RESUMEN
Given growing specialization in medical care, optimal care may require regionalization, which may create access barriers. We tested this within a large prostate cancer (PC) screening program in Brazil. In 2004-2007, Barretos Cancer Hospital prospectively screened men for PC throughout rural Brazil. Men with abnormal screen were referred for follow-up and possible biopsy. We tested the link between distance from screening site to Barretos Cancer Hospital and risk of noncompliance with showing up for biopsy, PC on biopsy and, among those with PC, PC grade using crude and multivariable logistic regression analysis. Among 10,467 men undergoing initial screen, median distance was 257 km (IQR: 135-718 km). On crude and multivariable analyses, farther distance was significantly linked with biopsy noncompliance (OR/100 km: 0.83, P < 0.001). Among men who lived within 150 km of Barretos Cancer Hospital, distance was unrelated to compliance (OR/100 km: 1.09, P=0.87). There was no association between distance and PC risk or PC grade (all P > 0.25). In Brazil, where distances to referral centers can be large, greater distance was related to reduced biopsy compliance in a PC screening cohort. Among men who lived within 150 km, distance was unrelated to compliance. Care regionalization may reduce access when distances are large.
RESUMEN
OBJECTIVE: To compare the number of children needed to screen to identify a case of childhood dyslipidemia and estimate costs under universal vs targeted screening approaches. STUDY DESIGN: We constructed a decision-analytic model comparing the health system costs of universal vs targeted screening for hyperlipidemia in US children aged 10 years over a 1-year time horizon. Targeted screening was defined by family history: dyslipidemia in a parent and/or early cardiovascular disease in a first-degree relative. Prevalence of any hyperlipidemia (low-density lipoprotein [LDL] ≥130 mg/dL) and severe hyperlipidemia (LDL ≥190 mg/dL or LDL ≥160 mg/dL with family history) were obtained from published estimates. Costs were estimated from the 2016 Maryland Medicaid fee schedule. We performed sensitivity analyses to evaluate the influence of key variables on the incremental cost per case detected. RESULTS: For universal screening, the number needed to screen to identify 1 case was 12 for any hyperlipidemia and 111 for severe hyperlipidemia. For targeted screening, the number needed to screen was 7 for any hyperlipidemia and 49 for severe hyperlipidemia. The incremental cost per case detected for universal compared with targeted screening was $1980 for any hyperlipidemia and $32 170 for severe hyperlipidemia. CONCLUSIONS: Our model suggests that universal cholesterol screening detects hyperlipidemia at a low cost per case, but may not be the most cost-efficient way to identify children with severe hyperlipidemia who are most likely to benefit from treatment.