The contribution of spatial analysis to understanding HIV/TB mortality in children: a structural equation modelling approach
Identifieur interne : 002380 ( Pmc/Corpus ); précédent : 002379; suivant : 002381The contribution of spatial analysis to understanding HIV/TB mortality in children: a structural equation modelling approach
Auteurs : Eustasius Musenge ; Penelope Vounatsou ; Mark Collinson ; Stephen Tollman ; Kathleen KahnSource :
- Global Health Action [ 1654-9716 ] ; 2013.
Abstract
South Africa accounts for more than a sixth of the global population of people infected with HIV and TB, ranking her highest in HIV/TB co-infection worldwide. Remote areas often bear the greatest burden of morbidity and mortality, yet there are spatial differences within rural settings.
The primary aim was to investigate HIV/TB mortality determinants and their spatial distribution in the rural Agincourt sub-district for children aged 1–5 years in 2004. Our secondary aim was to model how the associated factors were interrelated as either underlying or proximate factors of child mortality using pathway analysis based on a Mosley-Chen conceptual framework.
We conducted a secondary data analysis based on cross-sectional data collected in 2004 from the Agincourt sub-district in rural northeast South Africa. Child HIV/TB death was the outcome measure derived from physician assessed verbal autopsy. Modelling used multiple logit regression models with and without spatial household random effects. Structural equation models were used in modelling the complex relationships between multiple exposures and the outcome (child HIV/TB mortality) as relayed on a conceptual framework.
Fifty-four of 6,692 children aged 1–5 years died of HIV/TB, from a total of 5,084 households. Maternal death had the greatest effect on child HIV/TB mortality (adjusted odds ratio=4.00; 95% confidence interval=1.01–15.80). A protective effect was found in households with better socio-economic status and when the child was older. Spatial models disclosed that the areas which experienced the greatest child HIV/TB mortality were those without any health facility.
Low socio-economic status and maternal deaths impacted indirectly and directly on child mortality, respectively. These factors are major concerns locally and should be used in formulating interventions to reduce child mortality. Spatial prediction maps can guide policy makers to target interventions where they are most needed.
Url:
DOI: 10.3402/gha.v6i0.19266
PubMed: 23364095
PubMed Central: 3556702
Links to Exploration step
PMC:3556702Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">The contribution of spatial analysis to understanding HIV/TB mortality in children: a structural equation modelling approach</title><author><name sortKey="Musenge, Eustasius" sort="Musenge, Eustasius" uniqKey="Musenge E" first="Eustasius" last="Musenge">Eustasius Musenge</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0002">Division of Epidemiology and Biostatistics, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation></author><author><name sortKey="Vounatsou, Penelope" sort="Vounatsou, Penelope" uniqKey="Vounatsou P" first="Penelope" last="Vounatsou">Penelope Vounatsou</name><affiliation><nlm:aff id="AF0003">Swiss Tropical and Public Health Institute, University of Basel, Basel, Switzerland</nlm:aff></affiliation></author><author><name sortKey="Collinson, Mark" sort="Collinson, Mark" uniqKey="Collinson M" first="Mark" last="Collinson">Mark Collinson</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0004">Centre for Global Health Research, Umeå University, Umeå, Sweden</nlm:aff></affiliation><affiliation><nlm:aff id="AF0005">INDEPTH Network, Accra, Ghana</nlm:aff></affiliation></author><author><name sortKey="Tollman, Stephen" sort="Tollman, Stephen" uniqKey="Tollman S" first="Stephen" last="Tollman">Stephen Tollman</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0004">Centre for Global Health Research, Umeå University, Umeå, Sweden</nlm:aff></affiliation><affiliation><nlm:aff id="AF0005">INDEPTH Network, Accra, Ghana</nlm:aff></affiliation></author><author><name sortKey="Kahn, Kathleen" sort="Kahn, Kathleen" uniqKey="Kahn K" first="Kathleen" last="Kahn">Kathleen Kahn</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0004">Centre for Global Health Research, Umeå University, Umeå, Sweden</nlm:aff></affiliation><affiliation><nlm:aff id="AF0005">INDEPTH Network, Accra, Ghana</nlm:aff></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">PMC</idno><idno type="pmid">23364095</idno><idno type="pmc">3556702</idno><idno type="url">http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3556702</idno><idno type="RBID">PMC:3556702</idno><idno type="doi">10.3402/gha.v6i0.19266</idno><date when="2013">2013</date><idno type="wicri:Area/Pmc/Corpus">002380</idno><idno type="wicri:explorRef" wicri:stream="Pmc" wicri:step="Corpus" wicri:corpus="PMC">002380</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en" level="a" type="main">The contribution of spatial analysis to understanding HIV/TB mortality in children: a structural equation modelling approach</title><author><name sortKey="Musenge, Eustasius" sort="Musenge, Eustasius" uniqKey="Musenge E" first="Eustasius" last="Musenge">Eustasius Musenge</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0002">Division of Epidemiology and Biostatistics, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation></author><author><name sortKey="Vounatsou, Penelope" sort="Vounatsou, Penelope" uniqKey="Vounatsou P" first="Penelope" last="Vounatsou">Penelope Vounatsou</name><affiliation><nlm:aff id="AF0003">Swiss Tropical and Public Health Institute, University of Basel, Basel, Switzerland</nlm:aff></affiliation></author><author><name sortKey="Collinson, Mark" sort="Collinson, Mark" uniqKey="Collinson M" first="Mark" last="Collinson">Mark Collinson</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0004">Centre for Global Health Research, Umeå University, Umeå, Sweden</nlm:aff></affiliation><affiliation><nlm:aff id="AF0005">INDEPTH Network, Accra, Ghana</nlm:aff></affiliation></author><author><name sortKey="Tollman, Stephen" sort="Tollman, Stephen" uniqKey="Tollman S" first="Stephen" last="Tollman">Stephen Tollman</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0004">Centre for Global Health Research, Umeå University, Umeå, Sweden</nlm:aff></affiliation><affiliation><nlm:aff id="AF0005">INDEPTH Network, Accra, Ghana</nlm:aff></affiliation></author><author><name sortKey="Kahn, Kathleen" sort="Kahn, Kathleen" uniqKey="Kahn K" first="Kathleen" last="Kahn">Kathleen Kahn</name><affiliation><nlm:aff id="AF0001">MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</nlm:aff></affiliation><affiliation><nlm:aff id="AF0004">Centre for Global Health Research, Umeå University, Umeå, Sweden</nlm:aff></affiliation><affiliation><nlm:aff id="AF0005">INDEPTH Network, Accra, Ghana</nlm:aff></affiliation></author></analytic><series><title level="j">Global Health Action</title><idno type="ISSN">1654-9716</idno><idno type="eISSN">1654-9880</idno><imprint><date when="2013">2013</date></imprint></series></biblStruct></sourceDesc></fileDesc><profileDesc><textClass></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en"><sec id="st1"><title>Background</title><p>South Africa accounts for more than a sixth of the global population of people infected with HIV and TB, ranking her highest in HIV/TB co-infection worldwide. Remote areas often bear the greatest burden of morbidity and mortality, yet there are spatial differences within rural settings.</p></sec><sec id="st2"><title>Objectives</title><p>The primary aim was to investigate HIV/TB mortality determinants and their spatial distribution in the rural Agincourt sub-district for children aged 1–5 years in 2004. Our secondary aim was to model how the associated factors were interrelated as either underlying or proximate factors of child mortality using pathway analysis based on a Mosley-Chen conceptual framework.</p></sec><sec id="st3"><title>Methods</title><p>We conducted a secondary data analysis based on cross-sectional data collected in 2004 from the Agincourt sub-district in rural northeast South Africa. Child HIV/TB death was the outcome measure derived from physician assessed verbal autopsy. Modelling used multiple logit regression models with and without spatial household random effects. Structural equation models were used in modelling the complex relationships between multiple exposures and the outcome (child HIV/TB mortality) as relayed on a conceptual framework.</p></sec><sec id="st4"><title>Results</title><p>Fifty-four of 6,692 children aged 1–5 years died of HIV/TB, from a total of 5,084 households. Maternal death had the greatest effect on child HIV/TB mortality (adjusted odds ratio=4.00; 95% confidence interval=1.01–15.80). A protective effect was found in households with better socio-economic status and when the child was older. Spatial models disclosed that the areas which experienced the greatest child HIV/TB mortality were those without any health facility.</p></sec><sec id="st5"><title>Conclusion</title><p>Low socio-economic status and maternal deaths impacted indirectly and directly on child mortality, respectively. These factors are major concerns locally and should be used in formulating interventions to reduce child mortality. Spatial prediction maps can guide policy makers to target interventions where they are most needed.</p></sec></div></front><back><div1 type="bibliography"><listBibl><biblStruct></biblStruct><biblStruct></biblStruct><biblStruct></biblStruct><biblStruct><analytic><author><name sortKey="Bradshaw, D" uniqKey="Bradshaw D">D Bradshaw</name></author><author><name sortKey="Bourne, D" uniqKey="Bourne D">D Bourne</name></author><author><name sortKey="Nannan, N" uniqKey="Nannan N">N Nannan</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Doherty, T" uniqKey="Doherty T">T Doherty</name></author><author><name sortKey="Sanders, D" uniqKey="Sanders D">D Sanders</name></author><author><name sortKey="Goga, A" uniqKey="Goga A">A Goga</name></author><author><name sortKey="Jackson, D" uniqKey="Jackson D">D Jackson</name></author></analytic></biblStruct><biblStruct></biblStruct><biblStruct><analytic><author><name sortKey="Waller, La" uniqKey="Waller L">LA Waller</name></author><author><name sortKey="Gotway, Ca" uniqKey="Gotway C">CA Gotway</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Mosley, Wh" uniqKey="Mosley W">WH Mosley</name></author><author><name sortKey="Chen, Lc" uniqKey="Chen L">LC Chen</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Rutstein, So" uniqKey="Rutstein S">SO Rutstein</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Filteau, S" uniqKey="Filteau S">S Filteau</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Guerrero, L" uniqKey="Guerrero L">L Guerrero</name></author><author><name sortKey="La Valley, A" uniqKey="La Valley A">A La Valley</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Hargreaves, J" uniqKey="Hargreaves J">J Hargreaves</name></author><author><name sortKey="Collinson, M" uniqKey="Collinson M">M Collinson</name></author><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author><author><name sortKey="Clark, S" uniqKey="Clark S">S Clark</name></author><author><name sortKey="Tollman, S" uniqKey="Tollman S">S Tollman</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Marshall, Rj" uniqKey="Marshall R">RJ Marshall</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Gosoniu, L" uniqKey="Gosoniu L">L Gosoniu</name></author><author><name sortKey="Vounatsou, P" uniqKey="Vounatsou P">P Vounatsou</name></author><author><name sortKey="Tami, A" uniqKey="Tami A">A Tami</name></author><author><name sortKey="Nathan, R" uniqKey="Nathan R">R Nathan</name></author><author><name sortKey="Grundmann, H" uniqKey="Grundmann H">H Grundmann</name></author><author><name sortKey="Lengeler, C" uniqKey="Lengeler C">C Lengeler</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Moonan, Pk" uniqKey="Moonan P">PK Moonan</name></author><author><name sortKey="Bayona, M" uniqKey="Bayona M">M Bayona</name></author><author><name sortKey="Quitugua, Tn" uniqKey="Quitugua T">TN Quitugua</name></author><author><name sortKey="Oppong, J" uniqKey="Oppong J">J Oppong</name></author><author><name sortKey="Dunbar, D" uniqKey="Dunbar D">D Dunbar</name></author><author><name sortKey="Jost, Kc" uniqKey="Jost K">KC Jost</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Sartorius, B" uniqKey="Sartorius B">B Sartorius</name></author><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author><author><name sortKey="Vounatsou, P" uniqKey="Vounatsou P">P Vounatsou</name></author><author><name sortKey="Collinson, M" uniqKey="Collinson M">M Collinson</name></author><author><name sortKey="Tollman, S" uniqKey="Tollman S">S Tollman</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Middelkoop, K" uniqKey="Middelkoop K">K Middelkoop</name></author><author><name sortKey="Bekker, Lg" uniqKey="Bekker L">LG Bekker</name></author><author><name sortKey="Morrow, C" uniqKey="Morrow C">C Morrow</name></author><author><name sortKey="Zwane, E" uniqKey="Zwane E">E Zwane</name></author><author><name sortKey="Wood, R" uniqKey="Wood R">R Wood</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Tanser, F" uniqKey="Tanser F">F Tanser</name></author><author><name sortKey="Lesueur, D" uniqKey="Lesueur D">D LeSueur</name></author><author><name sortKey="Solarsh, G" uniqKey="Solarsh G">G Solarsh</name></author><author><name sortKey="Wilkinson, D" uniqKey="Wilkinson D">D Wilkinson</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Musenge, E" uniqKey="Musenge E">E Musenge</name></author><author><name sortKey="Vounatsou, P" uniqKey="Vounatsou P">P Vounatsou</name></author><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author><author><name sortKey="Tollman, Sm" uniqKey="Tollman S">SM Tollman</name></author><author><name sortKey="Collinson, Ma" uniqKey="Collinson M">MA Collinson</name></author><author><name sortKey="Clark, Sj" uniqKey="Clark S">SJ Clark</name></author><author><name sortKey="Twine, R" uniqKey="Twine R">R Twine</name></author><author><name sortKey="Clark, Bd" uniqKey="Clark B">BD Clark</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Clark, Sj" uniqKey="Clark S">SJ Clark</name></author><author><name sortKey="Collinson, Ma" uniqKey="Collinson M">MA Collinson</name></author><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author><author><name sortKey="Drullinger, K" uniqKey="Drullinger K">K Drullinger</name></author><author><name sortKey="Tollman, Sm" uniqKey="Tollman S">SM Tollman</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author><author><name sortKey="Tollman, Sm" uniqKey="Tollman S">SM Tollman</name></author><author><name sortKey="Garenne, M" uniqKey="Garenne M">M Garenne</name></author><author><name sortKey="Gear, Js" uniqKey="Gear J">JS Gear</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Tollman, Sm" uniqKey="Tollman S">SM Tollman</name></author><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author><author><name sortKey="Sartorius, B" uniqKey="Sartorius B">B Sartorius</name></author><author><name sortKey="Collinson, Ma" uniqKey="Collinson M">MA Collinson</name></author><author><name sortKey="Clark, Sj" uniqKey="Clark S">SJ Clark</name></author><author><name sortKey="Garenne, Ml" uniqKey="Garenne M">ML Garenne</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Meade, Ms" uniqKey="Meade M">MS Meade</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Lewis, Jjc" uniqKey="Lewis J">JJC Lewis</name></author><author><name sortKey="Donnelly, Ca" uniqKey="Donnelly C">CA Donnelly</name></author><author><name sortKey="Mare, P" uniqKey="Mare P">P Mare</name></author><author><name sortKey="Mupambireyi, Z" uniqKey="Mupambireyi Z">Z Mupambireyi</name></author><author><name sortKey="Garnett, Gp" uniqKey="Garnett G">GP Garnett</name></author><author><name sortKey="Gregson, S" uniqKey="Gregson S">S Gregson</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Boerma, J" uniqKey="Boerma J">J Boerma</name></author><author><name sortKey="Weir, Ss" uniqKey="Weir S">SS Weir</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Lopman, B" uniqKey="Lopman B">B Lopman</name></author><author><name sortKey="Nyamukapa, C" uniqKey="Nyamukapa C">C Nyamukapa</name></author><author><name sortKey="Mushati, P" uniqKey="Mushati P">P Mushati</name></author><author><name sortKey="Mupambireyi, Z" uniqKey="Mupambireyi Z">Z Mupambireyi</name></author><author><name sortKey="Mason, P" uniqKey="Mason P">P Mason</name></author><author><name sortKey="Garnett, Gp" uniqKey="Garnett G">GP Garnett</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Cressie, Nac" uniqKey="Cressie N">NAC Cressie</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Banerjee, S" uniqKey="Banerjee S">S Banerjee</name></author><author><name sortKey="Carlin, Bp" uniqKey="Carlin B">BP Carlin</name></author><author><name sortKey="Gelfand, Ae" uniqKey="Gelfand A">AE Gelfand</name></author></analytic></biblStruct><biblStruct></biblStruct><biblStruct><analytic><author><name sortKey="Kimani Murage, Ew" uniqKey="Kimani Murage E">EW Kimani-Murage</name></author><author><name sortKey="Manderson, L" uniqKey="Manderson L">L Manderson</name></author><author><name sortKey="Norris, Sa" uniqKey="Norris S">SA Norris</name></author><author><name sortKey="Kahn, K" uniqKey="Kahn K">K Kahn</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Belitz, C" uniqKey="Belitz C">C Belitz</name></author><author><name sortKey="Brezger, A" uniqKey="Brezger A">A Brezger</name></author><author><name sortKey="Kneib, T" uniqKey="Kneib T">T Kneib</name></author><author><name sortKey="Lang, S" uniqKey="Lang S">S Lang</name></author><author><name sortKey="Fronk, E" uniqKey="Fronk E">E Fronk</name></author><author><name sortKey="Heinzl, F" uniqKey="Heinzl F">F Heinzl</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Coetzee, D" uniqKey="Coetzee D">D Coetzee</name></author><author><name sortKey="Hilderbrand, K" uniqKey="Hilderbrand K">K Hilderbrand</name></author><author><name sortKey="Boulle, A" uniqKey="Boulle A">A Boulle</name></author><author><name sortKey="Draper, B" uniqKey="Draper B">B Draper</name></author><author><name sortKey="Abdullah, F" uniqKey="Abdullah F">F Abdullah</name></author><author><name sortKey="Goemaere, E" uniqKey="Goemaere E">E Goemaere</name></author></analytic></biblStruct><biblStruct></biblStruct><biblStruct><analytic><author><name sortKey="Ndirangu, J" uniqKey="Ndirangu J">J Ndirangu</name></author><author><name sortKey="Newell, M L" uniqKey="Newell M">M-L Newell</name></author><author><name sortKey="Tanser, F" uniqKey="Tanser F">F Tanser</name></author><author><name sortKey="Herbst, Aj" uniqKey="Herbst A">AJ Herbst</name></author><author><name sortKey="Bland, R" uniqKey="Bland R">R Bland</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Katapa, R" uniqKey="Katapa R">R Katapa</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Wiegers, E" uniqKey="Wiegers E">E Wiegers</name></author><author><name sortKey="Curry, J" uniqKey="Curry J">J Curry</name></author><author><name sortKey="Garbero, A" uniqKey="Garbero A">A Garbero</name></author><author><name sortKey="Hourihan, J" uniqKey="Hourihan J">J Hourihan</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Oni, S" uniqKey="Oni S">S Oni</name></author><author><name sortKey="Obi, C" uniqKey="Obi C">C Obi</name></author><author><name sortKey="Okorie, A" uniqKey="Okorie A">A Okorie</name></author><author><name sortKey="Thabede, D" uniqKey="Thabede D">D Thabede</name></author><author><name sortKey="Jordan, A" uniqKey="Jordan A">A Jordan</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Collinson, Ma" uniqKey="Collinson M">MA Collinson</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Posel, D" uniqKey="Posel D">D Posel</name></author><author><name sortKey="Casale, D" uniqKey="Casale D">D Casale</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Hahn, Mb" uniqKey="Hahn M">MB Hahn</name></author><author><name sortKey="Riederer, Am" uniqKey="Riederer A">AM Riederer</name></author><author><name sortKey="Foster, So" uniqKey="Foster S">SO Foster</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Beguy, D" uniqKey="Beguy D">D Beguy</name></author><author><name sortKey="Bocquier, P" uniqKey="Bocquier P">P Bocquier</name></author><author><name sortKey="Zulu, Em" uniqKey="Zulu E">EM Zulu</name></author></analytic></biblStruct><biblStruct></biblStruct><biblStruct><analytic><author><name sortKey="Kline, Rb" uniqKey="Kline R">RB Kline</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Bollen, Ka" uniqKey="Bollen K">KA Bollen</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Bollen, Ka" uniqKey="Bollen K">KA Bollen</name></author><author><name sortKey="Curran, Pj" uniqKey="Curran P">PJ Curran</name></author></analytic></biblStruct></listBibl></div1></back></TEI><pmc article-type="research-article"><pmc-dir>properties open_access</pmc-dir>
<front><journal-meta><journal-id journal-id-type="nlm-ta">Glob Health Action</journal-id><journal-id journal-id-type="iso-abbrev">Glob Health Action</journal-id><journal-id journal-id-type="publisher-id">GHA</journal-id><journal-title-group><journal-title>Global Health Action</journal-title></journal-title-group><issn pub-type="ppub">1654-9716</issn><issn pub-type="epub">1654-9880</issn><publisher><publisher-name>Co-Action Publishing</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="pmid">23364095</article-id><article-id pub-id-type="pmc">3556702</article-id><article-id pub-id-type="publisher-id">19266</article-id><article-id pub-id-type="doi">10.3402/gha.v6i0.19266</article-id><article-categories><subj-group subj-group-type="heading"><subject>Building New Knowledge Supplement</subject></subj-group></article-categories><title-group><article-title>The contribution of spatial analysis to understanding HIV/TB mortality in children: a structural equation modelling approach</article-title></title-group><contrib-group><contrib contrib-type="author"><name><surname>Musenge</surname><given-names>Eustasius</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0002">2</xref><xref ref-type="corresp" rid="cor1">*</xref></contrib><contrib contrib-type="author"><name><surname>Vounatsou</surname><given-names>Penelope</given-names></name><xref ref-type="aff" rid="AF0003">3</xref></contrib><contrib contrib-type="author"><name><surname>Collinson</surname><given-names>Mark</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0004">4</xref><xref ref-type="aff" rid="AF0005">5</xref></contrib><contrib contrib-type="author"><name><surname>Tollman</surname><given-names>Stephen</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0004">4</xref><xref ref-type="aff" rid="AF0005">5</xref></contrib><contrib contrib-type="author"><name><surname>Kahn</surname><given-names>Kathleen</given-names></name><xref ref-type="aff" rid="AF0001">1</xref><xref ref-type="aff" rid="AF0004">4</xref><xref ref-type="aff" rid="AF0005">5</xref></contrib></contrib-group><aff id="AF0001"><label>1</label>MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</aff><aff id="AF0002"><label>2</label>Division of Epidemiology and Biostatistics, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa</aff><aff id="AF0003"><label>3</label>Swiss Tropical and Public Health Institute, University of Basel, Basel, Switzerland</aff><aff id="AF0004"><label>4</label>Centre for Global Health Research, Umeå University, Umeå, Sweden</aff><aff id="AF0005"><label>5</label>INDEPTH Network, Accra, Ghana</aff><author-notes><corresp id="cor1"><label>*</label><bold>Eustasius Musenge</bold>, Division of Epidemiology and Biostatistics, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, 7 York Road, Parktown Johannesburg 2193, South Africa. Tel: +27 11 717 2610. Email: <email xlink:href="Eustasius.Musenge@wits.ac.za">Eustasius.Musenge@wits.ac.za</email></corresp></author-notes><pub-date pub-type="epub"><day>24</day><month>1</month><year>2013</year></pub-date><pub-date pub-type="collection"><year>2013</year></pub-date><volume>6</volume><elocation-id content-type="doi">10.3402/gha.v6i0.19266</elocation-id><history><date date-type="received"><day>02</day><month>8</month><year>2012</year></date><date date-type="rev-recd"><day>25</day><month>10</month><year>2012</year></date><date date-type="accepted"><day>26</day><month>10</month><year>2012</year></date></history><permissions><copyright-statement>© 2013 Eustasius Musenge et al.</copyright-statement><copyright-year>2013</copyright-year><license license-type="open-access" xlink:href="http://creativecommons.org/licenses/by/2.0/"><license-p>This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</license-p></license></permissions><abstract><sec id="st1"><title>Background</title><p>South Africa accounts for more than a sixth of the global population of people infected with HIV and TB, ranking her highest in HIV/TB co-infection worldwide. Remote areas often bear the greatest burden of morbidity and mortality, yet there are spatial differences within rural settings.</p></sec><sec id="st2"><title>Objectives</title><p>The primary aim was to investigate HIV/TB mortality determinants and their spatial distribution in the rural Agincourt sub-district for children aged 1–5 years in 2004. Our secondary aim was to model how the associated factors were interrelated as either underlying or proximate factors of child mortality using pathway analysis based on a Mosley-Chen conceptual framework.</p></sec><sec id="st3"><title>Methods</title><p>We conducted a secondary data analysis based on cross-sectional data collected in 2004 from the Agincourt sub-district in rural northeast South Africa. Child HIV/TB death was the outcome measure derived from physician assessed verbal autopsy. Modelling used multiple logit regression models with and without spatial household random effects. Structural equation models were used in modelling the complex relationships between multiple exposures and the outcome (child HIV/TB mortality) as relayed on a conceptual framework.</p></sec><sec id="st4"><title>Results</title><p>Fifty-four of 6,692 children aged 1–5 years died of HIV/TB, from a total of 5,084 households. Maternal death had the greatest effect on child HIV/TB mortality (adjusted odds ratio=4.00; 95% confidence interval=1.01–15.80). A protective effect was found in households with better socio-economic status and when the child was older. Spatial models disclosed that the areas which experienced the greatest child HIV/TB mortality were those without any health facility.</p></sec><sec id="st5"><title>Conclusion</title><p>Low socio-economic status and maternal deaths impacted indirectly and directly on child mortality, respectively. These factors are major concerns locally and should be used in formulating interventions to reduce child mortality. Spatial prediction maps can guide policy makers to target interventions where they are most needed.</p></sec></abstract><kwd-group><kwd>HIV/TB</kwd><kwd>child mortality</kwd><kwd>conceptual framework</kwd><kwd>spatial analysis</kwd><kwd>pathway analysis</kwd><kwd>South Africa</kwd></kwd-group></article-meta><notes><fn-group><fn><p><ext-link ext-link-type="uri" xlink:href="http://www.globalhealthaction.net/index.php/gha/rt/suppFiles/19266/0">Appendices</ext-link> available online under Reading Tools.</p></fn></fn-group></notes></front><body><p>Tuberculosis (TB) and the HIV epidemic are among the top public health priorities especially in sub-Saharan Africa. The sub-Saharan African region, which constitutes 10% of the world's population, is home to more than 67% (22.9 million) of people living with HIV (<xref ref-type="bibr" rid="CIT0001">1</xref>). TB, which is preventable and curable, was responsible for 25% of HIV-related deaths in 2009 and infected 33% of those living with HIV. As a result, the World Health Organization (WHO) aims to reduce TB-related deaths by 1 million by 2015 (<xref ref-type="bibr" rid="CIT0002">2</xref>). South Africa's population is about 7% [50.6 million (<xref ref-type="bibr" rid="CIT0003">3</xref>)] of the sub-Saharan African population, yet accounts for 24.6% (5.63 million of the 22.9 million) of people living with HIV and an estimated 67% of these have TB. South Africa ranks as the country with the highest co-infection rates (<xref ref-type="bibr" rid="CIT0001">1</xref>).</p><p>Children are affected in two ways: a very high proportion of child deaths (60% in the year 2000) are attributed to the HIV epidemic (<xref ref-type="bibr" rid="CIT0004">4</xref>), and approximately 1.9 million children have been orphaned by the HIV epidemic in South Africa (<xref ref-type="bibr" rid="CIT0001">1</xref>). The nation at the onset of the 21st century had an estimated 70,000 children [about 25% rate of mother-to-child-transmissions (MTCT)] newly infected, whereas 10 years later only 10,000 [3.5% rate of MTCT (<xref ref-type="bibr" rid="CIT0005">5</xref>)], which is indicative of an improved and more effective mother-to-child prevention programme (<xref ref-type="bibr" rid="CIT0006">6</xref>). In public health, determining patterns of disease occurrence provides the first steps toward increased understanding of determinants and potentially greater control of the disease (<xref ref-type="bibr" rid="CIT0007">7</xref>). Understanding the associated factors of child HIV/TB mortality and how these are geographically distributed is key for effective interventions.</p><p>Several determinants of child mortality have been reported over the past three decades. In the 1980s, researchers explored medical, socio-economic, and demographic causes of child mortality (<xref ref-type="bibr" rid="CIT0008">8</xref>). In the 1990s, fertility behaviour, nutritional status, breastfeeding and infant feeding, the use of health services by mothers and children, ecological variables, and socio-economic status were studied (<xref ref-type="bibr" rid="CIT0009">9</xref>). In the 21st century, the focus has been on HIV exposure, poor maternal health, inadequate infant care, increased exposure to infections, deaths of parents or caregivers, immune system abnormalities, poor nutrition, reduced breastfeeding and antiretroviral treatment exposure (<xref ref-type="bibr" rid="CIT0010">10</xref>). These determinants are interrelated and can be combined into a conceptual framework. This combined with spatial aspects forms social ecological models, which are a component of the fast developing systems science that examines multiple effects and interrelatedness of social elements in an environment (<xref ref-type="bibr" rid="CIT0011">11</xref>). Analysis of such frameworks can now be handled with the aid of structural equation models also known as pathway analysis.</p><p>The motivation for spatial analysis in epidemiological modelling is based on the notion that people living in a household and those in close proximity share similar exposures, which impact on the observed outcomes. In HIV/TB epidemiological research, the focus has largely been on drug development, societal integration, personal nutrition, physical health, and care giving (<xref ref-type="bibr" rid="CIT0008">8</xref>–<xref ref-type="bibr" rid="CIT0010">10</xref>, <xref ref-type="bibr" rid="CIT0012">12</xref>). Country-wide or region-specific pooled results are often reported, disregarding geographical confounding that commonly exists. Little has been done to interlink factors associated with child HIV/TB mortality or model their impact on areas of residence (households) (<xref ref-type="bibr" rid="CIT0013">13</xref>). The few studies implementing spatial analysis have looked across different infectious diseases (<xref ref-type="bibr" rid="CIT0014">14</xref>–<xref ref-type="bibr" rid="CIT0016">16</xref>), used different study designs (cross-sectional and longitudinal) (<xref ref-type="bibr" rid="CIT0017">17</xref>, <xref ref-type="bibr" rid="CIT0018">18</xref>), and considered varying levels of analysis (village, household, and individual) (<xref ref-type="bibr" rid="CIT0016">16</xref>, <xref ref-type="bibr" rid="CIT0018">18</xref>, <xref ref-type="bibr" rid="CIT0019">19</xref>). As this is an emerging field, few studies have focused on HIV/TB cause-specific mortality assessing geographical differentials in South Africa over time (<xref ref-type="bibr" rid="CIT0019">19</xref>).</p><p>The primary aim of this paper was to investigate the cause-specific (HIV/TB) mortality determinants and their spatial (geographical) distribution in the Agincourt sub-district of rural northeast South Africa for children aged 1–5 years in 2004. Our secondary aim was to model how the associated factors were interrelated as either direct (underlying) or indirect (proximal) factors of child HIV/TB mortality using pathway analysis based on the Mosley-Chen conceptual framework that incorporates social, biological, and ecological determinants of child mortality (<xref ref-type="bibr" rid="CIT0008">8</xref>).</p><sec sec-type="methods" id="S0001"><title>Methods</title><sec id="S20002"><title>Agincourt sub-district data and study design</title><p>The Agincourt health and socio-demographic surveillance system was established in 1992 in a remote rural sub-district in the Bushbuckridge Municipality of Mpumalanga Province, northeast South Africa near the Mozambique border (<xref ref-type="bibr" rid="CIT0020">20</xref>). There are annual census updates that collect socio-demographic information including in-migration and out-migration, mortality, fertility, health, and behavioural data (<xref ref-type="bibr" rid="CIT0020">20</xref>).</p><p>In this study, we extracted data for 2004 describing a cross-sectional study of households in the study site. In that year, the Agincourt health and socio-demographic surveillance system covered a sub-district population of over 70,000 persons living in approximately 12,000 households scattered throughout 21 neighbouring villages. There were several health facilities in the area as shown in <xref ref-type="fig" rid="F0001">Fig. 1</xref>.</p><fig id="F0001" position="float"><label>Fig. 1</label><caption><p>Agincourt sub-district in 2004 showing villages (grey shading) schools (flags), and health facilities (circled grey crosses).</p></caption><graphic xlink:href="GHA-6-19266-g001"></graphic></fig><p>Cause of death data were obtained through verbal autopsies conducted on every recorded death during the year (<xref ref-type="bibr" rid="CIT0021">21</xref>). Trained lay field workers conducted interviews within 1–11 months after a death, with the closest caregiver of the deceased in his/her mother tongue. Two medical practitioners independently determined cause of death; a third diagnosis was sought to resolve discrepancies. Their consensus cause of death was classified according to the WHO International Classification of Diseases 10 (ICD10) (<xref ref-type="bibr" rid="CIT0022">22</xref>). The reported signs and symptoms ascertained HIV/TB mortality in children, and in some instances this was substantiated by the mother's cause of death (<xref ref-type="bibr" rid="CIT0022">22</xref>). Over 90% of the health and socio-demographic surveillance system (HDSS) households were geo-coded, thus enabling spatial analyses at household level.</p></sec><sec id="S20003"><title>Dependent and independent variables</title><p>The study sample included all children aged between 1 and less than 5 years who were registered in the Agincourt HDSS between January and December 2004. The independent variables used were child's gender, nationality and age; mother's age, parity and death; gender of household head, cumulative deaths in household, number of household dwellers, minimum distance to health facility, and household socio-economic status. The latitude and longitude of each household were used to construct the variables for the spatial correlation (random effects), based on the straight line (Euclidean) distances between all pairs of households. The dependent variable was death due to HIV and TB determined by the ICD10 codes A16–A19<xref ref-type="fn" rid="FN0001">1</xref> for TB and B20–B24<xref ref-type="fn" rid="FN0002">2</xref> for HIV. The verbal autopsy cannot reliably distinguish HIV and TB due to low sensitivity and specificity; thus, these were combined in our analysis (<xref ref-type="bibr" rid="CIT0023">23</xref>).</p></sec><sec id="S20004"><title>Conceptual framework: determinants of child mortality</title><p>As a way of interrelating child mortality determinants, we formulated a hybrid conceptual framework for child mortality determinants and disease ecology as proposed by Mosley-Chen and Meade (<xref ref-type="bibr" rid="CIT0008">8</xref>, <xref ref-type="bibr" rid="CIT0024">24</xref>). This approach integrates ecological, social, and biological variables in the analysis, with the idea that these determinants operate through a common set of either ‘proximate’ (intermediate or indirect) or ‘underlying’ (direct) variables that impact on HIV/TB child mortality (<xref ref-type="bibr" rid="CIT0008">8</xref>, <xref ref-type="bibr" rid="CIT0025">25</xref>–<xref ref-type="bibr" rid="CIT0027">27</xref>). Variables are either endogenous (outcome variables) or exogenous (explanatory) or both, and these are modelled using structural equation models. Details of the structural equation model are given in <ext-link ext-link-type="uri" xlink:href="http://www.globalhealthaction.net/index.php/gha/rt/suppFiles/19266/0">Appendix 1</ext-link></p><p>We used the observed variables: child specific, mother specific, household related; and unobserved (latent) variables: socio-economic quintiles and spatial (ecological) variables. We tested the pathways shown in <xref ref-type="fig" rid="F0002">Fig. 2</xref>, to measure the direct and indirect determinants of child mortality for the Agincourt sub-district in 2004. Two approaches that complemented each other were used in our modelling. Firstly, we modelled the HIV/TB mortality determinants for children aged 1–5 years in 2004 and adjusted for their spatial random effects. In relation to our <xref ref-type="fig" rid="F0002">Fig. 2</xref>, we estimated the direct effects pathways (red arrows) from all categories of the explanatory (exogenous) variables controlling for confounders amongst them and their spatial or geographical effects (orange arrows). Results for these are presented in <xref ref-type="table" rid="T0001">Table 1</xref> and <xref ref-type="fig" rid="F0003">Fig. 3</xref>. Lastly, using the significant variables as outcomes (endogenous), we determined the indirect effects (proximal denoted by purple arrows) within and between categories to establish those factors impacting on child HIV/TB mortality.
</p><fig id="F0002" position="float"><label>Fig. 2</label><caption><p>Conceptual framework for modelling childhood HIV/TB-related mortality [adapted from Meade (<xref ref-type="bibr" rid="CIT0024">24</xref>)]. The red arrows reflect a direct effect on mortality and the purple (explanatory variables) and orange (spatial random effects) arrows depict an indirect effect.</p></caption><graphic xlink:href="GHA-6-19266-g002"></graphic></fig><fig id="F0003" position="float"><label>Fig. 3</label><caption><p>Child HIV/TB mortality: posterior adjusted odds ratio map, Agincourt sub-district, 2004.</p></caption><graphic xlink:href="GHA-6-19266-g003"></graphic></fig><table-wrap id="T0001" position="float"><label>Table 1</label><caption><p>Child HIV/TB mortality: descriptive statistics and results of two multiple logit regression analyses, Agincourt sub-district 2004</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1">Variable</th><th align="center" rowspan="1" colspan="1">Summary number (%)</th><th align="center" rowspan="1" colspan="1">Univariate logit coefficient (standard error)</th><th align="center" rowspan="1" colspan="1">Non-spatial multiple logit coefficient (standard error)</th><th align="center" rowspan="1" colspan="1">Spatial Multiple logit coefficient (standard error)</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Age of child in years<xref ref-type="table-fn" rid="TF0003">1</xref></td><td align="center" rowspan="1" colspan="1">2.77±1.20</td><td align="center" rowspan="1" colspan="1">−0.617 (0.131)<xref ref-type="table-fn" rid="TF0002">**</xref></td><td align="center" rowspan="1" colspan="1">−0.601 (0.146)<xref ref-type="table-fn" rid="TF0002">**</xref></td><td align="center" rowspan="1" colspan="1">−0.630 (0.137)<xref ref-type="table-fn" rid="TF0002">**</xref></td></tr><tr><td align="left" colspan="5" rowspan="1">Gender</td></tr><tr><td align="left" rowspan="1" colspan="1"> Female</td><td align="center" rowspan="1" colspan="1">3,379 (50.52)</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0</td></tr><tr><td align="left" rowspan="1" colspan="1"> Male</td><td align="center" rowspan="1" colspan="1">3,310 (49.48)</td><td align="center" rowspan="1" colspan="1">0.556 (0.283)<xref ref-type="table-fn" rid="TF0001">*</xref></td><td align="center" rowspan="1" colspan="1">0.698 (0.303)<xref ref-type="table-fn" rid="TF0001">*</xref></td><td align="center" rowspan="1" colspan="1">0.641 (0.291)<xref ref-type="table-fn" rid="TF0001">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1"><break></break></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" colspan="5" rowspan="1">Child's parents former refugee</td></tr><tr><td align="left" rowspan="1" colspan="1"> Yes</td><td align="center" rowspan="1" colspan="1">2,388 (35.71)</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"> No</td><td align="center" rowspan="1" colspan="1">4,299 (64.29)</td><td align="center" rowspan="1" colspan="1">−0.137 (0.280)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"><break></break></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1">Mother's age in years<xref ref-type="table-fn" rid="TF0003">1</xref></td><td align="center" rowspan="1" colspan="1">28.70±7.59</td><td align="center" rowspan="1" colspan="1">−0.003 (0.018)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1">Cumulative household deaths<xref ref-type="table-fn" rid="TF0004">2</xref></td><td align="center" rowspan="1" colspan="1">0.68±0.96 (0–6)</td><td align="center" rowspan="1" colspan="1">0.684 (0.086)<xref ref-type="table-fn" rid="TF0002">**</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" colspan="5" rowspan="1">Mother deceased by 2004</td></tr><tr><td align="left" rowspan="1" colspan="1"> No</td><td align="center" rowspan="1" colspan="1">6,634 (99.13)</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0</td></tr><tr><td align="left" rowspan="1" colspan="1"> Yes</td><td align="center" rowspan="1" colspan="1">58 (0.87)</td><td align="center" rowspan="1" colspan="1">1.508 (0.732)<xref ref-type="table-fn" rid="TF0001">*</xref></td><td align="center" rowspan="1" colspan="1">1.504 (0.867)<xref ref-type="table-fn" rid="TF0001">*</xref></td><td align="center" rowspan="1" colspan="1">1.386 (0.707)<xref ref-type="table-fn" rid="TF0001">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Minimum distance to health facility<xref ref-type="table-fn" rid="TF0003">1</xref></td><td align="center" rowspan="1" colspan="1">2.33±1.61</td><td align="center" rowspan="1" colspan="1">0.078 (0.083)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1">Household size<xref ref-type="table-fn" rid="TF0003">1</xref></td><td align="center" rowspan="1" colspan="1">4.88±2.92</td><td align="center" rowspan="1" colspan="1">−0.011 (0.048)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"><break></break></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" colspan="5" rowspan="1">Gender of household-head</td></tr><tr><td align="left" rowspan="1" colspan="1"> Female</td><td align="center" rowspan="1" colspan="1">2,434 (36.37)</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1">0</td></tr><tr><td align="left" rowspan="1" colspan="1"> Male</td><td align="center" rowspan="1" colspan="1">4,258 (63.63)</td><td align="center" rowspan="1" colspan="1">−0.865 (0.418)<xref ref-type="table-fn" rid="TF0002">**</xref></td><td align="center" rowspan="1" colspan="1">−0.880 (0.302)<xref ref-type="table-fn" rid="TF0002">**</xref></td><td align="center" rowspan="1" colspan="1">−0.895 (0.30)<xref ref-type="table-fn" rid="TF0002">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Parity<xref ref-type="table-fn" rid="TF0004">2</xref></td><td align="center" rowspan="1" colspan="1">2.12±1.43 (1–10)</td><td align="center" rowspan="1" colspan="1">0.085 (0.090)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"><break></break></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" colspan="5" rowspan="1">Socio-economic status quintiles</td></tr><tr><td align="left" rowspan="1" colspan="1"> Most poor</td><td align="center" rowspan="1" colspan="1">1,823 (27.24)</td><td align="center" rowspan="1" colspan="1">0</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"> Very Poor</td><td align="center" rowspan="1" colspan="1">1,073 (16.03)</td><td align="center" rowspan="1" colspan="1">−0.112 (0.393)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"> Moderately poor</td><td align="center" rowspan="1" colspan="1">1,120 (16.74)</td><td align="center" rowspan="1" colspan="1">−0.516 (0.444)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"> Poor</td><td align="center" rowspan="1" colspan="1">1,316 (19.67)</td><td align="center" rowspan="1" colspan="1">−0.425 (0.406)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1"> Least poor</td><td align="center" rowspan="1" colspan="1">1,360 (20.32)</td><td align="center" rowspan="1" colspan="1">−0.458 (0.406)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td></tr><tr><td align="left" rowspan="1" colspan="1">Non-spatial variance</td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.10; 95% BCI (0.001; 0.68)</td><td align="center" rowspan="1" colspan="1">0.04; 95% BCI (0.001–0.40)</td></tr><tr><td align="left" rowspan="1" colspan="1">Spatial variance estimate</td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.04; 95% CI (0.001–0.28)</td></tr><tr><td align="left" rowspan="1" colspan="1">Deviance information criteria</td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">457.26±15.56</td><td align="center" rowspan="1" colspan="1">555.87±11.65</td></tr></tbody></table><table-wrap-foot><fn id="TF0001"><label>*</label><p><italic>p</italic><0.05</p></fn><fn id="TF0002"><label>**</label><p><italic>p</italic><0.01.</p></fn><fn id="TF0003"><label>1</label><p>mean±standard deviation</p></fn><fn id="TF0004"><label>2</label><p>mean±standard deviation (range). 95% BCI=95% Bayesian credible interval.</p></fn></table-wrap-foot></table-wrap></sec><sec id="S20005"><title>Statistical methodology</title><p>We modelled the child HIV/TB mortality data to find associated factors and determine geographic disparities. Statistical techniques take into account spatial correlations by introducing spatially structured random effects into the model. These are usually based on independent observations. However, with multilevel correlated structures commonly used, maximum likelihood estimation (MLE) methods often underestimate the standard error; thus, the statistical significance is overestimated (<xref ref-type="bibr" rid="CIT0028">28</xref>). Analysis was carried out using a geospatial logit regression model, which catered for individual- and household-level random effects (<xref ref-type="bibr" rid="CIT0029">29</xref>). Inference was done using classical MLE and Bayesian inference techniques to fit the non-spatial and spatial models, respectively. Bayesian estimation of the many parameters was done using Markov Chain Monte Carlo (MCMC) simulations with the aid of Metropolis Hastings (MH) steps (<xref ref-type="bibr" rid="CIT0029">29</xref>). Geospatial prediction maps showing the geographical coordinates, adjusting for child HIV/TB mortality determinants, were used in the analysis.</p><p>Data extraction was carried out using Structured Query Language (SQL), data management was done in STATA 10.0 (<xref ref-type="bibr" rid="CIT0030">30</xref>), and structural equation modelling was performed in STATA 12.1 (Stata Corporation, College Station, Texas, USA). The Bayesian models were fit using software available in the public domain: BayesX version 2.1 for the spatial modelling and R-cran version 2.12.2 for geospatial mapping using the analysis and mapping packages, maptools, maps, spBayes, and Multilevel B-spline Approximation (MBA) (<xref ref-type="bibr" rid="CIT0031">31</xref>, <xref ref-type="bibr" rid="CIT0032">32</xref>).</p></sec><sec id="S20006"><title>Ethical clearance and informed consent</title><p>While this study used secondary data, the original study on which it was based was careful to consider ethical issues. Verbal informed consent was obtained when the Agincourt health and socio-demographic update rounds were conducted and also when verbal autopsy data were collected from a close relative of the deceased. To avoid emotionally stressing the interviewee, a culturally accepted period of 1 month was observed before the interview was conducted. Risks of stigmatisation were minimised through special training of fieldworkers on the importance of confidentiality. This work was granted ethical clearance by the University of the Witwatersrand's Committee for Research on Human Subjects (No. 960720 and M081145).</p></sec></sec><sec sec-type="results" id="S0007"><title>Results</title><p>Of the 6,692 children between 1 and less than 5 years, with a mean age of 2.8 years, 54 HIV/TB-related deaths were recorded in the HDSS in 2004. There were 5,084 fully geo-coded households that recorded children either living or dead. Just over half the children were female (51%), and more than a third (36%) were born to former Mozambican refugees. The mean age of mothers was 29 years, and the mean distance to the nearest health facility was 2.33 km. <xref ref-type="table" rid="T0001">Table 1</xref> shows the results for the descriptive, univariate, and multivariable logit regression analyses. Maternal death had the greatest effect on child HIV/TB mortality [logit=1.386 (0.707)], equivalent to an adjusted odds ratio (AOR) of 4.00 [95% Bayesian credible interval (BCI)=1.01–15.80] controlling for spatial correlation and keeping all other variables constant.</p><p>Multivariable analysis showed three further significant predictors of child mortality in 2004, which were child, maternal, and household specific. Adjusting for spatial correlation and keeping all other variables constant, there were several significant determinants of child HIV/TB mortality. Child-specific differentials showed that boys were almost twice as likely to die compared to girls [logit=0.641 (0.291); AOR=1.89; 95% BCI=1.07–3.36]. As the children got older, there was a protective effect against death; for every year the child aged, the likelihood of death decreased by about 50% [logit=−0.630 (0.137); AOR=0.53; 95% BCI 0.40–0.70]. Male-headed households had a 60% lower risk of child HIV/TB mortality compared to female-headed households [logit=−0.895 (0.30); AOR=0.41; 95% BCI 0.23–0.74].</p><p>Spatial disparities can be viewed on maps, which can show unexpected relationships that may have been overlooked in standard regression analysis. We obtained household-specific posterior odds ratios, which were adjusted for child-, maternal-, and household-related variables in the multivariable models. One main hotspot emerged with high mortality in the north, and there were two southerly areas with low mortality. The odds ratio posterior map in <xref ref-type="fig" rid="F0003">Fig. 3</xref> shows the hotspot area between longitude 20° and 25° and latitude −2,750<bold>°</bold> and −2,745<bold>°</bold>. This area is characterised by the following: higher rate of maternal deaths, more male child mortality, lack of a health facility, and being furthest from any health facility as shown in <xref ref-type="fig" rid="F0001">Fig. 1</xref>.</p><p>The results from the conceptual framework analysis of child HIV/TB mortality are shown in <xref ref-type="fig" rid="F0004">Fig. 4</xref> and <xref ref-type="table" rid="T0002">Table 2</xref>. All of the connected arrows show pathways that were statistically significant whether directly or indirectly. Being male (logit=5.2×10<sup>−3</sup>) and older child age (logit=−4.6×10<sup>−3</sup>) both impacted directly on HIV/TB mortality controlling for potential confounders related to the maternal and households variables. Upon treating male household headship as an exogenous variable, this directly impacted negatively on child HIV/TB deaths (logit=−7.6×10<sup>−3</sup>), adjusting for household variables. Moreover, when male headship was an endogenous variable, the following household variables indirectly impacted mortality: better socio-economic status and larger household size had a protective effect (logit=7.7×10<sup>−3</sup>) and increase in household deaths resulted in greater child HIV/TB mortality (logit=−8.3×10<sup>−2</sup>).
</p><fig id="F0004" position="float"><label>Fig. 4</label><caption><p>Conceptual framework pathways results for Agincourt sub-district child mortality determinants in 2004. Key for variable names: very poor versus most poor (<bold>vpoor_mostpoor</bold>), moderately poor versus most poor (<bold>modpoor_mostpoor</bold>), least poor versus most poor (<bold>leastpoor_mostpoor</bold>), poor versus most poor (<bold>poor_mostpoor)</bold>, age of child in years (<bold>age_ch</bold>), household size (<bold>hh_size</bold>), cumulative household deaths (<bold>cum_hh_deaths</bold>), gender of household head (<bold>hh_gender</bold>), mother age in years (<bold>age_mum</bold>), mother deceased by 2004 (<bold>mother_deceased</bold>), child born to former refugee parents (<bold>refuge</bold>), gender (<bold>gender</bold>), minimum distance to health facility (<bold>min_dist</bold>) parity (<bold>parity</bold>), child deceased in 2004 (<bold>dead</bold>). Other: the arrows pointing from the exogenous (explanatory) to endogenous (dependent) variables and the error terms (<italic><bold>ɛ</bold></italic>) placed on all five endogenous variables.</p></caption><graphic xlink:href="GHA-6-19266-g004"></graphic></fig><table-wrap id="T0002" position="float"><label>Table 2</label><caption><p>Direct, indirect and total effects of child HIV/TB mortality in 2004, Agincourt sub-district</p></caption><table frame="hsides" rules="groups"><thead><tr><th align="left" rowspan="1" colspan="1"></th><th align="center" colspan="5" rowspan="1">Direct effects on the deceased child as shown in <xref ref-type="fig" rid="F0002">Fig. 2</xref> conceptual framework</th><th align="center" rowspan="1" colspan="1"></th><th align="center" rowspan="1" colspan="1"></th></tr><tr><th align="left" rowspan="1" colspan="1"></th><th colspan="5" rowspan="1"><hr></hr></th><th align="center" rowspan="1" colspan="1"></th><th align="center" rowspan="1" colspan="1"></th></tr><tr><th align="left" rowspan="1" colspan="1"></th><th align="center" rowspan="1" colspan="1">Gender of household-head</th><th align="center" rowspan="1" colspan="1">Child a former refugee</th><th align="center" rowspan="1" colspan="1">Child deceased</th><th align="center" rowspan="1" colspan="1">Mother deceased by 2004</th><th align="center" rowspan="1" colspan="1">Age of child in years</th><th align="center" rowspan="1" colspan="1">Indirect effects on child HIV/TB death</th><th align="center" rowspan="1" colspan="1">Total effects on child HIV/TB death</th></tr></thead><tbody><tr><td align="left" rowspan="1" colspan="1">Very poor vs most poor</td><td align="center" rowspan="1" colspan="1">−0.029 (0.019)</td><td align="center" rowspan="1" colspan="1">0.140 (0.015)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">2.6×10<sup>−4</sup> (1.6×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0008">ns</xref><xref ref-type="table-fn" rid="TF0009">!</xref></td><td align="center" rowspan="1" colspan="1">2.6×10<sup>−4</sup> (1.6×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0008">ns</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Moderately poor vs most poor</td><td align="center" rowspan="1" colspan="1">−0.059 (0.020)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">4.5×10<sup>−4</sup> (2.1×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0006">**</xref></td><td align="center" rowspan="1" colspan="1">4.5×10<sup>−4</sup> (2.1×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0006">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Poor vs most poor</td><td align="center" rowspan="1" colspan="1">−0.052 (0.020)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">4.0×10<sup>−4</sup> (2.0×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0006">**</xref></td><td align="center" rowspan="1" colspan="1">4.0×10<sup>−4</sup> (2.0×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0006">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Least poor vs most poor</td><td align="center" rowspan="1" colspan="1">−0.111 (0.018)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">8.4×10<sup>−4</sup> (3.0 ×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1">8.4×10<sup>−4</sup> (3.0×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0007">***</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Household size vs most poor</td><td align="center" rowspan="1" colspan="1">0.008 (0.002)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−5.9×10<sup>−5</sup> (2.5×10<sup>−5</sup>)<xref ref-type="table-fn" rid="TF0006">**</xref></td><td align="center" rowspan="1" colspan="1">−5.9×10<sup>−5</sup> (2.5×10<sup>−5</sup>)<xref ref-type="table-fn" rid="TF0006">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Cumulative household deaths</td><td align="center" rowspan="1" colspan="1">−0.083 (0.006)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">6.4×10<sup>−4</sup> (2.1×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1">6.4×10<sup>−4</sup> (2.1×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0007">***</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Minimum distance to health facility</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.031 (0.004)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">7.72×10<sup>−6</sup> (4.88×10<sup>−6</sup>)<xref ref-type="table-fn" rid="TF0008">ns</xref></td><td align="center" rowspan="1" colspan="1">7.72×10<sup>−6</sup> (4.88×10<sup>−6</sup>)<xref ref-type="table-fn" rid="TF0008">ns</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Male household-head</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−0.0076 (0.002)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−0.0076 (0.002)<xref ref-type="table-fn" rid="TF0007">***</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Mother deceased by 2004</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.027 (0.012)<xref ref-type="table-fn" rid="TF0005">*</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.027 (0.012)<xref ref-type="table-fn" rid="TF0006">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Age of child in years</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−0.0046 (0.001)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−0.0046 (0.001)<xref ref-type="table-fn" rid="TF0007">***</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Gender</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.0052 (0.0023)<xref ref-type="table-fn" rid="TF0006">**</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="left" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.0052 (0.0023)<xref ref-type="table-fn" rid="TF0006">**</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Mother's age (years)</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">0.001 (0.000)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">1.9×10<sup>−5</sup> (9.99×10<sup>−6</sup>)<xref ref-type="table-fn" rid="TF0005">*</xref></td><td align="center" rowspan="1" colspan="1">1.9×10–05 (9.99×10<sup>−6</sup>)<xref ref-type="table-fn" rid="TF0005">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Parity</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−0.003 (0.001)<xref ref-type="table-fn" rid="TF0006">**</xref></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−7.0×10<sup>−5</sup> (4.0×10<sup>−5</sup>)<xref ref-type="table-fn" rid="TF0008">ns</xref></td><td align="center" rowspan="1" colspan="1">−7.0×10–05 (4.0×10<sup>−5</sup>)<xref ref-type="table-fn" rid="TF0008">ns</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Parents former refugee</td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1"></td><td align="center" rowspan="1" colspan="1">−0.054 (0.032)<xref ref-type="table-fn" rid="TF0005">*</xref></td><td align="center" rowspan="1" colspan="1">−2.5×10<sup>−4</sup> (1.5×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0005">*</xref></td><td align="center" rowspan="1" colspan="1">−2.5×10<sup>−4</sup> (1.5×10<sup>−4</sup>)<xref ref-type="table-fn" rid="TF0005">*</xref></td></tr><tr><td align="left" rowspan="1" colspan="1">Intercept</td><td align="center" rowspan="1" colspan="1">1.719 (0.017)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1">1.537 (0.011)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1">−0.002 (0.014)</td><td align="center" rowspan="1" colspan="1">0.995 (0.005)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="center" rowspan="1" colspan="1">2.814 (0.054)<xref ref-type="table-fn" rid="TF0007">***</xref></td><td align="left" rowspan="1" colspan="1"></td><td align="left" rowspan="1" colspan="1"></td></tr></tbody></table><table-wrap-foot><fn id="TF0005"><label>*</label><p><italic>p<</italic>0.10</p></fn><fn id="TF0006"><label>**</label><p><italic>p<</italic>0.05</p></fn><fn id="TF0007"><label>***</label><p><italic>p<</italic>0.01</p></fn><fn id="TF0008"><label>ns</label><p>ns: not significant</p></fn><fn id="TF0009"><label>!</label><p>indirect effects computed as the product along the related pathways, i.e −0.029×−0.0076+0.14×−0.054×−0.0046=2.55×10<sup>−4</sup>≈2.6×10<sup>−<bold>4</bold></sup>.</p></fn></table-wrap-foot></table-wrap><p><xref ref-type="table" rid="T0002">Table 2</xref> details the direct pathways for child HIV/TB deaths (arrows directly linked to the red square) and the indirect pathways showing all possible routes for child HIV/TB deaths as shown in <xref ref-type="fig" rid="F0004">Fig. 4</xref>. In our analysis, poverty indirectly impacted on HIV/TB child mortality as conferred by two pathways: first, male headship of households (protective), and second, refugee status (a risk) controlling for child's age (as shown on <xref ref-type="table" rid="T0002">Table 2</xref> foot note). The greatest direct impact on child HIV/TB mortality was maternal death (logit=2.7×10<sup><italic>−</italic>2</sup>), adjusting for the indirect effects of parity (logit=−2.6×10<sup><italic>−</italic>3</sup>) and mother's age (logit=6.8×10<sup><italic>−</italic>4</sup>).</p></sec><sec sec-type="discussion" id="S0008"><title>Discussion</title><p>In this study, we showed that HIV/TB child mortality is influenced by a combination of child, maternal, household, and spatial determinants. Over the years 2002–2005, HIV/TB alone accounted for 34% of all-cause mortality among children living in the area (<xref ref-type="bibr" rid="CIT0023">23</xref>). While being male was a risk factor for HIV/TB mortality as children got older, irrespective of sex they were less likely to die. A major finding in our study was that maternal death had the greatest direct impact on HIV/TB child deaths. A comparable study in the same area, adjusting for spatial random effects at village level, showed that infants whose mothers had died were at greater risk of death during the first 12 months of life (<xref ref-type="bibr" rid="CIT0016">16</xref>).</p><p>The main source of TB transmission among children are adults, as shown by a study on a high-risk spatial cluster in South Africa with an incidence three times higher than the WHO levels (<xref ref-type="bibr" rid="CIT0017">17</xref>). The most likely cause of HIV infection was mother-to-child transmission. Prevention of mother-to-child transmission (PMTCT), using a single-dose of nevirapine (sdNVP) administered to the child and mother, was found to reduce child mortality by about 9% in South Africa (<xref ref-type="bibr" rid="CIT0033">33</xref>). Between 2004 and 2012 the guidelines for PMTCT and ART eligibility have changed. There is also a seven-fold lower (from 25 to 3.5%) rate of infection from mother to child over the period 2000–2010 (<xref ref-type="bibr" rid="CIT0005">5</xref>). Kimani-Murage et al. (2009) found that knowledge of child HIV status influenced the manner of care-giving as well as acceptance and up-take of community-wide interventions (<xref ref-type="bibr" rid="CIT0031">31</xref>). The mother's survival significantly increases the likelihood of the child's survival, a finding that strengthens the current policy focus on maternal health. In South Africa, a comprehensive three stage process has been set up to improve maternal, foetal, infant, and ultimately child survival through antenatal care, labour and delivery, and postnatal care (<xref ref-type="bibr" rid="CIT0034">34</xref>). Reduction of TB morbidity and mortality is the focus of the WHO 2015 goal of zero new TB infections and zero TB-related deaths (<xref ref-type="bibr" rid="CIT0002">2</xref>). The positive impact of the PMTCT programme and ART access has also been demonstrated in another rural South Africa site (<xref ref-type="bibr" rid="CIT0035">35</xref>).</p><p>In Agincourt, female-headed households had a higher child mortality compared to those headed by males. Studies in Tanzania, Zambia, and Limpopo province South Africa showed similar results (<xref ref-type="bibr" rid="CIT0036">36</xref>–<xref ref-type="bibr" rid="CIT0038">38</xref>). The pathway analysis (<xref ref-type="fig" rid="F0004">Fig. 4</xref>) demonstrates that male-headship does not independently lead to lower mortality but is a proxy for better socio-economic status of households. Main breadwinners are usually male with females having fewer opportunities in the labour market and consequently earning less. The absence from the household of a male due to death or permanent out-migration tended to impact negatively on household socio-economic status, leading to 90% greater child mortality odds (<xref ref-type="bibr" rid="CIT0039">39</xref>). In other African countries, similar patterns were observed as men temporarily migrate within a country or to neighbouring countries due to either pull or push factors (<xref ref-type="bibr" rid="CIT0040">40</xref>–<xref ref-type="bibr" rid="CIT0042">42</xref>). HIV/TB deaths have lowered the life expectancy of working men from the area who often “return home to die” from their jobs (<xref ref-type="bibr" rid="CIT0021">21</xref>). A recent qualitative study found that a father's unemployment, inability to work, and reluctance to provide support were common factors in HIV-infected households (<xref ref-type="bibr" rid="CIT0031">31</xref>). From our pathway analysis (<xref ref-type="table" rid="T0002">Table 2</xref> and <xref ref-type="fig" rid="F0004">Fig. 4</xref>), we observed an indirect impact on child HIV/TB mortality: as the number of household deaths increased, the risk of child deaths was higher. This suggests that households with a greater burden of mortality are more vulnerable socio-economically through loss of breadwinners and high health care and funeral costs. This compromises a family's ability to feed and care for all children, including those who are HIV positive. We also observed that former Mozambicans are at greater risk of child HIV/TB deaths, indicating the need for health care interventions that target these marginalised and vulnerable households.</p></sec><sec id="S0009"><title>Strengths and limitations of the study</title><p>The strengths and limitations of this paper are discussed in four categories: classification of exposures and outcomes; local or global relevance of results; confounding and effect modification; and analysis deviance.</p><p>A major strength in our observed variables is that the data are consistently updated, corrected, and checked for inconsistencies. Nevertheless, there are potentially two variable-related biases, namely selection and recall bias. Selection bias may occur if permanent out-migrants differ from residents or temporary migrants, which results in incomparable groups. Recall bias on the outcome variable mortality is unlikely as death is a major household event. However, it may affect the quality of the verbal autopsy interview, conducted at least 1 month but up to 11 months after death, and hence the ability of medical experts to assign a likely cause of death.</p><p>Our study focused on one sub-district only, but demonstrated geographical differentials in child HIV/TB mortality. There was one main hotspot with highest mortality and a visible downward trend in mortality moving southwards. These patterns are comparable to those found in other studies carried out in the area (<xref ref-type="bibr" rid="CIT0016">16</xref>, <xref ref-type="bibr" rid="CIT0019">19</xref>). While many rural areas are not geo-coded to enable direct comparison of trends and patterns, our findings reflect similar patterns in other rural settlements at least in northeast South Africa.</p><p>To deal with confounding variables, we used multiple regression and added the spatial random effects in our modelling and also performed pathway analysis. Mediation was tested for in the pathway analysis as we determined factors directly and indirectly related to mortality and malnutrition. However, we were unable to model time varying covariates as this was a cross-sectional study design.</p><p>The robustness of our analysis procedures was demonstrated in that after catering for the spatial household random effects, we obtained narrower confidence bands compared to the non-spatial analogue. Therefore, catering for the spatial random effects further improves the model and parameter estimation compared to classical statistical models. Bayesian and pathway analysis combined were able to handle the complex structure of hierarchical data from the Agincourt health and socio-demographic surveillance system. Other advantages of Bayesian modelling include the ability to adjust for spatial confounding, and the ability to handle many parameters and inclusion of prior-known information on the distribution of data.</p></sec><sec id="S0010"><title>Conclusion</title><p>The growing field of spatio-temporal epidemiology brings in person, place, and time aspects in the risk factor analysis. Spatial Bayesian models provide better-fitting models compared to non-spatial models. This procedure also enables the drawing of risk maps, which can be used by policy makers to target and develop community-based interventions.</p><p>Going forward, further studies investigating trends over time using spatio-temporal Bayesian analysis will strengthen the evidence on associated factors and trends found in this study and demonstrate the value of this analytic approach. Studies comparing deaths due to other causes with those for HIV/TB in children would add policy value. Similarly, adopting a secondary analysis of a cohort study design would have enabled temporality to be investigated in greater detail and survival analysis to be undertaken. Moreover, a combination of qualitative and quantitative analysis methods is needed to understand the proximal and distal factors that influence child HIV/TB mortality. Findings that can be used for national policy would require sentinel sites located strategically across the country.</p></sec></body><back><ack><title>Acknowledgements</title><p>This work was made possible by doctoral fellowship support from the World Health Organization, Tropical Disease Research (WHO/TDR), and Swiss–South Africa Joint Research Programme. The MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt) is funded by the Wellcome Trust, UK, (grants 058893/Z/99/A, 069683/Z/02/Z and 085477/Z/08/Z), the University of the Witwatersrand and Medical Research Council, South Africa, and the Andrew W Mellon and William and Flora Hewlett Foundations, USA. Special thanks goes to Benn Sartorius for data extraction, David Rees for his valuable review of an early draft, community leaders, field workers, supervisors, and the Agincourt community.</p></ack><fn-group><fn id="FN0001"><label>1</label><p>A16=Respiratory tuberculosis; not confirmed bacteriologically or histologically; A17=Tuberculosis of nervous system; A18=Tuberculosis of other organs; A19=Miliary tuberculosis.</p></fn></fn-group><fn-group><fn id="FN0002"><label>2</label><p>B20=Human immunodeficiency virus (HIV) disease resulting in infectious and parasitic diseases; B21=Human immunodeficiency virus (HIV) disease resulting in malignant neoplasms; B22=Human immunodeficiency virus (HIV) disease resulting in other specified diseases; B23=Human immunodeficiency virus (HIV) disease resulting in other conditions; B24=Unspecified human immunodeficiency virus (HIV) disease.</p></fn></fn-group><sec id="S0011"><title>Conflict of interest and funding</title><p>The authors declare that they have no competing interests.</p></sec><ref-list><title>References</title><ref id="CIT0001"><label>1</label><element-citation publication-type="webpage"><collab>UNAIDS</collab><article-title>Report on the global AIDS epidemic 2011</article-title><comment>Available from: <ext-link ext-link-type="uri" xlink:href="http://www.avert.org/africa-hiv-aids-statistics.htm">http://www.avert.org/africa-hiv-aids-statistics.htm</ext-link> [cited 23 October 2012]</comment></element-citation></ref><ref id="CIT0002"><label>2</label><element-citation publication-type="webpage"><collab>WHO</collab><article-title>One million lives can be saved between now and the end of 2015 by preventing and treating tuberculosis among people living with HIV</article-title><comment>Available from: <ext-link ext-link-type="uri" xlink:href="http://www.who.int/tb/challenges/hiv/fastfactsbrochure.pdf">http://www.who.int/tb/challenges/hiv/fastfactsbrochure.pdf</ext-link> [cited 23 October 2012].</comment></element-citation></ref><ref id="CIT0003"><label>3</label><element-citation publication-type="webpage"><collab>Statistics South Africa (StatsSA)</collab><article-title>Statistical release P0302 2011</article-title><comment>Available from: <ext-link ext-link-type="uri" xlink:href="http://www.statssa.gov.za/publications/P0302/P03022011.pdf">http://www.statssa.gov.za/publications/P0302/P03022011.pdf</ext-link> [cited 23 October 2012]</comment></element-citation></ref><ref id="CIT0004"><label>4</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Bradshaw</surname><given-names>D</given-names></name><name><surname>Bourne</surname><given-names>D</given-names></name><name><surname>Nannan</surname><given-names>N</given-names></name></person-group><article-title>What are the leading causes of death among South African children</article-title><source>MRC Policy Brief & Unicef</source><year>2003</year><volume>3</volume><fpage>1</fpage><lpage>4</lpage></element-citation></ref><ref id="CIT0005"><label>5</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Doherty</surname><given-names>T</given-names></name><name><surname>Sanders</surname><given-names>D</given-names></name><name><surname>Goga</surname><given-names>A</given-names></name><name><surname>Jackson</surname><given-names>D</given-names></name></person-group><article-title>Implications of the new WHO guidelines on HIV and infant feeding for child survival in South Africa</article-title><source>Bull World Health Organ</source><year>2011</year><volume>89</volume><fpage>62</fpage><lpage>7</lpage><pub-id pub-id-type="pmid">21346892</pub-id></element-citation></ref><ref id="CIT0006"><label>6</label><element-citation publication-type="webpage"><collab>UNICEF</collab><article-title>State of the World's Children year 2009</article-title><comment>Available from: <ext-link ext-link-type="uri" xlink:href="http://www.unicef.org/publications/index_47127.html">http://www.unicef.org/publications/index_47127.html</ext-link> [cited 23 October 2012]</comment></element-citation></ref><ref id="CIT0007"><label>7</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Waller</surname><given-names>LA</given-names></name><name><surname>Gotway</surname><given-names>CA</given-names></name></person-group><source>Applied spatial statistics for public health data</source><year>2004</year><publisher-loc>Atlanta, GA</publisher-loc><publisher-name>Wiley</publisher-name></element-citation></ref><ref id="CIT0008"><label>8</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Mosley</surname><given-names>WH</given-names></name><name><surname>Chen</surname><given-names>LC</given-names></name></person-group><article-title>An analytical framework for the study of child survival in developing countries</article-title><source>Popul Dev Rev</source><year>1984</year><volume>10</volume><fpage>25</fpage><lpage>45</lpage></element-citation></ref><ref id="CIT0009"><label>9</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Rutstein</surname><given-names>SO</given-names></name></person-group><article-title>Factors associated with trends in infant and child mortality in developing countries in the 1990s</article-title><source>Bull World Health Organ</source><year>2000</year><volume>78</volume><fpage>1256</fpage><lpage>70</lpage><pub-id pub-id-type="pmid">11100620</pub-id></element-citation></ref><ref id="CIT0010"><label>10</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Filteau</surname><given-names>S</given-names></name></person-group><article-title>The HIV-exposed, uninfected African child</article-title><source>Trop Med Int Health</source><year>2009</year><volume>14</volume><fpage>276</fpage><lpage>87</lpage><pub-id pub-id-type="pmid">19171011</pub-id></element-citation></ref><ref id="CIT0011"><label>11</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Guerrero</surname><given-names>L</given-names></name><name><surname>La Valley</surname><given-names>A</given-names></name></person-group><person-group person-group-type="editor"><name><surname>Oetzel</surname><given-names>JG</given-names></name><name><surname>Ting-Toomey</surname><given-names>S</given-names></name></person-group><article-title>Conflict, emotion, and communication</article-title><source>The SAGE handbook of conflict communication: integrating theory, research, and practice</source><year>2006</year><publisher-loc>Thousand Oaks, CA</publisher-loc><publisher-name>SAGE</publisher-name><fpage>69</fpage><lpage>96</lpage></element-citation></ref><ref id="CIT0012"><label>12</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hargreaves</surname><given-names>J</given-names></name><name><surname>Collinson</surname><given-names>M</given-names></name><name><surname>Kahn</surname><given-names>K</given-names></name><name><surname>Clark</surname><given-names>S</given-names></name><name><surname>Tollman</surname><given-names>S</given-names></name></person-group><article-title>Childhood mortality among former Mozambican refugees and their hosts in rural South Africa</article-title><source>Int J Epidemiol</source><year>2004</year><volume>33</volume><fpage>1271</fpage><lpage>8</lpage><pub-id pub-id-type="pmid">15319405</pub-id></element-citation></ref><ref id="CIT0013"><label>13</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Marshall</surname><given-names>RJ</given-names></name></person-group><article-title>A review of methods for the statistical analysis of spatial patterns of disease</article-title><source>J Roy Stat Soc Series A (Statistics in Society)</source><year>1991</year><volume>154</volume><fpage>421</fpage><lpage>41</lpage></element-citation></ref><ref id="CIT0014"><label>14</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Gosoniu</surname><given-names>L</given-names></name><name><surname>Vounatsou</surname><given-names>P</given-names></name><name><surname>Tami</surname><given-names>A</given-names></name><name><surname>Nathan</surname><given-names>R</given-names></name><name><surname>Grundmann</surname><given-names>H</given-names></name><name><surname>Lengeler</surname><given-names>C</given-names></name></person-group><article-title>Spatial effects of mosquito bednets on child mortality</article-title><source>BMC Publ Health</source><year>2008</year><volume>8</volume><fpage>356</fpage></element-citation></ref><ref id="CIT0015"><label>15</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Moonan</surname><given-names>PK</given-names></name><name><surname>Bayona</surname><given-names>M</given-names></name><name><surname>Quitugua</surname><given-names>TN</given-names></name><name><surname>Oppong</surname><given-names>J</given-names></name><name><surname>Dunbar</surname><given-names>D</given-names></name><name><surname>Jost</surname><given-names>KC</given-names></name><etal></etal></person-group><article-title>Using GIS technology to identify areas of tuberculosis transmission and incidence</article-title><source>Int J Health Geogr</source><year>2004</year><volume>3</volume><fpage>23</fpage><pub-id pub-id-type="pmid">15479478</pub-id></element-citation></ref><ref id="CIT0016"><label>16</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Sartorius</surname><given-names>B</given-names></name><name><surname>Kahn</surname><given-names>K</given-names></name><name><surname>Vounatsou</surname><given-names>P</given-names></name><name><surname>Collinson</surname><given-names>M</given-names></name><name><surname>Tollman</surname><given-names>S</given-names></name></person-group><article-title>Young and vulnerable: spatial-temporal trends and risk factors for infant mortality in rural South Africa (Agincourt), 1992–2007</article-title><source>BMC Publ Health</source><year>2010</year><volume>10</volume><fpage>645</fpage></element-citation></ref><ref id="CIT0017"><label>17</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Middelkoop</surname><given-names>K</given-names></name><name><surname>Bekker</surname><given-names>LG</given-names></name><name><surname>Morrow</surname><given-names>C</given-names></name><name><surname>Zwane</surname><given-names>E</given-names></name><name><surname>Wood</surname><given-names>R</given-names></name></person-group><article-title>Childhood tuberculosis infection and disease: a spatial and temporal transmission analysis in a South African township</article-title><source>S Afr Med J</source><year>2009</year><volume>99</volume><fpage>738</fpage><lpage>43</lpage><pub-id pub-id-type="pmid">20128273</pub-id></element-citation></ref><ref id="CIT0018"><label>18</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Tanser</surname><given-names>F</given-names></name><name><surname>LeSueur</surname><given-names>D</given-names></name><name><surname>Solarsh</surname><given-names>G</given-names></name><name><surname>Wilkinson</surname><given-names>D</given-names></name></person-group><article-title>HIV heterogeneity and proximity of homestead to roads in rural South Africa: an exploration using a geographical information system</article-title><source>Trop Med Int Health</source><year>2000</year><volume>5</volume><fpage>40</fpage><lpage>6</lpage><pub-id pub-id-type="pmid">10672204</pub-id></element-citation></ref><ref id="CIT0019"><label>19</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Musenge</surname><given-names>E</given-names></name><name><surname>Vounatsou</surname><given-names>P</given-names></name><name><surname>Kahn</surname><given-names>K</given-names></name></person-group><article-title>Space-time confounding adjusted determinants of child HIV/TB mortality for large zero-inflated data in rural South Africa</article-title><source>Spat Spatio-temporal Epidemiol</source><year>2011</year><volume>2</volume><fpage>205</fpage><lpage>17</lpage></element-citation></ref><ref id="CIT0020"><label>20</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kahn</surname><given-names>K</given-names></name><name><surname>Tollman</surname><given-names>SM</given-names></name><name><surname>Collinson</surname><given-names>MA</given-names></name><name><surname>Clark</surname><given-names>SJ</given-names></name><name><surname>Twine</surname><given-names>R</given-names></name><name><surname>Clark</surname><given-names>BD</given-names></name><etal></etal></person-group><article-title>Research into health, population and social transitions in rural South Africa: data and methods of the Agincourt Health and Demographic Surveillance System</article-title><source>Scand J Publ Health</source><year>2007</year><volume>35</volume><fpage>8</fpage><lpage>20</lpage></element-citation></ref><ref id="CIT0021"><label>21</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Clark</surname><given-names>SJ</given-names></name><name><surname>Collinson</surname><given-names>MA</given-names></name><name><surname>Kahn</surname><given-names>K</given-names></name><name><surname>Drullinger</surname><given-names>K</given-names></name><name><surname>Tollman</surname><given-names>SM</given-names></name></person-group><article-title>Returning home to die: circular labour migration and mortality in South Africa</article-title><source>Scand J Public Health Suppl</source><year>2007</year><volume>69</volume><fpage>35</fpage><lpage>44</lpage><pub-id pub-id-type="pmid">17676501</pub-id></element-citation></ref><ref id="CIT0022"><label>22</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kahn</surname><given-names>K</given-names></name><name><surname>Tollman</surname><given-names>SM</given-names></name><name><surname>Garenne</surname><given-names>M</given-names></name><name><surname>Gear</surname><given-names>JS</given-names></name></person-group><article-title>Validation and application of verbal autopsies in a rural area of South Africa</article-title><source>Trop Med Int Health</source><year>2000</year><volume>5</volume><fpage>824</fpage><lpage>31</lpage><pub-id pub-id-type="pmid">11123832</pub-id></element-citation></ref><ref id="CIT0023"><label>23</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Tollman</surname><given-names>SM</given-names></name><name><surname>Kahn</surname><given-names>K</given-names></name><name><surname>Sartorius</surname><given-names>B</given-names></name><name><surname>Collinson</surname><given-names>MA</given-names></name><name><surname>Clark</surname><given-names>SJ</given-names></name><name><surname>Garenne</surname><given-names>ML</given-names></name></person-group><article-title>Implications of mortality transition for primary health care in rural South Africa: a population-based surveillance study</article-title><source>Lancet</source><year>2008</year><volume>372</volume><fpage>893</fpage><lpage>901</lpage><pub-id pub-id-type="pmid">18790312</pub-id></element-citation></ref><ref id="CIT0024"><label>24</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Meade</surname><given-names>MS</given-names></name></person-group><article-title>Medical geography as human ecology: the dimension of population movement</article-title><source>Geogr Rev</source><year>1977</year><volume>67</volume><fpage>379</fpage><lpage>93</lpage></element-citation></ref><ref id="CIT0025"><label>25</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Lewis</surname><given-names>JJC</given-names></name><name><surname>Donnelly</surname><given-names>CA</given-names></name><name><surname>Mare</surname><given-names>P</given-names></name><name><surname>Mupambireyi</surname><given-names>Z</given-names></name><name><surname>Garnett</surname><given-names>GP</given-names></name><name><surname>Gregson</surname><given-names>S</given-names></name></person-group><article-title>Evaluating the proximate determinants framework for HIV infection in rural Zimbabwe</article-title><source>Sex Transm Infect</source><year>2007</year><volume>83</volume><fpage>i61</fpage><lpage>9</lpage><pub-id pub-id-type="pmid">17215273</pub-id></element-citation></ref><ref id="CIT0026"><label>26</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Boerma</surname><given-names>J</given-names></name><name><surname>Weir</surname><given-names>SS</given-names></name></person-group><article-title>Integrating demographic and epidemiological approaches to research on HIV/AIDS: the proximate-determinants framework</article-title><source>J Infec Dis</source><year>2005</year><volume>191</volume><fpage>S61</fpage><pub-id pub-id-type="pmid">15627232</pub-id></element-citation></ref><ref id="CIT0027"><label>27</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Lopman</surname><given-names>B</given-names></name><name><surname>Nyamukapa</surname><given-names>C</given-names></name><name><surname>Mushati</surname><given-names>P</given-names></name><name><surname>Mupambireyi</surname><given-names>Z</given-names></name><name><surname>Mason</surname><given-names>P</given-names></name><name><surname>Garnett</surname><given-names>GP</given-names></name><etal></etal></person-group><article-title>HIV incidence in 3 years of follow-up of a Zimbabwe cohort—1998–2000 to 2001–03: contributions of proximate and underlying determinants to transmission</article-title><source>Int J Epidemiol</source><year>2008</year><volume>37</volume><fpage>88</fpage><lpage>105</lpage><pub-id pub-id-type="pmid">18203774</pub-id></element-citation></ref><ref id="CIT0028"><label>28</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Cressie</surname><given-names>NAC</given-names></name></person-group><source>Statistics for spatial data</source><year>1993</year><publisher-loc>New York</publisher-loc><publisher-name>Wiley</publisher-name></element-citation></ref><ref id="CIT0029"><label>29</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Banerjee</surname><given-names>S</given-names></name><name><surname>Carlin</surname><given-names>BP</given-names></name><name><surname>Gelfand</surname><given-names>AE</given-names></name></person-group><source>Hierarchical modeling and analysis for spatial data</source><year>2004</year><publisher-loc>Boca Raton, FL</publisher-loc><publisher-name>Chapman & Hall/CRC</publisher-name></element-citation></ref><ref id="CIT0030"><label>30</label><element-citation publication-type="book"><collab>Stata Corporation</collab><article-title>Stata statistical software release 10</article-title><year>2007</year><publisher-loc>College Station, TX</publisher-loc><publisher-name>Stata Corporation LP</publisher-name></element-citation></ref><ref id="CIT0031"><label>31</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Kimani-Murage</surname><given-names>EW</given-names></name><name><surname>Manderson</surname><given-names>L</given-names></name><name><surname>Norris</surname><given-names>SA</given-names></name><name><surname>Kahn</surname><given-names>K</given-names></name></person-group><article-title>‘You opened our eyes’: care-giving after learning a child's positive HIV status in rural South Africa</article-title><source>Health Soc Care Community</source><year>2010</year><volume>18</volume><fpage>264</fpage><lpage>71</lpage><pub-id pub-id-type="pmid">20039969</pub-id></element-citation></ref><ref id="CIT0032"><label>32</label><element-citation publication-type="webpage"><person-group person-group-type="author"><name><surname>Belitz</surname><given-names>C</given-names></name><name><surname>Brezger</surname><given-names>A</given-names></name><name><surname>Kneib</surname><given-names>T</given-names></name><name><surname>Lang</surname><given-names>S</given-names></name><name><surname>Fronk</surname><given-names>E</given-names></name><name><surname>Heinzl</surname><given-names>F</given-names></name><etal></etal></person-group><article-title>Bayesx-software for bayesian inference in structured additive regression models</article-title><year>2011</year><comment>Version 2.01. Available from: <ext-link ext-link-type="uri" xlink:href="http://www.stat.uni-muenchen.de/~bayesx">http://www.stat.uni-muenchen.de/∼bayesx</ext-link> [cited 23 October 2012].</comment></element-citation></ref><ref id="CIT0033"><label>33</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Coetzee</surname><given-names>D</given-names></name><name><surname>Hilderbrand</surname><given-names>K</given-names></name><name><surname>Boulle</surname><given-names>A</given-names></name><name><surname>Draper</surname><given-names>B</given-names></name><name><surname>Abdullah</surname><given-names>F</given-names></name><name><surname>Goemaere</surname><given-names>E</given-names></name></person-group><article-title>Effectiveness of the first district-wide programme for the prevention of mother-to-child transmission of HIV in South Africa</article-title><source>Bull World Health Organ</source><year>2005</year><volume>83</volume><fpage>489</fpage><lpage>94</lpage><pub-id pub-id-type="pmid">16175822</pub-id></element-citation></ref><ref id="CIT0034"><label>34</label><element-citation publication-type="webpage"><collab>National Department of Health</collab><article-title>Clinical guidelines: prevention of mother-to-child transmission (PMTCT) 2010</article-title><comment>Available from: <ext-link ext-link-type="uri" xlink:href="http://wwwfidssacoza/images/PMTCT_Guidelinespdf">http://wwwfidssacoza/images/PMTCT_Guidelinespdf</ext-link> [cited 23 October 2012].</comment></element-citation></ref><ref id="CIT0035"><label>35</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Ndirangu</surname><given-names>J</given-names></name><name><surname>Newell</surname><given-names>M-L</given-names></name><name><surname>Tanser</surname><given-names>F</given-names></name><name><surname>Herbst</surname><given-names>AJ</given-names></name><name><surname>Bland</surname><given-names>R</given-names></name></person-group><article-title>Decline in early life mortality in a high HIV prevalence rural area of South Africa: evidence of HIV prevention or treatment impact?</article-title><source>AIDS</source><year>2010</year><volume>24</volume><fpage>593</fpage><lpage>602</lpage><pub-id pub-id-type="pmid">20071975</pub-id></element-citation></ref><ref id="CIT0036"><label>36</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Katapa</surname><given-names>R</given-names></name></person-group><article-title>A comparison of female-and male-headed households in Tanzania and poverty implications</article-title><source>J Biosoc Sci</source><year>2006</year><volume>38</volume><fpage>327</fpage><pub-id pub-id-type="pmid">16613619</pub-id></element-citation></ref><ref id="CIT0037"><label>37</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Wiegers</surname><given-names>E</given-names></name><name><surname>Curry</surname><given-names>J</given-names></name><name><surname>Garbero</surname><given-names>A</given-names></name><name><surname>Hourihan</surname><given-names>J</given-names></name></person-group><article-title>Patterns of vulnerability to AIDS impacts in Zambian households</article-title><source>Dev Change</source><year>2006</year><volume>37</volume><fpage>1073</fpage><lpage>92</lpage></element-citation></ref><ref id="CIT0038"><label>38</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Oni</surname><given-names>S</given-names></name><name><surname>Obi</surname><given-names>C</given-names></name><name><surname>Okorie</surname><given-names>A</given-names></name><name><surname>Thabede</surname><given-names>D</given-names></name><name><surname>Jordan</surname><given-names>A</given-names></name></person-group><article-title>The economic impact of HIV/AIDS on rural households in Limpopo Province</article-title><source>S Afr J Econ</source><year>2002</year><volume>70</volume><fpage>551</fpage><lpage>62</lpage></element-citation></ref><ref id="CIT0039"><label>39</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Collinson</surname><given-names>MA</given-names></name></person-group><source>Striving against adversity: the dynamics of migration, health and poverty in rural South Africa</source><year>2009</year><publisher-loc>Umeå</publisher-loc><publisher-name>Medical Dissertations, Umeå University</publisher-name></element-citation></ref><ref id="CIT0040"><label>40</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Posel</surname><given-names>D</given-names></name><name><surname>Casale</surname><given-names>D</given-names></name></person-group><article-title>What has been happening to internal labour migration in South Africa, 1993–1999?</article-title><source>S Afr J Econ</source><year>2003</year><volume>71</volume><fpage>455</fpage><lpage>79</lpage></element-citation></ref><ref id="CIT0041"><label>41</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Hahn</surname><given-names>MB</given-names></name><name><surname>Riederer</surname><given-names>AM</given-names></name><name><surname>Foster</surname><given-names>SO</given-names></name></person-group><article-title>The livelihood vulnerability index: a pragmatic approach to assessing risks from climate variability and change--a case study in Mozambique</article-title><source>Global Environ Change</source><year>2009</year><volume>19</volume><fpage>74</fpage><lpage>88</lpage></element-citation></ref><ref id="CIT0042"><label>42</label><element-citation publication-type="journal"><person-group person-group-type="author"><name><surname>Beguy</surname><given-names>D</given-names></name><name><surname>Bocquier</surname><given-names>P</given-names></name><name><surname>Zulu</surname><given-names>EM</given-names></name></person-group><article-title>Circular migration patterns and determinants in Nairobi slum settlements</article-title><source>Demographic Res</source><year>2010</year><volume>23</volume><fpage>549</fpage><lpage>86</lpage></element-citation></ref><ref id="CIT0043"><label>43</label><element-citation publication-type="book"><collab>Stata Corporation</collab><article-title>Stata statistical software release 12</article-title><year>2011</year><publisher-loc>College Station, TX</publisher-loc><publisher-name>Stata Corporation LP</publisher-name></element-citation></ref><ref id="CIT0044"><label>44</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Kline</surname><given-names>RB</given-names></name></person-group><source>Principles and practice of structural equation modeling</source><year>2011</year><edition>3rd ed</edition><publisher-loc>New York</publisher-loc><publisher-name>Guilford Press</publisher-name></element-citation></ref><ref id="CIT0045"><label>45</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Bollen</surname><given-names>KA</given-names></name></person-group><source>Structural equations with latent variables</source><year>1989</year><publisher-loc>New York</publisher-loc><publisher-name>Wiley</publisher-name></element-citation></ref><ref id="CIT0046"><label>46</label><element-citation publication-type="book"><person-group person-group-type="author"><name><surname>Bollen</surname><given-names>KA</given-names></name><name><surname>Curran</surname><given-names>PJ</given-names></name></person-group><source>Latent curve models: a structural equation perspective</source><year>2006</year><publisher-loc>New York</publisher-loc><publisher-name>Wiley-Interscience</publisher-name></element-citation></ref></ref-list></back></pmc></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Sante/explor/SidaSubSaharaV1/Data/Pmc/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 002380 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Pmc/Corpus/biblio.hfd -nk 002380 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Sante
|area= SidaSubSaharaV1
|flux= Pmc
|étape= Corpus
|type= RBID
|clé= PMC:3556702
|texte= The contribution of spatial analysis to understanding HIV/TB mortality in children: a structural equation modelling approach
}}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/Pmc/Corpus/RBID.i -Sk "pubmed:23364095" \
| HfdSelect -Kh $EXPLOR_AREA/Data/Pmc/Corpus/biblio.hfd \
| NlmPubMed2Wicri -a SidaSubSaharaV1
| This area was generated with Dilib version V0.6.32. |  |