Study design.

Kenya was one of five countries conducting CaDMIA, working from a similar protocol to assess the feasibility and acceptability of the MIA approach. (14) We adapted the common protocol to local circumstances. Sensitization meetings were held with Community Advisory Boards, local chiefs, politicians and village reporters to inform about the study aims and objectives. After sensitization, data collection through interviews, focus groups and observation proceeded, as described below:

Data collection techniques:

• To conduct interviews with next-of-kin immediately after death, the study coordinator received notification via mobile text message or call from village reporters who learned of the death from within the community. The study team visited to request consent and conduct a brief (typically under thirty minutes) semi-structured interview.
• Additional next-of-kin in-depth interviews were also conducted with different households approximately one month after burial; visits to consent and interview were coordinated with the HDSS VA team. These interviews lasted between thirty minutes and one hour, depending on the willingness of the respondent to discuss the issues.
• Community key informants were identified through suggestions from village reporters, local persons at community meetings, and from snowball sampling from other respondents. These key informant interviews typically lasted from one to one-and-half hours.

Specialised key informants, such as facility- and community-based health care workers (medical doctors, clinical officers, nurses, lab technologists, community health workers and peer educators) and people whose work deals with death (pathologists, mortuary technicians, embalmers), were identified via suggestion from community members, community health workers and village reporters. We also made a purposive selection of staff at public and private health care facilities at various levels, based on specialist expertise, as well as workers at morgues and funeral homes serving the community. These key informant interviews typically lasted 30 minutes to longer than one hour, depending on the individual.

• Focus group discussions (FGDs) were held with several existing groups, including civic leaders and healthcare workers. Interviews as part of these groups lasted approximately 1–1 ½ hours.
• Study staff also attended events in the study community in order to observe rituals and interactions. Observations were made at home (home embalmment, night vigil, arrival of body from mortuary, or burial) or the mortuary (embalmment, viewing or removal for burial) or the church (for funeral prayers). Permission before the observation was sought from the event leader(s).
• Additionally, field notes were collected serendipitously, for example capturing comments made while waiting for a primary interview subject to be ready or when explaining the study to other people in the community. No identifying data were collected with any of these notes.

Interviews were conducted by either a single qualitative researcher or a two-person team from KEMRI, with the majority of the interviews conducted in Dholuo, the local language, or less frequently in Kiswahili or English. Interviews and FGDs followed an established guide which covered the topics of what happens around death in the community (e.g., procedures, practices and timing), perceived value and importance of knowing cause of death, and the respondent’s concerns about and perceived benefits of MIA procedures. The guides offered open-ended questions, but research staff were also able to follow lines of inquiry that were raised during in-depth interviews using probing and follow-up questions. “MIA” was explained briefly to participants as a procedure in which needles would be used to take samples from the body of a deceased, with samples examined and tested in a laboratory to help understand what may have been the person’s cause of death. Next-of-kin, specifically, were also asked whether hypothetically they would have permitted a MIA to have been performed on the recently deceased, why or why not, and under what circumstances they might have agreed. Observations were conducted by the qualitative research team witnessing body embalmment, night vigils, funerals, religious services and VA home visits.

Desire to know CoD.

Respondents described several situations in which there is an existing desire to know the CoD. This was discussed especially in cases of child and maternal death as well as in sudden death, when it was described that the family had a heightened interest to understand the cause. Respondents mentioned that when death occurs at a health facility there may be concerns and suspicions about treatment received, and therefore a desire to know the CoD to assess whether there was fault on the part of health care providers. Another example of the desire to know the CoD was when a death was suspected to be due to witchcraft or chiraa (contravening of tradition) and therefore considered to be “unnatural.” Death due to witchcraft has implications for the family and for the memory of the deceased and therefore, can be a cause of consternation. Community members are eager to hear the medical CoD, creating an incentive for the family to have this information.

"It is good in cases where there is a conflict or maybe someone thinks there was something [e.g., witchcraft] was done or you feel this is actually what happens … then it sort of resolves the conflict"

{Health care provider/Nurse, female}

“In my opinion I would have agreed for it [MIA] to be done, because most of the time when someone dies according to our belief someone had a hand in killing him or her. But instead the person died of sickness, so when MIA is done it will establish the cause of death, and this will be good. In our community someone doesn’t just die without someone being pin-pointed as the killer. So I support MIA.”

{Next-of-kin, Elderly male}

Parents spoke of a desire to protect their surviving children and other family members by having more knowledge about the cause of the child’s death, and possibly being able to respond better if such illness were to occur in the future to a sibling or other child. Additionally, it was explained that mothers may be blamed for the death of a child, either for poor parenting or by causation an “unnatural” death due to chiraa or witchcraft (when someone else may cause the death of the child due to ill wishes). Some female respondents saw the possibility that MIA could possibly assist a mother who wants to clear suspicion that she caused the death, through either poor caretaking or due to chiraa. It was also expressed that a mother’s own desire to know if she could have prevented the death would be a factor for acceptance.

“…like babies, whereby they just die, this baby could not say what he or she was feeling and the mother is asking herself whether she made the mistake or who made the mistake. Was it the health care worker? Or what was not done? I think such questions will be answered.”

{Community leader, male}

“He was examined and nothing could be found ailing the baby, [he] could not be diagnosed, so I do not know very well what caused his death. The grandmother was saying maybe it is chiraa, that is why [we] used herbal medicines to treat the child too [long]…”

{Next-of-kin/mother of deceased child}

We interviewed family members in ten cases of maternal death. Respondents spoke of the fact that in cases of maternal death the family may believe that the death occurred due to medical negligence or due to witchcraft, for example if it is believed that someone in the community or a co-wife wished ill to the mother. It was seen that MIA could provide some clarity in these situations, providing the husband and children with answers regarding the mother’s death. It was also mentioned by several participants that the idea of postmortem procedures (more invasive than MIA) was already familiar in cases of maternal death as per traditional burial requirements it was necessary to separate a mother’s and a fetus’s body:

“Yes [people can agree to MIA in maternal death cases] because those are simple procedures, you don’t open up the body you only injure… because even within the villages when … an expectant mother passes on, there is a belief that the mother is not buried with the fetus, so what normally happens there is someone who is brought to remove the fetus from the mothers body and that is done at home so why not just injecting and taking some specimen."

{Male key informant/ local public health researcher}

Knowing CoD offers future benefit to the community.

A small number of respondents, primarily health care providers and health administrators, described their belief that learning more about CoD could assist in improving medical care and public health for the community as a whole.

“They will see the benefit, because if they help someone who is still alive who was suffering from that illness, you must see the benefit.”

{Traditional birth attendant, female}

“Knowing cause of death has one benefit, because it will make me take precaution because I’m still alive, it will make me take precaution maybe from what was happening that caused the death”

{Male verbal autopsy interviewer, FGD participant}

“…it is of most benefit because sometimes we are unable to pin point what actually led to the death of someone, some patients die before we are able to make a definite diagnosis, we are working with a provisional diagnosis, we are yet to do some investigations, we are unable to do some investigations and in those cases, we can’t completely say we know what caused the death because of limitation of tests and limitations of specialization amongst us”

{Healthcare worker/clinician, male}

“It is beneficial to the health professionals; it can help in the continuity of management of patient. May be the cause was hidden, so in future if you notice the same symptoms on a patient it can help you the direction to take”

{Healthcare worker/nurse, female}

Some respondents mentioned that the results of MIA should be put to immediate use, including possibly providing treatment to family of the deceased if a specific risk was identified during MIA.

“Maybe support… it can be in terms of medicine to protect the remaining individuals, so maybe the death of so-and-so made the community to be given such drug to protect the community.”

{Next-of-kin, male}

Serving the immediate needs of the family.

There is a history in the community of participants receiving small financial or in-kind incentives for participation in health studies, though the HDSS does not offer an incentive for participation in census activities. Many CaDMIA respondents indicated that they were aware of the previous study in the region which had conducted CDA, and which offered the family 5000 Kenyan shillings (approximately 50 USD) to assist with the purchase of a coffin for the deceased. Respondents saw agreeing to participate in MIA as offering the potential to receive some type of incentive, which would help satisfy the pressing financial and logistical needs which arise at the time of death. Furthermore, there is a tradition throughout much of Kenya where friends and relatives of the family of the deceased make contribution toward funeral expenses. Referred to as “harambee” this tradition is used in a variety of other situations of urgent need as well, such as medical expenses and school fees. It was explained that Luo families face acute financial needs upon death, especially with the death of a senior family member, due to community expectations of an elaborate funeral to honor the deceased. In our study communities, funerals for adults typically occur on the weekend following death, or later if critical family members are still awaited for funeral events or some matters related to the deceased had yet to be settled. It was explained that this latter situation may arise, in particular, in the case of agreeing on how to settle the estate of a man with more than one wife.

Many respondents described assistance with coffin purchase or other financial assistance by the program conducting MIA as an important motivation for agreeing to participate. In addition, respondents explained that if the MIA process keeps a body of an adult in the mortuary for several days and provides embalming after the MIA procedure, this was seen as an added value of participation, as these actions satisfy existing needs (waiting for the funeral) and reduce the financial burden on the family.

“You know, nowadays, should someone face a problem and then receive assistance, the community will accept the study because people depend on assistance, and this is because of the challenges that people go through and the poor economic status.”

{Next-of-kin, male}

“…if you just go with nothing for the bereaved family, there are people who can say no to MIA.”

{Key informant, male}

Minimal interference with the body.

After learning about how MIA is conducted, respondents expressed that MIA would be accepted due the minimal effect on the body which they perceived as giving respect to the deceased. For some respondents MIA was contrasted to what they understood or had heard of previous complete autopsies conducted through another study—however few had personal experience of any postmortem procedure.

“I would go for MIA due to reasons one, there is that maximum respect to the body, maybe you have seen how postmortem is done, the process of opening that body, cutting it is an inhumane way, I think MIA is more efficient than the other one, because you can cut and chop most precious part of the body, but for MIA that small portion is enough for the entire process, just a tissue, that is enough for the [minimally invasive] autopsy.”

{Pharmacist, male}

Limited value or use to know the CoD.

Some respondents expressed the perception that learning the CoD would not serve a purpose for the family and therefore MIA did not serve a role. One reason given was the idea that death was “in the hands of God” and not under the control of humans, therefore making post-mortem investigations not useful. Others expressed their concern that MIA would not benefit the deceased and therefore did not serve a purpose.

“For one, death is irreversible, once death has occurred, would that postmortem bring back my loved one? But when one is gone, in my case he or she is already gone.”

{Healthcare worker/nurse, female}

“It is very rare [to learn the cause of death] because it has never happened with us to have the cause of death because we say we are all from God and we shall all return to God, we do not go deep to know what really killed the person, we leave it to God. He brings and takes away and no questioning of that.”

{Muslim community leader, male}

Complications to learning the CoD.

Concerns about learning about an HIV diagnosis as part of learning the CoD were mentioned frequently, bearing in mind that Siaya County has one of the highest rates of HIV infection in Kenya. Some respondents mentioned an emphasis on and concern about confidentiality of the results. It was described that while it is well known that there is HIV in the Luo community, there was nonetheless concern about having others in the community learn about HIV in one’s family.

“… you know we still love our dead, and in a case where we have known, like in my case I know my husband has died of HIV, and now you are coming up with a tool that is going to determine and everybody is going to know what killed my husband, so actually it will be received with mixed feelings”

{Next-of-kin and healthcare worker, female)

In the Luo community there is a tradition of widow-inheritance {Geissler, 2010 #34}, in which a brother or other male relative of the deceased may take on responsibility for a widow, one aspect of which is that they will engage in sexual relations. While somewhat diminishing, this practice is still in place especially in rural areas in part because women may have limited options for employment. Several respondents mentioned that findings about the husband’s CoD, particularly in the case of HIV, could have implications for his surviving widow. It was explained that she can be stigmatized or find it difficult to get an “inheritor,” and therefore may have difficulty to support herself and her children.

“… now there is this issue of inheritance, so somebody would like to know, if the deceased was a man and somebody would like to inherit the wife, they would really want to know the cause of death”

{Key Informant, male}

“There is stigma in the community … Because of stigma, the community would not want to know the cause of death, including the immediate family because if people know so-and-so died of HIV the widow might not be inherited which is a practice that they still do. No one will inherit her because she is [HIV] positive.”

{Key informant, male}.

Prioritization of resources.

Multiple respondents expressed concern that MIA would utilize resources which they felt would be better spent on “the living.” Respondents were concerned that the use of transport, medical equipment and trained personnel should be devoted to saving lives rather than assessing the CoD for those who were already deceased. Community respondents expressed their experience that autopsies are costly and inaccessible, and concerns about the costs of conducting post-mortem studies (including MIA) were also raised by community leaders and health officials. Some respondents shared their opinion that conducting MIA would only be of benefit to researchers, rather than to families themselves.

“That is what I’m trying to say… please try to find the illness when someone is still alive, stop looking for this thing [cause of death] when someone is dead. Please help people when they are still alive, my people.”

{Traditional Birth Attendant, female}

“She [the deceased] was bitten by a snake and CDC [CDC and KEMRI researchers working in the HDSS] did nothing about it. She was buying drugs for herself and sometimes she did not afford the drugs. I’m just wondering why the CDC could not help while she was still alive, until she is now dead is when the CDC comes to ask questions about her death.”

(Next-of-kin, male)

Negative impact on the family.

In some cases, respondents shared concerns that agreeing to MIA could contribute to conflict within the family, either due to agreeing to the procedures at all or because acceptance of money from the MIA team (e.g., funds toward a coffin) could be viewed as “selling the body”. It was also suggested that there could be resentment or jealousy toward the person(s) who would receive any funds that were provided.

“Even in my case there were concerns. It is my firm nature which helped me out when people were now asking me how dare I give my mother’s body to the CDC [researchers] who are going to take some body parts to sell and only pleasing me with 5000 shillings. I became so wild and told them the CDC did not take my mother’s body because I did not have money”.

{Next-of-kin, male, referring to participation in a previous postmortem study in the region}

Another setting in which participation in MIA was seen as problematic for the family was in the case of death of a child or a stillbirth. Respondents described that burial procedures for children in Luo communities vary significantly from adults; typically the child will be buried by sundown on the day of death or the following day. Generally, only immediate family members attend the burial, as a child is not yet regarded as a full member of the community. For this reason, there was some concern that MIA procedures could interfere with the normal timing of burial which some respondents saw as a barrier. It was also acknowledged the logistics of reaching the family in time to get consent for MIA when death had occurred at home may be more difficult in the case of children as compared to adults. This was particularly noted in the case of a stillborn child, as the deceased child would be taken home and buried very rapidly. It was described that this was seen as a means to spare the mother the emotional toll of having the deceased infant present.

“I think it is not easy [for MIA] to be carried [out] on stillbirth. This is because as Luos stillbirths are buried immediately….Unless it happens in the hospital. Suppose my daughter had a stillbirth in the morning, by around this time [1100hrs], the body [of the child] would have been buried.”

{Key Informant, female}

It was infrequent that respondents spoke about a specific concern that participating in MIA would increase the emotional toll of experiencing a death, however one respondent expressed how adding an additional procedure, especially one done in the morgue, would have affected him:

“I did not want my child to first go to the mortuary. Secondly, to stay long so that the body will turn, I wanted … my daughter to maintain her freshness. I did not want anybody to touch her [at the mortuary].”

{Next-of-kin, father of a deceased child}

Also with regard to the death of a child, another respondent indicated that it would be difficult to find him in a state of mind to speak with anyone about allowing such a procedure:

“I don’t know, but I don’t think if you came to ask me to do such a procedure on my child in about two or three hours [after death], I would not have anything to do with you.”

{Next-of-kin, male}

Some respondents reported that MIA would not be accepted because they saw it as a form of body mutilation and they felt that there were cultural rules against this. This was, however, a theme that arose in relatively few settings.

“That procedure is actually very much condemned, because if somebody goes on puncturing the body of the deceased it is not actually good, if the patient was still in the ward the body can be punctured to find out the illness, but somebody just putting a knife on the body, that one no! It is condemned.”

{Surgical theatre technician, male}

Religious concerns among Muslim respondents.

Muslims are a relatively small proportion of the population in western Kenya. We had the opportunity to speak with two Muslim respondents—a community leader and a clinical officer, who expressed some additional concerns with the conduct of MIA. They explained that taking the body of a deceased Muslim to the morgue could not be permitted, and there are specific times prescribed for burial, therefore introducing concerns with potential delay in burial. Both nonetheless indicated the possibility of allowing MIA through involvement of Muslim organization leaders and if Muslims’ concerns were addressed. It was emphasized that leaders in the community would need to approve of any procedures verses only getting individual approval.

“We look at the Islam perspective; after looking the Islam perspective then after that there is a discussion and then they will come to resolve, okay, let us try it. But the consent has to depend on them [community leaders].”

{Muslim community leader, male}

“We have Muslim organizations who give directions to the Muslims, so even in those organizations we have officials, …, so through these organizations where their leaders are you can talk to them then they can tell you yes or no, because if you go down to the community level, that will be at an individual level.”

{Clinical officer who is Muslim, male}