Evaluating the utility of mid-infrared spectral subspaces for predicting soil properties.

We propose four methods for finding local subspaces in large spectral libraries. The proposed four methods include (a) cosine angle spectral matching; (b) hit quality index spectral matching; (c) self-organizing maps and (d) archetypal analysis methods. Then evaluate prediction accuracies for global and subspaces calibration models. These methods were tested on a mid-infrared spectral library containing 1907 soil samples collected from 19 different countries under the Africa Soil Information Service project. Calibration models for pH, Mehlich-3 Ca, Mehlich-3 Al, total carbon and clay soil properties were developed for the whole library and for the subspace. Root mean square error of prediction was used to evaluate predictive performance of subspace and global models. The root mean square error of prediction was computed using a one-third-holdout validation set. Effect of pretreating spectra with different methods was tested for 1st and 2nd derivative Savitzky-Golay algorithm, multiplicative scatter correction, standard normal variate and standard normal variate followed by detrending methods. In summary, the results show that global models outperformed the subspace models. We, therefore, conclude that global models are more accurate than the local models except in few cases. For instance, sand and clay root mean square error values from local models from archetypal analysis method were 50% poorer than the global models except for subspace models obtained using multiplicative scatter corrected spectra with which were 12% better. However, the subspace approach provides novel methods for discovering data pattern that may exist in large spectral libraries.

An exploration of opportunities and challenges facing cervical cancer managers in Kenya.

BACKGROUND: Kenya like other developing countries is low in resource setting and is facing a number of challenges in the management of cervical cancer. This study documents opportunities and challenges encountered in managing cervical cancer from the health care workers' perspectives. A qualitative study was conducted among cervical cancer managers who were defined as nurses and doctors involved in operational level management of cervical cancer. The respondents were drawn from four provincial hospitals and the only two main National public referral hospitals in Kenya. Twenty one [21] nurse managers and twelve [12] medical doctors were interviewed using a standardized interview guide. The responses were audio recorded, transcribed verbatim and the content analyzed in emerging themes. FINDINGS: Four themes were identified. Patient related challenges included a large number of patients, presenting in the late stage of disease, low levels of knowledge on cancer of the cervix, low levels of screening and a poor attitude towards screening procedure. Individual health care providers identified a lack of specialised training, difficulty in disclosure of diagnosis to patients, a poor attitude towards cervical cancer screening procedure and a poor attitude towards cervical cancer patients. Health facilities were lacking in infrastructure and medical supplies. Some managers felt ill-equipped in technological skills while the majority lacked access to the internet. Mobile phones were identified as having great potential for improving the management of cervical cancer in Kenya. CONCLUSION: Kenya faces a myriad of challenges in the management of cervical cancer. The peculiar negative attitude towards screening procedure and the negative attitude of some managers towards cervical cancer patients need urgent attention. The potential use of mobile phones in cervical cancer management should be explored.

Self-reported use of internet by cervical cancer clients in two National Referral Hospitals in Kenya.

BACKGROUND: Cervical cancer remains a devastating disease in Kenya accounting for more than 2000 deaths each year. Lack of information on cervical cancer prevention and management has been attributed to the apathy among women in seeking health interventions. Use of internet-based and mobile e-health tools could increase information access among cervical cancer patients. The objective of the study was; to establish the extent of use of mobile phones and internet by cervical cancer patients in accessing information related to cancer treatment and management.; find out the characteristics of patients associated with internet use and identify barriers faced by the patients in internet use. A cross sectional descriptive survey of 199 cervical patients visiting the two main referral hospitals in Kenya was done. A structured questionnaire was used to collect data. FINDINGS: The average length of illness was 2.43 years (SD ± 3.0). Only 7.5 %( n=15) reported to having used the internet as a source of information. 92.5 %( n=184) did not use internet. With Multiple options, 70.9% did not know how to use a computer, 29.2% did not have access to a computer, 14.6% lacked the money to use computers at the local cyber cafe while other barriers identified accounted for 11.1%. Patients reported that the internet had an important role in the management of cancer of the cervix in health education (17.6%), online consultation (14.6%), booking of patients (13.6%), referrals (8.5%) and collecting data (7%). The 96.5% of the respondents who had access to a mobile phone, recommended mobile phones for health education messages (31.7%), reminder alerts for medication (29.7%) and booking appointments (21.6%). There was a statistically significant association between income of the patients and internet use (p = 0.026) in this study. CONCLUSIONS: There is low level use of the internet by cervical cancer clients attended in Public referral facilities in Kenya. This was attributed to; lack of knowledge on how to use computers and lack of access to a computer. High level of access to mobile phones was reported. This is an indicator of great potential for use of mobile phones in the management of cervical cancer through short messaging services (sms), without internet connectivity. There is even greater potential to internet use through web access via mobile phones.

Regression and simulation models for human and baboon brain parameters / Modelo regresivo y simulación para los parámetros cerebrales humanos y del babuino

The variations in morphometric parameter of mammalian brains may be influenced by process of functional complexity, evolution and adaptation. Comparative analysis of linear measurements of cerebrum in the human and baboon has shown morphometric differences. In the present study linear measurements from human and baboon cerebrum (n=10 each) were used to predict various values for human and baboon brain and body parameters through multiple regression models. The average brain weights were found to be 2.08 percent and 0.84 percent of the body weights for humans and baboons respectively. The elasticity of regression models revealed that unit percentage increase in Occipital-Frontal (OF) distance would increase the human brain weight by 66.19 percent, while the baboon brain weight would increase by 7.63 percent. The unit percentage increase in the Height of Temporal Lobe (HTL) would increase the human brain weight by 16.28 percent, while the baboon brain weight would increase by only 0.28 percent. Unit percentage increase in Frontal-Temporal (FT) distance would decrease the human and baboon brain weights by 14.04 percent and 0.46 percent respectively. Inter-species values were also predicted through simulation techniques by using the ratios of model parameters with application of programming language Python. The OF, FT and HTL values for human were found to be 2.01 times, 1.55 times and 1.91 times respectively to that of baboon.

Las variaciones en los parámetros morfométricos del cerebro de los mamíferos pueden estar influenciadas por el proceso de complejidad funcional de la evolución y adaptación. Análisis comparativo de las mediciones lineales del cerebro en el humano y babuino han puesto de manifiesto las diferencias morfométricas. En este estudio las mediciones lineales del cerebro humano y babuinos (n = 10 cada uno) fueron utilizados para predecir los valores distintivos para el cerebro de humanos y monos babuinos y los parámetros del cuerpo a través de modelos de regresión múltiple. El peso medio del cerebro resultó ser 2,08 por ciento y 0,84 por ciento del peso corporal de los seres humanos y los babuinos, respectivamente. La elasticidad de los modelos de regresión reveló que el aumento de una unidad porcentual en la distancia occipital-frontal (DE) aumentaría el peso del cerebro humano en 66,19 por ciento, mientras que el peso del cerebro babuino se incrementaría en 7,63 por ciento. El porcentaje de aumento en la altura de lóbulo temporal (HTL) aumentaría el peso del cerebro humano en 16,28 por ciento, mientras que el peso del cerebro babuino aumentaría en sólo el 0,28 por ciento. Si aumenta la distancia frontal-temporal (FT) se reduciría el peso del cerebro humano y babuinos en 14,04 por ciento y 0,46 por ciento, respectivamente. También se prevéen valores entre las especies a través de técnicas de simulación, mediante el uso de proporciones de los parámetros del modelo con la aplicación del lenguaje de programación Python. Los valores humanos de DE, FT y HTL resultaron ser 2,01, 1,55 y 1,91 veces, respectivamente con respecto a la de los babuinos.