BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] BMJ Open

A cluster randomised controlled trial to determine the effectiveness of bridging from emergency to regular contraception: The Bridge–It study protocol For peer review only Journal: BMJ Open

Manuscript ID bmjopen-2019-029978

Article Type: Protocol

Date Submitted by the 16-Apr-2019 Author:

Complete List of Authors: Cameron, Sharon; University of Edinburgh, Obstetrics and Gynaecology; NHS Lothian, Sexual and Reproductive Health Baraitser, Paula; King's College London School of Medical Education Glasier, Anna; University of Edinburgh McDaid, Lisa; University of Glasgow, MRC/CSO Social and Public Health Sciences Unit Norrie, John; Edinburgh Clinical Trials Unit, University of Edinburgh No. 9, Bioquarter Radley, Andrew; NHS Tayside, Directorate of Public Health; University of Dundee Division of Cardiovascular and Diabetes Medicine, Liver Group Stephenson, Judith; University College London, Instiute for Women's Health Trussell, James; Princeton university, Office of Population Research Battison, Claire; Edinburgh Clinical Trials Unit, University of Edinburgh No. 9, Bioquarter Cameron, Sarah; University of Aberdeen, Health Services Research Unit Cowle, Kathleen; Boots UK Ltd Forrest, Mark; University of Aberdeen, Health Services Research Unit Gilson, Richard; University College London, Institute for Global Health Goulao, Beatriz; University of Aberdeen, Health Services Research Unit Johnstone, Anne; University of Edinburgh, Obstetrics and Gynaecology McDonald, Alison; University of Aberdeen, Health Services Research Unit Morelli, Alessandra; King's College London School of Medical Education Patterson, Susan; University of Glasgow, MRC/CSO Social and Public Health Sciences Unit Sally, Deirdre; University College London, Instiute for Women's Health Stewart, Nicola; University College London, Instiute for Women's Health

contraception, emergency contraception, pharmacy, unintended Keywords: pregnancy, levonorgestrel, progestogen only pill

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 18 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 18

1 2 3 A cluster randomised controlled trial to determine the effectiveness of bridging 4 5 from emergency to regular contraception: The Bridge–It study protocol 6 Bridge-It Study Group* 7 8 *Sharon Cameron1, Paula Baraitser 2, Anna Glasier 1, Lisa McDaid 3, John Norrie 4, 9 Andrew Radley 5, Judith Stephenson 6, James Trussell 7†,Claire Battison 4, Sarah 10 8 9 8 10 8 11 Cameron , Kathleen Cowle , Mark Forrest , Richard Gilson , Beatriz Goulao , 12 Anne Johnstone 1, Alison McDonald 8, Alessandra Morelli 11, Susan Patterson 3, 13 Deirdre Sally 6, Nicola Stewart 6 14 15 16 1 17 Corresponding author: Professor Sharon Cameron, Consultant Gynaecologist, 18 Address: 2a ChalmersFor Street, peer Edinburgh, review EH3 9ESTel: 00only 44 131 536 2091 Email: 19 [email protected] 20 21 22 Word Count: 3767 23 24 25 Funder: HTA/NIHR 26 27 Funding Reference Number: HTA Project 15/113/01 28 29 In addition, the following disclaimer should be included in this paper: 30 31 The views expressed are those of the author(s) and not necessarily those of the 32 33 NIHR, the NHS or the Department of Health 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 18 BMJ Open

1 2 3 ABSTRACT: 4 5 Introduction: Oral emergency contraception (EC) can prevent unintended pregnancy 6 but it is important to start a regular method of contraception. Women in the UK 7 usually access EC from a pharmacy but then need a subsequent appointment with a 8 GP or a sexual and reproductive health (SRH) service to access regular contraception. 9 Unintended pregnancies can occur during this time. 10 11 Methods and analysis: Bridge-It is a pragmatic cluster randomised cohort crossover 12 trial designed to determine whether pharmacist provision of a bridging supply of a 13 progestogen only pill (POP) plus rapid access to a local SRH clinic, results in increased 14 uptake of effective contraception and prevents more unintended pregnancies than 15 16 provision of EC alone. Bridge-It involves 31 pharmacies in three UK regions (London, 17 Lothian and Tayside) aiming to recruit 626 to 737 women. Pharmacies will give EC 18 (levonorgestrel)For according peer to normal reviewpractice and recruit only women to both 19 intervention and the control phases of the study. In the intervention phase, 20 pharmacists will provide the POP (desogestrel) and offer rapid access to a SRH clinic. 21 22 In the control phase, pharmacists will advise women to attend a contraceptive 23 provider for contraception (standard care). 24 Women will be asked four months later about contraceptive use. Data linkage to 25 abortion registries will provide abortion rates over 12 months. The sample size is 26 27 calculated on the primary outcome of effective contraception use at four months 28 (yes/no) with 90% power and a 5% level of significance. Abortion rates will be an 29 exploratory secondary analysis. Process evaluation includes interviews with 30 pharmacists, SRH clinicians and women. Cost-effectiveness analysis will use a 31 healthcare system perspective and be expressed as incremental cost-effectiveness 32 33 ratio. 34 Ethics and dissemination: Ethical approval was received from South East Scotland 35 REC June 2017. Results will be published in peer-reviewed journals and conference 36 presentations. 37 38 Trial registration number ISRCTN 70616901 39 40 41 42 Strengths and limitations: 43 44  Innovative and efficient cluster cohort crossover design 45  Examines impact of the intervention on uptake of effective contraception at 46 four months 47  Examines the important outcome of abortion rates over 1 year as an 48 49 exploratory secondary analysis. 50  Applicable only to women receiving levonogrestrel EC followed by a 51 desogestrel POP 52  Not applicable to use of ulipristal acetate for EC, since hormonal methods of 53 54 contraception such as the desogestrel POP may interact with efficacy of 55 ulipristal acetate, if started within five days 56 57 58 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 18

1 2 3 INTRODUCTION 4 5 Unintended pregnancy is a major public health problem. The UK has among the 6 highest abortion rates in Europe [1]. In 2017 almost 200,000 pregnancies ended in 7 induced abortion [2, 3]. Unintended pregnancy also ends in childbirth; around 10% 8 of UK births are unintended and 25% mistimed [4]. Unintended pregnancy is costly 9 10 to the NHS (estimated to cost over £1 billion annually) [5] and can be distressing for 11 women. Unintended pregnancies are more common in young women from deprived 12 backgrounds, contributing to widening health inequalities for both mother and baby, 13 and their families [2,3]. Unintended childbirth can have both socioeconomic 14 consequences for women and their families and mental health consequences [6]. 15 16 17 Oral emergency contraception (EC) prevents pregnancy in individual women 18 following unprotectedFor sex peer or contraceptive review accidents. onlyEC is only effective if taken 19 before ovulation as it works by inhibiting or delaying ovulation [7]. Since EC became 20 available from pharmacies in the UK without the need for a prescription, there has 21 22 been a change in the pattern of access such that women who seek EC now choose to 23 obtain this from a pharmacy rather than a contraceptive provider such as a GP or 24 sexual and reproductive health (SRH) service [8]. Although trials have shown that this 25 facilitates access to EC and increases use, they have failed to show that this reduces 26 unintended pregnancy rates within the population [9]. 27 28 There are two types of EC: the most widely used EC contains the progestogen 29 levonorgestrel and should be taken within 72 hours of sex; the other EC contains the 30 progesterone receptor modulator ulipristal acetate and should be taken within 120 31 32 hours of sex [10]. Neither formulation of EC prevents conceptions from subsequent 33 acts of sex and the risk of pregnancy is increased up to threefold among women who 34 have further unprotected sex in the same menstrual cycle after using EC than those 35 who do not [10]. An effective method of contraception should therefore be started 36 as soon as possible [10, 11]. However the only contraceptives that can be obtained 37 38 from any pharmacy without a prescription are condoms, which have high failure 39 rates [12]. This means that women usually need to make an appointment with a 40 contraceptive provider (GP or SRH) and may experience delays in accessing regular 41 contraception or lose the motivation to access a regular method altogether, which in 42 turn may result in unintended pregnancies. In addition, in one UK study fewer than 43 44 half of pharmacists gave advice about ongoing contraception after EC [13]. It is 45 possible that if pharmacists could supply a temporary (bridging) method of 46 contraception to women along with EC, this would bridge the gap until women could 47 get an appointment with a contraceptive provider for contraceptive advice and 48 49 supplies. The progestogen only pill (POP) is an effective method of contraception 50 with few contraindications [14] making it safer than the combined oral contraceptive 51 pill for pharmacy provision. However, studies have shown that starting hormonal 52 contraception containing a progestogen within five days of ulipristal acetate may 53 reduce the efficacy of EC and so only EC containing levonorgestrel is suitable for use 54 55 in conjunction with a bridging method of hormonal contraception in this way [15, 56 16]. 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 18 BMJ Open

1 2 3 Pilot 4 5 In a pilot study in Edinburgh of 168 women presenting for EC [17] 11 pharmacies 6 were randomised to one of three groups to provide EC (levonorgestrel) and either 7 (i) standard advice on where to obtain ongoing contraception or (ii) one month of a 8 progestogen only pill (POP) or (iii) the offer of rapid access to a local SRH service. 9 Participants were contacted by telephone six to eight weeks later to determine their 10 11 current contraceptive use. Compared to standard care, the proportion of women 12 using effective contraception was significantly greater in both the POP (56% vs. 16% 13 p=0.001) and the rapid access groups (52% vs. 16% p=0.027). This suggests that a 14 supply of one month of POP after EC or rapid access from a pharmacy to a SRH 15 16 service might increase short-term uptake of effective contraception following EC. 17 We now propose a large randomised trial to determine whether a pharmacy-based 18 intervention designedFor to peer facilitate the review uptake of effective only contraception after EC 19 increases use of effective contraceptive methods including the most effective long 20 acting reversible contraceptive methods (LARC) such as the contraceptive implant 21 22 and intrauterine contraception[12] at four months when compared with standard 23 care. We will examine contraceptive uptake at four months as most POP 24 preparations are packaged as a three month supply and so by four months the 25 pharmacy provided supply will have ended. 26 27 28 Aim 29 The aim is to develop a simple and affordable intervention which facilitates the 30 uptake of effective ongoing contraception among women obtaining EC from 31 pharmacies thereby reducing unintended pregnancy. The primary objective is to 32 33 determine whether offering women attending a pharmacy for EC, a three month 34 bridging supply of POP plus the offer of rapid access to a local SRH service results in 35 increased uptake of effective contraception. This combined intervention (POP plus 36 rapid access) offers both temporary contraception and facilitates access to a 37 38 specialist contraceptive service where all methods of contraception including the 39 most effective LARC methods can be provided. If this intervention leads to increased 40 uptake of effective contraception including LARC methods compared to standard 41 care alone then we might expect that this would translate into fewer unintended 42 pregnancies for women. 43 44 45 METHODS AND ANALYSIS 46 47 Study design and setting 48 49 A pragmatic cluster randomised cohort crossover clinical and cost effectiveness trial 50 including process, outcome and economic evaluation involving 31 pharmacies in 51 three UK regions (15 in London (South and Central), 12 in Lothian (Edinburgh and 52 region) and four in Tayside (Dundee and region). 53 54 55 Patient and Public Involvement 56 The members of the Patient and Public Involvement (PPI) Group at the SRH service in 57 Edinburgh contributed to the design of this study. The study protocol and 58 documentation were reviewed and approved by the chair and members of the PPI 59 60 group. The plain English summary was edited by a PPI member and improved as a result. There are three PPI members that participate and contribute to the Bridge-It 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 18

1 2 3 Trial steering Committee meeting (TSC) that provides oversight of the study. PPI 4 5 group will be involved in the dissemination of the study results. 6 7 Intervention 8 The planned intervention is a composite intervention. Each woman in the 9 intervention phase will receive three months of POP (in a single package covering 3 10 11 months) and an invitation to attend a local participating SRH service to discuss and 12 obtain effective contraception (including LARC methods). Three packets of POP, (75 13 mcg desogestrel; UK) containing 28 tablets each will be provided at no cost to 14 women as a bridging method of contraception, providing them with three months of 15 16 contraception during which time they can get an appointment with a contraceptive 17 provider to obtain their preferred method of contraception. Locally approved Patient 18 Group DirectionsFor (i.e. strict peer criteria to review permit provision only of specified medicines by 19 non-prescribers) will permit participating pharmacists to dispense the POP to women 20 recruited to the study. Pre-study training will be undertaken with participating 21 22 pharmacists including identifying medical contraindications to POP, potential drug 23 interactions medications and ‘missed pill’ guidance for POP. Pharmacists will advise 24 women to start the POP the day following intake of EC [10]. 25 Pharmacists will encourage women to attend the participating local SRH service to 26 27 obtain the contraceptive method of their choice. Participants (intervention phase) 28 will be given a study card to alert staff at SRH services that they are in the Bridge-It 29 study and should be seen as a drop in for contraception that same day. This card will 30 also provide written information about the location and opening hours of the local 31 participating SRH service. These SRH services are within a 5-mile radius of the 32 33 participating pharmacies and provide all methods of contraception at no cost as is 34 the norm in the NHS. 35 36 Standard Care 37 38 A mystery shopper exercise [13] will be undertaken in all 31 participating pharmacies 39 to characterise ‘standard care’ (usually verbal advice to visit a clinic for 40 contraception, with/without written information) in the control phase. The mystery 41 shopper visits will be conducted when the pharmacy is not recruiting and just before 42 the control phase starts. The mystery shoppers and the scenario used will be chosen 43 44 by the Patient Public Involvement group. A simple scenario relating to request for EC 45 will be used. Immediately after leaving the pharmacy the mystery shopper will 46 complete a standard data collection proforma, recording any information given by 47 the pharmacist about use of contraception after taking EC, including provision of 48 49 written information. 50 51 Participants 52 We will recruit a total of 626 to 737 women presenting for EC. The final number will 53 be determined based on the observed ratio of the between-period within-cluster 54 55 variability (BPC) and the within-period within-cluster (WPC) variability – with the 56 larger sample size (near the 737 upper limit) required for values of BPC close to zero, 57 and the smaller sample size (near the 626 lower limits) required for values of BPC 58 close to WPC (18). Each pharmacy will be expected to recruit an average of 12-13 59 60 women to the intervention arm and 12-13 women to standard care. This allows for a 25% loss to follow up at four months (missing data on primary outcome). 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 18 BMJ Open

1 2 3 4 5 Randomisation 6 Each pharmacy will be randomised to either the intervention phase for 7 approximately 20 weeks followed by standard care phase for 20 weeks or vice-versa 8 with a wash-out period of two weeks between the two phases. The order in which 9 pharmacy allocation to each arm is undertaken (intervention or control first) will be 10 11 randomised (Figure 1). 12 This is a cluster cohort crossover design so it is the pharmacy that is the unit of 13 randomisation and the 'crossover' means that we are just randomising the order that 14 each pharmacy gives the intervention in. The ‘cohort’ label means that we expect 15 16 different women to be recruited within each site in the two periods (intervention 17 and control phases). 18 For peer review only 19 Recruitment 20 The pharmacist will assess medical eligibility of women presenting for EC for the 21 22 study, provide EC according to normal practice and invite eligible women to 23 participate. The EC used in this study is levonorgestrel and will be given in the 24 clinically indicated dose for the woman’s weight (1.5 mg or 3 mg levonorgestrel) 25 [10]. 26 27 Inclusion and exclusion criteria are shown in Table 1. Women who give written 28 consent will be recruited in the study. We recognise the importance of participant 29 retention and will offer a voucher of £10 at recruitment [19]. 30 31 Outcomes 32 33 A full list of study outcome measures is included in Table 2. Outcomes at four 34 months will primarily be collected via telephone interviews or via web based 35 questionnaires. However, participants will also have the option to provide the same 36 information by a postal questionnaire. The primary outcome is use of effective 37 38 contraception at four months (intervention vs standard care). 39 Secondary outcomes are proportion of participants having an abortion within 12 40 months of EC use using record linkage from participants to national registries and 41 cost effectiveness. 42 43 44 Process evaluation measures 45 A process evaluation will be conducted as part of the study to assess potential issues 46 concerning intervention implementation, the causal mechanisms of impact, and the 47 48 contextual factors that could affect these. The process evaluation will comprise of 49 quantitative and qualitative data measures, as detailed in Table 3. 50 51 52 Data Collection 53 Quantitative data 54 55 Participant flow: Participant flow through the study will be assessed and reported 56 following the CONSORT flow chart. 57 58 Baseline 59 Participant demographics and reproductive history is collected at recruitment by a 60 self-administered paper questionnaire given to them by the pharmacist.

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 18

1 2 3 Demographic data will also be reported for the process evaluation, protocol 4 5 adherence checklists and for recruitment screening forms (see Table 3). 6 7 Contraceptive use at 4 months 8 This will be based upon self-reported data from women at a telephone follow up 9 interview with a research nurse at four months after obtaining EC. If participants 10 11 prefer, the questions can be self-completed by a web based questionnaire or paper 12 questionnaire sent by post. Women will be asked what method of contraception 13 they are using (if any), if they attended a GP or SRH service for this, if they used the 14 POP (intervention phase only) and their pregnancy status. If pregnancy has occurred 15 16 since EC then the validated London Measure of Unintended Pregnancy tool will be 17 administered to measure intended-ness of the index pregnancy [20] (Table 2). 18 For peer review only 19 Abortion rates at 12 months 20 Information Services Division (ISD Scotland) and Department of Health (DOH 21 22 England) will provide the number of abortions occurring during the 12 month follow 23 up period in each arm by conducting linkage of the identifiers (collected at baseline) 24 from study participants. 25 26 27 Validation 28 We will check with data from local SRH services to determine the numbers of 29 participants from intervention and control phases who attend the local participating 30 SRH service, and which method of contraception they received. 31 32 33 Qualitative data 34 Semi-structured, qualitative telephone interviews of a purposive sample of up to 60 35 women who received the intervention (approximately 22 in London, 30 in Edinburgh 36 and 8 in Dundee) will be undertaken. Participants who consent to these telephone 37 38 interviews will be contacted by the Process Evaluation Research Assistant. Interviews 39 will explore experience of intervention acceptability in more depth and assess 40 experiences of bridging from EC to regular contraception, and reasons for doing so 41 or not (Table 3). Interviews will be conducted soon after the four month follow up. 42 43 44 Interviews with 27 pharmacists and 12 SRH service providers will explore their 45 perceptions of barriers and facilitators to implementation and more broadly, their 46 views on the intervention, the trial, and the target population. Interviews will be 47 conducted by the process evaluation research assistant soon after the intervention 48 49 phase has completed. 50 51 For the process evaluation, data collection also includes: review of training and 52 materials; observation of training; mapping of local contraceptive services within 10 53 miles of study sites and monitoring of contemporaneous events, such as relevant 54 55 high coverage media stories using Google Alerts (Table 3). 56 57 Sample size calculation 58 The study is a cluster randomised cohort crossover trial. Ideally the control and 59 60 intervention phases should be of roughly equal duration and size, and the sample size is calculated assuming an equal cluster size in both control and intervention 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 18 BMJ Open

1 2 3 periods, and equal across sites (the expected site/period average). In practice, there 4 5 is variability in EC demand across sites and over time: for example, demand is 6 affected by peak holiday periods (recruitment decreases) and student term times 7 (recruitment increases); and the ability of a pharmacy to translate that variable 8 demand into study recruits depends on many factors, including changes to 9 circumstances at individual pharmacies or loss of / change of pharmacists at multi- 10 11 pharmacist stores. 12 Informed by our pilot study, we have assumed that effective contraception use in 13 the control would be 30% and we were likely to achieve a 50% relative improvement 14 to 45%. This means the sample size is in the range 626 to 737, assuming 25% of 15 16 women do not provide four-month contraceptive use data, and an average cluster 17 size of 12-13 in each period, and around 25 pharmacies taking part, with 90% power 18 and a 5% level ofFor significance. peer The uncertainty review in the requiredonly sample size rests on 19 the assumed between-period within-cluster correlation (BPC) and its relationship to 20 the other component of variability, the within-period within-cluster correlation 21 22 (WPC)[18]. The WPC is the usual correlation (known as the Intra Class Correlation– 23 ICC in a standard parallel groups non-crossover cluster setting) of two individuals’ 24 outcomes within a cluster (in the same period). The BPC on the other hand is the 25 correlation between two individuals’ outcomes in the same cluster between the two 26 27 periods. If the BPC is zero, there is no advantage in a crossover design over the 28 standard parallel groups cluster design; if the BPC equals the WPC then the crossover 29 is as efficient as an individually randomised design. We will finalise the sample size 30 depending on the observed ratio of the BPC to WPC, and the observed rate of 31 attrition, but still assuming the same control rate and treatment effect, once we 32 33 have 4 month data on at least 500 participants. 34 35 Quantitative Analyses 36 There will be a single analysis at study end (there is no opportunity for any interim 37 38 analyses given the crossover design) although an independent Data Monitoring 39 Committee will monitor study progress and any safety issues. This will follow the 40 intention to treat principle and will use a hierarchical model appropriate for the 41 specific outcome. For the primary outcome this will be a mixed effects logistic 42 regression, using the hierarchical model approach as recommended by Turner for a 43 44 cluster crossover design [21]. We will pre-specify any individual level (or cluster 45 level) covariates that we intend to adjust for, and the comprehensive Statistical 46 Analysis Plan will specify the sensitivity type analyses that will explore how robust 47 the findings are to any missing data at the cluster level (probably unlikely) and the 48 49 individual level (expected to be substantial for the patient reported outcomes at four 50 months). As well as the usual assumption of missing at random, we will try to explore 51 possible mechanisms for non-ignorable (informative missingness) at the individual 52 level which may well be operating in this context. Subgroup analyses (appropriately 53 analysed by testing treatment by subgroup interaction) will explore the possible 54 55 effect modification by LARC (most effective contraceptive methods) vs non-LARC. 56 57 Qualitative and mixed-methods process evaluation analyses 58 All process data will be analysed independently of the outcome data and, 59 60 importantly, documented before the outcomes are known. Qualitative analysis of in -depth interviews will be recorded and transcribed verbatim. Transcription and 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 18

1 2 3 analysis (proceeding case by case) will start with the first interview and be ongoing 4 5 during the course of data collection, allowing for emergent themes to be identified 6 and explored in future interviews. The transcripts will be read repeatedly and coded 7 for analysis. Data management will be assisted by the software, QSR NVivo 10. 8 Analysis will be undertaken using ‘Framework Analysis’ a method of proven validity 9 and reliability where data are coded, indexed and charted systematically, then 10 11 organised using a matrix or framework [22]. Constant comparison will be carried out 12 to ensure that the analysis represents all perspectives and negative (‘deviant’) cases. 13 14 The multi-source process evaluation will be synthesised to address the three key 15 16 process evaluation questions: i) what was delivered, ii) how it was delivered, and iii) 17 what role context may have had in shaping the delivery/outcomes 18 For peer review only 19 Economic evaluation: An economic evaluation will be undertaken comparing the 20 intervention and control arms in a cost effectiveness analysis. A trial-based analysis 21 22 will be followed by the construction of a decision model to extrapolate future costs 23 and benefits beyond the completion of the trial. The overall perspective used will be 24 that of the health system. Costs will include the pharmacist training to provide POP, 25 direct and indirect costs of health service use, and the provision and dispensing of 26 27 POP. We will compare the costs to the NHS in the intervention and control arms. To 28 account for differences in the numbers of women in the two arms, we will compare 29 the cost per woman in each arm. In the control arm, the costs are (i) cost of EC, (ii) 30 cost of pharmacist provision of EC, (iii) cost of abortions. In the intervention arm, the 31 costs are in addition to these (iv) the cost of the POP, (v) cost of pharmacist training 32 33 to provide POP and (vi) cost of pharmacist provision of POP. The costs (i) and (ii) are 34 the same in both groups and so the extra cost of the intervention will be the sum of 35 (iv), (v) and (vi). The cost per women who has an abortion is the same in both groups 36 except that we hypothesise that the abortion rate will be lower in the intervention 37 38 group. We can then state the outcome as conventional incremental cost 39 effectiveness ratio i.e. for every £100 spent on the intervention results in x fewer 40 abortions for a savings of £Y. If Y is greater than 100 then the intervention is cost 41 effective. We will examine the sensitivity of the outcomes to variations in the costs 42 of iv, v, and vi. 43 44 45 46 Data Management and Clinical Trial Unit support 47 Data will be collected on a paper case report form and will be entered directly into 48 49 the trial database. Data will be entered into a trial database by pharmacists, research 50 nurses or staff at the trial co-ordinating centre. The data management and statistical 51 support for the study is provided by the UKCRC registered Clinical Trial Unit (CTU) 52 the Centre for Healthcare Randomised Trials (CHaRT) at the University of Aberdeen; 53 while the trial management is provided by another UKCRC registered CTU, the 54 55 Edinburgh Clinical Trials Unit (ECTU) at Edinburgh University. 56 57 Ethics and dissemination 58 The Bridge-It trial involves procedures and medications which are well established in 59 60 current NHS clinical practice and use. Adverse events may occur during or after the use of EC or POP and are well documented in the POP patient information leaflet. 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 18 BMJ Open

1 2 3 Serious adverse events will be recorded at the four month follow up interview. The 4 5 study will be conducted in accordance with the principles of Good Clinical Practice 6 (GCP). 7 A favorable ethical opinion has been obtained from the South East Scotland REC in 8 June 2017. Approvals have been obtained from NHS Research Scotland (NRS) and 9 Health Research Authority (HRA) England prior to commencement of the study. 10 11 Annual progress reports and a final report at the conclusion of the trial will be 12 submitted to REC within the timelines defined in the regulations. 13 The Bridge-It study website will include trial materials, trial progress, and summaries 14 of key findings. In addition, public engagement and dissemination will also be 15 16 undertaken via our Patient Public Involvement group. 17 The results of the study will be published in the academic journals and all 18 participants willFor be offered peer a lay summary review of the main only findings of the study. The 19 findings will also be presented at national and international conferences and 20 disseminated via social media. 21 22 23 Trial Status 24 As at 7th Feb 2019 the study had recruited 503 participants across 29 sites. 25 Recruitment to the first period completed on 13th Jan 2019, with 391 participants 26 27 recruited at 29 sites. Recruitment is scheduled to be completed by Jun 2019, with 28 analyses of the 4-month primary outcome expected to be available by Oct 2019. 29 30 DISCUSSION 31 Unintended pregnancy remains a major public health problem. The proposed 32 33 intervention in the Bridge-It study provides both temporary contraception (the POP) 34 and facilitates access to effective contraception at a local SRH clinic. With the cluster 35 crossover design, each cluster will act as its own control and fewer pharmacies are 36 required than with a parallel cluster design. 37 38 If our proposed intervention works, then this could prevent more unintended 39 pregnancies for more women. If the intervention is cost effective then it could have 40 cost savings for the NHS. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 18

1 2 3 REFERENCES. 4 5 1. Sedgh G, Singh S, Shah IH et al. Induced abortion: incidence and trends worldwide 6 from 1995 to 2008. Lancet. 2012Feb 18;379(9816):625-32. 7 8 2.Department of Health. Abortion Statistics, England and Wales: 2017. 9 https://www.gov.uk/government/publications/report-on-abortion-statistics-in- 10 11 england-and-wales-for-2014 (access date 11 Jan 2019) 12 13 3. Information Services Division. Termination of Pregnancy Statistics 2017 14 http://www.isdscotland.org/Health-Topics/Sexual-Health/Abortions/ (access date 11 15 16 Jan 2019) 17 18 4. Wellings K, JonesFor KG, Mercerpeer CH et review al. The prevalence only of unplanned pregnancy 19 and associated factors in Britain: findings from the third National Survey of Sexual 20 Attitudes and Lifestyles (Natsal-3). Lancet 2013;382:1807-16 21 22 23 5. Thomas C, Cameron ST. Can we reduce costs and prevent more unintended 24 pregnancies? A cost of illness study illustrating the cost-saving potential of a more 25 effective method of EHC. BMJ open 2013;3:e003815.doi:10.1136 26 27 28 6. SmithBattle L, Leonard V. Inequities compounded: explaining variations in the 29 transition to adulthood for teen mothers' offspring. J Fam Nurs. 2012;18(3):409-31. 30 31 7. Cameron ST, Li H R, Gemzell –Danielsson K. Current controversies with Emergency 32 33 Contraception: a review. BJOG 2017(13):1948-1956. 34 35 8. Marston C, Meltzer H, Majeed A. Impact on contraceptive practice of making 36 emergency hormonal contraception available over the counter in Great Britain: 37 repeated cross sectional surveys.BMJ. 2005Jul 30;331(7511):271. 38 39 40 9. Polis CB, Grimes DA, Schaffer K, et al . Advance provision of emergency 41 contraception for pregnancy prevention. Cochrane Database of Systematic Reviews 42 2007,Issue2.Art. No: CD005497. DOI:0.1002/14651858.CD005497.pub2. 43 44 45 10. Faculty of Sexual and Reproductive Healthcare, UK. Emergency contraception. 46 2017. www.fsrh.org (access date 11 January 2019) 47 48 49 11. Cheng L, Che Y, Gülmezoglu AM. Interventions for emergency contraception. 50 Cochrane Database of Systematic Reviews 2012;Issue 8. Art. No.: CD001324. DOI: 51 10.1002/14651858.CD001324.pub4. 52 53 54 12. Trussell J. Contraceptive efficacy. Global Library of Womens 55 56 Medicinehttp://www.glowm.com/section_view/heading/Contraceptive%20Efficacy/i 57 tem/374 ( access date 29 Jan 2019) 58 59 60

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 18 BMJ Open

1 2 3 13 .Glasier A, Manners RJ, Loudon JC, et al. Community pharmacists providing 4 5 emergency contraception give little advice about future contraceptive use: a mystery 6 shopper study. Contraception 2010;82: 538–42. 7 8 9 14. The Faculty of Sexual and Reproductive Healthcare (FSRH). UK medical Eligibility 10 criteria for Contraceptive Use. http://www.fsrh.org/standards-and-guidance/uk- 11 12 medical-eligibility-criteria-for-contraceptive-use/ (access date 11 Jan 2019) 13 14 15. Brache V, Cochon L, Duijkers IJ, et al. A prospective, randomized, 15 pharmacodynamic study of quick-starting a desogestrel progestin-only pill following 16 17 ulipristal acetate for emergency contraception. Hum Reprod 2015;30(12):2785-2793 18 For peer review only 19 16. Cameron ST, Berger C, Michie L,et al. The effects on ovarian activity of 20 ‘quickstarting’ a combined oral contraceptive pill after ulipristal acetate : 21 prospective, randomised, double-blind parallel-arm, placebo-controlled study. Hum 22 23 Reprod 2015;30(7):1566-1572. 24 25 17. Michie L, Cameron ST, Glasier A et al .Pharmacy based interventions for initiating 26 effective contraception following the use of emergency contraception: a pilot study. 27 28 Contraception 2014 ;90, 447-53 29 30 18. Arnup SJ, McKenzie JE, Hemming K, et al. Understanding the cluster crossover 31 design: a graphical illustration of the components of variation and a sample size 32 tutorial. Trials. 2017;18; 381. 33 34 35 19. Brueton VC, Tierney J, Stenning S, et al. Strategies to improve retention in 36 randomised trials. Cochrane Database Syst Rev. 2013 Dec 3; 12:MR000032. doi: 37 10.1002/14651858.MR000032.pub2. 21 38 39 40 20. Cameron ST ,Glasier A. Identifying women in need of further discussion about the 41 decision to have an abortion and eventual outcome. Contraception. 2013;88(1): 128- 42 32 43 44 45 21. Turner RM, Omar RZ, Thompson SG. Modelling multivariate outcomes in 46 hierarchical data, with application to cluster randomised trials. Biom J. 2006 47 Jun;48(3):333-45. 48 49 50 22. Ritchie J, Spencer L. Qualitative research practice: a guide for social science 51 students and researchers. Sage: London, 2003. 52 53 Authors’ contributions: 54 55 56 STC, AG, JN, LMc D, AR, PB, JS, JT developed the original protocol. CB contributed to 57 later stages of study design. STC wrote the draft of the protocol with significant 58 input from all authors at all stages. All authors contributed, read and approved the 59 final manuscript. 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 18

1 2 3 Author affiliations 4 5 6 1. Chalmers Centre, NHS Lothian, Edinburgh EH3 9ES and University of 7 Edinburgh 8 2. King’s College Hospital NHS Foundation Trust, London, SE5 9RS 9 3. MRC/CSO Social & Public Health Sciences Unit, University of Glasgow, 10 11 Glasgow G2 3AX 12 4. Edinburgh Clinical Trial Unit, Usher Institute, University of Edinburgh, 13 Edinburgh EH16 4UX 14 5. King’s Cross Hospital, NHS Tayside, Dundee DD3 8EA 15 16 6. University College London, EGA Institute for Women’s Health, London WC1E 17 6AU 18 7. †DeceasedFor December peer 2018. Princetonreview University only USA. 19 8. Health Service Research Unit, University of Aberdeen, Aberdeen, AB25 2ZD 20 9. Boots UK, North Hub Office, Falkirk FK1 1ES 21 22 10. University College London, Institute for Global Health, London WC1E 6JB 23 11. King’s College London, SE5 9RJ 24 25 26 27 28 Acknowledgements 29 30 Thank you to the members of the Patient and Public Involvement Group at the SRH 31 service in Edinburgh who contributed to the design of this study. Thanks to the 32 women who have participated in this study and to the community pharmacists, 33 34 research staff and health care professionals at the local SRH services who have 35 assisted with the implementation of this study. Thanks to Laura Flett for trial 36 management support. 37 38 The Bridge-It Study Steering Committee provides oversight for this trial on behalf of 39 40 the sponsor (University of Edinburgh and NHS Lothian) and the funder. The 41 members are: Professor Peter Brocklehurst (Chair), Birmingham Clinical Trials Unit; 42 Dr Lucy Michie, Sandyford Glasgow; Professor Kaye Wellings, London School of 43 Hygiene and Tropical Medicine; Joanna Loudon, PPI member, Edinburgh; Kirsten 44 45 Stuart PPI member Edinburgh; Emily Whittaker PPI member Edinburgh. 46 The Data Monitoring Committee is an independent multidisciplinary group consisting 47 of clinicians and statisticians. The members are: Professor Claire Anderson (Chair), 48 University of Nottingham; Professor Elizabeth Allen, London School of Hygiene and 49 Tropical Medicine; Professor Caroline Moreau, Johns Hopkins Bloomberg School of 50 51 Public Health, USA. 52 53 54 The paper presents independent research funded by the National Institute for Health 55 56 Research (NIHR). The views expressed are those of the authors and not necessarily 57 those of the NHS, the NIHR or the Department of Health. 58 59 Funding statement: The Bridge-It study is funded by the National Institute for Health 60 Research’s Health Technology Assessment Programme. HTA Project:15/113/01

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 18 BMJ Open

1 2 3 Competing interests statement. 4 5 6 AG is a member of HRA Pharma scientific advisory board. 7 8 PB is a Clinical Director of the not-for profit community interest company SH:24 that 9 provides online sexual health services in partnership with the NHS. 10 11 12 AR Research grants, educational grants and consultancy with Gilead 13 Research grants from Roche and BMS 14 Educational grants from Abbvie 15 16 17 JN is Deputy Chair of the NIHR/HTA General Board Committee. NIHR/HTA funded 18 this research. For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 18

1 2 3 Table 1. Inclusion and exclusion criteria 4 5 6 Inclusion Criteria 7  Intake of EC (levonorgestrel) 8 9  Capacity to give informed consent to participate in the trial which includes 10 adherence to trial requirements 11  Age 16 years or over 12 13  Willing to give contact details and be contacted at four months by phone or 14 text or e-mail or post 15 16  Willing to give identifying data sufficient to allow data linkage with NHS 17 registries 18 For peer review only 19 20 21 Exclusion Criteria 22  Contraindications to the POP 23  On medication that interacts adversely with POP 24 25  Already using a hormonal method of contraception 26  Require interpreting services 27 28  If pharmacist has concerns about non-consensual sex 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 18 BMJ Open

1 2 3 Table 2. Study outcomes 4 5 6 Data Source 7 Main outcome Use of effective Self reported 8 contraceptive method (telephone or self 9 10 (hormonal or completed survey) at 4 11 intrauterine) in months 12 intervention vs control 13 at 4 months 14 15 16 17 18 Secondary OutcomesFor peerNumbers review undergoing an onlyNational abortion 19 abortion within 12 months registries 20 21 intervention vs control 22 23 24 25 26 27 28 29 30 Incremental cost 31 Economic evaluation 32 effectiveness ratio for 33 pregnancies prevented 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 18

1 2 3 Table 3: Process evaluation methods and data 4 5 6 7 8 Theory 9 10 Theory of change model 11 12 Study Team 13 14 Pharmacy recruitment forms (study team members involved in recruitment will 15 16 routinely record decision making contributing to pharmacy selection, including: number 17 of contacts made; responses from potential pharmacists; rationales for 18 inclusion/exclusion;For and peer reasons for reviewrefusal) only 19 20 Pharmacists 21 22 23 Participant observation of training & review of training and intervention materials 24 25 Recruitment monitoring forms (n=100% of pharmacists) & protocol adherence 26 checklists (n=100% of pharmacists) 27 28 Follow-up semi-structured telephone interviews with pharmacists (n=27; one with each 29 30 pharmacy involved). 31 32 SRH Providers 33 34 Semi-structured telephone interviews with SRH providers (n=12; with Service Manager, 35 mix staff at 2x services in London; 1x service in Edinburgh; 1x service in Dundee). 36 37 38 Participants 39 40 Telephone questionnaire administered by Research Nurse at 4 months post- 41 intervention (n=100% participants) 42 43 Semi-structured telephone interviews at 4 months post-intervention (n=60; 22 in 44 London, 30 in Edinburgh and 8 in Dundee) 45 46 47 Context 48 49 Audit of local contraceptive services within 10 miles of study sites in London, Edinburgh 50 and Dundee 51 52 Monitoring of contemporaneous events, such as relevant high coverage media stories 53 54 using Google Alerts 55 56 57 58 59 60

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 18 BMJ Open

1 2 3 Figure 1: Bridge-It study flowchart 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open

A pragmatic cluster randomised cohort crossover trial to determine the effectiveness of bridging from emergency to regular contraception: The Bridge–It study protocol For peer review only Journal: BMJ Open

Manuscript ID bmjopen-2019-029978.R1

Article Type: Protocol

Date Submitted by the 09-Aug-2019 Author:

Primary Subject Sexual health Heading:

Secondary Subject Heading: Public health

contraception, emergency contraception, pharmacy, unintended Keywords: pregnancy, levonorgestrel, progestogen only pill

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 25 BMJ Open

1 2 3 A pragmatic cluster randomised cohort crossover trial to determine the 4 5 effectiveness of bridging from emergency to regular contraception: The Bridge–It 6 study protocol 7 Bridge-It Study Group* 8 9 *Sharon T Cameron1, Paula Baraitser 2, Anna Glasier 1, Lisa McDaid 3, John Norrie 4, 10 5 6 7† 4 11 Andrew Radley , Judith Stephenson , James Trussell ,Claire Battison , Sarah 12 Cameron 8 , Kathleen Cowle 9, Mark Forrest 8, Richard Gilson 10, Beatriz Goulao 8, 13 Anne Johnstone 1, Alison McDonald 8, Alessandra Morelli 11, Susan Patterson 3, 14 Deirdre Sally 6, Nicola Stewart 6 15 16 17 18 1CorrespondingFor author: Professorpeer Sharon review Cameron, Consultant only Gynaecologist, 19 Address: 2a Chalmers Street, Edinburgh, EH3 9ESTel: 00 44 131 536 2091 Email: 20 [email protected] 21 22 23 Word Count: 3767 24 25 26 27 Funder: HTA/NIHR 28 Funding Reference Number: HTA Project 15/113/01 29 Protocol version 5, 8 January 2019 30 31 In addition, the following disclaimer should be included in this paper: 32 33 34 The views expressed are those of the author(s) and not necessarily those of the 35 NIHR, the NHS or the Department of Health 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 25 BMJ Open

1 2 3 ABSTRACT: 4 5 Introduction: Oral emergency contraception (EC) can prevent unintended pregnancy 6 but it is important to start a regular method of contraception. Women in the UK 7 usually access EC from a pharmacy but then need a subsequent appointment with a 8 GP or a sexual and reproductive health (SRH) service to access regular contraception. 9 Unintended pregnancies can occur during this time. 10 11 Methods and analysis: Bridge-It is a pragmatic cluster randomised cohort crossover 12 trial designed to determine whether pharmacist provision of a bridging supply of a 13 progestogen only pill (POP) plus rapid access to a local SRH clinic, results in increased 14 uptake of effective contraception and prevents more unintended pregnancies than 15 16 provision of EC alone. Bridge-It involves 31 pharmacies in three UK regions (London, 17 Lothian and Tayside) aiming to recruit 626 to 737 women. Pharmacies will give EC 18 (levonorgestrel)For according peer to normal reviewpractice and recruit only women to both 19 intervention and the control phases of the study. In the intervention phase, 20 pharmacists will provide the POP (desogestrel) and offer rapid access to a SRH clinic. 21 22 In the control phase, pharmacists will advise women to attend a contraceptive 23 provider for contraception (standard care). 24 Women will be asked four months later about contraceptive use. Data linkage to 25 abortion registries will provide abortion rates over 12 months. The sample size is 26 27 calculated on the primary outcome of effective contraception use at four months 28 (yes/no) with 90% power and a 5% level of significance. Abortion rates will be an 29 exploratory secondary analysis. Process evaluation includes interviews with 30 pharmacists, SRH clinicians and women. Cost-effectiveness analysis will use a 31 healthcare system perspective and be expressed as incremental cost-effectiveness 32 33 ratio. 34 Ethics and dissemination: Ethical approval was received from South East Scotland 35 REC June 2017. Results will be published in peer-reviewed journals and conference 36 presentations. 37 38 Trial registration number ISRCTN 70616901; pre results 39 40 41 42 Strengths and limitations: 43 44  Examines the important outcome of abortion rates over 1 year as an 45 exploratory secondary analysis. 46  Applicable only to women receiving levonogrestrel EC followed by a 47 desogestrel POP 48 49  Not applicable to use of ulipristal acetate for EC, since hormonal methods of 50 contraception such as the desogestrel POP may interact with efficacy of 51 ulipristal acetate, if started within five days 52 53 54 55 56 57 58 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 25

1 2 3 INTRODUCTION 4 5 Unintended pregnancy is widely perceived as a major public health problem. 6 Unintended pregnancy commonly ends in abortion and the UK has among the 7 highest abortion rates in Europe [1]. In 2017 almost 200,000 pregnancies ended in 8 induced abortion [2, 3]. Unintended pregnancy also ends in childbirth; around 10% 9 10 of UK births are unintended and 25% mistimed [4]. Unintended pregnancy is costly 11 to the NHS (estimated to cost over £1 billion annually) [5] and can be distressing for 12 women. Unintended pregnancies are more common in young women from deprived 13 backgrounds, contributing to widening health inequalities for both mother and baby, 14 and their families [2,3]. Unintended childbirth can have both socioeconomic 15 16 consequences for women and their families and mental health consequences [6]. 17 18 Oral emergencyFor contraception peer (EC) preventsreview pregnancy only in individual women 19 following unprotected sex or contraceptive accidents. EC is only effective if taken 20 before ovulation as it works by inhibiting or delaying ovulation [7]. Since EC became 21 22 available from pharmacies in the UK without the need for a prescription, there has 23 been a change in the pattern of access such that women who seek EC now choose to 24 obtain this from a pharmacy rather than a contraceptive provider such as a general 25 practitioner (GP) or sexual and reproductive health (SRH) service [8]. Although trials 26 have shown that this facilitates access to EC and increases use, they have failed to 27 28 show that this reduces unintended pregnancy rates within the population [9]. 29 There are two types of EC: the most widely used EC contains the progestogen 30 levonorgestrel and should be taken within 72 hours of sex; the other EC contains the 31 32 progesterone receptor modulator ulipristal acetate and should be taken within 120 33 hours of sex [10]. Neither formulation of EC prevents conceptions from subsequent 34 acts of sex and the risk of pregnancy is increased up to threefold among women who 35 have further unprotected sex in the same menstrual cycle after using EC than those 36 who do not [10]. An effective method of contraception should therefore be started 37 38 as soon as possible [10, 11]. However, the only contraceptives that can be obtained 39 from any pharmacy without a prescription are condoms, which have high failure 40 rates [12]. This means that women usually need to make an appointment with a 41 contraceptive provider (GP or SRH) and may experience delays in accessing regular 42 contraception or lose the motivation to access a regular method altogether, which in 43 44 turn may result in unintended pregnancies. In addition, although pharmacists in the 45 UK are supposed to advise women on where to obtain ongoing contraception after 46 EC, in one study fewer than half of pharmacists did so [13]. It is possible that if 47 pharmacists could supply a temporary (bridging) method of contraception to women 48 49 along with EC, this would bridge the gap until women could get an appointment with 50 a contraceptive provider for contraceptive advice and supplies. The progestogen 51 only pill (POP) is an effective method of contraception with few contraindications 52 [14] making it safer than the combined oral contraceptive pill for pharmacy 53 provision. However, studies have shown that starting hormonal contraception 54 55 containing a progestogen within five days of ulipristal acetate may reduce the 56 efficacy of EC and so only EC containing levonorgestrel is suitable for use in 57 conjunction with a bridging method of hormonal contraception in this way [15, 16]. 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 25 BMJ Open

1 2 3 4 5 Pilot 6 In a pilot study in Edinburgh of 168 women presenting for EC [17] 11 pharmacies 7 were randomised to one of three groups to provide EC (levonorgestrel) and either 8 (i) standard advice on where to obtain ongoing contraception or (ii) one month of a 9 progestogen only pill (POP) or (iii) the offer of rapid access to a local SRH service. 10 11 Participants were contacted by telephone six to eight weeks later to determine their 12 current contraceptive use. Compared to standard care, the proportion of women 13 using effective contraception was significantly greater in both the POP (56% vs. 16% 14 p=0.001) and the rapid access groups (52% vs. 16% p=0.027). This suggests that a 15 16 supply of one month of POP after EC or rapid access from a pharmacy to a SRH 17 service might increase short-term uptake of effective contraception following EC. 18 We now proposeFor a large peerrandomised review trial to determine only whether a pharmacy-based 19 intervention designed to facilitate the uptake of effective contraception after EC 20 increases use of effective contraceptive methods including the most effective long 21 22 acting reversible contraceptive methods (LARC) such as the contraceptive implant 23 and intrauterine contraception [12] at four months when compared with standard 24 care. We will examine contraceptive uptake at four months as most POP 25 preparations are packaged as a three month supply and so by four months the 26 27 pharmacy provided supply will have ended. 28 29 Aim 30 The aim is to develop a simple and affordable intervention which facilitates the 31 uptake of effective ongoing contraception among women obtaining EC from 32 33 pharmacies thereby reducing unintended pregnancy. The primary objective is to 34 determine whether offering women attending a pharmacy for EC, a three month 35 bridging supply of POP plus the offer of rapid access to a local SRH service results in 36 increased uptake of effective contraception. The study POP (desogestrel) is 37 38 commonly used in the UK. In contrast to other POP s, the desogestrel POP reliably 39 inhibits ovulation and has similar effectiveness to a combined hormonal oral 40 contraceptive pill (COCP), yet fewer contraindications than a COCP [14,15]. This 41 combined intervention (POP plus rapid access) offers both a highly safe temporary 42 method of contraception and facilitates access to a specialist contraceptive service 43 44 where all methods of contraception including the most effective LARC methods can 45 be provided. If this intervention leads to increased uptake of effective contraception 46 including LARC methods compared to standard care alone then we might expect that 47 this would translate into fewer unintended pregnancies for women. 48 49 50 METHODS AND ANALYSIS 51 Study design and setting 52 A pragmatic cluster randomised cohort crossover trial with cost effectiveness 53 including process, outcome and economic evaluation involving 31 pharmacies in 54 55 three UK regions (15 in London (South and Central), 12 in Lothian (Edinburgh and 56 region) and four in Tayside (Dundee and region). 57 58 Patient and Public Involvement 59 60 The members of the Patient and Public Involvement (PPI) Group at the SRH service in Edinburgh contributed to the design of this study. The study protocol and 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 25

1 2 3 documentation were reviewed and approved by the chair and members of the PPI 4 5 group. The plain English summary was edited by a PPI member and improved as a 6 result. There are three PPI members that participate and contribute to the Bridge-It 7 Trial steering Committee meeting (TSC) that provides oversight of the study. PPI 8 group will be involved in the dissemination of the study results. 9 10 11 Intervention 12 The planned intervention is a composite intervention. Each woman in the 13 intervention phase will receive three months of POP (in a single package covering 3 14 months) and an invitation to attend a local participating SRH service to discuss and 15 16 obtain effective contraception (including LARC methods). Three packets of POP, (75 17 mcg desogestrel; UK) containing 28 tablets each will be provided at no cost to 18 women as a bridgingFor method peer of contraception, review providing only them with three months of 19 contraception during which time they can get an appointment with a contraceptive 20 provider to obtain their preferred method of contraception. Locally approved Patient 21 22 Group Directions (i.e. strict criteria to permit provision of specified medicines by 23 non-prescribers) will permit participating pharmacists to dispense the POP to women 24 recruited to the study. Pre-study training will be undertaken with participating 25 pharmacists including identifying medical contraindications to POP, potential drug 26 27 interactions medications and ‘missed pill’ guidance for POP. Pharmacists will advise 28 women to start the POP the day following intake of EC [10] and provide women with 29 a patient information booklet on the POP from the family planning association 30 (www.fpa.org). 31 Pharmacists will encourage women to attend the participating local SRH service to 32 33 obtain the contraceptive method of their choice. Participants (intervention phase) 34 will be given a study card to alert staff at SRH services that they are in the Bridge-It 35 study and should be seen as a drop in for contraception that same day. This card will 36 also provide written information about the location and opening hours of the local 37 38 participating SRH service. These SRH services are within a 5-mile radius of the 39 participating pharmacies and provide all methods of contraception at no cost as is 40 the norm in the NHS. 41 42 Standard Care 43 44 A mystery shopper exercise [13] will be undertaken in all 31 participating pharmacies 45 to characterise ‘standard care’ (usually verbal advice to visit a clinic for 46 contraception, with/without written and verbal information) in the control phase. 47 The mystery shopper visits will be conducted when the pharmacy is not recruiting 48 49 and just before the control phase starts. The mystery shoppers and the scenario 50 used will be chosen by the Patient Public Involvement group. A simple scenario 51 relating to request for EC will be used. Immediately after leaving the pharmacy the 52 mystery shopper will complete a standard data collection proforma, recording any 53 information given by the pharmacist about use of contraception after taking EC, 54 55 including provision of written information. 56 57 Participants 58 We will recruit a total of 626 to 737 women presenting for EC. The final number will 59 60 be determined based on the observed ratio of the between-period within-cluster correlation (BPC) and the within-period within-cluster correlation (WPC) – with the 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 25 BMJ Open

1 2 3 larger sample size (near the 737 upper limit) required for values of BPC close to zero, 4 5 and the smaller sample size (near the 626 lower limits) required for values of BPC 6 close to WPC (18). Each pharmacy will be expected to recruit an average of 12-13 7 women to the intervention arm and 12-13 women to standard care. This allows for a 8 25% loss to follow up at four months (missing data on primary outcome). 9 10 11 Randomisation 12 Each pharmacy will be randomised to either the intervention phase for 13 approximately 20 weeks followed by standard care phase for 20 weeks or vice-versa 14 with a wash-out period of two weeks between the two phases. The order in which 15 16 pharmacy allocation to each arm is undertaken (intervention or control first) will be 17 randomised (Figure 1).The order of delivery of intervention or control for each 18 pharmacy was randomisedFor peer for this cluster review crossover design only from a randomisation file 19 prepared by the study statistician in the Data Centre at the Centre for Healthcare 20 Randomised Trials (CHaRT), University of Aberdeen, using SAS 6.4 for Windows. The 21 22 method used for generating the random unpredictable mix of permuted blocks was 23 a computer software algorithm that randomly allocated blocks of size 2, 4 and 6; 24 blocking was used to ensure balanced group sizes. 25 This is a cluster cohort crossover design so it is the pharmacy that is the unit of 26 27 randomisation and the 'crossover' means that we are just randomising the order that 28 each pharmacy gives the intervention in. The ‘cohort’ label means that we expect 29 different women to be recruited within each site in the two periods (intervention 30 and control phases). 31 32 33 Recruitment 34 The pharmacist will assess medical eligibility of women presenting for EC for the 35 study, provide EC according to normal practice and invite eligible women to 36 participate. A detailed Patient Information Sheet (PIS) will be provided to all women 37 38 and informed consent will be obtained by participating community pharmacists. The 39 EC used in this study is levonorgestrel and will be given in the clinically indicated 40 dose for the woman’s weight (1.5 mg or 3 mg levonorgestrel) [10]. 41 Inclusion and exclusion criteria are shown in Table 1. Women who give written 42 consent will be recruited in the study. We recognise the importance of participant 43 44 retention and will offer a voucher of £10 at recruitment [19]. 45 46 Outcomes 47 A full list of study outcome measures is included in Table 2. Outcomes at four 48 49 months will primarily be collected via telephone interviews or via web-based 50 questionnaires. However, participants will also have the option to provide the same 51 information by a postal questionnaire. The primary outcome is use of effective 52 contraception at four months (intervention vs standard care). 53 Secondary outcomes are proportion of participants having an abortion within 12 54 55 months of EC use using record linkage from participants to national registries and 56 cost effectiveness. 57 58 Process evaluation measures 59 60 A process evaluation will be conducted as part of the study to assess potential issues concerning intervention implementation, the causal mechanisms of impact, and the 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 25

1 2 3 contextual factors that could affect these. The process evaluation will comprise of 4 5 quantitative and qualitative data measures, as detailed in Table 3. 6 7 8 Data Collection 9 Quantitative data 10 Participant flow: Participant flow through the study will be assessed and reported 11 12 following the CONSORT flow chart. 13 14 Baseline 15 Participant demographics and reproductive history is collected at recruitment by a 16 self-administered paper questionnaire given to them by the pharmacist. 17 18 Demographic dataFor will also peer be reported review for the process only evaluation, protocol 19 adherence checklists and for recruitment screening forms (see Table 3). 20 21 Contraceptive use at 4 months 22 23 This will be based upon self-reported data from women at a telephone follow up 24 interview with a research nurse at four months after obtaining EC. If participants 25 prefer, the questions can be self-completed by a web-based questionnaire or paper 26 questionnaire sent by post. Women will be asked what method of contraception 27 they are using (if any), if they attended a GP or SRH service for this, if they used the 28 29 POP (intervention phase only) and their pregnancy status. If pregnancy has occurred 30 since EC then the validated London Measure of Unintended Pregnancy tool will be 31 administered to measure intended-ness of the index pregnancy [20] (Table 2). 32 33 34 Abortion rates at 12 months 35 Information Services Division (ISD Scotland) and Department of Health (DOH 36 England) will provide the number of abortions occurring during the 12 month follow 37 up period in each arm by conducting linkage of the identifiers (collected at baseline) 38 from study participants. 39 40 41 Validation 42 We will check with data from local SRH services to determine the numbers of 43 participants from intervention and control phases who attend the local participating 44 SRH service, and which method of contraception they received. 45 46 47 Qualitative data 48 Semi-structured, qualitative telephone interviews of a purposive sample of up to 60 49 women who received the intervention (approximately 22 in London, 30 in Edinburgh 50 51 and 8 in Dundee) will be undertaken. Participants who consent to these telephone 52 interviews will be contacted by the Process Evaluation Research Assistant. Interviews 53 will explore experience of intervention acceptability in more depth and assess 54 experiences of bridging from EC to regular contraception, and reasons for doing so 55 or not (Table 3). Interviews will be conducted soon after the four month follow up. 56 57 58 Interviews with 27 pharmacists and 12 SRH service providers will explore their 59 perceptions of barriers and facilitators to implementation and more broadly, their 60 views on the intervention, the trial, and the target population. Interviews will be

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 25 BMJ Open

1 2 3 conducted by the process evaluation research assistant soon after the intervention 4 5 phase has completed. 6 7 For the process evaluation, data collection also includes: review of training and 8 materials; observation of training; mapping of local contraceptive services within 10 9 miles of study sites and monitoring of contemporaneous events, such as relevant 10 11 high coverage media stories using Google Alerts (Table 3). 12 13 Sample size calculation 14 The study is a pragmatic cluster randomised cohort crossover trial. Ideally the 15 16 control and intervention phases should be of roughly equal duration and size, and 17 the sample size is calculated assuming an equal cluster size in both control and 18 intervention periods,For and peer equal across review sites (the expected only site/period average). In 19 practice, there is variability in EC demand across sites and over time: for example, 20 demand is affected by peak holiday periods (recruitment decreases) and student 21 22 term times (recruitment increases); and the ability of a pharmacy to translate that 23 variable demand into study recruits depends on many factors, including changes to 24 circumstances at individual pharmacies or loss of / change of pharmacists at multi- 25 pharmacist stores. 26 27 Informed by our pilot study, we have assumed that effective contraception use in 28 the control would be 30% and we were likely to achieve a 50% relative improvement 29 to 45%. This means the sample size is in the range 626 to 737, assuming 25% of 30 women do not provide four-month contraceptive use data, and an average cluster 31 size of 12-13 in each period, and around 25 pharmacies taking part, with 90% power 32 33 and a 5% level of significance. The uncertainty in the required sample size rests on 34 the assumed between-period within-cluster correlation (BPC) and its relationship to 35 the other component of variability, the within-period within-cluster correlation 36 (WPC)[18]. The WPC is the usual correlation (known as the Intra Class Correlation– 37 38 ICC in a standard parallel groups non-crossover cluster setting) of two individuals’ 39 outcomes within a cluster (in the same period). The BPC on the other hand is the 40 correlation between two individuals’ outcomes in the same cluster between the two 41 periods. If the BPC is zero, there is no advantage in a crossover design over the 42 standard parallel groups cluster design; if the BPC equals the WPC then the crossover 43 44 is as efficient as an individually randomised design. We will finalise the sample size 45 depending on the observed ratio of the BPC to WPC, and the observed rate of 46 attrition, but still assuming the same control rate and treatment effect, once we 47 have 4 month data on at least 500 participants. 48 49 50 Quantitative Analyses 51 There will be a single analysis at study end (there is no opportunity for any interim 52 analyses given the crossover design) although an independent Data Monitoring 53 Committee will monitor study progress and any safety issues. This will follow the 54 55 intention to treat principle and will use a hierarchical model appropriate for the 56 specific outcome. For the primary outcome this will be a mixed effects logistic 57 regression, using the hierarchical model approach as recommended by Turner for a 58 cluster crossover design [21]. We will pre-specify any individual level (or cluster 59 60 level) covariates that we intend to adjust for, and the comprehensive Statistical Analysis Plan will specify the sensitivity type analyses that will explore how robust 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 25

1 2 3 the findings are to any missing data at the cluster level (probably unlikely) and the 4 5 individual level (expected to be substantial for the patient reported outcomes at four 6 months). As well as the usual assumption of missing at random, we will try to explore 7 possible mechanisms for non-ignorable (informative missingness) at the individual 8 level which may well be operating in this context. Subgroup analyses will explore the 9 possible effect modification by LARC (most effective contraceptive methods) vs non- 10 11 LARC vs no use of contraception. 12 13 Qualitative and mixed-methods process evaluation analyses 14 All process data will be analysed independently of the outcome data and, 15 16 importantly, documented before the outcomes are known. Qualitative analysis of in 17 -depth interviews will be recorded and transcribed verbatim. Transcription and 18 analysis (proceedingFor case peer by case) will review start with the first only interview and be ongoing 19 during the course of data collection, allowing for emergent themes to be identified 20 and explored in future interviews. The transcripts will be read repeatedly and coded 21 22 for analysis. Data management will be assisted by the software, QSR NVivo 10. 23 Analysis will be undertaken using ‘Framework Analysis’ a method of proven validity 24 and reliability where data are coded, indexed and charted systematically, then 25 organised using a matrix or framework [22]. Constant comparison will be carried out 26 27 to ensure that the analysis represents all perspectives and negative (‘deviant’) cases. 28 The multi-source process evaluation will be synthesised to address the three key 29 process evaluation questions: i) what was delivered, ii) how it was delivered, and iii) 30 what role context may have had in shaping the delivery/outcomes 31 32 33 Economic evaluation: An economic evaluation will be undertaken comparing the 34 intervention and control arms in a cost effectiveness analysis. A trial-based analysis 35 will be followed by the construction of a decision model to extrapolate future costs 36 and benefits beyond the completion of the trial. The overall perspective used will be 37 38 that of the health system. Costs will include the pharmacist training to provide POP, 39 direct and indirect costs of health service use, and the provision and dispensing of 40 POP. We will compare the costs to the NHS in the intervention and control arms. To 41 account for differences in the numbers of women in the two arms, we will compare 42 the cost per woman in each arm. In the control arm, the costs are (i) cost of EC, (ii) 43 44 cost of pharmacist provision of EC, (iii) cost of abortions. In the intervention arm, the 45 costs are in addition to these (iv) the cost of the POP, (v) cost of pharmacist training 46 to provide POP and (vi) cost of pharmacist provision of POP. The costs (i) and (ii) are 47 the same in both groups and so the extra cost of the intervention will be the sum of 48 49 (iv), (v) and (vi). The cost per women who has an abortion is the same in both groups 50 except that we hypothesise that the abortion rate will be lower in the intervention 51 group. We can then state the outcome as conventional incremental cost 52 effectiveness ratio i.e. for every £100 spent on the intervention results in x fewer 53 abortions for a savings of £Y. If Y is greater than 100 then the intervention is cost 54 55 effective. We will examine the sensitivity of the outcomes to variations in the costs 56 of iv, v, and vi. 57 58 Data Management and Clinical Trial Unit support 59 60 Data will be collected on a paper case report form and will be entered directly into the trial database. Data will be entered into a trial database by pharmacists, research 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 25 BMJ Open

1 2 3 nurses or staff at the trial co-ordinating centre. The data management and statistical 4 5 support (including responsibility of data and final dataset) for the study is provided 6 by the UKCRC registered Clinical Trial Unit (CTU) the Centre for Healthcare 7 Randomised Trials (CHaRT) at the University of Aberdeen while the trial 8 management is provided by another UKCRC registered CTU, the Edinburgh Clinical 9 Trials Unit (ECTU) at Edinburgh University. 10 11 12 Ethics and dissemination 13 The Bridge-It trial involves procedures and medications which are well established in 14 current NHS clinical practice and use. Adverse events may occur during or after the 15 16 use of EC or POP and are well documented in the POP patient information leaflet. 17 Serious adverse events will be recorded at the four month follow up interview and 18 reported to theFor study sponsor. peer The studyreview will be conducted only in accordance with the 19 principles of Good Clinical Practice (GCP). 20 A favorable ethical opinion has been obtained from the South East Scotland REC in 21 22 June 2017. Approvals have been obtained from NHS Research Scotland (NRS) and 23 Health Research Authority (HRA) England prior to commencement of the study. 24 Annual progress reports and a final report at the conclusion of the trial will be 25 submitted to REC within the timelines defined in the regulations. Protocol 26 27 modifications are communicated by the ECTU to study sites via email and electronic 28 newsletters. 29 The Bridge-It study website will include trial materials, trial progress, and summaries 30 of key findings. In addition, public engagement and dissemination will also be 31 undertaken via our Patient Public Involvement group. 32 33 The results of the study will be published in the academic journals and all 34 participants will be offered a lay summary of the main findings of the study. The 35 findings will also be presented at national and international conferences and 36 disseminated via social media. 37 38 39 Trial Status 40 As at 7th Feb 2019 the study had recruited 503 participants across 29 sites. 41 Recruitment to the first period completed on 13th Jan 2019, with 391 participants 42 recruited at 29 sites. Recruitment is scheduled to be completed by Jun 2019, with 43 44 analyses of the 4-month primary outcome expected to be available by Oct 2019. 45 46 DISCUSSION 47 Unintended pregnancy is widely perceived as a major public health problem. The 48 49 proposed intervention in the Bridge-It study provides both temporary contraception 50 (the POP) and facilitates access to effective contraception at a local SRH clinic. The 51 cluster design was felt necessary for logistical reasons and confirmed in the 52 qualitative work of our pilot study [17, 23] that an individually randomised trial 53 would simply not recruit, as it was not feasible for pharmacists within a busy 54 55 pharmacy to take additional time to randomise each individual. The crossover nature 56 of the cluster design was chosen for efficiency, and by having a different set of 57 women recruited at a pharmacy in the two different periods we avoided 58 contamination by the participant. The purpose of the washout out period (was to 59 60 minimise intervention effect carrying over from one period to another, as part of any contamination effect mediated by the pharmacist. With the cluster crossover design, 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 25

1 2 3 each cluster will act as its own control and fewer pharmacies are required than with 4 5 a parallel cluster design. 6 If our proposed intervention works, then this could prevent more unintended 7 pregnancies for more women. If the intervention is cost effective then it could have 8 cost savings for the NHS. 9 10 11 12 REFERENCES. 13 1. Sedgh G, Singh S, Shah IH et al. Induced abortion: incidence and trends worldwide 14 from 1995 to 2008. Lancet. 2012Feb 18;379(9816):625-32. 15 16 17 2.Department of Health. Abortion Statistics, England and Wales: 2017. 18 https://www.gov.uk/government/publications/report-on-abortion-statistics-in-For peer review only 19 england-and-wales-for-2014 (access date 11 Jan 2019) 20 21 22 3. Information Services Division. Termination of Pregnancy Statistics 2017 23 http://www.isdscotland.org/Health-Topics/Sexual-Health/Abortions/ (access date 11 24 Jan 2019) 25 26 27 4. Wellings K, Jones KG, Mercer CH et al. The prevalence of unplanned pregnancy 28 and associated factors in Britain: findings from the third National Survey of Sexual 29 Attitudes and Lifestyles (Natsal-3). Lancet 2013;382:1807-16 30 31 5. Thomas C, Cameron ST. Can we reduce costs and prevent more unintended 32 33 pregnancies? A cost of illness study illustrating the cost-saving potential of a more 34 effective method of EHC. BMJ open 2013;3:e003815.doi:10.1136 35 36 6. SmithBattle L, Leonard V. Inequities compounded: explaining variations in the 37 38 transition to adulthood for teen mothers' offspring. J Fam Nurs. 2012;18(3):409-31. 39 40 7. Cameron ST, Li H R, Gemzell –Danielsson K. Current controversies with Emergency 41 Contraception: a review. BJOG 2017(13):1948-1956. 42 43 8. Marston C, Meltzer H, Majeed A. Impact on contraceptive practice of making 44 45 emergency hormonal contraception available over the counter in Great Britain: 46 repeated cross sectional surveys.BMJ. 2005Jul 30;331(7511):271. 47 48 49 9. Polis CB, Grimes DA, Schaffer K, et al . Advance provision of emergency 50 contraception for pregnancy prevention. Cochrane Database of Systematic Reviews 51 2007,Issue2.Art. No: CD005497. DOI:0.1002/14651858.CD005497.pub2. 52 53 10. Faculty of Sexual and Reproductive Healthcare, UK. Emergency contraception. 54 55 2017. www.fsrh.org (access date 11 January 2019) 56 57 11. Cheng L, Che Y, Gülmezoglu AM. Interventions for emergency contraception. 58 Cochrane Database of Systematic Reviews 2012;Issue 8. Art. No.: CD001324. DOI: 59 60 10.1002/14651858.CD001324.pub4. 12. Trussell J. Contraceptive efficacy. Global Library of Womens 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 25 BMJ Open

1 2 3 Medicinehttp://www.glowm.com/section_view/heading/Contraceptive%20Efficacy/i 4 5 tem/374 ( access date 29 Jan 2019) 6 7 13 .Glasier A, Manners RJ, Loudon JC, et al. Community pharmacists providing 8 emergency contraception give little advice about future contraceptive use: a mystery 9 shopper study. Contraception 2010;82: 538–42. 10 11 12 13 14. The Faculty of Sexual and Reproductive Healthcare (FSRH). UK medical Eligibility 14 criteria for Contraceptive Use. http://www.fsrh.org/standards-and-guidance/uk- 15 medical-eligibility-criteria-for-contraceptive-use/ (access date 11 Jan 2019) 16 17 18 15. Brache V, CochonFor L, peerDuijkers IJ, etreview al. A prospective, only randomized, 19 pharmacodynamic study of quick-starting a desogestrel progestin-only pill following 20 ulipristal acetate for emergency contraception. Hum Reprod 2015;30(12):2785-2793 21 22 23 16. Cameron ST, Berger C, Michie L,et al. The effects on ovarian activity of 24 ‘quickstarting’ a combined oral contraceptive pill after ulipristal acetate : 25 prospective, randomised, double-blind parallel-arm, placebo-controlled study. Hum 26 Reprod 2015;30(7):1566-1572. 27 28 29 17. Michie L, Cameron ST, Glasier A et al .Pharmacy based interventions for initiating 30 effective contraception following the use of emergency contraception: a pilot study. 31 Contraception 2014 ;90, 447-53 32 33 34 18. Arnup SJ, McKenzie JE, Hemming K, et al. Understanding the cluster crossover 35 design: a graphical illustration of the components of variation and a sample size 36 tutorial. Trials. 2017;18; 381. 37 38 39 19. Brueton VC, Tierney J, Stenning S, et al. Strategies to improve retention in 40 randomised trials. Cochrane Database Syst Rev. 2013 Dec 3; 12:MR000032. doi: 41 10.1002/14651858.MR000032.pub2. 21 42 43 20. Cameron ST ,Glasier A. Identifying women in need of further discussion about the 44 45 decision to have an abortion and eventual outcome. Contraception. 2013;88(1): 128- 46 32 47 48 21. Turner RM, Omar RZ, Thompson SG. Modelling multivariate outcomes in 49 50 hierarchical data, with application to cluster randomised trials. Biom J. 2006 51 Jun;48(3):333-45. 52 53 22. Ritchie J, Spencer L. Qualitative research practice: a guide for social science 54 students and researchers. Sage: London, 2003. 55 56 57 23. Michie L, Cameron ST, Glasier A et al. Providing bridging contraception after 58 emergency contraception from the pharmacy: a qualitative study. Public Health 59 Public Health 2016; 135: 97-103. 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 25

1 2 3 Authors’ contributions: STC, AG, JN, LMc D, AR, PB, JS, JT developed the original 4 5 protocol. SC, CB, MF, RG, AMcD, BG, AJ, AM, SP,DS,NS, KC contributed to later stages 6 of study design. STC wrote the draft of the protocol with significant input from all 7 authors at all stages. All authors contributed, read and approved the final 8 manuscript. 9 10 11 12 Author affiliations: 13 14 1. Chalmers Centre, NHS Lothian, Edinburgh EH3 9ES and University of 15 Edinburgh 16 17 2. King’s College Hospital NHS Foundation Trust, London, SE5 9RS 18 3. MRC/CSOFor Social &peer Public Health review Sciences Unit, onlyUniversity of Glasgow, 19 Glasgow G2 3AX 20 4. Edinburgh Clinical Trial Unit, Usher Institute, University of Edinburgh, 21 22 Edinburgh EH16 4UX 23 5. King’s Cross Hospital, NHS Tayside, Dundee DD3 8EA 24 6. University College London, EGA Institute for Women’s Health, London WC1E 25 6AU 26 7. †Deceased December 2018. Princeton University USA. 27 28 8. Health Service Research Unit, University of Aberdeen, Aberdeen, AB25 2ZD 29 9. Boots UK, North Hub Office, Falkirk FK1 1ES 30 10. University College London, Institute for Global Health, London WC1E 6JB 31 11. King’s College London, SE5 9RJ 32 33 34 Acknowledgements: Thank you to the members of the Patient and Public 35 Involvement Group at the SRH service in Edinburgh who contributed to the design of 36 this study. Thanks to the women who have participated in this study and to the 37 community pharmacists, research staff and health care professionals at the local SRH 38 39 services who have assisted with the implementation of this study. Thanks to Laura 40 Flett for trial management support. 41 42 The Bridge-It Study Steering Committee provides oversight for this trial on behalf of 43 the sponsor (University of Edinburgh and NHS Lothian) and the funder. The 44 45 members are: Professor Peter Brocklehurst (Chair), Birmingham Clinical Trials Unit; 46 Dr Lucy Michie, Sandyford Glasgow; Professor Kaye Wellings, London School of 47 Hygiene and Tropical Medicine; Joanna Loudon, PPI member, Edinburgh; Kirsten 48 Stuart PPI member Edinburgh; Emily Whittaker PPI member Edinburgh. 49 50 51 The Data Monitoring Committee is an independent multidisciplinary group consisting 52 of clinicians and statisticians. The members are: Professor Claire Anderson (Chair), 53 University of Nottingham; Professor Elizabeth Allen, London School of Hygiene and 54 Tropical Medicine; Professor Caroline Moreau, Johns Hopkins Bloomberg School of 55 56 Public Health, USA. A copy of the DMC charter is held in Edinburgh Clinical Trials 57 Unit. 58 59 60

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 25 BMJ Open

1 2 3 The study has co-sponsorship between The University Court of the University of 4 5 Edinburgh and Lothian Health Board. The sponsors representative is 6 [email protected] 7 8 The paper presents independent research funded by the National Institute for Health 9 Research (NIHR). The views expressed are those of the authors and not necessarily 10 11 those of the NHS, the NIHR or the Department of Health. 12 13 Funding statement: The Bridge-It study is funded by the National Institute for Health 14 Research’s Health Technology Assessment Programme. HTA Project:15/113/01 15 16 17 Competing interests statement. 18 For peer review only 19 AG is a member of HRA Pharma scientific advisory board. 20 21 22 PB is a Clinical Director of the not-for profit community interest company SH:24 that 23 provides online sexual health services in partnership with the NHS. 24 25 AR Research grants, educational grants and consultancy with Gilead 26 Research grants from Roche and BMS 27 28 Educational grants from Abbvie 29 30 JN is Deputy Chair of the NIHR/HTA General Board Committee. NIHR/HTA funded 31 this research. 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 25

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 25 BMJ Open

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 25

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 25 BMJ Open

1 2 3 4 5 Figure Legend 6 7 Figure 1. Bridge-it flowchart 8 9 POP= progestogen only pill 10 11 IDI= in depth interviews 12 SRH= sexual and reproductive health 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 25

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 25 BMJ Open

1 2 3 4 5 6 7 8 SPIRIT 2013 Checklist: Recommended items to address in a clinical trial protocol and related documents* 9 10 Section/item Item Description: Bridge it study Addressed on 11 No page number 12 For peer review only 13 14 Administrative information 15 16 Title 1 Descriptive title identifying the study design, population, interventions, and, if applicable, trial acronym ______1____ 17 18 Trial registration 2a Trial identifier and registry name. If not yet registered, name of intended registry ______1___ 19 20 2b All items from the World Health Organization Trial Registration Data Set ______N/A_ 21 22 Protocol version 3 Date and version identifier _____1______23 24 Funding 4 Sources and types of financial, material, and other support ______1 25 26 Roles and 5a Names, affiliations, and roles of protocol contributors ______1,13____ 27 responsibilities __ 28 29 5b Name and contact information for the trial sponsor ______13______30 31 5c Role of study sponsor and funders, if any, in study design; collection, management, analysis, and 32 interpretation of data; writing of the report; and the decision to submit the report for publication, including ______13___ 33 whether they will have ultimate authority over any of these activities 34 35 5d Composition, roles, and responsibilities of the coordinating centre, steering committee, endpoint ______13_ 36 adjudication committee, data management team, and other individuals or groups overseeing the trial, if 37 38 applicable (see Item 21a for data monitoring committee) 39 40 41 42 1 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 22 of 25

1 Introduction 2 3 Background and 6a Description of research question and justification for undertaking the trial, including summary of relevant ______3, 4 rationale studies (published and unpublished) examining benefits and harms for each intervention 4 5 6 6b Explanation for choice of comparators _3______7 8 Objectives 7 Specific objectives or hypotheses __4______9 10 Trial design 8 Description of trial design including type of trial (eg, parallel group, crossover, factorial, single group), 2,4 11 12 allocation ratio,For and framework peer (eg, superiority, review equivalence, noninferiority, only exploratory) ______13 14 Methods: Participants, interventions, and outcomes 15 16 Study setting 9 Description of study settings (eg, community clinic, academic hospital) and list of countries where data will _4______17 18 be collected. Reference to where list of study sites can be obtained 19 Eligibility criteria 10 Inclusion and exclusion criteria for participants. If applicable, eligibility criteria for study centres and Table1______20 21 individuals who will perform the interventions (eg, surgeons, psychotherapists) __ 22 23 Interventions 11a Interventions for each group with sufficient detail to allow replication, including how and when they will be ___5______24 administered 25 26 11b Criteria for discontinuing or modifying allocated interventions for a given trial participant (eg, drug dose __N/A______27 change in response to harms, participant request, or improving/worsening disease) _ 28 29 11c Strategies to improve adherence to intervention protocols, and any procedures for monitoring adherence ______6_____ 30 (eg, drug tablet return, laboratory tests) 31 32 11d Relevant concomitant care and interventions that are permitted or prohibited during the trial ______N/A___ 33 34 _ 35 36 Outcomes 12 Primary, secondary, and other outcomes, including the specific measurement variable (eg, systolic blood 37 pressure), analysis metric (eg, change from baseline, final value, time to event), method of aggregation (eg, ___6______38 median, proportion), and time point for each outcome. Explanation of the clinical relevance of chosen 39 efficacy and harm outcomes is strongly recommended 40 41 42 2 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 23 of 25 BMJ Open

1 Participant timeline 13 Time schedule of enrolment, interventions (including any run-ins and washouts), assessments, and visits for _____Fig1______2 participants. A schematic diagram is highly recommended (see Figure) __ 3 4 Sample size 14 Estimated number of participants needed to achieve study objectives and how it was determined, including ______8____ 5 clinical and statistical assumptions supporting any sample size calculations 6 7 Recruitment 15 Strategies for achieving adequate participant enrolment to reach target sample size ______6____ 8 9 10 Methods: Assignment of interventions (for controlled trials) 11 12 Allocation: For peer review only 13 14 Sequence 16a Method of generating the allocation sequence (eg, computer-generated random numbers), and list of any ____6______15 generation factors for stratification. To reduce predictability of a random sequence, details of any planned restriction 16 (eg, blocking) should be provided in a separate document that is unavailable to those who enrol participants 17 or assign interventions 18 19 Allocation 16b Mechanism of implementing the allocation sequence (eg, central telephone; sequentially numbered, ______6____ 20 21 concealment opaque, sealed envelopes), describing any steps to conceal the sequence until interventions are assigned 22 mechanism 23 24 Implementation 16c Who will generate the allocation sequence, who will enrol participants, and who will assign participants to ____6 25 interventions _N/A as cluster 26 randomised_____ 27 28 ___ 29 30 Blinding (masking) 17a Who will be blinded after assignment to interventions (eg, trial participants, care providers, outcome __N/A______31 assessors, data analysts), and how _ 32 33 17b If blinded, circumstances under which unblinding is permissible, and procedure for revealing a participant’s _____N/A______34 allocated intervention during the trial _ 35 36 37 Methods: Data collection, management, and analysis 38 39 40 41 42 3 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 24 of 25

1 Data collection 18a Plans for assessment and collection of outcome, baseline, and other trial data, including any related ______7,9____ 2 methods processes to promote data quality (eg, duplicate measurements, training of assessors) and a description of _ 3 study instruments (eg, questionnaires, laboratory tests) along with their reliability and validity, if known. 4 Reference to where data collection forms can be found, if not in the protocol 5 6 18b Plans to promote participant retention and complete follow-up, including list of any outcome data to be ______6__ 7 8 collected for participants who discontinue or deviate from intervention protocols 9 10 Data management 19 Plans for data entry, coding, security, and storage, including any related processes to promote data quality ______9_ 11 (eg, double data entry; range checks for data values). Reference to where details of data management 12 procedures canFor be found, peer if not in the protocol review only 13 14 Statistical methods 20a Statistical methods for analysing primary and secondary outcomes. Reference to where other details of the ______8,9 15 statistical analysis plan can be found, if not in the protocol 16 17 20b Methods for any additional analyses (eg, subgroup and adjusted analyses) _8,9______18 19 _ 20 20c Definition of analysis population relating to protocol non-adherence (eg, as randomised analysis), and any 21 22 statistical methods to handle missing data (eg, multiple imputation) _8______23 24 Methods: Monitoring 25 26 Data monitoring 21a Composition of data monitoring committee (DMC); summary of its role and reporting structure; statement of _13______27 28 whether it is independent from the sponsor and competing interests; and reference to where further details 29 about its charter can be found, if not in the protocol. Alternatively, an explanation of why a DMC is not 30 needed 31 32 21b Description of any interim analyses and stopping guidelines, including who will have access to these interim ____N/A______33 results and make the final decision to terminate the trial _ 34 35 Harms 22 Plans for collecting, assessing, reporting, and managing solicited and spontaneously reported adverse ______10____ 36 37 events and other unintended effects of trial interventions or trial conduct 38 39 Auditing 23 Frequency and procedures for auditing trial conduct, if any, and whether the process will be independent __N/A______40 from investigators and the sponsor _ 41 42 4 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 25 of 25 BMJ Open

1 Ethics and dissemination 2 3 Research ethics 24 Plans for seeking research ethics committee/institutional review board (REC/IRB) approval ______2______4 approval 5 6 Protocol 25 Plans for communicating important protocol modifications (eg, changes to eligibility criteria, outcomes, ______10_____ 7 8 amendments analyses) to relevant parties (eg, investigators, REC/IRBs, trial participants, trial registries, journals, 9 regulators) 10 11 Consent or assent 26a Who will obtain informed consent or assent from potential trial participants or authorised surrogates, and ______6____ 12 how (see ItemFor 32) peer review only 13 14 26b Additional consent provisions for collection and use of participant data and biological specimens in ancillary ___N/A______15 studies, if applicable _ 16 17 Confidentiality 27 How personal information about potential and enrolled participants will be collected, shared, and maintained _____This is in 18 19 in order to protect confidentiality before, during, and after the trial ethical application- 20 on bridge it 21 website _____ 22 23 Declaration of 28 Financial and other competing interests for principal investigators for the overall trial and each study site 14 24 interests Also BMJ open 25 IJMEC 26 27 forms______28 29 Access to data 29 Statement of who will have access to the final trial dataset, and disclosure of contractual agreements that __page 9 30 limit such access for investigators 31 32 Ancillary and post- 30 Provisions, if any, for ancillary and post-trial care, and for compensation to those who suffer harm from trial __ PIS, and 33 trial care participation consent__uploade 34 d as 35 36 supplementary___ 37 ______38 39 Dissemination policy 31a Plans for investigators and sponsor to communicate trial results to participants, healthcare professionals, __10______40 the public, and other relevant groups (eg, via publication, reporting in results databases, or other data 41 sharing arrangements), including any publication restrictions 42 5 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 26 of 25

1 31b Authorship eligibility guidelines and any intended use of professional writers ___12______2 3 31c Plans, if any, for granting public access to the full protocol, participant-level dataset, and statistical code ______N/A____ 4 _ 5 6 Appendices 7 8 Informed consent 32 Model consent form and other related documentation given to participants and authorised surrogates __uploaded as 9 10 materials supplementary 11 files 12 For peer review only Participant consent 13 form 14 15 Participant information 16 sheet 17 Pharmacist/SRH 18 provider information 19 20 sheet 21 Pharmacist/SRH 22 provided consent form 23 GDPR participant 24 25 Information sheet 26 27 28 Biological 33 Plans for collection, laboratory evaluation, and storage of biological specimens for genetic or molecular N/a 29 specimens analysis in the current trial and for future use in ancillary studies, if applicable 30 31 *It is strongly recommended that this checklist be read in conjunction with the SPIRIT 2013 Explanation & Elaboration for important clarification on the items. 32 Amendments to the protocol should be tracked and dated. The SPIRIT checklist is copyrighted by the SPIRIT Group under the Creative Commons 33 34 “Attribution-NonCommercial-NoDerivs 3.0 Unported” license. 35 36 37 38 39 40 41 42 6 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open

Manuscript ID bmjopen-2019-029978.R2

Article Type: Protocol

Date Submitted by the 06-Sep-2019 Author:

Primary Subject Sexual health Heading:

Secondary Subject Heading: Public health

contraception, emergency contraception, pharmacy, unintended Keywords: pregnancy, levonorgestrel, progestogen only pill

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 25 BMJ Open

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 25 BMJ Open

1 2 3 ABSTRACT: 4 5 Introduction: Oral emergency contraception (EC) can prevent unintended pregnancy 6 but it is important to start a regular method of contraception. Women in the UK 7 usually access EC from a pharmacy but then need a subsequent appointment with a 8 GP or a sexual and reproductive health (SRH) service to access regular contraception. 9 Unintended pregnancies can occur during this time. 10 11 Methods and analysis: Bridge-It is a pragmatic cluster randomised cohort crossover 12 trial designed to determine whether pharmacist provision of a bridging supply of a 13 progestogen only pill (POP) plus rapid access to a local SRH clinic, results in increased 14 uptake of effective contraception and prevents more unintended pregnancies than 15 16 provision of EC alone. Bridge-It involves 31 pharmacies in three UK regions (London, 17 Lothian and Tayside) aiming to recruit 626 to 737 women. Pharmacies will give EC 18 (levonorgestrel)For according peer to normal reviewpractice and recruit only women to both 19 intervention and the control phases of the study. In the intervention phase, 20 pharmacists will provide the POP (desogestrel) and offer rapid access to a SRH clinic. 21 22 In the control phase, pharmacists will advise women to attend a contraceptive 23 provider for contraception (standard care). 24 Women will be asked four months later about contraceptive use. Data linkage to 25 abortion registries will provide abortion rates over 12 months. The sample size is 26 27 calculated on the primary outcome of effective contraception use at four months 28 (yes/no) with 90% power and a 5% level of significance. Abortion rates will be an 29 exploratory secondary analysis. Process evaluation includes interviews with 30 pharmacists, SRH clinicians and women. Cost-effectiveness analysis will use a 31 healthcare system perspective and be expressed as incremental cost-effectiveness 32 33 ratio. 34 Ethics and dissemination: Ethical approval was received from South East Scotland 35 REC June 2017. Results will be published in peer-reviewed journals and conference 36 presentations. 37 38 Trial registration number ISRCTN 70616901; pre results 39 40 41 42 Strengths and limitations: 43 44  Examines the important outcome of abortion rates over 1 year as an 45 exploratory secondary analysis. 46  Applicable only to women receiving levonogrestrel EC followed by a 47 desogestrel POP 48 49  Not applicable to use of ulipristal acetate for EC, since hormonal methods of 50 contraception such as the desogestrel POP may interact with efficacy of 51 ulipristal acetate, if started within five days 52 53 54 55 56 57 58 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 25

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 25 BMJ Open

1 2 3 larger sample size (near the 737 upper limit) required for values of BPC close to zero, 4 5 and the smaller sample size (near the 626 lower limits) required for values of BPC 6 close to WPC [18]. Each pharmacy will be expected to recruit an average of 12-13 7 women to the intervention arm and 12-13 women to standard care. This allows for a 8 25% loss to follow up at four months (missing data on primary outcome). 9 10 11 Randomisation 12 Each pharmacy will be randomised to either the intervention phase for 13 approximately 20 weeks followed by standard care phase for 20 weeks or vice-versa 14 with a wash-out period of two weeks between the two phases. The order in which 15 16 pharmacy allocation to each arm is undertaken (intervention or control first) will be 17 randomised (Figure 1).The order of delivery of intervention or control for each 18 pharmacy was randomisedFor peer for this cluster review crossover design only from a randomisation file 19 prepared by the study statistician in the Data Centre at the Centre for Healthcare 20 Randomised Trials (CHaRT), University of Aberdeen, using SAS 6.4 for Windows. The 21 22 method used for generating the random unpredictable mix of permuted blocks was 23 a computer software algorithm that randomly allocated blocks of size 2, 4 and 6; 24 blocking was used to ensure balanced group sizes. 25 This is a cluster cohort crossover design so it is the pharmacy that is the unit of 26 27 randomisation and the 'crossover' means that we are just randomising the order that 28 each pharmacy gives the intervention in. The ‘cohort’ label means that we expect 29 different women to be recruited within each site in the two periods (intervention 30 and control phases). 31 32 33 Recruitment 34 The pharmacist will assess medical eligibility of women presenting for EC for the 35 study, provide EC according to normal practice and invite eligible women to 36 participate. A detailed Patient Information Sheet (PIS) will be provided to all women 37 38 and informed consent will be obtained by participating community pharmacists. The 39 EC used in this study is levonorgestrel and will be given in the clinically indicated 40 dose for the woman’s weight (1.5 mg or 3 mg levonorgestrel) [10]. 41 Inclusion and exclusion criteria are shown in Table 1. Women who give written 42 consent will be recruited in the study. We recognise the importance of participant 43 44 retention and will offer a voucher of £10 at recruitment [19]. 45 46 Outcomes 47 A full list of study outcome measures is included in Table 2. Outcomes at four 48 49 months will primarily be collected via telephone interviews or via web-based 50 questionnaires. However, participants will also have the option to provide the same 51 information by a postal questionnaire. The primary outcome is use of effective 52 contraception at four months (intervention vs standard care). 53 Secondary outcomes are proportion of participants having an abortion within 12 54 55 months of EC use using record linkage from participants to national registries and 56 cost effectiveness. 57 58 Process evaluation measures 59 60 A process evaluation will be conducted as part of the study to assess potential issues concerning intervention implementation, the causal mechanisms of impact, and the 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 25