Natural language processing offers complete view of cancer patients' pain
Natur al language processing (NLP) offers tools to improve monitoring of patient pain statuses, according to research published online Nov. 9 by the Journal of the American Medical Informatics Association.
A research team including software developers, medical subject matter experts and statisticians constructed an NLP tool relying on 42 pain severity contextual rules and a pain categorization scale, and retroactively applied it to EHRs documenting 4,409 clinical encounters of 33 men participating in a 15-year study of metastatic prostate cancer. The NLP tool identified 6,387 pain and 13,827 drug mentions in unstructured text, which were translated and normalized as structured text using terms in the Unified Medical Language System and the pain categorization scale. The results also were rendered as a graphical representation to demonstrate how the tool could be used to provide clinicians with a view of patients' pain over time.
Previous research has manually normalized and classified pain to demonstrate statistical correlations and the results returned by the NLP tool were consistent with existing clinical expectations, such as the associations of severe pain with the receipt of opioids and palliative radiation, according to lead author Norris H. Heintzelman, MS, a systems engineer at Lockheed Martin, and her colleagues. The importance of their research is to demonstrate the feasibility of an automated NLP tool’s ability to mine text for the purposes of charting pain, which could be useful for monitoring pain management and identifying novel cancer phenotypes.
“Text in longitudinal data is valuable for the study of symptoms such as pain, where the clinical unstructured description may be more complete than it is in structured data,” Heintzelman et al wrote. “NLP techniques convert such unstructured data into structures data, which is typically more amenable to rigorous analysis and display.”
“Relief of pain is essential in the management of many acute and chronic diseases, and convenient automated monitoring of patient pain status could provide a valuable new tools for improving quality of life and car,” they concluded. “Real-time, easy-to-interpret views of the pain status history of an individual patient or group of patients could allow busy clinicians to identify patients most in need of increased pain management intensity, and allow researchers to perform visual and quantitative comparison of groups of subjects participating in clinical trials of novel therapies or novel clinical interventions.”