Automated anesthesia artifact analysis: can machines be trained to take out the garbage?

The electronic health record: does it enhance or distract from patient safety?

PURPOSE OF REVIEW

The electronic health record (EHR) is an invaluable tool that may be used to improve patient safety. With a variety of different features, such as clinical decision support and computerized physician order entry, it has enabled improvement of patient care throughout medicine. EHR allows for built-in reminders for such items as antibiotic dosing and venous thromboembolism prophylaxis.

RECENT FINDINGS

In anesthesiology, EHR often improves patient safety by eliminating the need for reliance on manual documentation, by facilitating information transfer and incorporating predictive models for such items as postoperative nausea and vomiting. The use of EHR has been shown to improve patient safety in specific metrics such as using checklists or information transfer amongst clinicians; however, limited data supports that it reduces morbidity and mortality.

SUMMARY

There are numerous potential pitfalls associated with EHR use to improve patient safety, as well as great potential for future improvement.

A Comparison of Five Algorithmic Methods and Machine Learning Pattern Recognition for Artifact Detection in Electronic Records of Five Different Vital Signs: A Retrospective Analysis

BACKGROUND

Research on electronic health record physiologic data is common, invariably including artifacts. Traditionally, these artifacts have been handled using simple filter techniques. The authors hypothesized that different artifact detection algorithms, including machine learning, may be necessary to provide optimal performance for various vital signs and clinical contexts.

METHODS

In a retrospective single-center study, intraoperative operating room and intensive care unit (ICU) electronic health record datasets including heart rate, oxygen saturation, blood pressure, temperature, and capnometry were included. All records were screened for artifacts by at least two human experts. Classical artifact detection methods (cutoff, multiples of SD [z-value], interquartile range, and local outlier factor) and a supervised learning model implementing long short-term memory neural networks were tested for each vital sign against the human expert reference dataset. For each artifact detection algorithm, sensitivity and specificity were calculated.

RESULTS

A total of 106 (53 operating room and 53 ICU) patients were randomly selected, resulting in 392,808 data points. Human experts annotated 5,167 (1.3%) data points as artifacts. The artifact detection algorithms demonstrated large variations in performance. The specificity was above 90% for all detection methods and all vital signs. The neural network showed significantly higher sensitivities than the classic methods for heart rate (ICU, 33.6%; 95% CI, 33.1 to 44.6), systolic invasive blood pressure (in both the operating room [62.2%; 95% CI, 57.5 to 71.9] and the ICU [60.7%; 95% CI, 57.3 to 71.8]), and temperature in the operating room (76.1%; 95% CI, 63.6 to 89.7). The CI for specificity overlapped for all methods. Generally, sensitivity was low, with only the z-value for oxygen saturation in the operating room reaching 88.9%. All other sensitivities were less than 80%.

CONCLUSIONS

No single artifact detection method consistently performed well across different vital signs and clinical settings. Neural networks may be a promising artifact detection method for specific vital signs.

EDITOR’S PERSPECTIVE

A pragmatic methodology to extract anesthetic and physiological data from the electronic health record

BACKGROUND/AIMS

Traditional manual methods of extracting anesthetic and physiological data from the electronic health record rely upon visual transcription by a human analyst that can be labor-intensive and prone to error. Technical complexity, relative inexperience in computer coding, and decreased access to data warehouses can deter investigators from obtaining valuable electronic health record data for research studies, especially in under-resourced settings. We therefore aimed to develop, pilot, and demonstrate the effectiveness and utility of a pragmatic data extraction methodology.

METHODS

Expired sevoflurane concentration data from the electronic health record transcribed by eye was compared to an intermediate preprocessing method in which the entire anesthetic flowsheet narrative report was selected, copy-pasted, and processed using only Microsoft Word and Excel software to generate a comma-delimited (.csv) file. A step-by-step presentation of this method is presented. Concordance rates, Pearson correlation coefficients, and scatterplots with lines of best fit were used to compare the two methods of data extraction.

RESULTS

A total of 1132 datapoints across eight subjects were analyzed, accounting for 18.9 h of anesthesia time. There was a high concordance rate of data extracted using the two methods (median concordance rate 100% range [96%, 100%]). The median time required to complete manual data extraction was significantly longer compared to the time required using the intermediate method (240 IQR [199, 482.5] seconds vs 92.5 IQR [69, 99] seconds, p = .01) and was linearly associated with the number of datapoints (rmanual  = .97, p < .0001), whereas time required to complete data extraction using the intermediate approach was independent of the number of datapoints (rintermediate  = -.02, p = .99).

CONCLUSIONS

We describe a pragmatic data extraction methodology that does not require additional software or coding skills intended to enhance the ease, speed, and accuracy of data collection that could assist in clinician investigator-initiated research and quality/process improvement projects.

A Spatiotemporal and Multisensory Approach to Designing Wearable Clinical ICU Alarms

In health care, auditory alarms are an important aspect of an informatics system that monitors patients and alerts clinicians attending to multiple concurrent tasks. However, the volume, design, and pervasiveness of existing Intensive Care Unit (ICU) alarms can make it difficult to quickly distinguish their meaning and importance. In this study, we evaluated the effectiveness of two design approaches not yet explored in a smartwatch-based alarm system designed for ICU use: (1) using audiovisual spatial colocalization and (2) adding haptic (i.e., touch) information. We compared the performance of 30 study participants using ICU smartwatch alarms containing auditory icons in two implementations of the audio modality: colocalized with the visual cue on the smartwatch's low-quality speaker versus delivered from a higher quality speaker located two feet away from participants (like a stationary alarm bay situated near patients in the ICU). Additionally, we compared participant performance using alarms with two sensory modalities (visual and audio) against alarms with three sensory modalities (adding haptic cues). Participants were 10.1% (0.24s) faster at responding to alarms when auditory information was delivered from the smartwatch instead of the higher quality external speaker. Meanwhile, adding haptic information to alarms improved response times to alarms by 12.2% (0.23s) and response times on their primary task by 10.3% (0.08s). Participants rated learnability and ease of use higher for alarms with haptic information. These small but statistically significant improvements demonstrate that audiovisual colocalization and multisensory alarm design can improve user response times.