Maternal and Newborn Health Priority Setting Partnership in

Protocol version: 1.5 Protocol Date: 01 Dec 2019

Signed by (on behalf of the steering group) Protocol Chairs: Academic Chair: Andrew D Weeks, MD, MRCOG Professor of International Maternal Health University of Liverpool, UK

Steering Group Chair Toto Anne Gronlund, (TBA) James Lind Alliance University of Southampton, UK

2nd Academic Chair: Anthony Mbonye, MBChB, MA, MPH, PhD Professor of Public Health, Makerere University, Uganda

Principal investigator James Ditai, PhD research fellow University of Liverpool, UK Sanyu Africa Research Institute, Uganda

Study sponsor: University of Liverpool, UK Study coordinator: Sanyu Africa Research Institute, Uganda Study Funder: European and Developing countries Clinical Trials Partnership (EDCTP)

Study protocol version 1.5 protocol date 01 Dec 2019 1 Protocol Revision History Version Effective Description of change number Date Version 1.0 05 Nov 2018 Initial protocol Version 1.2 22 Mar 2019 Made a separate protocol from the main study Version 1.3 22 May 2019 Included changes from the REC, trial information Version 1.4 25 July 2019 The protocol was re-written structurally and submitted to the University of Liverpool ethics committee and sponsorship Version 1.5 01 Dec 2019 Amendment of the steering group composition, PSP timeline

List of abbreviations Abbreviations Definition BUFHS Faculty of Health Sciences WHO World Health Organisation HC Health Centre UNCST Uganda National Council for Science and Technology UK United Kingdom UCU Uganda Christian University GCC Grand Challenges Canada VHT Village Health Team member REC Research Ethics Committee JLA James Lind Alliance PSP Priority setting partnership PPI Patient and public involvement SAfRI Sanyu Africa Research Institute MNH-UG PSP Maternal and newborn health priority setting partnership in Uganda LC1 Local council one NGO Non-government organisation UTODA Uganda Taxi owners and Drivers association RA Research Assistant NIN National Identification Number

Study protocol version 1.5 protocol date 01 Dec 2019 2 1.0 GENERAL INFORMATION

The purpose of this protocol is to set out the aims, objectives and commitments of the Maternal and Newborn health priority setting partnership in Uganda (mnhugpsp) working in association with the James Lind Alliance. It also stipulates the basic roles and responsibilities of the partners and wider stakeholders therein. The protocol details procedures for setting up the initiative, recruiting partners and stakeholders into the initiative, as well as conducting and monitoring it. It is not a guide for management of any specific condition or complications during pregnancy, childbirth and or after childbirth.

The protocol will be reviewed and approved by the Steering Group. The most up-to-date version of this document can be found at www.mnhpsp.ac.ug

The PSP will be co-ordinated by Sanyu Africa Research Institute, , Uganda. Any queries should be addressed to the PSP coordinator or James Ditai, the principal investigator (email: [email protected] and Phone: +256782620193).

Study protocol version 1.5 protocol date 01 Dec 2019 3 Table of Contents

PROTOCOL REVISION HISTORY ...... 2 LIST OF ABBREVIATIONS ...... 2 1.0 GENERAL INFORMATION ...... 3 2.0 THE STUDY ADMINISTRATION ...... 5 2.1 STUDY STEERING GROUP ...... 5 2.2 LAY MOTHERS’ GROUP ...... 6 2.3 PARTNERS ...... 6 2.4 EXCLUSION CRITERIA...... 7 2.5 OPERATIONAL RELATIONSHIP BETWEEN STEERING GROUP, PARTNERS AND THE PUBLIC ...... 8 3.0 SYNOPSIS ...... 9 4.0 BACKGROUND ...... 10 4.1 BURDEN OF DISEASE ...... 10 4.2 RESEARCH PRIORITY SETTING ...... 11 4.3 JAMES LIND ALLIANCE ...... 11 4.4 MATERNAL AND NEWBORN HEALTH PRIORITY SETTING PARTNERSHIP IN UGANDA ...... 12 4.5 AIMS AND OBJECTIVES OF THE PSP...... 13 4.6 SCOPE OF THE PSP ...... 13 4.7 THEORETICAL FRAMEWORK ...... 15 4.8 RATIONALE FOR MATERNAL AND NEWBORN HEALTH PRIORITY SETTING PARTNERSHIP IN UGANDA ...... 16 5.0 METHODS ...... 17 5.1 INTRODUCTION ...... 17 5.2 STUDY DESIGN ...... 17 5.3 STUDY SETTING ...... 18 5.6 THE JLA PROCESS ...... 20 Step 1: Identification and invitation of potential partners ...... 20 Step 2: Inaugural Steering Group meeting ...... 20 Step 2i: Initial local awareness meeting ...... 21 Step 3: Initial survey for gathering questions...... 22 Step 4: Data processing, and refining questions...... 27 Step 5 Interim prioritisation ...... 30 Step 6: Final Prioritisation workshop...... 32 5.7 Dissemination of findings and research ...... 33 6.0 PATIENT AND PUBLIC INVOLVEMENT ...... 35 6.1 PPI IN THE PSP PROCESS ...... 35 6.2 EVALUATION OF THE PATIENT AND PUBLIC INVOLVEMENT IN THE PSP IN UGANDA ...... 36 7.0 ETHICS AND FUNDING ...... 37 7.1 ETHICS ...... 37 6.2 FUNDING ...... 37 6.3 OVERALL PSP TIMELINE ...... 37 REFERENCE: ...... 38

Study protocol version 1.5 protocol date 01 Dec 2019 4 2.0 The study administration 2.1 Study steering group The steering group will lead and manage the Maternal and Newborn health priority setting partnership in Uganda. The group includes membership of mothers, carers, health workers and academics1, as individuals or representatives from a relevant group or organisations. No. Names Area/ organisation Country Title/email Phone A Mother and carer representatives 1 Enid Nazziwa Kaliro Uganda Mother representative 2 Rachael Madaya Mbale Uganda Social scientist/mother rep 3 Susan Watuwa Primary school/Mbale Uganda Teacher/ mother rep 4 Godfrey Butoto Busiu Uganda VHT /ambulance 5 Rev Alice Wataka UCU-Mbale chaplain Uganda Chaplain/ Mother Rep 6 Cynthia Nalumansi Step radio station/Mbale Uganda Radio presenter/mother rep 7 James Kidulu Christian Childcare programme Uganda Director/carer rep. B Clinical representatives 1 Grace Nabulo Busiu HCIV Uganda Midwife 2 Florence Ouchi Uganda Neonatal nurse 3 Auma Proscovia Mbale Hospital Uganda Midwife 4 Mugabe Kenneth Mbale Hospital Uganda Obstetrician 5 Rebecca Nekaka Women in Medical world Uganda Patron, mother rep 6 Stella Kisolo District Health office Uganda ADHO (MCH) 7 Martin Chebet SAfRI/BUFHS Uganda Paediatrician C Academic representatives 1 Anthony Mbonye Makerere University Uganda 2 Andrew Weeks University of Liverpool UK D 1 James Ditai Sanyu Africa Research Institute Uganda 2 David Mukunya Sanyu Africa Research Institute Uganda 3 Enid Kawala Sanyu Africa Research Institute Uganda E 1 Toto A Gronlund University of Southampton UK Table 1: Maternal and newborn PSP Steering group composition

1 In this case, academic researchers are represented on the Steering Group, to advise on the shaping of research questions, as they are supervisors to the PhD fellow. However, these researchers shall not participate in the prioritisation exercise. This is to ensure that the final prioritised research questions are those agreed by mothers, carers and health workers only, in line with the JLA’s principles.

Study protocol version 1.5 protocol date 01 Dec 2019 5 The steering group is diverse; some members have extensive knowledge and experience of the James Lind Alliance (JLA) priority setting partnership (PSP), patient and public involvement (PPI) in research and their wealth of knowledge will be invaluable during every stage of the project. The Steering Group will agree the resources, including time and expertise that they will be able to contribute to each stage of the process, with input and advice from the JLA, in line with the agreed terms of reference (appendix 1). 2.2 Lay mothers’ group The lay mothers will contribute lay perspectives to the maternal and newborn research priority setting partnership in Uganda. The group of 5 members includes typical village women, whose voices are hardly heard of, majority of which are illiterate but with previous experience or adversity in maternal and or neonatal health. It is a sub-group, that will be chaired by SG member to relay their thoughts and perspectives to the Steering group.

2.3 Partners

Organisations and individuals will be invited to be involved with the maternal and newborn health priority setting partnership as partners. Partners are defined as ‘organisations or groups who commit to supporting the priority setting partnership by disseminating the survey, promoting the initiative and encouraging their represented groups or members to participate in gathering questions and uncertainties of practical clinical importance relating to the treatment and management of the health problem in question’(The James Lind Alliance Priority Setting Partnerships, 2018). Partners that represent the following groups will be identified and recruited:

• Women who have experience with pregnancy, childbirth and or newborn care in Uganda • Carers of women who have had pregnancy, childbirth or newborn care experience in Uganda • Midwives, nurses, doctors and other allied health professionals with experience of skilled birth attendance or newborn care in Uganda.

Organisations which can reach and advocate for these groups will be invited to become involved in the PSP. Working in association with the JLA adviser or chair, the principal investigator and PSP coordinator will ensure that various stakeholder groups contribute equitably to the process. Table 2 shows the proposed partners.

Study protocol version 1.5 protocol date 01 Dec 2019 6 Name of partner Contact person Position Contacts

Busitema University Prof Julius Faculty Dean Faculty of Health sciences Wandabwa Uganda Medical Dr Medi President Association Elgon Chapter Mulongo Uganda Paediatric Julian Abeso Lead Association National Midwives Acam Vice secretary Association of Uganda Janefrances Mbale Regional Referral Dr Emmanuel Hospital Director Hospital Tugaineyo Health Dr Wangisi District Health Office Jonathan Officer Mt Elgon Hospital Salem Kolonyi Hospital Rural Empowerment on Sydney President Health Management Nsubuga (REHEMA) Montana Hospital Dr Kasoro Andrew IntraHealth RHITES-E Simiyu MNCH/FP/eMTCT Christine Technical Advisor St Andrews Cathedral, Rev Juliet Reverend Mbale Wabwire Boda Boda Association Mbale stage Chairperson Table 2: List of partners for the Maternal and Newborn health PSP

2.4 Exclusion criteria Some organisations may be judged by the JLA or the Steering Group to have conflicts of interest. These may be perceived to potentially cause unacceptable bias as a member of the Steering Group. As this bias is likely to affect the ultimate findings of the PSP, those organisations will not be invited to participate. It is possible, however, that interested parties may participate in a purely observational capacity when the Steering Group considers it may be helpful.

Study protocol version 1.5 protocol date 01 Dec 2019 7 2.5 Operational relationship between steering group, partners and the public

Though the steering group identifies the partners, the partners will support and promote the PSP alongside identifying potential members to participate in the survey. The overall structure is illustrated in figure 1 below

A small manageable group Coordinates and implements the PSP Steering Group activity Includes mothers, carers, health workers, academics and JLA advisor • Lay Mothers Group

Organizations and groups with interest in maternal and newborn health Partners

Mothers, carers Wider stakeholders and health Mothers workers Carers Health workers Community representatives

Figure 1: Operational relationship between steering group, partners and the public

Study protocol version 1.5 protocol date 01 Dec 2019 8 3.0 SYNOPSIS

Title Maternal and Newborn Health Priority setting Partnership in Uganda Short title Maternal and Neonatal health research priorities in Uganda Acronym mnhUG-PSP website www.mnhpsp.ac.ug Objectives Purpose: To identify the unanswered questions about maternal health and newborn care from mothers, carers and clinical perspectives in Uganda and then prioritise those that mothers, carers and health workers agree are the most important for the research to address. Specific objectives: • To work with mothers, caregivers and health workers to identify uncertainties about maternal health and newborn care in rural Uganda • To agree by consensus on a prioritised list of those uncertainties about maternal health and newborn care, for research in Uganda • To establish the research priorities for maternal health over those of neonatal health in Uganda • To publicise the process and results of the Maternal and Newborn Health Priority Setting Partnership in Uganda • To share the results with the ministry of health, academic researchers and funding bodies Research design Multiphase mixed methods design Research The James Lind Alliance Priority setting partnership strategy Setting Mbale Hospital and Mbale Municipality Kolonyi HCIV and 2 villages Busiu HCI and 2 villages Budaka HCIV and 2 villages Population Seldom heard groups of mothers, caregivers and health workers in the study sites Duration This will be a short time PSP lasting a duration of 6-8 months (i.e. initial survey 2-month, collation 1-month, interim survey 2-month, final workshop 1 month). Outcome Primary Outcome: measures A top 10 list of prioritised maternal and newborn health priorities. Secondary Outcomes: Maternal: A top 30 list of maternal health research priorities in Uganda, a list of all maternal health research priorities. Neonatal: A top 30 list of neonatal health research priorities in Uganda, a list of all neonatal health research priorities.

Study protocol version 1.5 protocol date 01 Dec 2019 9 4.0 BACKGROUND 4.1 Burden of disease

Globally, an estimated 303,000 maternal deaths occur, with an average lifetime risk of 1 in 180 (World Health Organization, 2015b). However, almost all, 99% (an estimated 302,000) maternal deaths occurring in low-income regions (LIRs) with the average lifetime risk of 1 in 150 compared to 1 in 4900 in high-income regions (Alkema et al., 2016; Kassebaum et al., 2016; World Health Organization, 2015b). That enormous hidden difference not only exists between regions and continents but also within continent-specific regions and countries. Sub- Saharan Africa (SSA) bears the highest maternal deaths globally, an estimated 66% (201,000 deaths) of the global total with the highest lifetime risk of 1 in 37 (Alkema et al., 2016; Kassebaum et al., 2016; World Health Organization, 2015b). In Africa, further disparities exist between countries; Sierra Leone has the highest estimated maternal mortality ratio (MMR) of 1360 per 100,000 live births, followed by the Central African Republic (MMR of 882/100,000 live births) and Chad (856/100,000 live births). Even within a country, differences are not uncommon. In Uganda, the maternal mortality ratio is 336/100,000 live births, with great variation in region- specific mortality ratios (Uganda Bureau of Statistics (UBOS) and ICF, 2018). Maternal mortality is a tragedy for an individual woman, for her family and the community.

Considering the high maternal mortality in the developing countries, WHO in 1987 conceived the idea of safe motherhood initiative at a conference in Nairobi, Kenya. This is a global effort to reduce the maternal deaths by at least half by 2000 (O’Loughlin, 1997), which was later extended to 2015 via the UN Millennium Development Goals (MDGs) and now extended to 2030 via the sustainable development goals (SDGs)(United Nations, 2018). The objectives are to enhance the quality and safety of girls and women’s lives through adoption of a combination of health and non-health strategies. Maternal and child health promotion is a key commitment in the WHO constitution (WHO, 2006).

Diversity and divergence define poor maternal health in the 21st century(Graham et al., 2016). Health workers need to provide interventions based on region-specific priorities that address the variations in maternal outcomes. However, the health research priority setting mostly happens at the highest international levels and in developed countries, with little in LIRs (Youngkong, Kapiriri, & Baltussen, 2009). Hence, directing the limited available resources into the most effective interventions and health research (Rudan et al., 2007, 2010)

Study protocol version 1.5 protocol date 01 Dec 2019 10 still remains a challenge in SSA. Often policies and technologies or solutions seek to resolve a problem without considering these experts’ complexities of everyday life. To adequately address maternity health challenges faced in LIRs, research therefore needs to be designed to address the local conditions of local women and communities, working with women to choose what research is undertaken (Chalmers, 1991). The proposed priorities need to come from the users of the services at the grassroots level and meet their demands (Aranda-jan, Jagtap, & Moultrie, 2016).

4.2 Research priority setting

There is an increasing level of emphasis being placed on health care providers and funders to incorporate patient-centred care into research. There is a remarkable rise in public involvement in research between 1995 and to date (Boote, Wong, & Booth, 2012). Public involvement in establishing research priorities ensures the relevance of the research produced. Priority setting is a process that can be adopted to produce a robust set of research questions to be addressed over time (Barnieh, Jun, Laupacis, Manns, & Hemmelgarn, 2014). A few groups in low and middle-income countries (LMICs), have developed their own practices of priority setting for health research (Tomlinson, Chopra, Hoosain, & Rudan, 2011) but these have methodological limitations. The common methodologies, developed to be implemented at a national level, do not generally provide a transparent algorithm to distinguish between competing research investment options. Furthermore, in those LMIC settings that have used one of the transparent methods (Yoshida, 2016), especially James Lind Alliance (JLA) (Cowan & Oliver, 2013) and Child Health and Nutrition Research Initiative (CHNRI) methodology (Rudan et al., 2008), the initial methods of data collection are only applicable to the affluent and with access to internet facilities. The public is rarely involved in the priority setting process. The research priorities are therefore decided mainly by skilled professionals or experts who do not live the life of those affected by many of the health problems that affect the poorest.

4.3 James Lind Alliance

The James Lind Alliance (JLA) is a non-profit making initiative, established in 2004 (Cowan & Oliver, 2013). It brings patients, carers and clinicians together in Priority Setting Partnerships (PSPs). These partnerships identify and prioritise uncertainties, or ‘unanswered questions’, about the effects of treatments that they agree are the most important. The aim of

Study protocol version 1.5 protocol date 01 Dec 2019 11 this is to help ensure that those who fund health research are aware of what really matters to both patients and clinicians. The UK’s National Institute for Health Research (NIHR – www.nihr.ac.uk) funds the infrastructure of the JLA to oversee the processes for priority setting partnerships, based at the NIHR Evaluation, Trials and Studies Coordinating Centre (NETSCC), University of Southampton.

4.4 Maternal and Newborn Health priority setting partnership in Uganda

The ministry of health has demonstrated some prioritisation for sexual, reproductive, maternal neonatal and child health in Uganda (Wallace & Kapiriri, 2019). However, not all actors participate in the process and there is no single systematic process for identifying priorities for research. Over the past 7 years, Sanyu Africa Research Institute (SAfRI) has been exploring models for involving the women in designing interventions and research (J. Ditai, Kanyago, et al., 2018; James Ditai et al., 2018; Lissauer et al., 2018; Weeks et al., 2015) and we have found women and communities to be very responsive in generating interventions that are locally responsive to their needs in these settings.

Over the past two years, a transnational team was identified, and its members aim to develop global maternity research agenda with mothers, families, health workers, and researchers in Mbale, eastern Uganda. Our team represents research interests of Sanyu Africa Research Institute, Makerere University, Ministry of Health and Mbale Regional Referral Hospital. This research initiative brings together technical experts and ‘experts by lived experience’ of maternal and newborn services in rural Uganda, to develop locally appropriate research priorities.

We will conduct a Maternal and Newborn health priority setting partnership in Uganda. We are working in association with the James Lind Alliance on involving strata of the population whose voices are often not heard, even though they are the ‘experts by experience’ of local maternal and neonatal services. The primary aim is to conduct a Priority Setting Partnership using the JLA methodology and guidance to lead the identification of clinically important maternity and newborn research questions and emerging priorities for key Ugandan stakeholders, including patients, caregivers, health workers (knowledge users and policy makers) and researchers. Identifying Ugandan maternity research priorities will help focus our research agenda and link to patient-oriented research efforts.

Study protocol version 1.5 protocol date 01 Dec 2019 12 4.5 Aims and objectives of the PSP

The aim of the Maternal and Newborn health priority setting partnership in Uganda is to identify the unanswered questions about maternal health and newborn care with mothers, carers and clinical care teams in Uganda and then prioritise those that mothers, carers and health workers agree are the most important for the research to address.

The objectives of the Maternal and Newborn health priority setting partnership in Uganda are to:

• work with mothers, caregivers and health workers to identify uncertainties about maternal health and newborn care in rural Uganda

• to agree by consensus on a prioritised list of uncertainties about maternal health and newborn care, for research in Uganda • to establish the research priorities for maternal health over those of neonatal health in Uganda • to publicise the process and the results of the Maternal and Newborn Health Priority Setting Partnership in Uganda • to share the results with the ministry of health, academic researchers and funding bodies

4.6 Scope of the PSP

The Maternal and Newborn health priority setting partnership in Uganda will cover aspects of care during pregnancy and the immediate postpartum period, and the management of complications that develop during this period (World Health Organisation, 2017a). The initial scope, developed in June 2018 and agreed at the first Steering Group meeting, also covered gynaecological conditions. Since then, limitations in time and funding have led to the scope being narrowed and gynaecological disease removed.

Study protocol version 1.5 protocol date 01 Dec 2019 13 Up to 27 weeks of amenorrhea Up to 6 weeks after childbirth

First Trimester Second Trimester Third Trimester Postpartum period Up to 13 weeks of Up to 40 weeks of 7-days Up to 28 days amenorrhea amenorrhea Neonatal period

Up 42 to days after birth Labour/ childbirth normalLast menstrual period

Figure 2: PSP scope

Inclusion criteria

The scope in this phase of the PSP is to include broadly questions about ‘maternal health’ and ‘newborn health’. The initiative’s broad remit will help us to gather more evidence uncertainties (The James Lind Alliance Priority Setting Partnerships, 2018), as well as design future PSPs with a tighter remit e.g. birth asphyxia PSP, pre-eclampsia PSP.

For the purpose of this PSP, maternal health is defined as ‘the health of women during pregnancy, childbirth, and the postpartum period(World Health Organisation, 2017b). It encompasses the healthcare dimensions of family planning, preconception, prenatal, and postnatal care in order to ensure a positive and fulfilling experience, in most cases, and reduce maternal morbidity and mortality, in other cases (World Health Organization, 2015a). For the purpose of this PSP, a postpartum (or postnatal) period commences immediately after the childbirth until 6 weeks. The terms puerperium or puerperal period, or immediate postpartum period are commonly used to refer to the first six weeks following childbirth(Romano, Cacciatore, Giordano, & La Rosa, 2010). The postnatal period is the most critical when most maternal and newborn deaths occur (World Health Organisation, 2013). During this period, newborn health is embedded and corresponding questions related to the newborns will be explored (World Health Organisation, 2017c). We will include questions about promoting and managing maternal and newborn health, and about preventing and treating the complications in maternal and newborn health. There are successfully completed PSPs with a broad remit to provide guidance and lessons for the proposed initiative, available at http://www.jla.nihr.ac.uk/priority-setting- partnerships/. The examples include the Sight Loss and Vision PSP, adult social work PSP,

Study protocol version 1.5 protocol date 01 Dec 2019 14 the anaesthesia PSP, the palliative and end of life care PSP (The James Lind Alliance Priority Setting Partnerships, 2018).

Exclusion criteria: As this priority setting partnership has its focus on “maternal health” and “newborn health”, we will exclude comments, questions unrelated to maternal and newborn care, for example the gynaecological related questions.

4.7 Theoretical Framework

The PSP is underpinned by a socio-ecological model (SEM) (Elder et al., 2007; Salihu, Wilson, King, Marty, & Whiteman, 2015), a system of classifying the main influences in research priority setting in low income regions (figure 2). We will apply an SEM-based ‘framework for holistic contextual design for low-resource settings’ whenever the focus is on the design of medical devices (Aranda-jan et al., 2016). The model has its foundation in the social choice theory (King, 1987), combining individual preferences to reach a collective decision, as well as the utility theory (Fishburn, 1968), understanding the interest of individuals before talking of the interest of the community. This model has the potential to enable women themselves to explore ‘why maternal and newborn adverse outcomes continue to occur?’ and what would be the priority interventions, thus exploring the contribution of social determinants to health inequalities (Roos & von Xylander, 2016). In this way, it can identify inclusive research priorities applicable to the local context using a formalised, transparent methodology (Cowan & Oliver, 2013). In the proposed study, the seldom heard groups of people will determine their maternity and neonatal health research priorities regardless of social status, age, literacy level and internet access.

Policy/Enabling Environment (national, state, local laws)

Organizational (organizations and social institutions

Community (relationships between organizations)

Interpersonal (families, friends, social networks)

Individual (knowledge, attitudes, behaviors)

` Figure 3: The Social Ecological Model (conceptual framework) for maternity research priorities

Study protocol version 1.5 protocol date 01 Dec 2019 15 4.8 Rationale for Maternal and Newborn Health Priority Setting Partnership in Uganda The study will inform the establishment of a protocol for prioritisation of the Maternal and neonatal health (MNH) in Uganda. If proven effective and appropriate in MNH, this methodology will ultimately be adopted as one of the methodologies in identifying research priorities for other disease conditions in Uganda or any LMICs. The results from this study will inform the design of studies out of women’s voices or seldom heard population. The results of maternity-newborn research priorities with the inclusion of the “experts by experience opinion” will guide Ministry of Health Uganda and current funders in resource allocation, as well as the RCOG global health division and other researchers in research agenda development. We anticipate the findings could be rapidly incorporated into the future WHO research priorities and or prioritisation exercise (facilitated by University of Liverpool Department of Women’s and Children’s health, WHO Collaborating Centre status).

Study protocol version 1.5 protocol date 01 Dec 2019 16 5.0 METHODS 5.1 Introduction This section describes a schedule of proposed steps through which the PSP aims to meet its objectives. The process is iterative and dependent on the active participation and contribution of different groups. The methods used in any step will be agreed through consultation between the Steering Group members, guided by the PSP’s aims and objectives. More details of the method are in the Guidebook section of the JLA website at www.jla.nihr.ac.uk where examples of the work of other JLA PSPs can be seen (The James Lind Alliance Priority Setting Partnerships, 2018). 5.2 Study design The study will adopt a multiphase mixed methods design (Creswell, 2014). The initial phase of the study will mainly apply a qualitative methodology to generate raw questions. This will be followed by the quantitative approach to rank a list of questions in the order of priority. Then the last phase will apply both qualitative and quantitative approaches to prioritise and agree on the top ten research questions.

The design follows the James Lind Alliance (JLA) priority setting partnership methodology. This is a multi-stepped pragmatic methodology for priority research setting (Barnieh et al., 2014; Cowan & Oliver, 2013). The JLA priority setting partnership (PSP) brings together patients, carers and health professionals to identify treatment uncertainties, which become research questions (Barnieh et al., 2014; Cowan & Oliver, 2013; Yoshida, 2016). The method considers the opinions of ‘experts by experience’. The method is “focussed on being inclusive, transparent, and evidence-based” (Barnieh et al., 2014; The James Lind Alliance Priority Setting Partnerships, 2018).

In this study, we have made the JLA PSP locally appropriate to the differing needs of research priority in a low-income region in Africa. We are using a non-internet-based system for collection of opinions from the lay members of the public and health workers. We plan to go door-to-door to conduct interviews with members of the public using smart phone data collection systems, followed by village focus groups in a cross-section of Uganda community settings. We will target especially the poorest and those whose voices are often unheard, before convening a final prioritisation exercise. An iterative, interpretative phenomenological approach might be applied in exploring the lived experience of these experts, like in other studies (Smith & Osborn, 2015).

Study protocol version 1.5 protocol date 01 Dec 2019 17 5.3 Study setting The PSP will be conducted in 5 targeted hubs (east, north, central, west and south) of Mbale town in Eastern Uganda. The point of reference for the hubs will be the Health facility purposively selected based on the compass direction (guided by main roads out of Mbale town. We will map out villages around each facility in each hub using the National Identity (NIN) Database Registry at the National Identification and Registration Authority (NIRA) or ministry of internal affairs or manually drawn villages present at the health facility. Our plan is to select 2 villages surrounding each hub by simple random sampling, as illustrated in figure 4.

Three of the Hubs or health facilities will be the participating sites for the BabyGel trial, which has been funded by European and Developing countries Clinical Trials Partnership (EDCTP) while the Sanyu Africa Research Institute has been implementing the clinical trials in all these study areas for at least 3 years.

Study protocol version 1.5 protocol date 01 Dec 2019 18

Figure 4: levels to sample population for maternity and neonatal health research priority setting in Mbale, Uganda

Study protocol version 1.5 protocol date 01 Dec 2019 19 5.6 The JLA Process The maternal and newborn health PSP in Uganda will follow the JLA research strategy, illustrated in the visual model below

Initial survey Gathering Questions (n=200)

Refining Questions Collation Verifying Questions

A short list of questions: Top Interim Prioritisation 20-25 (n=60-100)

Final Prioritisation Top 10 questions (n=27)

Output Reporting Dissemination

Figure 5: JLA process for the maternal and newborn health PSP in Uganda

Step 1: Identification and invitation of potential partners The established Steering Group will identify potential partner organisations through a process of peer knowledge and consultation, and through the Steering Group members’ networks. The principal investigator will contact the potential partners, inform them of the establishment and aims of the Maternal and Newborn Health Priority Setting Partnership in Uganda and invite them to take part in the initial awareness meeting. We expect at least 1 professional and 1 health consumers’ organisation to participate as partners.

The Steering Group will approve the invitation (appendix 2) and use this to invite identified contacts to join the initiative as partners. This work will formally commence at the launch of the PSP and continue during Step 3: initial survey for gathering uncertainties. Step 2: Inaugural Steering Group meeting

The inaugural Steering Group (SG) meeting will be planned on an agreed date and times after being granted permission to proceed from the University of Liverpool sponsorship

Study protocol version 1.5 protocol date 01 Dec 2019 20 committee. The meeting will occur in the month of December 2019. This meeting’s key objectives will include:

• To welcome and introduce members of the Maternal and Newborn Health Priority Setting Partnership in Uganda PSP Steering Group • To present the proposed plan for the Maternal and Newborn Health Priority Setting Partnership in Uganda • To discuss and agree the scope of the Maternal and Newborn Health Priority Setting Partnership in Uganda and ways to maximise the responses to step 3. • To initiate discussion, answer questions and address concerns related to the protocol, initial survey tool • To identify additional partner organisations which may commit to the PSP and confirm individuals who will be those organisations’ representatives and the PSP’s principal contacts • To establish principles upon which an open, inclusive and transparent mechanism can be based for contributing to, reporting and recording the work and progress of the PSP in this specific setting.

The agenda for this inaugural SG meeting will be shared well in advance to all members (appendix 3). SG members will be active in raising awareness of the PSP among their networks and communities, in order to secure support and participation. This shall take the form of conversations, mentions at institutional meetings, events and talks, signposting people to the project webpage and using the #maternalpspuganda hashtag on social channels. An initial meeting with the entire team was held in London on the 5th June 2018 to discuss this study, and there have been ongoing teleconferences between the investigators and the JLA. Step 2i: Initial local awareness meeting After the inaugural SG meeting, the members of SG will organise a half day awareness meeting at the Sanyu Africa Research Institute. The meeting will include partners and stakeholders within Mbale as participants and led by the principal investigator. The objectives of the initial local awareness meeting will include • To introduce the proposed plan for the Maternal and Newborn Health Priority Setting Partnership in Uganda. • To welcome and introduce the representatives of the partner organisations

Study protocol version 1.5 protocol date 01 Dec 2019 21 During this initial local meeting, members will be encouraged to comment on the tools and process. Further, five members of the public will be invited to participate in the piloting of the initial survey tool. Data from this pilot will be used to revise the tool for the initial survey. We will too launch the initial survey in this meeting (appendix 4 for the poster). Step 3: Initial survey for gathering questions. Objective: The initial survey will be a consultation process, which will produce “raw” unanswered questions about prevention, diagnosis, treatment, management, quality of life, and support in maternal and newborn health in Uganda. Research design: Exploratory research design/ survey Study Population A purposive sample in each SEM level will be recruited using agreed inclusion criteria (table 3). We already hear the voices of international NGOs and policymakers, but this survey is instead focussed on local rural women, health workers and organisations relevant to them. This initiative includes an important population, whose first language is not English, where more than 50% of women completed only primary education and below at the village level in this setting (J. Ditai et al., 2019; Weeks et al., 2015). SEM Level Description Individual • Women of reproductive age, regardless of religious identity, socio- (Mother) economic status, or literacy. • Pregnant and postnatal mothers • Mothers with history of any maternal or newborn morbidity Interpersonal Social networks and social support systems, including (carers) • Family (husbands, parents of mothers with any experience of maternal or neonatal morbidity / mortality) • friends, peers, or co-workers of the above individuals. • religious networks (Religious leaders) • customs or traditions (Clan leaders, Traditional birth attendants. Community Relationships among organizations, institutions, and informational networks (wider within defined boundaries, including stakeholders) • village associations (women’s groups) • community leaders (Local council I-III) • transportation (UTODA leaders, motorcyclist and car drivers) • Village health teams (VHTs)

Study protocol version 1.5 protocol date 01 Dec 2019 22 Organizational Organizations or social institutions with rules and regulations for operations (Health and that affect how, or how well, maternity and newborn services are provided to social care an individual or group; professionals) • Schools that include women’s health in the curriculum (primary, secondary). • Tertiary institutions (Universities, colleges) • Health facilities (private and public clinics/ hospitals) • Local NGOs focused on Women’s Health, IntraHealth Table 3: Study population and inclusion criteria.

The individual or ‘patient as indicated in the JLA guidebook’ will be the mother in this initiative, and the different categories of mothers have been included for inclusiveness of viewpoints. The carers include the wider family members. We have too included the community as there are stakeholders who contribute directly to the health of mothers and new- borns and may have a viewpoint (VHTs, motorcyclists). These are relevant but seldom-heard groups, which is in line with the JLA guidance (The James Lind Alliance Priority Setting Partnerships, 2018).

The organisational SEM level is composed by health workers mainly. Our sample will include health workers with experience in maternity and newborn care, namely the midwives, nurses, clinical officers, doctors. We have further included other social care workers to capture a wider range of viewpoints in line with the JLA process (The James Lind Alliance Priority Setting Partnerships, 2018).

Sample selection Individuals from each SEM level will be selected to participate. At the individual level, participants will be selected from the towns and villages by simple random sampling using the sampling frame from the National Identification Number (NIN) Registry. Whilst in homes of recruited participants, the Research Assistant (RA) will inquire about the occurrence of any maternal mortality or severe maternal morbidity in the village and the immediate relatives recruited for participation. An exponential non-discriminative snowball sampling will be used to sample all mothers or families that have experienced any adverse maternal or newborn outcomes for participation in the survey. In this way, participants at interpersonal level will be identified and recruited. At the Community level, religious and political leaders, community groups or associations in each village will participate. At an organisation level, potential local

Study protocol version 1.5 protocol date 01 Dec 2019 23 organisations will be those who have direct or indirect experience of severe maternal-neonatal morbidity and mortality, those representing health consumers, medical doctors, midwives, nurses and professionals allied to medicine with clinical experience of maternity and newborn services. Sample Size We shall target an average sample of 120 (100-140) seldom-heard participants (mothers and their families), 20 (10-30) wider stakeholders and 80 (60-100) health and social care professionals) for face to face data collection. The survey will also be uploaded online, and we expect to reach about 100 persons with access to internet. Hence, we expect an overall total of 320 participants in the survey.

Though this is a smaller sample size for initial survey compared to previous JLA PSPs (Gadsby et al., 2012; Hall, Mohamad, Firkins, Fenton, & Stockdale, 2013; Koblinsky et al., 2016; Rees et al., 2017; The James Lind Alliance Priority Setting Partnerships, 2018; Van Middendorp et al., 2016; Victora et al., 2016). We expect at least 600 responses to be generated from this sample. This is within the range of 100s to 1000s for the initial survey responses from previous PSPs (The James Lind Alliance Priority Setting Partnerships, 2018).

Unlike the previous studies that have used online surveys, this study will involve face-to-face interviews, which could get complicated to reach the same target population. We have included a diverse range of participants rather than looking for numbers. The details of the sample size are summarised below

Sample Sample size Mothers- General 115 Family members (focus groups) 5 with size of 5-8 participants Wider stakeholders Community 20 (priests, imams, VHTs, TBA, cyclists, LCI) Health workers 30 (midwives, doctors, clinical officers, nurses) Social care workers 50 (teachers, engineers, other professionals) Online users – general public 100 Table 4: Sample size distribution.

Study protocol version 1.5 protocol date 01 Dec 2019 24 The sample size will be attained in a rotational recruitment across the 5 hubs, that is, recruiting 8 mothers plus a focus group of family members in one village per Hub in the first cycle, then repeat in the second cycle until data saturation has been reached. Recruitment Home recruitment The research assistant will identify the mother for the face-to-face interview in her home. This will then be followed by a focus group with the members at the interpersonal SEM level (husband, mother in law, sister in laws, friend, neighbour, church leader). The research assistant or principal investigator will give verbal explanation of the study to the potential participant and a copy of the participant information sheet (appendix 5) at this stage of the study. Community recruitment The RA will recruit other wider stakeholders in the village for a face-to-face semi-structured interview. These will include the motorcyclist, the VHT member, LCI secretary for women and public health and any women’s groups’ leaders. Facility recruitment The RA will recruit a nursing assistant, midwife, clinical officer, doctor at a nearby health facility serving the participating village and administer the structured interview. SG recruitment Each local member of the Steering Group (SG) will be requested to identify and recruit at least 5 participants to participate in the initial survey. Data collection Development of the data collection tool. The Steering Group have developed the initial priority setting survey tool to identify unanswered questions relating to maternity and newborn health (appendix 6). Questions within the tool have been developed iteratively with women, mothers, carers and professionals. This survey tool contains six questions, asking the public to consider their own experiences across six specific domains. The specific domains include pregnancy/antenatal care, labour/childbirth, treatment of complications of pregnancy and childbirth, postnatal care, Newborn care, and gynaecological care. These areas were derived from the existing literature and textbooks for obstetrics, gynaecology, and neonatology. These questions will be uploaded onto the ODK, an electronic data capture system (Baker & Kenny, 2011; Edmonds, 2008; F. Gary Cunningham et al., 2010; Monga, 2006). The initial survey tool has been translated into the local language to allow access to the rural population who neither speak nor understand English.

Study protocol version 1.5 protocol date 01 Dec 2019 25 Pilot testing of the tool. Initial piloting of the tool was conducted in Uganda amongst five stakeholders of Sanyu Africa research Institute (SAfRI) for design and clarity of questions, namely 2-mothers, 1-midwife, 1-priest and 1-motorcyclist. However, we plan to further pilot the tool to five participants a prior or during the initial awareness meeting. Data collection process The methods of data collection will be individual interviews and group interviews with some members of the family. Unlike the UK’s online surveys for this step (Barnieh et al., 2014; Gadsby et al., 2012; Hall et al., 2013; Koblinsky et al., 2016; Van Middendorp et al., 2016; Victora et al., 2016), the researchers will physically visit (Pollock, George, Fenton, Crowe, & Firkins, 2014) the selected individuals in their homes door-to-door in each village, clinics and organisations or during meetings or workshops to collect their opinions. Face-to-face visits engage more patients than an online survey (Pollock et al., 2014). He/she will administer the survey questions in the appropriate language using Open Data Kit (ODK; www.opendatakit.org ) loaded onto a mobile smartphone.

The researcher will ask the participants to propose questions that they think should be considered as potential maternity or neonatal priority research topics in an open-ended format to a broad question about each of the six specific domains. Participants will be offered a seventh open-ended comment box for any other items not considered in the six thematic domain questions. Further, the researcher will collect demographic questions from each participant, a privacy notice will be shared with the participant (appendix 7).

The interviews will be audio-recorded automatically into the ODK phone, transcribed and translated from the local languages. Extra notes taken by research assistant will be incorporated into final transcripts and imported into the dataset for further analysis using NVivo Qualitative software. Where need arises, images and videos will be taken using the ODK system. Duration of initial survey: The research assistants will administer the questions for a period of 45-60 days. We expect this to be sufficient time to generate enough responses to achieve the data saturation in this survey.

The principal investigator will download the survey responses weekly to start the cleaning and categorisation of data while the survey is still open. An interim sub-analysis of demographic details will be performed fortnightly to identify under-represented study population and employ strategies to reach the specific groups.

Study protocol version 1.5 protocol date 01 Dec 2019 26

The Steering Group members and all partners will be encouraged to promote the survey through their social groups, WhatsApp groups, departmental or village meetings.

Step 4: Data processing, and refining questions. Objective: To assemble, categorise and refine the raw questions from the initial survey into a list of unique researchable questions.

Participants. The principal investigator will be responsible for this stage. The JLA adviser will advise on the amount of time likely to be required for its execution. The JLA adviser will participate in this process as an observer, to ensure accountability and transparency.

The data processing The process will occur in a series of stages, with different SG members taking different roles, as described below

Stage 1: Update and download the survey data All survey responses will be collated into the ODK electronic data capture system. Any paper responses will be added into the survey ODK system, while any audio-recordings will be transcribed verbatim, translated (if in local language) and entered into the ODK system before the weekly download. The principal investigator will transfer the downloaded data into the data handling excel file (appendix 8) and clean and organise the data for discussion with the steering group. At the end of the survey, a complete and clean dataset will be available for refining. Stage 2: Remove out-of-scope survey submissions During the cleaning phase, submissions which are out of scope of this initiative will be moved and kept in a separate list of the data handling excel file. These may include questions seeking further information or advice on a topic, or issues around awareness. These questions will be discussed by the SG members and a relevant partner identified to address the out-of- scope questions. Submissions already answered by research at this stage will also be removed. We will analyse these separately and pass over any clinical care questions to the Mbale regional referral hospital, Ministry of health and any other interested partner. Stage 3: developing a framework to categorise eligible survey submissions The principal investigator, will develop an initial framework to categorise the raw submissions and organise them into themes following the principles of thematic content approach for

Study protocol version 1.5 protocol date 01 Dec 2019 27 qualitative analysis (Green & Thorogood, 2009). A sub-sample of eligible questions data (comments, one or large tracts of questions) will be exported into NVivo 12 software and analysed using the principles of thematic content approach for analysis (Green & Thorogood, 2009) whilst interpretative phenomenological analysis (that is, interpretation of the insights of how a given person, in a given context, makes sense of events and relationships) will be applied for sub-sample of questions on lived experiences.

He will capture the emerging themes from the selected sub-sample of questions during this process, identifying the common elements in questions relating to maternal and neonatal health. This initial framework will then be piloted with two other SG members independently coding questions and comparing the category labels. This shall include one person with lived experience and one professional. SG members will be asked to volunteer to do this while 4 or 6 SG members will help review the categories, as well as support subsequent data analysis. Disagreements will be resolved through discussion and consensus. The principal investigator will further test the framework with other remaining questions, which will be reviewed and amended. The final framework will be presented to the Steering group for discussion and approval. This process may further identify ineligible submissions, which will be shifted to the out-of-scope section. Stage 4: applying the framework to the eligible survey submissions The principal investigator will then apply the agreed framework in stage 3 to the entire dataset. Checks will be instituted to ensure consistency of the approach. A second member of the steering group (project coordinator or academic chair) will peer review the reliability of emerging themes and issues regularly; 5 participants will be reviewed for every 20 participants, swapping a portion of their respective data and comparing findings for consistency. Once this is done, one SG member with lived experience and one professional will check the produced list of emerging themes. Any discrepancies or issues arising from specific responses will be adjudicated by the SG Chair, JLA adviser and discussed by the steering group if necessary. Stage 5: Synthesis of eligible survey submissions This stage will include combining of duplicate questions, splitting responses that contain multiple questions and further removal of those out of scope. Questions about specific maternal or neonatal services will be re-phrased to relate to the exact terminology referred to. The principal investigator will break down longer responses into several key excerpts as appropriate and have multiple questions created. He will merge questions with common themes into broader questions, where appropriate, as well as remove duplicates. Responses judged to be

Study protocol version 1.5 protocol date 01 Dec 2019 28 irrelevant to maternal and newborn health research will be classed as “out of scope” and considered for another analysis. Stage 6: formulating indicative questions This stage involves rephrasing questions to be clear, focused (if possible), addressable by research and understandable to all. The eligible submitted questions, from the coded data in the framework, will be re-written as collated focused questions in the PICOTs format (Glasziou, Mar, & Salisbury, 2007) or PCC format by the principal investigator. He will group all the indicative questions under maternal health and neonatal health separately to form the long list of questions. The principal investigator will manage the process, discuss and agree on research questions with steering group consistently. A one-off face-to-face meeting will be held to discuss the long list of indicative questions in March 2020.

Stage 7: verifying the Indicative questions. Objective: To produce a list of verified research questions to be put forward for the interim prioritisation. The verification process The principal investigator will check each indicative question against existing literature or systematic reviews through a rapid review of literature. This will allow inclusion of questions in the interim survey only if they have not been adequately answered. He will consult with the librarian at University of Liverpool for results from the search.

We will develop our search strategy by reviewing, combining search strategies from the MASCOT/Wotro Map of Maternal Health Research (Chersich & Martin, 2017). We will run the search, prior the interim prioritisation, and across MASCOT database (http://eppi.ioe.ac.uk/webdatabases4/Intro.aspx?ID=11), MEDLINE, EMBASE, Cochrane Database of Systematic Reviews, Social Sciences Citation Index and ERIC.

The principal investigator will map the findings of relevance, up to date systematic reviews to the identified questions, before being reviewed by two members of the steering group. JLA defines up to date literature to include publications within the last three years (The James Lind Alliance Priority Setting Partnerships, 2018).

Study protocol version 1.5 protocol date 01 Dec 2019 29 We will apply the GRADE system for rating the quality of a body of evidence as a guideline wherever evidence is not clear (Guyatt et al., 2011). A question verification form (appendix 9) will be filled and sent back to JLA.

Indicative questions not adequately answered by existing evidence will be collated and prepared for entry into a data management template (appendix 10). The long list of questions will be reviewed by the steering group to agree on the number of questions to go forward for the interim prioritisation.

Duration: This step will occur for a period of 4-6 weeks

Step 5 Interim prioritisation Introduction: We will perform a second survey to prioritise and shortlist the verified questions. Objective. The aim of this stage of the priority setting process is to rank the long list of verified uncertainties in order of those considered most important by mothers with experience of pregnancy and childbirth, their carers and the health workers in Uganda. Design: This will be a cross-sectional study and the approach to involvement will be consultation with the seldom-heard population (INVOLVE, 2012). Sample Participation to this survey will not be restricted to participants in the initial survey. Our target is to include participants similar in characteristics at each SEM level to those in the initial survey. Sample size We aim to reach a sample of 100 participants. Though this sample size is not based on the average of the previous JLA studies (Gadsby et al., 2012; Hall et al., 2013; Koblinsky et al., 2016; Van Middendorp et al., 2016; Victora et al., 2016) for this step of the JLA process, it is above the childhood disability PSPs (The James Lind Alliance Priority Setting Partnerships, 2018). Unlike, the previous studies that have used online surveys, this study will involve face-to-face recruitment, which would be costly if we attempted to reach the same target population. However, we shall test the adequacy of the sample size iteratively. The principal investigator will review the results from the 40% data against the next 20% of the data for any significant change in the result. Number of indicative questions for inclusion

Study protocol version 1.5 protocol date 01 Dec 2019 30 In this PSP, we will consider up to 70 indicative questions for maternal health and newborn health for interim prioritisation. Previous PSPs have considered an average of 65 (range 40- 114) indicative questions. The following criteria will be used to reduce the large list of indicative questions to the proposed number. • The question based on submission by utmost two participants will be prioritised lower than questions from several participants in the initial survey • The question based on submission from one socio-ecological model level or group will be prioritised lower than questions from more SEM levels in the initial survey The removed questions will be checked to ensure that those questions known to be coming from ‘seldom heard’ persons are retained. Data collection process: The agreed longlist of 70 indicative questions for maternal health and newborn health will be arranged in a format to allow ease of printing on the paper or presenting from the ODK mobile phone. The research assistant will then visit individuals to administer the questionnaires with individual participants from each of the selected SEM levels as above. This will facilitate input from the ‘seldom heard’ groups that are unlikely to do an online survey and definitely need support to participate. We will ask each member of the public to rank the top 30 most important research questions per category of maternal or neonatal health from the long list. Ranking process We will adhere to the JLA approach described in the JLA Guidebook (The James Lind Alliance Priority Setting Partnerships, 2018) to choose the top 10 most important questions, as outlined below. Based on experience of the JLA adviser, asking people to actually rank their top 10 does not add much to the value of the data, hence they will only be asked to choose and not rank. There are examples of PSPs that have applied this ranking approach at www.jla.nihr.ac.uk (The James Lind Alliance Priority Setting Partnerships, 2018). 1. The research assistant will read through the long list of questions for both maternal health and newborn health, and then ask the participant to choose any top 10 questions. 2. Either the participant or research assistant will tick the respondent’s 10 questions in the left-hand box of the questionnaire (appendix 11). 3. RA will then present to participant a specific long list for either maternal or newborn health alone 4. Then Steps 1-2 will be repeated for maternal health and newborn health separately. We have previously used a glass model to demonstrate the orders of importance or satisfaction from carers/ mothers locally in rural Uganda (J. Ditai, Mudoola, et al., 2018). We will then

Study protocol version 1.5 protocol date 01 Dec 2019 31 generate an overall or combined short list of 20-25 top questions to be taken for final priority setting workshop. Duration: This exercise will take a duration of 8 weeks.

Step 6: Final Prioritisation workshop. Aim: To determine the top 10 research priorities for maternal and neonatal health in a low- income region in Uganda. Design: This will be a participatory mixed methods study. There will be a face-to-face group workshop for each category (maternal health and the neonatal health). The approach to involvement will be user controlled research (INVOLVE, 2012). Sample size: An average sample size of 27 participants will be invited by the steering group to take part in a one-day workshop. The JLA method recommends a minimum of 12 and a maximum of 30 patients, caregivers, and clinicians (Barnieh et al., 2014). Previous studies have used a sample size of 20 to 43 (Gadsby et al., 2012; Hall et al., 2013; Koblinsky et al., 2016; Van Middendorp et al., 2016; Victora et al., 2016). Each participant will be an ‘expert by lived experience’ in regard to maternal and neonatal health. The participants will include, 4 pregnant women, 3 postnatal mothers, 8 carers (husbands, mothers in law), 4 midwives, 2 obstetricians, 2 neonatal nurses and 1 paediatrician, Selection: The participants will be identified purposively at the time of initial survey. Procedure: The workshop will occur at the Sanyu Africa Research Institute, a central meeting place for all the sites. We will follow the standard JLA approach as described in detail in the JLA Guidebook (www.jla.nihr.ac.uk/ jla-guidebook/). The process will encourage open discussion and involvement of all group members guided by independent JLA chair and assisted by the principal investigator. The ranking exercises will be based on nominal group techniques (Cowan & Oliver, 2013) illustrated in figure 5. All participants will be made aware of the purpose of the workshop at the beginning, to generate the top 10 research priorities representing the views of all those who participated. The workshop will begin with small-group debate and discussion to challenge and explore the final questions on the short list before the final plenary ranking session.

Study protocol version 1.5 protocol date 01 Dec 2019 32 1. Introduce and explain

2. Participants 5. Aggregation of rank research rank order questions independently

4. Group Participants share discussion of rankings (round ranking table)

Figure 6: Nominal Group Technique

The JLA will facilitate this process and ensure transparency, accountability and fairness. Participants will be expected to declare their interests in advance of this meeting.

Duration: The entire exercise will occur for a period of 2 weeks, including one day of workshop, and most of the time taken up in report writing.

5.7 Dissemination of findings and research The steering group members will publish the findings of the Maternal and Newborn Health Priority Setting Partnership in Uganda using both internal and external communication mechanisms. The results of this study will be disseminated through the PSP website, the Sanyu Africa Research Institute website, twitter, WhatsApp and Facebook, peer reviewed publications, academic conferences, and through formal presentations to the stakeholders. It is anticipated that the findings of the Maternal and Newborn Health Priority Setting Partnership in Uganda PSP will be reported to funding and research agenda setting organisations such as the Medical Research Council (MRC), NIHR, Bill and Melinda Gates Foundation, Global Fund, as well as any other major research funding bodies. The JLA may also capture and publicise the results, through descriptive reports of the process itself.

Steering group members, with input from partners where appropriate, will work to develop the prioritised uncertainties into research questions, and to work to establish the research needs of

Study protocol version 1.5 protocol date 01 Dec 2019 33 those unanswered questions to use when approaching potential funders, or when allocating funding for research themselves, if applicable.

Study protocol version 1.5 protocol date 01 Dec 2019 34 6.0 Patient and Public Involvement 6.1 PPI in the PSP process The JLA method goes beyond patient and public involvement. The method is inclusive and

transparent. JLA simply assumes that people with lived experience and professionals are all

experts in their own right. We are not ‘involving patients’, we are working with all people as

equals, and in equal numbers. For purposes of reporting, table 5 summarises some of the

approaches and methods of public involvement demonstrated in this PSP.

Stage of JLA process Public involvement Public involvement Specific involvement approach method activities 1. Initial enquiry None N/A N/A 2. Identify and invite collaboration Patient representatives Recruit wider stakeholders partners 3. Inaugural steering collaboration PPI advisory group Meetings group meeting PSP advisory group Teleconferences 4. Initial awareness collaboration Workshop Stakeholders workshop meeting Pilot testing the data tool Comment on questions Launch initial survey 5. Initial survey to gather Consultation Individual interviews e-data capture (mobile) questions Focus groups Paper based survey 6. Data processing & User controlled SG review Meetings refining questions research Phone calls, WhatsApp Emails Zoom 7. Verifying indicative User controlled SG review Meetings questions research Phone calls, WhatsApp Emails Zoom 8. Presentation of raw User controlled SG review Meetings submissions for interim research Phone calls, WhatsApp survey Emails Zoom 9. Interim prioritisation Consultation Face to face survey e-data capture (mobile) Focus groups Paper based survey 10. Identify top 30 Collaboration Face to face meeting Analysis Zoom 11. Final priority setting Collaboration Workshop in Mbale Small groups Identify top 10 Plenary sessions 12. Next steps Collaboration Dissemination Checking data Quality assurance 13. Communication User controlled Dissemination stakeholders meeting research Preparation of newspaper article Table 5: approaches and methods of public involvement in maternal and Newborn health PSP in Uganda

Study protocol version 1.5 protocol date 01 Dec 2019 35 6.2 Evaluation of the patient and public involvement in the PSP in Uganda

The principal investigator will track and report on PPI processes within this research including the experiences and impact of public involvement from the perspectives of the steering group members, partners and other stakeholders. We will use an impact log for recording PPI outcomes during each SG meeting (appendix 12). Partners will also be encouraged to complete an evaluation form embedded in the impact log during the end of the initial awareness and dissemination meetings.

All participants in the initial and interim surveys will have their satisfaction assessed about the process of involving them to identify and rank questions respectively. During the final priority setting workshop, participants will complete an evaluation form.

Comparative analyses of the responses from mothers, carers and health workers will also be performed to determine the impact of different types of participants on proposed research priorities and establish how the maternal research priorities are ranked over neonatal health research priorities.

Study protocol version 1.5 protocol date 01 Dec 2019 36 7.0 Ethics and funding 7.1 Ethics The protocol endorsed by the steering group will be submitted to the Cure Children’s Hospital of Uganda Research ethics committee (CCHU-REC) in Uganda for approval as part of the main study (titled Patient and Public Involvement in maternal-neonatal health research in Uganda). Administrative clearance and registration will be sought from the Uganda National Council for Science and Technology in Uganda. Ethics and sponsorship approvals will be sought from the University of Liverpool. Written informed consent for participation in the study will be obtained from all participants before recruitment in line with the international GCP standards (ICH Harmonised Guideline, 2016). 6.2 Funding

The Maternal and Newborn Health Priority Setting Partnership in Uganda will be financially supported by European and Developing countries Clinical Trials Partnership (EDCTP) through a grant to the BabyGel cluster randomised trial (ref: RIA2017MC-2029-BabyGel). EDCTP has no any part in the execution of the study and write-up. University of Southampton has provided partial contribution towards the JLA adviser’s time. 6.3 Overall PSP Timeline

Study protocol version 1.5 protocol date 01 Dec 2019 37 Reference: Alkema, L., Chou, D., Hogan, D., Zhang, S., Moller, A., Gemmill, A., … Boerma, T. (2016). Global, regional, and national levels and trends in maternal mortality between 1990 and 2015 , with scenario-based projections to 2030: a systematic analysis by the UN Maternal Mortality Estimation Inter-Agency Group. The Lancet, 387(10017), 462–474. https://doi.org/10.1016/S0140-6736(15)00838-7 Aranda-jan, C. B., Jagtap, S., & Moultrie, J. (2016). Towards A Framework for Holistic Contextual Design for Low-Resource Settings. International Journal of Design, 10(3), 43–63. Baker, P. N., & Kenny, L. C. (2011). Obstetrics by Ten Teachers (19th ed). https://doi.org/10.1136/pgmj.57.669.472 Barnieh, L., Jun, M., Laupacis, A., Manns, B., & Hemmelgarn, B. (2014). Determining Research Priorities Through Partnership with Patients: An Overview. Seminars in Dialysis, 28(2). https://doi.org/10.1111/sdi.12325 Boote, J., Wong, R., & Booth, A. (2012). “Talking the talk or walking the walk?” A bibliometric review of the literature on public involvement in health research published between 1995 and 2009. Health Expectations, 18(1), 44–57. https://doi.org/10.1111/hex.12007 Chalmers, I. (1991). The perinatal research agenda: whose priorities?... including discussion. Birth: Issues in Perinatal Care, 18(3), 137-145 9p. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=rzh&AN=107508706&site=eho st-live Chersich, M. F., & Martin, G. (2017). Priority gaps and promising areas in maternal health research in low- and middle-income countries: summary findings of a mapping of 2292 publications between 2000 and 2012. Globalization and Health, 13(1), 6. https://doi.org/10.1186/s12992-016-0227-z Cowan, K., & Oliver, S. (2013). James Lind Alliance Guidebook (Version 5) (James Lind Alliance, ed.). Retrieved from www.jla.nihr.ac.uk Creswell, J. W. (2014). Research design: qualitative, quantitative, and mixed methods approaches (4th ed). Washington, DC: SAGE Publications. Ditai, J., Abeso, J., Odeke, N. M., Mobbs, N., Dusabe-Richards, J., Mudoola, M., … Weeks, A. D. (2019). BabyGel pilot: a pilot cluster randomised trial of the provision of alcohol handgel to postpartum mothers to prevent neonatal and young infant infection-related morbidity in the community. Pilot and Feasibility Studies, 5(49).

Study protocol version 1.5 protocol date 01 Dec 2019 38 https://doi.org/10.1186/s40814-019-0432-7 Ditai, J., Kanyago, J., Nambozo, M. R., Odeke, N. M., Abeso, J., Dusabe-Richards, J., … Weeks, A. D. (2018). Optimising informed consent for participants in a randomised controlled trial in rural Uganda: a comparative prospective cohort mixed-methods study. Trials, 19(1), 699. https://doi.org/10.1186/s13063-018-3030-8 Ditai, J., Mudoola, M., Gladstone, M., Abeso, J., Dusabe-Richards, J., Adengo, M., … Weeks, A. D. (2018). Preventing neonatal sepsis in rural Uganda: a cross-over study comparing the tolerance and acceptability of three alcohol-based hand rub formulations. BMC Public Health, 18(1), 1279. https://doi.org/10.1186/s12889-018-6201-3 Ditai, James, Mudoola, M., Gladstone, M., Abeso, J., Dusabe-Richards, J., Adengo, M., … Weeks, A. D. (2018). Preventing neonatal sepsis in rural Uganda: a cross-over study comparing the tolerance and acceptability of 3 alcohol-based hand rub formulations. BMC Public Health, 18(1279), 1–10. https://doi.org/10.1186/s12889-018-6201-3 Edmonds, D. K. (2008). Dewhurst’s Textbook of Obstetrics & Gynaecology (7th ed). https://doi.org/10.1002/9780470753354 Elder, J. P., Lytle, L., Sallis, J. F., Young, D. R., Steckler, A., Simons-morton, D., … Ribisl, K. (2007). A description of the social–ecological framework used in the trial of activity for adolescent girls (TAAG). Health Educ Res, 22(2), 155–165. https://doi.org/10.1093/her/cyl059.A F. Gary Cunningham, John C. Hauth, Kenneth J. Leveno, Larry Gilstrap Iii, Steven L. Bloom, & Katharine D. Wenstrom. (2010). Williams obstetrics. In Williams Obstetrics. https://doi.org/10.1097/00001888-193609000-00027 Fishburn, P. C. (1968). UTILITY THEORY. Management Science, 14(5), 335–378. Retrieved from http://www.jstor.org/stable/2628674 Gadsby, R., Snow, R., Daly, A. C., Crowe, S., Matyka, K., Hall, B., & Petrie, J. (2012). Setting research priorities for Type 1 diabetes. Diabetic Medicine, 29(10), 1321–1326. https://doi.org/10.1111/j.1464-5491.2012.03755.x Glasziou, P., Mar, C. Del, & Salisbury, J. (2007). Evidence-based practice workbook: bridging the gap between health care ... (2nd ed.). Retrieved from http://books.google.com/books?id=Avvs2DgX0QEC&pgis=1 Graham, W., Woodd, S., Byass, P., Filippi, V., Gon, G., Virgo, S., … Singh, S. (2016). Diversity and divergence: the dynamic burden of poor maternal health. The Lancet. https://doi.org/10.1016/S0140-6736(16)31533-1 Green, J., & Thorogood, N. (2009). Qualitative Methods for Health Research (2nd ed).

Study protocol version 1.5 protocol date 01 Dec 2019 39 London: SAGE Publications. Guyatt, G., Oxman, A. D., Akl, E. A., Kunz, R., Vist, G., Brozek, J., … Sch€ Unemann, H. J. (2011). GRADE guidelines: 1. IntroductiondGRADE evidence profiles and summary of findings tables. Journal of Clinical Epidemiology, 64, 383–394. https://doi.org/10.1016/j.jclinepi.2010.04.026 Hall, D. a, Mohamad, N., Firkins, L., Fenton, M., & Stockdale, D. (2013). Identifying and prioritizing unmet research questions for people with tinnitus: the James Lind Alliance Tinnitus Priority Setting Partnership. Clinical Investigation, 3(1), 21–28. https://doi.org/10.4155/cli.12.129 ICH Harmonised Guideline. (2016). Integrated Addendum to ICH E6(R1): Guideline for Good Clinical Practice E6(R2). International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH), Step 4(November). INVOLVE. (2012). Briefing notes for researchers: public involvement in NHS, public health and social care research. Retrieved from http://www.invo.org.uk/wp- content/uploads/2012/03/INVOLVEBriefingnotesforresearchersMarch2012.pdf Kassebaum, N. J., Barber, R. M., Dandona, L., Hay, S. I., Larson, H. J., Lim, S. S., … Zuhlke, L. J. (2016). Global, regional, and national levels of maternal mortality, 1990– 2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet, 388(10053), 1775–1812. https://doi.org/10.1016/S0140-6736(16)31470-2 King, D. S. (1987). Social Choice Theory. Political Studies, 35(2), 301–310. https://doi.org/10.1111/j.1467-9248.1987.tb01890.x Koblinsky, M., Moyer, C. A., Calvert, C., Campbell, J., Campbell, O. M. R., Feigl, A. B., … Doherty, T. (2016). Quality maternity care for every woman, everywhere: a call to action. The Lancet, 0(0), 189–205. https://doi.org/10.1016/S0140-6736(16)31333-2 Lissauer, D., Wilson, A., Daniels, J., Middleton, L., Bishop, J., Hewitt, C., … Coomarasamy, A. (2018). Prophylactic antibiotics to reduce pelvic infection in women having miscarriage surgery - The AIMS (Antibiotics in Miscarriage Surgery) trial: study protocol for a randomized controlled trial. Trials, 19(1), 245. https://doi.org/10.1186/s13063-018-2598-3 Monga, A. (2006). Gynecology by ten teachers (18th ed). Southampton: UK: Hodder Arnold. O’Loughlin, J. (1997). Safe motherhood: impossible dream or achievable reality? The Medical Journal of Australia, 167(December), 622–625. Pollock, A., George, B. S., Fenton, M., Crowe, S., & Firkins, L. (2014). Development of a new model to engage patients and clinicians in setting research priorities. Journal of

Study protocol version 1.5 protocol date 01 Dec 2019 40 Health Services Research & Policy, 19(1), 12–18. https://doi.org/10.1177/1355819613500665 Rees, S. E., Chadha, R., Donovan, L. E., Guitard, A. L. T., Koppula, S., Laupacis, A., … Johnson, J. A. (2017). Engaging Patients and Clinicians in Establishing Research Priorities for Gestational Diabetes Mellitus. Canadian Journal of Diabetes, 41(2), 156– 163. https://doi.org/10.1016/j.jcjd.2016.08.219 Romano, M., Cacciatore, A., Giordano, R., & La Rosa, B. (2010). Postpartum period: three distinct but continuous phases. Journal of Prenatal Medicine, 4(2), 22–25. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22439056 Roos, N., & von Xylander, S. R. (2016). Why do maternal and newborn deaths continue to occur? Best Practice and Research Clinical Obstetrics and Gynaecology, 36, 30–44. https://doi.org/10.1016/j.bpobgyn.2016.06.002 Rudan, I., Gibson, J., Kapiriri, L., Lansang, M. A., Hyder, A. A., Lawn, J., … Black, R. E. (2007). Setting Priorities in Global Child Health Research Investments: Assessment of Principles and Practice. Croat Med J, 48, 595–604. Rudan, I., Gibson, J. L., Ameratunga, S., Arifeen, S. El, Bhutta, Z. A., Black, M., … Tsai, A. C. (2008). Setting Priorities in Global Child Health Research Investments: Guidelines for Implementation of CHNRI Method. Croat Med J, 49, 720–733. https://doi.org/10.3325/cmj.2008.49.720 Rudan, I., Kapiriri, L., Tomlinson, M., Balliet, M., Cohen, B., & Chopra, M. (2010). Evidence-Based Priority Setting for Health Care and Research : Tools to Support Policy in Maternal , Neonatal , and Child Health in Africa. PLoS Medicine, 7(7). https://doi.org/10.1371/journal.pmed.1000308 Salihu, H. M., Wilson, R. E., King, L. M., Marty, P. J., & Whiteman, V. E. (2015). Socio- ecological Model as a Framework for Overcoming Barriers and Challenges in Randomized Control Trials in Minority and Underserved Communities. International Journal of MCH and AIDS, 3(1), 85–95. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/27621990%0Ahttp://www.pubmedcentral.nih.gov /articlerender.fcgi?artid=PMC4948176 Smith, J. A., & Osborn, M. (2015). Interpretative phenomenological analysis as a useful methodology for research on the lived experience of pain. British Journal of Pain, 9(1), 41–42. https://doi.org/10.1177/2049463714541642 The James Lind Alliance Priority Setting Partnerships. (2018). The James Lind Alliance Guidebook Version 8 (version 8). Retrieved from www.jla.nihr.ac.uk

Study protocol version 1.5 protocol date 01 Dec 2019 41 Tomlinson, M., Chopra, M., Hoosain, N., & Rudan, I. (2011). A review of selected research priority setting processes at national level in low and middle income countries : towards fair and legitimate priority setting. Health Research Policy and Systems, 9(19), 1–7. https://doi.org/10.1186/1478-4505-9-19 Uganda Bureau of Statistics (UBOS) and ICF. (2018). Uganda Demographic and Health Survey 2016. : Uganda and Rockville, Maryland, USA: UBOS and ICF. United Nations. (2018). The Sustainable Development Goals Report 2018. https://doi.org/10.18356/4d038e1e-en Van Middendorp, J. J., Allison, H. C., Ahuja, S., Bracher, D., Dyson, C., Fairbank, J., … Cowan, K. (2016). Top ten research priorities for spinal cord injury: The methodology and results of a British priority setting partnership. Spinal Cord, 54(5), 341–346. https://doi.org/10.1038/sc.2015.199 Victora, C. G., Requejo, J. H., Barros, A. J. D., Berman, P., Bhutta, Z., Boerma, T., … Chopra, M. (2016). Countdown to 2015: a decade of tracking progress for maternal, newborn, and child survival. Lancet (London, England), 387(10032), 2049–2059. https://doi.org/10.1016/S0140-6736(15)00519-X Wallace, L. J., & Kapiriri, L. (2019). Priority setting for maternal, newborn and child health in Uganda: a qualitative study evaluating actual practice. BMC Health Services Research, 19(1), 465. https://doi.org/10.1186/s12913-019-4170-6 Weeks, A. D., Ditai, J., Ononge, S., Faragher, B., Frye, L. J., Durocher, J., … Alfirevic, Z. (2015). The MamaMiso study of self-administered misoprostol to prevent bleeding after childbirth in rural Uganda: A community-based, placebo-controlled randomised trial. BMC Pregnancy and Childbirth, 15(1). https://doi.org/10.1186/s12884-015-0650-9 WHO. (2006). Constitution of the World Health Organisation: Basic Documents, Forty-fifth edition, Supplement, (pp. 1–18). pp. 1–18. https://doi.org/12571729 World Health Organisation. (2013). WHO recommendations on Postnatal care of the mother and newborn. Retrieved from https://apps.who.int/iris/bitstream/handle/10665/97603/9789241506649_eng.pdf?sequen ce=1 World Health Organisation. (2017a). Managing Complications in Pregnancy and Childbirth: A guide for midwives and doctors (2nd ed). Retrieved from https://apps.who.int/iris/bitstream/handle/10665/255760/9789241565493- eng.pdf?sequence=1 World Health Organisation. (2017b). WHO Recommendations on maternal health:

Study protocol version 1.5 protocol date 01 Dec 2019 42 Guidelines approved by the WHO Guidelines review committee. Retrieved from https://apps.who.int/iris/bitstream/handle/10665/259268/WHO-MCA-17.10- eng.pdf;jsessionid=924847FACC8C438384135B38947339E2?sequence=1 World Health Organisation. (2017c). WHO recommendations on newborn health guidelines approved by the WHO guidelines review committee. Retrieved from https://apps.who.int/iris/bitstream/handle/10665/259269/WHO-MCA-17.07- eng.pdf?sequence=1 World Health Organization. (2015a). Pregnancy, Childbirth, Postpartum and New Born Care: A Guilde for Essential Practice (3rd ed). https://doi.org/10.1017/CBO9781107415324.004 World Health Organization. (2015b). Trends in maternal mortality: 1990 to 2015: Estimates by WHO, UNICEF, UNFPA, World Bank Group and the United Nations Population Division. Retrieved from http://www.afro.who.int/sites/default/files/2017-05/trends-in- maternal-mortality-1990-to-2015.pdf Yoshida, S. (2016). Approaches, tools and methods used for setting priorities in health research in the 21. Journal of Global Health, 6(1). https://doi.org/10.7189/jogh.06.010507 Youngkong, S., Kapiriri, L., & Baltussen, R. (2009). Review Setting priorities for health interventions in developing countries : a review of empirical studies. Tropical Medicine & International Health : TM & IH, 14(8), 930–939. https://doi.org/10.1111/j.1365- 3156.2009.02311.x

Study protocol version 1.5 protocol date 01 Dec 2019 43