Combining an expert-based medical entity recognizer to a machine-learning system: methods and a case study

A Novel Approach towards Medical Entity Recognition in Chinese Clinical Text

Medical entity recognition, a basic task in the language processing of clinical data, has been extensively studied in analyzing admission notes in alphabetic languages such as English. However, much less work has been done on nonstructural texts that are written in Chinese, or in the setting of differentiation of Chinese drug names between traditional Chinese medicine and Western medicine. Here, we propose a novel cascade-type Chinese medication entity recognition approach that aims at integrating the sentence category classifier from a support vector machine and the conditional random field-based medication entity recognition. We hypothesized that this approach could avoid the side effects of abundant negative samples and improve the performance of the named entity recognition from admission notes written in Chinese. Therefore, we applied this approach to a test set of 324 Chinese-written admission notes with manual annotation by medical experts. Our data demonstrated that this approach had a score of 94.2% in precision, 92.8% in recall, and 93.5% in F-measure for the recognition of traditional Chinese medicine drug names and 91.2% in precision, 92.6% in recall, and 91.7% F-measure for the recognition of Western medicine drug names. The differences in F-measure were significant compared with those in the baseline systems.

Using Supervised Machine Learning to Classify Real Alerts and Artifact in Online Multisignal Vital Sign Monitoring Data

OBJECTIVE

The use of machine-learning algorithms to classify alerts as real or artifacts in online noninvasive vital sign data streams to reduce alarm fatigue and missed true instability.

DESIGN

Observational cohort study.

SETTING

Twenty-four-bed trauma step-down unit.

PATIENTS

Two thousand one hundred fifty-three patients.

INTERVENTION

Noninvasive vital sign monitoring data (heart rate, respiratory rate, peripheral oximetry) recorded on all admissions at 1/20 Hz, and noninvasive blood pressure less frequently, and partitioned data into training/validation (294 admissions; 22,980 monitoring hours) and test sets (2,057 admissions; 156,177 monitoring hours). Alerts were vital sign deviations beyond stability thresholds. A four-member expert committee annotated a subset of alerts (576 in training/validation set, 397 in test set) as real or artifact selected by active learning, upon which we trained machine-learning algorithms. The best model was evaluated on test set alerts to enact online alert classification over time.

MEASUREMENTS AND MAIN RESULTS

The Random Forest model discriminated between real and artifact as the alerts evolved online in the test set with area under the curve performance of 0.79 (95% CI, 0.67-0.93) for peripheral oximetry at the instant the vital sign first crossed threshold and increased to 0.87 (95% CI, 0.71-0.95) at 3 minutes into the alerting period. Blood pressure area under the curve started at 0.77 (95% CI, 0.64-0.95) and increased to 0.87 (95% CI, 0.71-0.98), whereas respiratory rate area under the curve started at 0.85 (95% CI, 0.77-0.95) and increased to 0.97 (95% CI, 0.94-1.00). Heart rate alerts were too few for model development.

CONCLUSIONS

Machine-learning models can discern clinically relevant peripheral oximetry, blood pressure, and respiratory rate alerts from artifacts in an online monitoring dataset (area under the curve > 0.87).

Real alerts and artifact classification in archived multi-signal vital sign monitoring data: implications for mining big data

Huge hospital information system databases can be mined for knowledge discovery and decision support, but artifact in stored non-invasive vital sign (VS) high-frequency data streams limits its use. We used machine-learning (ML) algorithms trained on expert-labeled VS data streams to automatically classify VS alerts as real or artifact, thereby "cleaning" such data for future modeling. 634 admissions to a step-down unit had recorded continuous noninvasive VS monitoring data [heart rate (HR), respiratory rate (RR), peripheral arterial oxygen saturation (SpO₂) at 1/20 Hz, and noninvasive oscillometric blood pressure (BP)]. Time data were across stability thresholds defined VS event epochs. Data were divided Block 1 as the ML training/cross-validation set and Block 2 the test set. Expert clinicians annotated Block 1 events as perceived real or artifact. After feature extraction, ML algorithms were trained to create and validate models automatically classifying events as real or artifact. The models were then tested on Block 2. Block 1 yielded 812 VS events, with 214 (26 %) judged by experts as artifact (RR 43 %, SpO₂ 40 %, BP 15 %, HR 2 %). ML algorithms applied to the Block 1 training/cross-validation set (tenfold cross-validation) gave area under the curve (AUC) scores of 0.97 RR, 0.91 BP and 0.76 SpO₂. Performance when applied to Block 2 test data was AUC 0.94 RR, 0.84 BP and 0.72 SpO₂. ML-defined algorithms applied to archived multi-signal continuous VS monitoring data allowed accurate automated classification of VS alerts as real or artifact, and could support data mining for future model building.