BackgroundChildhood pneumonia is a major cause of childhood illness and the second leading cause of child death globally. Understanding the costs associated with the management of childhood pneumonia is essential for resource allocation and priority setting for child health.MethodsWe conducted a systematic review to identify studies reporting data on the cost of management of pneumonia in children younger than 5 years old. We collected unpublished cost data on non–severe, severe and very severe pneumonia through collaboration with an international working group. We extracted data on cost per episode, duration of hospital stay and unit cost of interventions for the management of pneumonia. The mean (95% confidence interval, CI) and median (interquartile range, IQR) treatment costs were estimated and reported where appropriate.ResultsWe identified 24 published studies eligible for inclusion and supplemented these with data from 10 unpublished studies. The 34 studies included in the cost analysis contained data on more than 95 000 children with pneumonia from both low– and–middle income countries (LMIC) and high–income countries (HIC) covering all 6 WHO regions. The total cost (per episode) for management of severe pneumonia was US$ 4.3 (95% CI 1.5–8.7), US$ 51.7 (95% CI 17.4–91.0) and US$ 242.7 (95% CI 153.6–341.4)–559.4 (95% CI 268.9–886.3) in community, out–patient facilities and different levels of hospital in–patient settings in LMIC. Direct medical cost for severe pneumonia in hospital inpatient settings was estimated to be 26.6%–115.8% of patients’ monthly household income in LMIC. The mean direct non–medical cost and indirect cost for severe pneumonia management accounted for 0.5–31% of weekly household income. The mean length of stay (LOS) in hospital for children with severe pneumonia was 5.8 (IQR 5.3–6.4) and 7.7 (IQR 5.5–9.9) days in LMIC and HIC respectively for these children.ConclusionThis is the most comprehensive review to date of cost data from studies on the management of childhood pneumonia and these data should be helpful for health services planning and priority setting by national programmes and international agencies.

Skin detection is used in applications ranging from face detection, tracking body parts and hand gesture analysis, to retrieval and blocking objectionable content. In this paper, we investigate and evaluate (1) the effect of color space transformation on skin detection performance and finding the appropriate color space for skin detection, (2) the role of the illuminance component of a color space, (3) the appropriate pixel based skin color modeling technique and finally, (4) the effect of color constancy algorithms on color based skin classification. The comprehensive color space and skin color modeling evaluation will help in the selection of the best combinations for skin detection. Nine skin modeling approaches (AdaBoost, Bayesian network, J48, Multilayer Perceptron, Naive Bayesian, Random Forest, RBF network, SVM and the histogram approach of Jones and Rehg [15]) in six color spaces (IHLS, HSI, RGB, normalized RGB, YCbCr and CIELAB) with the presence or absence of the illuminance component are compared and evaluated. Moreover, the impact of five color constancy algorithms on skin detection is reported. Results on a database of 8991 images with manually annotated pixel-level ground truth show that (1) the cylindrical color spaces outperform other color spaces, (2) the absence of the illuminance component decreases performance, (3) the selection of an appropriate skin color modeling approach is important and that the tree based classifiers (Random forest, J48) are well suited to pixel based skin detection. As a best combination, the Random Forest combined with the cylindrical color spaces, while keeping the illumi- nance component outperforms other combinations, and (4) the usage of color constancy algorithms can improve skin detection performance.

Skin detection is used in applications ranging from face detection, tracking body parts and hand gesture analysis, to retrieval and blocking objectionable content. For robust skin segmentation and detection, we investigate color classification based on random forest. A random forest is a statistical framework with a very high generalization accuracy and quick training times. The random forest approach is used with the IHLS color space for raw pixel based skin detection. We evaluate random forest based skin detection and compare it to Bayesian network, Multilayer Perceptron, SVM, AdaBoost, Naive Bayes and RBF network. Results on a database of 8991 images with manually annotated pixel-level ground truth show that with the IHLS color space, the random forest approach outperforms other approaches. We also show the effect of increasing the number of trees grown for random forest. With fewer trees we get faster training times and with 10 trees we get the highest F-score.

BackgroundThe prevalence of hypertension is increasing in much of the South Asian region, including Nepal. This paper reports the prevalence and risk factors of hypertension and pre-hypertension among adult women in a rural community of Nepal.MethodsCross-sectional data on socioeconomic status (SES), lifestyle factors and blood pressure (BP) were collected from a cohort of 15,934 women in rural Nepal in 2006–08. Among a subsample (n = 1679), anthropometry and biomarkers of cardiovascular risk were measured.ResultsThe mean age of women was 34.2 years (range 16.4-71.2 years). More than three percent (3.3%) had hypertension and 14.4% had pre-hypertension. In an adjusted analysis, lower SES, especially lower household farm assets and storage of food for long term consumption, was associated with increased odds of hypertension (OR = 1.14 for mid-level SES and OR = 1.40 for low SES; p for trend < 0.01). Smoking, alcohol use and not working outside the home were also associated with higher risk. In a subsample, both systolic BP (SBP) and diastolic BP (DBP) were positively associated with high triglycerides (SBP β = 4.1 mm Hg; DBP β =3.6 mm Hg), high HbA1c (SBP β = 14.0; DBP β = 9.2), raised fasting glucose (SBP β = 10.0; DBP β = 6.9), high BMI (SBP β = 6.7; DBP β = 5.1) and high waist circumference (SBP β = 6.2; DBP β = 5.3) after adjusting for potential confounders (p for all <0.01).ConclusionsAlthough the prevalence of hypertension was low in this cohort, it was more prevalent among the poorer women and was strongly associated with other cardiovascular risks. These associations at a relatively young age may confer greater risk for cardiovascular disease among women in later life, indicating the need for interventions to reduce the progression from pre-hypertension to hypertension.