Contextual property detection in Dutch diagnosis descriptions for uncertainty, laterality and temporality

Abstract

Background Accurate, coded problem lists are valuable for data reuse, including clinical decision support and research. However, healthcare providers frequently modify coded diagnoses by including or removing common contextual properties in free-text diagnosis descriptions: uncertainty (suspected glaucoma), laterality (left glaucoma) and temporality (glaucoma 2002). These contextual properties could cause a difference in meaning between underlying diagnosis codes and modified descriptions, inhibiting data reuse. We therefore aimed to develop and evaluate an algorithm to identify these contextual properties. Methods A rule-based algorithm called UnLaTem (Uncertainty, Laterality, Temporality) was developed using a single-center dataset, including 288,935 diagnosis descriptions, of which 73,280 (25.4%) were modified by healthcare providers. Internal validation of the algorithm was conducted with an independent sample of 980 unique records. A second validation of the algorithm was conducted with 996 records from a Dutch multicenter dataset including 175,210 modified descriptions of five hospitals. Two researchers independently annotated the two validation samples. Performance of the algorithm was determined using means of the recall and precision of the validation samples. The algorithm was applied to the multicenter dataset to determine the actual prevalence of the contextual properties within the modified descriptions per specialty. Results For the single-center dataset recall (and precision) for removal of uncertainty, uncertainty, laterality and temporality respectively were 100 (60.0), 99.1 (89.9), 100 (97.3) and 97.6 (97.6). For the multicenter dataset for removal of uncertainty, uncertainty, laterality and temporality it was 57.1 (88.9), 86.3 (88.9), 99.7 (93.5) and 96.8 (90.1). Within the modified descriptions of the multicenter dataset, 1.3% contained removal of uncertainty, 9.9% uncertainty, 31.4% laterality and 9.8% temporality. Conclusions We successfully developed a rule-based algorithm named UnLaTem to identify contextual properties in Dutch modified diagnosis descriptions. UnLaTem could be extended with more trigger terms, new rules and the recognition of term order to increase the performance even further. The algorithm’s rules are available as additional file 2. Implementing UnLaTem in Dutch hospital systems can improve precision of information retrieval and extraction from diagnosis descriptions, which can be used for data reuse purposes such as decision support and research.

To assess the impact of problem lists' accuracy in electronic health records on clinical decision-making: protocol for a randomized controlled trial

“…Problems can be added, closed and removed and healthcare providers working at our institution can add free-text details in the display name. For instance, if a diagnosis is not certain, a healthcare provider might temporarily add suspected to the diagnosis 'eye infection' in a free-text display eld [52]. Once the diagnosis is con rmed, these details should be removed.…”

“…A problem can also be removed completely although this will be logged. Additionally, details added to a display eld can include laterality (left eye infection), temporality (eye infection 2002) and other properties specifying a diagnosis [52]. Details can also be included in problem list notes, for instance details that describe the care and diagnosis process in a chronological order.…”

Applying Natural Language Processing to Textual Data From Clinical Data Warehouses: Systematic Review

“…Linguistic occurrences are particularly relevant where medical information is concerned, such as negation, temporality, uncertainty, or experiencer (ie, determine whether the identified information is related to the patient or a third party, such as a family member). In the included studies, rule-based methods were often used to detect contextual information in clinical text [178,180,182]. Although these methods offer good results (with an approximate F 1 -measure value of 0.90), they rely on handmade resources, such as terminologies and regular expressions, and customization is often needed for specific use cases.…”

“…Publications, n (%) Data language [15-17,19-25,27-38,41-48,50,51,53-68,71,72,74,75,78-80,83-88,90-92,96,97,99-112,114,116,119-124,126-135, 137,140,142-149,151-154,156-159,161,162,165-176,179,181,184-187,189-192,194-196,198,200-208] 153 (78.9) English [39,49,52,73,76,77,81,89,93,94,113,115,118,125,136,138,139,155,163,164,177,178,182,183,188,193,197] 27 (13.9) French [18,26,69,95,117,150,160,180,199] 9 (4.6) German [40,65,82] 3 (1.5) Korean…”