JMIR Publications

ABSTRACT

Background:

Sickle cell disease (SCD) is the most common inherited blood disorder, affecting millions of people worldwide. Most patients with SCD experience repeated, unpredictable episodes of severe pain. These pain episodes are the leading cause of emergency department visits among SCD patients and may last for as long as several weeks. Arguably, the most challenging aspect of treating pain episodes in SCD is assessing and interpreting a patient's pain intensity level.

Objective:

This study aims to learn deep feature representations of subjective pain trajectories from objective physiological signals collected from electronic health records (EHRs).

Methods:

This study utilized electronic health record (EHR) data collected from 496 Duke University Medical Center participants over five consecutive years. Each record contained measures for six vital signs and the patient's self-reported pain score with an ordinal range from 0 (no pain) to 10 (severe and unbearable pain). We also extracted three features related to medication: Medication Type, Medication Status (Given/Applied or Missed/Removed/Due), and Total Medication Dosage (in mg/ml). We used variational autoencoders for representation learning and designed machine learning classification algorithms to build pain prediction models. We evaluate our results using an accuracy and confusion matrix and visualize the qualitative data representations.

Results:

We designed a classification model on raw data as well as using deep representational learning to predict subjective pain scores with an average accuracy of 82.8%, 70.6%, 49.3%, and 47.4% for 2-point, 4-point, 6-point, and 11-point pain ratings respectively. We observed that Random Forest classification models trained on deep represented features outperformed models trained on unrepresented data for all pain rating scales. We observed that at varying Likert scales, our models performed better when provided with medication data along with vital signs data. We visualized the data representations to understand the underlying latent representations indicating neighboring representations for similar pain scores with higher resolution of pain ratings.

Conclusions:

Our results demonstrate that medication information (which medication, how much total medication dosage, whether it was given or missed) can significantly improve subjective pain prediction modeling compared to modeling with only vital signs. This study shows promise in data-driven estimated pain scores that will help clinicians with additional information about the patient condition other than the patient's self-reported pain scores.