ImportanceIn resource-constrained settings where the neonatal mortality rate (NMR) is high due to preventable causes and health systems are underused, community-based interventions can increase newborn survival by improving health care practices.ObjectivesTo develop and evaluate the effectiveness of a community-based maternal and newborn care services package to reduce perinatal and neonatal mortality in rural Pakistan.Design, Setting, and ParticipantsThis cluster randomized clinical trial was conducted between November 1, 2012, and December 31, 2013, in district Rahim Yar Khan in the province of Punjab. A cluster was defined as an administrative union council. Any consenting pregnant resident of the study area, regardless of gestational age, was enrolled. An ongoing pregnancy surveillance system identified 12 529 and 12 333 pregnancies in the intervention and control clusters, respectively; 9410 pregnancies were excluded from analysis due to continuation of pregnancy at the end of the study, loss to follow-up, or miscarriage. Participants were followed up until the 40th postpartum day. Statistical analysis was performed from January to May 2014.InterventionA maternal and newborn health pack, training for community- and facility-based health care professionals, and community mobilization through counseling and education sessions.Main Outcomes and MeasuresThe primary outcome was perinatal mortality, defined as stillbirths per 1000 births and neonatal death within 7 days per 1000 live births. The secondary outcome was neonatal mortality, defined as death within 28 days of life per 1000 live births. Systematic random sampling was used to allocate 10 clusters each to intervention and control groups. Analysis was conducted on a modified intention-to-treat basis.ResultsFor the control group vs the intervention group, the total number of households was 33 188 vs 34 315, the median number of households per cluster was 3092 (IQR, 3018-3467) vs 3469 (IQR, 3019-4075), the total population was 229 155 vs 234 674, the mean (SD) number of residents per household was 6.9 (9.5) vs 6.8 (9.6), the number of males per 100 females (ie, the sex ratio) was 104.2 vs 103.7, and the mean (SD) number of children younger than 5 years per household was 1.0 (4.2) vs 1.0 (4.3). Altogether, 7598 births from conrol clusters and 8017 births from intervention clusters were analyzed. There was no significant difference in perinatal mortality between the intervention and control clusters (rate ratio, 0.86; 95% CI, 0.69-1.08; P = .19). The NMR was lower among the intervention than the control clusters (39.2/1000 live births vs 52.2/1000 live births; rate ratio, 0.75; 95% CI, 0.58-0.95; P = .02). The frequencies of antenatal visits and facility births were similar between the 2 groups. However, clean delivery practices were higher among intervention clusters than control clusters (63.2% [2284 of 3616] vs 13.2% [455 of 3458]; P < .001). Chlorhexidine use was also more common among intervention clusters than control clusters (55.9% [4271 of 7642] vs 0.3% [19 of 7203]; P < .001).Conclusions and RelevanceThis pragmatic cluster randomized clinical trial demonstrated a reduction in NMR that occurred in the background of improved household intrapartum and newborn care practices. However, the effect of the intervention on antenatal visits, facility births, and perinatal mortality rates was inconclusive, highlighting areas requiring further research. Nevertheless, the improvement in NMR underscores the effectiveness of community-based programs in low-resource settings.Trial RegistrationClinicalTrials.gov Identifier: NCT01751945