Increasing digitisation has put pressure on institutions of higher learning around the globe to find effective solutions to better prepare their students for the 21 st century world of work. The article presents the findings of a qualitative inquiry into the attitudes, feelings, and experiences of entrepreneurship students in a selected private institution of higher learning on the use of eportfolios as a tool to enhance the learning experience. Data was collected using a combination of focus groups and semi-structured interviews. A thematic analysis conducted for this study revealed five (5) emergent themes on the contributions of e-portfolios in enhancing the learning experiences of the modern-day scholar. The resourcefulness, flexibility, engagement, entrepreneurial skills improvement and collaboration emerged as the outstanding contributions of e-portfolios towards enhancing learning. The emergent themes are in line with the entrepreneurial behaviour, imperative for every 21 st century world of work. The value the findings of this study lies in its exploratory utility in exposing the lived-experiences, attitudes, and feelings of modern day students on the role of e-portfolios as a tool to enhance their learning experience and cultivate entrepreneurial behaviour in the modern world of work. Recommendations to explore the technological advantage presented by e-portfolios are presented.

Background: The process of determining causes of death in areas where there is limited clinical services using verbal autopsies has become a key issue in terms of accuracy on cause of death (prone to errors and subjective), quality of data among many drawbacks. This is mainly because there is no proper standard available in performing verbal autopsy, even though it is important for civil registration systems and strengthening of health priorities. Physician diagnosis is the only gold standard in reviewing verbal autopsy narratives. In practice, conventional statistical methods are used to perform verbal autopsies due to their simplicity and transparency. However, in literature complex machine learning models can be found that can replace the traditional statistical methods. There has not been much application of machine learning techniques in verbal autopsy to determine cause of death, despite the advances in technology. As such, there is a need for a thorough survey of recent literature on statistical and machine learning approaches applied in verbal autopsy to determine cause of death. Methods: A systematic review was conducted and included a search from six databases. Our study only included scientific articles published in last decade that reported on verbal autopsy and: (1) algorithms; (2) statistical techniques; (3) machine learning and (4) deep learning. The search yielded 110 articles, after meta analysis, we identified 85 articles as being relevant and discarded the other 25. We investigated and compared the most commonly used statistical and machine learning techniques in VAs, identified limitations of each of these techniques, proposed a guiding machine learning framework and pointed to future directions. Results: Eighty five studies met the inclusion criteria. Apart from physician diagnosis, statistical methods are the most currently applied tools to determine cause of death from verbal autopsies. However, there has been little application of traditional machine learning and emerging techniques, even though they have shown promising results in other domains. Conclusions: Technological application of machine learning to determine cause of death, should focus on effective ideal strategies of pre-processing, transparency, robust feature engineering techniques and data balancing in order to attain optimal model performance.

Computer Coded Verbal Autopsy (CCVA) algorithms are commonly used to determine the cause of death (CoD) from questionnaire responses extracted from verbal autopsies (VAs). However, they can only operate on structured data and cannot effectively harness information from unstructured VA narratives. Machine Learning (ML) algorithms have also been applied successfully in determining the CoD from VA narratives, allowing the use of auxiliary information that CCVA algorithms cannot directly utilize. However, most ML-based studies only use responses from the structured questionnaire, and the results lack generalisability and comparability across studies. We present a comparative performance evaluation of ML methods and CCVA algorithms on South African VA narratives data, using data from Agincourt Health and Demographic Surveillance Site (HDSS) with physicians' classifications as the gold standard. The data were collected from 1993 to 2015 and have 16,338 cases. The random forest and extreme gradient boosting classifiers outperformed the other classifiers on the combined dataset, attaining accuracy of 96% respectively, with significant statistical differences in algorithmic performance (p < 0.0001). All our models attained Area Under Receiver Operating Characteristics (AUROC) of greater than 0.884. The InterVA CCVA attained 83% Cause Specific Mortality Fraction accuracy and an Overall Chance-Corrected Concordance of 0.36. We demonstrate that ML models could accurately determine the cause of death from VA narratives. Additionally, through mortality trends and pattern analysis, we discovered that in the first decade of the civil registration system in South Africa, the average life expectancy was approximately 50 years. However, in the second decade, life expectancy significantly dropped, and the population was dying at a much younger average age of 40 years, mostly from the leading HIV related causes. Interestingly, in the third decade, we see a gradual improvement in life expectancy, possibly attributed to effective health intervention programmes. Through a structure and semantic analysis of narratives where experts disagree, we also demonstrate the most frequent terms of traditional healer consultations and visits. The comparative approach also makes this study a baseline that can be used for future research enforcing generalization and comparability. Future study will entail exploring deep learning models for CoD classification.

Background Prostate cancer (PCa) is the leading male neoplasm in South Africa with an age-standardised incidence rate of 68.0 per 100,000 population in 2018. The Gleason score (GS) is the strongest predictive factor for PCa treatment and is embedded within semi-structured prostate biopsy narrative reports. The manual extraction of the GS is labour-intensive. The objective of our study was to explore the use of text mining techniques to automate the extraction of the GS from irregularly reported text-intensive patient reports. Methods We used the associated Systematized Nomenclature of Medicine clinical terms morphology and topography codes to identify prostate biopsies with a PCa diagnosis for men aged > 30 years between 2006 and 2016 in the Gauteng Province, South Africa. We developed a text mining algorithm to extract the GS from 1000 biopsy reports with a PCa diagnosis from the National Health Laboratory Service database and validated the algorithm using 1000 biopsies from the private sector. The logical steps for the algorithm were data acquisition, pre-processing, feature extraction, feature value representation, feature selection, information extraction, classification, and discovered knowledge. We evaluated the algorithm using precision, recall and F-score. The GS was manually coded by two experts for both datasets. The top five GS were reported, with the remaining scores categorised as “Other” for both datasets. The percentage of biopsies with a high-risk GS (≥ 8) was also reported. Results The first output reported an F-score of 0.99 that improved to 1.00 after the algorithm was amended (the GS reported in clinical history was ignored). For the validation dataset, an F-score of 0.99 was reported. The most commonly reported GS were 5 + 4 = 9 (17.6%), 3 + 3 = 6 (17.5%), 4 + 3 = 7 (16.4%), 3 + 4 = 7 (14.7%) and 4 + 4 = 8 (14.2%). For the validation dataset, the most commonly reported GS were: (i) 3 + 3 = 6 (37.7%), (ii) 3 + 4 = 7 (19.4%), (iii) 4 + 3 = 7 (14.9%), (iv) 4 + 4 = 8 (10.0%) and (v) 4 + 5 = 9 (7.4%). A high-risk GS was reported for 31.8% compared to 17.4% for the validation dataset. Conclusions We demonstrated reliable extraction of information about GS from narrative text-based patient reports using an in-house developed text mining algorithm. A secondary outcome was that late presentation could be assessed.

Background: Prostate cancer (PCa) is the leading male neoplasm in South Africa with an age-standardised incidence rate of 68.0 per 100,000 population in 2018. The Gleason score (GS) is the strongest predictive factor for PCa treatment and is embedded within semi-structured prostate biopsy narrative reports. The manual extraction of the GS is labour-intensive. The objective of our study was to explore the use of text mining techniques to automate the extraction of the GS from irregularly reported text-intensive patient reports.Methods: We used the associated Systematized Nomenclature of Medicine clinical terms morphology and topography codes to identify prostate biopsies with a PCa diagnosis for men aged >30 years between 2006 and 2016 in the Gauteng Province, South Africa. We developed a text mining algorithm to extract the GS from 1,000 biopsy reports with a PCa diagnosis from the National Health Laboratory Service database and validated the algorithm using 1,000 biopsies from the private sector. The logical steps for the algorithm were data acquisition, pre-processing, feature extraction, feature value representation, feature selection, information extraction, classification, and discovered knowledge. We evaluated the algorithm using precision, recall and F-score. The GS was manually coded by two experts for both datasets. The top five GS were reported, with the remaining scores categorised as “Other” for both datasets. The percentage of biopsies with a high-risk GS (≥8) was also reported. Results: The first output reported an F-score of 0.99 that improved to 1.00 after the algorithm was amended (the GS reported in clinical history was ignored). For the validation dataset, an F-score of 0.99 was reported. The most commonly reported GS were 5+4=9 (17.6%), 3+3= 6 (17.5%), 4+3=7 (16.4%), 3+4=7 (14.7%) and 4+4=8 (14.2%). For the validation dataset, the most commonly reported GS were: (i) 3+3=6 (37.7%), (ii) 3+4=7 (19.4%), (iii) 4+3=7 (14.9%), (iv) 4+4=8 (10.0%) and (v) 4+5=9 (7.4%). A high-risk GS was reported for 31.8% compared to 17.4% for the validation dataset. Conclusions: We demonstrated reliable extraction of information about GS from narrative text-based patient reports using an in-house developed text mining algorithm. A secondary outcome was that late presentation could be assessed.