Study design

We undertook a pilot study to explore the feasibility and acceptability of teleVAs.27 Feasibility refers to the practicality of implementing an intervention, including the extent to which its components and activities can be carried out effectively, considering factors such as the delivery methods used. Acceptability, on the other hand, relates to how suitable and appropriate the intervention is perceived to be by both patients and healthcare professionals, as well as its alignment with daily life practices.28 Over a 2-month period (September 2021 and October 2021), we conducted teleVAs with NoK contacts of deceased individuals, recording data via interviewers’ call log sheets. We also conducted surveys to assess the quality, acceptability and feasibility of each teleVA, followed by in-depth interviews (IDIs) with both respondents and interviewers after the pilot to gain further insights into their perceptions of conducting teleVAs.

This pilot was embedded within the larger Strengthening Mortality Surveillance (SMS) study, which was developed to respond to a gap in the district health information system regarding access to timeous and accurate CoD data during the COVID-19 epidemic. There were several subprojects within the SMS which included a rapid assessment of CoD data collection and public health use in South Africa, the teleVA pilot and a scoping study of online capture of the medical certificate of CoD in South Africa.29

Study setting

The teleVA pilot was conducted in the Cape Town metro which falls under the Western Cape Provincial and the City of Cape Town Health Departments. Furthermore, two nodes of the South African Population Research Infrastructure Network (SAPRIN) HDSS (Africa Health Research Institute (AHRI) and Agincourt)) were included in the qualitative part of the study.22 The AHRI node is located in rural KwaZulu-Natal and Agincourt is located in the rural province of Mpumalanga. As a result of the COVID epidemic, both nodes had shifted their routine in-person VA activities to be conducted over the phone starting in 2021. This enabled us to explore the opportunities and barriers for teleVA in both metropolitan and more rural settings in South Africa.

Intervention: teleVA pilot

Six interviewers experienced in conducting face-to-face VAs were recruited for this teleVA pilot study. Interviewers were provided with 5 days of refresher VA training as well as training on conducting teleVAs. The teleVAs were conducted over a 7-week period between August 2021 and October 2021. The V.1.5.3 of the electronic (Open Data Kit) 2016 WHO VA instrument was used, incorporating the COVID-19 package.30 A data system for electronic data capture and storage was set up using KoboToolbox which was installed on the Western Cape Provincial Health Data Centre (WCPHDC) server.31

TeleVA respondents (NoK) were eligible if they were aged 18 years or older and if their deceased family member was (1) aged 18 years or older, (2) their death occurred outside a health facility and (3) the CoD was not due to an unnatural cause (eg, an accident or suicide).

We used three different methods to identify deaths that occurred outside a health facility with contact details for NoK for study recruitment and included:

1. Out-of-facility deaths were retrospectively identified through linkage of known COVID-19, tuberculosis, HIV and diabetes cases from the WCPHDC with the National Population Register data held by the South African Medical Research Council (SAMRC). These linkages were restricted to cases with South African identity numbers. Contact information for relatives was sought from other available data sets in the data centre and those with information for relatives were selected (n=3977).

2. Emergency medical services (EMS) identified deaths outside a health facility to which they attended during August 2021 and obtained contact details for the person who called EMS at the time of the death (n=358).

3. Funeral parlours (FPs) were approached to assist with identifying out-of-facility deaths and obtaining NoK contact details and their permission for project staff to make contact to arrange an interview (n=7).

During the period from December 2020 to September 2021, a total of 4342 deaths outside a health facility were identified included contact details for NoK. A convenience non-probability sample of 600 deaths was targeted. We assumed a refusal rate of around 60%, slightly lower than expected for a telephone survey (70–80%) as we felt that respondents might be more likely to respond positively to an interview request coming from the Western Cape Health Department.32 Thus, 1571 recent deaths that occurred outside a health facility were selected, ensuring at least a 1-month gap since death had transpired before an interview was attempted (more detail in figure 1). Since the contact information for home deaths identified by the WCPHDC took some time to obtain and the fieldwork time was limited, we included cases as the information became available.

A call centre was established in August 2021 at the Western Cape Department of Health with a team of six trained interviewers. These interviewers used the KoBoToolbox app to capture data while contacting NoK via phone to obtain permission for teleVA interviews and identify the most suitable respondent. Once the NoK agreed to the interview, the interview proceeded immediately, or an appointment was set up. Verbal consent was obtained over the phone and recorded using an Android device. Initially, a maximum of 10 attempts were made to contact each respondent before a final outcome (successfully interviewed, refusal to be interviewed, unable to be contacted) was declared. After the first month, due to fieldwork time constraints, the maximum number of attempted calls was reduced to three per respondent.

NoK were reimbursed with a R100 (∼US$5.4) e-voucher after successful completion of the teleVA and only Western Cape respondents who completed an IDI received a R50 e-voucher. The e-voucher was designed to compensate for the respondent’s time and the inconvenience of participating in the teleVA interview. The vouchers were disbursed through the Flash mobile application (https://flash.co.za/).

A call log was maintained by the teleVA interviewers. Data from the call log described the date, time and outcomes for each individual call, the total number of calls and the final outcome for each potential participant.

Feasibility and acceptability evaluation

Feasibility was rigorously assessed by analysing response rates and the quality of the entire VA process was assessed by comparing valid responses (‘yes’ or ‘no’) from the teleVAs with face-to-face VAs from the NCoDV Study.33 Concurrently, acceptability was evaluated through a comprehensive post-teleVA survey and IDIs conducted with both the respondents and interviewers. This dual approach aimed to shed light on the viability and receptiveness of conducting teleVAs within this context.

Post-teleVA survey

After each teleVA, the teleVA interviewers filled in a survey to further assess the feasibility of teleVA. The survey captured the following information: (1) the number of calls made to the NoK before successfully conducting an interview, (2) the occurrence of any technical challenges, (3) sound quality, (4) if any interruptions occurred during the call, (5) the level of comfort of the interviewer, (6) the respondent’s attitude as perceived by the interviewer and (7) the overall satisfaction of the interviewer during the call. This survey had been previously implemented in Tanzania.26 As this was a pilot study, the level of missing data was documented but no imputation was undertaken.

Quality of the VA responses

The root (entry level) questions in the VA are closed-ended allowing the selection of one of the following options ‘yes’, ‘no’, ‘don’t know’ or ‘missing’. Questions with responses ‘yes’ or ‘no’ were defined as informative responses, with ‘don’t know’ and ‘missing’ defined as non-informative responses. To assess the quality of the teleVA responses, we compared the proportions of non-informative responses, for each root (entry level) VA question with those for face-to-face VAs conducted during a previous study (NCoDV2017/18).25 The difference in the proportions of non-informative responses between teleVA and face-to-face VA was tested for each of the VA root questions using a χ² test.

Individual in-depth interviews

We conducted semistructured interviews with teleVA respondents as well as interviewers to understand the feasibility and acceptability of teleVA. We used the VA call log to purposively recruit individuals from a diverse group of teleVA respondents for the qualitative interviews. We also recruited VA interviewers from the Western Cape pilot, as well as from the SAPRIN nodes, in order to capture experiences in a different setting and from VA interviewers who had experience in VA and teleVA.

All participants in the study gave informed verbal consent after being sent a written study information sheet via WhatsApp and a verbal explanation of the consenting process at the beginning of the interview. Verbal informed consent was recorded during the phone interviews.

All interviews with Western Cape respondents and interviewers were conducted by CSF (first author). The interview was conducted in English, except for two that were conducted in Xhosa and Afrikaans, respectively, with the support of a translator. In the SAPRIN nodes, SN conducted three interviews in Shangaan (Agincourt) and three were conducted by XN in Zulu (AHRI). The interviews were conducted based on a question guide that had been developed and tested for the study (see the full guide as online supplemental materials). Semistructured interviews were used to contextualise the findings of the survey. Probes were used to elicit additional information based on the course of conversations and the interviewer’s observations.

All interviews were conducted by phone, using MS Teams and audio-recorded, after the interviewee’s consent. Only the interviews conducted by SN and XN were transcribed and translated into English to allow CSF, who led the analysis, to immerse in the data.

The IDIs were analysed using rapid qualitative methods.34 The information collected during the interviews was compiled and summarised using a rapid assessment procedure (RAP) sheet by the researcher conducting the interview. The RAP sheet was created using the domains of acceptability, that is, comfort and satisfaction and feasibility, that is, logistics, technical challenges and sound quality which allowed the findings to be categorised using a deductive approach.35 This allowed the researchers to immerse in the data and permitted early identification of findings as data were being collected. Thereafter, inductive coding was used to allow for emerging themes to be identified. Representative quotes were extracted from the audio recordings and categorised in the RAP sheet.

Patient and public involvement

This pilot study actively involved the public by assessing the feasibility and acceptability of teleVAs through next-of-kin in-depth interviews. Their feedback shaped the understanding of teleVA implementation, participant burden and alignment with daily practices, ensuring future developments are guided by public experience.

Table 1

Success rate for eligible respondents contacted by source. Strengthening Mortality Surveillance (SMS) for improved monitoring of HIV/AIDS, TB and COVID-19 in South Africa, August–September 2021

Supplemental material

Theme 1: interviewers needed time and training to conduct VA over the phone instead of in-person

In the SAPRIN nodes and the Western Cape, interviewers had prior experience conducting face-to-face VA interviews before the teleVA pilot. However, transitioning to the telephonic format posed a challenge for them. One difficulty they encountered was discussing sensitive topics covered in a VA over the phone. According to several interviewers, expressing condolences for the loss of a family member seemed more effective during in-person interviews but this proved to be more difficult during phone interviews. Interviewers frequently encountered difficulties in finding suitable methods to provide comfort to respondents experiencing distress during the phone interviews.

It was a bit strange at first, because the questions we ask are very personal. It was very difficult to understand the reactions of the respondents, especially me being a male, asking very personal questions. ID2, male teleVA interviewer, Western Cape

Some when they were giving information, they started crying - I didn’t know what to say to the person, because it was on the phone. I stayed quiet and waited, then asked if we should continue. ID5, male teleVA interviewer Western Cape

In the Western Cape, interviewers encountered an additional challenge related to managing rejection during phone interviews. They observed that some respondents were less respectful on the phone compared with face-to-face interactions with families. Surprisingly, this experience of rejection led to the interviewers forming a closer team bond and providing peer support. This support system enabled them to develop strategies to enhance their approach over time when contacting NoK. As a result, interviewers underscored the importance of thorough and specialised training for conducting effective VA interviews over the phone.

It’s much more difficult to reject when you’re in person. There’s nothing else you can do if a person hangs up. When they see your face and you will say it won’t take long…, they see this person is at work, so maybe they need help… but over the phone there’s not much that you can do. ID3, female teleVA interviewer, Western Cape

In contrast, the interviewers from the SAPRIN nodes mentioned the openness of the family members to participate in the teleVA. This could have been due to the communities of the SAPRIN nodes being very familiar with the HDSS work and call centre staff calling for health-related interviews twice a year, versus the ‘cold’ calling in the Western Cape.

Theme 2: perceptions of the social and individual benefits of participating in a teleVA

In general, the interviewed respondents appreciated the contact of government officials to inquire about the death of a family member. They saw the VA as a mechanism of the Department of Health to improve the health of the population which many perceived as caring about the passing of their family member. Some respondents who lost family members to an unknown cause mentioned that they would have appreciated hearing feedback from the VA, indicating the probable CoD.

This is awesome to have understanding of what is happening. I appreciated that someone took the time to know we’re ok. Gave me hope in our health system. ID5, female respondent, Western Cape

In the Western Cape, all participants in the pilot study received a small voucher as compensation for their participation, while in the SAPRIN nodes, there was no such compensation. There were mixed responses regarding the importance of the voucher in the Western Cape. Among the respondents, some mentioned that they appreciated and understood the importance of the VA for public health policy and therefore were happy to participate even without a voucher. Other respondents mentioned that considering the time it took to do the VA, the provision of a voucher was very relevant.

It was a nice thing to have, but not essential. I would have done it anyway. ID3, male respondent, Western Cape

The interviewers also provided similar mixed responses. Some had the feeling that without vouchers the response rates would have been lower, while others thought this was not important for the participation rate.

Sometimes you get the feeling. If we remove the voucher that we give the respondents, the response rate would have gone down. ID2, male teleVA interviewer, Western Cape

Theme 3: building trust over the phone to conduct a VA

In all three settings, the interviewers generally described the process of finding the right respondent as a difficult process. The telephone numbers provided or details of the NoK were not always correct. However, those who would pick up the phone would usually help to try to find the right respondent.

An interesting difference between the SAPRIN nodes and Western Cape was building rapport with respondents. In the Western Cape, the interviewers required several attempts of calling and sending information via WhatsApp or SMS before managing to conduct a VA. While in the SAPRIN nodes, this was quite straightforward as community members were used to participating in research and often had previously met the interviewers in person.

Some people who are young, they understand better talking over the phone, but older people don’t. I think people who are 18–40 [years old] they are more ok with this. But lots of our respondents are older people, because they’re usually the ones taking care of the deceased. ID6, male teleVA interviewer, SAPRIN node

Theme 4: teleVA respondents felt comfortable discussing the VA over the phone

During the interviews, we also explored the level of comfort that the respondents had in discussing the VA topics on the phone. There were mixed opinions on this topic, some respondents mentioned that they preferred doing a VA on the phone as they felt less pushed to participate at the time of the first contact, while others mentioned the importance of discussing these topics face-to-face. Furthermore, several respondents mentioned that due to the COVID-19 situation, they understood why the interviews were done telephonically.

I was very comfortable doing the interview, but would have preferred in person, to see face expressions. The good thing on phone, I could ask questions and not being judged. But face-to-face is a bit more personal. ID 1, female respondent, Western Cape

The interviewers also mentioned that despite the initial challenges to discuss sensitive topics on the phone, the teleVAs were often more comfortable as the respondents would limit the expressing of their emotions which was sometimes very challenging to manage during the face-to-face interviews. Lastly, several respondents also mentioned that during the teleVA they felt less shy to ask questions if there was something they would not understand.

She made me feel very comfortable and that helps. There was no pushiness. She asked and appreciated my time. Gives you the sense of being valued and your time also being important. ID5, female respondent, Western Cape

In one of the SAPRIN nodes, there was a period where face-to-face and phone interviews were combined as part of the routine surveillance at the node. The interviewer would reach out to respondents on the phone and if they preferred to answer in person, she would travel and conduct an in-person face-to-face interview. She described this as an efficient model, as she could decide to visit those people who felt uncomfortable on the phone or would not pick up when she was calling. This was also the case for elderly people who had some challenges in hearing her speak on the phone.

Some people don’t hear well over the phone or prefer discussing these topics in person. From 50 participants, maybe 1 prefers home. ID7, female teleVA interviewer, SAPRIN node

Theme 5: phone interviews were perceived as a safer and more flexible tool by the interviewers

The interviewers raised several benefits of doing the VAs over the phone, most being related to the fact that the teleVAs were more flexible and less time-consuming than doing the VA in person. Several interviewers, as well as respondents, indicated that they appreciated being flexible about the time of day the interview could be conducted.

…Cheaper, because you save time, no going to households. Sometime you can go there and the respondent isn’t there. Over the phone you can call another time. Easier to get people. ID6, male teleVA interviewer, SAPRIN node

None expected the numbers to grow as quickly as they are. It’s working well. If we were in the field, we wouldn’t be able to do 5 interviews. In theory, we can do more now, but depends how many potential respondents answer. ID3, female teleVA interviewer, Western Cape

The respondents said it was easier to ask to postpone the interview if it was by phone than in person and they appreciated having less pressure.

Furthermore, the interviewers, as well as some respondents highlighted the safety aspect as an important consideration when choosing between conducting phone versus in-person VAs.

Some of the people we interview are living in areas that are dangerous. Doing it over the phone is much safer. Once you are out there you are on your own. It’s been positive in that sense. ID1, male teleVA interviewer, Western Cape

Theme 6: respondents found emotional support in participating in VA interviews

Several respondents described how participating in the VA process helped them process their trauma and grief. They described that the discussion with the teleVA interviewer was at times very painful, but it also helped them open up and discuss the death with other family members.

The questions made me very emotional, as we never spoke about this. After the interview, I broke down and cried. After waking up, I talked with my mom about this and we opened up. ID1, female teleVA respondent, Western Cape

Limitations

A major limitation of this pilot was the identification and recruitment of NoK, due to the lack of access to current contact details. This information is routinely collected by the Department of Home Affairs during the registration of a death. However, for this pilot, these data were not available, necessitating the use of alternative recruitment methods, such as engaging with EMS and FPs. Addressing this challenge requires collaboration between sectors, including the Department of Health and the Department of Home Affairs, to effectively collect crucial information for successful teleVA implementations in the future.6

It is important to acknowledge the limitation that not everyone has access to smartphones which may restrict the feasibility of this approach for some individuals. Therefore, it would be important for future endeavours in this field to remain adaptive and consider the unique needs and infrastructural variations of each country and setting. By adopting a flexible and context-specific approach, we can further refine and tailor virtual assessments to ensure their optimal impact on healthcare and research practices.

Other limitations included the lack of community engagement and funding constraints which limited the number of calls to respondents. These issues are both likely to have impacted the response rate. Community engagement will be important in any future implementation of teleVA.

Additionally, the study’s comparison between telephonic and in-person VAs is limited by the differences between the two cohorts which may limit the results. While this study provides valuable insights, future research would require a comparative design to rigorously assess the differences in performance and quality between teleVAs and face-to-face VAs. Such an approach would help determine the most effective and scalable methods for capturing high-quality mortality data in diverse settings.

Patient consent for publication

Not applicable.

Ethics approval

The study received approval from the SAMRC ethics committee (EC005-2/2021) and an ethical exemption from the Swiss Tropical and Public Health Institute.