Soofi 2017.
| Methods | Community‐based cluster‐randomised trial in rural Pakistan | |
| Participants | All households in the selected districts were included in this study. There were 27 randomised clusters, including 35 healthcare facilities and their affiliated Lady Health Workers (LHWs), with the entire population of the clusters included in the trial. 13 clusters were allocated to the intervention group (n = 242,740) and 14 to the control group (n = 256,985). During the trial intervention period, 51,436 pregnancies were identified | |
| Interventions | Intervention: package focused on immediate household management of intrapartum events, recognition of low birth weight and suspected serious neonatal infection, and prompt referral to public sector hospitals LHWs from the intervention clusters were trained at their affiliated health facility and were provided with a bag and mask for neonatal resuscitation and oral amoxicillin. LHWs were also given pictorial guides depicting the management of asphyxia, thermal care, co‐bedding, breastfeeding of low birth weight babies, and recognition of suspected pneumonia and administration of oral amoxicillin before referral. LHWs were also encouraged to maintain links with traditional birth attendants, to keep records of expected births, and to attend home deliveries. Pregnant women in the intervention clusters were provided with clean delivery kits during health education sessions delivered by LHWs and the importance of provision of urgent neonatal care at birth. LHWs were also trained as per national and project guidelines to do additional postnatal visits (days 3, 5, 14, 28 after birth). With regard to traditional birth attendants in the intervention arm, a programme in basic immediate maternal and newborn care was organised, and attendants were trained in the use of clean delivery kits. Separate training sessions on health education and community mobilisation were held for male community members, during which meetings aimed to promote antenatal and postnatal care and facility births (n = 19,984 deliveries) Control: received routine care through the existing national programme, with the LHW programme functioning as usual (n = 18,325 deliveries) |
|
| Outcomes | Primary outcome: all‐cause neonatal mortality Secondary outcomes: cause‐specific neonatal mortality due to intrapartum events, prematurity, and sepsis; stillbirth rate |
|
| Notes | Funding: Saving Newborn Lives, Save the Children USA. There is no statement related to involvement of these funding parties in design and conduct of the study | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Quote: "to ensure reasonable balance between the two arms, we used stratified, restricted randomisation to allocate clusters"; "1 million random allocation schemes were generated by the study statistician (SC), who used a computer algorithm" Comment: probably done |
| Allocation concealment (selection bias) | Low risk | Comment: as this is a cluster‐randomised trial, allocation concealment should not be an issue because the clusters are randomised at once |
| Blinding of participants and personnel (performance bias) All outcomes | High risk | Quote: "delivery of the intervention was not blinded for practical reasons" Comment: probably not done |
| Blinding of outcome assessment (detection bias) All outcomes | Low risk | Quote: "verbal autopsies of all stillbirths and neonatal deaths were done by a separate team of trained anthropologists within 2–16 weeks of the event with standard WHO recommended instruments" Comment: probably done |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Reasons given (migrated or lost to follow‐up). Attrition (%) was 9.5% in each of intervention and control |
| Selective reporting (reporting bias) | Low risk | Comment: this is a registered trial, and this study has reported all outcomes mentioned in the protocol |
| Other bias | Low risk | Comment: study seems to be free from other biases |