Improving Evaluation of International Public Health Programs Through the Use of a Geographically Informed Data Model

Improving Public Health Programs Through The Use of A Geographically Informed Data

Model: A Strategy for Low Capacity Environments

John SpencerSr. Geospatial AnalystMEASURE Evaluation

American Evaluation Association MeetingWashington D.C.October 16, 2013

A health worker reviews health records in a Kenyan district health facility, 2012

Using data for evaluation and evidence based decision making

Transitioning to a data rich environment

Data has become stovepiped

Barriers to using reporting data

Barriers to using reporting data

Technical• Do I have to clean the data?• Is it in a compatible format?• Can the data link to other data?

Non-Technical• Who has the data?• How do I get a copy of the data?• When was it collected?

A geographically informed data model can address both technical and non-technical barriers

Data modelsCommon in other sectors but not so common in global public health.

Data Model

1. Location of program2. Service provided3. Number of beneficiaries4. Implementing organization5. Year or date

Linked using Geography

Use numeric geographic identifiers

District Population

Coast 79,133

Mountain 66,251

North 23,415

ID District Population

101 Coast 79,133

103 Mountain 66,251

105 North 23,415

• Facilitates linking• Many countries have district codes, but they may not be

widely used• If there aren’t codes, there are international standards that

can be used to create codes.

Hard to link Easier to link

Not rocket science

Geography can be the common link across data

Non-technical side

• Creates consistency with data• Can help achieve buy-in about sharing data– Makes it easier for data producers to link data

themselves– More involved in the data community

At least 4 Steps1. Standardize names

• Le Tierge

2. Spelling• Karatu

3. Identify changes in boundaries• Totou

4. Definitional Changes• OVC

Illustrative data linkingOrphan and Vulnerable

Children Programs

District Orphan Est. 07 OVC Served by PEPFAR

CT HH 2013

Koratu 21821 54 1604

Le Tiergé 21804 5015 2000

Salamansa 471204 2500 2229

Totou 108109 7074 -999

East Totou -999 -999 3473

District Orphan Est. 07

Koratu 21821

Le Tierge 21804

Salamansa 471204

Totou 108109

District OVC Served by PEPFAR

Koratu 54Letierge 5015Salamansa 2500Totou 7074

PEPFARHIV Prevalence Report

District CT HH 2013

Karatu 1604

Le Tiergé 2000

Salamansa 2229

East Totou 3473

Cash Transfer Database

Integrated Data Table

National Gove

rnment

PEPFAR

Multilateral

• Numeric code for districts

• Spelling variation not an issue

• Accommodates changes in geography

Using a data model

ID District Orphan Est. 07

OVC Served by PEPFAR

CT HH 2013

101 Koratu 21821 54 1604

103 Le Tiergé 21804 5015 2000

105 Salamansa 471204 2500 2229

107 Totou 108109 7074 -999

108 East Totou -999 -999 3473

ID District Orphan Est. 07

101 Koratu 21821

103 Le Tierge 21804

105 Salamansa 471204

107 Totou 108109

ID District OVC Served by PEPFAR

101 Koratu 54103 Letierge 5015105 Salamansa 2500107 Totou 7074

PEPFARHIV Prevalence Report

ID District CT HH 2013

101 Karatu 1604

103 Le Tiergé 2000

105 Salamansa 2229

108 East Totou 3473

Cash Transfer Database

Integrated Data Table

Not rocket science

Geography can be the common link across data

Growing focus on data

Building blocks for better M&E

Photo by ogimogi http://flic.kr/p/4r9zSK

Coming soon

• Upcoming publication that goes into more detail

The research presented here has been supported by the President’s Emergency Plan for AIDS Relief (PEPFAR) through the United States Agency for International Development (USAID) under the terms of MEASURE Evaluation cooperative agreement GHA-A-00-08-00003-00. Views expressed are not necessarily those of PEPFAR, USAID or the United States government.

MEASURE Evaluation is implemented by the Carolina Population Center at the University of North Carolina at Chapel Hill in partnership with Futures Group, ICF International, John Snow, Inc., Management Sciences for Health, and Tulane University.

www.measureevaluation.org