Paediatric and Perinatal Epidemiology, 2025; 39:698–710 https://doi.org/10.1111/ppe.70067 698 Paediatric and Perinatal Epidemiology ORIGINAL ARTICLE OPEN ACCESS Major Causes of Perinatal and Paediatric Mortality in Sub-Saharan Africa and South Asia: Adjustment for Selection Bias in the CHAMPS Network Kartavya J. Vyas1,2,3   | Jonathan A. Muir1,2  | Zachary J. Madewell2,4  | Priya M. Gupta1,2  | Dianna M. Blau2,4  | Shams E. Arifeen5  | Emily S. Gurley6  | Atique I. Chowdhury5  | Kazi M. Islam5  | Afruna Rahman5  | J. Anthony G. Scott7,8  | Nega Assefa8   | Lola Madrid7,8  | Yohanis A. Asefa8  | Yasir Y. Abdullahi8   | Dickens Onyango9  | Victor Akelo10,11  | Beth A. Tippett-Barr10,11  | George Aol12  | Samba O. Sow13  | Karen L. Kotloff14  | Milagritos D. Tapia14  | Adama M. Keita13  | Kiranpreet Chawla13  | Quique Bassat15,16   | Inacio Mandomando16,17  | Ariel Nhacolo16  | Charfudin Sacoor16  | Ikechukwu Ogbuanu18  | Dickens Kowuor18  | Babatunde Duduyemi18  | Andrew Moseray18  | James S. Squire18   | Shabir Madhi19  | Sana Mahtab19  | Yasmin Adam19  | Amy Wise19  | Takwanisa Machemedza19   | Cynthia G. Whitney1,2  | on behalf of the CHAMPS Network 1Rollins School of Public Health, Emory University, Atlanta, Georgia, USA  |  2Emory Global Health Institute, Emory University, Atlanta, Georgia, USA  |  3U.S. Military HIV Research Program, Walter Reed Army Institute of Research, Bethesda, Maryland, USA  |  4Center for Global Health, US Centers for Disease Control and Prevention, Atlanta, Georgia, USA  |  5International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh  |  6Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland, USA  |  7London School of Hygiene & Tropical Medicine, London, UK  |  8College of Health and Medical Sciences, Haramaya University, Harar, Ethiopia  |  9Kisumu County Department of Health, Kisumu, Kenya  |  10US Centers for Disease Control and Prevention-Kenya, Kisumu, Kenya  |  11US Centers for Disease Control and Prevention-Kenya, Nairobi, Kenya  |  12KEMRI-Wellcome Trust Research Programme, Kilifi, Kenya  |  13Centre pour le Développement des Vaccins, Ministère de la Santé, Bamako, Mali  |  14School of Medicine, University of Maryland, Baltimore, Maryland, USA  |  15ISGlobal, Hospital Clínic, Universitat de Barcelona, Barcelona, Spain  |  16Centro de Investigação em Saúde de Manhiça, Maputo, Mozambique  |  17Instituto Nacional de Saúde, Maputo, Mozambique  |  18Crown Agents, Freetown, Sierra Leone  |  19South African Medical Research Council Vaccines and Infectious Diseases Analytics Research Unit, University of the Witwatersrand, Johannesburg, South Africa Correspondence: Kartavya J. Vyas (kvyas@global-id.org) Received: 7 December 2024  |  Revised: 5 August 2025  |  Accepted: 27 August 2025 Funding: This work was supported by Bill and Melinda Gates Foundation (OPP1126780). Keywords: cause of death | child mortality | selection bias | South Asia | stillbirths | Sub-Saharan Africa ABSTRACT Background: Studies of child mortality that employ minimally invasive tissue sampling (MITS) produce highly accurate cause of death data; however, selection bias may render these as non-representative of their underlying populations. Objectives: Estimate cause-specific mortality fractions and rates for the five most frequent causes—underlying and others in the chain of events leading to death—among stillbirths, neonatal, infant and child deaths—in Sub-Saharan Africa and South Asia, adjusted for any identified selection biases. Methods: The Child Health and Mortality Prevention Surveillance (CHAMPS) Network collects standardised, population- based, longitudinal data on causes of death among stillbirths and under-five children in 12 catchments in seven countries in Sub- Saharan Africa and South Asia. Cause-specific mortality fractions and rates were calculated for the five most frequent causes among stillbirths, neonatal, infant and child deaths, and for the five most frequent maternal conditions among perinatal deaths; This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2025 The Author(s). Paediatric and Perinatal Epidemiology published by John Wiley & Sons Ltd. Data from this manuscript were presented at the virtual American Society of Tropical Medicine and Hygiene (ASTMH) meeting, 17–21 November 2021. *See Appendix for the members of CHAMPS Network. https://doi.org/10.1111/ppe.70067 https://doi.org/10.1111/ppe.70067 mailto:kvyas@global-id.org https://orcid.org/0000-0001-7443-7940 https://orcid.org/0000-0003-0341-2329 https://orcid.org/0000-0003-3402-0029 https://orcid.org/0000-0003-0875-7596 https://orcid.org/0000-0002-8593-7972 https://orcid.org/0000-0002-9308-7525 mailto:kvyas@global-id.org http://creativecommons.org/licenses/by/4.0/ http://crossmark.crossref.org/dialog/?doi=10.1111%2Fppe.70067&domain=pdf&date_stamp=2025-09-04 699Paediatric and Perinatal Epidemiology, 2025 all estimates were subsequently adjusted for selection bias. Selection probabilities were estimated from membership in subgroups defined by factors hypothesised to affect selection. Results: In 2017–2020, of 10,122 deaths ascertained, 5847 (57.8%) were enrolled in CHAMPS and 2654 (26.2%) additionally con- sented to MITS. Estimates were calculated for 265 and 65 site/age-specific causes of death and maternal conditions, respectively; five (1.9%) and four (6.2%) required adjustment, respectively, but they did not meaningfully change. Estimates were calculated for 34 site-specific causes of death among all stillbirths and under-five deaths combined; 28 (82.4%) required adjustment (all included age at death), and change-in-estimates demonstrated considerable variability. Conclusions: Selection bias is not a concern in the CHAMPS Network. Deaths where MITS were performed accurately represent the distribution of causes of death in their respective target populations, specifically when stratified by age or adjusted accordingly. Future studies of child mortality that employ MITS should consider adjusting for age at death for their measures of frequency. 1   |   Introduction Sub-Saharan Africa and South Asia accounted for 77% of 2.0 mil- lion stillbirths [1] and 82% of 5.2 million under-five deaths [2] worldwide in 2019. Current preventive efforts are informed by child mortality estimates based on vital registration systems, pop- ulation censuses, verbal autopsies and household surveys—all of which are often incomplete and prone to systematic error [2–4]. Global estimates of child mortality aetiologies have traditionally relied on the underlying cause, neglecting other causes in the chain of events leading to death and omitting maternal conditions that may have precipitated or indirectly contributed to perinatal deaths [3, 5–7]. In response, the Child Health and Mortality Prevention Surveillance (CHAMPS) Network was established in 2015 to col- lect detailed, standardised, population-based, longitudinal data within a network of sites in areas with high child mortality, with the overarching objectives of understanding and tracking prevent- able causes of stillbirths and under-five deaths globally [3, 8]. CHAMPS employs minimally invasive tissue sampling (MITS) to facilitate post-mortem pathology, microbiology, molecular and other diagnostic testing, including the identification of specific pathogens; however, not all those enrolled consent to this proce- dure [5, 9]. MITS has been shown to be less physically disruptive, faster, more culturally acceptable, less resource-intensive and nearly as valid as a complete diagnostic autopsy for determining causes of death [3, 10–13]. However, only a subset of all known eli- gible deaths in these areas enrol in CHAMPS, and an even smaller subset undergo MITS, as consent may not have been granted or burial may have occurred before the family was approached for enrolment [14]. Reasons for MITS refusal may include having had a negative experience with hospital care, concerns with disfigur- ing the body or delaying the funeral, feeling there is no need for further examination and religious or cultural concerns [15–18]. Families of deceased children who do not consent to MITS tend to be less educated and less likely to be employed [16]. Moreover, infant and child deaths and deaths that occur in the community tend to be underrepresented in CHAMPS compared to stillbirths and neonatal deaths and in-hospital deaths, possibly due to is- sues of logistics and feasibility [14, 19]. Cause-specific mortality estimates that rely solely on deaths where MITS were performed may therefore not accurately represent their underlying target populations [20]. Selection bias is introduced when deaths where MITS were and were not performed differ in terms of character- istics that affect both selection and the outcome of interest; these associations must be independently assessed for each exposure- outcome pair [21, 22]. Previously published data from CHAMPS have reported cause-specific mortality frequencies or fractions that only represent deaths where MITS were performed, without adjusting for possible selection bias and extrapolating to all deaths in the catchment areas [3, 5]. Objectives of this current work include: (1) estimating crude mortality fractions and rates for the five most frequent causes of death—underlying and others in the chain of events leading to death—among stillbirths, neonatal, infant and child deaths in each site; (2) estimating crude fractions and rates for the five most frequent maternal conditions for stillbirths and neonatal deaths in each site; (3) adjusting all estimates for any identified selection biases so that they are representative of their target populations; and (4) comparing adjusted mortality fractions and rates with those published in the scientific literature. 2   |   Methods 2.1   |   Mortality and Demographic Surveillance CHAMPS is an ongoing study conducted in 12 catchments in seven countries in Sub-Saharan Africa and South Asia: Baliakandi and Faridpur, Bangladesh; Haramaya, Harar and Kersa, Ethiopia; Manyatta and Siaya, Kenya; Bamako, Mali; Manhiça and Quelimane, Mozambique; Makeni, Sierra Leone; and Soweto, South Africa. Site characteristics and selection criteria have been described elsewhere [14]. Of note, not all CHAMPS sites began enrolment at the same time; Mozambique began in 2016; South Africa, Kenya, Mali and Bangladesh began in 2017; and Sierra Leone and Ethiopia began in 2019. Most catchments—except Faridpur, Bangladesh; Quelimane, Mozambique; and Makeni, Sierra Leone—carry out mortality surveillance within a demographic surveillance system (DSS) that captures sociodemographic characteristics and data on births, deaths, pregnancies and in- and out-migration episodes within a geographically defined area [23–26]. Eligible deaths captured in the DSS but never enrolled in CHAMPS are referred to as ‘DSS only deaths’. Ethics committees overseeing investi- gators at each site and at Emory University (Atlanta, GA, USA) approved overall and site-specific protocols, as appropriate. According to CHAMPS procedures, attempts were made to no- tify staff of stillbirths and under-five deaths within the first 24 h. Shortly thereafter, CHAMPS staff approached families for eli- gibility screening. All stillborn foetuses and deceased children who resided within the catchment were eligible for inclusion. 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 700 Paediatric and Perinatal Epidemiology, 2025 Families provided consent for MITS [27], verbal autopsy (VA) [28] and clinical data abstraction. MITS eligibility required that CHAMPS staff were notified within 24 h of death (or ≤ 72 h if post-mortem refrigeration was used) and that the body of the stillborn foetus or deceased child was available for specimen collection. Families of stillborn foetuses and deceased children who were not eligible for the MITS procedure were asked to consent for only the VA interview and clinical data abstraction. Deaths enrolled in CHAMPS whose families have and have not granted consent for MITS are referred to as MITS and non-MITS cases, respectively. 2.2   |   Specimen and Data Collection All deaths between 1 January 2017 and 31 December 2020 were included in the current analysis. Methods used for specimen and data collection for the MITS procedure have been described previously [3]. Briefly, trained staff took photographs and an- thropometric measurements before specimen collection. Tissue specimens were collected from the lungs, heart, brain, liver and bone marrow; peripheral blood, cerebrospinal fluid, stool and nasopharyngeal secretions were also collected [27]. Blood sam- ples were tested for HIV DNA or RNA by PCR and for malaria using thick and thin smears and rapid diagnostic tests; blood and cerebrospinal fluid underwent culture for bacteria. Five custom TaqMan Array Cards (TAC; ThermoFisher Scientific, Waltham, MA, USA) with specific molecular assays were used to detect 116 pathogen targets [27, 29]. Tissues were examined using histopathology techniques, including routine and special stains [30]. Data were abstracted from all available clinical re- cords of deceased children, and relevant maternal health records were abstracted for stillbirths and neonatal deaths. Families were interviewed using translations of the 2016 World Health Organization (WHO) VA instrument [14, 28]. 2.3   |   Cause of Death Determination A complete description of the determination of cause of death (DeCoDe) process has been described elsewhere [3, 6]. Briefly, all available data for each case were reviewed by DeCoDe panels convened at each site, consisting of paediatricians, obstetricians, epidemiologists, pathologists, microbiologists and other subject- matter experts [6]. The panels reviewed available case data and determined the chain of events (immediate, underlying, comor- bid causes—hereafter referred to as the causal chain) leading to death and assigned WHO International Classification of Diseases, Tenth Revision (ICD-10) codes [31]. Panels also iden- tified and assigned maternal conditions contributing to perina- tal deaths using WHO ICD-Perinatal Mortality (PM) codes [32]. Causes and maternal conditions were grouped according to the Global Burden of Disease (GBD) categories for analysis [33]. MITS cases that have been reviewed by a DeCoDe panel and as- signed a cause of death are referred to as DeCoDed MITS cases. 2.4   |   Statistical Analysis Details of the analytic method have been described elsewhere [34]. Descriptive analyses were performed to characterise all ascertained stillbirths and under-five deaths in the DSS areas by whether the deaths were enrolled in CHAMPS and whether MITS were performed. Factors of interest included: site of death; age at death (stillbirth [no spontaneous breathing or move- ment at time of delivery and (1) weighing > 1 kg and/or (2) es- timates gestational age ≥ 28 weeks], neonate [0–28 days], infant [29–364 days] or child [1–5 years]); sex at birth (male or female); location of death (community or facility); season of death (dry or rainy); VA cause of death (Inter-VA [35]: infection, trauma or other); and maternal education (none, primary, secondary or tertiary). If DSS data were available, it is assumed that all CHAMPS cases were also captured in the DSS. The five most frequent underlying causes and causes anywhere in the causal chain were identified by age group (including still- births and under-five deaths combined—hereafter referred to as total under-five [36]) and site, and the five most frequent ma- ternal conditions were identified for stillbirths and neonates by site. Cause-specific mortality fractions (CSMF) were calculated for each of the five most frequent causes or maternal conditions, by age group and site. Crude CSMFs (cCSMF) were calculated as the proportion of age-specific deaths attributed to each cause among all age-specific DeCoDed MITS cases. Selection bias was hypothesised to be the greatest threat to the external validity of CHAMPS MITS cases to the underlying target populations; other biases were not explored. To calculate adjusted CSMFs (aCSMF), factors hypothesised to affect selection (age at death, sex at birth, location of death, season of death, VA cause of death and maternal education) had to meet four a priori criteria for adjustment: (1) associated with MITS consent; (2) missing < 20% data when comparing MITS and non-MITS cases; (3) associated with the specific cause of death; and (4) missing < 20% data when comparing cause-specific deaths and deaths due to all other causes. Factors were selected for adjustment if one or at most two factors (where age at death must be one of the two, due to data limitations) met all four criteria. If three or more factors met all four criteria, the top two were selected based on the fol- lowing a priori hierarchy: age at death, season of death, location of death, VA cause of death, sex at birth and maternal educa- tion. Age-specific CHAMPS cases (non-MITS and MITS), MITS cases and all deaths in the target population were stratified by the factors that met selection. Selection probabilities were cal- culated as the proportion of age-specific MITS cases among all eligible age-specific deaths in the target population for each stra- tum. Direct standardisation was then performed for factors that had met all four criteria. The target population for most sites was all eligible age-specific deaths ascertained in the DSS in each respective site. However, for sites where DSS data were unavail- able for one or more catchments, the target population consisted of all age-specific CHAMPS deaths regardless of MITS consent. Crude and adjusted cause-specific mortality rates (cCSMR and aCSMR, respectively) were calculated as the product of the cCSMF or aCSMF and the all-cause age-specific mortality rate, respectively, for each of the five most frequent causes or maternal conditions. Where DSS data were available, the all- cause age-specific mortality rate in the target population was calculated as the number of age-specific deaths among all live- births (and stillbirths, when appropriate) [37]. Where DSS data were unavailable (Mali and Sierra Leone), the all-cause age- specific mortality rate in the target population was sourced from 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 701Paediatric and Perinatal Epidemiology, 2025 the Demographic and Health Surveys (DHS) Program [38], in which case the all-cause age-specific mortality rate may repre- sent a larger geographic region than the catchment itself, and the year of data collection may not exactly coincide with that of CHAMPS. We calculated 90% Bayesian credible intervals (CrI) based on a non-informative prior distribution for all estimates. Analyses were performed in R 4.2.1 (R Core Team, Vienna, Austria) [39]. An R package called champsmortality was specifically devel- oped and validated to help calculate these estimates [40]. All re- search conformed to the principles embodied in the Declaration of Helsinki. 2.5   |   Missing Data Among all ascertained deaths in the DSS and CHAMPS, 1.0% were missing site of death (DSS only deaths, non-MITS cases and MITS cases: 2.4%, 0.0% and 0.0%, respectively), 7.2% were missing sex at birth (16.1%, 0.8% and 0.2%, respectively), 11.4% were missing location of death (26.0%, 0.7% and 0.0%, respec- tively), 1.8% were missing season of death (4.1%, 0.0% and 0.0%, respectively), 43.3% were missing VA cause of death (78.5%, 16.3% and 17.8%, respectively), and 53.6% were missing maternal education (34.7%, 61.2% and 75.4%, respectively). 2.6   |   Ethics Approval Ethics committees overseeing investigators at each site and at Emory University (Atlanta, GA, USA) approved overall and site- specific protocols, as appropriate. The U.S. Centers for Disease Control and Prevention (CDC; Atlanta, GA, USA) relied on eth- ical review committees at Emory University and at individual sites where CDC staff were directly engaged to review the pro- tocol. All research conformed to the principles embodied in the Declaration of Helsinki. 3   |   Results 3.1   |   Enrolment and Consent A total of 15,106 unique death notifications were received in 2017–2020, of which 7844 (51.9%) were deemed eligible, 6968 (46.1%) were requested to consent and 5847 (38.7%) enrolled in CHAMPS. Of those enrolled, 3024 (51.7%) were requested to con- sent for MITS, 2707 (46.3%) granted consent, 2654 (45.4%) had the procedure performed and 2547 (43.6%) have been DeCoDed. The analytic sample consists of 10,122 deaths: 4275 DSS only deaths and 5847 deaths enrolled in CHAMPS (Figure S1). Venn diagrams of the proportions of eligible deaths in the DSS en- rolled in CHAMPS and consented to MITS are depicted by site and age group in Figure S2. 3.2   |   Sample Characteristics Most CHAMPS cases were stillbirths (33.6%) or neonates (36.7%), male (55.2%), died in a facility (74.8%), died during the dry season (57.3%), had a VA cause of death other than infec- tion or trauma (77.6%) and were born to mothers educated at the primary level (40.5%; Table 1). DSS only deaths were predom- inantly children (39.5%), male (55.3%), died in the community (54.4%), during the dry season (54.0%), had a VA cause of death other than infection or trauma (52.9%) and were born to mothers educated at the secondary level (36.8%). 3.3   |   Stillbirths Regarding cause-specific mortality, the criteria for adjustment were never met for any site (Figure 1A and Figure S3A). Regarding contributing maternal conditions, the criteria for adjustment for MITS ascertainment were only met for umbilical cord compli- cations in Kenya (crude fraction [90% CrI] = 7.5% [4.4, 11.9], ad- justed fraction [90% CrI] = 10.7% [8.0, 13.9], change-in-estimate (CIE) = 3.2%; crude rate [90% CrI] = 1.8 per 1000 births [1.1, 2.9], adjusted rate [90% CrI] = 2.6 per 1000 births [1.9, 3.3], CIE = 0.8; adjusted for location of death; Figure 2A and Figure S4A). 3.4   |   Neonates Regarding cause-specific mortality, the criteria for adjustment were only met for perinatal asphyxia/hypoxia in Kenya (cCSMF [90% CrI] = 34.0% [27.9, 40.6], aCSMF [90% CrI] = 35.5% [32.1, 39.1], CIE = 1.5%; cCSMR [90% CrI] = 6.5 per 1000 live-births [5.4, 7.8], aCSMR [90% CrI] = 6.8 per 1000 live-births [6.2, 7.5], CIE = 0.3; adjusted for season of death) and neonatal preterm birth complications in South Africa (cCSMF [90% CrI] = 62.2% [58.1, 66.2], aCSMF [90% CrI] = 61.6% [57.8, 65.4], CIE = −0.6%; cCSMR [90% CrI] = 26.6 per 1000 live-births [24.8, 28.3], aCSMR [90% CrI] = 26.3 per 1000 live-births [24.7, 27.9], CIE = −0.3; adjusted for location of death; Figure 1B and Figure S3B). Regarding contrib- uting maternal conditions, the criteria for adjustment were only met for premature rupture of membranes in Bangladesh (crude fraction [90% CrI] = 9.2% [5.4, 14.5], CIE = 1.1%; crude rate [90% CrI] = 2.2 per 1000 live-births [1.3, 3.5], CIE = 0.3; adjusted for sea- son of death), other labour and delivery complications in Kenya (crude fraction [90% CrI] = 14.1% [9.9, 19.3], adjusted fraction [90% CrI] = 14.9% [12.4, 17.6], CIE = 0.8%; crude rate [90% CrI] = 3.4 per 1000 live-births [2.4, 4.6], adjusted rate [90% CrI] = 3.6 per 1000 live-births [3.0, 4.2], CIE = 0.2; adjusted for season of death) and preterm labour or delivery in Kenya (crude fraction [90% CrI] = 7.4% [4.5, 11.5], CIE = 0.6%; crude rate [90% CrI] = 1.8 per 1000 live-births [1.1, 2.8], CIE = 0.1; adjusted for season of death; Figure 2B and Figure S4B). 3.5   |   Infants and Children Criteria for adjustment were never met for infants (Figure 1D and Figure  S3D) or children (Figure  1D and Figure  S3D) for any site. 3.6   |   Total Under-Five Twenty-eight out of 34 (82.4%) causes examined met the criteria for adjustment; of which, 15 (53.6) were adjusted for age at death 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 702 Paediatric and Perinatal Epidemiology, 2025 TABLE 1    |    Characteristics of all ascertained stillbirths and under-five deaths in demographic surveillance systems by CHAMPS enrolment and MITS performed (2017–2020). DSS onlya CHAMPS Non-MITSb MITSc N = 4380 N = 3193 N = 2654 n (%)d n (%)d n (%)d Case characteristics Site of deathe Bangladesh 161 (3.8) 918 (28.8) 212 (8.0) Ethiopia 1156 (27.0) 223 (7.0) 173 (6.5) Kenya 1394 (32.6) 185 (5.8) 500 (18.8) Malif — 839 (26.3) 194 (7.3) Mozambique 1289 (30.2) 581 (18.2) 646 (24.3) Sierra Leoneg — 303 (9.5) 229 (8.6) South Africa 275 (6.4) 144 (4.5) 700 (26.4) Age at deathe,f,g Stillbirths 718 (16.4) 1124 (35.2) 838 (31.6) Neonates 875 (20.0) 1058 (33.1) 1087 (41.0) Infants 1059 (24.2) 535 (16.8) 382 (14.4) Children 1728 (39.5) 476 (14.9) 347 (13.1) Sex at birthe Male 2034 (55.3) 1737 (54.8) 1476 (55.7) Female 1642 (44.7) 1431 (45.2) 1174 (44.3) Location of deathe Community 1763 (54.4) 1170 (36.9) 299 (11.3) Facility 1477 (45.6) 2002 (63.1) 2354 (88.7) Season of deathe,h Dry 2267 (54.0) 1747 (54.7) 1605 (60.5) Rainy 1932 (46.0) 1446 (45.3) 1049 (39.5) Verbal autopsy cause of deathe,i Infection 410 (43.6) 559 (20.9) 430 (19.7) Trauma 33 (3.5) 75 (2.8) 25 (1.1) Other 497 (52.9) 2040 (76.3) 1728 (79.2) Maternal characteristicsj Educationg None 809 (28.3) 307 (24.8) 110 (16.9) Primary 918 (32.1) 509 (41.1) 256 (39.3) Secondary 1053 (36.8) 339 (27.4) 184 (28.2) Tertiary 79 (2.8) 83 (6.7) 102 (15.6) Abbreviations: CHAMPS, Child Health and Mortality Prevention Surveillance Network; DSS, demographic surveillance system; ICD-10, International Classification of Diseases, Revision 10; MITS, minimally invasive tissue sampling; VA, verbal autopsy. aEligible stillbirths and under-five deaths captured in the DSS but never enrolled in CHAMPS. bStillbirths and under-five deaths enrolled in CHAMPS but for whom MITS was not performed. cStillbirths and under-five deaths enrolled in CHAMPS and for whom MITS was performed. dPercentages (column distributions) may not sum to 100% due to rounding. eMissing: [DSS only]: site (n = 105), age at death (n = 0), sex at birth (n = 704), location of death (n = 1140), season of death (n = 181), VA cause of death (n = 3440), maternal education (n = 1521); [Non-MITS]: sex at birth (n = 25), location of death (n = 21), VA cause of death (n = 519), maternal education (n = 1955); [MITS]: sex at birth (n = 4), location of death (n = 1), VA cause of death (n = 471), maternal education (n = 2002). f[Mali]: DSS data included in non-MITS; [Sierra Leone]: no DSS data available. gStillbirths (no spontaneous breathing or movement at time of delivery and [1] weighing > 1 kg and/or [2] estimated gestational age ≥ 28 weeks); neonates (0–28 days); infants (29–365 days); children (1–5 years). hDry, rainy: BD (November–May, June–October); ET (October–May, June–September); KE (July and December–March, April–June and August–November); ML (November–May, June–October); MZ (May–October, November–April); SL (October–May, June–September); ZA (March–November, December–February). iInter-VA algorithm: Infection (ICD-10 codes 01, 10.3–10.5); trauma (ICD-10 code 12). jAll characteristics pertinent to time of pregnancy. 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 703Paediatric and Perinatal Epidemiology, 2025 FIGURE 1    |    Rates of the most frequent perinatal and paediatric causes of death in the CHAMPS Network, 2017–2020. Estimates are adjusted for selection bias due to enrolment and post-mortem sampling. (A) Stillbirths, no spontaneous breathing or movement at time of delivery and (1) weigh- ing > 1 kg and/or (2) estimated gestational age ≥ 28 weeks. (B) Neonates (0–28 days). (C) Infants (29–365 days). (D) Children (1–5 years). (E) Under- five, includes stillbirths, neonates, infants and children. CHAMPS, Child Health and Mortality Prevention Surveillance Network. 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 704 Paediatric and Perinatal Epidemiology, 2025 FIGURE 1    |     (Continued) 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 705Paediatric and Perinatal Epidemiology, 2025 and location of death, 10 (35.7%) were adjusted for age at death only, 2 (7.1%) were adjusted for age at death and VA cause of death, and 1 (3.6%) was adjusted for age at death and season of death (Figure 1E and Figure S3E). The largest CSMF and CSMR CIE occurred for sepsis in Ethiopia (cCSMF [90% CrI] = 9.0% [5.9, 13.2], aCSMF [90% CrI] = 52.7% [50.5, 54.9], CIE = 43.7%; cCSMR [90% CrI] = 3.8 per 1000 births [2.5, 5.6], aCSMR [90% CrI] = 22.5 per 1000 births [21.6, 23.4], CIE = 18.7; adjusted for age at death and location of death). On average, adjustment increased the CSMF by 3.4% (standard deviation [SD] = 15.6, range = [−33.6, 43.7]) and increased the CSMR by 1.4 per 1000 births (SD = 6.6, range = [−14.4, 18.7]). 4   |   Comment 4.1   |   Principal Findings The work presented represents a robust analysis to extrapolate age-specific fractions and rates of cause-specific mortality, and of maternal conditions, from non-representative samples of DeCoDed MITS cases to larger target populations. Previous esti- mates from CHAMPS and other studies that employ MITS have reported cause-specific mortality frequencies or fractions that only represent MITS cases, without adjusting for selection bias [3, 5, 41–44]. That said, the results here suggest that such an ad- justment may not always be necessary, that crude estimates are often just as valid, particularly when stratified by age [20, 45]. 4.2   |   Strengths of the Study Strengths of this study include its standardised, population- based, longitudinal approach; employment of MITS for post-mortem evaluation; standardised DeCoDe process; deter- mination of not only the underlying cause, but also the causal chain; determination of the main maternal conditions; use of DSS data to characterise deaths never enrolled in CHAMPS; and application of a method to extrapolate sparse gold standard cause of death data to characterise broader catchment areas, thereby addressing the issue of external validity, a common crit- icism of CHAMPS data. 4.3   |   Limitations of the Data Limitations of this study include unavailable or discrepant DSS data for some sites; all-cause mortality rates sourced from DHS when DSS data were unavailable, which may not coin- cide geographically or temporally with CHAMPS; sparse data, particularly for uncommon outcomes; underlying assump- tions—missing at random, ad-hoc criteria to identify factors for adjustment, no residual bias due to categorisation of age, hierar- chy of factors when more than 2 factors met criteria, no unmea- sured factors that affect selection and all CHAMPS deaths are a subset of DSS deaths—presumed true; and extrapolation beyond the target population was not performed, therefore adjusted esti- mates do not represent the entire site. FIGURE 1    |     (Continued) 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 706 Paediatric and Perinatal Epidemiology, 2025 FIGURE 2    |    Rates of the most frequent contributing maternal conditions in the CHAMPS Network, 2017–2020. Estimates are adjusted for selec- tion bias due to enrolment and post-mortem sampling. (A) Stillbirths, no spontaneous breathing or movement at time of delivery and (1) weighing > 1 kg and/or (2) estimated gestational age ≥ 28 weeks. (B) Neonates (0–28 days). CHAMPS, Child Health and Mortality Prevention Surveillance Network. 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 707Paediatric and Perinatal Epidemiology, 2025 4.4   |   Interpretation For the major age-specific underlying and causal chain causes of death in each site, there were only five instances where ad- justment for selection bias was necessary; even still, both the CSMF and CSMR did not meaningfully change. For the major age-specific maternal conditions in each site, there were only four instances where adjustment for selection bias was neces- sary; even still, both the CSMF and CSMR did not meaning- fully change. In contrast, for the major stillbirth and under-five causal chain causes of death in each site, there were 53 instances where adjustment for selection bias was necessary, all of which included age at death; here, we find considerable variability in the change of both CSMF and CSMR. The major country-level causes of death among neonatal, infant and child deaths reported by the Institute for Health Metrics and Evaluation (IHME) GBD Study [2, 46, 47] are mostly consistent with those reported here, in that both report them as highly fre- quent causes and, although their point estimates do not corre- spond exactly, they have overlapping measures of uncertainty. Notably, results are less consistent for infant and child deaths. This discrepancy, however, is likely because some causes are more likely to be attributed as an immediate or comorbid cause of death among infants and children, not necessarily an underly- ing cause, and therefore would not be captured in the IHME GBD data. Estimates for CSMRs published by the WHO Global Health Observatory [48–50] (GHO) are also largely consistent with those reported here, although comparisons are difficult as they do not provide a measure of uncertainty. And yet, considerable differ- ences also exist. Again, this inconsistency is likely because the WHO GHO data do not capture immediate or comorbid causes of death. Thus, due to differing methodological approaches, opera- tional definitions, and geographical representations, direct com- parisons should be made with abundant caution, if at all. 5   |   Conclusions The CHAMPS Network can better inform and help target pre- ventive efforts to curtail child mortality in Sub-Saharan Africa and South Asia. The work presented here has shown that MITS cases appear to accurately represent the distribution of causes of death in their respective underlying target populations, when stratified by age or otherwise adjusted accordingly. Future stud- ies of child mortality that employ MITS should consider adjust- ing for age at death for their measures of frequency. Moreover, efforts should be made to further extrapolate these CHAMPS data so inferences can be made at the national level. Author Contributions D.M.B., S.E.A., E.S.G., J.A.G.S., N.A., D.O., V.A., S.O.S., K.L.K., Q.B., I.M., I.O. and S.Madhi designed the protocol and led the involvement of sites in the CHAMPS Network. A.I.C., K.M.I., A.R., L.M., Y.A.A., Y.Y.A., D.O., B.A.T.-B., G.A., M.D.T., A.M.K., K.C., A.N., C.S., D.K., B.D., A.M., J.S.S., S.M., Y.A.A., A.W. and T.M. helped coordinate and implement CHAMPS procedures in the sites. K.J.V., J.A.M., Z.J.M., P.M.G. and C.G.W. conceptualised the manuscript. K.J.V., J.A.M., Z.J.M. and P.M.G. directed data management, conducted data analy- ses and directly accessed and verified the underlying data reported in the manuscript. K.J.V. drafted the manuscript. All authors revised the manuscript. K.J.V., J.A.M., Z.J.W., P.M.G., D.M.B. and C.G.W. had final responsibility for the decision to submit it for publication. K.J.V. is the corresponding author and guarantor for this manuscript. All authors have reviewed this manuscript and have approved the decision to sub- mit it for publication. Acknowledgements CHAMPS would like to extend its deepest appreciation to all the fam- ilies who participated. The Network would like to acknowledge mem- bers who comprise the MITS, DSS, SBS, IT and lab teams and local communities across all seven sites. CHAMPS Bangladesh—Sanwarul Bari, Shahana Pareveen, Farzana Islam, Mohammed Kamal, A. S. M. Nawshad U. Ahmed, Mahbubul Hoque, Saria Tasnim, Ferdousi Islam, Farida Ariuman, K. Zaman, Mustafizur Rahman, Dilruba Ahmed and Meerjady S. Flora. CHAMPS Ethiopia—Joseph Oundo, Alexader M. Ibrahim, Fikremelekot Temesgen, Tadesse Gure, Addisu Alemu, Melisachew M. Yeshi, Mahlet A. Gizaw, Stian Orlien and Fentabil Getnet. CHAMPS Kenya—Emily Rogena, Florence Murila, Gutunduru Revathi, Paul K. Mitei, Magdalene Kuria, Jennifer K. Verani, Janet Agaya, Thomas Misore, Richard Omore, Solomon Sava and Aggrey Igunza. CHAMPS Mali—J. Kristie Johnson, Tatiana Keita, Adama M. Keita, Rima Koka, Karen D. Fairchild, Diakaridia Kone, Boubou Tamboura, Sharon M. Tennant, Carol L. Greene, Ashka Mehta, Diakaridia Sidibe, Doh Sanogo, Uma U. Onwuchekwa, Nana Kourouma, Seydou Sissoko, Cheick B. Traore, Kiranpreet Chawla and Jane Juma. CHAMPS Mozambique—Carla Carrilho, Fabiola Fernandes, Justina Bramugy, Celso Monjane, Sheila Nhachungue, Elisio Xerinda, Dercio Chitungo, Jaume Ordi, Juan C. Hurtado, Khatia Munguambe, Clara Menendez, Natalia Rakislova, Sara Ajanovic, Antonio Sitoe, Rita Mabunda, Arsenia Massinga, David Torres, Amicar Magaco, Marta Valente, Maria Maixenchs, Sozinho Acacio, Tacilta Nhampossa, Milton Kincardett and Rosauro Varo. CHAMPS Program Office—Betsy Dewey, Mischka Garel, Navit T. Salzber, Rebecca P. Philipsborn, Jeffrey P. Koplan, Margaret Baskett and Robert H. Lyles. CDC Central Pathology Lab—Megan Bias, Jana Ritter, Sherif Zaki, Tais Wilson and Josilene N. Seixas. CDC TaqMan Array Team—Jonas Winchell and Jakob Witherbee. CHAMPS Sierra Leone—Ima-Abasi Bassey, Sandra Lako, Julius Ojulong, Ramatu Sesay, Babatunde Duduyemi, James Bunn, Alim Swaray-Deen, Joseph Bangura, Prince Masuba, Amara Jambai, Margaret Mannah, Fatu Fona, okokon Ita, Cornell Chukwuegbo, Sulaiman Sannoh, Princewill Nwajiobi, Dickens Kowuor, Erick Kaluma, Oluseyi Balogun, Carrie J. Cain, Baindu Kosia, Solomon Samura, Samuel Pratt, Foday Sesay, Francis Moses, Tom Sesay, James S. Squire, Mohammed Sheku, Mahawa Dumbuya, Binyam Halu, Hailemariam Legesse and Sartie Kenneh. CHAMPS South Africa—Jeanie Dutoit, Fatima Solomon, Gillian Sorour, Sana Mahtab, Jeannette Wadula, Ziyaad Dangor, Karen Petersen, Martin Hale, Nelesh P. Govender, Peter J. Swart, Sanjay G. Lala, Sithembiso Velaphi, Vuyelwa Baba, Yasmin Adam, Amy Wise, Vicky Baillie, Nellie Myburgh and Cleopas Hwinya. Preva Group— Ryan Hafen and John Hathway. ISGlobal receives support from the Spanish Ministry of Science and Innovation through the ‘Centro de Excelencia Severo Ochoa 2019–2023’ Program (CEX2018-000806-S) and support from the Generalitat de Catalunya through the CERCA Program. CISM is supported by the Government of Mozambique and the Spanish Agency for International Development. Several authors are employed by the CDC. The findings and conclusions in this report are those of the author(s) and do not necessarily represent the official posi- tion of the CDC. A. S. M. Nawshad Uddin Ahmed, Mahbubul Hoque, Mohammed Kamal, Mohammad Mosiur, Ferdousi Begum, Saria Tasnim, Meerjady Sabrina Flora, Farida Arjuman, Iqbal Ansary Khan, Tahmina Shirin, Mahbubur Rahman, Sanwarul Bari, Shahana Parveen, Farzana Islam, Mohammad Zahid Hossain, Kazi Munisul Islam, Mohammad Sabbir Ahmed, K. Zaman, Mustafizur Rahman, Dilruba Ahmed, Md. Atique Iqbal Chowdhury, Muntasir Alam, Kyu Han Lee, Ferdousi Islam, Joseph O. Oundo, Fikremelekot Temesgen, Melisachew Mulatu Yeshi, Alexander M. Ibrahim, Tadesse Gure, Yunus Edris, Addisu Alemu, Dadi Marami, Ephrem Lemma, Ayantu Mekonnen, Henok Wale, Tseyon Tesfaye, Haleluya Leulseged, Tadesse Dufera, 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense 708 Paediatric and Perinatal Epidemiology, 2025 Anteneh Belachew, Fentabil Getnet, Surafel Fentaw, Yenework Acham, Stian M. S. Orlien, Mahlet Abayneh Gizaw, Emily Rogena, Florence Murila, Gunturu Revathi, Paul K. Mitei, Magdalene Kuria, Jennifer R. Verani, Aggrey Igunza, Peter Nyamthimba, Elizabeth oele, Karen D. Fairchild, Carol L. Greene, Rima Koka, Sharon M. Tennant, Ashka Mehta, J. Kristie Johnson, Adama Mamby Keita, Nana Kourouma, Uma U. Onwuchekwa, Awa Traore, Doh Sanogo, Diakaridia Sidibe, Seydou Sissoko, Diakaridia Kone, Milton Kindcardett, Khátia Munguambe, Ariel Nhacolo, Tacilta Nhampossa, Elisio Xerinda, Justina Bramugy, Celso Monjane, Sheila Nhachungue, Juan Carlos Hurtado, Maria Maixenchs, Clara Menéndez, Jaume Ordi, Natalia Rakislova, Marta Valente, Dercio Chitungo, Zara Manhique, Sibone Mocumbi, Fabiola Fernandes, Carla Carrilho, Rebecca Pass Philipsborn, Jeffrey P. Koplan, Mischka Garel, Betsy Dewey, Shailesh Nair, Navit T. Salzberg, Lucy Liu, Rebecca Alkis-Ramirez, Jana M. Ritter, Sherif R. Zaki, Joy Gary, Jonas M. Winchell, Jacob Witherbee, Jessica L. Waller, Ruby Fayorsey, Ronita Luke, Ima-Abasi Bassey, Dickens Kowuor, Foday Sesay, Baindu Kosia, Samuel Pratt, Carrie-Jo Cain, Solomon Samura, Portia Mutevedzi, Fatima Solomon, Ashleigh Fritz, Noluthando Dludlu, Constance Ntuli, Richard Chawana, Karen Petersen, Sanjay G. Lala, Sithembiso Velaphi, Yasmin Adam, Jeannette Wadula, Martin Hale, Peter J. Swart, Hennie Lombaard, Gillian Sorour. Conflicts of Interest The authors declare no conflicts of interest. Data Availability Statement Summarised data are publicly available through the CHAMPS website: https://​champ​sheal​th.​org/​data/​enrol​led-​popul​ation-​summa​ry/​. Requests for further detailed data, for research and evaluation purposes, can be made at: https://​champ​sheal​th.​org/​data/​. References 1. L. Hug, D. You, H. Blencowe, et al., “Global, Regional, and National Estimates and Trends in Stillbirths From 2000 to 2019: A Systematic Assessment,” Lancet 398, no. 10302 (2021): 772–785. 2. D. Sharrow, L. Hug, D. You, et al., “Global, Regional, and National Trends in Under-5 Mortality Between 1990 and 2019 With Scenario- Based Projections Until 2030: A Systematic Analysis by the UN Inter- Agency Group for Child Mortality Estimation,” Lancet Global Health 10, no. 2 (2022): e195–e206. 3. A. W. Taylor, D. M. Blau, Q. Bassat, et  al., “Initial Findings From a Novel Population-Based Child Mortality Surveillance Ap- proach: A Descriptive Study,” Lancet Global Health 8, no. 7 (2020): e909–e919. 4. L. Liu, S. Oza, D. Hogan, et  al., “Global, Regional, and National Causes of Child Mortality in 2000-13, With Projections to Inform Post- 2015 Priorities: An Updated Systematic Analysis,” Lancet 385, no. 9966 (2015): 430–440. 5. R. F. Breiman, D. M. Blau, P. Mutevedzi, et al., “Postmortem Inves- tigations and Identification of Multiple Causes of Child Deaths: An Analysis of Findings From the Child Health and Mortality Prevention Surveillance (CHAMPS) Network,” PLoS Medicine 18, no. 9 (2021): e1003814. 6. D. M. Blau, J. P. Caneer, R. P. Philipsborn, et al., “Overview and De- velopment of the Child Health and Mortality Prevention Surveillance Determination of Cause of Death (DeCoDe) Process and DeCoDe Diagnosis Standards,” Clinical Infectious Diseases 69, no. S4 (2019): S333–S341. 7. E. Quincer, R. Philipsborn, D. Morof, et al., “Insights on the Differen- tiation of Stillbirths and Early Neonatal Deaths: A Study From the Child Health and Mortality Prevention Surveillance (CHAMPS) Network,” PLoS One 17, no. 7 (2022): e0271662. 8. S. F. Dowell, A. Zaidi, and P. Heaton, “Why Child Health and Mor- tality Prevention Surveillance?,” Clinical Infectious Diseases 69, no. S4 (2019): S260–S261. 9. C. R. Paganelli, N. J. Goco, E. M. McClure, et al., “The Evolution of Minimally Invasive Tissue Sampling in Postmortem Examination: A Narrative Review,” Global Health Action 13, no. 1 (2020): 1792682. 10. K. Ben-Sasi, L. S. Chitty, L. S. Franck, et al., “Acceptability of a Min- imally Invasive Perinatal/Paediatric Autopsy: Healthcare Profession- als' Views and Implications for Practice,” Prenatal Diagnosis 33, no. 4 (2013): 307–312. 11. P. Byass, “Minimally Invasive Autopsy: A New Paradigm for Under- standing Global Health?,” PLoS Medicine 13, no. 11 (2016): e1002173. 12. Q. Bassat, P. Castillo, M. J. Martinez, et al., “Validity of a Minimally Invasive Autopsy Tool for Cause of Death Determination in Pediatric Deaths in Mozambique: An Observational Study,” PLoS Medicine 14, no. 6 (2017): e1002317. 13. C. Menendez, P. Castillo, M. J. Martinez, et al., “Validity of a Mini- mally Invasive Autopsy for Cause of Death Determination in Stillborn Babies and Neonates in Mozambique: An Observational Study,” PLoS Medicine 14, no. 6 (2017): e1002318. 14. N. T. Salzberg, K. Sivalogan, Q. Bassat, et  al., “Mortality Surveil- lance Methods to Identify and Characterize Deaths in Child Health and Mortality Prevention Surveillance Network Sites,” Clinical Infectious Diseases 69, no. S4 (2019): S262–S273. 15. M. Bunei, P. Muturi, F. Otiato, et al., “Factors Influencing Accep- tance of Post-Mortem Examination of Children at a Tertiary Care Hos- pital in Nairobi, Kenya,” Annals of Global Health 85, no. 1 (2019): 95. 16. S. S. Tikmani, S. Saleem, J. L. Moore, et  al., “Factors Associated With Parental Acceptance of Minimally Invasive Tissue Sampling to Identify the Causes of Stillbirth and Neonatal Death,” Clinical Infectious Diseases 73, no. S5 (2021): S422–S429. 17. M. K. Das, N. K. Arora, P. Debata, et  al., “Why Parents Agree or Disagree for Minimally Invasive Tissue Sampling (MITS) to Identify Causes of Death in Under-Five Children and Stillbirth in North India: A Qualitative Study,” BMC Pediatrics 21, no. 1 (2021): 513. 18. S. Lawrence, D. Namusanya, S. B. Mohamed, et al., “Primary Moti- vations for and Experiences With Paediatric Minimally Invasive Tissue Sampling (MITS) Participation in Malawi: A Qualitative Study,” BMJ Open 12, no. 6 (2022): e060061. 19. K. Munguambe, M. Maixenchs, R. Anselmo, et al., “Consent to Min- imally Invasive Tissue Sampling Procedures in Children in Mozam- bique: A Mixed-Methods Study,” PLoS One 16, no. 11 (2021): e0259621. 20. M. Johnson, W. A. Adewole, V. Alegana, C. E. Utazi, N. McGrath, and J. Wright, “A Scoping Review of the Methods Used to Estimate Health Facility Catchment Populations for Child Health Indicators in Sub-Saharan Africa,” Population Health Metrics 23, no. 1 (2025): 11. 21. G. Biele, K. Gustavson, N. O. Czajkowski, et al., “Bias From Self Se- lection and Loss to Follow-Up in Prospective Cohort Studies,” European Journal of Epidemiology 34, no. 10 (2019): 927–938. 22. P. Pahwa, C. Karunanayake, L. Hagel, et al., “Self-Selection Bias in an Epidemiological Study of Respiratory Health of a Rural Population,” Journal of Agromedicine 17, no. 3 (2012): 316–325. 23. K. Adazu, K. A. Lindblade, D. H. Rosen, et al., “Health and Demo- graphic Surveillance in Rural Western Kenya: A Platform for Evaluat- ing Interventions to Reduce Morbidity and Mortality From Infectious Diseases,” American Journal of Tropical Medicine and Hygiene 73, no. 6 (2005): 1151–1158. 24. A. Q. Nhacolo, D. A. Nhalungo, C. N. Sacoor, J. J. Aponte, R. Thompson, and P. Alonso, “Levels and Trends of Demographic Indices in Southern Rural Mozambique: Evidence From Demographic Surveil- lance in Manhica District,” BMC Public Health 6 (2006): 291. 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://champshealth.org/data/enrolled-population-summary/ https://champshealth.org/data/ 709Paediatric and Perinatal Epidemiology, 2025 25. F. O. Odhiambo, K. F. Laserson, M. Sewe, et  al., “Profile: The KEMRI/CDC Health and Demographic Surveillance System—Western Kenya,” International Journal of Epidemiology 41, no. 4 (2012): 977–987. 26. S. A. Cunningham, N. I. Shaikh, A. Nhacolo, et al., “Health and De- mographic Surveillance Systems Within the Child Health and Mortality Prevention Surveillance Network,” Clinical Infectious Diseases 69, no. S4 (2019): S274–S279. 27. N. Rakislova, F. Fernandes, L. Lovane, et  al., “Standardization of Minimally Invasive Tissue Sampling Specimen Collection and Pathol- ogy Training for the Child Health and Mortality Prevention Surveillance Network,” Clinical Infectious Diseases 69, no. S4 (2019): S302–S310. 28. WHO, Verbal Autopsy Standards: The 2016 WHO Verbal Autopsy Instrument (World Health Organization, 2016), https://​www.​who.​int/​ publi​catio​ns/m/​item/​verbal-​autop​sy-​stand​ards-​the-​2016-​who-​verbal-​ autop​sy-​instr​ument​. 29. M. H. Diaz, J. L. Waller, M. J. Theodore, et al., “Development and Implementation of Multiplex TaqMan Array Cards for Specimen Test- ing at Child Health and Mortality Prevention Surveillance Site Labora- tories,” Clinical Infectious Diseases 69, no. S4 (2019): S311–S321. 30. R. B. Martines, J. M. Ritter, J. Gary, et al., “Pathology and Telepathol- ogy Methods in the Child Health and Mortality Prevention Surveillance Network,” Clinical Infectious Diseases 69, no. S4 (2019): S322–S332. 31. WHO, ICD-10: International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, 2nd ed. (World Health Or- ganization, 2004), https://​iris.​who.​int/​handle/​10665/​​42980​. 32. WHO, The WHO Application of ICD-10 to Deaths During the Peri- natal Period: ICD-PM (World Health Organization, 2016), https://​www.​ who.​int/​publi​catio​ns/i/​item/​97892​41549752. 33. Global Burden of Disease 2019 Viewpoint Collaborators, “Five In- sights From the Global Burden of Disease Study 2019,” Lancet 396, no. 10258 (2020): 1135–1159. 34. R. H. Lyles, S. A. Cunningham, S. Kundu, et  al., “Extrapolating Sparse Gold Standard Cause of Death Designations to Characterize Broader Catchment Areas,” Epidemiologic Methods 9, no. 1 (2020): 20190031. 35. P. Byass, L. Hussain-Alkhateeb, L. D'Ambruoso, et  al., “An Inte- grated Approach to Processing WHO-2016 Verbal Autopsy Data: The InterVA-5 Model,” BMC Medicine 17, no. 1 (2019): 102. 36. L. Liu, K. Hill, S. Oza, et al., “Levels and Causes of Mortality Under Age Five Years,” in Reproductive, Maternal, Newborn, and Child Health: Disease Control Priorities, Third Edition (Volume 2), ed. R. E. Black, R. Laxminarayan, M. Temmerman, and N. Walker (International Bank for Reconstruction and Development/The World Bank, 2016). 37. WHO/CDC/ICBDSR, Birth Defects Surveillance: A Manual for Pro- gramme Managers (World Health Organization, 2014), https://​www.​ who.​int/​publi​catio​ns/i/​item/​97892​41548724. 38. “The Demographic and Health Surveys (DHS) Program” [Internet] (U.S. Agency for International Development [USAID], 2022), https://​ dhspr​ogram.​com/​. 39. R Core Team, R: A Language and Environment for Statistical Com- puting (R Foundation for Statistical Computing, 2021). 40. R. Hafen and J. Hathaway, “champsmortality: Calculate Mortal- ity Fractions and Rates for CHAMPS Sites,” R Package Version 0.0.1 (2022), https://​github.​com/​ki-​tools/​​champs-​morta​lity. 41. N. Rakislova, M. T. Rodrigo-Calvo, L. Marimon, et al., “Minimally Invasive Tissue Sampling Findings in 12 Patients With Coronavirus Disease 2019,” Clinical Infectious Diseases 73, no. S5 (2021): S454–S464. 42. G. Guruprasad, S. Dhaded, S. Yogesh Kumar, et al., “Lung Findings in Minimally Invasive Tissue Sampling (MITS) Examinations of Fetal and Preterm Neonatal Deaths: A Report From the PURPOSe Study,” Clinical Infectious Diseases 73, no. S5 (2021): S430–S434. 43. R. Chawana, V. Baillie, A. Izu, et  al., “Potential of Minimally In- vasive Tissue Sampling for Attributing Specific Causes of Childhood Deaths in South Africa: A Pilot, Epidemiological Study,” Clinical Infec- tious Diseases 69, no. S4 (2019): S361–S373. 44. S. A. Madhi, J. Pathirana, V. Baillie, et  al., “Unraveling Specific Causes of Neonatal Mortality Using Minimally Invasive Tissue Sam- pling: An Observational Study,” Clinical Infectious Diseases 69, no. S4 (2019): S351–S360. 45. M. A. Hernan, S. Hernandez-Diaz, and J. M. Robins, “A Structural Approach to Selection Bias,” Epidemiology 15, no. 5 (2004): 615–625. 46. “Global Burden of Disease Study 2019 (GBD 2019) Results” [Inter- net] (Institute for Health Metrics and Evaluation [IHME], 2020), https://​ vizhub.​healt​hdata.​org/​gbd-​resul​ts/​. 47. Global Burden of Disease 2019 Under-5 Mortality Collaborators, “Global, Regional, and National Progress Towards Sustainable Devel- opment Goal 3.2 for Neonatal and Child Health: All-Cause and Cause- Specific Mortality Findings From the Global Burden of Disease Study 2019,” Lancet 398, no. 10303 (2021): 870–905. 48. “Global Health Observatory (GHO)” [Internet] (World Health Orga- nization [WHO], 2016), https://​www.​who.​int/​data/​gho. 49. E. Vardell, “Global Health Observatory Data Repository,” Medical Reference Services Quarterly 39, no. 1 (2020): 67–74. 50. R. F. Terry, J. F. Salm, Jr., C. Nannei, and C. Dye, “Creating a Global Observatory for Health R&D,” Science 345, no. 6202 (2014): 1302–1304. Supporting Information Additional supporting information can be found online in the Supporting Information section. Figure S1: Flow diagram from ascertainment to CHAMPS enrolment, MITS performed and cause of death determina- tion, by site. DSS only refers to deaths captured in the DSS but never enrolled in CHAMPS; non-MITS refers to deaths enrolled in CHAMPS but for whom MITS was not performed; MITS refers to deaths enrolled in CHAMPS and for whom MITS was performed; DeCoDed refers to deaths for whom MITS was performed and reviewed by the DeCoDe panel as of 24 May 2022. All-cause age-specific mortality rates from the DHS were substituted during calculations for catchments without DSS data availability. CHAMPS, Child Health and Mortality Prevention Surveillance Network; children (1–5 years), DeCoDe, determination of cause of death; DHS, Demographic and Health Surveys Program; DSS, demographic surveillance system; infants (29–365 days); MITS, min- imally invasive tissue sampling; neonates, neonates (0–28 days); still- births (no spontaneous breathing or movement at time of delivery and [1] weighing > 1 kg and/or [2] estimated gestational age ≥ 28 weeks). Figure S2: Venn diagrams of enrolment and MITS performed among all ascertained deaths in the CHAMPS Network, by site and age. In sites with available DSS data, it is assumed all CHAMPS cases are also captured in the DSS system. (1) DSS data are ignored due to discordant availability among catchments. (2) DSS data included in count of non- MITS CHAMPS cases. (3) DSS data are not available. (4) Stillbirths (no spontaneous breathing or movement at time of delivery and [1] weigh- ing > 1 kg and/or [2] estimated gestational age ≥ 28 weeks); neonates (0–28 days); infants (29–365 days); children (1–5 years). (5) Combined for all catchments with available DSS data. CHAMPS, Child Health and Mortality Prevention Surveillance Network; DSS, demographic surveil- lance system; MITS, minimally invasive tissue sampling. Figure S3: Fractions for the most frequent perinatal and paediatric causes of death in the CHAMPS Network, 2017–2020. (A) Stillbirths, no spontaneous breathing or movement at time of delivery and (1) weighing > 1 kg and/ or (2) estimated gestational age ≥ 28 weeks. (B) Neonates (0–28 days). (C) Infants (29–365 days). (D) Children (1–5 years). (E) Under-five, includes stillbirths, neonates, infants and children. CHAMPS, Child Health and Mortality Prevention Surveillance Network. Figure S4: Fractions for the most frequent contributing maternal conditions in the CHAMPS Network, 2017–2020. (A) Stillbirths, no spontaneous breathing or move- ment at time of delivery and (1) weighing > 1 kg and/or (2) estimated 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://www.who.int/publications/m/item/verbal-autopsy-standards-the-2016-who-verbal-autopsy-instrument https://www.who.int/publications/m/item/verbal-autopsy-standards-the-2016-who-verbal-autopsy-instrument https://www.who.int/publications/m/item/verbal-autopsy-standards-the-2016-who-verbal-autopsy-instrument https://iris.who.int/handle/10665/42980 https://www.who.int/publications/i/item/9789241549752 https://www.who.int/publications/i/item/9789241549752 https://www.who.int/publications/i/item/9789241548724 https://www.who.int/publications/i/item/9789241548724 https://dhsprogram.com/ https://dhsprogram.com/ https://github.com/ki-tools/champs-mortality https://vizhub.healthdata.org/gbd-results/ https://vizhub.healthdata.org/gbd-results/ https://www.who.int/data/gho 710 Paediatric and Perinatal Epidemiology, 2025 gestational age ≥ 28 weeks. (B) Neonates (0–28 days). CHAMPS, Child Health and Mortality Prevention Surveillance Network. Appendix CHAMPS Network CHAMPS Bangladesh—A. S. M. Nawshad Uddin Ahmed and Mahbubul Hoque of Dhaka Shishu Hospital and Institute; Mohammed Kamal, Mohammad Mosiur and Ferdousi Begum of Bangabandhu Sheikh Mujib Medical University; Saria Tasnim of Dhaka Community Medical College and Hospital; Meerjady Sabrina Flora of the Directorate General of Health Services in Bangladesh; Farida Arjuman of Institute of Cancer Research and Hospital, Iqbal Ansary Khan, Tahmina Shirin and Mahbubur Rahman of Institute of Epidemiology, Disease Control and Research; Sanwarul Bari, Shahana Parveen, Farzana Islam, Mohammad Zahid Hossain, Kazi Munisul Islam, Mohammad Sabbir Ahmed, K. Zaman, Mustafizur Rahman, Dilruba Ahmed, Md. Atique Iqbal Chowdhury, Muntasir Alam of International Centre for Diarrhoeal Disease Research, Bangladesh; Kyu Han Lee of the Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health; and Ferdousi Islam of Popular Medical College and Hospital in Dhaka, Bangladesh. CHAMPS Ethiopia— Joseph O. Oundo of London School of Hygiene & Tropical Medicine; Fikremelekot Temesgen of Addis Ababa University in Addis Ababa, Ethiopia; Melisachew Mulatu Yeshi of Mekelle University, Mekele, Ethiopia; Alexander M. Ibrahim, Tadesse Gure, Yunus Edris, Addisu Alemu, Dadi Marami, Ephrem Lemma, Ayantu Mekonnen, Henok Wale, Tseyon Tesfaye, Haleluya Leulseged, Tadesse Dufera, Anteneh Belachew of the College of Health and Medical Sciences at Haramaya University; Fentabil Getnet, Surafel Fentaw, Yenework Acham of the Ethiopian Public Health Institute; Stian M. S. Orlien of the University of Hargeisa, Somaliland and Vestfold Hospital Trust, Tønsberg, Norway; and Mahlet Abayneh Gizaw of St. Paul's Hospital Millennium Medical College in Addis Ababa, Ethiopia. CHAMPS Kenya—Emily Rogena of Jomo Kenyatta University of Agriculture and Technology; Florence Murila of University of Nairobi; Gunturu Revathi of University of Aga Khan Medical College; Paul K. Mitei and Magdalene Kuria of Kisumu County Department of Health; and Jennifer R. Verani of the National Center for Immunization and Respiratory Diseases Branch at the Centers for Disease Control and Prevention; Aggrey Igunza of Kenya Medical Research Institute; Peter Nyamthimba of Kenya Medical Research Institute; Elizabeth oele of Kisumu County Health Department, Kisumu Kenya. CHAMPS Mali—Karen D. Fairchild of University of Virginia; Carol L. Greene, Rima Koka, Sharon M. Tennant, Ashka Mehta and J. Kristie Johnson of University of Maryland School of Medicine; Tatiana Keita of Clinique Pasteur in Bamako Mali; Adama Mamby Keita, Nana Kourouma, Uma U. Onwuchekwa, Awa Traore, Doh Sanogo, Diakaridia Sidibe and Seydou Sissoko of Centre pour le Développement des Vaccins; and Diakaridia Kone of CSRef Commune I in Bamako, Mali. CHAMPS Mozambique—Milton Kindcardett, Khátia Munguambe, Ariel Nhacolo, Tacilta Nhampossa, Elisio Xerinda and Justina Bramugy of Centro de Investigação em Saúde de Manhiça in Maputo; Celso Monjane and Sheila Nhachungue of Instituto Nacional de Saúde in Maputo; Juan Carlos Hurtado, Maria Maixenchs, Clara Menéndez, Jaume Ordi, Natalia Rakislova and Marta Valente of ISGlobal Hospital Clinic at Universitat de Barcelona; Dercio Chitungo and Zara Manhique of Quelimane Central Hospital; Sibone Mocumbi, Fabiola Fernandes and Carla Carrilho of Eduardo Mondlane University and Maputo Central Hospital. CHAMPS Program Office—Rebecca Pass Philipsborn of Emory University and Children's Healthcare of Atlanta; Jeffrey P. Koplan, Mischka Garel and Betsy Dewey of Emory Global Health Institute; Shailesh Nair, Navit T. Salzberg and Lucy Liu of the Public Health Informatics Institute at the Task Force for Global Health in Atlanta, Georgia, USA. CDC Central Pathology Lab—Rebecca Alkis- Ramirez of Center for Global Health at the Centers for Disease Control and Prevention; Jana M. Ritter, Sherif R. Zaki and Joy Gary of Infectious Diseases Pathology Branch, National Center for Emerging and Zoonotic Diseases at the Centers for Disease Control and Prevention. CDC TaqMan Array Team—Jonas M. Winchell, Jacob Witherbee and Jessica L. Waller of the National Center for Immunization and Respiratory Diseases at the Centers for Disease Control and Prevention. CHAMPS Sierra Leone—Ruby Fayorsey of ICAP at Columbia University and Harlem Hospital Centers; Ronita Luke of Ministry of Health and Sanitation, Sierra Leone; Ima-Abasi Bassey and Dickens Kowuor of Crown Agents; Foday Sesay of Sierra Leone's Ministry of Health & Sanitation; Baindu Kosia and Samuel Pratt of FOCUS 1000; Carrie-Jo Cain and Solomon Samura of World Hope International. CHAMPS South Africa—Portia Mutevedzi, Fatima Solomon, Ashleigh Fritz, Noluthando Dludlu, Constance Ntuli and Richard Chawana of South African Council Vaccines and Infectious Diseases Analytics Research Unit, University of Witwatersrand; Karen Petersen, Sanjay G. Lala, Sithembiso Velaphi and Yasmin Adam of Chris Hani Baragwanath Academic Hospital and University of Witwatersrand; Jeannette Wadula, Martin Hale and Peter J. Swart of National Health for Laboratory Service in South Africa; Hennie Lombaard of Rahima Moosa Mother and Child Hospital; and Gillian Sorour of Wits Health Consortium. 13653016, 2025, 8, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/ppe.70067 by U niversity O f W itw atersrand, W iley O nline L ibrary on [06/03/2026]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Major Causes of Perinatal and Paediatric Mortality in Sub-Saharan Africa and South Asia: Adjustment for Selection Bias in the CHAMPS Network ABSTRACT 1   |   Introduction 2   |   Methods 2.1   |   Mortality and Demographic Surveillance 2.2   |   Specimen and Data Collection 2.3   |   Cause of Death Determination 2.4   |   Statistical Analysis 2.5   |   Missing Data 2.6   |   Ethics Approval 3   |   Results 3.1   |   Enrolment and Consent 3.2   |   Sample Characteristics 3.3   |   Stillbirths 3.4   |   Neonates 3.5   |   Infants and Children 3.6   |   Total Under-Five 4   |   Comment 4.1   |   Principal Findings 4.2   |   Strengths of the Study 4.3   |   Limitations of the Data 4.4   |   Interpretation 5   |   Conclusions Author Contributions Acknowledgements Conflicts of Interest Data Availability Statement References Appendix