Evaluation of geospatial methods to generate subnational HIV prevalence estimates for local level planning

The Subnational Estimates Working Group of the HIV Modelling Consortium

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Objective: There is evidence of substantial subnational variation in the HIV epidemic. However, robust spatial HIV data are often only available at high levels of geographic aggregation and not at the finer resolution needed for decision making. Therefore, spatial analysis methods that leverage available data to provide local estimates of HIV prevalence may be useful. Such methods exist but have not been formally compared when applied to HIV. Design/methods: Six candidate methods-including those used by the Joint United Nations Programme on HIV/AIDS to generate maps and a Bayesian geostatistical approach applied to other diseases-were used to generate maps and subnational estimates of HIV prevalence across three countries using cluster level data from household surveys. Two approaches were used to assess the accuracy of predictions: internal validation, whereby a proportion of input data is held back (test dataset) to challenge predictions; and comparison with location-specific data from household surveys in earlier years. Results: Each of the methods can generate usefully accurate predictions of prevalence at unsampled locations, with the magnitude of the error in predictions similar across approaches. However, the Bayesian geostatistical approach consistently gave marginally the strongest statistical performance across countries and validation procedures. Conclusions: Available methods may be able to furnish estimates of HIV prevalence at finer spatial scales than the data currently allow. The subnational variation revealed can be integrated into planning to ensure responsiveness to the spatial features of the epidemic. The Bayesian geostatistical approach is a promising strategy for integrating HIV data to generate robust local estimates.

Original languageEnglish (US)
Pages (from-to)1467-1474
Number of pages8
JournalAIDS
Volume30
Issue number9
DOIs
StatePublished - 2016

Fingerprint

HIV
Bayes Theorem
Spatial Analysis
United Nations
Decision Making
Acquired Immunodeficiency Syndrome
Joints
Surveys and Questionnaires

Keywords

  • Health planning/organization and administration
  • Health policy
  • HIV infections/epidemiology
  • HIV seroprevalence
  • HIV/infections prevention and control
  • Population surveillance/methods

ASJC Scopus subject areas

  • Immunology and Allergy
  • Immunology
  • Infectious Diseases

Cite this

Evaluation of geospatial methods to generate subnational HIV prevalence estimates for local level planning. / The Subnational Estimates Working Group of the HIV Modelling Consortium.

In: AIDS, Vol. 30, No. 9, 2016, p. 1467-1474.

Research output: Contribution to journalArticle

The Subnational Estimates Working Group of the HIV Modelling Consortium 2016, 'Evaluation of geospatial methods to generate subnational HIV prevalence estimates for local level planning', AIDS, vol. 30, no. 9, pp. 1467-1474. https://doi.org/10.1097/QAD.0000000000001075
The Subnational Estimates Working Group of the HIV Modelling Consortium. / Evaluation of geospatial methods to generate subnational HIV prevalence estimates for local level planning. In: AIDS. 2016 ; Vol. 30, No. 9. pp. 1467-1474.
@article{a6a5970840ba48769f2f69e7ac9edd86,
title = "Evaluation of geospatial methods to generate subnational HIV prevalence estimates for local level planning",
abstract = "Objective: There is evidence of substantial subnational variation in the HIV epidemic. However, robust spatial HIV data are often only available at high levels of geographic aggregation and not at the finer resolution needed for decision making. Therefore, spatial analysis methods that leverage available data to provide local estimates of HIV prevalence may be useful. Such methods exist but have not been formally compared when applied to HIV. Design/methods: Six candidate methods-including those used by the Joint United Nations Programme on HIV/AIDS to generate maps and a Bayesian geostatistical approach applied to other diseases-were used to generate maps and subnational estimates of HIV prevalence across three countries using cluster level data from household surveys. Two approaches were used to assess the accuracy of predictions: internal validation, whereby a proportion of input data is held back (test dataset) to challenge predictions; and comparison with location-specific data from household surveys in earlier years. Results: Each of the methods can generate usefully accurate predictions of prevalence at unsampled locations, with the magnitude of the error in predictions similar across approaches. However, the Bayesian geostatistical approach consistently gave marginally the strongest statistical performance across countries and validation procedures. Conclusions: Available methods may be able to furnish estimates of HIV prevalence at finer spatial scales than the data currently allow. The subnational variation revealed can be integrated into planning to ensure responsiveness to the spatial features of the epidemic. The Bayesian geostatistical approach is a promising strategy for integrating HIV data to generate robust local estimates.",
keywords = "Health planning/organization and administration, Health policy, HIV infections/epidemiology, HIV seroprevalence, HIV/infections prevention and control, Population surveillance/methods",
author = "{The Subnational Estimates Working Group of the HIV Modelling Consortium} and Hallett, {Timothy B.} and Anderson, {Sarah Jane} and Asante, {Cynthia Adobea} and Noah Bartlett and Victoria Bendaud and Samir Bhatt and Burgert, {Clara R.} and Cuadros, {Diego Fernando} and Janet Dzangare and Daniela Fecht and Gething, {Peter William} and Ghys, {Peter D.} and Guwani, {James M.} and Heard, {Nathan Joseph} and Ezekiel Kalipeni and Kandala, {Ngianga Bakwin} and Kim, {Andrea A.} and Kwao, {Isaiah Doe} and Joseph Larmarange and Manda, {Samuel O.M.} and Moise, {Imelda K.} and Imelda Moise and Mwai, {Daniel N.} and Samuel Mwalili and Ashton Shortridge and Frank Tanser and Ian Wanyeki and Leo Zulu",
year = "2016",
doi = "10.1097/QAD.0000000000001075",
language = "English (US)",
volume = "30",
pages = "1467--1474",
journal = "AIDS",
issn = "0269-9370",
publisher = "Lippincott Williams and Wilkins",
number = "9",

}

TY - JOUR

T1 - Evaluation of geospatial methods to generate subnational HIV prevalence estimates for local level planning

AU - The Subnational Estimates Working Group of the HIV Modelling Consortium

AU - Hallett, Timothy B.

AU - Anderson, Sarah Jane

AU - Asante, Cynthia Adobea

AU - Bartlett, Noah

AU - Bendaud, Victoria

AU - Bhatt, Samir

AU - Burgert, Clara R.

AU - Cuadros, Diego Fernando

AU - Dzangare, Janet

AU - Fecht, Daniela

AU - Gething, Peter William

AU - Ghys, Peter D.

AU - Guwani, James M.

AU - Heard, Nathan Joseph

AU - Kalipeni, Ezekiel

AU - Kandala, Ngianga Bakwin

AU - Kim, Andrea A.

AU - Kwao, Isaiah Doe

AU - Larmarange, Joseph

AU - Manda, Samuel O.M.

AU - Moise, Imelda K.

AU - Moise, Imelda

AU - Mwai, Daniel N.

AU - Mwalili, Samuel

AU - Shortridge, Ashton

AU - Tanser, Frank

AU - Wanyeki, Ian

AU - Zulu, Leo

PY - 2016

Y1 - 2016

N2 - Objective: There is evidence of substantial subnational variation in the HIV epidemic. However, robust spatial HIV data are often only available at high levels of geographic aggregation and not at the finer resolution needed for decision making. Therefore, spatial analysis methods that leverage available data to provide local estimates of HIV prevalence may be useful. Such methods exist but have not been formally compared when applied to HIV. Design/methods: Six candidate methods-including those used by the Joint United Nations Programme on HIV/AIDS to generate maps and a Bayesian geostatistical approach applied to other diseases-were used to generate maps and subnational estimates of HIV prevalence across three countries using cluster level data from household surveys. Two approaches were used to assess the accuracy of predictions: internal validation, whereby a proportion of input data is held back (test dataset) to challenge predictions; and comparison with location-specific data from household surveys in earlier years. Results: Each of the methods can generate usefully accurate predictions of prevalence at unsampled locations, with the magnitude of the error in predictions similar across approaches. However, the Bayesian geostatistical approach consistently gave marginally the strongest statistical performance across countries and validation procedures. Conclusions: Available methods may be able to furnish estimates of HIV prevalence at finer spatial scales than the data currently allow. The subnational variation revealed can be integrated into planning to ensure responsiveness to the spatial features of the epidemic. The Bayesian geostatistical approach is a promising strategy for integrating HIV data to generate robust local estimates.

AB - Objective: There is evidence of substantial subnational variation in the HIV epidemic. However, robust spatial HIV data are often only available at high levels of geographic aggregation and not at the finer resolution needed for decision making. Therefore, spatial analysis methods that leverage available data to provide local estimates of HIV prevalence may be useful. Such methods exist but have not been formally compared when applied to HIV. Design/methods: Six candidate methods-including those used by the Joint United Nations Programme on HIV/AIDS to generate maps and a Bayesian geostatistical approach applied to other diseases-were used to generate maps and subnational estimates of HIV prevalence across three countries using cluster level data from household surveys. Two approaches were used to assess the accuracy of predictions: internal validation, whereby a proportion of input data is held back (test dataset) to challenge predictions; and comparison with location-specific data from household surveys in earlier years. Results: Each of the methods can generate usefully accurate predictions of prevalence at unsampled locations, with the magnitude of the error in predictions similar across approaches. However, the Bayesian geostatistical approach consistently gave marginally the strongest statistical performance across countries and validation procedures. Conclusions: Available methods may be able to furnish estimates of HIV prevalence at finer spatial scales than the data currently allow. The subnational variation revealed can be integrated into planning to ensure responsiveness to the spatial features of the epidemic. The Bayesian geostatistical approach is a promising strategy for integrating HIV data to generate robust local estimates.

KW - Health planning/organization and administration

KW - Health policy

KW - HIV infections/epidemiology

KW - HIV seroprevalence

KW - HIV/infections prevention and control

KW - Population surveillance/methods

UR - http://www.scopus.com/inward/record.url?scp=84959186542&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84959186542&partnerID=8YFLogxK

U2 - 10.1097/QAD.0000000000001075

DO - 10.1097/QAD.0000000000001075

M3 - Article

VL - 30

SP - 1467

EP - 1474

JO - AIDS

JF - AIDS

SN - 0269-9370

IS - 9

ER -