Provision of Surgical Care for Children Across Somaliland: Challenges and Policy Guidance

Existing data suggest a large burden of surgical conditions in low- and middle-income countries (LMICs). However, surgical care for children in LMICs remains poorly understood. Our goal was to define the hospital infrastructure, workforce, and delivery of surgical care for children across Somaliland and provide policy guidance to improve care. We used two established hospital assessment tools to assess infrastructure, workforce, and capacity at all hospitals providing surgical care for children across Somaliland. We collected data on all surgical procedures performed in children in Somaliland between August 2016 and July 2017 using operative logbooks. Data were collected from 15 hospitals, including eight government, five for-profit, and two not-for-profit hospitals. Children represented 15.9% of all admitted patients, and pediatric surgical interventions comprised 8.8% of total operations. There were 0.6 surgical providers and 1.2 anesthesia providers per 100,000 population. A total of 1255 surgical procedures were performed in children in all hospitals in Somaliland over 1 year, at a rate of 62.4 surgical procedures annually per 100,000 children. Care was concentrated at private hospitals within urban areas, with a limited number of procedures for many high-burden pediatric surgical conditions. We found a profound lack of surgical capacity for children in Somaliland. Hospital-level surgical infrastructure, workforce, and care delivery reflects a severely resource-constrained health system. Targeted policy to improved essential surgical care at local, regional, and national levels is essential to improve the health of children in Somaliland.


Introduction 1 2
Although the global health agenda increasingly recognizes the importance of surgical care in low-and 3 middle-income countries (LMICs), [1][2][3] surgical care delivery for children remains poorly understood. Existing 4 data suggests a large burden of surgical conditions in children in many LMICs. [4][5][6][7] Although precise data on the 5 burden of surgical conditions in children in LMICs is limited, existing data suggest that the global burden of 6 surgical disease in children is high, with congenital anomalies and injuries in children comprising a large 7 proportion of the overall disease burden. [4,[8][9][10][11][12] However, detailed data on the delivery of surgical care for 8 children remains limited due to lack of high-quality data, reliance on small cohort studies, and focus on urban 9 areas.[5,6,10, [13][14][15][16][17][18] Additionally, studies which focus on government health facilities often underreport the large 10 portion of surgical care which is provided at private facilities in many LMICs. [10,13,19-22] 11 Somaliland faces many challenges as it rebuilds its health care system following long-standing civil 12 conflicts, and is the fourth poorest country in the world.
[23] Using a national community-based survey, our group 13 recently revealed a high prevalence of surgical conditions in children across the country (12.2%), translating to 14 approximately 256,000 children nationally who require surgical care. [24] The goal of our current study is to 15 assess hospital-level infrastructure, human resources, and delivery of surgical care for children across Somaliland. 16 To examine hospital infrastructure, we used two standardized assessment tools to define each institution's 17 capacity, workforce, and surgical caseload. To measure the delivery of surgical care, we surveyed case logs from 18 all hospitals across Somaliland and summarized data on pediatric surgical procedures. In conjunction with our 19 previous burden of disease analysis [24]

15
Hospitals were included in this study if they had the capacity to perform surgery, which was defined as 16 the presence of at least one operating room. Although there is no national registry of public and private hospitals 17 in the country, a total of 16 hospitals with surgical capacity for children across the country were identified based 18 on information from local collaborators and conversations with Ministry of Health officials. Hospitals which had 19 no surgical capacity were excluded. One private hospital, International Hospital in Hargeisa, chose not to 20 participate in this study and was also excluded. Hospitals that performed surgical procedures on adults but had not 21 performed procedures on children in the past one year were still included in hospital capacity analysis in order to 22 understand the potential surgical capacity across these hospitals (Figure 1). 23 Hospital Infrastructure, Workforce, and Capacity Survey 1 We combined the World Health Organization's (WHO) Surgical Assessment Tool -Hospital 2

Walkthrough[31] and the Global Initiative for Children's Surgery's (GICS') Global Assessment in Pediatric 3
Surgery (GAPS) [32] to create a single-page hospital assessment tool to assess surgical infrastructure and capacity 4 (Supplementary File). Surveys were completed by either the lead surgeon or administrator at each hospital. The 5 survey included data on hospital infrastructure, workforce, service delivery, financial indicators, and payment 6 methods for patients. [33] To account for the reality that the same providers work at multiple hospitals,[34,35] on-7 site coordinators compiled a list of surgical and anesthesia providers and providers were associated with their 8 primary hospital of work. We chose not to have an external definition of provider, as providers in Somaliland 9 often have different levels and years of training and the locally reported classification more accurately represents 10 the level of care they provide. 11

12
To assess the delivery of surgical care for children, we reviewed all hospital-based surgical records from 13 surgical logbooks for all procedures on children ages 0 to 15 years who underwent surgical care between August 14 1, 2016 and July 31, 2017. Non-surgical inpatient medical records and surgical outpatient records were not 15 included. We collected data on patient age, gender, surgical diagnosis; surgical procedure(s) performed, and date 16 of surgery. As data was collected only from surgical logbooks, it did not include admission information, 17 comorbidities, complications, or outcome. 18

19
Responses to the hospital capacity assessment were aggregated using Microsoft Excel (Microsoft Corp, 20 Redmond, WA). Surgical diagnoses were coded by both a research assistant and a research manager, with 21 diagnostic codes reviewed and classified into nine surgical specialties by a pediatric surgeon. Of the 15 hospitals 22 included in the analysis, 11 hospitals had complete records available for 12 months, while 4 hospitals had between 23 1 and 8 months of data. Missing data for the child's age, gender, town of origin, type of anesthesia, surgical 24 provider, and anesthesia provided were accounted for through multiple imputations using multivariate normal 25 distribution using the SAS Markov Chain Monte Carlo computation. Data were summarized using frequency 1 tables and median values with associated interquartile ranges (IQR). Child demographics, surgical procedure 2 categories, and surgical diagnoses were stratified by hospital as well as by region and hospital type (private for-3 profit, not-for-profit, or government). Surgical indicators that are NOT child specific are reported "per 100,000 4 population" and surgical indicators that ARE child specific are reported "per 100,000 children". P-values were 5 determined using the Mantel-Haenszel Chi-Square test statistic. Data were analyzed in SAS 9.4 (SAS, Cary, NC) 6 and Microsoft Excel 2010 (Microsoft Corp, Redmond, WA). 7

Ethical Considerations
8 Institutional Review Board (IRB) approval was granted from Duke University. Since Somaliland does not 9 have a national IRB, a letter of approval for the study was obtained from the Somaliland Ministry of Health. 10

12
A total of 1,255 surgical procedures were performed in children over one year at all 15 hospitals in 13 Somaliland. There were 0.6 surgical providers and 1.2 anesthesia providers per 100,000 population, and hospital 14 characteristics and surgical provision varied greatly between hospitals and regions. 15

16
Of the 15 hospitals included in this analysis, eight are government, five private for-profit, and two private 17 not-for-profit hospitals. Of the eight government hospitals, one is a national referral hospital (Hargeisa Group  18 Hospital), six are regional hospitals, and one is a district hospital (Figure 2). Overall, children made up 15.9% of 19 all patients admitted monthly. All hospitals reported having electricity, running water, internet, phone service 20 more than 75% of the time, while oxygen was available more than 50% of the time (Supplementary File). surgical workforce had a total number of 23 surgical (1 pediatric surgeon) and 47 anesthesia providers across the 23 country, translating to a workforce density of 0.6 and 1.2 (per 100,000 population) (Figure 3). Over half of all the 24 surgical and anesthesia providers were located at hospitals in Maroodi Jeex. There was a total of 42 operating 25 rooms across all hospitals in Somaliland, with hospitals in Maroodi Jeex having the highest number per hospital 1 (3.8). (Figure 3). 2

3
Over half of all procedures were performed on children over 5 years old (52.8%) and only 0.2% of 4 procedures were performed by a pediatric surgeon. We classified the 1,255 surgical procedures, general surgery 5 (22.9%), otolaryngology (19.4%), orthopedic surgery (15.8%), ophthalmology (15.1%), and neurosurgery 6 (11.3%). Slightly over half of the procedures were elective (58.6%). Of all surgical procedures reported, the most 7 common were tonsillectomy (18.3%), various procedures for traumatic injury (11.2%), and shunting for 8 hydrocephalus (7.5%) ( Table 1). There were large variations in the ratio between elective/emergency procedures 9 between regions; with more urban regions had elective cases while more rural regions having more emergent 10 cases ( Table 2). There was also a difference between age groups and urgency ratios (p< 0.003), with children 0-11 28 days and 1-12 months old receiving primarily emergency procedures (60.7% and 67.5%, respectively) rather 12 than elective procedures (Figure 4). 13 The most common procedure within each surgical specialty category was trauma in general surgery 14 (19.9%), tonsillitis in otolaryngology (94.7%), fracture within orthopedics (40.1%), cataract within 15 ophthalmology (44.4%), hydrocephalus within neurology (66.2%), cleft lip within plastic surgery (46.4%), 16 circumcision within urology (41.1%) and prolonged labor in obstetrics (100%) ( Table 3). Of the 1,255 procedures 17 in children, 80% were performed at private hospitals (private for-profit 28.3%, private not-for-profit 51.5%) and 18 only 20% at public hospitals. (Table 4). There were differences in age groups receiving surgery (p<0.001) 19 between hospital types, with over half of children at not-for-profit hospitals being 5 years or younger (53.0%) 20 while for-profit and public hospitals had less than half of children in this age group (44.8% and 35.9% 21 respectively). There was also a significant difference between hospital types in the emergency/elective ratio 22 (p<0.001) with private for-profit and private not-for-profit hospitals performing mostly elective surgeries (72.8% 23 and 61.7% respectively), while public hospitals performed more emergency cases (68.0%) ( Table 4). 24

1
As Somaliland rebuilds its health care system following civil conflict, there is a profound lack of surgical 2 capacity for children. Through use of a national hospital-based survey, we found that only 1,255 pediatric surgical 3 procedures were performed over one year in Somaliland, translating to a rate of 62.4 surgical procedures per 4 100,000 children annually. Although the optimal rate of pediatric surgical procedures for children in LMICs has 5 not been defined, this rate of surgical procedures is far below the minimum recommended operative volume of other LMICs. [6,7,15,20,24,[41][42][43][44][45][46] A focus on essential surgical procedures can serve to develop surgical 2 infrastructure and capacity.[36,47] Although such "bellwether procedures" [33] have not been defined for children 3 and might differ among pediatric surgical subspecialties, they likely include care for conditions such as inguinal 4 hernia, congenital anomalies, and traumatic injuries.
[32] 5 There are several limitations to our study, many which are common to the use of hospital-based surveys. 6 First, although the use of a single-page hospital survey limited our ability to collect granular data on hospital 7 infrastructure and capacity, it reduced the burden of data collection and allowed us to travel to remote hospitals 8 and complete the survey in a timely manner. Additionally, this single-page assessment has not been validated on 9 its own. Validation of a low-burden hospital assessment tool is an important step to quantify hospital capacity on 10 a large scale. Secondly, little information was available in the surgical logbook records other than name, age, 11 procedure, and name of the surgeon or physician. Therefore, we were not able to capture data on clinical 12 outcomes, quality of care provided, or severity of the conditions prior to surgery. Additionally, we were not able 13 to collect any information on surgical procedures performed outside of the operating theater. Further studies 14 would need to rely on expanded data collection to enhance data quality and information, such as detailing health 15 seeking behavior, surgical complications, costs of care, and other outcomes.
[48] 16 In summary, we found that the number of surgical procedures performed in children across Somaliland is 17 low and insufficient for the burden of surgical disease. Surgical care is largely absent in rural regions, and the 18 types of surgical procedures are skewed to low burden conditions (such as tonsillitis) and not reflective of the 19 burden of surgical conditions in the population, particularly the high burden conditions such as congenital 20 anomalies. This study is the first of its kind to systematically assess the delivery of pediatric surgical care across 21 Somaliland and complements our prior analysis of the burden of pediatric surgical conditions across 22 Somaliland. [24,49] 23 We offer several policy recommendations to address these gaps in surgical care for children: 24 • A national health care plan, such as NSOAPs, should include strategic support for basic surgical 25 services addressing the health needs of children 26 • Implementation of a set of essential surgical procedures at district and regional hospitals is "low-1 hanging fruit" to increase access to surgical care, particularly in rural areas of Somaliland, in 2 conjunction with advocacy and strong collaboration with in-country government and health 3 officials. 4 • Expansion of surgical capacity, manpower, and infrastructure at all public hospitals across 5 Somaliland is required to support a comprehensive health system for the surgical care of children 6 7 ACKNOWLEDGEMENTS 1 We want to thank the Global Initiative for Children's Surgery (GICS) for its support of this work. GICS 2 (www.globalchildrenssurgery.org) is a network of children's surgical and anesthesia providers from low-, middle-, 3 and high-income countries collaborating for the purpose of improving the quality of surgical care for children 4 globally. 5 6 7 No table of figures entries found.