BACKGROUND: In 2009, on the basis of promising evidence from trials in south Asia, WHO and UNICEF issued a joint statement about home visits as a strategy to improve newborn survival. In the Newhints trial, we aimed to test this home-visits strategy in sub-Saharan Africa by assessing the effect on all-cause neonatal mortality rate (NMR) and essential newborn-care practices. METHODS: The Newhints cluster randomised trial was undertaken in 98 zones in seven districts in the Brong Ahafo Region, Ghana. 49 zones were randomly assigned to the Newhints intervention and 49 to the control intervention by use of restricted randomisation with stratification to ensure comparability between interventions. Community-based surveillance volunteers (CBSVs) in Newhints zones were trained to identify pregnant women in their community and to make two home visits during pregnancy and three in the first week of life to promote essential newborn-care practices, weigh and assess babies for danger signs, and refer as necessary. Primary outcomes were NMR and coverage of key essential newborn-care practices. Analyses were by intention to treat. This study is registered with ClinicalTrials.gov, number NCT00623337. FINDINGS: 16,168 (99%) of 16,329 deliveries between November, 2008, and December, 2009, were livebirths; the status at 1 month was known for 15,619 (97%) livebirths. 482 neonatal deaths were recorded. Coverage data were available from 6029 women in Newhints zones; of these 4358 (72%) reported having CBSV visits during pregnancy and 3815 (63%) reported having postnatal visits. This coverage increased substantially from June, 2009, after the introduction of new implementation strategies and reached almost 90% for pregnancy visits by the end of the trial and 75% for postnatal visits. The Newhints intervention significantly increased coverage of key essential newborn-care behaviours, except for four or more antenatal-care visits (5975 [76%] of 7859 vs 5988 [74%] of 8121, respectively; relative risk 1·02, 95% CI 0·96-1·09; p=0·52) and baby delivered in a facility (5373 [68%] vs 5539 [68%], respectively; 0·97, 0·81-1·14; p=0·69). The largest increase was for care-seeking, with 102 (77%) of 132 sick babies in Newhints zones taken to a hospital or clinic compared with 77 (55%) of 139 in control zones (1·43, 1·17-1·76; p=0·001). Increases were also noted in bednet use during pregnancy (5398 [69%] of 7859 vs 5135 [63%] of 8121, respectively; 1·12, 1·03-1·21; p=0·005), money saved for delivery or emergency (5730 [86%] of 6681 vs 5525 [80%] of 6941, respectively; 1·09, 1·05-1·12; p<0·0001), transport arranged in advance for facility (2496 [37%] vs 2061 [30%], respectively; 1·30, 1·12-1·49; p=0·0004), birth assistant for home delivery washed hands with soap (1853 [93%] of 1992 vs 1817 [87%] of 2091, respectively; 1·05, 1·02-1·09; p=0·001), initiation of breastfeeding in less than 1 h of birth (3743 [49%] of 7673 vs 3280 [41%] of 7921, respectively; 1·22, 1·07-1·40; p=0·004), skin to skin contact (3355 [44%] vs 1931 [24%], respectively; 2·30, 1·85-2·87; p=0·0002), first bath delayed for longer than 6 h (3131 [41%] vs 2269 [29%], respectively; 1·65, 1·27-2·13; p<0·0001), exclusive breastfeeding for 26-32 days (1217 [86%] of 1414 vs 1091 [80%] of 1371; 1·10, 1·04-1·16; p=0·001), and baby sleeping under bednet for 8-56 days (4548 [79%] of 5756 vs 4291 [73%] of 5846; 1·09, 1·03-1·15; p=0·002). There were 230 neonatal deaths in the Newhints zones compared with 252 in the control zones. The overall NMRs per 1000 livebirths were 29·8 and 31·9, respectively (0·92, 0·75-1·12; p=0·405). INTERPRETATION: The reduction in NMR with Newhints is consistent with the reductions achieved in three trials undertaken in programme settings in south Asia. Because there is no suggestion of any heterogeneity (p=0·850) between these trials and Newhints, the meta-analysis summary estimate of a reduction of 12% (95% CI 5-18) provides the best evidence for the likely effect of the home-visits strategy delivered within programmes in sub-Saharan Africa and in south Asia. Improvements in the quality of delivery and neonatal care in health facilities and development of innovative, effective strategies to increase coverage of home visits on the day of birth could lead to the achievement of more substantial reductions. FUNDING: WHO, Bill & Melinda Gates Foundation, and UK Department for International Development.