Persson 2013.
| Study characteristics | ||
| Methods |
Study design: Cluster‐RCT; Unit of randomisation: commune (geopolitical sites)
Sample size/power calculation: "The sampling strategy was one‐stage cluster sampling with a probability proportional to size of the clusters. The Probability proportional to size, in this case number of births per year, was chosen to obtain similar distribution of sizes of clusters across intervention and control communes. Sampling was neither blocked nor paired. A sampling frame was established with a cumulative list of number of births in each of the communes in the 2005 survey. ... The design effect was arbitrarily estimated to be 1.5 [taking into account] the high number clusters and low average cluster size. A 3‐year sample would allow detection of 7/1000 neonatal mortality rate with 80% at 0.05 significance". (p 2‐3, Persson 2013) Country of study: Quang Ninh province, Vietnam (Middle‐income country) Setting: Commune Health Centres in with 3‐6 staff members provide primary health care including reproductive and antenatal care to approximately 1000‐18,000 people. Delivery care is offered by Commune Health Centres, or by hospital at district, province and regional levels. Each Commune Health Centre has a Village (community) Health Worker who provides basic healthcare in the villages. Private alternatives for antenatal care in addition to governmental health care system exist but in this province the private or non‐governmental sector plays a limited role in relation to delivery services. Study duration: 2008 ‐ 2011 Broader consumer involvement: Members from the Women's Union were also members of the intervention steering board (chaired by the hospital director of the regional Uong Bi hospital and the Provincial Health Bureau) |
|
| Participants |
Consumer partnership participants: (n = 11), Pairs of lay women were facilitators for the 44 Maternal and Newborn Health Groups (MNHGs). Each facilitator was responsible for 5‐8 MNHGs. Recruitment criteria included being an experienced Women's Union member, having completed secondary school and having children. Facilitators were recruited using local newspaper adverts and recommendations from communes. “Originally the Women's Union office in each of the 8 study districts selected two individuals among the applicants for further interview [by two researchers and the chairwomen of the Women's Union]… Out of the 16 applicants interviewed 8 were selected for further training" (p 5, Eriksson 2016). Three additional facilitators were recruited during the trial to replace facilitators leaving because of childbirth or new employment. *The Women's Union is an organisation with high national coverage, working on a range of welfare issues relevant for women in Vietnam, especially those related to healthcare. Data below relates to only the facilitators (not the Women's Union members of the MNHGs)
Provider partnership participants: (n = 352; each of the 44 MNHGs had 8 members), Commune Health Centre staff (physicians, midwifes, nurses); a commune Village Health Worker, a population collaborator, a chairperson/vice chairperson of the commune; and two Women's Union representatives
Health service user trial participants: (n = 1243 mothers randomly sampled from 22,377 live births located within 90 communes during first 3 years of trial July 2008‐June 2011). The demographic details below were from a random sample of mothers (398/7033) with live births during first year of trial. ("A 6% random sample of all registered live births, surviving the neonatal period, was continuously selected (each month) in order to represent the entire birth cohort") (p 4, Persson 2013). The demographic details below are for Intervention: n = 213; Control: n = 185 mothers
Health service provider trial participants: none |
|
| Interventions |
Randomised to intervention: 44 communes; analysed during years 1‐3: 656 mothers randomly sampled from 11,818 live births in 44 communes. Nature of intervention: laywomen facilitation of MNHGs during which members collaborated to decide which problems to focus on and what actions to take directed towards pregnant women and their households, health services, or general public in order to address those problems. Intervention aim: to reduce neonatal mortality and improve maternal delivery, and newborn care indicators. Stillbirth was defined as birth of a dead foetus after an estimated 28 weeks of gestation. Live birth was defined as birth of a foetus with any sign of viability. Neonatal death was defined as death of a live birth during the first 28 days of life. Context of partnering: "MNHGs were constituted in each intervention commune (by directives from the Provincial Health Bureau). ...The facilitators primarily used the plan‐do‐act cycle in mobilising the groups in identifying and prioritising local perinatal health problems and accomplishing improvement cycles that included concrete actions on prioritized problems directed towards pregnant women and their households, the health services, or the general public. Such improvement cycles on different problems were performed continuously over the intervention period in all MNHGs. Where possible one facilitator performed monthly meetings with the same MNHG over the 3 year period. Each facilitator was responsible for 5‐8 MNHGs. Twenty groups kept the same facilitator the whole period, 22 changed facilitator once, and two groups changed facilitator twice. When appropriate, facilitators were recommended to highlight recommendations provided by the Vietnamese National Standard and Guidelines on Reproductive Health Care Services." (p 4, Persson 2013) Decision‐making activity: In MNHGs Plan‐Do‐Study‐Act discussions centred on individual and common experiences in the local setting, the facilitator supported the group in critical reflection, problem identification, finding solutions and developing change strategies. The intervention aimed to achieve local ownership and "bottom‐up" approach in empowering healthcare staff to improve practice. When appropriate, the facilitators would highlight recommendations in the National Guidelines. "The groups identified 32 unique problems and implemented 39 unique actions. The identified problems targeted health issues concerning both women and neonates. Actions implemented were mainly communication activities." (p 1, Eriksson 2016) Meeting format, duration, frequency and location: Monthly face‐to‐face meetings lasting on average 2 hrs (110 mins), meetings located at commune centre or health care centre. Partnership duration: 3 years (31 months) Training/support: The facilitators were trained in problem solving, participatory and enabling approach in a 10 day training program which "included theoretical sessions, group discussion, role‐plays, and field practice. It covered topics such as group dynamics, quality improvement methods (e.g. brainstorming and the plan‐do‐study‐act cycle); and evidence based perinatal care. A facilitation manual and a specific diary were developed to guide the work of the facilitators. Two research team members coordinated the facilitation process and acted as supervisors of the facilitators, i.e. field supervision and performing 2 day meetings with all facilitators once a month during the entire trial period." (p 3‐4, Persson 2013) Decision‐making process, attempts to resolve conflict: facilitators were trained on brainstorming, the nominal group technique, the plan‐do‐study‐act cycle and the strengths‐weaknesses‐opportunities‐threats diagnostic tool. Diversity and ratio of consumer and provider participants: MNHGs consisted of 8 members: 3 Commune Health Centre staff (physician, midwife, nurse); a commune Village Health Worker, a population collaborator, the chairperson/vice chairperson of the commune; and two Women's Union representatives (from village and commune levels). Attempts to address intrinsic power imbalances: The MNHG facilitators were paid on a full‐time basis for the 3 years of the intervention. Except for the Village Health Worker and the Women's Union employee (who were reimbursed travel expenses to and from the meetings), other members were neither paid nor received allowances ‐ as implementation work was assumed to be part of their normal duties. Theoretical basis for partnering: study draws from “the Promoting Action on Research Implementation in Health Services Model. This middle range theory highlights three major ingredients for being successful in implementing research into practice: 1) the nature of the evidence being used, 2) the quality of the context in terms of coping with changes, and 3) the types of facilitation needed to ensure a successful change process. Implementation is conceived as a multifaceted intervention, rather than a more straightforward, linear process of translating knowledge from experts to the local level. In the trial the authors analysed the effect of facilitation of local stakeholder groups focusing maternal and neonatal health problems and actions." (p 7, Persson 2013) Tailoring/modification/adapting: Each MNHG was context‐specific and continuously negotiated and interpreted among stakeholders. Fidelity/integrity: "Two research team members coordinated the facilitation process and acted as supervisors of the facilitators; i.e. field supervision and performing 2 day meetings with all facilitators once a month during the entire trial period. The intervention process was monitored continuously. Issues like MNHG's choice of improvement topics, activities for improving practice, the interaction between facilitators and group members, and progress of the facilitation process at all intervention sites were examined using several approaches, like interviews with facilitators and focus group discussions with MNHG members, analyses of facilitators diaries from MNHG meetings and the notes from the monthly meetings with the supervisors." (p 5, Eriksson 2016); 1508 (out of 1584; 95%) of the planned MNHG meetings with a facilitator were completed; during the intervention a facilitator joined the MNHG activity in a commune 294 times; a facilitator supported a co‐facilitator at a MNHG meeting 122 (8%) of the time; a NeoKIP researcher supported a facilitator at a 95 MNHG meetings (6%); and 35 (97%) monthly meetings between supervisor and facilitators took place (p 12, Eriksson 2016). Overall attendance at the meetings was 86%; the Head of the Women's Union (village level) attended 97%; the Head of the Community Health Centre attended 95%; midwives attended 94%; Community (Village) Health Workers attended 90%; Nurse attended (88%); Population collaborator (87%); Chairwomen of the Women's Union (commune level) (87%); and Vice chairperson of peoples committee attended 61% (p 7, Eriksson 2016). "No deviations from the protocol were observed except from one out of 44 MNHGs that ended the facilitation intervention two‐thirds of the way into the trial. The facilitated intervention with MNHGs maintained a high activity with a large number of problems identified and actions taken, in spite of no extra financial benefits to the group members" (p 6, Eriksson 2016). "However, the NeoKIP intervention was a new approach for local stakeholders, who were not used to collaborate in this kind of group activity. This type of approach requires active and disciplined stakeholders who assess, discuss, and find ways to overcome contextual barriers that may impede the process of implementation." (p 8, Eriksson 2016) Randomised to control: 46 communes; analysed during years 1‐3: 587 mothers randomly sampled from 10,559 live births in 46 communes Nature of comparison: usual practice; baseline perinatal care comprised the following: 'At primary care level communal health centres (CHC) provide antenatal care (ANC), delivery services and postnatal care staffed by 3–6 persons (physician, nurses, midwife). The primary health care level is also supported by at least one part‐time village health worker (VHW) in each village, who mainly takes part in preventive services. Delivery services are also offered at hospitals at district, provincial and regional levels with more skilled staff and resources available at the higher levels' (p.1479, Nga 2010) Most women (76%) received 3+ ANC visits, in line with national guidelines. 92% of births occurred in health care facilities (regional, provincial or district hospitals, community health centres); 8% occurred at home (accounting for 20% of neonatal deaths). Co‐intervention: none |
|
| Outcomes |
Outcomes measured at Baseline (T0) and follow‐up (during first year of intervention ‐ T1; second year T2; third year T3; sixth year T4): Outcome measures relevant to the review (reporting available comparative data for longest time point measured):
Other study outcome measures (data extracted but not reported in review):
|
|
| Notes |
Funding: Swedish International Development Cooperation Agency (SIDA), Swedish Research Council and Uppsala University. Conflict of interest: Authors have declared that no competing interests exist. Other: For other outcomes reported by the trial but not reported in the review, these were judged, through author consensus, to be outside the scope of the selection criteria (e.g. neonatal mortality) and/or less relevant measures of an outcome for which multiple measures were reported. |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | "A random number list was used to subsequently allocate "intervention" or "control" to the list of communes, and 44 out of the 90 communes were allocated to intervention and 46 to control. The randomisation was preformed by one of the involved researchers at Uppsala University." (p 3, Persson 2013). |
| Allocation concealment (selection bias) | Low risk | "The sequence generation was concealed until the intervention was assigned; otherwise allocation was not masked." (p 3, Persson 2013) |
| Blinding of participants and personnel (performance bias) All outcomes | High risk | Unclear whether study participants were blind from knowledge of which intervention cluster they were assigned to; not able to blind personnel from knowledge of which intervention cluster they were assigned to. |
| Blinding of outcome assessment (detection bias) All outcomes | Low risk | No description of measures used, if any, to blind outcome assessors from knowledge of which intervention a participant received; The following relates to the outcome neonatal births/mortality "One (out 10) of the Village health Workers in the commune were involved in the Maternal and Newborn Health groups while all Village Health Workers were informants in the data collection system. The information on births and/or neonatal deaths provided by the village health workers who were involved in the intervention was most likely not biased since the updated list of pregnant women enabled a systematic enquiry on a monthly basis on pregnancy outcomes, as well as a triangulation of information sources and cross‐checking of data." (p 6, Perrson 2013) "Six data collectors were trained for 2 weeks. They attended monthly meetings at all commune health centres, all district hospitals, and the two provincial level hospitals in the area to collect data on live births and neonatal deaths. Further the data collectors met all Village Health Workers in the communes to collect data on live births and neonatal deaths in their villages. Triangulation was systematically performed of live births and neonatal deaths from all included sources (records and reports from the district, provincial and regional hospitals, commune health centres and village health workers) to ascertain that all data were registered and secure that no duplication of information occurred." (p 777, Eriksson 2018). Author contact confirms that data collectors were blinded to group allocation and had no contact with those involved in intervention delivery. |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Zero losses to follow‐up. |
| Selective reporting (reporting bias) | Low risk | Outcomes appear to be reported as planned based on those identified in protocol. |
| Other bias | Low risk | Groups were mostly balanced at baseline (except slightly more poor households in intervention group). Data collection was separated and verified by triangulation from different sources (and by analysts unaware of group assignment for at least some outcomes). |
| Selective recruitment of participants | Unclear risk | Not sure when participants were recruited to clusters. |