Prediction of postpartum depression using multilayer perceptrons and pruning.

OBJECTIVE. The main goal of this paper is to obtain a classification model based on feed-forward multilayer perceptrons in order to improve postpartum depression prediction during the 32 weeks after childbirth with a high sensitivity and specificity and to develop a tool to be integrated in a decision support system for clinicians. MATERIALS AND METHODS. Multilayer perceptrons were trained on data from 1397 women who had just given birth, from seven Spanish general hospitals, including clinical, environmental and genetic variables. A prospective cohort study was made just after delivery, at 8 weeks and at 32 weeks after delivery. The models were evaluated with the geometric mean of accuracies using a hold-out strategy. RESULTS. Multilayer perceptrons showed good performance (high sensitivity and specificity) as predictive models for postpartum depression. CONCLUSIONS. The use of these models in a decision support system can be clinically evaluated in future work. The analysis of the models by pruning leads to a qualitative interpretation of the influence of each variable in the interest of clinical protocols.

Long-term efficacy and safety of atazanavir/ritonavir treatment in a cohort of treatment-naïve HIV patients: An interim analysis of the REMAIN study.

Purpose: Combined antiretroviral therapy has dramatically improved HIV-infected individuals survival. Long-term strategies are currently needed to achieve the goal of durable virologic suppression. However, long-term available data for specific antiretrovirals (ARV) are limited. In clinical trials, boosted atazanavir (ATV/r) regimens has shown good efficacy and tolerability in ARV-naïve patients for up to 4 years. The REMAIN study aimed to evaluate the long-term outcomes of ATV/r regimens in ARV-naïve patients in a real life setting. Methods: Non-comparative, observational study conducted in Germany, Portugal and Spain. Historical and longitudinal follow-up data was extracted six monthly from the medical record of HIV-infected, treatment-naïve patients, who initiated an ATV/r-regimen between 2008 and 2010. The primary endpoint was the proportion of patients remaining on ATV treatment over time. Secondary endpoints included virologic response (HIV-1 RNA <50 c/mL and <500 c/mL), reasons for discontinuation and long-term safety. The duration of treatment and time to virologic failure (VF) were analyzed using the Kaplan- Meier method. Data from an interim analysis including patients with at least one year of follow-up are reported here. Results: A total of 411 patients were included in this interim analysis [median (Q1, Q3) follow-up: 23.42 (16.25, 32.24) months≥: 77% male; median age 40 years [min, max: 19, 78≥; 16% IDUs; 18% CDC C; 18% hepatitis C. TDF/FTC was the most common backbone (85%). At baseline, median (Q1, Q3) HIV-RNA and CD4 cell count were 4.91 (4.34, 5.34) log10 c/mL and 256 (139, 353) cells/mm3, respectively. The probability of remaining on treatment was 0.84 (95% CI: 0.80, 0.87) and 0.72 (95% CI: 0.67, 0.76) for the first and second year, respectively. After 2 years of follow-up, 84% (95% CI: 0.79, 0.88) of patients were virologically suppressed (<50 c/mL). No major protease inhibitors mutations were observed at VF. Overall, 125 patients (30%) discontinued ATV therapy [median (Q1, Q3) time to discontinuation: 11.14 (6.24, 19.35) months]. Adverse events (AEs) were the main reason for discontinuation (n =47, 11%). Hyperbilirubinaemia was the most common AE leading to discontinuation (14 patients). No unexpected AEs were reported. Conclusions: In a real life clinical setting, ATV/r regimens showed durable virologic efficacy with good tolerability in an ARV-naïve population. Data from longer follow-up will provide additional valuable information.