Table 2 Common challenges faced in MPDSR implementation identified across humanitarian settings

Acute crises

Reporting anecdotal during acute disaster response

Decline in reporting after onset of insecurity

Reported mortality rates likely to be large under-estimates, and paradoxically reduce while true mortality rates rise

Remote reporting via SMS may be more resilient

Often very limited reporting of community deaths

Low skilled birth attendance rates early after displacement

Very challenging to conduct formal death reviews during acute response period

Death reviews were feasible in IDP camps a few months after acute conflict/displacement

External facilitation (mentorship visits) to support reviews was useful but often interrupted by insecurity

Experienced staff trained to conduct reviews often leave region and system may collapse after insecurity

Indirectly/informally highlighting learning points from recent deaths during trainings may be more feasible than conducting formal mortality reviews

Protracted crises

Insecurity and access limit community reporting

Stillbirth and neonatal death misclassification common in community death reporting

Community surveillance often does not include maternal or perinatal deaths

Simplified definitions in community surveillance may miss indirect, early pregnancy and postpartum deaths in particular

Deaths ‘in transit’ between facilities often not reported by either referring or receiving facility

Accurate cause of death determination for maternal and perinatal deaths identified via community-surveillance was challenging

Verbal autopsy may not be possible or prioritized

Facility-based review of maternal deaths revealed unexpectedly high proportion of deaths due to unsafe abortions

Defensive approach, reviews evolving into HR processes, linked to disciplinary procedures

Even where maternal death review well established (with no-blame culture), perinatal death reviews have had limited implementation, and limited engagement of some staff and challenges with ICD-PM coding

Security of health workers may be at risk if blamed for deaths, particularly when confidentiality and anonymity is challenging

Community tensions may increase sensitivity of death reviews

Disease epidemics

Significant underreporting of maternal deaths

Those with potential symptoms of the epidemic disease prioritized above other causes

Non-epidemic health issues often deprioritized

Mistrust and suspicion of health services reduces reporting

Linking maternal death reporting with integrated disease surveillance and response (IDSR) may improve reporting

Short term funding during epidemics may improve mortality surveillance, but this may not be sustained without predictable investment in health system strengthening

As in other crises formal death reviews impacted by crisis

Focus on disease epidemic means reviewing other causes of death requires explicit focus and investment

Funding should continue to support reporting and review of non-epidemic related deaths

Refugee camps or camp-like settings

Fear of loss of household rations if death reported

Deaths often not captured if woman referred outside refugee camp for care

Refugee mortality often underestimated, and national statistics may exclude or not differentiate refugee deaths

Accountability and focus on improving care within camp may be reduced if mortality statistics are exported outside the camp

Lack of experienced staff with capacity and authority to do mortality reviews (particularly with prescriptive guidelines on composition of review committees)

High staff turnover

Deaths often occur outside camps in higher facilities often not reviewed by health services within camps, and lessons learned often not fed back to referring facilities

Any setting

Mortality statistics unlikely to represent true mortality rates (underreporting of facility and community deaths)

Concerns about negative consequences discourages reporting

Concerns about reputational damage if facility deaths reported

Data collection may raise suspicion; medical records destroyed and false names used due to fear of data use by military/government actors

Poor engagement/punitive approach towards TBAs discourages reporting of community deaths

Feasibility of review limited by data available and capacity/interests of staff involved

Blame culture limits quality of reviews

Limited attention to or documentation of response and follow-up

Large numbers of perinatal death reviews difficult to manage

Large committees with prolonged meetings are resource intensive, particularly in areas with strained human resources

Sustainability of reviews without external donor support is questionable in some settings

Limitations of clinical documentation limit value of reviews

Reviews conducted by an external agencies may miss opportunity for local teams to be involved in defining solutions

Some recommendations useful, others had tendency to be generic and non-specific