Using Natural Language Processing to Improve Discrete Data Capture From Interpretive Cervical Biopsy Diagnoses at a Large Health Care Organization.

CONTEXT.­: The terminology used by pathologists to describe and grade dysplasia and premalignant changes of the cervical epithelium has evolved over time. Unfortunately, coexistence of different classification systems combined with nonstandardized interpretive text has created multiple layers of interpretive ambiguity. OBJECTIVE.­: To use natural language processing (NLP) to automate and expedite translation of interpretive text to a single most severe, and thus actionable, cervical intraepithelial neoplasia (CIN) diagnosis. DESIGN.­: We developed and applied NLP algorithms to 35 847 unstructured cervical pathology reports and assessed NLP performance in identifying the most severe diagnosis, compared to expert manual review. NLP performance was determined by calculating precision, recall, and F score. RESULTS.­: The NLP algorithms yielded a precision of 0.957, a recall of 0.925, and an F score of 0.94. Additionally, we estimated that the time to evaluate each monthly biopsy file was significantly reduced, from 30 hours to 0.5 hours. CONCLUSIONS.­: A set of validated NLP algorithms applied to pathology reports can rapidly and efficiently assign a discrete, actionable diagnosis using CIN classification to assist with clinical management of cervical pathology and disease. Moreover, discrete diagnostic data encoded as CIN terminology can enhance the efficiency of clinical research.

Pregnancy, hormonal contraceptive use, and HIV-related death in Rwanda.

BACKGROUND: Few studies have examined the influence of pregnancy and hormonal contraception on HIV-related deaths in African women. Rwanda is a country with high fertility, high HIV prevalence, and frequent use of hormonal contraception in urban areas. METHODS: Data from a prospective cohort study of 460 HIV-infected urban childbearing (18-35 years) women followed at 6-monthly intervals for 6 years in Kigali, Rwanda, were analyzed. The relationship of time-dependent measures of pregnancy and hormonal contraceptive use to death from HIV disease was assessed with multivariate models. RESULTS: Incident pregnancy was not associated with elevated risk of death among HIV-infected women. Oral and injectable hormonal contraceptive use had borderline protective effects associated with reduced mortality (HR 0.40, 95% CI 0.15-1.07 and HR 0.48, 95% CI 0.21-1.08 for mortality, respectively) in a multivariate model including time-dependent measures. CONCLUSIONS: The results point to the benefits of integrating family planning and HIV services. In a highly pronatalist society, such as Rwanda, which is experiencing high HIV prevalence, service integration affords an opportunity to provide HIV testing to women at risk of pregnancy and to promote family planning among HIV-positive women.

The influence of informed consent content on study participants' contraceptive knowledge and concerns.

Little is known about how the information presented in the informed consent process influences study outcomes among participants. This study examines the influence of informed consent content on reported baseline contraceptive knowledge and concerns among two groups of HIV-serodiscordant and seroconcordant HIV-positive couples enrolled in research projects at an HIV research center in Lusaka, Zambia. We found significant differences in the reporting of contraceptive knowledge and concerns between couples viewing consent materials that included detailed information about contraception and those viewing consent materials that lacked the detailed information. We conclude that the design of informed consent materials should strike a balance between ensuring that participants give truly informed consent and educating participants in ways that do not compromise the assessment of the impact of behavioral interventions.