Using NLP workflows to identify key social determinants of health
A year has gone by since I last blogged about the importance of social determinants of health (SDoH). Since then so much has happened in our lives. The COVID-19 pandemic has affected us all, in ways we could never imagine and it will continue to do so into the future. When you review the SDoH risk factors, you realize that, due to the pandemic, responsible citizens are now enforcing one of these risk factors for safety reasons: social isolation. It wasn’t too long ago that social isolation was branded a warning sign for mental health issues - not a recommendation for maintaining physical health.
Combating the enemy
Many of us have unhealthy habits we turn to during times of crisis (or boredom) - it may be drinking alcohol, smoking or eating unhealthily - as a coping mechanism. So it’s easy to see why alcohol and tobacco sales are spiking. Unfortunately, such comforts can put you at higher risk of developing a more complicated case of the COVID-19 virus, if exposed, and possibly on a slippery slope to increased mental/physical health issues, even if you aren’t exposed to the virus. Individuals such as those struggling with addiction, mental health issues, the elderly and those under heightened stress (especially essential workers), are most susceptible during these times - especially if you add a “vice” like smoking tobacco. Keep in mind we are all struggling during these times of social isolation. So it may not be as obvious who really needs help. Parents now have to work and be full time teachers, unemployment rates are skyrocketing, and healthcare workers are already burdened with immediate dangers. Healthcare workers simply don’t have time to be proactive and look for warning signs. So what can we do to help?
The Best Defense is a Good Offense - Utilize NLP to “Augment Intelligence”
One way would be to identify key linguistic patterns within clinical notes - which is easier said than done unless you have the right technology in place to help. Early signs of detection, which are often documented as text-based information by dutiful clinicians during patient encounters, are key to intervening before the situation gets dire. This is where automated informatics workflow solutions can help. One example would be to look for patterns from multiple data sources (like patient portal messages, nurse notes, telehealth transcripts and even social media) and analyze the data. Data could be extracted using Natural Language Processing (NLP) via an Augmented Intelligence (AI) process which allows you to access and combine both structured and unstructured text content to get a clear depiction of the potential issues arising.
Analytical reporting supported by NLP can help by doing the heavy lifting for you, closely monitoring all sources of unstructured patient data, and sending providers healthcare alerts. Examples of reporting workflows could include:
- Flagging patients with known mental health diagnoses of high risk that could signal danger signs i.e. if a patient has major depression look for signs of hopelessness, substance abuse, high stress and social isolation
- Identification of elderly patients with chronic disease, who are also experiencing feelings of loneliness, inability to access proper medications, etc
- Identifying people with complicated comorbidities such as diabetes, hypertension and obesity, along with issues such as limited access to proper medications and healthy foods, and barriers to physical activity.
This Too Shall Pass
Until we develop a successful vaccine or develop herd immunity life is going to be different for a while. We all should be considerate of others by practicing social distancing and wearing masks in common areas. We will all be happy when this is over - but by utilizing technology AND embracing humanity (and utilizing good sense) we will move past this to a new normal where we are ready and able to be proactive, the next time something like this happens...so that together, we may all see the world in a better place.
