Shared Sanitation in Madhya Pradesh, India Kali Nelson, MPH Candidate Department of Epidemiology and Biostatistics The George Washington University
Jul 16, 2015
Shared Sanitation in Madhya Pradesh, India
Kali Nelson, MPH Candidate Department of Epidemiology and Biostatistics The George Washington University
Outline
• Background
• Objectives
• Methods
• Results
• Discussion
http://www.nationsonline.org/oneworld/india_map.html
http://www.wssinfo.org/definitions-methods/watsan-categories/
http://thefarthesthousecall.files.wordpress.com/2011/04/vip-latrine.jpg
IMPROVED VENTILATED IMPROVED
PIT (VIP) LATRINE
UNIMPROVED HANGING LATRINE
http://www.flickr.com/photos/waterdotorg/3695499855/sizes/l/in/photostream/
Background: Global Sanitation Trends
• Millennium Development Goals Sanitation Target: 75% covered with sustainable access to improved sanitation by 2015 1
• 37% of the global population lacked access to improved sanitation 2
• 15% of the global population practiced open defecation 2
• 86% rural
1 WHO/UNICEF 2006, 2 WHO/UNICEF 2012
Sanitation in India • 66% of India lacked access to improved
sanitation 1
• 42% urban
• 77% rural
• 51% of India practiced open defecation 1
• 14% urban
• 67% rural
• Madhya Pradesh 2
• 2006 - 27% households with toilets
• 2010 - 54% households with toilets 1 WHO/UNICEF 2012, 2 WHO/UNICEF 2011
Sanitation Ladder - India
http://www.wssinfo.org/fileadmin/user_upload/resources/IND_san.pdf
= open defecation
= other unimproved facilities
= shared
= improved
URBAN RURAL TOTAL
Background: Shared Sanitation
• Shared sanitation definition
• In 2010, 11% of the global population utilized shared sanitation
• 39% rural
• Shared sanitation in India: 9% in 2010
• 19% urban
• 4% rural
Background: Health Outcomes • Research on shared sanitation is very limited:
• Alexandria, Egypt: infection with soil-transmitted helminthes 1
• Rural Tanzania: trachoma risk 2
• Dhaka, Bangladesh: parasite and diarrheal disease prevalence 3
• Botswana, Ghana, and Zambia: infection with intestinal parasites 4
• Bhopal, India: open defecation 5
• Urban Bangladesh: weight-for-height scores 6
1 Curtale, et. al., 1998 2 Montgomery, et. al., 2010 3 Khan, 1987
4 Feachem, et. al., 1983 5 Biran, et. al., 2011 6 WHO/UNICEF 2012
Objectives
• Describe sanitation access among households in Madhya Pradesh, India.
• Analyze the relationship between sanitation access and diarrheal disease prevalence, safety of female users, cleanliness, and user satisfaction.
• Recommend changes to the current definition of improved sanitation based on research results.
Methods
• World Bank Water and Sanitation Program
Global Scaling Up Rural Sanitation intervention
• 2009 baseline survey
• Cross-sectional study of Impact Evaluation survey:
• Household questionnaire
• Children < 5 years health questionnaire
Methods
• Sanitation facility characteristics
• Sanitation status
• Sharing status
• Health outcome
• Two-week prevalence of diarrheal disease
• 3 or more bowel movements per day
• Cleanliness outcome
• Presence of flies
Methods
• Female safety outcomes
• Safety of sanitation facility during the day
• Safety of sanitation facility during the night
• Privacy
• Satisfaction with sanitation facility
• Potential confounders
Data Analysis
• Descriptive statistics
• Pearson’s Chi-square tests for association among sanitation variables and health, safety, and cleanliness outcomes
• Logistic regression analysis
• Proportional Odds Model
Descriptive Statistics Madhya Pradesh n (%)
Survey data
Households 1,978
Children 3,464
Sanitation status
Improved 261 (14.1%)
Unimproved 1,856 (85.9%)
Sharing status
Not shared 253 (12.8%)
Shared 1,717 (87.2%)
Health Outcome
Two-week diarrheal disease
prevalence
510 (14.7%)
Madhya Pradesh n (%)
Female safety
Daytime 968 (49.0%)
Nighttime 452 (22.9%)
Privacy 542 (27.5%)
Cleanliness – presence of flies
Always 1,671 (84.5%)
Sometimes 201 (10.2%)
Rarely 104 (5.3%)
Satisfaction
Satisfied 774 (39.3%)
Not satisfied 1,193 (60.7%)
Child Health Results
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Shared Shared (v. not shared) 1.534 0.765, 3.074
Location Household latrine/ less
than 10 min. (v. no
designated area)
1.328 0.761, 2.319
Location More than 10 min. (v. no
designated area)
1.408 1.032, 1.921*
Visible feces One or more (v. none) 0.522 0.353, 0.773*
Feces odor Yes (v. no) 1.264 0.857, 1.864
Logistic regression for two-week diarrheal disease prevalence among children
*Statistically significant at p < 0.05
Safety Results
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Sanitation
status
Improved (v.
unimproved)
14.921 3.695, 60.247*
Sharing
status
Not shared (v.
shared)
3.119 1.016, 9.573*
Toilet
location
Household latrine/
less than 10 min. (v.
no designated area)
0.927 0.511, 1.684
Logistic regression for female safety – daytime
*Statistically significant at p < 0.05
Safety Results
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Sanitation
status
Improved (v.
unimproved)
30.438 7.388,
125.409*
Sharing
status
Not shared (v.
shared)
1.830 0.581, 5.767
Toilet
location
Household latrine/
less than 10 min. (v.
no designated area)
5.663 2.778, 11.546*
Toilet
location
More than 10 min. (v.
no designated area)
1.995 1.233, 3.228*
Logistic regression for female safety – nighttime
*Statistically significant at p < 0.05
Safety Results
*Statistically significant at p < 0.05
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Sanitation
status
Improved (v.
unimproved)
15.824 4.874, 51.379*
Sharing
status
Not shared (v.
shared)
3.981 1.468, 10.796*
Toilet
location
Household latrine/
less than 10 min. (v.
no designated area)
3.892 2.008, 7.545*
Toilet
location
More than 10 min. (v.
no designated area)
1.955 1.315, 2.907*
Logistic regression for female privacy
Cleanliness Results
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Sanitation
status
Improved (v.
unimproved)
5.221 1.963, 13.883*
Sharing status Not shared (v. shared) 2.449 1.036, 5.790*
Toilet location Household latrine/ less
than 10 min. (v. no
designated area)
6.248 2.588, 15.086*
Toilet location More than 10 min. (v.
no area)
1.992 1.014, 3.911*
Visible feces None (v. one or more) 1.286 0.719, 2.302
Feces odor No (v. yes) 0.693 0.385, 1.246
Open pit/ open
drain nearby
No (v. yes) 1.269 0.823, 1.958
*Statistically significant at p < 0.05
Proportional Odds Model for presence of flies
Satisfaction Results
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Sanitation
status
Improved (v.
unimproved)
1.085 0.270, 4.353
Sharing status Not shared (v. shared) 9.897 2.713, 36.097*
Presence of
flies
Rarely (v. always) 1.931 0.559, 6.673
Presence of
flies
Sometimes (v. always) 0.721 0.362, 1.436
Toilet location Household latrine/ less
than 10 min. (v. no
designated area)
0.662 0.326, 1.346
Toilet location More than 10 min. (v.
no designated area)
0.381 0.271, 0.536*
Logistic regression for satisfaction
*Statistically significant at p < 0.05
Satisfaction Results
Variable Outcome modeled Odds
Ratio
95%
Confidence
Interval
Feces odor No (v. yes) 1.094 0.674, 1.775
Open pit/ open
drain nearby
No (v. yes) 0.541 0.386, 0.757*
Visible feces None (v. one or more) 0.496 0.298, 0.827*
Female safety
– day
Yes (v. no) 6.489 4.539, 9.276*
Female safety
– night
Yes (v. no) 1.496 0.940, 2.382
Female privacy Yes (v. no) 1.397 0.903, 2.163
Logistic regression for satisfaction
*Statistically significant at p < 0.05
Conclusions
• Health outcomes
• Improved sanitation
• Shared facilities
Limitations
• Weak indicators
• Two-week diarrheal prevalence
• Number of households sharing
• Cross-sectional data
Strengths
• Large dataset
• One of first studies to analyze shared sanitation and user satisfaction
Discussion
• Number of households sharing
• How shared facilities affect user satisfaction • Analysis of household survey data from
Bangladesh, Tanzania, and Indonesia
Acknowledgments
• Dr. Angelo Elmi
Dept. of Epidemiology and Biostatistics
• Dr. Jay Graham
Dept. of Environmental and Occupational Health
• Craig Kullmann
World Bank Water and Sanitation Program
• Prof. Ann Goldman
Dept. of Epidemiology and Biostatistics
Questions?