SHARED SANITATION AND DIARRHEA: EVIDENCE FROM 51 COUNTRIES

SHARED SANITATION AND DIARRHEA: EVIDENCE FROM 51

COUNTRIESJames A. FullerDepartment of EpidemiologyUniversity of Michigan School of Public Health

Co-authors: Thomas Clasen, Marike Heijnen, Joseph Eisenberg

Shared Facilities

Currently classified by JMP as ‘unimproved’ due to: Accessibility Cleanliness

Little evidence linking sharing to diarrhea

Research Questions1. Is a child more likely to have diarrhea if

his/her household uses a shared facility (compared to a facility that is not shared)?

2. Is there a safe threshold for the number of households using a facility (i.e. < 5)

Demographic and Health Surveys 51 Surveys

Children < 5

Diarrhea prevalence in the past 2 weeks

Potential ConfoundersChild-level variables

Age Health Card

Household-level variables Toilet facility (improved/unimproved, ignoring sharing) Water source (improved/unimproved) Ownership of assets (refrigerator, bicycle, motorcycle/scooter) Mother’s education (6 categories) Mother’s age (6 categories) Number of children < 5 in the household Urban/Rural

Unadjusted (Crude) EffectsSharing is harm

fulSharing is protective

PROTECTIVE IN A FEW

NO EFFECT IN A FEW

HARMFUL IN MOST

Adjusted EffectsSharing is harm

fulSharing is protective

ATTENUATION OF THE EFFECT

Adjusted for: Household assets, mother’s age, mother’s education, child’s health card

Crude Prevalence Ratios

Pooled Results

RegionCrude PR (95% CI)

Adjusteda PR (95% CI)

Africa 1.07 (1.03-1.10) 1.05 (1.01-1.09)Latin America and the Caribbean 1.11 (1.04-1.19) 1.02 (0.96-1.10)South-East Asia and Western Pacific 1.16 (1.06-1.26) 1.09 (1.00-1.18)Eastern Mediterranean and Europe 1.26 (1.11-1.42) 1.22 (1.08-1.38)

All Regions Combined 1.09 (1.06-1.12) 1.05 (1.02-1.08)PR, Prevalence Ratio; 95%CI, 95% confidence interval.aAdjusted for mother’s age, mother’s educational attainment, asset ownership, and whether the child has a health card.

“Modest” Effect

Attenuation

Number of Households JMP is considering <5 HH as a safe

threshold

Different dose-response relationships have different policy implications

Not Shared

Shared with < 5

Shared with 5+

Not Shared

Shared with < 5

Shared with 5+

Not Shared

Shared with < 5

Shared with 5+

Number of Households

RegionSharing Category Crude PR Adjusted PR

All Regions Combined < 5 HH 1.07 (1.04-1.11) 1.04 (1.00-1.07)5 or more HH 1.06 (1.00-1.12) 1.02 (0.97-1.08)

Africa < 5 HH 1.06 (1.02-1.10) 1.04 (1.00-1.08)5 or more HH 1.02 (0.96-1.09) 1.03 (0.97-1.09)

Latin Am & Car < 5 HH 1.08 (0.99-1.18) 1.02 (0.93-1.11)5 or more HH 1.14 (0.96-1.35) 1.01 (0.85-1.20)

SEA & WP < 5 HH 1.13 (1.02-1.25) 1.07 (0.96-1.18)5 or more HH 1.22 (1.02-1.48) 1.12 (0.93-1.36)

Eastern Med & Eur < 5 HH 1.21 (0.89-1.65) 1.14 (0.84-1.56)5 or more HH 1.71 (0.89-3.30) 1.75 (0.97-3.16)

Table 4. The number of households sharing a toilet facility and the prevalence ratios for diarrhea among children < 5 years of age. Data from 39 Demographic and Health Surveys, 2001-2011.

The 2 groups appear to be similar

Some evidence of a dose-response

Summary Pooled analysis shows a modest effect

(5-10%)

Geographic heterogeneity

Confounding via socioeconomic status

Number of HH sharing has no clear effect

Strengths of this approach Broad scope captures virtually every

sharing scenario and setting

Adjusting for confounders

Data is readily available

Limitations Broad scope misses the details

Public vs. private ownership Cleanliness and Accessibility Fecal Sludge Management

Residual confounding