Introduction: Epilepsy is one of the most common neurological conditions worldwide, with large variation in prevalence across sub-Saharan African countries. Northern Uganda is one of the poorest areas of the country and has seen a high density of pigs and a prevalence of Taenia solium, a zoonotic tapeworm transmitted which causes neurocysticercosis in humans. The objective of this study was to estimate the population-level prevalence of active epilepsy in 25 sub-counties of northern Uganda. Methods: This cross-sectional study was conducted in 2010-2011 in 25 sub-counties of Moyo, Adjumani, and Gulu districts, northern Uganda. Participants were sampled using a multistage cluster sampling strategy including sub-counties, parishes, villages, and households as sampling levels. Eligible individuals were interviewed using a previously validated screening questionnaire for epilepsy. Screen positive individuals were further examined by a team of neurologists for confirmation of active epilepsy. Sampling weights and post-stratification to account for sex distribution in each of the 25 sub-counties sampled based on projected 2010 population sizes were applied. Results: A total of 38,303 individuals were sampled across 299 villages from 25 sub-counties. The overall weighted and post-stratified prevalence estimate of active epilepsy was 3.7% (95% confidence interval [CI]: 3.4%–3.9%). However, there was large variation across sex (4.6% (95% CI: 4.2%–5.0%) in men and 2.7% (95% CI: 2.4%–3.0%) in women) and across sub-counties ranging from 1.7% in Pece Division (Gulu District) and Moyo Town Council (Moyo District) to 6.6% in Awach (Gulu District). People aged between 10 and 19 were the most affected. Conclusions: In northern Uganda, active epilepsy was very prevalent but varied largely across sub-counties. Males were a lot more affected than women, making the use of weighted and post-stratified methods to estimate the prevalence of epilepsy essential. Implementing programs and interventions targeting the control of local risk factors of epilepsy such as neurocysticercosis and improving population health care access could help reduce the rather high prevalence of epilepsy in this area of the country.

Epilepsy results from a combination of genetic and acquired causal mechanisms [1], and affects people of all ages, ethnicities, social classes, and geographical backgrounds [2]. It is one of the most common neurological conditions [3] affecting almost 46 million people worldwide (active epilepsy, idiopathic, and acquired) in 2016 [4], nearly 80% of whom live in low- and middle-income countries (LMICs) [5].

The prevalence of lifetime and active epilepsy is reported to be higher in LMICs, particularly in rural areas, than in high-income countries (HICs). The median prevalence of lifetime epilepsy was estimated at 15.4 per 1,000 and 10.3 per 1,000 in rural and urban LMIC areas, respectively, while it was lower at 5.8 per 1,000 in HICs [6]. In a recent systematic review and meta-analysis, the prevalence of lifetime epilepsy was 8.75 and 5.18 per 1,000 and that of active epilepsy 6.68 and 5.49 per 1,000 for LMICs and HICs, respectively [7]. In sub-Saharan Africa (SSA), the median prevalence of epilepsy based on two reviews with meta-analysis of population-based door-to-door surveys was reported to be 15.4 per 1,000 in 2005 (not specified whether lifetime or active) [8] and 14.2 per 1,000 in 2014 (both active and lifetime) [9]. Another systematic review and meta-analysis conducted in 2019 estimated the pooled prevalence of lifetime and active epilepsy in SSA at 16 per 1,000 and 9 per 1,000 [10], respectively, with important variation between countries in SSA, ranging from 2.2 per 1,000 in Ghana to 49 per 1,000 in Cameroon [10].

The wide variability in the prevalence estimates of epilepsy across studies could be explained by variations in the distribution of epilepsy risk factors across populations, demographic structure of the population studied, quality and access to health care, knowledge, and attitudes toward epilepsy, as well as the heterogeneity of the study methods used, which may result from differences in the definitions of (active) epilepsy, the type (lifetime or active) of epilepsy studied, and the population sampled (general or selected) [2, 8, 9, 11‒14]. For example, neurocysticercosis (NCC), an infection of the central nervous system caused by Taenia solium, a tapeworm transmitted between humans and pigs, has been found in 29% (95% confidence interval [CI]: 22.9%–35.5%) of people with epilepsy in a meta-analysis conducted between 1990 and 2008 [15].

Population-based neuro-epidemiological studies in low-resource settings often sample the source population through a clustered method, whereby a set of locations and sometimes households are sampled followed by door-to-door surveys during which consenting participants, constituting the study population, are administered a screening questionnaire for epilepsy [13]. Those screened positive then go through clinical confirmation by a physician or neurologist. However, such sampling approaches may lead to biased population-level prevalence estimates if the sampled population is not reflective of the population of interest. For example, if males are underrepresented in the study population but the prevalence of epilepsy is higher, then the observed prevalence would underestimate the prevalence in the target population. Therefore, adjusting simultaneously for the sample’s selection probability and for the demographic characteristics of the target population, the data using census data will lead to prevalence estimates generalizable to the population of interest [16‒18].

Northern Uganda is one of the regions with very high poverty levels aggravated by the brutal rebel insurgency of the Lord’s Resistance Army from 1986 to 2006, that disrupted known public health control measures in the region and has seen an important increase in traditional pig raising in the past few decades. It was thus suspected that the prevalence of epilepsy would be higher than elsewhere due to T. solium infection of pigs and humans, which may lead to NCC, one of the leading secondary causes of epilepsy in T. solium endemic areas of SSA [19], but also due to other causes of poverty [20, 21]. However, the prevalence of epilepsy so far has not been estimated in a large-scale population-based survey in this area at risk of NCC. Therefore, the objective of this study was to estimate and compare the population-level prevalence of active epilepsy in 25 sub-counties of three districts of northern Uganda.

Study Design and Setting

This study was a cross-sectional multistage cluster random survey conducted from April 2010 to March 2011 in 25 sub-counties from three districts in northern Uganda, namely, Moyo, Adjumani, and Gulu. All the sub-counties of Adjumani and Gulu districts were included in the study. In Moyo, we included only four (Dufile, Metu, Moyo, and Moyo Town Council) of the eight sub-counties as the other four sub-counties mainly comprised refugee camps, and their population was considered unstable. Based on the 2010 district population projection data provided by Moyo, Gulu, and Adjumani districts Local Governments Planning Department, the total population living in the 25 sampled sub-counties was estimated at 704,900 people, with an equal distribution between the sexes (online suppl. File 1; for all online suppl. material, see https://doi.org/10.1159/000543472).

Sampling Strategy

The source population was people living in the districts of Moyo, Adjumani, and Gulu in northern Uganda. All parishes in the 25 selected sub-counties were included and visited, except for two parishes in Adjumani which could not be reached due to poor road network. In each parish, a list of villages was generated in alphabetical order and a sampling interval was calculated using the total number of villages in the parish and the desired number of villages to be sampled, and then systematic random sampling of the villages was done. For small parishes with few villages of 2 or 3, all villages were included in the sample. In each village, households were also sampled using systematic random sampling techniques. It started at a random point in the village, a bottle was spun, and sampling was started with the household where the bottle neck faced. The total sampled household population was divided by the desired sample size of household to obtain the sampling interval. Households were selected within this village according to the obtained sampling interval. In the households visited, an eligible individual was every member of the sampled household from 1 month of age up to the oldest, who was available during the study period. Online supplementary File 2 provides additional information on the sampling procedure and statistics on the population and on the people screened in the three districts surveyed.

Definitions, Tools, and Participant Evaluation

All individuals older than 1 month and present at the time the study team visited the household were deemed eligible upon consent. Parental or caregiver consent was provided for participants aged 1 month to 17 years of age and those mentally challenged, and assent requested for those aged 8–17 years. The evaluation of participants involved two phases: a screening phase (first phase) and an in-depth examination phase (second phase). Participants that answered positively to any of the questions 1–9 of the screening questionnaire went on to the second phase of the study and individuals with febrile seizures only as assessed through question 10 of the screening questionnaire were excluded a priori and did not proceed to the second phase of the study (online suppl. File 3). The current article reports the results from the screening phase.

During the screening phase, eligible individuals were interviewed using the screening questionnaire to identify those with epileptic seizures (online suppl. File 3). The screening questionnaire used in this study was developed, pre-tested and extensively applied in the African context. The content of the questionnaire was inspired by works by Placencia et al. [22] and Birbeck and Kalichi [23], who conducted epidemiological studies on epilepsy and adapted by Andrea S. Winkler, senior author of the article, through experiences gained while working with people with epilepsy in northern Tanzania 2002–2004. For validation purposes, the questionnaire was tested on 106 known people with epilepsy from the Epilepsy Clinic at the Haydom Lutheran Hospital (HLH) in northern Tanzania. The questionnaire was further tested on 104 patients in hospital. Some of them had been admitted because of epileptic seizures; the interviewers, however, were blinded to the positive cases. The calculated sensitivity was 100% and the specificity 97.5% [24]. Afterward, a pilot study was conducted in the wider catchment area of HLH on 400 people in November 2003, in which the practicability of the screening questionnaire was tested. Over the years and through the use of different community-based studies throughout SSA, the questionnaire was slightly adjusted to fit the needs of the respective studies. A suspected epilepsy case in our study was defined if any of the questions 1–9 were answered positively and question 10 was answered negatively. Question 10 refers to febrile seizures, which were excluded a priori from the study population (online suppl. File 3).

Parents or guardians answered the questionnaire on behalf of their children. All suspected cases of epileptic seizures were invited to the hospital and or nearest health center III, where they were re-interviewed and examined by the researchers and a team of neurologists from Germany and Austria. People who did not come to the health facilities were visited at home in the communities by the research team. Once the screening questionnaire was answered positively, an in-depth clinical questionnaire was applied to confirm people with epilepsy, and more specifically with active epilepsy. For our study, epilepsy was defined as two or more seemingly unprovoked epileptic seizures at least 24 h apart. Active epilepsy was defined as at least one epileptic seizure during the preceding year among people with epilepsy, irrespective of whether participants were on anti-seizure medication.

Statistical Methods

To obtain prevalence estimates at the sub-county level, we applied sampling weights to the data to account for the unequal selection probability of each individual in the study. Sampling weights were calculated for each study participant as the inverse of the individual’s selection probability [25]. In each sub-county, the individual’s selection probability was calculated as the product of the probability of selection of their village (ratio between the number of sampled villages within a parish divided by the total number of villages in that parish) and their own probability of being selected (ratio between the number of participants selected from the sampled villages divided by the total village population).

In addition to the sampling weights, the data were post-stratified to resemble the sex distribution in each of the 25 sub-counties based on the 2010 population projections. Post-stratification is a technique that adjusts the sampling weights to account for under-represented subgroups in the population [17]. The post-stratification was done to account for the under- and over-sampling of males and females in our study and to obtain prevalence estimates that are generalizable at the sub-county level.

We describe the quantitative and categorical variables by mean and percentage with their 95% CIs, respectively. We report unweighted and weighted and post-stratified prevalence estimates of active epilepsy and their corresponding 95% CI for each of the 25 sub-counties, and according to 10-year age group and sex. All analyses were conducted using Stata/MP 17.0 statistical software (StataCorp: College Station, TX, USA) and the map describing the geographical distribution of epilepsy in the 25 sampled sub-counties was drawn using ArcGIS Pro 2.8 software.

Study Population Characteristics

The study sample consisted of 38,303 individuals sampled across 299 villages from 25 sub-counties. Based on the 2010 population projections, the sampled study population corresponds to 5.4% of the individuals living across the 25 sub-counties. The study participation proportion was 61.7% (12,503/20,259) in Moyo district, 68.2% (12,800/18,763) in Adjumani district and 6.7% (13,000/194,080) in Gulu district (online suppl. File 2).

Men were underrepresented and constituted 42.7% (95% CI: 42.2%–43.2%) of the sample population, with large variations between sub-counties, ranging between 33.5% (95% CI: 31.6%–35.5%) in Adropi, Adjumani District, and 53.4% (95% CI: 49.1%–57.7%) in Bar-Dege, Gulu District. When weighted and post-stratified, the sex distribution in each sub-county was similar to that of the 2010 population projections, with each sex representing half of the analytical sample, overall (online suppl. File 1). The mean age of the sampled population was 19.3 years (95% CI: 19.1–19.5 years), with no difference between men (mean age 19.2 years [95% CI: 18.9–19.5 years]) and women (mean age 19.4 years [95% CI: 19.1–19.6 years]) (Table 1). Overall, people aged less than 10 years were the most represented with a weighted and post-stratified proportion of 38.5% (95% CI: 37.8%–39.1%). Over three-quarters (76.5%) of the study sample were aged less than 30 years. No significant difference in age group distribution was observed between females and males (Table 2).

Table 1.

Age and sex distribution of study population per sub-county using data from 38,303 sampled individuals across 25 sub-counties in three districts of northern Uganda, 2010–2011

Sub-countiesFemaleMale
n% (95% CI)mean age (95% CI), yearsn% (95% CI)mean age (95% CI), years
All 21,952 57.3 (56.8–57.8) 19.4 (19.1–19.6) 16,351 42.7 (42.2–43.2) 19.2 (18.9–19.5) 
Adjumani TC 1,308 57.0 (55.0–59.0) 19.6 (18.7–20.5) 986 43.0 (41.0–45.0) 19.8 (18.7–20.9) 
Adropi 1,468 66.5 (64.5–68.4) 20.1 (19.1–21.1) 741 33.5 (31.6–35.5) 20.2 (18.9–21.6) 
Awach 335 58.4 (54.3–62.3) 18.3 (16.8–19.8) 239 41.6 (37.7–45.7) 18.5 (16.7–20.3) 
Bar-Dege 241 46.6 (42.4–50.9) 20.1 (17.7–22.5) 276 53.4 (49.1–57.7) 18.1 (16.3–19.9) 
Bobi 419 60.3 (56.6–63.9) 20.4 (18.8–22.1) 276 39.7 (36.1–43.4) 21.1 (19.0–23.3) 
Bungatira 638 59.4 (56.4–62.2) 19.9 (18.6–21.2) 437 40.6 (37.8–43.6) 20.8 (19.2–22.5) 
Ciforo 1,020 61.8 (59.5–64.1) 19.8 (18.7–20.8) 630 38.2 (35.9–40.6) 19.6 (18.2–21.0) 
Dufile 1,884 56.8 (55.1–58.4) 19.0 (18.2–19.7) 1,436 43.2 (41.6–45.0) 18.9 (18.1–19.8) 
Dzaipi 962 55.4 (53.0–57.7) 17.7 (16.5–18.8) 776 44.6 (42.3–47.0) 16.2 (15.0–17.4) 
Koro 638 56.5 (53.6–59.4) 20.6 (19.1–22.2) 491 43.5 (40.6–46.4) 18.5 (16.9–20.2) 
Lakwana 397 56.8 (53.1–60.4) 16.7 (15.0–18.3) 302 43.2 (39.6–46.9) 18.0 (15.9–20.1) 
Lalogi 864 60.1 (57.6–62.6) 16.4 (15.3–17.6) 573 39.9 (37.4–42.4) 15.9 (14.5–17.2) 
Laroo 197 49.4 (44.5–54.3) 17.8 (15.6–19.9) 202 50.6 (45.7–55.5) 19.4 (17.2–21.6) 
Layibi 303 58.3 (54.0–62.4) 24.0 (21.7–26.3) 217 41.7 (37.6–46.0) 23.1 (20.4–25.8) 
Metu 1,500 49.6 (47.8–51.4) 20.0 (19.1–20.9) 1,523 50.4 (48.6–52.2) 20.2 (19.3–21.1) 
Moyo 2,003 58.7 (57.0–60.4) 18.3 (17.5–19.1) 1,409 41.3 (39.7–43.0) 17.3 (16.4–18.2) 
Moyo TC 1,775 64.6 (62.8–66.4) 20.5 (19.7–21.4) 973 35.4 (33.6–37.2) 20.6 (19.4–21.7) 
Odek 1,012 62.2 (59.8–64.5) 21.4 (20.3–22.6) 615 37.8 (35.5–40.2) 21.2 (19.7–22.7) 
Ofua 1,085 52.0 (49.8–54.1) 19.8 (18.7–21.0) 1,003 48.0 (45.9–50.2) 19.8 (18.6–21.0) 
Ongako 1,161 58.9 (56.8–61.1) 19.6 (18.6–20.6) 809 41.1 (38.9–43.3) 19.9 (18.7–21.1) 
Paicho 596 56.3 (53.3–59.3) 18.8 (17.5–20.1) 462 43.7 (40.7–46.7) 18.6 (17.1–20.1) 
Pakelle 1,417 50.2 (48.4–52.1) 19.4 (18.5–20.3) 1,404 49.8 (47.9–51.6) 19.6 (18.7–20.5) 
Palaro 194 59.2 (53.7–64.4) 15.5 (13.7–17.2) 134 40.8 (35.7–46.3) 17.6 (15.0–20.2) 
Patiko 195 57.7 (52.4–62.9) 17.1 (14.8–19.4) 143 42.3 (37.1–47.7) 18.4 (15.4–21.3) 
Pece 340 53.6 (49.7–57.5) 18.3 (16.6–20.0) 294 46.4 (42.5–50.3) 18.7 (16.9–20.6) 
Sub-countiesFemaleMale
n% (95% CI)mean age (95% CI), yearsn% (95% CI)mean age (95% CI), years
All 21,952 57.3 (56.8–57.8) 19.4 (19.1–19.6) 16,351 42.7 (42.2–43.2) 19.2 (18.9–19.5) 
Adjumani TC 1,308 57.0 (55.0–59.0) 19.6 (18.7–20.5) 986 43.0 (41.0–45.0) 19.8 (18.7–20.9) 
Adropi 1,468 66.5 (64.5–68.4) 20.1 (19.1–21.1) 741 33.5 (31.6–35.5) 20.2 (18.9–21.6) 
Awach 335 58.4 (54.3–62.3) 18.3 (16.8–19.8) 239 41.6 (37.7–45.7) 18.5 (16.7–20.3) 
Bar-Dege 241 46.6 (42.4–50.9) 20.1 (17.7–22.5) 276 53.4 (49.1–57.7) 18.1 (16.3–19.9) 
Bobi 419 60.3 (56.6–63.9) 20.4 (18.8–22.1) 276 39.7 (36.1–43.4) 21.1 (19.0–23.3) 
Bungatira 638 59.4 (56.4–62.2) 19.9 (18.6–21.2) 437 40.6 (37.8–43.6) 20.8 (19.2–22.5) 
Ciforo 1,020 61.8 (59.5–64.1) 19.8 (18.7–20.8) 630 38.2 (35.9–40.6) 19.6 (18.2–21.0) 
Dufile 1,884 56.8 (55.1–58.4) 19.0 (18.2–19.7) 1,436 43.2 (41.6–45.0) 18.9 (18.1–19.8) 
Dzaipi 962 55.4 (53.0–57.7) 17.7 (16.5–18.8) 776 44.6 (42.3–47.0) 16.2 (15.0–17.4) 
Koro 638 56.5 (53.6–59.4) 20.6 (19.1–22.2) 491 43.5 (40.6–46.4) 18.5 (16.9–20.2) 
Lakwana 397 56.8 (53.1–60.4) 16.7 (15.0–18.3) 302 43.2 (39.6–46.9) 18.0 (15.9–20.1) 
Lalogi 864 60.1 (57.6–62.6) 16.4 (15.3–17.6) 573 39.9 (37.4–42.4) 15.9 (14.5–17.2) 
Laroo 197 49.4 (44.5–54.3) 17.8 (15.6–19.9) 202 50.6 (45.7–55.5) 19.4 (17.2–21.6) 
Layibi 303 58.3 (54.0–62.4) 24.0 (21.7–26.3) 217 41.7 (37.6–46.0) 23.1 (20.4–25.8) 
Metu 1,500 49.6 (47.8–51.4) 20.0 (19.1–20.9) 1,523 50.4 (48.6–52.2) 20.2 (19.3–21.1) 
Moyo 2,003 58.7 (57.0–60.4) 18.3 (17.5–19.1) 1,409 41.3 (39.7–43.0) 17.3 (16.4–18.2) 
Moyo TC 1,775 64.6 (62.8–66.4) 20.5 (19.7–21.4) 973 35.4 (33.6–37.2) 20.6 (19.4–21.7) 
Odek 1,012 62.2 (59.8–64.5) 21.4 (20.3–22.6) 615 37.8 (35.5–40.2) 21.2 (19.7–22.7) 
Ofua 1,085 52.0 (49.8–54.1) 19.8 (18.7–21.0) 1,003 48.0 (45.9–50.2) 19.8 (18.6–21.0) 
Ongako 1,161 58.9 (56.8–61.1) 19.6 (18.6–20.6) 809 41.1 (38.9–43.3) 19.9 (18.7–21.1) 
Paicho 596 56.3 (53.3–59.3) 18.8 (17.5–20.1) 462 43.7 (40.7–46.7) 18.6 (17.1–20.1) 
Pakelle 1,417 50.2 (48.4–52.1) 19.4 (18.5–20.3) 1,404 49.8 (47.9–51.6) 19.6 (18.7–20.5) 
Palaro 194 59.2 (53.7–64.4) 15.5 (13.7–17.2) 134 40.8 (35.7–46.3) 17.6 (15.0–20.2) 
Patiko 195 57.7 (52.4–62.9) 17.1 (14.8–19.4) 143 42.3 (37.1–47.7) 18.4 (15.4–21.3) 
Pece 340 53.6 (49.7–57.5) 18.3 (16.6–20.0) 294 46.4 (42.5–50.3) 18.7 (16.9–20.6) 

CI, confidence interval; TC, Town Council.

Table 2.

Ten-year age group unweighted and weighted and post-stratified distribution of study population by sex using data from 38,303 sampled individuals across 25 sub-counties in three districts of northern Uganda, 2010–2011

Age group, yearsFemale (n = 21,952)Male (n = 16,351)Total (n = 38,303)
nunweighted % (95% CI)weighted and post-stratified % (95% CI)nunweighted % (95% CI)weighted and post-stratified % (95% CI)nunweighted % (95% CI)weighted and post-stratified % (95% CI)
<10 8,784 40.0 (39.4–40.7) 39.2 (38.3–40.0) 6,308 38.6 (37.8–39.3) 37.7 (36.8–38.7) 15,092 39.4 (38.9–39.9) 38.5 (37.8–39.1) 
10–19 5,392 24.6 (24.0–25.1) 25.5 (24.7–26.3) 4,350 26.6 (25.9–27.3) 27.6 (26.7–28.5) 9,742 25.4 (25.0–25.9) 26.5 (26.0–27.1) 
20–29 2,498 11.4 (11.0–11.8) 11.5 (10.9–12.0) 1,865 11.4 (10.9–11.9) 11.5 (10.9–12.2) 4,363 11.4 (11.1–11.7) 11.5 (11.1–11.9) 
30–39 1,589 7.2 (6.9–7.6) 7.5 (7.1–8.0) 1,308 8.0 (7.6–8.4) 8.0 (7.5–8.5) 2,897 7.6 (7.3–7.8) 7.8 (7.4–8.1) 
40–49 1,551 7.1 (6.7–7.4) 6.6 (6.2–7.0) 1,020 6.2 (5.9–6.6) 6.0 (5.6–6.5) 2,571 6.7 (6.5–7.0) 6.3 (6.0–6.6) 
50–59 1,250 5.7 (5.4–6.0) 5.6 (5.2–5.9) 776 4.8 (4.4–5.1) 4.7 (4.3–5.1) 2,026 5.3 (5.1–5.5) 5.1 (4.9–5.4) 
60–69 712 3.2 (3.0–3.5) 3.2 (2.9–3.6) 582 3.6 (3.3–3.9) 3.6 (3.2–4.0) 1,294 3.4 (3.2–3.6) 3.4 (3.2–3.7) 
≥70 176 0.8 (0.7–0.9) 1.0 (0.8–1.2) 142 0.9 (0.7–1.0) 0.9 (0.7–1.1) 318 0.8 (0.7–0.9) 0.9 (0.8–1.1) 
Age group, yearsFemale (n = 21,952)Male (n = 16,351)Total (n = 38,303)
nunweighted % (95% CI)weighted and post-stratified % (95% CI)nunweighted % (95% CI)weighted and post-stratified % (95% CI)nunweighted % (95% CI)weighted and post-stratified % (95% CI)
<10 8,784 40.0 (39.4–40.7) 39.2 (38.3–40.0) 6,308 38.6 (37.8–39.3) 37.7 (36.8–38.7) 15,092 39.4 (38.9–39.9) 38.5 (37.8–39.1) 
10–19 5,392 24.6 (24.0–25.1) 25.5 (24.7–26.3) 4,350 26.6 (25.9–27.3) 27.6 (26.7–28.5) 9,742 25.4 (25.0–25.9) 26.5 (26.0–27.1) 
20–29 2,498 11.4 (11.0–11.8) 11.5 (10.9–12.0) 1,865 11.4 (10.9–11.9) 11.5 (10.9–12.2) 4,363 11.4 (11.1–11.7) 11.5 (11.1–11.9) 
30–39 1,589 7.2 (6.9–7.6) 7.5 (7.1–8.0) 1,308 8.0 (7.6–8.4) 8.0 (7.5–8.5) 2,897 7.6 (7.3–7.8) 7.8 (7.4–8.1) 
40–49 1,551 7.1 (6.7–7.4) 6.6 (6.2–7.0) 1,020 6.2 (5.9–6.6) 6.0 (5.6–6.5) 2,571 6.7 (6.5–7.0) 6.3 (6.0–6.6) 
50–59 1,250 5.7 (5.4–6.0) 5.6 (5.2–5.9) 776 4.8 (4.4–5.1) 4.7 (4.3–5.1) 2,026 5.3 (5.1–5.5) 5.1 (4.9–5.4) 
60–69 712 3.2 (3.0–3.5) 3.2 (2.9–3.6) 582 3.6 (3.3–3.9) 3.6 (3.2–4.0) 1,294 3.4 (3.2–3.6) 3.4 (3.2–3.7) 
≥70 176 0.8 (0.7–0.9) 1.0 (0.8–1.2) 142 0.9 (0.7–1.0) 0.9 (0.7–1.1) 318 0.8 (0.7–0.9) 0.9 (0.8–1.1) 

CI, confidence interval.

Prevalence of Active Epilepsy

The overall prevalence estimate of active epilepsy across the 25 sub-counties was 3.3% (95% CI: 3.1%–3.4%) when unweighted and 3.7% (95% CI: 3.4%–3.9%) when weighted and post-stratified. The lower unweighted prevalence estimate was due to the under-representation of males in the study sample, and the fact that they were more likely to have epilepsy. Indeed, the overall unweighted prevalence of active epilepsy was found to be 2.5% (95% CI: 2.3%–2.7%) among females and 4.2% (95% CI: 4.0%–4.6%) among males, and the corresponding weighted and post-stratified prevalence was 2.7% (95% CI: 2.4%–3.0%) and 4.6% (95% CI: 4.2%–5.0%), respectively. Males showed a higher prevalence of epilepsy in all Sub-Counties, except for Bobi and Laroo in Gulu District, and Ofua in Adjumani District (Table 3). The prevalence estimates of epilepsy varied across sub-counties from 1.7% in Pece (Gulu District) and Moyo Town Council (Moyo District) to 6.6% in Awach (Gulu District) for the weighted and post-stratified estimate. The sub-counties in the highest prevalence category were Awach (Gulu District) and Dzaipi (Adjumani District), and those in the lowest prevalence category are Bobi, Koro and Pece in Gulu District, Moyo TC (Moyo District), and Pakelle (Adjumani District) (Fig. 1).

Table 3.

Unweighted and weighted and post-stratified prevalence estimates (95% CI) of active epilepsy by sub-county and sex using data from 38,303 sampled individuals across 25 sub-counties in three districts of northern Uganda, 2010–2011

Sub-CountiesFemaleMaleTotal
unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)
All 2.5 (2.3–2.7) 2.7 (2.4–3.0) 4.2 (4.0–4.6) 4.6 (4.2–5.0) 3.3 (3.1–3.4) 3.7 (3.4–3.9) 
Adjumani TC 3.5 (2.6–4.7) 3.5 (2.7–4.7) 5.0 (3.8–6.5) 5.0 (3.8–6.6) 4.1 (3.4–5.0) 4.3 (3.5–5.2) 
Adropi 1.5 (1.0–2.3) 1.5 (1.0–2.3) 3.1 (2.1–4.6) 3.2 (2.1–4.7) 2.0 (1.5–2.7) 2.4 (1.7–3.2) 
Awach 4.2 (2.5–6.9) 4.5 (2.7–7.4) 8.4 (5.5–12.6) 8.9 (5.8–13.5) 5.9 (4.3–8.2) 6.6 (4.8–9.1) 
Bar-Dege 3.7 (2.0–7.0) 3.8 (2.0–7.1) 6.9 (4.4–10.5) 7.1 (4.6–10.9) 5.4 (3.8–7.7) 5.5 (3.8–7.8) 
Bobi 2.6 (1.5–4.7) 2.7 (1.5–4.8) 1.5 (0.6–3.8) 1.4 (0.5–3.9) 2.2 (1.3–3.6) 2.1 (1.2–3.4) 
Bungatira 2.4 (1.4–3.9) 2.4 (1.5–4.0) 3.4 (2.1–5.6) 3.4 (2.0–5.5) 2.8 (2.0–4.0) 2.9 (2.0–4.1) 
Ciforo 2.9 (2.1–4.2) 3.0 (2.1–4.3) 4.1 (2.8–6.0) 4.1 (2.8–6.0) 3.4 (2.6–4.4) 3.6 (2.7–4.7) 
Dufile 2.9 (2.2–3.7) 2.6 (2.0–3.5) 4.0 (3.1–5.2) 3.9 (3.0–5.1) 3.4 (2.8–4.1) 3.3 (2.7–4.0) 
Dzaipi 4.7 (3.5–6.2) 4.7 (3.5–6.3) 8.2 (6.5–10.4) 8.3 (6.6–10.5) 6.3 (5.2–7.5) 6.5 (5.4–7.8) 
Koro 1.4 (0.7–2.7) 1.5 (0.8–2.8) 2.4 (1.4–4.3) 2.4 (1.4–4.3) 1.9 (1.2–2.8) 1.9 (1.3–3.0) 
Lakwana 2.3 (1.2–4.3) 2.3 (1.2–4.3) 5.6 (3.5–8.9) 5.6 (3.5–8.8) 3.7 (2.6–5.4) 3.9 (2.6–5.6) 
Lalogi 3.2 (2.3–4.7) 3.3 (2.3–4.7) 5.4 (3.8–7.6) 5.3 (3.7–7.4) 4.1 (3.2–5.3) 4.2 (3.3–5.4) 
Laroo 5.6 (3.1–9.8) 6.0 (3.3–10.7) 4.5 (2.3–8.3) 4.6 (2.4–8.7) 5.0 (3.3–7.6) 5.3 (3.4–8.2) 
Layibi 2.3 (1.1–4.8) 2.4 (1.1–4.9) 8.3 (5.3–12.8) 8.6 (5.4–13.2) 4.8 (3.3–7.0) 5.4 (3.7–7.9) 
Metu 5.0 (4.0–6.2) 5.0 (4.0–6.2) 5.5 (4.5–6.8) 5.5 (4.5–6.8) 5.3 (4.5–6.1) 5.3 (4.5–6.1) 
Moyo 1.9 (1.3–2.5) 1.9 (1.3–2.5) 4.4 (3.5–5.6) 4.4 (3.5–5.6) 2.9 (2.4–3.5) 3.2 (2.6–3.8) 
Moyo TC 0.6 (0.3–1.0) 0.6 (0.3–1.0) 2.9 (2.0–4.1) 2.9 (2.0–4.1) 1.4 (1.0–1.9) 1.7 (1.3–2.4) 
Odek 3.1 (2.2–4.3) 3.0 (2.1–4.3) 6.0 (4.4–8.2) 5.8 (4.3–8.0) 4.2 (3.3–5.3) 4.4 (3.5–5.6) 
Ofua 2.9 (2.0–4.0) 2.9 (2.0–4.1) 2.6 (1.8–3.8) 2.6 (1.7–3.8) 2.7 (2.1–3.5) 2.7 (2.1–3.5) 
Ongako 1.1 (0.7–1.9) 1.1 (0.7–2.0) 3.6 (2.5–5.1) 3.7 (2.6–5.3) 2.1 (1.6–2.9) 2.4 (1.8–3.3) 
Paicho 1.9 (1.0–3.3) 1.8 (1.0–3.3) 3.3 (2.0–5.3) 3.2 (1.9–5.2) 2.5 (1.7–3.6) 2.5 (1.7–3.6) 
Pakelle 1.5 (1.0–2.3) 1.4 (0.9–2.1) 2.2 (1.6–3.1) 2.3 (1.6–3.2) 1.8 (1.4–2.4) 1.8 (1.4–2.4) 
Palaro 2.6 (1.1–6.0) 2.6 (1.1–6.0) 3.7 (1.6–8.7) 3.7 (1.6–8.7) 3.1 (1.7–5.6) 3.2 (1.7–5.9) 
Patiko 2.1 (0.8–5.3) 2.0 (0.7–5.2) 2.8 (1.1–7.2) 2.9 (1.1–7.6) 2.4 (1.2–4.7) 2.4 (1.2–4.8) 
Pece 0.9 (0.3–2.7) 0.9 (0.3–2.7) 2.7 (1.4–5.4) 2.7 (1.3–5.3) 1.7 (1.0–3.1) 1.7 (1.0–3.1) 
Sub-CountiesFemaleMaleTotal
unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)
All 2.5 (2.3–2.7) 2.7 (2.4–3.0) 4.2 (4.0–4.6) 4.6 (4.2–5.0) 3.3 (3.1–3.4) 3.7 (3.4–3.9) 
Adjumani TC 3.5 (2.6–4.7) 3.5 (2.7–4.7) 5.0 (3.8–6.5) 5.0 (3.8–6.6) 4.1 (3.4–5.0) 4.3 (3.5–5.2) 
Adropi 1.5 (1.0–2.3) 1.5 (1.0–2.3) 3.1 (2.1–4.6) 3.2 (2.1–4.7) 2.0 (1.5–2.7) 2.4 (1.7–3.2) 
Awach 4.2 (2.5–6.9) 4.5 (2.7–7.4) 8.4 (5.5–12.6) 8.9 (5.8–13.5) 5.9 (4.3–8.2) 6.6 (4.8–9.1) 
Bar-Dege 3.7 (2.0–7.0) 3.8 (2.0–7.1) 6.9 (4.4–10.5) 7.1 (4.6–10.9) 5.4 (3.8–7.7) 5.5 (3.8–7.8) 
Bobi 2.6 (1.5–4.7) 2.7 (1.5–4.8) 1.5 (0.6–3.8) 1.4 (0.5–3.9) 2.2 (1.3–3.6) 2.1 (1.2–3.4) 
Bungatira 2.4 (1.4–3.9) 2.4 (1.5–4.0) 3.4 (2.1–5.6) 3.4 (2.0–5.5) 2.8 (2.0–4.0) 2.9 (2.0–4.1) 
Ciforo 2.9 (2.1–4.2) 3.0 (2.1–4.3) 4.1 (2.8–6.0) 4.1 (2.8–6.0) 3.4 (2.6–4.4) 3.6 (2.7–4.7) 
Dufile 2.9 (2.2–3.7) 2.6 (2.0–3.5) 4.0 (3.1–5.2) 3.9 (3.0–5.1) 3.4 (2.8–4.1) 3.3 (2.7–4.0) 
Dzaipi 4.7 (3.5–6.2) 4.7 (3.5–6.3) 8.2 (6.5–10.4) 8.3 (6.6–10.5) 6.3 (5.2–7.5) 6.5 (5.4–7.8) 
Koro 1.4 (0.7–2.7) 1.5 (0.8–2.8) 2.4 (1.4–4.3) 2.4 (1.4–4.3) 1.9 (1.2–2.8) 1.9 (1.3–3.0) 
Lakwana 2.3 (1.2–4.3) 2.3 (1.2–4.3) 5.6 (3.5–8.9) 5.6 (3.5–8.8) 3.7 (2.6–5.4) 3.9 (2.6–5.6) 
Lalogi 3.2 (2.3–4.7) 3.3 (2.3–4.7) 5.4 (3.8–7.6) 5.3 (3.7–7.4) 4.1 (3.2–5.3) 4.2 (3.3–5.4) 
Laroo 5.6 (3.1–9.8) 6.0 (3.3–10.7) 4.5 (2.3–8.3) 4.6 (2.4–8.7) 5.0 (3.3–7.6) 5.3 (3.4–8.2) 
Layibi 2.3 (1.1–4.8) 2.4 (1.1–4.9) 8.3 (5.3–12.8) 8.6 (5.4–13.2) 4.8 (3.3–7.0) 5.4 (3.7–7.9) 
Metu 5.0 (4.0–6.2) 5.0 (4.0–6.2) 5.5 (4.5–6.8) 5.5 (4.5–6.8) 5.3 (4.5–6.1) 5.3 (4.5–6.1) 
Moyo 1.9 (1.3–2.5) 1.9 (1.3–2.5) 4.4 (3.5–5.6) 4.4 (3.5–5.6) 2.9 (2.4–3.5) 3.2 (2.6–3.8) 
Moyo TC 0.6 (0.3–1.0) 0.6 (0.3–1.0) 2.9 (2.0–4.1) 2.9 (2.0–4.1) 1.4 (1.0–1.9) 1.7 (1.3–2.4) 
Odek 3.1 (2.2–4.3) 3.0 (2.1–4.3) 6.0 (4.4–8.2) 5.8 (4.3–8.0) 4.2 (3.3–5.3) 4.4 (3.5–5.6) 
Ofua 2.9 (2.0–4.0) 2.9 (2.0–4.1) 2.6 (1.8–3.8) 2.6 (1.7–3.8) 2.7 (2.1–3.5) 2.7 (2.1–3.5) 
Ongako 1.1 (0.7–1.9) 1.1 (0.7–2.0) 3.6 (2.5–5.1) 3.7 (2.6–5.3) 2.1 (1.6–2.9) 2.4 (1.8–3.3) 
Paicho 1.9 (1.0–3.3) 1.8 (1.0–3.3) 3.3 (2.0–5.3) 3.2 (1.9–5.2) 2.5 (1.7–3.6) 2.5 (1.7–3.6) 
Pakelle 1.5 (1.0–2.3) 1.4 (0.9–2.1) 2.2 (1.6–3.1) 2.3 (1.6–3.2) 1.8 (1.4–2.4) 1.8 (1.4–2.4) 
Palaro 2.6 (1.1–6.0) 2.6 (1.1–6.0) 3.7 (1.6–8.7) 3.7 (1.6–8.7) 3.1 (1.7–5.6) 3.2 (1.7–5.9) 
Patiko 2.1 (0.8–5.3) 2.0 (0.7–5.2) 2.8 (1.1–7.2) 2.9 (1.1–7.6) 2.4 (1.2–4.7) 2.4 (1.2–4.8) 
Pece 0.9 (0.3–2.7) 0.9 (0.3–2.7) 2.7 (1.4–5.4) 2.7 (1.3–5.3) 1.7 (1.0–3.1) 1.7 (1.0–3.1) 

CI, confidence interval; TC, Town Council.

Fig. 1.

Map of the weighted and post-stratified prevalence of active epilepsy across 25 sub-counties in three districts of northern Uganda, 2010–2011, using data from 38,303 sampled individuals. TC, Town Council.

Fig. 1.

Map of the weighted and post-stratified prevalence of active epilepsy across 25 sub-counties in three districts of northern Uganda, 2010–2011, using data from 38,303 sampled individuals. TC, Town Council.

Close modal

Overall, the prevalence was found to be the lowest for those aged less than 10 years and highest for those between 10 years and less than 20 years of age, with the weighted and post-stratified estimates of 1.2% (95% CI: 1.0%–1.5%) and 6.8% (95% CI: 6.1%–7.5%), respectively. From the age group 20–29, the prevalence of epilepsy declined, reaching 2.1% (95% CI: 1.4%–3.0%) between the ages of 50–59, and rose again gradually in the older age groups, although estimates are less precise in these groups due to the limited number of elderly people in the study in line with Uganda’s population structure. Similar pattern by age group was observed among males and females, even though among the latter, the age group between 60 and 69 had the lowest prevalence (Table 4).

Table 4.

Unweighted and weighted and post-stratified prevalence estimates of active epilepsy (95% CI) by 10-year age groups and sex using data from 38,303 sampled individuals across 25 sub-counties in three districts of northern Uganda, 2010–2011

Age group, yearsFemaleMaleTotal
unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)
<10 0.9 (0.7–1.1) 1.0 (0.7–1.3) 1.4 (1.1–1.7) 1.4 (1.1–1.8) 1.1 (0.9–1.3) 1.2 (1.0–1.5) 
10–19 4.8 (4.3–5.4) 5.1 (4.4–5.9) 7.4 (6.7–8.2) 8.4 (7.3–9.5) 6.0 (5.5–6.5) 6.8 (6.1–7.5) 
20–29 4.7 (3.9–5.6) 4.6 (3.7–5.7) 6.3 (5.3–7.5) 5.9 (4.7–7.3) 5.4 (4.7–6.1) 5.2 (4.5–6.1) 
30–39 3.0 (2.3–4.0) 4.3 (2.9–6.6) 4.9 (3.8–6.2) 5.6 (4.2–7.6) 3.9 (3.2–4.6) 5.0 (3.9–6.4) 
40–49 1.5 (1.0–2.2) 1.4 (0.9–2.2) 4.9 (3.7–6.4) 4.6 (3.4–6.3) 2.8 (2.3–3.6) 2.9 (2.3–3.8) 
50–59 1.5 (1.0–2.4) 1.1 (0.7–1.9) 3.1 (2.1–4.6) 3.2 (2.0–5.2) 2.1 (1.6–2.9) 2.1 (1.4–3.0) 
60–69 0.6 (0.2–1.5) 0.7 (0.2–1.8) 4.0 (2.6–5.9) 3.8 (2.2–6.5) 2.1 (1.4–3.0) 2.3 (1.4–3.8) 
≥70 2.8 (1.2–6.6) 1.8 (0.7–4.5) 4.2 (1.9–9.1) 3.7 (1.6–8.5) 3.5 (1.9–6.1) 2.7 (1.4–5.1) 
Age group, yearsFemaleMaleTotal
unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)unweighted % (95% CI)weighted and post-stratified % (95% CI)
<10 0.9 (0.7–1.1) 1.0 (0.7–1.3) 1.4 (1.1–1.7) 1.4 (1.1–1.8) 1.1 (0.9–1.3) 1.2 (1.0–1.5) 
10–19 4.8 (4.3–5.4) 5.1 (4.4–5.9) 7.4 (6.7–8.2) 8.4 (7.3–9.5) 6.0 (5.5–6.5) 6.8 (6.1–7.5) 
20–29 4.7 (3.9–5.6) 4.6 (3.7–5.7) 6.3 (5.3–7.5) 5.9 (4.7–7.3) 5.4 (4.7–6.1) 5.2 (4.5–6.1) 
30–39 3.0 (2.3–4.0) 4.3 (2.9–6.6) 4.9 (3.8–6.2) 5.6 (4.2–7.6) 3.9 (3.2–4.6) 5.0 (3.9–6.4) 
40–49 1.5 (1.0–2.2) 1.4 (0.9–2.2) 4.9 (3.7–6.4) 4.6 (3.4–6.3) 2.8 (2.3–3.6) 2.9 (2.3–3.8) 
50–59 1.5 (1.0–2.4) 1.1 (0.7–1.9) 3.1 (2.1–4.6) 3.2 (2.0–5.2) 2.1 (1.6–2.9) 2.1 (1.4–3.0) 
60–69 0.6 (0.2–1.5) 0.7 (0.2–1.8) 4.0 (2.6–5.9) 3.8 (2.2–6.5) 2.1 (1.4–3.0) 2.3 (1.4–3.8) 
≥70 2.8 (1.2–6.6) 1.8 (0.7–4.5) 4.2 (1.9–9.1) 3.7 (1.6–8.5) 3.5 (1.9–6.1) 2.7 (1.4–5.1) 

CI, confidence interval.

Quantifying the prevalence of epilepsy is key to estimating its burden and assisting health authorities to prioritize prevention and control efforts. Moreover, measuring local variations in epilepsy prevalence is important to not only allocate sufficient resources where it is most needed, which is often among the most vulnerable populations, but also explore ecological associations with potential risk factors [26]. Our study estimated the population-level prevalence of active epilepsy in 25 sub-counties of Northern Uganda, an area facing greater deprivation than the Central, Eastern, and Western regions [27] and also endemic to T. solium [20, 28].

Our results showed that, after adjusting the data for sampling selection probability, to represent the target population, and post-stratification, to represent the source population, the prevalence estimate of active epilepsy was higher than the corresponding unweighted prevalence estimate. This may be explained by the fact that men were under-sampled in our study population but were more likely to suffer from epilepsy than females. Applying post-stratification allowed us to better represent the prevalence in the source population of the 25 sub-counties sampled where the proportion of males and females were similar (50%). This suggests that omitting adjusting for the under-representation or over-representation of certain groups (such as sex) in the analytic sample with respect to their distribution in the target population, may distort estimates of epilepsy prevalence, and calls for a cautious interpretation of several prevalence estimates of epilepsy reported in the literature.

Our estimated prevalence of active epilepsy in northern Uganda, even unadjusted, was among the higher values reported in the literature (please refer to introduction). Within the East African sub-region, to which Uganda belongs, substantial variation in prevalence estimates was observed across countries, ranging from 0.29% in Kenya to 4.1% in Rwanda [24, 29‒33]. Two studies have reported the prevalence of active epilepsy in Uganda (eastern Uganda 1.03% and western Uganda 1.3%), both resulting in lower estimates than ours [14, 34]. The approximately three times higher prevalence of active epilepsy in our study population may be due to NCC, which was around 8% in our study districts [35]. However, a house-to-house survey carried out in three districts of northern Uganda in 2017 by Gumisiriza et al. [36], including one of the districts (Moyo) that was also visited in our study, reported age-standardized epilepsy prevalence estimates of 4.6%, 5.1%, and 3.7% in Moyo, Kitgum and Pader districts, respectively, and confirmed the higher epilepsy prevalence in northern Uganda. Unfortunately, the authors did not specify whether they assessed active epilepsy or lifetime epilepsy. In Moyo district, the weighted and post-stratified prevalence of active epilepsy in our study ranged from 1.7% to 5.3%. In fact, the two villages targeted by the Gumisiriza et al. [36] study in 2017 in Moyo district, which were the village of Pakarukwe in Dufile sub-county and Pajakiri-North in Metu sub-county, showed an average age-standardized epilepsy estimate of 4.6%. In our study, we found a weighted and post-stratified prevalence of active epilepsy in Dufile sub-county of 3.3% and in Metu sub-county of 5.3%, which is not too far from the prevalence estimate reported by Gumisiriza et al. [36], although their study in Moyo was much smaller than ours (961 vs. 12,503 participants). However, we showed that there was a wide variation of prevalence estimates of active epilepsy across sub-counties, ranging from 1.7% to 6.6% across all three districts and all 25 sub-counties, which suggests a more nuanced risk factor profile and calls for micro-level analyses. Of note, sub-counties with higher prevalence estimates were rural and those with lower prevalence estimate urban, so that socioeconomic determinants seem to play a role.

Our study also showed a variation in prevalence estimates of active epilepsy with participant age, with a peak between the ages of 10 and 19. The prevalence of active epilepsy was lowest in people aged less than 10 years. The shape of the distribution of the prevalence of active epilepsy in our study was similar to that reported by GBD 2016 Epilepsy Collaborators [37], although this refers to the global prevalence of idiopathic epilepsy only and shows a peak at 5–9 years of age, followed by a very slight decrease, and then an increase from age 40–45. In the meta-analysis by Fiest et al. [7], the highest point prevalence estimate of active epilepsy globally was between 20 and 29 years. Thijs et al. [38] described the distribution of global epilepsy incidence to be bimodal, with a peak in children aged less than 1 year and another in people over the age of 50 years. In people over 50 years, incidence increases with age, reaching a maximum in those over 70 [38]. In our study, we have missed the early peak in infants and younger children as we were rather rigorous in excluding children with slightest doubt of suffering from febrile seizures only.

In our study, both unweighted and weighted and post-stratified epilepsy prevalence estimates were higher in men than in women, regardless of age group. In the literature, evidence on the sex distribution of epilepsy prevalence in SSA countries is inconsistent, with several studies suggesting a higher prevalence in males [37, 39], while others report higher frequency in females [40, 41]. A literature review by SSA reported large variation in the prevalence estimates of epilepsy across studies, ranging from 0.32% to 12.3% in men and 0.25% to 9% in women [42]. A meta-analysis published in 2017 reported a pooled point prevalence estimate of active epilepsy slightly higher in men (0.73%; 95% CI: 0.60–0.88) than in women (0.68%; 95% CI: 0.55–0.84) [7]. The specific biological mechanisms that may explain the sex differences in the prevalence of epilepsy are not completely understood [43‒46]. Some researchers suggested that those differences might be explained by the different risk factors, e.g., head trauma, and by the fact that women may conceal the disease because of social stigma [2, 39]. Some researchers suggested that those differences might be explained by the different risk factors, e.g., head trauma, and by the fact that women may conceal the disease because of social stigma [2, 39].

The strengths of this study include the use of a large random sample of residents from 25 sub-counties in three districts of northern Uganda, which increased the precision of estimates. In addition to including participants from multiple age groups, our analyses simultaneously adjusted for the sample’s selection probability and post-stratified the data to reflect the sex distribution of the source population of the 3 districts. Finally, by reporting a weighted and post-stratified prevalence estimate of epilepsy greater than the unweighted estimate, our study highlighted the need to adjust sampling weights to account for the unequal distribution of certain population subgroups, which could prevent bias in the epilepsy prevalence estimate.

Some limitations are also present. The study was conducted over 10 years ago and due to lack of resources the analysis, including the application of sampling weights with post-stratification, was majorly delayed. However, the socioeconomic situation of northern Uganda and in particular in the districts where the study took place has not changed over the last decade, and despite government interventions within the region, the area remains among the poorest regions in the country. According to the Uganda multidimensional poverty index report 2022, multidimensional poverty at the regional level is highest in the northern region, which stands at 63% [27]. Under those circumstances, it is very unlikely that epilepsy, including NCC control, has been improved. We therefore believe that the results of our study are still valid today. Also, our study gives very granular data on epilepsy prevalence estimates in SSA from one of the largest study populations, totaling 38,303 people in 25 sub-counties in three districts of northern Uganda, which is rather unique, and hence definitely contributes to the existing body of epilepsy literature in SSA. The low participant proportion reported in the Gulu district is another limitation of our study and may hamper the generalizability of active epilepsy prevalence estimates for the fifteen selected sub-counties in Gulu district to the entire population of these sub-counties. The lower participation rate in Gulu district can be explained by the urban structure compared to Moyo and Adjumani and the associated higher mobility of people, the fact that many people were still living in Internally Displaced People camps but left them during the day after the end of the war and the timing of the study during the harvest season. Also, our information was limited to the distribution of epilepsy according to age, sex, and geographical location (sub-counties). Extending such analysis to other potential risk factors, such as housing location (urban vs. rural) or household characteristics, including socioeconomic and environmental determinants, would provide further micro-level insight into the distribution of epilepsy prevalence in this population. Finally, we cannot rule out possible misclassification of epilepsy cases in our study, given the absence of a perfect diagnostic test for epilepsy [38].

In conclusion, accurate estimation of the health and socioeconomic impacts of epilepsy requires reliable epidemiological data on the extent of epilepsy in the region. Our study showed that active epilepsy is highly prevalent in northern Uganda, and the population-level prevalence estimates may be underestimated if the study does not correct for subgroup representation errors in the analytical sample. Active epilepsy was more prevalent in men than in women, and affected people of all ages, mainly children and young adults. Implementing programs and interventions targeting the control of local risk factors of epilepsy such as NCC and improving population health care access could help reduce the prevalence of active epilepsy in this area of the country.

Foremost, we would like to thank the study teams, including enumerators, nurses, clinical officers, doctors, administrative personnel, and laboratory personnel, without whom the study would not have been possible. Our heartfelt thank you also goes out to the many individuals who took part in the study, for their compliance, patience, and flexibility. We would also like to thank Prof. Emmanuel Ondongo-Aginya and Dr. Simon Peter Alarokol for maintaining the T. solium laboratory and Prof. Charles Karamagi for epidemiological advice.

Ethical approval was obtained from Gulu University Faculty of Medicine Institutional Research Committee (IRC) and Uganda National Council for Science and Technology (UNCST) (Ref. No. 543). District local authorities in all the 25 sub-counties were informed about the initiation of the research study and presented with the necessary research clearance. Administration clearance was then obtained from the district leadership. Local council (LC1) leaders of the various villages were informed and they, in turn, informed and mobilized the population for the study. Oral informed consent was obtained from the local village leaders. Informed consent was obtained from all study participants after explaining the purpose, procedures, benefits, and potential risks of participating in the study. Study participants indicated their consent orally and by endorsing pre-designed consent forms using their signatures or thumbprints in the presence of the investigators. Children above 8 years of age assented to participate in the study in the presence of their parents/guardians. Parents/guardians of children below 8 years of age or mentally challenged individuals consented for them.

The authors have no conflicts of interest to declare.

This study was funded by the DFG (German Research Foundation) within the research grant (WI 3427/1-1) “Neurocysticercosis in sub-Saharan Africa.” The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Joyce Moriku Kaducu: data collection, data analysis, writing – original draft, and writing – review and editing. Fiston Ikwa Ndol Mbutiwi: data analysis, writing – original draft, and writing – review and editing. Luise Keller, Gabriele Escheu, Peter Hauke, Bettina Pfausler, and Erich Schmutzhard: data collection, patient examination, and writing – review and editing. Veronika Schmidt and Emilio Ovuga: concept and design of the study, data collection, and writing – review and editing. Hélène Carabin: concept and design of the study, data analysis, writing – original draft, and writing – review and editing. Andrea S. Winkler: concept and design of the study, data collection, data analysis, writing – original draft, and writing – review and editing.

Additional Information

Hélène Carabin, Emilio Ovuga, and Andrea S. Winkler contributed equally to senior authorship.

Data will be made available upon request to the corresponding author.

1.
Shorvon
SD
.
The etiologic classification of epilepsy
.
Epilepsia
.
2011
;
52
(
6
):
1052
7
.
2.
Beghi
E
.
The epidemiology of epilepsy
.
Neuroepidemiology
.
2020
;
54
(
2
):
185
91
.
3.
GBD 2021 Nervous System Disorders Collaborators
,
Steinmetz
JD
,
Seeher
KM
,
Schiess
N
,
Nichols
E
,
Cao
B
.
Global, regional, and national burden of disorders affecting the nervous system, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021
.
Lancet Neurol
.
2024
;
23
(
4
):
344
81
.
4.
GBD 2016 Epilepsy Collaborators
,
Giussani
G
,
Nichols
E
,
Abd-Allah
F
,
Abdela
J
,
Abdelalim
A
.
Global, regional, and national burden of epilepsy, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016
.
Lancet Neurol
.
2019
;
18
(
4
):
357
75
.
5.
WHO
.
Epilepsy [Internet]
.
Geneva
:
World Health Organization
;
2024
. [cited 2024 February 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/epilepsy/?gad_source=1&gclid=EAIaIQobChMIjYuI2oKDhQMVIuXjBx2x-gxqEAAYASAAEgJd7_D_BwE
6.
Ngugi
AK
,
Bottomley
C
,
Kleinschmidt
I
,
Sander
JW
,
Newton
CR
.
Estimation of the burden of active and life-time epilepsy: a meta-analytic approach
.
Epilepsia
.
2010
;
51
(
5
):
883
90
.
7.
Fiest
KM
,
Sauro
KM
,
Wiebe
S
,
Patten
SB
,
Kwon
CS
,
Dykeman
J
, et al
.
Prevalence and incidence of epilepsy: a systematic review and meta-analysis of international studies
.
Neurology
.
2017
;
88
(
3
):
296
303
.
8.
Preux
PM
,
Druet-Cabanac
M
.
Epidemiology and aetiology of epilepsy in sub-Saharan Africa
.
Lancet Neurol
.
2005
;
4
(
1
):
21
31
.
9.
Ba-Diop
A
,
Marin
B
,
Druet-Cabanac
M
,
Ngoungou
EB
,
Newton
CR
,
Preux
PM
.
Epidemiology, causes, and treatment of epilepsy in sub-Saharan Africa
.
Lancet Neurol
.
2014
;
13
(
10
):
1029
44
.
10.
Owolabi
LF
,
Adamu
B
,
Jibo
AM
,
Owolabi
SD
,
Isa
AI
,
Alhaji
ID
, et al
.
Prevalence of active epilepsy, lifetime epilepsy prevalence, and burden of epilepsy in Sub-Saharan Africa from meta-analysis of door-to-door population-based surveys
.
Epilepsy Behav
.
2020
;
103
(
Pt A
):
106846
.
11.
Burneo
JG
,
Jette
N
,
Theodore
W
,
Begley
C
,
Parko
K
,
Thurman
DJ
, et al
.
Disparities in epilepsy: report of a systematic review by the north American commission of the international league against epilepsy
.
Epilepsia
.
2009
;
50
(
10
):
2285
95
.
12.
Banerjee
PN
,
Filippi
D
,
Allen Hauser
W
.
The descriptive epidemiology of epilepsy-a review
.
Epilepsy Res
.
2009
;
85
(
1
):
31
45
.
13.
Bharucha
N
,
Odermatt
P
,
Preux
PM
.
Methodological difficulties in the conduct of neuroepidemiological studies in low- and middle-income countries
.
Neuroepidemiology
.
2014
;
42
(
1
):
7
15
.
14.
Ngugi
AK
,
Bottomley
C
,
Kleinschmidt
I
,
Wagner
RG
,
Kakooza-Mwesige
A
,
Ae-Ngibise
K
, et al
.
Prevalence of active convulsive epilepsy in sub-Saharan Africa and associated risk factors: cross-sectional and case-control studies
.
Lancet Neurol
.
2013
;
12
(
3
):
253
63
.
15.
Ndimubanzi
PC
,
Carabin
H
,
Budke
CM
,
Nguyen
H
,
Qian
YJ
,
Rainwater
E
, et al
.
A systematic review of the frequency of neurocyticercosis with a focus on people with epilepsy
.
PLoS Negl Trop Dis
.
2010
;
4
(
11
):
e870
.
16.
Glasgow
G
.
Stratified sampling types
.
Encyclopedia Soc Meas
.
2005
;
3
:
683
8
.
17.
StataCorp
.
Stata 18 survey data reference manual
.
College Station
:
Texas: Stata Press
;
2023
[Internet] [cited 2024 February 20]. Available from: https://www.stata.com/bookstore/survey-data-reference-manual/
18.
Kulas
JT
,
Robinson
DH
,
Smith
JA
,
Kellar
DZ
.
Post-stratification weighting in organizational surveys: a cross-disciplinary tutorial
.
Hum Resour Manag
.
2018
;
57
(
2
):
419
36
.
19.
Millogo
A
,
Kongnyu Njamnshi
A
,
Kabwa-PierreLuabeya
M
.
Neurocysticercosis and epilepsy in sub-Saharan Africa
.
Brain Res Bull
.
2019
;
145
:
30
8
.
20.
Zulu
G
,
Stelzle
D
,
Mwape
KE
,
Welte
TM
,
Strømme
H
,
Mubanga
C
, et al
.
The epidemiology of human Taenia solium infections: a systematic review of the distribution in Eastern and Southern Africa
.
PLoS Negl Trop Dis
.
2023
;
17
(
3
):
e0011042
.
21.
Winkler
AS
.
Neurocysticercosis in sub-Saharan Africa: a review of prevalence, clinical characteristics, diagnosis, and management
.
Pathog Glob Health
.
2012
;
106
(
5
):
261
74
.
22.
Placencia
M
,
Sander
JW
,
Shorvon
SD
,
Ellison
RH
,
Cascante
SM
.
Validation of a screening questionnaire for the detection of epileptic seizures in epidemiological studies
.
Brain
.
1992
;
115
(
Pt 3
):
783
94
.
23.
Birbeck
GL
,
Kalichi
EM
.
Epilepsy prevalence in rural Zambia: a door-to-door survey
.
Trop Med Int Health
.
2004
;
9
(
1
):
92
5
.
24.
Winkler
AS
,
Kerschbaumsteiner
K
,
Stelzhammer
B
,
Meindl
M
,
Kaaya
J
,
Schmutzhard
E
.
Prevalence, incidence, and clinical characteristics of epilepsy: a community-based door-to-door study in northern Tanzania
.
Epilepsia
.
2009
;
50
(
10
):
2310
3
.
25.
Spittal
MJ
,
Carlin
JB
,
Currier
D
,
Downes
M
,
English
DR
,
Gordon
I
, et al
.
The Australian longitudinal study on male health sampling design and survey weighting: implications for analysis and interpretation of clustered data
.
BMC Public Health
.
2016
;
16
(
Suppl 3
):
1062
.
26.
Duggan
MB
.
Epilepsy in rural Ugandan children: seizure pattern, age of onset and associated findings
.
Afr Health Sci
.
2010
;
10
(
3
):
218
25
.
27.
UBOS
.
Multidimensional poverty index report 2022
.
Kampala
:
Uganda Bureau of Statistics
;
2022
[Internet] [cited 2024 February 20]. Available from: https://www.ubos.org/wp-content/uploads/publications/08_2022Multi_Poverty_Dimensional_Index_Report_2022.pdf
28.
Ngwili
N
,
Ahimbisibwe
S
,
Sentamu
DN
,
Thomas
LF
,
Ouma
E
.
Structure of the pork value chain in Northern Uganda: implications for Taenia solium control interventions
.
Front Vet Sci
.
2023
;
10
:
1177526
.
29.
Edwards
T
,
Scott
AG
,
Munyoki
G
,
Odera
VM
,
Chengo
E
,
Bauni
E
, et al
.
Active convulsive epilepsy in a rural district of Kenya: a study of prevalence and possible risk factors
.
Lancet Neurol
.
2008
;
7
(
1
):
50
6
.
30.
Almu
S
,
Tadesse
Z
,
Cooper
P
,
Hackett
R
.
The prevalence of epilepsy in the Zay Society, Ethiopia: an area of high prevalence
.
Seizure
.
2006
;
15
(
3
):
211
3
.
31.
Burton
KJ
,
Rogathe
J
,
Whittaker
R
,
Mankad
K
,
Hunter
E
,
Burton
MJ
, et al
.
Epilepsy in Tanzanian children: association with perinatal events and other risk factors
.
Epilepsia
.
2012
;
53
(
4
):
752
60
.
32.
Sebera
F
,
Munyandamutsa
N
,
Teuwen
DE
,
Ndiaye
IP
,
Diop
AG
,
Tofighy
A
, et al
.
Addressing the treatment gap and societal impact of epilepsy in Rwanda--Results of a survey conducted in 2005 and subsequent actions
.
Epilepsy Behav
.
2015
;
46
:
126
32
.
33.
Keller
L
,
Stelzle
D
,
Schmidt
V
,
Carabin
H
,
Reinhold
AK
,
Keller
C
, et al
.
Community-level prevalence of epilepsy and of neurocysticercosis among people with epilepsy in the Balaka district of Malawi: a cross-sectional study
.
PLoS Negl Trop Dis
.
2022
;
16
(
9
):
e0010675
.
34.
Kaiser
C
,
Kipp
W
,
Asaba
G
,
Mugisa
C
,
Kabagambe
G
,
Rating
D
, et al
.
The prevalence of epilepsy follows the distribution of onchocerciasis in a west Ugandan focus
.
Bull World Health Organ
.
1996
;
74
(
4
):
361
7
.
35.
Stelzle
D
,
Schmidt
V
,
Keller
L
,
Ngowi
BJ
,
Matuja
W
,
Escheu
G
, et al
.
Characteristics of people with epilepsy and Neurocysticercosis in three eastern African countries-A pooled analysis
.
PLoS Negl Trop Dis
.
2022
;
16
(
11
):
e0010870
.
36.
Gumisiriza
N
,
Mubiru
F
,
Siewe Fodjo
JN
,
Mbonye Kayitale
M
,
Hotterbeekx
A
,
Idro
R
, et al
.
Prevalence and incidence of nodding syndrome and other forms of epilepsy in onchocerciasis-endemic areas in northern Uganda after the implementation of onchocerciasis control measures
.
Infect Dis Poverty
.
2020
;
9
(
1
):
12
.
37.
GBD 2016 Epilepsy Collaborators
.
Global, regional, and national burden of epilepsy, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016
.
Lancet Neurol
.
2019
;
18
(
4
):
357
75
.
38.
Thijs
RD
,
Surges
R
,
O’Brien
TJ
,
Sander
JW
.
Epilepsy in adults
.
Lancet
.
2019
;
393
(
10172
):
689
701
.
39.
Bharucha
NE
,
Bharucha
EP
,
Bharucha
AE
,
Bhise
AV
,
Schoenberg
BS
.
Prevalence of epilepsy in the parsi community of Bombay
.
Epilepsia
.
1988
;
29
(
2
):
111
5
.
40.
Mung’ala-Odera
V
,
White
S
,
Meehan
R
,
Otieno
GO
,
Njuguna
P
,
Mturi
N
, et al
.
Prevalence, incidence and risk factors of epilepsy in older children in rural Kenya
.
Seizure
.
2008
;
17
(
5
):
396
404
.
41.
Dupont
F
,
Trevisan
C
,
Moriku Kaducu
J
,
Ovuga
E
,
Schmidt
V
,
Winkler
AS
, et al
.
Human health and economic impact of neurocysticercosis in Uganda
.
Trop Med Int Health
.
2022
;
27
(
1
):
99
109
.
42.
Dedeken
P
,
Sebera
F
,
Mutungirehe
S
,
Garrez
I
,
Umwiringirwa
J
,
Van Steenkiste
F
, et al
.
High prevalence of epilepsy in Northern Rwanda: exploring gender differences
.
Brain Behav
.
2021
;
11
(
11
):
e2377
.
43.
Mao
Y
,
Ahrenfeldt
LJ
,
Christensen
K
,
Wu
C
,
Christensen
J
,
Olsen
J
, et al
.
Risk of epilepsy in opposite-sex and same-sex twins: a twin cohort study
.
Biol Sex Differ
.
2018
;
9
(
1
):
21
.
44.
Ziemka-Nalecz
M
,
Pawelec
P
,
Ziabska
K
,
Zalewska
T
.
Sex differences in brain disorders
.
Int J Mol Sci
.
2023
;
24
(
19
):
14571
.
45.
Kight
KE
,
McCarthy
MM
.
Using sex differences in the developing brain to identify nodes of influence for seizure susceptibility and epileptogenesis
.
Neurobiol Dis
.
2014
.
72
(
Pt B
);
136
43
.
46.
Taubøll
E
,
Sveberg
L
,
Svalheim
S
.
Interactions between hormones and epilepsy
.
Seizure
.
2015
;
28
:
3
11
.