Global Surgery indicators and pediatric hydrocephalus: a multicenter cross-country comparative study building the case for health systems strengthening

Kevin Paul Ferraris (  kpferraris@gmail.com ) Jose R Reyes Memorial Medical Center https://orcid.org/0000-0002-0919-6658 Eric Paolo Palabyab José R. Reyes Memorial Medical Center: Jose R Reyes Memorial Medical Center Sergei Kim FSBI Federal Neurosurgical Center of the Ministry of Health of the Russian Federation: FGBU Federal'nyj centr nejrohirurgii Ministerstva zdravoohranenia RF Hideaki Matsumura University of Tsukuba School of Medicine: Tsukuba Daigaku Igaku Iryokei Maria Eufemia Yap ThinkWell Global Venus Oliva Cloma-Rosales 101Health Research German Letyagin FSBI Federal Neurosurgical Center of the Ministry of Health of the Russian Federation: FGBU Federal'nyj centr nejrohirurgii Ministerstva zdravoohranenia RF Ai Muroi University of Tsukuba Faculty of Medicine: Tsukuba Daigaku Igaku Iryokei Ronnie Baticulon University of the Philippines Manila College of Medicine Jose Carlos Alcazaren Jose R Reyes Memorial Medical Center Kenny Seng University of the Philippines Manila College of Medicine Joseph Erroll Navarro Jose R Reyes Memorial Medical Center


Introduction
The Global Surgery movement has ushered in an awareness of the existing inequities in surgical care the world over [1][2][3][4]. A call was made for nation-level solutions in improving access to essential surgery, in order to ultimately achieve health, welfare, and economic development by year 2030 [1]. The Global Surgery indicators-with corresponding working de nitions and targets for countries-became the standardized metrics by which the extent of the problem in a country's healthcare system can be de ned and elucidated in relation to surgical processes and outcomes [5]. These six core indicators are: 1) access to timely essential surgery, 2) specialist surgical workforce density, 3) surgical volume, 4) perioperative mortality, and protection against 5) impoverishing and 6) catastrophic expenditures.

Outcome Disparities Borne from Country of Origin and Social Determinants
Few studies have examined the association between patient-level socioeconomic factors and outcomes following treatment for pediatric hydrocephalus [6-8]-and these studies are mostly situated in a single country. And yet, neurosurgical outcomes are also affected by systemic factors particularly the availability or absence of resources and investments within a health system. Cross-country disparities in patient outcomes are especially apparent in diseases like stroke and cancer [9,10]. The worse outcomes appear to be explained not only by patient factors but also by the perioperative care systems situated in a given hospital or country. The level of functionality of a country's health system -which depends on economic infrastructure [11], public policy [12,13], responsive governance [14] and effective nancing [15] arrangements-can in turn mitigate barriers and increase access to neurosurgical care. In this respect, the Global Surgery indicators can become useful in assessing and providing insight into the effectiveness and responsiveness of the surgical system in a neurosurgical center of a given country.

Surgery for Pediatric Hydrocephalus as Situated in Country Contexts
LMICs have a disproportionately larger case volume of pediatric hydrocephalus than high-income countries (HICs), owing to differences in crude birth rate and incidence of congenital and post-infectious etiologies [11,16]. The greater burden of this disease in LMICs is further compounded by the fact that access to neurosurgical care and resources for health service delivery are limited [11]. As CSF diversion surgery for hydrocephalus is considered a highly equitable and cost-effective bellwether procedure [17,18], exploring the nation-level barriers to this kind of care can provide a snapshot of the functioning of a country's healthcare and surgical systems-and give insight to the disparities in outcomes of patients undergoing the procedure. Using the lens of nation-level social determinants of health and the framework of Global Surgery indicators, we aim to determine whether differences in outcomes exist between speci c institutions and countries of varying income levels in relation to the neurosurgical management of allcause pediatric hydrocephalus.

Study Setting and Population
After approval by the institutional review boards from the three participating centers, we conducted a retrospective, cross-sectional study across countries of differing income levels. The study was conducted in three neurosurgical centers that are non-children's hospitals: 1) the University of Tsukuba Hospital, in Tsukuba, Ibaraki, Japan, 2) Jose R. Reyes Memorial Medical Center, in Manila, Philippines, and 3) the Federal Center of Neurosurgery, Federal State Budget Institution, in Novosibirsk, Russian Federation. The population included in the study were pediatric patients admitted in a period between January 1, 2019 to December 31, 2019 at the three centers, and who had either obstructive or communicating hydrocephalus with etiologies including the following, but not limited to: infantile-posthemorrhagic, post-infectious, congenital-structural, associated or related to central nervous system tumors, or associated with neural tube defects like myelomeningocele. They should have undergone the minimum operation of CSF diversion, that were any one of the following: ventriculoperitoneal shunt (VPS) insertion with or without revisions, external ventricular drain (EVD) insertion, endoscopic third ventriculostomy (ETV), Ommaya reservoir insertion, or combinations thereof. Children with hydrocephalus but who underwent another surgery without or other than CSF diversion were excluded.

Study Variables and Other Data
Patient-level variables consisted of clinical characteristics of individual patients collected from chart review. The outcome variables were patient-level covariates 1) time from admission-to-surgery and 2) perioperative mortality, while the explanatory variables included patient age, sex, socioeconomic status, type of hydrocephalus, timing of CSF diversion, and type of CSF diversion surgery. The time to CSF diversion surgery was the outcome of interest for cross-country comparison of institutions and was subdivided into descriptive categories based on the reasonable timeframes of hydrocephalus management by Mansouri and colleagues [19]. Nation-level data consisted of country-speci c metrics of health-system functions based on Global Surgery indicators, as well as surrogate measures of the economy and growth. Secondary data from the World Bank [20], literature on global health [21][22][23][24], and other studies of each country's health systems [25][26][27] were obtained for comparison. However, these nation-level data were not subjected to the statistical analyses.

Statistical Analysis
Descriptive statistics was used to summarize the general and clinical characteristics of the participants. Frequency and proportion were used for categorical variables. Shapiro-Wilk test was used to determine the normality distribution of continuous variables. Continuous quantitative data that did not meet the normality assumption of distribution were described using median and range. For timing of CSF diversion surgery as a time-to-event variable, Cox proportional hazards model was used to plot the Kaplan-Meier curves. For perioperative mortality as a dichotomous variable, logistic regression was done. The nal results from regression were presented as hazard ratios and odds ratios with their associated con dence intervals. Missing data were neither replaced nor estimated. Null hypothesis was rejected at 0.05α-level of signi cance. Stata 15.0 (StataCorp, College Station, TX, US) was used for data analysis.

Clinical and Socioeconomic Characteristics
The medical records of 159 children-99 (62%) from the Philippines, 47 (30%) from Russia and 13 (8%) from Japan-were reviewed (Table 1). Their median age was 2.4 (range 0-17) years, with proportions of sexes roughly similar. Most children in the Philippines belonged to poor households (77%), whereas 72% and 100% of the Russian and Japanese patients came from non-poor families, respectively. Almost 6 in 10 hydrocephalus cases in the Philippines were either post-infectious, congenital, or tumor-related. About 7 in 10 of those in Russia were infantile/post hemorrhagic (49%) or tumor-related (21%). Two-thirds of the cases from Japan were congenital (38%) or infantile/post hemorrhagic (23%) hydrocephalus. VPS insertion was the most common procedure for CSF diversion in the Philippine center (92%). Likewise in the Russian center, VPS was the surgery for the majority (64%), while varied methods of CSF diversion were done in the Japanese center. The revision rates were 7.7%, 11.1%, and 25.5% in the Japanese, Philippine, and Russian centers, respectively.

Primary Outcomes: Timing of CSF Diversion Surgery and Mortality
For timeliness of access to care, majority of the cases from the Russian (85%) and Japanese (70%) centers were able to access CSF diversion within 2 days of con nement. In the Philippine center, only 36% of patients were able to undergo surgery by the second hospital day. The median time to CSF diversion was 6 days in the Philippine center, whereas it was 1 day for both the Japanese and Russian centers.
(Table 2, Figure 1). The delivery of surgical care for pediatric hydrocephalus was signi cantly more e cient in the Russian center compared to that in the Philippine center (HR=2.94, 95%CI 1.99-4.35, p<0.001). In terms of all-cause mortality, three children died in the Philippine cohort (proportion of 3.03, 95%CI 0.63-8.60), but none in the Russian and Japanese cohorts ( Table 2).

Patient-Level Determinants of Primary Outcomes
For the cohort from the Philippines, non-poor patients were more likely to receive CSF diversion at an earlier time (HR=4.74, 95%CI 2.34-9.61, p<0.001). In the same center, those with infantile or posthemorrhagic hydrocephalus (HR=3.72, 95%CI 1.70-8.15, p=0.001) were more likely to receive CSF diversion earlier compared to those with congenital hydrocephalus, and those with post-infectious (HR=0.39, 95%CI 0.22-0.70, p=0.002) or myelomeningocele-associated hydrocephalus (HR=0.46, 95%CI 0.22-0.95, p=0.037) were less likely to undergo surgery at an earlier time. For Russia, older patients were more likely to receive or require early CSF diversion (HR=1.07, 95%CI 1.01-1.14, p=0.035). For Japan, evidence was insu cient to identify signi cant factors associated with the timing of CSF diversion (Table   3). Because the cohorts of Japanese and Russian children had zero mortality, only the cohort of Filipino children was analyzed for logistic regression (Table 4). EVD insertion was found to be signi cantly associated with mortality (cOR 14.45, 95% CI 1.28-162.97, p = 0.031).

Other Global Surgery Indicators and Social Determinants of Health
In terms of metrics that represent health-system indicants, the various indices appear to depend and follow the trend of the income level of each respective country. The Philippines as an LMIC, when compared with the higher-income countries (Table 5) has lesser absolute number, proportions, and percentages of the variables that are generally accepted as social determinants of health. In terms of the Global Surgery indicators, Japan has the best neurosurgical workforce density, and provides better nancial risk protection as evidenced by low out-of-pocket payment shares and a high universal health coverage (UHC) effective coverage index. The purely neurosurgical center in Russia has the highest neurosurgical volume for the duration of this study period (Table 6).

Discussion
Hydrocephalus and Treatment Realities in Low-Resource Settings Disparities in Health Systems: Workforce, Infrastructure, and Financing Japan and Russia have surplus resources, better universal health coverage (UHC), and higher public nancing for health (Table 5)-borne from deliberate policies of placing a prime value on public health [26,27]-and this becomes important given that increased expenditure on surgery can easily allow the proportional expansion of surgical capacity [1,10,22,33]. Studies that analyzed large patient databases [34,35] have examined the effect of nation-level socioeconomic factors on conditions treated by neurosurgeons. Remick and colleagues [34], after multivariate mixed-effects logistic regression, identi ed two nation-level variables-physician density and mean GDP growth-as signi cantly associated with good seizure outcomes following pediatric epilepsy surgery. Similarly, Guha and colleagues [35] identi ed a higher country GDP and a greater neurosurgeon-to-population density as two nation-level variables that are independent predictors of good outcome following treatment for aneurysmal subarachnoid hemorrhage. The Philippines, in contrast to the two countries, has a lower neurosurgeon workforce density (Table 5), thus expectedly restricting the breadth of access to care for children needing CSF diversion surgery.
The increased likelihood of inability to undergo an early CSF diversion surgery for the Filipino cohort of patients ( Figure 1 and Table 2), especially the poor and extremely-poor subsets of Filipino patients, can be explained chie y by nancial barriers [1,5,19,24,33,36]. These barriers-direct and indirect costs relating to treatment-cause economic hardships to the household in which a patient belongs to. Families in which one of the members is a patient needing neurosurgical care are at risk for nancial catastrophe and impoverishment, and this is especially true in LMICs like the Philippines [1,24,28]. Furthermore, out-ofpocket expenditures and the risks for catastrophic and impoverishing expenditures are higher in the Philippines than in Japan and in Russia (Table 5). This is particularly disadvantageous given that our results show that a higher socioeconomic status is signi cantly associated with an earlier time-interval of CSF diversion surgery (Table 3). Financial risk protection, therefore, is important for the acceptability and accessibility of any surgical intervention [33], especially in countries where a signi cant proportion of the population are poverty-stricken [1,21,24,37]. This necessitates countrywide UHC for health insurance in any form or combination-precisely the kind of social structure that Japan excels at [21,27]-that would subsidize the treatment-related costs incurred by the patients' families. While all patients from the three centers in this study have some form of health insurance coverage, those from the Japanese and Russian neurosurgical centers receive the broad range of inpatient and outpatient services with negligible out-of-pocket expenses after substantial subsidies by public insurance and government funding. These institutional features improve health-seeking behavior and make consultations during initial presentation of disease more likely. Furthermore, Japan leads in having a high UHC Effective Coverage Index (Table 5), in what could be considered as having an effective social safety net that offsets household expenditures against costly neurosurgical management [38]. Financial risk protection is indeed important because inadequate or poor-quality health insurance coverage poses an increased likelihood of poor outcomes following CSF diversion procedures for pediatric patients [7,39].

Governance Structures and Social Determinants of Health of a Country
While a comprehensive review of the health systems of Japan [27], the Philippines [25], and the Russian Federation [26] are beyond the scope of this article, our results show that across the majority of nationlevel metrics, the Philippines lags behind Japan and Russia in terms of both the Global Surgery indicators and social determinants of health (Table 5, 6). The provision of adequate standard of care is also shaped by social determinants of health within a country. Several studies have shown that a country's economic robustness and level of resources, i.e., being a HIC, translates into better patient-level outcomes particularly when investments on perioperative care and surgical systems are made [10,22,[34][35][36].
Perennial lack of resources in public hospitals contributes to the inability to provide safe and quality surgical services thus diminishing the capacity to rescue patients from avoidable deaths due to treatment complications. At the level of the neurosurgical centers, effective domestic resource mobilization in healthcare institutions is necessary in securing the health nancing needed to achieve improved patient outcomes. Investing in surgical services is therefore paramount, but this responsibility lies beyond the sphere of in uence of organized neurosurgery.

Policy Work and Resource Management Can Be The Way Forward
Policies that increase government expenditure on health appear to improve the composite metric that re ects nation-level performance of a health system [22,34]. Advocacy for more strategic policies and investments that address social determinants of health can strengthen governance and nancing arrangements [14,15]. These in turn help to re-shape more responsive and equitable health and surgical systems, as certain strategies can be undertaken to reduce variation in the use of surgery [40]. LMICs have the task of providing the full range of a responsive neurosurgical system-from as simple as the availability of shunt catheter kits to more capital-intensive measures such as comprehensive facility development, progressive hospital billing, strategic purchasing, and catastrophic case packages [41]that all in all, curbs out-of-pocket payments and helps achieve universal health coverage. A multi-level systems approach [12,40] by the involved policymakers can result in improvements in care processes of the surgical and health systems, which in turn result in better quality of care, and upon which hinges the hope of ultimately translating to better patient outcomes.

Limitations of the Study and Future Directions
The study includes patients with considerable heterogeneity in terms of etiology of hydrocephalus.
Additionally, the neurosurgical centers are not entirely representative of their countries because of inherent intranational heterogeneity of institutions, especially between the public and private sectors. Selection bias and information bias may have been present because of limitations of a retrospective review. Attributing certain outcomes to a policy when they are in fact owed to unmeasurable variables runs the risk of secular trend bias as well. Due to the limited sample size, limited regression analysis, and the non-randomized and unmatched observational study design, confounding factors and their impact may not have been adequately lessened. Regardless, our study ventures into a cross-country comparison of outcomes and explores issues that are of larger socioeconomic context as related to the granularity of patient outcomes for a particular disease entity after neurosurgical intervention. If and when the Global Surgery indicator-targets are met, the outcomes of neurosurgical patients in low-resource centers of LMICs after certain policy changes can be compared using the difference-in-differences study design [42]. Finally, the authors look forward to increasing center recruitment or prospectively gathering further primary patient-level data for the next phase or form of the present study. We recommend further studies with large sample sizes that allow the inclusion of nation-level covariates into a hierarchical mixedeffects statistical analysis [34,35] that can in turn determine the magnitude of effect of those variables.

Conclusion
In this study, we compared Global Surgery outcomes following CSF diversion surgery for pediatric hydrocephalus among neurosurgical centers from different countries of varying income levels. We found that the cohort of Filipino children underwent late time-interval of CSF diversion surgery compared to their Japanese and Russian counterparts. The differences in timeliness of surgery were signi cantly related to the etiology of hydrocephalus, as well as the socioeconomic status of the household that the child belongs to. In the cohort from the Philippines, from which three children suffered mortality, EVD insertion was associated with mortality. The variation in these outcomes may re ect the robustness of a country's health system. Certain institution-and nation-level factors may explain the differences when viewed through the lenses of the Global Surgery indicators and social determinants of health.   upper middle-income economies are those with a GNI per capita between $3,996 and $12,375; highincome economies are those with a GNI per capita of $12,376 or more.