Table 1. Summary of characteristics and methods used to identify deaths, and the number of deaths identified.

Source	Persons responsible for reporting deaths	Method for death reporting (paper, phone, etc.)	Persons responsible for investigating	Method for investigating & tools	Time frame for investigation	No. of deaths reported	No. of deaths investigated
Maternal deaths
Dikid et al. [25] India	Accredited social health activists (ASHAs) & Anganwadi workers investigate community reports of suspected maternal deaths	Not specified	Government staff investigated deaths using verbal autopsy (VA)	VA Maternal death enquiry questionnaire	March to December 2009	1120 maternal deaths were identified in four states	1103 (98.5%) of the identified maternal deaths were investigated
Singh et al. [44] India	Frontline health workers routinely collect and notify maternal deaths to the primary healthcare medical officer within 24 hours	Line list of maternal deaths	Senior Public Health Officer conducted the VA	VA Government of India standardised VA forms	April to September 2012	284 maternal deaths were recorded	193 (68%) maternal deaths were reviewed
Halim et al. [37] Bangladesh	Family welfare assistants and health assistants visit households to confirm the death of any woman known to be pregnant who died	Informants completed a notification slip	Health inspectors and family inspectors conducted VA	VA Modified WHO VA tool	January 2011 to January 2012	571 maternal deaths identified in four districts	VA performed for all 571 deaths (100%)
Biswas et al. [38] Bangladesh	Family welfare assistants and health assistants visit households to confirm the death of any woman known to be pregnant who died	Health Assistants and Family Welfare Assistants completed a community death notification slip	Health Inspector, Assistant health inspector and family planning inspector conducted VA interviews	VA Modified WHO VA tool	January to March 2018	34 maternal deaths were reported	VA performed for all 34 maternal deaths
Mir et al. [26] Pakistan	Lady health workers and religious leaders collected information related to deaths in women	Women of reproductive age death listing form	MADE-FOR study team	VA Revised WHO VA questionnaire	January 2012 to December 2013	In Chikwal, 62 pregnancy-related deaths identified by lady health workers, 38 by religious leaders & 28 by both. In Talagang, 37 deaths were identified by lady health workers, 26 religious leaders & 18 by both	VA was performed for 1808 (90.3%) deaths out of the 2001 deaths identified
Naik et al. [78] India	Health workers record and notify maternal deaths routinely to the Medical Officer	Line list of maternal deaths	Senior Public Health Officer conducted the VA	VA Government of India standardised VA forms	August-November 2014	67% of deaths investigated through district meetings	VA performed for 22 maternal deaths
Moshabela et al. [40] Senegal	CHWs identified maternal deaths through active household-level surveillance of pregnancies, births & deaths	mHealth platform Childcare+	Trained fieldworker conducted the verbal & social autopsy (SA) interview	VASA MVP VASA 00tool	January 2007 to December 2012	Two maternal deaths were identified in 2007, 2 in 2008, 1 in 2009, 3 in 2010, 5 deaths in 2011 and 2 in 2012	Five maternal deaths were investigated
Bayley et al. [31] Malawi	Community team members (Health Surveillance Assistants (HSAs), group village headman & volunteers) identify & notify a maternal death in the community	Death review form	Community team	VA Community VA form	July 2011 to June 2012	52 maternal deaths were identified–of these 25 (48%) were detected by the existing notification system at the district hospital, the community team identified 43 (83%) including 27 more (52%) which were unreported at the hospital	45 maternal deaths (86.5%) were investigated
Adomako et al. [32] Ghana	Community-based surveillance volunteers identified deaths of women of reproductive age and administered RAMOS 4+2 interviews	RAMOS interview record sheets	A community health nurse performed the VA	VA Modified WHO VA form	May to August 2013	132 deaths of women of reproductive age identified through RAMOS. VA found 64 maternal deaths. Identified 13 more deaths that occurred at home which were not included in the facility records	118 deaths (89.4%) investigated using VA
Mgawadere et al [27] Malawi Mgawadere et al. [28] Malawi	Death of women aged 15–49 years were identified by: Health providers; Heads of households; Village leaders; Traditional healers; HSAs; Burial sites; Village registers; Traditional birth attendants and Police	Deaths of women of reproductive age (WRA) were reported in writing or by telephone to the research staff based at the district hospital	Trained research staff performed VA	VA WHO VA tool	1 December 2011 to 30 November 2012	424 deaths of women of reproductive age were identified, 151 were maternal deaths. Of these, only 86 had been recorded via the health management information system–all of which were facility-based deaths	All 151 deaths investigated using VA
Zaba et al. [52] Kenya, Tanzania, Uganda, Malawi, Zimbabwe, South Africa	Community-based informants notified pregnancy-related deaths at Masaka (Uganda) & Manicaland (Zimbabwe) while at all sites in Kenya, Tanzania, Malawi & South Africa, VAs were triggered by reports of deaths collected during demographic surveillance	Paper-based reporting	Not specified	VA	Between June 1989 and April 2012	235 pregnancy related deaths identified. 40 (17%) were identified as pregnancy-related by both VA and demographic surveillance data. 144 were identified as pregnancy-related based on VA reports alone (61.3%). The remaining 51 (21.7%) were identified through demographic surveillance alone	VA was performed for 184 deaths (78.3%)
Akosah & Dapaah [80] Ghana	CHWs registered household members within their community including pregnant women & children under -five years	Data was sent via text messages to a central server	VA specialists conducted in-depth VA investigations	Not specified	2010 to 2014	160 deaths of children under five reported and 1 maternal death	All deaths were investigated
Gilmartin & Levin [59] Burkina Faso	Health agents identified & notified maternal deaths	Reported via mobile phone	Health agents investigated causes of maternal deaths	Not specified	2012 to 2014	1746 maternal deaths have been reported to the national level	63.2% of maternal deaths were audited in 2014
Child deaths
Rai et al. [42] India	ASHAs identify infant deaths during routine home visits	Not specified	A health worker performed the VA, and a medical doctor administered a SA	VASA Ballabgarh VA tool INDEPTH-WHO SA tool	From 2008 to 2012 SA conducted between 1 January 2012 & 31 June 2012	514 infant deaths reported	All 514 infant deaths (100%) were investigated using VA. 91 deaths (17.7%) were investigated using SA
Shikha et al. [77] India	ASHAs, Family health workers & Anganwadi workers notified the concerned primary healthcare Medical Officer of an infant’s death	Telephonically and sends a primary informant form filled within 24 hours to the relevant Medical Officer	Medical officers of each primary health centre conducted VA	VA VA form	2012–2013	345 infant deaths	302 VA forms were available (87.5%)
Kakoty et al. [46] India	Anganwadi workers reported deaths	Meeting with Anganwadi workers	Not specified	VA VA questionnaire	1 January to 31 March 2016	Not specified	90 neonatal deaths (11.5%) were purposively selected for VA
Soofi et al. [9] Pakistan	CHW identified deaths	CHWs recorded information on the VA questionnaire	CHWs conducted VA	VA WHO/LSTMH/John Hopkins University modified VA tool 2000	August 2006 –February 2008	784 neonatal deaths identified	VA conducted for 626 neonatal deaths (79.8%)
Willcox et al. [39] Mali & Uganda	Village Health Teams reported incidents of child deaths to the study team	Deaths reported by mobile phone to a fieldworker at the sub-district level	A fieldworker performed the VASA	VASA Modified QUARITE questionnaire	August-October 2011 & 2012–2014	762 deaths of children under five years were identified in Mali and 442 in Uganda	VA done for all deaths (100%)
Nabukalu et al. [50] Uganda	Village Health Teams (CHWs) & village council chairman identified deaths in households	VA questionnaire	Trained VHT of that respective village conducted VA interview	VA WHO 2014 VA questionnaire	1st of January 2016 to 31st December 2016	230 deaths identified (53.5% were not recorded in the facility-based surveillance system), 77 among children under five years	VA conducted for all identified deaths (100%)
Hutain et al. [35] Sierra Leone	CHWs registered under-five child deaths during routine monthly household visits	Paper forms issued by the Ministry of Health	The VA Officer (clinician) located families of the deceased with the help of the reporting CHW and Community Development Officer and conducted VAs	VA WHO 2007 VA tool; WHO 2014 VA tool; Population Health Metrics Research Consortium Shortened Questionnaire	October 2015 to May 2017	CHWs reported 582 deaths of children under five, 243 of these were recorded in the vital events database	VA was conducted for 222 deaths (38.1%)
Gupta et al. [47] Rwanda	Deaths identified through facility registers, CHW reports, CHW-held community death records, & phone-based reporting system	CHW report & death records	Trained data collectors performed VA	VA 2012 WHO VA tool	1 March 2013 to 28 February 2014	618 deaths of children under five years, of which 174 were neonatal deaths	All deaths were investigated (100%)
Bogale et al. [6] Ethiopia	Community data collectors record death in the 28 days of life as part of the ongoing Health & Demographic Surveillance System	VA forms	HDSS supervisors & trained data collectors conducted SA for deaths identified through VA	SA INDEPTH Network SA tool	October 2013 to September 2017	VA identified 39 neonatal deaths	SA was conducted for 37 neonatal deaths (94.9%)
Kallander et al. [51] Mozambique	CHWs visit & record all vital events including child deaths & report in their monthly report	Monthly reports	Research assistant investigated the circumstances of the death	VASA WHO VA tool; INDEPTH network SA tool	1 January to 31 December 2015	117 deaths of children under five years	VA was conducted for 115 deaths (98.3%)
Roder-Dewan et al. [45] Rwanda	Deaths were identified through health records, Ministry of Health reporting systems & Monitoring of Vital Events using Information Technology by CHWs	CHWs reported vital events telephonically	Trained interviewers conducted interviews	VASA 2012 WHO VASA tool; Rwanda Ministry of Health Death Audit tool	March 2013 to February 2014	259 deaths of children under five	77 VAs were analysed (29.7%)
Igumbor et al. [41] South Africa	Deaths were identified by community health workers (Mentor Mothers) based on information obtained from street committee members and facilities	Community health workers captured information on smartphones (REDCap software)	Trained CHWs conducted VASA interviews	VASA VASA questionnaire	January 2017 to July 2019	19 neonates and infant deaths and 3 maternal deaths	19 deaths were investigated using VASA (86.4%)
Keenan et al. [79] Niger	Deaths were identified through house-to-house censuses	Child deaths notified through household censuses done every 6 months	VA interviews were done by Medical Officers and Field Personnel	VA 2007 WHO VA questionnaire	26 May to 17 May 2018	3615 deaths of children aged 1–59 months	3301 VAs performed (91.3%)