Att_24_stunting_presentation_workshop.pptx

Trends and Determinants of Undernutrition in Ghana

Anirban Chatterjee and Michael Freeman August 31, 2012 Results from the 2011 MICS Survey

Outline Conceptual Framework Stunting Levels Complementary Feeding Practices Statistical Model Counterfactual Scenarios

Conceptual Framework Stunting Levels Complementary Feeding Practices Statistical Model

Indicator Levels

Indicator Levels

Stunting Levels Complementary Feeding Practices Statistical Model

Regional Survey Comparison Tool Link

Changes in Stunting Levels

Stunting over Age

Stunting By Wealth Quintile

Differences Across Quintiles

Stunting By Maternal Education

Differences Across Education Levels

Stunting By Urban/Rural

Stunting Levels Complementary Feeding Practices Statistical Model

Breastfeeding and IYCF

Stunting Levels Complementary Feeding Practices Statistical Model

Statistical Model Model 1: Proximal Factors Only

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection Model 2: Distal Factors Only

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection Model 2: Distal Factors Only Each increase in wealth or education is associated with a lower level of stunting: odds of stunting ~.5 if mother has secondary education, ~.3 for wealthiest quintile

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection Model 2: Distal Factors Only Each increase in wealth or education is associated with a lower level of stunting: odds of stunting ~.5 if mother has secondary education, ~.3 for wealthiest quintile Model 3: Proximal and Distal Factors

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection Model 2: Distal Factors Only Each increase in wealth or education is associated with a lower level of stunting: odds of stunting ~.5 if mother has secondary education, ~.3 for wealthiest quintile Model 3: Proximal and Distal Factors The effect of improved water and toilet facilities are soaked up by the more distal determinants of education and wealth

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection Model 2: Distal Factors Only Each increase in wealth or education is associated with a lower level of stunting: odds of stunting ~.5 if mother has secondary education, ~.3 for wealthiest quintile Model 3: Proximal and Distal Factors The effect of improved water and toilet facilities are soaked up by the more distal determinants of education and wealth Insurance maintained a protective effect in the presence of education and wealth

Insurance, improved toilet facilities, improved drinking water are all significantly associated with stunting: odds of stunting ~.75 compared to those without each protection Model 2: Distal Factors Only Each increase in wealth or education is associated with a lower level of stunting: odds of stunting ~.5 if mother has secondary education, ~.3 for wealthiest quintile Model 3: Proximal and Distal Factors The effect of improved water and toilet facilities are soaked up by the more distal determinants of education and wealth Insurance maintained a protective effect in the presence of education and wealth Only the highest levels of wealth and education remain protective once more downstream effects are accounted for

Stunting Levels Complementary Feeding Practices Statistical Model

Simulated Environment Purpose Translate model results into actionable items Measure the expected changes in stunting given changes in the predictors of interest Approach Change the values of a given predictor to capture the observed response

Simulated First Differences Expected gains from shifting from unexposed category to exposed category (ie, no insurance to insurance).

Variable Change in Probability of Stunting

Expected gains from shifting from unexposed category to exposed category (ie, no insurance to insurance).

Expected gains from shifting from unexposed category to exposed category (ie, no insurance to insurance).

Expected gains from shifting from unexposed category to exposed category (ie, no insurance to insurance).

Simulated Change in Stunting

Conclusions

Progress has been made on stunting, but not enough

Proximal Factors: Improvements in water and sanitation and higher levels of insurance are strongly associated with stunting at the individual level

Proximal Factors: Improvements in water and sanitation and higher levels of insurance are strongly associated with stunting at the individual level Distal Factors: Higher levels of education and wealth are associated with lower levels of stunting at the individual level

Acknowledgements Josephine Agborsor (UNICEF) Maria Dyrberg (UNICEF) Emmanuela Gakidou (IHME,UW) Gloria Obeng (UNICEF) Etienne Rusamira (UNICEF) Lilian Selenje (UNICEF)

Comments/Questions Thank you for your time

Contact: mikefree@uw.edu image2.jpeg image3.png image4.jpg image5.jpg image6.png image7.png image8.png image9.png image10.png image11.png image12.png image13.png image14.emf image15.png image16.png image1.jpeg