Skip to main content

Substance Abuse: Rising from the Fall - Can Natural Language Processing Help?

helping hand

"The greatest glory in living lies not in never falling, but in rising every time we fall." - Nelson Mandela

Difficult observations

It is never easy to observe someone hit “rock bottom”, especially when it is more of a routine than a one-time fall. Yet - so many of us have experienced those we love repeatedly falling into the familiar pattern of addiction. There are many forms of substance abuse. Some are more socially accepted, like alcohol consumption and tobacco use. Others like opioids, and other narcotic substances, are more taboo.

Whatever the flavor, substance abuse rarely appears in one day. There are patterns when abuse is forming and common clues once the addiction is established. The current pandemic has pushed many into a state of ill-health and substance abuse. Many nations are experiencing booming alcohol sales. And according to a recently published article in the Journal of the American Medical Association (JAMA),since March 2020 US hospitals are reporting an increase in substances found in urine samples nationwide: 67% fentanyl, 23 % methamphetamines, and 19% in cocaine.

Early identification is key. More so than we ever realized. For example, as a physician, I was taught that opioids are safe for short-term use. But the meaning of the term ‘short-term’ is shrinking drastically – some studies show that dependency starts within just a handful of days. As a former research scientist, I have reviewed thousands of patient charts - and the majority of opioids I have seen prescribed are for a minimum of five to seven days.

How do we stop these descents into the abyss?

Identify patterns - easier said than done, right? However, with vigilance and informatics solutions, patterns will form to provide a clearer picture. For example, analytical reporting techniques supported with robust data extracted by using Natural Language Processing (NLP) over a multitude of data sources allow you to understand the whole picture. With up to 80% of data hidden in the healthcare narrative, both structured and unstructured free text content are necessary to identify patterns.

Five examples of NLP-supported analytical reporting

  1. Closely monitor, clinician notes, telehealth transcripts of patients with chronic pain for symptoms that may identify addiction or withdrawal from substances such as opioids.
  2. Create pain registries within hospital systems that identify timelines for prescribed medications used to combat pain. Include metrics such as dosage, reason prescribed and factors that put them at risk for abuse, such as feelings of hopelessness.
  3. Identify people that seem to be ‘accident’ prone that have a history of injuries requiring pain medications. For example: Does the patient have medical issues that are causing so many accidents? Should the patient have a referral to an otolaryngologist for balance? Or are there signs of addiction present in the clinician's notes that might be clues of self-harm in order to receive medications? 
  4. Identify individuals that often visit urgent care type settings for various reasons asking for pain relief medications - check clinician observations within notes that may point to drug-seeking behaviors.  
  5. Look for patterns of highs and lows that may be a danger sign of substance abuse; such as high risk behavior, agitation, erratic behaviors, and so many more.  

Adding known risk factors to reporting is helpful to assist in clinical decision making, especially if someone has multiple risk-factors.  Like in opioid abuse, identify issues, such as a personal or family history of alcohol abuse, illicit drug use, sexual abuse, and common psychiatric disorders (bipolar, mood, anxiety, personality and stress disorders). Most importantly, once set up, institutions need to utilize the reports. I can’t emphasize how important it is to have a workflow in place for proper follow-up.

Addicts are creative, so we must be too. Our efforts would be optimized if we could go beyond single institutions, by building registries to combine information from different data pools (pharmacies, social media, health records) - but given the challenging legalities required, this is unfortunately less realistic. Regardless, efforts under the ‘roof’ of single entities such as hospital systems, national pharmacies, etc. are a critical component.  Community efforts such as this exemplar initiative in North Carolina, where police help instead of prosecute addicts, are also vital in order to combat substance abuse. 

The bottom line - we must all work together to do what we can and should do. While we cannot keep everyone from falling - we can put our best hand forward to lift others back up when we can.

Contact us to find out more

Ready to get started?

Request a Demo

Questions? Ask our experts