x

We are happy to announce that all trial registrations will now be issued DOIs (digital object identifiers). For more information, see here.
Credit with health education in Benin: A cluster randomized trial examining impacts on knowledge and behavior
Last registered on June 01, 2015

Pre-Trial

Trial Information
General Information
Title
Credit with health education in Benin: A cluster randomized trial examining impacts on knowledge and behavior
RCT ID
AEARCTR-0000726
Initial registration date
June 01, 2015
Last updated
June 01, 2015 4:08 PM EDT
Location(s)
Region
Primary Investigator
Affiliation
Northwestern University
Other Primary Investigator(s)
PI Affiliation
Ghent University; Innovations for Poverty Action
PI Affiliation
Freedom From Hunger
Additional Trial Information
Status
Completed
Start date
2007-10-01
End date
2010-03-01
Secondary IDs
Abstract
Background:
We evaluate whether health education integrated into microcredit lending groups reduces health risks by improving health knowledge and behaviors among urban and rural borrowers in eastern Benin.

Methods:
In 2007, we randomly assigned 138 villages in the Plateau region of Benin to one of four variations of a group liability credit product, varying lending groups’ gender-composition and/or inclusion of health education using a 2x2 design. Groups 1 (female-only) and 2 (mixed-gender) included health education on malaria, childhood illness danger signs, and HIV/AIDS. Groups 3 (female-only) and 4 (mixed-gender) received the credit product without health education. In 2010, a follow-up survey in 121 of 138 villages (n=3,625) interviewed a random sample of women living near each village meeting site to measure respondents’ knowledge of malaria, childhood illness, and HIV/AIDS; prevention and treatment behaviors; and empowerment and social capital. We estimate treatment effects on indices of key outcome variables using ordinary least squares.

Findings:
Women in villages receiving health education showed improved knowledge of malaria and of HIV/AIDS, but not of childhood illness danger signs. No significant changes in health behavior were observed except an increase in HIV/AIDS prevention behavior, a result predominantly driven by an increase in respondents’ self-reported ability to procure a condom, likely an indicator of increased perceived access rather than improved preventative behavior. Women in villages assigned to mixed-gender groups had significantly lower levels of social capital, compared to villages assigned to female-only groups.

Interpretation:
Bundling health education with microcredit may not be an effective way to improve health behavior.
External Link(s)
Registration Citation
Citation
Gray, Bobbi, Dean Karlan and Bram Thuysbaert. 2015. "Credit with health education in Benin: A cluster randomized trial examining impacts on knowledge and behavior." AEA RCT Registry. June 01. https://doi.org/10.1257/rct.726-1.0.
Former Citation
Gray, Bobbi, Dean Karlan and Bram Thuysbaert. 2015. "Credit with health education in Benin: A cluster randomized trial examining impacts on knowledge and behavior." AEA RCT Registry. June 01. https://www.socialscienceregistry.org/trials/726/history/4344.
Sponsors & Partners

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

Request Information
Experimental Details
Interventions
Intervention(s)
Intervention Start Date
2007-12-01
Intervention End Date
2009-12-01
Primary Outcomes
Primary Outcomes (end points)
Health outcomes:
1) malaria knowledge
2) HIV/AIDS knowledge
3) Integrated Management of Childhood Illness (IMCI) knowledge
4) mosquito bednet behavior
5) HIV/AIDS access and behavior
6) IMCI behavior

Social outcomes:
1) women's empowerment
2) social network
3) social capital
Primary Outcomes (explanation)
We group outcome indicators into nine indices of health and social outcomes, with each index standardized so the control group (all-female, no education group) has a mean of zero and a standard deviation of one. For each index we follow the method employed by Kling, Liebman and Katz (2007).

Where relevant, incorrect answers (not specified here) are assigned the value 0, and un-answered survey questions are coded as missing.

Health knowledge and behavior:
(1) Malaria knowledge score, 3 components: Q1: What causes malaria? =1 if respondent lists "mosquito bites" only; Q2: How does one protect him/herself from getting malaria? =1 if respondent lists "sleeping under ITN" only; Q3: Which two groups of people are most vulnerable to the effects of Malaria? =1 if respondent lists both "young children" and "pregnant women" only.
(2) HIV/AIDS knowledge score, 8 components: All questions take the value 1 if respondent answers "Yes." Q1: Have you heard of AIDS? Q2: In the last few months, have you heard or seen any messages on HIV/AIDS? Q3: Do you personally know someone who is suspected to have the AIDS virus or who has the AIDS virus? Q5: Can people reduce their chances of getting the AIDS virus by using a condom every time they have sex? Q6: Can people reduce their chance of getting the AIDS virus by abstaining from sexual intercourse? Q7: Are there any special medications that people infected with the AIDS virus can get from a doctor or nurse? Q8: Do you know where people can go to get tested for the virus that causes AIDS? Q8: Do you know where to get condoms?
(3) IMCI knowledge score, 7 components: Q1: What are the critical danger signs of serious illness in children 2 months to 5 years of age? 1 point for each correct answer, up to 6; Q2: How much liquid should you give a child with diarrhea? =1 if respondent selects "more than usual"; Q3: When a child has diarrhea, what symptoms indicate that you should take him/her for medical care? 1 point for each correct answer, up to 8; Q4: When a child has a cough, what symptoms indicate that you should take him/her for medical care? 1 point for each correct answer, up to 10; Q5: How do you know a child has malaria? =1 if respondent lists "fever"; Q6: If a child has a fever, how would you recommend that child be treated at home? 1 point for each correct answer, up to 5; Q7: What actions should a doctor/nurse take when assessing the health of your child? 1 point for each correct answer, up to 10.
(4) Bednet behavior score, 5 components: Q1: Number of installed bednets in household; Q2: Proportion of household members under age five who slept under a net the previous night; Q3: Household has a net, 1 year or newer; Q4: Household has an installed net, 1 year or newer; Q5: Household has a net treated within the past year
(5) HIV/AIDS behavior score, 3 components: All questions take the value 1 if respondent answers "Yes." Q1: Have you spoken with your husband/companion about ways to avoid contracting HIV/AIDS?; Q2: The last time you had sexual intercourse, was a condom used?; Q3: If you wanted to, could you yourself get a condom?
(6) IMCI behavior score, 2 components: Q1: Respondent has given Orasel or other oral rehydration salts to a child to treat diarrhea; Q2: Respondent sought treatment within 3 days for a child who had fever during the previous month

Social outcomes:
(1) Empowerment score, 8 components: Q1: Participates in decisions about children's school attendance; Q2: Participates in decisions about children's membership in groups; Q3: Participates in buying and selling decisions for the household; Q4: Participates in decisions about working outside the household; Q5: How often do conflicts with your spouse lead to a) verbal abuse, b) physical abuse, or c) physical abuse against children? For each of a, b and c, 0=often, 1=sometimes, 2=rarely, 3=never. Combined score: 0=often to 9=never; Q6: Ability to go out to a) the market, b) the health center, c) walk with friends, d) the religious center? For each of a, b, c, d, 0=never, 1=cannot when alone, 2=can when alone. Combined score: 0=never to 8=always able to travel alone; Q7: Views on women’s empowerment: a) Decisions in the family should be taken by men, b) If a woman works outside the home, then the husband should help with household chores, c) A married woman must be able to work outside the home if she wishes, d) A married woman should be allowed to express her opinion even when she disagrees with her husband, e) A woman should accept beatings by her husband in order to preserve the unity of the family, f) It is better to send boys to school than girls. For b, c, d, Yes=2, No=0, Depends=1, and for a, e, and f, Yes=0, No=2, Depends=1. Combined score: 0=low empowerment, 12=high empowerment; Q8: Frequency of voting: 0=never, 1=sometimes, 2=always
(2) Social network score, 3 components: Q1: Number of matched respondents who have given economic support, out of 10; Q2: Number of matched respondents she could ask for small money, out of 10; Q3: Number of matched respondents she would lend small money, out of 10
(3) Social capital score, 6 components: Q1: Number of community groups that the respondent belongs to; Q2: Number of groups from whom the respondent got economic support in the last 12 months; Q3: Number of groups from whom the respondent got advice in the last 12 months; Q4: Has spoken out in a community meeting in last 12 months; Q5: Has run for or held an elected community position or office in last 12 months; Q6: Feels at least somewhat influential in changing her village
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
Implementation occurred over three waves, with assignment to wave chosen non-randomly by the implementing partner based on operational considerations. Stratifying by three implementation waves, we randomly assigned 138 communities to receive one of four group loan products from PADME, with equal probability: credit with health education to female-only groups; credit with health education to mixed-gender groups; credit only to female-only groups; and, credit only to mixed-gender groups. Seventeen villages were dropped post-randomization because PADME was unable to offer the new product in those villages.

Experimental Design Details
Randomization Method
The randomization was conducted using a computerized random number generator.
Randomization Unit
Village level
Was the treatment clustered?
Yes
Experiment Characteristics
Sample size: planned number of clusters
138 villages
Sample size: planned number of observations
Endline survey planned for 30 woman per village cluster, for total of 4,140 women
Sample size (or number of clusters) by treatment arms
35 villages assigned to credit with health education in female-only groups; 34 villages assigned to credit with health education in mixed-gender groups; 35 villages assigned to credit only in female-only groups (control); 34 villages assigned to credit only in mixed-gender groups
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

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

Request Information
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
Human Research Protection Program at Yale University
IRB Approval Date
2007-05-10
IRB Approval Number
IRB Protocol # 0705002650
Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
Yes
Intervention Completion Date
December 01, 2009, 12:00 AM +00:00
Is data collection complete?
Yes
Data Collection Completion Date
March 01, 2010, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
121
Was attrition correlated with treatment status?
No
Final Sample Size: Total Number of Observations
3,625 women surveyed in 121 villages that received treatment
Final Sample Size (or Number of Clusters) by Treatment Arms
32 villages in health education & female-only groups treatment; 27 villages in health education & mixed-gender groups treatment; 32 villages in credit only & female-only groups control; 30 villages in credit only & mixed-gender groups treatment
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
No
Reports and Papers
Preliminary Reports
Relevant Papers