Quantifying the pediatric surgical need in Uganda: results of a nationwide cross-sectional, household survey

Over 50 % of Uganda’s 34.9 million inhabitants are under 18 years of age. The country has the third highest birth rate in the world at 44 births per 1000 population [15]. Only 8.0 % of the nation’s GDP is spent on health compared with 17.9 % in the United States and 9.4 % in the United Kingdom [16]. Uganda is organized into 11 sub-regions and the capital city of Kampala. The majority of health care is delivered at a district hospital level, which is able to perform basic emergency surgical procedures, including Cesarean sections and hernia repairs. There are 15 referral hospitals, which deliver a higher level of care and two national referral hospitals, one which can deliver specialized surgical care, while the second one focuses primarily on psychiatric disorders [17]. A few mission and non-governmental organization hospitals also provide elective specialty surgical care.

Using the SOSAS survey, we completed a two-stage, cluster-based nationwide, household survey, as previously described [18]. In short, the Uganda Bureau of Statistics assisted in randomly selecting 105 enumeration areas (EA), clustered by geographic sub-region. In each EA, 24 households were randomly selected from a list of all households within the EA. In each household, two individuals were randomly selected to complete the SOSAS survey, independent of age or gender. The SOSAS survey contains: (1) a household component gathering household demographics and information on any deaths that occurred in the household in the previous year and (2) an individual component querying for surgical conditions occurring in each anatomic location. The SOSAS Uganda team used the Performance Monitoring and Evaluation 2020 (PMA2020) Uganda team for data collection. PMA2020 is a survey team with trained enumerators, already living in each EA who speak each local language, as well as English. The SOSAS team trained PMA2020 enumerators on the SOSAS survey and how to identify surgical conditions.

Makerere University School of Medicine Research and Ethics Committee, Duke University Health System Institutional Review Board, and University of Minnesota Institutional Review Board approved this study prior to implementation. Enumerators obtained informed consent from each head of household and each individual participating in the survey, prior to initiation of the survey. For children less than 18, a parent or guardian provided informed consent, and children 8–18 provided assent. When necessary, parents or guardians assisted children in answering survey questions.

Children were defined as survey respondents between the ages of 0 and 18 years. Because our enumerators were not medically trained, six blinded surgical trainees and surgeons assessed each documented condition and assigned a rating on a defined four-point scale [0-definitely not surgical, 1-likely not surgical, 2-likely surgical, and 3-definitely surgical (received surgery)]. For discordant coding, the case was discussed, and a final score was assigned based on consensus. Cases with a score of 2 or 3 were included as surgically treatable conditions. In addition, each case was coded as treated or untreated based on whether the patient received appropriate surgical care. The lifetime prevalence of pediatric surgical conditions included all children with a surgically treatable condition and the unmet pediatric surgical need included those with an untreated surgically treatable condition. We analyzed the data using SAS 9.4 (SAS Institute Inc, Cary, NC, USA) and stored the data in spreadsheets on Microsoft Excel 2010 (Microsoft Corp, Redmond, WA, USA). Household and individual cases were weighted using design weights for each EA, household-level and individual-level response rates, and known population counts of gender and age groupings from the Uganda Census 2014 data. Demographic data and mortality estimates were compared between children who reported a surgical condition and those that did not report a surgical condition and between children with treated surgical conditions and those with untreated surgical conditions, using a weighted model. Childhood prevalence of surgical conditions was mapped by district, stratified by age group, using QGIS version 2.8 [19]. Clinical characteristics among children who reported a surgical condition were stratified and compared by age group and treatment status, using a weighted model. Proportions and frequencies described pediatric household deaths by age group, anatomic location, and the type of condition.

SOSAS Uganda interviewed a random sample of 4248 individuals (97.1 % response rate) in 2315 households (96.4 % response rate) with a total of 11,148 household members. Of the 4428 respondents, 2176 (49.1 %) were 18 years of age or less (Table 1). Of pediatric respondents, 305 (14.0 %) reported having a surgical condition at some point in their life, of which 51.1 % were untreated (156; 7.2 % of total). The median age of the children interviewed was 6.1 (IQR 2.5, 11.4) years, and the 0–5 year age group was the most prevalent. Among children interviewed, about half (50.2 %) were female and the majority (81.8 %) lived in a rural area. Most children reported being healthy in the 12 months prior to the interview; however, those with surgical conditions were more likely to report poor health compared with those without surgical conditions (25.3 versus 9.2 %, respectively; p < 0.0001). Children with surgical conditions made more visits to health facilities in the 12 months prior to the interview than children without surgical conditions (p = 0.0005). No differences were noted between children with surgical conditions and those without surgical conditions in regard to transportation time, wait time, and transportation costs to reach a primary, secondary, or tertiary care facility. Of children with surgical conditions, children with untreated conditions trended towards being younger and living in an urban area (p = 0.14 and p = 0.06, respectively). Children with untreated surgical conditions reported being less healthy (p < 0.0001) and making more frequent visits to health facilities (p < 0.0001). They were also more likely to have longer transportation times to primary care facilities (17.9 versus 9.5 min, p = 0.01). There was no difference in cost of transport to care facilities when comparing treatment status.

Overall, pediatric surgical conditions had similar prevalence among the Central, Northern, and Western regions, but were less prevalent in the Eastern region. The geographic distribution of pediatric surgical conditions varied by age group (Fig. 1). Most of the surgical conditions among children less than 5 years were concentrated in the Central and Western regions of Uganda, while surgical conditions among 6 to 9 years old were more prevalent in the Western and Northern regions. Surgical conditions among older children were more evenly dispersed throughout Uganda. Although not statistically significant, children with treated surgical conditions trended towards residing in the Central region, compared to those with untreated surgical conditions, who were more likely to live in the Northern region (p = 0.14).

A total of 305 children reported 356 surgical conditions (Table 2). Overall, the most common anatomical areas were the face and neck (36.2 %) and the extremities and back (31.7 %). Younger children (less than 9 years) trended toward more face and neck conditions, and older children (greater than 9 years) had more extremity and back conditions (p = 0.066). The majority of children older than 5 years of age had been living with their condition for over a year. Masses (23.8 %), wounds due to injury (19.7 %), and acquired deformities (16.6 %) were the most common type of surgical conditions. The type of surgical condition varied by age group. Among children less than 6 years of age, masses and burns were most common, but among children older than 6 years, wounds, masses, and acquired deformities were more common (p = 0.058).

The majority of children sought formal healthcare (74.2 %, 264/356), and only 19.7 % sought care from a traditional healer (69/356). Of those who sought formal care, 71.2 % underwent a minor procedure and 17.0 % received no care. Of those who sought care, but did not receive any, the primary reasons were lack of money (57.1 %), no perceived need for surgical care (25.0 %), and lack of available personnel and equipment (16.1 %). Of those who did not seek care, the primary reasons included lack of money (51.1 %), no perceived need (28.3 %), and fear (20.7 %).

Of the 356 surgical conditions, 174 (48.9 %) were untreated (Table 3). The majority of abdomen and chest conditions were untreated (71.9 and 72.0 %, respectively). Nearly, all congenital deformities remained untreated (91.7 %), and masses and abdominal problems were less likely to be treated compared with burns and wounds (p < 0.0001). Those with untreated conditions had higher rates of traditional healer use (25.9 versus 13.2 %, p = 0.002) and lower rates of seeking formal health care (54.6 versus 92.9 %, p < 0.0001). Children with untreated conditions who did seek health care cited lack of money as the primary reason for not receiving care.

Of 90 reported pediatric deaths, 31.1 % (28) were due to a surgical condition. The majority of deaths were male (71.4 %) and less than 6 years of age (57.1 %). Stillbirth and acquired deformity were the most common causes of death. Over half of deaths occurred at home, but 92.8 % sought health care prior to death. Of those who sought health care, 69.3 % underwent either a major procedure or minor procedure, but 23.0 % did not receive any care. Lack of money and time were the most frequent reasons for not receiving care.