A Comparison Of Text Mining Versus Diagnostic Codes To Identify Opioid Use Problem: A Retrospective Study​

Abstract

Background As opioid prescriptions have risen, there has also been a rise in opioid overdose deaths and substance use disorders. Public health systems have tried to improve their ability to detect and intervene in opioid use disorders to prevent death due to overdose. The objective of this study is to compare two approaches to identify opioid use problems (OUP) using electronic health record data- text mining versus diagnostic codes.Methods Our sample consisted of adults on long-term opioid therapy (LTOT), defined as at least ≥ 70 days of supply within 90 days, and who visited a large multi-hospital network within a two-year period, between 1 January 2013 and 31 December 2014. We excluded patients with active cancer or schizophrenia. Text mining results were validated by a semi-assisted human review process and positive predictive value and level of agreement was reported. Each algorithm sought to identify patients who visited a health care facility due to an opioid poisoning event, opioid abuse, or opioid dependence. Population characteristics for positive OUP identified by text mining and ICD cohorts were compared. Chi-square and Fishers exact test were used for categorical data analysis and independent t-test was used to compare means for continuous variables. We further compared the demographics of the cohorts identified by the two methods.Results We identified 14,298 eligible LTOT patients. Text mining of relevant electronic clinical notes yielded 127 positive OUP cases compared to 45 cases using International Classification of Disease (ICD)-9 codes for the same population. Just eight OUP patients were identified using both methods. The two cohorts differed significantly with respect to age, gender, and other characteristics.Conclusions Compared to diagnostic codes, text mining identified more OUP cases with distinct characteristics. Incorporating text-mining techniques into OUP surveillance methods may support better detection of OUP and more accurate estimates of prevalence.

Classifying Characteristics of Opioid Use Disorder From Hospital Discharge Summaries Using Natural Language Processing

“…Next, we selected three open-source NLP systems/knowledge bases to encode semantic features from the annotated sentences: Empath, Unified Medical Language System (UMLS), and ConText ( Table 1 ). We chose these approaches and systems to encode features based on their demonstrated informativeness in prior studies from the OUD literature [e.g., ( 7 , 11 , 14 , 15 )]. We applied Empath ( 21 ), a tool that draws connotations between words and phrases based on neural word embeddings from over 1.8 billion words of modern fiction, to generate semantic categories based on lay terms, including categories describing clinical, behavioral, and environmental factors such as pain, alcohol, crime, and family.…”

“…Prior studies have utilized NLP to identify problematic opioid use ( 7 , 11 – 15 ) and opioid overdose ( 16 ) from EHR and paramedic response documentation ( 17 ). However, several gaps remain in the development of NLP systems to identify problematic opioid use and OUD.…”

“…However, several gaps remain in the development of NLP systems to identify problematic opioid use and OUD. Symbolic rule-based systems that rely on keyword lists, regular expressions, and term co-occurrence have been most commonly developed, such as nDepth TM ( 11 ) and MediClass ( 16 ), among other tools ( 7 , 12 , 13 ). More contemporary NLP approaches remain limited, with only three previous studies having applied machine learning methods to identify opioid misuse ( 14 , 15 , 17 ).…”

Developing a Framework to Infer Opioid Use Disorder Severity From Clinical Notes to Inform Natural Language Processing Methods: Characterization Study

“…This step can be critical to informing intelligent search of EHRs and limiting the note types necessary for operationalizing the algorithm, thereby reducing computational effort and potentially improving accuracy. Although prior studies have used NLP to identify problematic opioid use from EHRs [21][22][23][24][25][26], few have described an annotation process and none have reported documentation patterns for OUD-relevant information within clinical notes.…”