Low Back Pain: AI Can Expedite Pain Relief Recommendations Through Electronic Record

PainRelief.com Interview with:

Ismail Nabeel MD, MPH
Associate Professor
Public Health and General Preventive Medicine
Mount Sinai Medical Center

Dr. Nabeel

PainRelief.com:  What is the background for this study?

Response: Acute and chronic low back pain (LBP) are different conditions with different treatments. However, they are coded in electronic health records with the same International Classification of Diseases, 10th revision (ICD-10) code (M54.5) and can be differentiated only by retrospective chart reviews. This prevents an efficient definition of data-driven guidelines for billing and therapy recommendations, such as return-to-work options, etc.

In this feasibility study, we evaluated if Artificial intelligence can automatically distinguish the quality of Low Back Pain (LBP) episodes by analyzing free-text clinical notes from the treating providers. 

These clinical notes were collected during a previous pilot study evaluating an RTW tool based on EHR data that included nearly 40,000 encounters for 15,715 patients spanning from 2016 to 2018 and clinical notes written by 81 different providers. We used a dataset of 17,409 clinical notes from different primary care practices; of these, 891 documents were manually annotated as “acute low back pain” and 2,973 were generally associated with LBP via the recorded ICD-10 code. 

PainRelief.com: What are the main findings?

Response: We compared supervised and unsupervised AI/Machine learning strategies to identify the quality of Low Back Pain in clinical practice. Deep learning Artificial intelligent (AI) model ConvNet (Convolutional Neural Network) trained using manually annotated clinical notes obtained the best results with an AUC-ROC of 0.98 and an F-score of 0.70. In the absence of manual annotations for supervised learning, topic models performed better than other methods trained using ICD-10 codes, which were noted to be inadequate in identifying the quality of the Low Back Pain (LBP).

PainRelief.com: What should readers take away from your report?

Response: Determination of the characteristic quality of the low back pain in practice has a profound implication in finding the right therapy, assisting patients in the return to normal activities and decreasing the economic burdens on clinical facilities to adequately treat and manage this subset of the patient population. 

The acuity of low back pain is usually reported in clinical notes, requiring a retrospective chart review process to characterize LBP events in the practice, which is time-consuming and not scalable. This paper is the first to explore the use of automated approaches based on Artificial intelligence/machine learning to analyze free-text clinical notes from multiple provider practices to identify the quality and characteristics of low back pain episodes. It has the potential to expedite the care of this complex condition at the point of care level in clinical practice. 

PainRelief.com: What recommendations do you have for future research as a result of this work?

Response: With this study, we have demonstrated the ineffectiveness of using ICD-10 codes to determine the acuity of pain in LBP.  The study provides a pathway to look at alternative ways to access low back pain acuity in patients where targeted interventions can substantially improve the care of the patient suffering from LBP.  Evidence-based recommendations can be deployed through Electronic Health Record (EHR) within the clinical domain for the frontline primary care providers. 

PainRelief.com: Is there anything else you would like to add?

Response: AI-based Machine learning methods for EHR data processing are enabling improved understanding of patient clinical trajectories, creating opportunities to derive new clinical insights. In this study, the approaches provide a generalizable framework for learning to differentiate disease acuity for low back pain in primary care practice. It also provides a clear path toward improving the accuracy of coding and billing of clinical encounters for LBP.

Disclosures: None 

Citation:

Miotto R, Percha BL, Glicksberg BS, Lee HC, Cruz L, Dudley JT, Nabeel I
Identifying Acute Low Back Pain Episodes in Primary Care Practice From Clinical Notes: Observational Study
JMIR Med Inform 2020;8(2):e16878

DOI: 10.2196/16878

[wysija_form id=”3″]

[last-modified]

The information on PainRelief.com is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition. Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.