Natural language processing for structuring clinical text data on depression using UK-CRIS.

BACKGROUND: Utilisation of routinely collected electronic health records from secondary care offers unprecedented possibilities for medical science research but can also present difficulties. One key issue is that medical information is presented as free-form text and, therefore, requires time commitment from clinicians to manually extract salient information. Natural language processing (NLP) methods can be used to automatically extract clinically relevant information. OBJECTIVE: Our aim is to use natural language processing (NLP) to capture real-world data on individuals with depression from the Clinical Record Interactive Search (CRIS) clinical text to foster the use of electronic healthcare data in mental health research. METHODS: We used a combination of methods to extract salient information from electronic health records. First, clinical experts define the information of interest and subsequently build the training and testing corpora for statistical models. Second, we built and fine-tuned the statistical models using active learning procedures. FINDINGS: Results show a high degree of accuracy in the extraction of drug-related information. Contrastingly, a much lower degree of accuracy is demonstrated in relation to auxiliary variables. In combination with state-of-the-art active learning paradigms, the performance of the model increases considerably. CONCLUSIONS: This study illustrates the feasibility of using the natural language processing models and proposes a research pipeline to be used for accurately extracting information from electronic health records. CLINICAL IMPLICATIONS: Real-world, individual patient data are an invaluable source of information, which can be used to better personalise treatment.

Management of paediatric burns in low- and middle-income countries: assessing capacity using the World Health Organization Surgical Assessment Tool.

BACKGROUND: Burns are a leading cause of global disease burden, with children in low- and middle-income countries (LMICs) disproportionately affected. Effective management improves outcomes; however, the availability of necessary resources in LMICs remains unclear. We evaluated surgical centres in LMICs using the WHO Surgical Assessment Tool (SAT) to identify opportunities to optimize paediatric burn care. METHODS: We reviewed WHO SAT database entries for 2010-2015. A total of 1121 facilities from 57 countries met the inclusion criteria: facilities with surgical capacity in LMICs operating on children. Human resources, equipment and infrastructure relevant to paediatric burn care were analysed by WHO Regional and World Bank Income Classifications and facility type. RESULTS: Facilities had an average of 147 beds and performed 485 paediatric operations annually. Discrepancies existed between procedures performed and resource availability; 86% of facilities performed acute burn care, but only 37% could consistently provide intravenous fluids. Many, particularly tertiary, centres performed contracture release and skin grafting (41%) and amputation (50%). CONCLUSIONS: LMICs have limited resources to provide paediatric burn care but widely perform many interventions necessary to address the burden of burns. The SAT may not capture innovative and traditional approaches to burn care. There remains an opportunity to improve paediatric burn care globally.

Recurrent sampling and ventriculostomy-associated infections: a case-control study.

BACKGROUND: External ventricular drain (EVD) insertion is a common neurosurgical procedure with a significant risk of ventriculostomy-associated infections (VAIs), other morbidities and mortality. Several risk factors have been identified but their effect is unclear. OBJECTIVE: Our primary objective was to assess whether recurrent EVD sampling increased the risk of VAIs. The secondary objective was to explore the effect of sampling frequency, duration of EVD stay, presence of cerebrospinal fluid leak and concurrent infections. METHODS: Ours was a retrospective, single-centre, age-matched, case control study of 83 patients and 249 controls who underwent EVD insertion between 1 January 2010 - 31 December 2016. Patients with primary CNS infections, age under 1 year and death within 5 days of EVD insertion were excluded. Blinded abstractors collected all patient data through the electronic patient record system and regression analysis was used to compare the two groups. RESULTS: Sampling the EVD more than once did not increase the risk of infection (OR 0.47, 95% CI 0.27-0.83, p = 0.01). Risk factors significantly associated with VAI included CSF leak (OR 2.06, 95% CI 1.16-3.67, p = 0.01), concurrent infection (OR 1.85, 95% CI 1.09-3.16, p = 0.02), and an EVD duration of >10 days (OR 2.28, 95% CI 1.24-4.18, p = 0.01). CONCLUSION: Our findings do not support the notion that increased sampling of EVDs is associated with a higher risk of VAI. CSF leaks, concurrent infection and long-term EVDs are also statistically significant risk factors for infection.

Using the SAPAS to identify risk for personality disorders among psychiatric outpatients in India: A feasibility study.

Personality disorders (PDs) are common among psychiatric outpatients and are associated with increased morbidity and worse treatment outcomes. Epidemiological research conducted among this population in Asian countries is limited, reflecting a significant gap in the current literature. One barrier to this research is the lack of appropriate screening tools. The current research assessed the feasibility of using the SAPAS (Standardized Assessment of Personality-Abbreviated Scale) screening tool to identify individuals at high risk of PD in an Indian psychiatric outpatient population and provides an initial estimate of PD prevalence by using a validated diagnostic interview, the ICD-10 International Personality Disorder Examination. The findings suggest that whilst use of the SAPAS was feasible, acceptable to patients and led to clinically useful findings, when using the recommended cut-off score of 4, the SAPAS largely overdiagnoses the risk for PD in psychiatric outpatients in India (positive predictive value = 26.3%). The estimated prevalence of personality disorder in the sample was 11.1%, based on administering the International Personality Disorder Examination diagnostic interview to high-risk patients scoring 4 and above on the SAPAS, which is higher than previous estimates for this population and still likely to be an underestimation. Future studies should translate the measure into Bengali and evaluate its sensitivity and specificity at different cut-off points in order to optimize its use in Indian populations. Copyright © 2017 John Wiley & Sons, Ltd.

Designing adverse event forms for real-world reporting: participatory research in Uganda.

The wide-scale roll-out of artemisinin combination therapies (ACTs) for the treatment of malaria should be accompanied by continued surveillance of their safety. Post-marketing pharmacovigilance (PV) relies on adverse event (AE) reporting by clinicians, but as a large proportion of treatments are provided by non-clinicians in low-resource settings, the effectiveness of such PV systems is limited. To facilitate reporting, AE forms should be easily completed; however, most are challenging for lower-level health workers and non-clinicians to complete. Through participatory research, we sought to develop user-friendly AE report forms to capture information on events associated with ACTs.Following situation analysis, we undertook workshops with community medicine distributors and health workers in Jinja, Uganda, to develop a reporting form based on experiences and needs of users, and communication and visual perception principles. Participants gave feedback for revisions of subsequent versions. We then conducted 8 pretesting sessions with 77 potential end users to test and refine passive and active versions of the form.The development process resulted in a form that included a pictorial storyboard to communicate the rationale for the information needed and facilitate rapport between the reporter and the respondent, and a diary format to record the drug administration and event details in chronological relation to each other. Successive rounds of pretesting used qualitative and quantitative feedback to refine the form, with the final round showing over 80% of the form completed correctly by potential end users.We developed novel AE report forms that can be used by non-clinicians to capture pharmacovigilance data for anti-malarial drugs. The participatory approach was effective for developing forms that are intuitive for reporters, and motivating for respondents. The forms, or their key components, could be adapted for use in other low-literacy settings to improve quality and quantity of drug safety reports as new medicines are scaled-up.

Molecules and mating: positive selection and reproductive behaviour in primates.

Sexual reproduction is generally thought to be more costly than asexual reproduction. However, it does have the advantage of accelerating rates of adaptation through processes such as recombination and positive selection. Comparative studies of the human and nonhuman primate genomes have demonstrated that positive selection has played an important role in the evolutionary history of humans and other primates. To date, many dozens of genes, thought to be affected by positive selection, have been identified. In this chapter, we will focus on genes that are associated with mating behaviours and reproductive processes, concentrating on genes that are most likely to enhance reproductive success and that also show evidence of positive selection. The genes encode phenotypic features that potentially influence mate choice decisions or impact the evolution and function of genes involved in the perception and regulation of, and the response to, phenotypic signals. We will also consider genes that influence precopulatory behavioural traits in humans and nonhuman primates, such as social bonding and aggression. The evolution of post-copulatory strategies such as sperm competition and selective abortion may also evolve in the presence of intense competition and these adaptations will also be considered. Although behaviour may not be solely determined by genes, the evidence suggests that the genes discussed in this chapter have some influence on human and nonhuman primate behaviour and that positive selection on these genes results in some degree of population differentiation and diversity.