Notwithstanding recent improvements in diagnostic technologies, full-body autopsies remain the gold standard in determining cause of death. This is because discordancy rates between clinical and autopsy diagnoses for communicable and non-communicable conditions range from 17% to 65%, and up to 40% of post-mortem (PM) examinations reveal significant information regarding cause of death that was not previously known [1234].

Clinical autopsies are not frequently performed as part of routine care [45]. Furthermore, autopsy rates across the world have declined, and rates are especially low in sub-Saharan Africa, where low perceived benefit, fears of body disfigurement, the practicalities of funeral arrangements, and cultural beliefs lead to autopsy refusals by next of kin [4567]. Aside from refusals to consent by next of kin, the cost and expertise of carrying out a full clinical autopsy is an impediment to its use, especially in Africa [89]. There are few trained pathologists in the African region, leading to insufficient human resource capacity to cover autopsy workloads [8]. In these settings, minimally invasive autopsies (MIA), which involve less invasive PM examination procedures such as minimally invasive tissue sampling (MITS) techniques, have been shown to provide a cause of death determination in four out of five cases [10]. However, they are more sensitive in accurately determining cause of death in infectious conditions as compared to non-infectious conditions [1011]. MIAs do not require removal of organs or major incision marks on the body. As it is a less intensive procedure, the turnaround time for results can be faster and it does not lead to major delays in burial, which would be particularly important in some communities where these constitute reasons for refusal [1213]. In Kenya, PM examinations are not frequently performed [14], although with adequate community mobilization, autopsy acceptance rates of 80% have been observed [15].

We set out to identify local factors associated with consent and refusal of pediatric autopsies at a tertiary teaching and referral hospital in Nairobi, Kenya, and examined local preferences for minimally invasive tissue sampling (MITS) techniques over the conventional autopsy. Understanding the reasons for consent and refusal, as well as preferences for MITS, could assist health authorities in developing strategies to increase local autopsy rates in order to obtain more accurate data on the causes of mortality and thereby guide health policies.

In this study, we interviewed family members who had received bereavement counseling following the death of their child and were asked for permission to conduct a PM examination to identify their child’s cause of death. We included those who consented to the PRESS study and those who did not. Regardless of their participation in the previous PM study, a majority of family members in both groups thought that PM examination could help identify the cause of death. Despite the small sample size, 81% of parents who previously refused PM examination of their child would most likely have accepted it if MITS was used instead of conventional autopsy. MITS could be an alternative approach to PM examination to determine infectious causes of death when cultural driven factors regarding timing of burial or cutting of the body preclude the use of conventional autopsy.

In a social behavioral study conducted in five low- and middle-income countries in Asia and Africa, the hypothetical acceptance of MIA (which includes MITS techniques) on deceased relatives was >70% [14]. In Bangladesh, a predominantly Muslim country, MITS techniques were deemed more acceptable than conventional autopsies because they did not require major delays in burial, removal of organs, or cutting or stitching of the body [12]. In a Belgian study, researchers were also able to demonstrate that Muslim parents preferred MITS to the conventional autopsy [17]. In our study setting, where the population was mainly Christian and the main reasons for decline did not include concerns of opening up the body, MITS techniques would likely not have as much as an effect on improving overall PM examination acceptance rates (previous studies have described conventional autopsy acceptance rates as high as 80% in rural Kenya [15]). Nevertheless, acceptance of MITS is encouraging especially in resource-constrained set-ups like ours, where routine standard autopsy is not always available.

In Nigeria, younger parents are less likely to consent to PM examination of their children than older parents [7], while non-Muslims and those with higher levels of education are more likely to consent [9]. In Belgium, maternal age, parity and education has no association with autopsy acceptance rates [17]. In our study, we were able to demonstrate that unemployed people, those with less education and those with less knowledge about autopsies were least likely to consent to autopsies. This suggests that the effects of socio-demographic factors on autopsy acceptance rates are largely context specific, and generalizations from studies conducted in different settings should be avoided.

Similar to our findings, one of the most common reasons given by family members for consenting to autopsy studies is to know the cause of death [914]. Aside from finding out why their child had died, parents are also motivated to consent to autopsies to prevent similar illnesses in their other children and by a strong sense of altruism [181920]. Parents hope to advance medical knowledge in order to help other families avoid the suffering associated with the death of a child. Clinicians could make use of these motivators to increase autopsy rates.

Studies have shown that PM examinations are often not conducted because the deceased’s family members are not asked to consider this option [2122]. In our study, none of the participating families were advised to undertake a PM examination by their attending doctor. Simple measures such as involvement of the attending physician in discussions with relatives about PM examination have the potential to significantly increase autopsy rates, as documented previously [23].

The majority of those who declined PM examination of their child did so because they felt there was no need for it (they had already lost their child) or because they were satisfied with the clinical diagnosis received from the attending clinician. Similarly, in a study conducted in Scotland, parents felt that there was no need for PM examination given that their child had died [18]. In Zambia, most parents declined PM examination because they did not see the benefit of it [7]. This seems to be a common theme: pragmatism surrounding death. Nonetheless, PM studies are important to identify potentially preventable deaths and advise health policies and medical staff on improvements in patient management [424]. Health care providers’ perceptions of the difficulties in obtaining consent from relatives, the administrative challenges associated with autopsies, perceived poor quality of autopsies conducted, delays in obtaining autopsy reports, fears of revealing clinical mismanagement of patients and the threat of malpractice litigation are some of the reasons health care providers are hesitant to talk to family members about PM examinations [914]. Improving autopsy rates would therefore require interventions targeted at concerns raised by the clinician as well as the deceased’s family members.

Our study had some limitations. Although we attempted to contact all the families previously approached for the PRESS study, most of our respondents were families who accepted autopsy investigation of their deceased child. We had 94% of consenters and 45% of non-consenters from PRESS participating in this survey. Those not interviewed could have different perspectives about PM examination (e.g., not inclined to agree to PM examination even if MITS procedures are used) or different demographics. The responses elicited from the phone interviews versus face-to-face interviews could also have been different. Moreover, our findings need to be interpreted within the context of the study setting. The study was based at a tertiary care referral hospital; most of the families would have sought care in other health care facilities before arriving in the study hospital and may have different attitudes and expectations towards understanding the cause of death of their child when compared to the general population and cases of death in the community. Even so, the study offers an insight into this population and may guide future initiatives to investigate causes of death in similar settings.

A desire to know the cause of death and an altruistic attitude among the deceased’s family members are great motivators in consenting for autopsy. There is potential to increase autopsy rates by strengthening reasons for acceptance and addressing modifiable reasons for refusals. Although MITS techniques have the potential to improve autopsy rates, they may not be significantly preferred over conventional autopsy in settings where opening up the body and removal of organs are not one of the major reasons for refusal by the deceased’s next of kin.

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.