Search

The persistence and emergence of new multi-drug resistant Mycobacterium tuberculosis (M. tb) strains continues to advance the devastating tuberculosis (TB) epidemic. Robust systems are needed to accurately and rapidly perform drug-resistance profiling, and machine learning (ML) methods combined with genomic sequence data may provide novel insights into drug-resistance mechanisms. Using 372 M. tb isolates, the combined utility of ML and bioinformatics to perform drug-resistance profiling is demonstrated. SNPs, InDels, and dinucleotide frequencies are explored as input features for three ML models, namely Decision Trees, Random Forest, and the eXtreme Gradient Boosted model. Using SNPs and InDels, all three models performed equally well yielding a 99% accuracy, 97% recall, and 99% F1-score. Using dinucleotide frequencies, the XGBoost algorithm was superior with a 97% accuracy, 94% recall and 97% F1-score. This study validates the use of variants and presents dinucleotide features as another effective feature encoding method for ML-based phenotype classification.

Purpose: This qualitative study was conducted to explore the perceptions and concerns of Liberian RNs who work for international nongovernmental organizations (NGOs) in Liberia, and to elicit insights and suggestions about how to improve collaboration between national and expatriate nursing staff. Methods Five focus groups involving 19 nurses and an individual interview with one nurse were conducted. A semistructured interview guide was used to gather perspectives and garner suggestions. A phenomenological approach was used to identify emergent themes. Results Regarding positive experiences, participants from all 10 NGOs reported that working with expats gave them opportunities to be mentored, learn new skills, and try out new techniques and equipment. Negative experiences included problems with disrespect and expats' lack of cultural competency. Participants spoke of adverse patient outcomes that resulted when expats made assumptions and wouldn't listen to national staff. Conclusions The findings indicate that changes are warranted in the current work culture between national and expatriate nursing staff. Improving collaboration between national and expatriate RNs could lead to higher-quality care and better patient outcomes. Predeployment training for expatriates that incorporates the voices of national staff is recommended; this would raise awareness of the realities of nursing in low-resource settings and promote respect for national staff.

Purpose: This qualitative study was conducted to explore the perceptions and concerns of Liberian RNs who work for international nongovernmental organizations (NGOs) in Liberia, and to elicit insights and suggestions about how to improve collaboration between national and expatriate nursing staff.Five focus groups involving 19 nurses and an individual interview with one nurse were conducted. A semistructured interview guide was used to gather perspectives and garner suggestions. A phenomenological approach was used to identify emergent themes.Regarding positive experiences, participants from all 10 NGOs reported that working with expats gave them opportunities to be mentored, learn new skills, and try out new techniques and equipment. Negative experiences included problems with disrespect and expats' lack of cultural competency. Participants spoke of adverse patient outcomes that resulted when expats made assumptions and wouldn't listen to national staff.The findings indicate that changes are warranted in the current work culture between national and expatriate nursing staff. Improving collaboration between national and expatriate RNs could lead to higher-quality care and better patient outcomes. Predeployment training for expatriates that incorporates the voices of national staff is recommended; this would raise awareness of the realities of nursing in low-resource settings and promote respect for national staff.

The effect of reproductive history on the risk of cervical, colorectal and thyroid cancers and melanoma has been explored but the results to date are inconsistent. We aimed to examine in a record- linkage cohort study the risk of developing these cancers, as well as breast, ovarian and endometrial cancers, among mothers who had given birth to twins compared with those who had only singleton pregnancies. Women who delivered a baby in Sweden between 1961 and 1996 and who were 15 years or younger in 1961 were selected from the Swedish civil birth register and linked with the Swedish cancer registry. We used Poisson regression to assess associations between reproductive factors and cancer. Twinning was associated with reduced risks of breast, colorectal, ovarian and uterine cancers, although no relative risks were statistically significant. The delivery of twins did not increase the risk of any cancers studied. Increasing numbers of maternities were associated with significantly reduced risks of all tumors except thyroid cancer. We found positive associations between a later age at first birth and breast cancer and melanoma, while there were inverse associations with cervix, ovarian, uterine and colorectal cancers. These findings lend weight to the hypothesis that hormonal factors influence the etiology of colorectal cancer in women, but argue against any strong effect of hormones on the development of melanoma or tumors of the thyroid.

Because the sources of fine-grained sediment problems in river systems are diffuse, it is essential to assemble catchment scale information for informing management strategies. Sediment source-tracing procedures have increasingly been adopted in this respect. Accordingly, a recently refined composite tracing procedure was used to investigate contemporary sources of fine-grained channel bed sediment in 11 sub-catchments (364km2) of the River Wensum Demonstration Test Catchment, in the eastern UK. The procedure incorporated a combination of statistical tests for discriminating source end members, plus numerical mass balance modelling incorporating weightings for within-source tracer variations and tracer-specific discriminatory power, as well as a combination of local and genetic algorithm optimisation coupled with Monte Carlo uncertainty analysis. Replicate Monte Carlo runs demonstrated the convergence of the modelling outputs within +/- 1% of the average medians. Relative frequency-weighted average median source type contributions were estimated to range between 27%69% (agricultural topsoils; predicted deviate median inputs 0%98% and 9%100%), 0%38% (damaged road verges; predicted deviate median inputs 0%58% and 0%100%) and 21%48% (channel banks/subsurface sources; predicted deviate median inputs 0%50% and 4%50%). The study provides further evidence of the importance of channel banks and damaged road verges as sediment sources and the need to include such areas in catchment sediment management strategies. Copyright (c) 2012 John Wiley & Sons, Ltd.