The data in electronic health records can be mined to pull patient-centered outcome information from unstructured narrative text, according to an article in eGEMS (generating evidence and methods to improve patient outcomes).
Many diseases have patient-centered or reported outcomes that are not reliably recorded as coded data but rather in a clinician's free text, such as clinical notes. The narrative text is where one could find patient preferences, concerns and patient-centered outcomes.
The researchers, from Stanford University, wanted to test the feasibility and accuracy of identifying patient centered outcomes within the EHR. They used data from patients with localized prostate cancer undergoing prostatectomy to develop and test an algorithm to identify post-operative urinary incontinence, a common side effect. They mined the EHRs of 5,349 patients at Stanford between 1998-2013. They identified patients with such incontinence using both structured data (ICD-9 CM: 788.30) and unstructured free text such as "incontinence," "urinary leakage" and "wears adult diapers."
They found that 30.3 percent had a text mention of incontinence within 90 days after the operation, compared to less than 1 percent from data within the structured data field. The researchers noted that with a disease like prostate cancer, which has a high survival rate, patient-centered outcomes might be among the best quality measures of healthcare delivered.
"This report provides evidence that patient-centered outcomes are recorded in EHRs and that these data can be efficiently and accurately extracted. Extracting and analyzing patient-centered outcome data in a precise and timely manner is the first step in creating treatment pathways that reflect the patients' individual risk values," the researchers concluded. The study also noted the importance of good clinical documentation.
To learn more:
- read the study