O'Connor 2019.

Study characteristics
Methods	Study design: Cluster‐RCT; Unit of randomisation: community sites Unclear how many assessed for eligibility Unclear how many eligible Unclear how many not randomised 10 community sites randomised Sample size/power calculation: "The sampling methodology provided a 95% confidence interval of +/‐8% or less for the prevalence of indicators measure in the population." (p 7, O'Connor 2019) "The number of CHWs trained was calculated using a Child Survival Project‐ and community‐led census and policy‐mandated population‐to‐CHW ratios." (p 4, O'Connor 2019) Country of study: Freetown, Western Urban District, Sierra Leone Setting: Low‐income Urban slum communities within Freetown Municipal area with the highest maternal mortality ratio in the world (136/10,000 live births) and under‐5 mortality rate was fourth highest in the world (114/1000 live births) (p 3, O'Connor 2019). Communities are served by government primary health care facilities (Peripheral Health Units) at the sub‐district level and these are supervised by District Health Management Team (consisting of 15 members) which is also responsible for coordinating public health interventions in the community, the Freetown City Council also supports the delivery of community health services. Each Peripheral Health Unit is supported by the Health Management Committee which liaises between the service and the community. Each community spans 1‐2 wards. Wards within Freetown Municipal area have an elected Ward Development Committee responsible for engaging community members on general development activities; the chair of the Ward Development Committee is also a Councillor on the Freetown City Council. Study duration: 2011 ‐ 2017 Broader consumer involvement: None described
Participants	Consumer partnership participants: (n = 49), peer supervisors of CHWs participating in Community Health Data Review (CHDR) meetings, each community had 5‐12 Peer Supervisors. Peer Supervisors were selected by Child Survival Project staff together with community leaders, based on their performance during initial CHW training. CHW were selected by a community‐level process according to the National CHW Program selection criteria. Criteria included currently living in and have close connections to their geographical community; accepted by community members, and ability and motivation to serve their community. Note that the data presented here is overall for all CHW (not just Peer Supervisors) Age: about two thirds were 18‐34; 21% 35‐54; and 4% 55+ years Sex: 46% female; 54% male Education level: Almost 60% completed some secondary school Socioeconomic status (SES): not described Diagnosis: not described Other: 5% classified themselves as non‐literate Provider partnership participants: (n = unknown 30‐50 participants in each of the 10 CHDR meetings; includes Peer Supervisors), Peripheral Health Unit (government primary health care facility) staff; Health Management Committee members, and Ward Development Committee members participating in CHDR meetings. Age: not described Sex: not described Education level: not described SES: not described Years practising: not described Other: not described Health service user trial participants: (n = 599 at baseline; 792 at 21 months), household respondents were pregnant women and mothers of children under 5 years targeted by CHW. Household respondents were selected randomly from 10 random households in each of the 60 clusters. Clusters were also randomly selected using a probability proportional to size methodology based on population projections from the most recent Government of Sierra Leone census. Age: not described Sex: not described Education level: not described SES: not described Other: not described Health service provider trial participants: none
Interventions	Randomised to intervention: 5 communities. Baseline data were collected from: 299 household respondents; 509 Community Health Workers (CHW); 49 Peer Supervisors; 80 Ward Development Committees and 75 Health Management Committees. Follow‐up data were collected from: 379 household respondents; 509 CHW; 49 Peer Supervisors; 80 Ward Development Committees and 75 Health Management Committees. Nature of intervention: In the Operation Research Study the Participatory Community‐based Health Information System intervention consisted of Community Health Data Review meetings every two months to support Health Management Committees, Ward Development Committees and CHW Peer Supervisors to undertake two activities (1. review of data collected by CHWs and actions in response to this data and 2. after 6 months, verbal autopsies for deaths under 5 years) in addition to the Child Survival Project activities. Intervention aim: “to 1) assess the extent to which the Participatory Community‐based Health Information System facilitated local community structures to use data to plan and implement actions for improving maternal, neonatal and child health and 2) assess the extent to which this contributed to improved community‐level maternal, neonatal and child health outcomes." (p 4, O'Connor 2019) Context of partnering: Findings from monthly reports submitted by volunteer community health workers and verbal autopsy findings for deaths of children aged less than 5 years were processed and shared at Community Health Data Review (CHDR) meetings in each intervention community. These bimonthly meetings were attended by community leaders, including members of the Ward Development Committee and Health Management Committee, the Peer Supervisors, and representatives of the Peripheral Health Unit. Following a review of the information, attendees proposed actions to strengthen community‐based health services in their community. Decision‐making activity: Typically, data for the preceding 4‐6 months were reviewed in CHDR meeting. The Operation Research Study staff analysed CHW collected data with the Child Survival Project staff prior to the meeting and determined topics and data to present in CHDR meetings. "The Operation Research Study staff prepared simple data sheets to be used by participants, and participants used them to draw and interpret bar charts in front of the group. Records were kept of discussion topics." (p5, O'Connor 2019) "Following the review of data, CHDR participants developed action points. Action points were documented during the meeting on flip chart paper which the Health Management Committee chairman kept after the meeting. Action points from previous meetings were reviewed in subsequent meetings and discussions held on the extent to which actions had been completed." (p 6, O'Connor 2019) Meeting format, duration, frequency and location: face‐to‐face, duration unknown, every two months, location unclear Partnership duration: 20 months Training/support: Based on their performance during initial training, community leaders and Child Survival Project staff together selected 106 Peer supervisors from the CHWs receiving training. Peer supervisors were given additional training by project staff and were assigned 8‐12 CHWs to supervise. At least one Health Management Committee and Ward Development Committee member from the same zone as the Peer Supervisor provided oversight and assistance. Decision‐making process, attempts to resolve conflict: not described Diversity and ratio of consumer and provider participants: Each community had 5‐12 Peer Supervisors and there were 30‐50 participants in Community Health Data Review meetings. Generally, the same District Health Management Committee members, Ward Development Committee members and Peer Supervisors attended each meeting. Although Government primary health care facility (Peripheral Health Unit) In‐Charges rarely attended meetings, they generally sent the same representative to each meeting. At the beginning, Peripheral Health Unit staff attendance was not strong, but attendance improved after Health Management Committee members engaged Peripheral Health Unit staff. Ratio of consumer and provider participants not described. Attempts to address intrinsic power imbalances: not described Theoretical basis for partnering: Sierra Leone Government highlights the need for community engagement in policy document "Basic Package of Essential Health Services, 2015‐2020", it focus on CHWs to fulfil this role. The policy document also recognises Health Management Committees (that support each Peripheral Health Unit) and Ward Development Committees (that are responsible for engaging community members on general development activities and the chair is also a Councillor on the Freetown City Council) but does not outline roles or the ways in which they should fit into the health system. Tailoring/modification/adapting: not described Fidelity/integrity: not described Randomised to control: 5 communities. Baseline data were collected from: 300 household respondents, 710 CHW; 57 Peer Supervisors; 75 Ward Development Committees and 75 Health Management Committees. Follow‐up data were collected from: 413 household respondents, 710 CHW; 57 Peer Supervisors; 75 Ward Development Committees and 75 Health Management Committees Nature of comparison: multifaceted intervention minus partnership: control sites of the Operation Research Study consisted of usual practice in addition to Child Survival Project activities Co‐intervention: The Concern Worldwide Child Survival Project was implemented in all 10 communities. In the broader Child Survival Project, 1325 volunteer CHWs were recruited and trained with the Ministry of Health and Sanitation 2012 National CHW Program training materials by Child Survival Project staff and the Western Area District Health Management Teams. CHWs were assigned 25 households to visit monthly and disseminate health messages, check for danger signs of illness and collect vital event and morbidity data using Ministry of Health and Sanitation registers.
Outcomes	Outcomes measured at Baseline (T0) and follow‐up (21 months post‐intervention ‐ T1): Outcome measures relevant to the review (reporting available comparative data for longest time point measured): Health service alterations data (changes to services resulting from partnership decisions): Number of mothers who have ever had a community health worker (CHW) visit (T1); and Number of mothers who had a home health visit from a CHW in the last year in which the CHW performed all roles (T1) Health service user (trial participant) health service performance ratings (local accountability): Number of mothers who had a CHW visit in the past year who found the visit helpful or somewhat helpful (T1) Consumer (partnership participant) reported behaviours/attitudes outcomes: Number of peer supervisors reporting/number of peer supervisors trained (T1) Provider (partnership participant) reported behaviours/attitudes outcomes: Health Management Committee review and contribute to CHW activity plans (T1) Other study outcome measures (data extracted but not reported in review): Health Institution capacity Assessment Process (assessed separately by Ward Development Committees and Health Management Committees); Measures of functionality of the CHW program; CHW and Peer Supervisory reporting rates; Key household level survey results on Maternal, Neonatal and Child health practices; and Household survey results on illness care seeking.
Notes	Funding: United States Agency for International Development (USAID); Irish Aid and Concern Worldwide. Conflict of interest: "The authors completed the Unified Competing Interests Form … and declare no competing interests" Other: For other outcomes reported not reported further in this review, choice of outcomes was made jointly by review authors according to measure(s) most representative of the outcome concept(s) sought. Data for other outcomes was reported by the trial but not extracted and considered further for analysis in this review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"The study was a cluster‐randomised controlled trial, with [10 previously selected] communities randomly selected [assigned] to either the intervention or comparison area. Within the 10 communities 30 clusters in the intervention and comparison areas each were selected at random using a probability proportional to size methodology based on population projections from the most recent Government of Sierra Leone census. Ten interviews were conducted in each cluster through a random selection of households and a random selection of the respondent within the household." (p 7, O'Connor 2019).
Allocation concealment (selection bias)	Unclear risk	Method use to conceal the allocation sequence not described.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No information on blinding of participants or personnel provided.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information on blinding of outcome assessors.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No information about attrition or exclusions were reported.
Selective reporting (reporting bias)	Unclear risk	No protocol identified; all results for all outcomes mentioned in trial reported.
Other bias	Unclear risk	There was substantial movement between communities (in and out).
Selective recruitment of participants	Unclear risk	No information on recruitment bias (differential participant recruitment in clusters)/loss of clusters.