Sierra Leone using 'verbal autopsies' to make the dead count

Sierra Leone has one of the lowest life expectancies in the world at 55 years old. A majority of Sierra Leoneans – 63 percent – die before age 70, mostly from preventable or treatable causes. But improving the odds that people will lead long and healthy lives requires a better sense of what ails them.

That can be a challenge in low- and middle-income countries, where those in rural areas do not access the formal health system. In Sierra Leone, for example, an estimated half of deaths occur in rural areas or at home, outside the formal health system.

“Sierra Leone has had a lot of health challenges, from Ebola to malaria,” said Dr. Rashid Ansumana, Dean School of Community Health Sciences, Njala University, Bo Campus. “Our country is at the bottom of the United Nations' ranking on many health indicators. And for other health indicators, it is just blank. There is no data for things that need to be measured.”

In 2019, Sierra Leone launched an ambitious program to resolve this quandary. Researchers from Njala University and the Centre for Global Health Research at the University of Toronto, in close partnership with the country's Ministry of Health and Sanitation, began conducting surveys in 678 randomly selected, nationally representative villages and urban neighborhoods throughout the country. More than 60 surveyors traveling by foot, canoe, and motorbike visit these communities once a year and identify who has died over the previous 12 months, then gather data on their symptoms by interviewing surviving family members. Two physicians use the information to assign a probable cause of death, completing a process known as a verbal autopsy.

The data has surprised researchers and ministry officials, and is already changing health care strategies and policies in important ways.

Researchers found that malaria was the leading cause of death, not only for children under five and pregnant women, but also for people of all ages except newborns. Previously, the death curve for malaria was thought to resemble a steep L curve, with most deaths occurring in children under five. Verbal autopsy data painted a different picture. The malaria death rate resembled more of a U curve, with especially high mortality in young children and the elderly. Malaria was the leading cause of death at 38 percent for children ages 1 to 59 months. But even among middle-aged and older adults, malaria caused 14 percent of deaths. Even in the country's largely urban Western Area, malaria was a leading cause of death. As a result of this new data, Sierra Leone’s malaria directorate is adjusting its strategies, including expanding indoor residual spraying efforts in areas vulnerable to the disease.

The researchers also determined that Sierra Leone, with an estimated maternal mortality ratio of over 1,100 per 100,000 live births, and once identified by World Health Organization computer models as the most dangerous place for a pregnant woman to give birth may no longer hold that dubious title.

“The real maternal mortality ratio is about half of what WHO has estimated, but still unacceptably high,” said Dr. Ansumana, who is also a co-investigator in the research project. “With this survey, we can see that the country has achieved some success with its efforts to reduce maternal mortality. No doubt it is still difficult for pregnant women in Sierra Leone. But it is getting better.”

These insights are important in helping identify what’s working, where it is working, where people are still not accessing health care, and what sort of health care is most needed in each area of the country. This can help in planning and targeting delivery.

The research “illuminates the high value of reliable cause-of-death information for improving life expectancy in Sierra Leone and other low-income countries in Africa and Asia,” wrote researchers in a Lancet article published earlier this year.

The country’s verbal autopsy surveys have also found that many people whose deaths were not registered by the health system had nonetheless been treated previously within the formal health system. They may have traveled to a health clinic and been diagnosed with malaria or tuberculosis or cancer and then gone home and died of that illness at a later date. That their death was never registered illustrates another problem, one that exists outside the Ministry of Health and Sanitation.

"The paperwork required to register a death is too complicated,” said Dr. Ansumana. “So people don’t bother registering the death of their family member. They just want to bury their dead. Registering a death should be made simpler and more accessible.”

These findings have prompted the National Civil Registration Authority, which manages birth and death records, to simplify the process and expand their touchpoints in rural areas.

Dr. Ansumana said his team is also planning to reach out to the Ministry of Transport to discuss the high number of traffic-related deaths they discovered. “Drivers here are a menace,” said Dr. Ansumana. “We see it in the data and I see it when I am driving on our roads.”

Sierra Leone’s Ministry of Health and Sanitation launched the verbal autopsy effort as part of the Countrywide Mortality Surveillance for Action (COMSA), a national platform that aims to generate important mortality birth and data to inform public health policies and priorities. The data gathered through electronic verbal autopsies becomes part of the Sierra Leone Sample Registration System (SL-SRS) of births and deaths. SL-SRS aims to track about five percent of the total population, deaths, stillbirths, and pregnancies.

To ensure the data is leveraged by those who need it most, leaders of the country’s malaria directorate, its Chief Medical Officer, and the Ministry of Health’s directorate of policy, planning and information are deeply involved in the effort. Also, COMSA briefs national health officials and district medical officers regularly and issues reports twice a year.

To ensure quality control, the surveyors, who often walk more than ten miles a day to reach villagers, are tracked using GPS. Each surveyor’s reports are reviewed weekly to ensure they are detailed enough to support reliable conclusions on cause of death. And communities are regularly resurveyed to spot check findings.

“Our quality control measures aim to eliminate shortcuts,” said Dr. Ansumana.

Sierra Leone’s program is modeled on the influential India’s Million Death Study, or MDS, which involved six monthly in- person surveys of more than one million nationally representative households across the country between 2001 and 2014. The study assigned probable cause of death through verbal autopsies for individuals in 1.3 million households – about one percent of the deaths across the country. India’s Million Death Study, like Sierra Leone’s surveys, identified a previously underestimated impact of malaria across all age groups. The MDS found about 200,000 malaria deaths before age 70 years, whereas the WHO had estimated only 15,000.

“The Indian malaria findings were very controversial at the time,” said Dr. Prabhat Jha, principal investigator of the study and the founding director of the Centre for Global Health Research at St. Michael’s Hospital in Toronto, Ontario. “This finding spurred a tremendous amount of action to ramp up malaria programming.”

It also demonstrated that India has about one million deaths attributable to smoking each year – a figure three times higher than previous WHO estimates. It found shockingly high numbers of people dying of snake bites – 50,000. That was the number of snakebite deaths the WHO had previously estimated for the entire world. This data point prompted the WHO to double its estimate for global snakebite deaths and add this cause of death to its list of priority neglected tropical diseases. Meanwhile, India’s Ministry of Health worked to improve access to snake bite venom in high-risk areas.

The India research spurred global and national health leaders to wonder if there were tools they might use to more regularly understand mortality in lower- and middle-income countries and to decrease the reliance on uncertain modeling. The discussion led to the establishment of COMSA in Sierra Leone and Mozambique and the CHAMPS program, which analyzes biological samples to better understand child mortality.

“This shows the promise of the approach,” said Dr. Jha. “There is the possibility that Sierra Leone might be able to transform its mortality rates.”

Dr. Jha added that the Sierra Leone COMSA program continued throughout COVID, helping leaders understand the impact of the pandemic. The program has piloted an expansion, asking urban residents and villagers for permission to prick their finger and take just five drops of blood to determine infection rates of dengue, malaria, typhoid, and COVID to expand evidence and improve health services.

Verbal autopsy studies are powerful and cost-effective tools for filling data gaps in countries with limited vital statistics or weak vital records performance, particularly where most deaths will continue to occur out of hospital for decades, said Dr. Jha.

In an effort to share learnings across the continent and around the world, COMSA, in collaboration with the Sierra Leone Ministry of Health and Sanitation and Njala University, is organizing an Africa Mortality Symposium in Freetown on November 29-30, 2022. To aid countries in ensuring that their verbal autopsy programs minimize bias and generate valid and useful data, the World Health Organization has also developed tools and guidelines for health leaders.