The Impact of Gender on Pediatric Surgical Access and Outcomes in Africa

Introduction

Children remain among the most vulnerable populations worldwide. As their numbers rapidly expand, the global burden of disease also increases [1–9]. In fact, estimates suggest that currently 1.7 billion children lack access to surgery that is both safe and affordable [10]. A significant proportion of these children reside in low- and middle-income countries (LMICs), that are frequently operating with limited resources and infrastructure [11,12]. In Africa, the population demographic is approximately 50% pediatric; by corollary, the surgical disease burden is critical [11,12]. However, the surgical care of children is an aspect of child health that continues to be inadequate, despite its potential for reducing or even eliminating child morbidity and/or mortality [13].

Gender is a social construct that may further compound the issue. It can be defined as a set of characteristics, roles, and behaviors ascribed to perceptions of masculinity and femininity [13,14], potentially mutable according to times and/or society norms and values. While gender may intersect with sex, they are not necessarily the same, as the latter tends to focus on biological factors like chromosomes, hormones, or reproductive organs. [14,15] Historically, gender has been implicated in hierarchical power imbalances, whereby, within the male-female dichotomy, females were deemed inferior [15]. As a result, gender norms may influence children’s health and well-being. For example, in certain African societies, gender inequity has been reported to manifest as social and/or structural barriers negatively impacting females [15]. Moreover, recent reports have suggested that, in some African countries, women seek care for their surgical conditions less commonly than men, and those who do undergo surgery experience poorer surgical outcomes [16–20].

Despite the evidence of gender disparity in the adult African population, potential gender differentials in the surgical care of African children have not been documented to date. This study sought to address the knowledge gap by exploring whether there were gender disparities in access, and postoperative outcomes for African children requiring surgery.

Methods

Results

Meta-Analysis

A) Access

For the 54 studies included in the analysis, involving conditions that should have had the same values for boys and girls, males were more likely (94.0%) to have access to surgery than females (Figure 2a). Based on the random effects model, the average ratio was calculated, and found to be significantly different from zero (p < 0.0001), suggesting the results were not due to random chance. It revealed an odds ratio of 3.6 (95% CI: 2.6, 4.9), suggesting that males received surgery 3.6 times more frequently than females. Further testing indicated that there was high heterogeneity among the studies (Tau² = 1.2, I² = 98.7%, H² = 76.5, Q = 1174.5; p < 0.001). As a result, a 95% prediction interval, or range was calculated (−0.9 – 3.5). According to studentized residuals, one study may be an outlier [24*]; and in concert with another study [15*] may be overly influential. Exploration of publication bias via funnel plot did not indicate asymmetry (Figure 2b). This was reinforced via Egger’s regression (0.6, p = 0.6), Spearman rank correlation (0.1, p = 0.5), and fail-safe N testing (41154, p < 0.001).

• Download figure
• Open in new tab

Fig. 2

(a) Forest plot of access (log scale) odds ratio (n = 54); (b) Funnel plot representing publication bias

B) Adverse Postoperative Outcomes - Mortality

For the 29 studies included in the analysis, involving conditions that should have had the same values for boys and girls, mortality was higher for females than males. (76.0%). (Figure 3a). Based on the random effects model, the average ratio was calculated, and found to be significantly different from zero (p < 0.0001), suggesting the results were not due to random chance. It revealed an odds ratio of 1.6 (95% CI: 1.3, 2.0), suggesting that females died 1.6 times more frequently than males. Further testing indicated that there was moderate heterogeneity among the studies (Tau² = 0.1, I² = 41.7%, H² = 1.7, Q = 41.2; p = 0.05). As a result, a 95% prediction interval, or range was calculated (−0.1 – 1.1). According to studentized residuals, there were no outliers; while another study [15*] may be overly influential [45*]. Exploration of publication bias via funnel plot did not indicate asymmetry (Figure 3b). This was reinforced via Egger’s regression (−0.2, p = 0.8), Spearman rank correlation (0.1, p = 0.7), and fail-safe N testing (336, p < 0.001).

• Download figure
• Open in new tab

Fig. 3

(a) Forest plot of access (log scale) odds ratio (n = 29); (b) Funnel plot representing publication bias

Discussion

Surgery is an indispensable component of healthcare, and essential to address the burden of disease worldwide [9,26]. For every child to be granted the right to safe, timely, and affordable care, access to surgery and good postoperative outcomes are crucial. The current study has for the first time explored, via systematic review, gender disparities in surgical access and outcomes for African children. This exploration was predicated on the suspicion that gender inequities may adversely affect the health and well-being of girls.

Gender disparities have been documented to affect children in many areas, including education. As a result, the United Nations established its Sustainable Development Goals (SDGs) #4 and #5 (equitable quality education, gender equality and empowerment) to promote universal access to, and completion of, primary education, thus closing the gender gap [27]. Notwithstanding, many African countries have yet to achieve educational parity by gender [28]. According to UNESCO, one in four girls are likely to be out of school in sub-Saharan Africa [28]. In addition, these girls are more likely than their male counterparts to never enroll in school [28]. Friedman et al., in their modeling and forecast of educational inequality as it pertains to the 2030 SDG targets, reported that the education of females is predicted to lag that of males in sub-Saharan Africa through 2030 [29]. Oruko et al., in their exploration of the educational access and outcomes of Kenyan children, reported that despite the government providing free primary education, fewer girls reached secondary school level; and those that remained in school were less likely than boys to achieve a grade of C+ or better in their national exams [30]. These examples demonstrate the disadvantaged position of girls compared to boys in the educational realm.

Gender disparities in the surgical care of children have been documented tangentially in a few isolated reports. For example, Sakita et al., in their exploration of clinical profiles and outcomes of children presenting to the emergency department of a Tanzanian hospital with undifferentiated abdominal pain, reported a low proportion of girls (32.6%) in their study population [26]. Moreover, girls were associated with higher in-hospital mortality in their regression analysis [26]. Peiffer et al., in their exploration of factors associated with adverse outcomes (e.g. prolonged length of stay, readmission, and mortality) following surgery in a Ghanaian regional hospital, found that although only 10% of the cohort were girls, female gender was significantly associated with in-hospital postoperative mortality [15*]. When looking at pediatric burn care in their investigation of potential relationships among patient demographics, thermal injury and mortality in a Malawi burn unit, Purcell et al. demonstrated that fewer females presented (44.1%), and that those who did were in more severe states (% total body surface area) [9*]. Perhaps due in part to the severity, the authors reported higher rates of surgical intervention for these girls, and interestingly, a lack of increased odds of mortality due to female gender [9*]. Allorto et al., while investigating factors associated with delayed children’s surgery and mortality following burn injuries in a South African burn service, reported fewer females in their cohort (44.5%); yet they were less likely to undergo early surgical intervention, they had higher mortality, and lengths of stay [12*]. Pediatric ophthalmology studies have identified a significant male predominance at presentation. For example, in reports concerning the proportions of children presenting for cataract surgery in a Nigerian ophthalmology clinic, Ugalahi et al. noted the paltry number of girls (44.4%) [31]. In their evaluation of several African child eye health tertiary facilities (CEHTFs), Adhisesha Reddy et al. found access disparities favouring boys for children receiving cataract surgery and follow up; they also noted little effective means to reduce the inequity [32]. Courtright et al., in their analysis of child recipients of cataract surgery at two Tanzanian CEHTFs found a sex ratio of 1.48 [33]. Eriksen et al. reported few girls (36.4%) in their examination of an active surgical service for children in a Tanzanian tertiary eye care referral facility [34]. They also found girls were less likely to attend follow up visits [34].

When discussing pediatric surgical disease, it must be acknowledged that there are certain diseases that affect boys more than girls, and vice versa. For example, congenital inguinal hernia is a very common pediatric condition, for which its surgical repair has been described in the literature as having a sex ratio prevalence (male:female) ranging from 2 to 18:1 [35, 36]. In Africa specifically, Igwe et al. reported a sex ratio of 5:1 in their comparison of laparoscopic versus open inguinal hernia repair, in a Nigerian university teaching hospital [37]. In addition, trauma may occur more often in males than females, as noted by Aluisio et al. and Halawa et al. in their Rwandan and Egyptian reports [38, 39]. Nevertheless, this current study excluded conditions that affected predominantly boys or girls, as a result of unique anatomy or physiology. As indicated in the methods definitions and caveats sub-section, this study incorporated conditions with equal sex ratios, meaning they occurred equally in boys and girls.

In this systematic review and meta-analysis of sex-disaggregated studies it was demonstrated that surgical interventions were more likely to be performed on males, and that females were more likely to have poor postoperative outcomes, including death. These findings are based on an analysis of 54 papers, mostly retrospective and single-site in nature, focusing on pediatric cohorts of up to 200 children, stemming from four African countries, mostly focusing on pediatric surgery. The male predominance in surgical access, and female predominance in adverse postoperative outcomes including mortality, were highlighted in sex ratio calculations, then reinforced through forest plot odds ratios in the meta-analysis.

There are 2 types of models for a meta-analysis: a fixed and random effects model. The latter was more appropriate for this study, because it assumed that the true effect could vary by study due to the differences among them (heterogeneity) [40, 41]. For example, the included studies varied by the size of their cohorts, location, types of surgical intervention, and other demographics. This could have resulted in different effect sizes. In the meta-analysis, there was high heterogeneity when evaluating access, and moderate heterogeneity when examining postoperative mortality. As a result, a prediction interval was calculated, and the vast majority of studies’ true effects were positive. The results were also supported by both the lack of statistical significance for statistical indices, and the absence of funnel plot asymmetry, both refuting publication bias.

In contrast to the Cochrane Risk of Bias tool (RoB2) [42], the recommended tool for randomized control trials, the Risk of Bias in Non-Randomized Studies – of Interventions (ROBINS-I) tool is used to evaluate the quality of the evidence for non-randomized studies [25]. In this study, only 22 of the studies were eligible for ROBINS-I assessment. These studies’ risk of bias level by domain and intervention were described (risk of bias assessment results sub-section) and illustrated (Table 2).

Unfortunately all of the studies assessed had a risk of bias in at least one domain, which limited the strength of the result interpretation. Within the domain for confounding, there was both critical and serious risk of bias in approximately 50% of the studies, which contributed to serious risk of bias across the pre-/at-intervention domains. As a result, the possibility of at least one domain not being appropriately measured or controlled cannot be excluded - nor can the possibility of confounding be entirely overlooked.

This study is a systematic review, providing a snapshot of the literature located in selected databases, published from specific African countries, during a discrete period of time. Although the analysis of publication bias was not significant according to the funnel plot (and associated statistical indices), the included studies represented approximately a third of all the countries in Africa. This study was a summary of what was reported. At no point did this study place judgment on African values, their family unit, their healthcare providers’ intentions or capabilities. Moreover, this study should not be viewed as an indictment of Africa’s treatment of its children, as similar findings have been reported in high-income countries. For example, in their thirteen year review of the epidemiology of pediatric surgery in the US, Rabbitts et al. reported female gender among other factors associated with decreased likelihood of surgical intervention on multivariate analysis [43]. In their examination of social determinants on pediatric trauma outcomes in the US South, Dickens et al. reported that girls had lower odds of being referred for rehabilitation services, and Black girls lowest of all groups [44]. In their analysis of access to renal transplantation, through a multinational European collaboration, Hogan et al. reported girls had lower probability of receiving preemptive transplantation, for reasons only partially attributed to medical factors [45]. In essence, what cannot be overstated is that findings of gender differentials in surgical access and/or postoperative outcomes have similarly been reported in high income countries; so gender disparities are not solely specific to Africa.

Other study limitations include: generalizability, the small number of included studies, the quality of data they presented, their age, and potential confounders (e.g. geographical location of participants, lack of infrastructure, etc). Studies with potential risk of bias were not removed from the analysis, nor was the one study that may have been a highly influential outlier.

This study was conducted in constant consultation with several African collaborators, who have attained positions of considerable distinction in pediatric surgery. In terms of the next phase in this study, collaboration with these colleagues and other African stakeholders (e.g. healthcare providers, caregivers, former patients, etc), is underway to uncover the obstacles to gender equity in pediatric surgical care, and to develop a comprehensive, sustainable mitigation plan to reduce the disparities via a mixed methods approach in Africa.

Conclusion

African girls are a vulnerable population who, in addition to facing the limitations of residing in many low-resource settings, may be subjected to additional barriers of gender inequity. This study’s findings suggest that although surgeries are performed more often on males, females have a higher likelihood of adverse postoperative outcomes. This first-time summary of evidence of gender disparities in pediatric surgical access and outcomes in Africa is an urgent call to action, hopefully prompting not only further investigation, but above all tangible change.

Disclosures

Author Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis was performed by SW, DP. The first draft of the manuscript was written by SW, and all authors edited, read, and approved the final manuscript.

Appendix A Search Strategy

Appendix B PRISMA Checklist

Appendix C Access, adverse postoperative outcomes, and mortality calculations

Appendix D Finalized list of studies for meta-analysis (n = 54)