Cohort Profile: The PROLIFE study in Kerala, India

Population Registry of Lifestyle Diseases (PROLIFE) is a prospective cohort study aimed at the long-term follow-up of the residents of seven contiguous villages in Kerala, India. There were 161 942 subject enrolments as of 1 July 2002. The study was initiated by Health Action by People (HAP), a non-profit public health research organization.1 Kerala, with a population of >31 million in 2001, has the best health indicators in India.2,3 The ‘Kerala health experience’ has been much discussed in academic circles.4–6 The dramatic declines in fertility and mortality in the state have resulted in a rapid epidemiological transition characterized by high morbidity and mortality from chronic diseases. However, valid epidemiological information on the burden of chronic diseases in the population does not exist. We have to rely on secondary data from hospitals. HAP decided to bridge the data gap by setting up PROLIFE. At the time of set up, PROLIFE was more in the nature of an exploratory study and not based on a single hypothesis. Our first goal was to assemble a cohort to study the epidemiology of chronic diseases in the selected rural community. In its second phase, the study proposes to estimate the burden of risk factors, both conventional and novel, and look for association with major causes of death in Kerala. We also propose to investigate the social determinants of diseases and deaths in our community. As we are collecting detailed information on births and deaths, life course epidemiology would be a future focus of the study. The study is set up in the Varkala rural development block in Thiruvananthapuram district, Kerala (Figure 1). The Varkala block comprises seven panchayats (local governments) with population ranging from 15 869 to 28 309. These panchayats are served by 105 childcare centres (anganvadis), as part of India's ‘Integrated Child Development Services’. Each anganvadi caters to approximately 100 children and is managed by a female health worker. These health workers served as enumerators for the study with the consent and support of the local administration. Oral consent was obtained from the head of each household for the initial baseline survey. Written, informed consent was obtained from all respondents in subsequent resurveys. The study was approved by the institutional ethics committee. Study area All residents living in the Varkala block were eligible to participate in the study. The block has an area of 86.67 km2 and the density of population was 1873/km2, more than double the state average. The study area was predominantly rural (73% of population) and the per capita annual income was 90%. We found that sensitivity increases considerably without affecting specificity when broad categories like CVD are used. Some authors consider that 50% sensitivity and 90% specificity as reasonable in population studies.11,12 The use of semi-skilled health workers may be viewed as a weakness. However, we decided to persist with the use of health workers, as one of the principal aims of the study was to improve local level health data collection. Regular interaction with the health workers ensures that their knowledge and skill levels remain high. Refresher training programmes were conducted between December 2003 and August 2004 to improve the sensitivity of the cause of death assignment. We are planning a repeat validation study of cause of deaths and another training programme in 2009. Owing to resource limitations, all baseline risk factors have not yet been mapped. We hope to overcome this problem soon. The study's data set is not freely available. However, we welcome specific queries and proposals for collaboration; these should be addressed to the corresponding author. Kerala Research Program for Local Level Development (KRPLLD) and HAP.1 The authors thank the residents of Manamboor, Ottoor, Edava, Chemmaruthy, Elakamon, Cheruniyoor and Vettoor panchayats who participated in the study. They are thankful for the help provided by officials in the seven panchayats. They are also grateful to the 105 health workers and the seven supervisors. Conflict of interest: None declared.