Social Accountability, Information and Networks for Health Care Delivery

Last registered on August 17, 2016


Trial Information

General Information

Social Accountability, Information and Networks for Health Care Delivery
Initial registration date
August 10, 2016

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
August 10, 2016, 10:25 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
August 17, 2016, 5:15 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.



Primary Investigator

Duke University

Other Primary Investigator(s)

PI Affiliation
Stanford University
PI Affiliation
University of North Carolina, Chapel Hill

Additional Trial Information

On going
Start date
End date
Secondary IDs
In several low and middle-income countries, Social Accountability (SA) interventions have been introduced as an innovative approach to governance, aiming to improve delivery of public services. These interventions typically include information provision to citizens regarding their rights/entitlements and local provider performance, and additionally, facilitation of community engagement with providers and officials.

This study aims to: (a) measure the causal effect of SA interventions on key outcomes (health status, quality of service); (b) test the effectiveness of social networks based strategies to disseminate information for community engagement; and (c) study individuals’ decisions to participate in collective action efforts in the context of social networks and information interventions. In addition to evaluating the impact of SA interventions, the study aims to generate new knowledge on relative strengths of information seeding strategies, identifying those that maximize the spread of information through the village network, and subsequently estimate peer effects on participation decisions.

The state government and UPHSSP have identified 12 districts where the social accountability impact evaluation will be introduced on a priority basis. The study has two concurrent designs – one evaluates the impact of policy intervention that was randomized across 87 blocks in 10 districts (population ~27 million) in Uttar Pradesh, India, and another that was randomized across 120 villages in 2 additional districts to study mechanisms through which information and collective action lead to improved accountability and outcomes.
External Link(s)

Registration Citation

Chandrasekhar, Arun, Manoj Mohanan and Harsha Thirumurthy. 2016. "Social Accountability, Information and Networks for Health Care Delivery." AEA RCT Registry. August 17.
Former Citation
Chandrasekhar, Arun, Manoj Mohanan and Harsha Thirumurthy. 2016. "Social Accountability, Information and Networks for Health Care Delivery." AEA RCT Registry. August 17.
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details


The overarching research objectives in our evaluation are to assess the impact of social accountability (SA) interventions in UP on: (1) objective measures of health service quality in practice, (2) village-level satisfaction with health services, and (3) village-level health outcomes.
Importantly, the evaluation will seek to identify whether information provision, which is a standard part of social accountability interventions, has an independent effect on outcomes that is comparable to the combined effect of information provision and community engagement / facilitation. We also test the effectiveness of alternative models of delivering information in order to inform implementation of accountability interventions at large scale in public policy settings. Given that the community health workers targeted by the intervention focus on maternal and child health, our measures of health system performance will also emphasize maternal and child health. Measures of health service quality include availability of services such as immunization and primary care services, distribution of food and nutritional supplementation as recommended in the national nutrition program, and provider absenteeism. Measures of satisfaction with local health services will include process measures such as availability of service, waiting time for services and whether or not community members perceive that they are treated by providers with respect as well as general subjective assessments of satisfaction with services. Health outcome measures will include child anthropometrics (weight-for-age and weight-for-height), self-reported morbidity in the preceding two weeks (diarrhea, cough, fever, headache, days of usual activities lost due to illness), and neonatal (0-28 days), infant (under age one), and child (under age five) mortality as well as maternal health indicators such as percentage of facility deliveries among mothers who gave birth in the past year.
The SA interventions were designed to address two specific aims: (1) to test the impact of SA interventions when implemented at scale as a policy instrument, and (2) to study mechanisms through which the information and facilitation interventions lead to increasing collective action efforts at the village level and improve governance of public service delivery at the local level. We investigate these two specific aims in two parallel sets of interventions.

51-Block intervention: The SA intervention will be implemented in 51 out of 87 blocks in 10 districts (covering a population of over 27 million in these districts). This intervention will include information dissemination as well as facilitation of community participation in Village Health, Sanitation and Nutrition Committees (VHSNCs). These activities will be led by Gram Panchayat Coordinators (GPCs), recruited and trained by the State Institute of Rural development (SIRD).

120 village intervention: In order to study mechanisms through which social accountability interventions might lead to improvement in outcomes, the intervention is also implemented in 120 villages that were selected at random within 2 districts. The SA interventions aim to distinguish the effect of provision of information and facilitated engagement of community members, from that of the effect of information alone. The 120 villages are randomized to either a control arm, or one of two treatment arms described below.

Community members will receive information about their health care rights and entitlements, and about certain health outcomes in their village. Additionally, they will receive notifications of upcoming health-related activities happening in their village. Importantly, besides assessing the effect of providing information on health service delivery and various health outcomes, our evaluation will also determine how best to provide the information.
Another innovation in this project is to disseminate monthly information about health system related issues using interactive voice response messages (IVRs), phone calls, or home visits to households in the treatment villages. IVRs include a brief message about a health (or health system) indicator that is specific to the village, or information regarding upcoming VHSNC meetings or Village Health and Nutrition Days (VHNDs), as well as a response option that can be used to collect data on what information households have received and about their participation in VHSNC meetings/VHNDs. These IVRs will be sent out every month to cover approximately 24,000 households in 80 treatment villages. During preparatory phase that was conducted in parallel with the baseline survey, we collected detailed data on social networks in each village and identified central individuals in the village.
We will use the data on networks in villages, combined with the information that we disseminate on a monthly basis to econometrically estimate how information dissemination within networks affects awareness and participation in social accountability activities as well as in utilization of health services.

The community engagement component aims to enhance the participation of the community in creating social accountability. The intervention will provide trained facilitators to help community members engage in a participatory process with VHSNCs and PRIs and identify key deficiencies for improvement in health services that most concern community members. The facilitators are trained to help organize meetings and are provided a detailed checklist of activities that need to be undertaken prior to the day of the meetings such as informing all community representatives about the date, inviting the block level officers and ensuring that logistics requirements for VHNDs are conveyed to VHSNC members in advance of the meetings. The facilitated meetings with healthcare providers and local and block level representatives aim to empower community members to demand better health services and convey these demands more effectively to providers and officials. The three key health workers at the village level (ASHA, ANM, and AWW) report to the local (village level) elected representatives and block level authorities, who receive feedback from the community in the accountability interventions. Moreover, through repeated community meetings village-level health workers are expected to respond to transparency and accountability innovations by improving quality of services delivered to their local constituents resulting in improvements in population health outcomes.
These interventions will focus on services delivered by village-level health workers including those providing primary care and maternal and child health services. The households surveyed in the project will be those with children less than 5 years of age. The interventions will cover and include all population subgroups, castes, and religious minorities in the treatment villages, and also collect data on all of these subgroups.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Primary Outcomes: (1)Weight-for-height Z scores, (2) U5 mortality rate, (3) incidence of diarrhea, (4) duration of diarrhea illness

Secondary Outcome Measures: (1) Satisfaction with local healthcare providers and services, (2) Participation, (3) Household participation in Village Health, Sanitation, and Nutrition Committee (VHSNC) meetings and attendance at Village Health and Nutrition Days (VHNDs)

Other Pre-specified Outcome Measures: (1) Spread of information (proportion of the village that received/retained information spread through the different info dissemination strategies)
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
As outline in the intervention description above, 40 villages will be assigned to the information only arm, 40 to the information plus community engagement arm, and the remaining 40 to the control arm. Additionally, we will implement a number of “networks” interventions that aim to identify effective information dissemination strategies as well as peer effects on participatory behavior and collective action. The networks interventions are described briefly below.

The main objective of selectively “seeding” information will be to track how that information gets disseminated throughout the village network and to identify which seeding strategy is more effective in helping disseminate information about health-related activities in the village. In the first two months of interventions in the project, the 80 intervention villages, stratified by treatment arm, will be randomized to receive information through one of the following strategies:
(1) Broadcast messaging – where every household (that has a mobile phone) receives an interactive voice response (IVR) message giving information regarding VHSNC activity and relevant health-related statistics.
(2) Information “seeds” – households selected as “seeds” will receive a phone call relaying similar information but in more detail (as the IVR messages are more time constrained) that they will be asked to share with others in the village.
In the first month, broadcast messages (strategy 1) will be disseminated to all households in 40 randomly selected villages and in the remaining 40 villages, 10 network central individuals will serve as the information seeds. These network central seeds will be selected using previously collected network data. The network survey asked “If we want to spread information about health services available to everyone in your village, to whom do you suggest we speak?”, and “If we want to spread information in the village about tickets to a music event, drama, or fair that we would like to organize in your village, to whom should we speak?”. We selected the 10 households most nominated as people with whom we should speak as our network seeds.
Similarly, in the second month, the 80 villages will be randomized again into 40 that will receive broadcast messages and 40 villages that will receive detailed information disseminated through phone calls to government information seeds. These seeds will be elected local officials and AAAs (ASHAs, Auxiliary Nurse-Midwives, and Anganwadi workers).
During both months, in the villages randomized to receive broadcast messages, 10 individuals will also be randomly selected as information “seeds” (and will receive detailed informational phone calls).

In the subsequent two months after the information intervention strategies are tested, we study households’ decision to participate in village level collective action (participation in monthly VHSNC meetings). Specifically, we aim to understand how this participation decision is influenced by the flow of information in the household’s social network, and by the share of their network that participates in the meeting. In the information plus community engagement (treatment arm 2) villages, 30-40% of households (village shares selected at random and households selected at random) will be visited and provided additional information regarding ongoing VHSNC activities. A subset of households will also get an incentive (in the form of a scratch card lottery) to attend the next VHSNC meeting in their village. The random assignment of share of households visited with meeting reminders, and random assignment of households to receive incentives to attend the meeting will enable us to study households’ decision to participate in collective action as a function of information in their social networks, threshold effects and free riding behavior.
Experimental Design Details
Randomization Method
Two districts of Uttar Pradesh (out of 12 implementation districts) were selected for the evaluation. From these two districts, 120 villages (60 per district) were randomly selected for participation. The villages were randomly selected into the three arms (info only, info plus community engagement, and control).

For the networks interventions:
The 80 intervention villages (arms 1 and 2) were randomized each month. In month 1, the 80 villages were randomized to receive information either through broadcast messaging or through "network central" individuals. Villages were then re-randomized in month 2 to receive information either through broadcast messaging or through local elected officials and AAA employees.

Randomization was done in office by a computer.
Randomization Unit
The unit of randomization is at the village-level.
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
120 villages
Sample size: planned number of observations
60,000 mothers and 120,000 children under 5
Sample size (or number of clusters) by treatment arms
40 villages control, 40 villages information only, 40 villages information and facilitation
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
(1) Weight-for-height Z scores: using an intracluster correlation coefficient (ICC) of 0.1 (Fenn, Morris, et al., [2004]), we have 90% power to detect a 0.225 SD change in Z scores (80% power to detect a 0.19 SD change). (2) Under 5 (U5) mortality: Using an ICC of 0.0175 (Mann, Veble et al., [2010]), we have 90% power to detect and 0.45 SD change in U5 mortality (80% power to detect a 0.3 SD change) (3) Childhood diarrhea incidence: Using an ICC of 0.03 (***CITE***) , we have 90% power to detect a 0.06 SD change in incidence of diarrhea (80% power to detect a 0.05 SD change)
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Institutional Review Boards (IRBs)

IRB Name
Duke University
IRB Approval Date
IRB Approval Number
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials