My research career began in 1981 with the team of clinicians, physiologists and physicists led by Professor Osmond Reynolds FRS studying the use of magnetic resonance spectroscopy in understanding energy metabolism of the infant brain. My work involved developing neurodevelopmental tests for high risk, premature infants and ultrasonographic and spectroscopic scanning.
As a medical officer in a remote rural district of Nepal, leading a public health programme for mothers and children in a population of 300,000, I addressed the problems of mothers and children in the developing world.  In 1989 a research fellowship allowed me to study diagnostics for childhood tuberculosis with Professor Graham Rook and Professor John Stanford at UCL.
From 1990, as a senior lecturer in international child health at ICH, my collaboration in Nepal with Professor Dharma Manandhar began. We founded MIRA (Mother and Infant Research Activities), which is now the largest health research organisation in the country, employing, at one time, up to 800 field staff.  Our first studies were hospital-based, focusing on the epidemiology of common and preventable newborn problems, such as the epidemiology of birth asphyxia and the effects of low birth weight.
In 1995 our randomised controlled trial of conventional health education aimed at postnatal mothers demonstrated surprisingly little impact on behavioural and health outcomes.  So we tested a peer education approach using women’s groups, through another trial in remote mountainous populations.  Larger than expected benefits of this intervention on both neonatal and maternal mortality raised many questions about international policy for maternal and child health.  Published in the Lancet in 2004, the paper emphasised the importance of community participation in the amplification of behavioural change in poor populations.
Since 2002, supported by DFID and a Wellcome Trust Strategic Award, we have established six new trials of women’s groups and other low cost interventions in southern Nepal, Bangladesh, and east India where tribal populations are predominant, in urban slums in Mumbai, and in central Malawi.  Our trial in a tribal population of 230,000 in east India, published in the Lancet in 2010, showed a 45% reduction in newborn mortality and a 57% reduction in maternal depression.  We have also conducted multiple micronutrient trials in pregnancy in southern Nepal showing effects on birth weight, blood pressure and childhood nutrition. A meta-analysis of the 7 trials published in the Lancet in May 2013 showed that maternal mortality was halved and newborn mortality cut by one third in populations where >30% of pregnant women joined the women's group programme.

This portfolio of trials has established surveillance systems for maternal and newborn mortality covering a combined population of more than 2 million worldwide.  The studies have enabled us to examine prospective maternal and infant mortality rates, risk factors and social determinants of ill health and death. 

Quality of care is a new priority.   In the Lancet we reported modelling of the impact of new strategies to deliver life-saving drugs to women in Africa, with a new strategy recommendation about how to reduce deaths from sepsis and haemorrhage.  Several other trials of community interventions are ongoing which focus on survival and nutrition. We are exploring new approaches to integrated development which look at health, nutrition and environmental outcomes such as ecosystems and agricultural yields.