Indicators and Making Inferences from Indicators Prof Richard

Indicators and Making Inferences from Indicators Prof Richard Lilford, University of Warwick UN-Habitat Sustainable Development Goals Technical Meeting, Lake Naivasha February 2017

Example: Water and Sanitation Indicators SDG 45 “Percentage of population using safely managed water services by urban/rural. ” SDG 46 “Percentage of population using safely managed sanitation services by urban/rural. ”

Issues with Indicators 1. Selection of Indicators. 2. Measurement of indicator: – Numerator. – Denominator. 3. Use of indicator – making inferences.

Selection of Indicators Policy Improve water & sanitation (SDG 45 & 46) Enablers Processes Finance & contracting & community engagement Intermediate variable Water & sewage installations of a certain type SDG indicators Levels of faecal contamination Outcome Diarrhoea, death & wellbeing

Issues with Indicators 1. Selection of Indicators. 2. Measurement of indicator: – Numerator. – Denominator. 3. Use of indicator – making inferences.

Measurement of Indicator 1. Numerator, e. g. “safely manage sewage services. ” – Type of installation. – Coverage. – Hygiene / usage aspects. Census and household surveys – Maintenance. 2. Denominator (urban, schools, slums). See also Methodological note on monitoring WASH and wastewater for the SDGs. https: //www. wssinfo. org/fileadmin/user_upload/resources/Methodological-note-on-monitoring-SDGtargets-for-WASH-and-wastewater_WHO-UNICEF_8 October 2015_Final. pdf

The Case for Slum Health Neighbourhood effects Ezeh, et al. Lancet. 2016. Lilford, et al. Lancet. 2016. Photo from Kibera. org. uk

Neighbourhood Effects Type of neighbourhood effect Example from slum context Physical environment Muslim mortality paradox. Contaminated environment. [Geruso & Spears. 2014. ] [Bain, et al. 2014. ] Social interactions Moving to Opportunity experiment. [Chetty, et al. 2016. ] Variable crime rates / cultures. Geographic factors Poor people in rich cities have better health than poor people Shared geographic hazards. [Landrigan, et al. 2015. ] in poor cities. [Chetty, et al. 2016. ] Institutional factors Teacher expectations and poor Stigmatised residents. neighbourhoods. [shiree-DSK. 2012. ] [Galster. 2012. ] [UN-HABITAT. 2003. ]

The Health of People in Slums: Child Mortality (under 5) Slums Rural Poor All Urban Bangladesh 2006/07 81 77 86 63 Bangladesh 2013/14 57 49 64 37 Nairobi 2000 151 117 144 93 Nairobi 2012/13 80 56 53 57 Data from Urban Health Survey 2006 and 2013; Nairobi data from Nairobi Cross-sectional Slum Survey 2000 and 2012; and Demographic Health Surveys. See also Cutler D & Miller G. The Role of Public Health Improvements in Health Advances: The 20 th Century United States. 2004.

Issues with Indicators 1. Selection of Indicators. 2. Measurement of indicator: – Numerator. – Denominator. 3. Use of indicator – making inferences.

Use of Indicator: Inferences Brown, et al. Qual Saf Health Care. 2008; 17: 158 -62. Brown, et al. Qual Saf Health Care. 2008; 17: 162 -9. Brown, et al. Qual Saf Health Care. 2008; 17: 170 -7. Brown, et al. Qual Saf Health Care. 2008; 17: 178 -81. Chen, et al. BMJ Qual Saf. 2016; 25: 303 -10.

Hemming, et al. BMJ. 2015; 350: h 391. Hemming, Lilford & Girling. Stat Med. 2015; 34(2): 181 -96.

Connecting Research & Action: The Role of Models Theoretical Model Data Support for Model

Conclusion Careless measurement and casual inference is worse than no measurement at all.