reported. This provides some support for the reliability of self-reporting of contraception. Loss to follow-up for the study was 35%, which was larger than the 25% we expected. However, there was no differential loss to follow-up between groups and the characteristics of women with and without follow-up data were similar.

This study was undertaken with EC containing LNG, which was the most commonly used EC at that time and suitable for immediate start of oral contraception. 35 However, any future rollout of the intervention does not need to be limited to this EC, as it would also be possible for community pharmacists to provide bridging contraception of the POP containing ulipristal acetate (a more effective EC). 36 The only issue would be that women would be advised to wait 5 days after taking ulipristal acetate before starting the POP (rather than immediately or the next day after EC containing LNG) because it has been shown that starting hormonal contraception sooner than 5 days after ulipristal acetate might potentially impair the ability of ulipristal acetate to delay ovulation and therefore render EC less effective. 10,37,38

Overall evidence

With the exception of our earlier pilot study,15 to the best of our knowledge, the only published study that examined the uptake of effective contraception after EC offered women (in Jamaica) a discount on the subsequent cost of contraceptive pills and had no effect on subsequent contraceptive use. 14 The only other intervention that has demonstrated a large impact on the use of effective contraception was the Contraceptive CHOICE Project in the USA,39 in which women requiring contraception in a range of settings were counselled about LARC and provided with their chosen method free of charge in a setting where contraception is usually not free. LARC uptake was 67% in the Contraceptive CHOICE Project cohort, compared with 3% nationally. The Bridge-it study, by comparison, assessed a simple intervention delivered by community pharmacists at EC request.

Recommendations for research

Future research should be directed at a robust evaluation of the impact of the Bridge-it study intervention following any wide-scale implementation (to include impact on abortion rates, cost-effectiveness and on experiences of women and community pharmacists following implementation). This should be cognisant of existing contextual challenges in the pharmacy and SRH context.

The findings of the Bridge-it study are particularly relevant to the NHS during the current Covid-19 pandemic. Pharmacists are already embedded in SRH care in the UK and could provide an alternative way for women to access regular effective contraception. 40 As part of the NHS Long Term Plan, NHS England plans to work with the government to make greater use of community pharmacists’ skills and opportunities to engage patients. 41,42 As part of the deal, the NHS is introducing an expanded clinical role for local pharmacists, beginning ‘a revolution in patient care’ that could see community pharmacy becoming the first port of call for minor illness and health advice.

Therefore, we propose that future research should also focus on the feasibility, acceptability and impact of more contraceptive services being delivered from community pharmacies. Our process evaluation has highlighted the need for sufficient resources for such services to be embedded in routine practice. In doing so, and alongside appropriate advertising and promotion of the service and an emphasis on well-being during contraceptive consultations, services could overcome persistent barriers to accessing and using effective contraception.

Future research should also consider ways to better support research in the community pharmacy setting.

Implications for decision-makers

Impact of patient and public involvement, challenges of patient and public involvement and lessons learnt

List of study pharmacies

Newington Pharmacy, Edinburgh; Boots Princes Street, Edinburgh; Boots Shandwick Place, Edinburgh; Boots Earl Grey Street, Edinburgh; Boots Gyle, Edinburgh; Boots St Patrick Street, Edinburgh; Boots Multrees Walk, Edinburgh; Boots Ocean Terminal, Edinburgh; Boots Edinburgh Fort Retail Park, Edinburgh; Boots Cameron Toll, Edinburgh; Boots Craigleith, Edinburgh; Bristo Square Pharmacy, Edinburgh; Asda Pharmacy, Perth; Boots High Street, Dundee; Boots Perth Road, Dundee; Peace Pharmacy, London; Westbury Chemist, London; Baba Chemist, London; Lings Chemist, London; Streatham Day Lewis, London; Morrisons, Aylesham Centre, London; Evergreen Pharmacy, London; Greenlight Pharmacy, London; Sandylight Pharmacy, London; Greenfields (Day Lewis), London; JP Pharmacy, London; Boots Goodge Street, London; Boots Tottenham Court Road, London; and Boots Holborn, London.

List of study sexual and reproductive health clinics

Chalmers Sexual and Reproductive Health Service, Edinburgh; Tayside Sexual and Reproductive Health Service, Ninewells Hospital, Dundee; Camberwell Sexual Health Centre, London; and The Margaret Pyke/Mortimer Market Centres or the Archway Centre, London.

Collaborators

The Bridge-it study TSC provided oversight for this trial on behalf of the sponsor (University of Edinburgh and NHS Lothian) and the funder. The members are Peter Brocklehurst (chairperson), Birmingham Clinical Trials Unit; Lucy Michie, NHS Ayrshire and Arran; Kaye Wellings, London School of Hygiene and Tropical Medicine; Joanna Loudon, PPI member, Edinburgh; Kirsten Stuart, PPI member, Edinburgh; and Emily Whitaker, PPI member, Edinburgh.

The DMC is an independent multidisciplinary group, consisting of clinicians and statisticians. The members are Claire Anderson (chairperson), University of Nottingham; Elizabeth Allen, London School of Hygiene and Tropical Medicine; and Caroline Moreau, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA. A copy of the DMC charter is held in Edinburgh Clinical Trials Unit.

The study has co-sponsorship between the University Court of the University of Edinburgh and Lothian Health Board. The sponsors’ representative is the Academic and Clinical Central Office for Research and Development (ACCORD) (Edinburgh, UK).

Contributions of authors

Sharon T Cameron (https://orcid.org/0000-0002-1168-2276) was responsible for the study design, wrote the first draft of the report, with significant input from all authors at all stages, and contributed to, read and approved the final report.

Anna Glasier (https://orcid.org/0000-0002-5697-255X) was responsible for the study design, wrote the first draft of the report, with significant input from all authors at all stages, and contributed to, read and approved the final report.

Lisa McDaid (https://orcid.org/0000-0002-7711-8723) was responsible for the study design, wrote the process evaluation component, and contributed to, read and approved the final report.

Andrew Radley (https://orcid.org/0000-0003-4772-2388) was responsible for the study design, and contributed to, read and approved the final report.

Susan Patterson (https://orcid.org/0000-0003-1846-5749) contributed to later stages of study design and analysis, wrote the process evaluation component, and contributed to, read and approved the final report.

Paula Baraitser (https://orcid.org/0000-0002-3354-6494) was responsible for the study design, and contributed to, read and approved the final report.

Judith Stephenson (https://orcid.org/0000-0002-8852-0881) was responsible for the study design, and contributed to, read and approved the final report.

Richard Gilson (https://orcid.org/0000-0002-9572-193X) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Claire Battison (https://orcid.org/0000-0001-9549-4522) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Kathleen Cowle contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Thenmalar Vadiveloo (https://orcid.org/0000-0001-5531-6289) was responsible for the statistical analyses and contributed to data interpretation, and contributed to, read and approved the final report.

Anne Johnstone (https://orcid.org/0000-0003-1886-2138) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Alessandra Morelli (https://orcid.org/0000-0002-9803-2136) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Beatriz Goulao (https://orcid.org/0000-0003-1490-7183) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Mark Forrest (https://orcid.org/0000-0002-2395-8823) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

Alison McDonald (https://orcid.org/0000-0002-0256-2889) contributed to later stages of study design and analysis, and contributed to, read and approved the final report.

John Norrie (https://orcid.org/0000-0001-9823-9252) was responsible for the study design, wrote the first draft of the report, with significant input from all authors at all stages, and contributed to, read and approved the final report.

Publications

Cameron ST, Baraitser P, Glasier A, McDaid L, Norrie J, Radley A, et al. Pragmatic cluster randomised cohort cross-over trial to determine the effectiveness of bridging from emergency to regular contraception: the Bridge-It study protocol. BMJ Open 2019;9:e029978.

Cameron ST, Glasier A, McDaid L, Radley A, Baraitser P, Stephenson J, et al. A pragmatic cluster randomised crossover trial to determine use of effective contraception following provision of the progestogen-only pill to women presenting to community pharmacies for emergency contraception: the Bridge-it study. Lancet 2020;396:1585–94.

Glasier A, Baraitser P, McDaid L, Norrie J, Radley A, Stephenson JM, et al. Emergency contraception from the pharmacy 20 years on: a mystery shopper study. BMJ Sex Reprod Health 2021;47:55–60.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to available anonymised data may be granted following review.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care.