OBJECTIVE: This pilot study investigated the feasibility of delivering a package of community-based interventions for improving perinatal care using lady health workers (LHWs) and traditional birth attendants (Dais) in rural Pakistan. METHODS: The intervention was implemented in four of eight village clusters (315 villages, total population 138,600), while four served as a comparison group. The LHWs in intervention clusters received additional training focused on essential maternal and newborn care, conducted community education group sessions, and were encouraged to link up with local Dais. The intervention was delivered within the regular government LHW programme and was supported by the creation of voluntary community health committees. FINDINGS: In intervention villages, there were significant reductions from baseline in stillbirth (from 65.9 to 43.1 per 1000 births, P < 0.001) and neonatal mortality rates (from 57.3 to 41.3 per 1000 live births, P < 0.001). The proportion of deliveries conducted by skilled attendants at public sector facilities also increased, from 18% at baseline to 30%, while the proportion of home births decreased from 79% to 65%. A household survey indicated a higher frequency of key behaviours (e.g. early and exclusive breastfeeding, delayed bathing and cord care) in intervention villages. CONCLUSION: The improved stillbirth and neonatal mortality rates observed indicate that community health workers (i.e. LHWs and Dais) can be effective in implementing a community and outreach package that leads to improved home care practices by families, increased care-seeking behaviour and greater utilization of skilled care providers. These preliminary observations require confirmation in an adequately powered trial.