Digital AI Operations: Could Healthcare Benefit?

 This week I watched this video on WSJ: https://www.wsj.com/video/series/travel-guides/how-ai-is-transforming-air-traffic-control-towers/1E2C984E-F047-47F7-B1AB-1F7D1104F9E1?mod=Searchresults_pos15&page=1

The video shows how digital AI traffic control towers are changing operations at Heathrow Airport in London. The system is powered by algorithms that identifies plans and then overlay satellite information, and can function in clouds or even "see through" buildings, improvements over human air traffic control skills. 


This made me think about what other operations could be improved and made more efficient using AI. I work in healthcare, and one area of inefficiency is the process of consulting specialists for a patient. Currently, it works like this:

  1. The patient's primary team identifies a problem that they don't have the knowledge or skills to address (a mental health need, an infection, etc.) 
  2. The primary team identifies the best specialist service to consult
  3. Someone on the primary team places the consult and sends out a page
  4. The specialist team receives the page and calls back the primary service to discuss the patient's need
  5. The specialist visits the patient
  6. The specialist communicates a plan and recommendations back to the primary team, often by writing a note in the electronic health record and/or by sending a message or page again

The issue with this process is that there are many points of failure. For instance, the primary team may not correctly identify the need for a specialist. This is a problem in mental health, where physicians are notorious for either over-estimating or under-estimating a patient's need for mental health support. The other issue is that the primary team must value the specialist's service and recognize their own limitations, which can be difficult for some physicians. Operationally there are failure points as well. For instance what if the primary team doesn't pick up the phone when the specialist calls back the page? And what if the primary team is so busy they don't read the recommendations in the note carefully?

All this made me wonder whether AI could improve this process operationally. For instance, an algorithm could trawl the medical record, looking for pre-identified triggers to consult a specialist. This could be something like elevated white blood counts, fever, or wounds. For mental health it could look for a history of psychiatric admissions, a current depression score on a depression measure, or look for words in the notes like "tearful" or "crying". Then, the AI could add that patient to the specialty team's patient list to evaluate.

Another benefit is that this would reduce workload for teams. Because they would not have to engage in the administrative process of identifying the problem, deciding to place a consult and physically entering the consult order in the computer, the specialist could get to the patient faster.

My hope is that as hospital operations catch up with the rest of the business operations in the US, healthcare systems will consider ways to reduce burden on physicians by using AI. Not only would this benefit the system, but may even improve staff turnover and satisfaction. 

Comments

Post a Comment

Popular posts from this blog

Why are urban, professional women buying construction worker overalls?

Image
 This week I read this really interesting style piece on WSJ: https://www.wsj.com/style/fashion/high-fashion-or-hunting-gear-why-chic-city-women-are-shopping-at-traditional-sporting-stores-ca4de233?mod=style_feat2_fashion_pos4 The article states that there is a rise in interest in brands that have traditionally been considered entirely functional, and even overtly ugly. Brands like Ariat and Barbour have been taken up by the fashion set for quite some time (I own Ariats and was *this close* to getting my husband to buy himself waxed Barbour jacket). Extending this trend, I've noticed an uptick in my friends showing up to parties in Carhartt jackets. What's going on? From a marketing perspective, it appears that traditionally functional brands are identifying new segments to target. We can see this in a range of products. The ubiquitous Stanley cup once belonged to a brand that made twist-top thermoses that my dad swears by. Today, I can't find the Stanley in the light pink ...
Read more

Why Can't We Fix Mental Health Care?

Image
 This week, I read this WSJ article about suicide rates during and after the pandemic: https://www.wsj.com/health/healthcare/americans-suicide-highest-level-2022-02eb10ea?mod=healthcare_more_article_pos44  This blog post is a long one, so bear with me. At the end of this article, you'll see why mental health care is so hard to access and how medical systems seeking to reduce their costs should counterintuitively hire more psychologists, not less, even though psychologists don't generate enough revenue to cover their  salaries and benefits.  Image from PlanStreet The article discusses ongoing populations at risk of suicide (older, white men), and how suicide rates are finally declining for younger people. Importantly though, the article points out that mental health care - which was hard to access before COVID - is now even harder to find. This begs the question, how can it be with all the increased attention on mental health and increased need for mental health ...
Read more