Influence of health-insurance on treatment outcome of childhood cancer in Western Kenya.

BACKGROUND: Few governments in low and middle-income countries (LMIC) have responded favourably to the international plea for Universal Health Coverage. Childhood cancer survival in LMIC is often below 20%. Limited health-insurance coverage may contribute to this poor survival. Our study explores the influence of health-insurance status on childhood cancer treatment outcomes in a Kenyan academic hospital. METHODS: This was a retrospective medical records review of all children diagnosed with cancer at Moi Teaching and Referral Hospital between 2010 and 2016. Socio-demographic and clinical data was collected using a structured data collection form. Fisher's exact test, chi-squared test, Kaplan-Meier method, log-rank test and Cox proportional hazard model were used to evaluate relationships between treatment outcomes and patient characteristics. Study was approved by Institutional Research Ethics Committee. FINDINGS: From 2010-2016, 879 children were newly diagnosed with cancer. Among 763 patients whose records were available, 28% abandoned treatment, 23% died and 17% had progressive/relapsed disease resulting in 32% event-free survival. In total 280 patients (37%) had health-insurance at diagnosis. After active enrolment during treatment, total health-insurance registration level reached 579 patients (76%). Treatment outcomes differed by health-insurance status (P < 0.001). The most likely treatment outcome in uninsured patients was death (49%), whereas in those with health-insurance at diagnosis and those who enrolled during treatment it was event-free survival (36% and 41% respectively). Overall survival (P < 0.001) and event-free survival (P < 0.001) were higher for insured versus uninsured patients. The hazard-ratio for treatment failure was 0.30 (95% CI:0.22-0.39; P < 0.001) for patients insured at diagnosis and 0.32 (95% CI:0.24-0.41; P < 0.001) for patients insured during treatment in relation to those without insurance. INTERPRETATION: Our study highlights the need for Universal Health Coverage in LMIC. Children without health-insurance had significantly lower survival. Childhood cancer treatment outcomes can be ameliorated by strategies that improve health-insurance access.

Sex differences in incidence of out-of-hospital cardiac arrest across ethnic and socioeconomic groups: A population-based cohort study in the Netherlands.

BACKGROUND: Insight into the occurrence of out-of-hospital cardiac arrest (OHCA) within general populations may help to target prevention strategies. Case registries suggest that there may be substantial differences in emergency medical service (EMS)-attended OHCA incidence between men and women, but relative sex differences across ethnic groups and socioeconomic (SES) groups have not been studied. We investigated sex differences in OHCA incidence, overall and across these subgroups. METHODS: We performed a retrospective population-based cohort study, combining individual-level data on ethnicity and income (as SES measure) from Statistics Netherlands of all men and women aged ≥25 years living in one study region in the Netherlands on 01-01-2009 (n = 1,688,285) with prospectively collected EMS-attended OHCA cases (n = 5676) from the ARREST registry until 31-12-2015. We calculated age-standardised incidence rates of OHCA. Sex differences were assessed with Cox proportional hazards regression analyses, adjusted for age, ethnicity and income, in the overall population, and across ethnic and SES groups. RESULTS: The age-standardised incidence rate of OHCA was lower in women than in men (30.9 versus 87.3 per 100,000 person-years), corresponding with a hazard ratio (HR) of 0.33 (95% confidence interval [CI] 0.31-0.35). These sex differences in hazard for OHCA existed in all income quintiles (HR range: 0.30-0.35) and ethnic groups (HR range: 0.19-0.40), except among Moroccans (HR 0.89, 95% CI 0.51-1.57). CONCLUSION: Women have a substantial, yet lower OHCA incidence rate than men. The magnitude of these sex differences did not vary across social strata.