Preventability of injury-related morbidity & mortality at four hospitals in Cameroon: A systematic approach to trauma quality improvement.

BACKGROUND: Adverse events reviews are a fundamental component of trauma quality improvement (QI) that facilitate the correction of systemic issues in care. Although injury-related mortality in Cameroon is substantial, to our knowledge, opportunities for QI have not been formally assessed. Thus, a formal review of adverse events in Cameroonian trauma patients was implemented as a first step toward identifying targets for systems modification. METHODS: A QI committee composed of multidisciplinary experts at four hospitals in Cameroon was formed to review adverse events including deaths among trauma patients from 2019 to 2021. Events were discussed at newly established morbidity and mortality conferences and committee meetings to identify contributing factors and overall preventability. RESULTS: During 50 meetings, 95 adverse events were reviewed, including 58 deaths (61%). Other adverse events were delays in diagnosis/treatment (22%) and surgical site infections (17%). Overall, 34 deaths (59%) were classified as preventable, 21% potentially preventable, and 21% not preventable. Over half (52%) of the 46 preventable or potentially preventable deaths occurred in the emergency department (ED); while brain injury (57%), respiratory failure (41%), and hemorrhage (39%) were the most frequent physiologic factors associated with mortality. Contributory factors identified include lack of a structured approach to patient management, absence of continuous training for personnel, and locally adapted protocols. CONCLUSIONS: Basic improvements in evaluation and management of life-threatening issues in the ED can significantly reduce the high rate of preventable trauma-related deaths across Cameroon. Formal trauma QI methods can be utilized in low-resource environments to determine mortality root causes and identify intervention targets.

Evaluating Shifts in Perception After a Pilot Trauma Quality Improvement Training Course in Cameroon.

INTRODUCTION: Trauma is a major contributor to the global burden of disease, with low- and middle-income countries (LMICs) being disproportionately affected. Trauma Quality Improvement (QI) initiatives could potentially save an estimated two million lives each year. Successful trauma QI initiatives rely on adequate training and a culture of quality among hospital staff. This study evaluated the effect of a pilot trauma QI training course on participants' perceptions on leadership, medical errors, and the QI process in Cameroon. METHODS: Study participants took part in a three-day, eight-module course training on trauma QI methods and applications. Perceptions on leadership, medical errors, and QI were assessed pre and post-course using a 15-item survey measured on a five-point Likert scale. Median pre- and post-course scores were compared using the Wilcoxon signed-rank test. Knowledge retention and course satisfaction were also evaluated in a post-course survey and evaluation. RESULTS: A majority of the 25 course participants completed pre-course (92%) and post-course (80%) surveys. Participants' perceptions of safety and comfort discussing medical errors at work significantly increased post-course (pre-median = 5, IQR [4-5]; post-median = 5, IQR [5-5]; P = 0.046). The belief that individuals responsible for medical error should be held accountable significantly decreased after the course (pre-median = 3, IQR [2-4]; post-median = 1, IQR [1-2]; P < 0.001). Overall satisfaction with the course was high with median scores ≥4. CONCLUSIONS: These initial results suggest that targeted trauma QI training effectively influences attitudes about QI. Further investigation of the effect of the trauma QI training on hospital staff in larger courses is warranted to assess reproducibility of these findings.

Baseline Patient Safety Culture in Cameroon: Setting a Foundation for Trauma Quality Improvement.

BACKGROUND: Trauma quality improvement (QI) has resulted in decreased trauma mortality and morbidity in high-income countries and has the potential to do the same in low- and middle-income countries. Effective implementation of QI programs relies on a foundational culture of patient safety; however, studies on trauma-related patient safety culture in Sub-Saharan Africa remain scarce. This study assesses baseline patient safety culture in Cameroon to best identify opportunities for improvement. MATERIALS AND METHODS: Over a 3-week period, the Hospital Survey on Patient Safety Culture was administered in three hospitals in the Littoral region of Cameroon. Percentages of positive responses (PPRs) were calculated across 42 items in 12 survey dimensions. A mixed-effects logistic regression model was used to summarize dimension-level percentages and confidence intervals. RESULTS: A total of 179 trauma-related hospital personnel were surveyed with an overall response rate of 76.8%. High PPRs indicate favorable patient safety culture. Of the 12 dimensions evaluated by the Hospital Survey on Patient Safety Culture, nine had a PPR below 50%. Dimensions particularly pertinent in the context of QI include Nonpunitive Response to Errors with a PPR of 25.8% and Organization Learning-Continuous Improvement with a PPR of 64.7%. CONCLUSIONS: The present study elucidates an opportunity for the development of trauma patient safety culture in Cameroon. Low PPR for Nonpunitive Response to Errors indicates a need to shift cultural paradigms from ascribing individual blame to addressing systemic shortcomings of patient care. Moving forward, data from this study will inform interventions to cultivate patient safety culture in partnering Cameroonian hospitals.

Financial Risk Protection and Hospital Admission for Trauma in Cameroon: An Analysis of the Cameroon National Trauma Registry.

BACKGROUND: Half of the global population is at risk for catastrophic health expenditure (CHE) in the event that they require surgery. Universal health coverage fundamentally requires protection from CHE, particularly in low- and middle-income countries (LMICs). Financial risk protection reports in LMICs covering surgical care are limited. We explored the relationship between financial risk protection and hospital admission among injured patients in Cameroon to understand the role of health insurance in addressing unmet need for surgery in LMICs. METHODS: The Cameroon National Trauma Registry, a database of all injured patients presenting to the emergency departments (ED) of three Cameroonian hospitals, was retrospectively reviewed between 2015 and 2017. Multivariate regression analysis identified predictors of hospital admission after injury and of patient report of cost inhibiting their care. RESULTS: Of the 7603 injured patients, 95.7% paid out-of-pocket to finance ED care. Less than two percent (1.42%) utilized private insurance, and more than half (54.7%) reported that cost inhibited their care. In multivariate analysis, private insurance coverage was a predictor of hospital admission (OR 2.17, 95% CI: 1.26, 3.74) and decreased likelihood of cost inhibiting care (OR 0.34, 95% CI: 0.20, 0.60) when compared to individuals paying out-of-pocket. CONCLUSION: The prevalence of out-of-pocket spending among injured patients in Cameroon highlights the need for financial risk protection that encompasses surgical care. Patients with private insurance were more likely to be admitted to the hospital, and less likely to report that cost inhibited care, supporting private health insurance as a potential financing strategy.

A cross-sectional survey of emergency and essential surgical care capacity among hospitals with high trauma burden in a Central African country.

BACKGROUND: As the overwhelming surgical burden of injury and disease steadily increases, disproportionately affecting low- and middle-income countries, adequate surgical and trauma care systems are essential. Yet, little is known about the emergency and essential surgical care (EESC) capacity of facilities in many African countries. The objective of this study was to assess the EESC capacity in different types of hospitals across Cameroon. METHODS: This cross-sectional survey used the WHO Tool for Situational Analysis to Assess EESC, investigating four key areas: infrastructure, human resources, interventions, and equipment and supplies. Twelve hospitals were surveyed between August and September 2009. Facilities were conveniently sampled based on proximity to road traffic and sociodemographic composition of population served in four regions of Cameroon. To complete the survey, investigators interviewed heads of facilities, medical advisors, and nursing officers and consulted hospital records and statistics at each facility. RESULTS: Seven district hospitals, two regional hospitals, two general hospitals, and one missionary hospital completed the survey. Infrastructure for EESC was generally inadequate with the largest gaps in availability of oxygen concentrator supply, an on-site blood bank, and pain relief management guidelines. Human resources were scarce with a combined total of six qualified surgeons, seven qualified obstetrician/gynecologists, and no anesthesiologists at district, regional, and missionary hospitals. Of 35 surgical interventions, 16 were provided by all hospitals. District hospitals reported referring patients for 22 interventions. Only nine of the 67 pieces of equipment were available at all hospitals for all patients all of the time. CONCLUSIONS: Severe shortages highlighted by this survey demonstrate the significant gaps in capacity of hospitals to deliver EESC and effectively address the increasing surgical burden of disease and injury in Cameroon. This data provides a foundation for evidence-based decision-making surrounding appropriate allocation and provision of resources for adequate EESC in the country.

Bystander intervention is associated with reduced early mortality among injury victims in Cameroon.

Despite high injury mortality rates, Cameroon currently lacks a formal prehospital care system. In other sub-Saharan African low and middle-income countries, Lay First Responder (LFR) programs have trained non-medical professionals with high work-related exposure to injury in principles of basic trauma care. To develop a context-appropriate LFR program in Cameroon, we used trauma registry data to understand current layperson bystander involvement in prehospital care and explore associations between current non-formally trained bystander-provided prehospital care and clinical outcomes. The Cameroon Trauma Registry (CTR) is a longitudinal, prospective, multisite trauma registry cohort capturing data on injured patients presenting to four hospitals in Cameroon. We assessed prevalence and patterns of prehospital scene care among all patients enrolled the CTR in 2020. Associations between scene care, clinical status at presentation, and outcomes were tested using univariate and multivariate logistic regression. Injury severity was measured using the abbreviated injury score. Data were analyzed using Stata17. Of 2212 injured patients, 455 (21%) received prehospital care (PC) and 1699 (77%) did not receive care (NPC). Over 90% (424) of prehospital care was provided by persons without formal medical training. PC patients were more severely injured (p<0.001), had markers of increased socioeconomic status (p = 0.01), and longer transport distances (p<0.001) compared to NPC patients. Despite increased severity of injury, patients who received PC were more likely to present with a palpable pulse (OR = 6.2, p = 0.02). Multivariate logistic regression adjusted for injury severity, socioeconomic status and travel distance found PC to be associated with reduced emergency department mortality (OR = 0.14, p<0.0001). Although prehospital injury care in Cameroon is rarely performed and is provided almost entirely by persons without formal medical training, prehospital intervention is associated with increased early survival after injury. Implementation of LFR training to strengthen the frequency and quality of prehospital care has considerable potential to improve trauma survival.

Gender-based violence and its health risks on women in Yaoundé, Cameroon.

INTRODUCTION: Gender-based violence (GBV) is a major public health problem that disproportionately affects women. In Cameroon, as well as other countries worldwide, GBV has immediate effects on women's health, with one in three women experiencing physical or sexual violence from an intimate partner, affecting their physical and reproductive health. The objective of this study was to determine the health risks associated with GBV among women in Yaoundé. METHODS: A cross-sectional study was conducted in Yaoundé (Cameroon), from August to October 2022. Adverse health outcome included mental disorders, physical trauma, gynaecological trauma, behavioral disorders, and any other disorder. Tests of associations were used to establish relationships between qualitative variables. Associations were further quantified using crude odds ratio (OR) for univariate analysis and adjusted odds ratio (aOR) for multivariate analysis with 95% confidence interval (CI). Independent variables included: Physical violence, Sexual violence, Economic violence, Emotional violence, Age, Number of children, and Marital status. Variables with p-value˂0.05 were considered statistically significant. RESULTS: A total of 404 women aged 17 to 67 years were interviewed. Emotional violence was the most commonly reported violence (78.8%), followed by economic violence (56.9%), physical violence (45.8%) and sexual violence (33.7%). The main reasons for violence were jealousy (25.7%), insolence (19.3%) and the refusal to have sexual intercourse (16.3%). The prevalences of adverse health outcomes were physical trauma (90.9%), followed by mental disorders (70,5%), gynaecological trauma (38.4%), behavioral disorders (29.7%), and other (5.5%). Most victims reported at least one of the above-mentioned conditions (80.2%). Women who were victims of any kind of violence had a higher likelihood of experiencing adverse health outcomes: physical violence [OR = 34.9, CI(10.8-112.9), p < 0.001]; sexual violence [OR = 1.5, CI(0.9-2.7), p = 0.11]; economic violence [OR = 2.4, CI(1.4-3.9), p = 0.001]; and emotional violence [OR = 2.9, CI(1.7-4.9), p < 0.001]. Using multiple binary logistic regression, only physical violence [aOR = 15.4, CI(6.7-22.5), p = 0.001] remained highly associated with an increased likelihood of having adverse health outcomes. CONCLUSION: This study underscores the urgent need for comprehensive interventions to address GBV, including improved reporting and documentation of cases, increased awareness among healthcare providers, the establishment of support networks for victims, primary and secondary prevention of GBV. It is essential that the Government of Cameroon, through the Ministries in charge of Health and Women's Empowerment, minimizes the health effects of GBV through early identification, monitoring, and treatment of GBV survivors by providing them with high-quality health care services.

Crystalloid resuscitation is associated with decreased treatment delays and improved systolic blood pressures in a blood-constrained setting.

Objectives: We analyzed resuscitation practices in Cameroonian patients with trauma as a first step toward developing a context-appropriate resuscitation protocol. We hypothesized that more patients would receive crystalloid-based (CB) resuscitation with a faster time to administration than blood product (BL) resuscitation. Methods: We included patients enrolled between 2017 and 2019 in the Cameroon Trauma Registry (CTR). Patients presenting with hemorrhagic shock (systolic blood pressure (SBP) <100 mm Hg and active bleeding) were categorized as receiving CB, BL, or no resuscitation (NR). We evaluated differences between cohorts with the Kruskal-Wallis test for continuous variables and Fisher's exact test for categorical variables. We compared time to treatment with the Wilcoxon rank sum test. Results: Of 9635 patients, 403 (4%) presented with hemorrhagic shock. Of these, 278 (69%) patients received CB, 39 (10%) received BL, and 86 (21%) received NR. BL patients presented with greater injury severity (Highest Estimated Abbreviated Injury Scale (HEAIS) 4 BL vs 3 CB vs 1 NR, p<0.001), and lower median hemoglobin (8.0 g/dL BL, 11.4 g/dL CB, 10.6 g/dL NR, p<0.001). CB showed greater initial improvement in SBP (12 mm Hg CB vs 9 mm Hg BL vs 0 NR mm Hg, p=0.04) compared with BL or no resuscitation, respectively. Median time to treatment was lower for CB than BL (12 vs 131 min, p<0.01). Multivariate logistic regression adjusted for injury severity found no association between resuscitation type and mortality (CB adjusted OR (aOR) 1.28, p=0.82; BL aOR 1.05, p=0.97). Conclusions: CB was associated with faster treatment, greater SBP elevation, and similar survival compared with BL in Cameroonian patients with trauma with hemorrhagic shock. In blood-constrained settings, treatment delays associated with blood product transfusion may offset the physiologic benefits of an early BL strategy. CB prior to definitive hemorrhage control in this resource-limited setting may be a necessary strategy to optimize perfusion pressure. Level of evidence and study type: III, retrospective study.

Who seeks care after intimate partner violence in Cameroon? sociodemographic differences between a hospital and population sample of women.

INTRODUCTION: Little is known regarding health care seeking behaviors of women in sub-Saharan Africa, specifically Cameroon, who experience violence. The proportion of women who experienced violence enrolled in the Cameroon Trauma Registry (CTR) is lower than expected. METHODS: We concatenated the databases from the October 2017-December 2020 CTR and 2018 Cameroon Demographic and Health Survey (DHS) into a singular database for cross-sectional study. Continuous and categorical variables were compared with Wilcoxon rank-sum and Fisher's exact test. Multivariable logistic regression examined associations between demographic factors and women belonging to the DHS or CTR cohort. We performed additional classification tree and random forest variable importance analyses. RESULTS: 276 women (13%) in the CTR and 197 (13.1%) of women in the DHS endorsed violence from any perpetrator. A larger percentage of women in the DHS reported violence from an intimate partner (71.6% vs. 42.7%, p<0.001). CTR women who experienced IPV demonstrated greater university-level education (13.6% vs. 5.0%, p<0.001) and use of liquid petroleum gas (LPG) cooking fuel (64.4% vs. 41.1%, p<0.001). DHS women who experienced IPV reported greater ownership of agricultural land (29.8% vs. 9.3%, p<0.001). On regression, women who experienced IPV using LPG cooking fuel (aOR 2.55, p = 0.002) had greater odds of belonging to the CTR cohort while women who owned agricultural land (aOR 0.34, p = 0.007) had lower odds of presenting to hospital care. Classification tree variable observation demonstrated that LPG cooking fuel predicted a CTR woman who experienced IPV while ownership of agricultural land predicted a DHS woman who experienced IPV. CONCLUSION: Women who experienced violence presenting for hospital care have characteristics associated with higher SES and are less likely to demonstrate factors associated with residence in a rural setting compared to the general population of women experiencing violence.

Mobile telephone follow-up assessment of postdischarge death and disability due to trauma in Cameroon: a prospective cohort study.

OBJECTIVES: In Cameroon, long-term outcomes after discharge from trauma are largely unknown, limiting our ability to identify opportunities to reduce the burden of injury. In this study, we evaluated injury-related death and disability in Cameroonian trauma patients over a 6-month period after hospital discharge. DESIGN: Prospective cohort study. SETTING: Four hospitals in the Littoral and Southwest regions of Cameroon. PARTICIPANTS: A total of 1914 patients entered the study, 1304 were successfully contacted. Inclusion criteria were patients discharged after being treated for traumatic injury at each of four participating hospitals during a 20-month period. Those who did not possess a cellular phone or were unable to provide a phone number were excluded. PRIMARY AND SECONDARY OUTCOME MEASURES: The Glasgow Outcome Scale-Extended (GOSE) was administered to trauma patients at 2 weeks, 1 month, 3 months and 6 months post discharge. Median GOSE scores for each timepoint were compared and regression analyses were performed to determine associations with death and disability. RESULTS: Of 71 deaths recorded, 90% occurred by 2 weeks post discharge. At 6 months, 22% of patients still experienced severe disability. Median (IQR) GOSE scores at the four timepoints were 4 (3-7), 5 (4-8), 7 (4-8) and 7 (5-8), respectively, (p<0.01). Older age was associated with greater odds of postdischarge disability (OR: 1.23, 95% CI: 1.07 to 1.41) and mortality (OR: 2.15, 95% CI: 1.52 to 3.04), while higher education was associated with decreased odds of disability (OR: 0.65, 95% CI: 0.58 to 0.73) and mortality (OR: 0.38, 95% CI: 0.31 to 0.47). Open fractures (OR: 1.73, 95% CI: 1.38 to 2.18) and closed fractures (OR: 1.83, 95% CI: 1.42 to 2.36) were associated with greater postdischarge disability, while higher Injury Severity Score (OR: 2.44, 95% CI: 2.13 to 2.79) and neurological injuries (OR: 4.40, 95% CI: 3.25 to 5.96) were associated with greater odds of postdischarge mortality. CONCLUSION: Mobile follow-up data show significant morbidity and mortality, particularly for orthopaedic and neurologic injuries, up to 6 months following trauma discharge. These results highlight the need for reliable follow-up systems in Cameroon.

Associations between social determinants of health and interpersonal violence-related injury in Cameroon: a cross-sectional study.

INTRODUCTION: Risk factors for interpersonal violence-related injury (IPVRI) in low-income and middle-income countries (LMICs) remain poorly defined. We describe associations between IPVRI and select social determinants of health (SDH) in Cameroon. METHODS: We conducted a cross-sectional analysis of prospective trauma registry data collected from injured patients >15 years old between October 2017 and January 2020 at four Cameroonian hospitals. Our primary outcome was IPVRI, compared with unintentional injury. Explanatory SDH variables included education level, employment status, household socioeconomic status (SES) and alcohol use. The EconomicClusters model grouped patients into household SES clusters: rural, urban poor, urban middle-class (MC) homeowners, urban MC tenants and urban wealthy. Results were stratified by sex. Categorical variables were compared via Pearson's χ2 statistic. Associations with IPVRI were estimated using adjusted odds ratios (aOR) with 95% confidence intervals (95%CI). RESULTS: Among 7605 patients, 5488 (72.2%) were men. Unemployment was associated with increased odds of IPVRI for men (aOR 2.44 (95% CI 1.95 to 3.06), p<0.001) and women (aOR 2.53 (95% CI 1.35 to 4.72), p=0.004), as was alcohol use (men: aOR 2.33 (95% CI 1.91 to 2.83), p<0.001; women: aOR 3.71 (95% CI 2.41 to 5.72), p<0.001). Male patients from rural (aOR 1.45 (95% CI 1.04 to 2.03), p=0.028) or urban poor (aOR 2.08 (95% CI 1.27 to 3.41), p=0.004) compared with urban wealthy households had increased odds of IPVRI, as did female patients with primary-level/no formal (aOR 1.78 (95% CI 1.10 to 2.87), p=0.019) or secondary-level (aOR 1.54 (95% CI 1.03 to 2.32), p=0.037) compared with tertiary-level education. CONCLUSION: Lower educational attainment, unemployment, lower household SES and alcohol use are risk factors for IPVRI in Cameroon. Future research should explore LMIC-appropriate interventions to address SDH risk factors for IPVRI.

Right population, right resources, right algorithm: Using machine learning efficiently and effectively in surgical systems where data are a limited resource.

There is a growing interest in using machine learning algorithms to support surgical care, diagnostics, and public health surveillance in low- and middle-income countries. From our own experience and the literature, we share several lessons for developing such models in settings where the data necessary for algorithm training and implementation is a limited resource. First, the training cohort should be as similar as possible to the population of interest, and recalibration can be used to improve risk estimates when a model is transported to a new context. Second, algorithms should incorporate existing data sources or data that is easily obtainable by frontline health workers or assistants in order to optimize available resources and facilitate integration into clinical practice. Third, the Super Learner ensemble machine learning algorithm can be used to define the optimal model for a given prediction problem while minimizing bias in the algorithm selection process. By considering the right population, right resources, and right algorithm, researchers can train prediction models that are both context-appropriate and resource-conscious. There remain gaps in data availability, affordable computing capacity, and implementation studies that hinder clinical algorithm development and use in low-resource settings, although these barriers are decreasing over time. We advocate for researchers to create open-source code, apps, and training materials to allow new machine learning models to be adapted to different populations and contexts in order to support surgical providers and health care systems in low- and middle-income countries worldwide.

Barriers to the National Onchocerciasis Control Programme at operational level in Cameroon: a qualitative assessment of stakeholders' views.

BACKGROUND: The global burden of onchocerciasis is the heaviest in sub-Saharan Africa. Studies have shown the importance of the role of Community-Directed Distributors (CDDs) and nurses in onchocerciasis control, but little is known about their experience in implementing onchocerciasis control programmes. Our aim was to document the barriers that CDDs and local health administrators face in implementing onchocerciasis control activities. METHODS: We conducted a qualitative survey consisting of 16 in-depth interviews and 8 focus group discussions (FGDs) across three health districts of Cameroon. We interviewed a total of 9 local health officials at the district and Health Area levels, and 7 CDDs. Eight FGDs were conducted with CDDs and Health Committee members. RESULTS: The major barriers to the implementation of Community Directed Treatment with Ivermectin that we identified were linked and interrelated. Examples of these barriers included: contextual factors (geographical and cultural background), top-to-bottom planning, insufficient human and material resources, and lack of transparency in the management of the programme's funds. CONCLUSIONS: The CDTI at operational level still faces many obstacles which negatively affect therapeutic coverages. This can lead to the non-adhesion of the communities to the programme, consequently jeopardizing the sustainability of the onchocerciasis elimination programme. We recommend that the national programme planners put in place a transparent management and planning system for onchocerciasis elimination activities, with better communication with local programme stakeholders.

Do Communities Really "Direct" in Community-Directed Interventions? A Qualitative Assessment of Beneficiaries' Perceptions at 20 Years of Community Directed Treatment with Ivermectin in Cameroon.

Recent studies in Cameroon after 20 years of implementation of the Community Directed Treatment with ivermectin (CDTI) strategy, revealed mixed results as regards community ownership. This brings into question the feasibility of Community Directed Interventions (CDI) in the country. We carried out qualitative surveys in 3 health districts of Cameroon, consisting of 11 individual interviews and 10 Focus Group Discussions (FGDs) with specific community members. The main topic discussed during individual interviews and FGDs was about community participation in health. We found an implementation gap in CDTI between the process theory in the 3 health districts. Despite this gap, community eagerness for health information and massive personal and financial adhesion to interventions that were perceived important, were indicators of CDI feasibility. The concept of CDI is culturally feasible in rural and semi-urban settlements, but many challenges hinder its actual implementation. In the view of community participation as a process rather than an intervention, these challenges include real dialogue with communities as partners, dialogue and advocacy with operational level health staff, and macroeconomic and political reforms in health, finance and other associated sectors.

Machine learning without borders? An adaptable tool to optimize mortality prediction in diverse clinical settings.

BACKGROUND: Mortality prediction aids clinical decision making and is necessary for quality improvement initiatives. Validated metrics rely on prespecified variables and often require advanced diagnostics, which are unfeasible in resource-constrained contexts. We hypothesize that machine learning will generate superior mortality prediction in both high-income and low- and middle-income country cohorts. METHODS: SuperLearner, an ensemble machine-learning algorithm, was applied to data from three prospective trauma cohorts: a highest-activation cohort in the United States, a high-volume center cohort in South Africa (SA), and a multicenter registry in Cameroon. Cross-validation was used to assess model discrimination of discharge mortality by site using receiver operating characteristic curves. SuperLearner discrimination was compared with standard scoring methods. Clinical variables driving SuperLearner prediction at each site were evaluated. RESULTS: Data from 28,212 injured patients were used to generate prediction. Discharge mortality was 17%, 1.3%, and 1.7% among US, SA, and Cameroonian cohorts. SuperLearner delivered superior prediction of discharge mortality in the United States (area under the curve [AUC], 94-97%) and vastly superior prediction in Cameroon (AUC, 90-94%) compared with conventional scoring algorithms. It provided similar prediction to standard scores in the SA cohort (AUC, 90-95%). Context-specific variables (partial thromboplastin time in the United States and hospital distance in Cameroon) were prime drivers of predicted mortality in their respective cohorts, whereas severe brain injury predicted mortality across sites. CONCLUSIONS: Machine learning provides excellent discrimination of injury mortality in diverse settings. Unlike traditional scores, data-adaptive methods are well suited to optimizing precise site-specific prediction regardless of diagnostic capabilities or data set inclusion allowing for individualized decision making and expanded access to quality improvement programming. LEVEL OF EVIDENCE: Prognostic and therapeutic, level II and III.

Adherence to ivermectin is more associated with perceptions of community directed treatment with ivermectin organization than with onchocerciasis beliefs.

BACKGROUND: The fight against onchocerciasis in Africa has boomed thanks to the Community Directed Treatment with Ivermectin (CDTI) program. However, in Cameroon, after more than 15 years of mass treatment, onchocerciasis prevalence is still above the non-transmission threshold. This study aimed to explore a possible association between people's beliefs/perceptions of onchocerciasis and of CDTI program, and their adherence to ivermectin in three regions of Cameroon. METHODOLOGY/PRINCIPAL FINDINGS: A cross sectional survey was carried out in three health districts with persistent high onchocerciasis prevalence. Participants were randomly selected in 30 clusters per district. Adherence to ivermectin was comparable between Bafang and Bafia (55.0% and 48.8%, respectively, p>0.05) and lower in Yabassi (40.7%). Among all factors related to program perceptions and disease representations that were studied, perceptions of the program are the ones that were most determinant in adherence to ivermectin. People who had a "not positive" opinion of ivermectin distribution campaigns were less compliant than those who had a positive opinion about the campaigns (40% vs 55% in Bafang, and 48% vs 62% in Bafia, p<0.01), as well as those who had a negative appreciation of community drug distributors' commitment (22% vs 53% in Bafang, 33% vs 59% in Bafia, 27% vs 47% in Yabassi; p<0.01). The most common misconception about onchocerciasis transmission was the lack of hygiene, especially in Bafia and Yabassi. In Bafang, high proportions of people believed that onchocerciasis was due to high consumption of sugar (31% vs less than 5% in Bafia and Yabassi, p<0.001). CONCLUSION/SIGNIFICANCE: There are still frequent misconceptions about onchocerciasis transmission in Cameroon. Perceptions of ivermectin distribution campaigns are more strongly associated to adherence. In addition to education/sensitisation on onchocerciasis during the implementation of the CDTI program, local health authorities should strive to better involve communities and more encourage community distributors' work.