Promoting latrine construction and use in rural villages practicing open defecation: process evaluation in connection with a randomised controlled trial in Orissa, India.

BACKGROUND: Our group conducted a cluster-randomised trial in 100 villages of Orissa, India to measure the impact of a rural sanitation intervention implemented under the government of India's Total Sanitation Campaign, on diarrhoea and soil-transmitted helminth infections. This paper reports on a process evaluation conducted in the context of the trial. METHODS: Process evaluation data were collected through review of key documentation, quantitative surveys, direct observations, and semi-structured interviews with staff from implementing NGOs and community members. Between March 2011 and March 2012, trained enumerators recorded observations on latrine construction status every 6-8 weeks in the 50 intervention villages and noted activities reported to have taken place based on NGO staff interviews and review of NGO records. A survey among 10% of households in intervention and control villages was conducted to compare levels of awareness of key intervention components. In addition, 10% of village water and sanitation committee (VWSC) members were interviewed to measure their level of involvement in the intervention delivery. RESULTS: The percentage of households with a latrine (completed or under construction) increased from 8% at baseline to 66% one year after the start of the intervention in March 2012. Almost none of the intervention households recall any form of participatory community-level activities at the start of the programme, although intervention households were generally more aware of the Total Sanitation Campaign (91% versus 49%, p < 0.001), VWSCs (51% versus 9%, p < 0.001), adolescent girls groups (23% versus 8%, p < 0.01), wall paintings (44% versus 7%, p < 0.001) and were more likely to report a household visit on sanitation during the past three months (65% versus 3%, p < 0.001). We found no strong evidence of an association between levels of awareness of or participation in mobilisation activities and levels of latrine coverage in intervention villages. CONCLUSIONS: The levels of coverage achieved and the levels of awareness of the mobilisation process in our intervention villages were lower than planned, but similar to those reported elsewhere in India under the TSC. Our process evaluation highlights important gaps between the TSC guidelines and their implementation on the ground. TRIAL REGISTRATION: Number on clinicaltrial.gov: NCT01214785.