There is A LOT OF ACTION right now as the tech industry, insurance
companies, governments, and major health systems cast about trying to
figure out how to lower costs and/or provide better services to mental
health clients.
*We REALLY need our professional associations watching all of this.
*
As I read these articles I see benefits, yet I see opportunities for:
Misunderstandings of how mental health actually works,
Turning therapists into plug and play component cogs in large machines.
With regards to this particular article, I have questions/issues:
I see the value of using AI for triage to determine who needs to be seen
more quickly. What I don't understand is how this works after the
initial intake. It's almost like the CEO being interviewed sees the
entire therapy process as akin to an Emergency Room intake process.
Does he think therapists are plug and play with a different one assigned
each session? Or is he only discussing psychiatrists (and that is bad
too)? Does he think there is no such thing as regular weekly sessions
and you can just switch times and days around constantly by AI
needs-assessment algorithms?
I may be misunderstanding his approach given its a short, edited interview.
Where is the value of RELATIONSHIP and human connection in this?
It reads like his vision of the future is a workforce of therapists on
shifts, and their work day varies in real time with different clients
being plugged constantly in and out of their schedules in real time.
(Again, my nightmare of therapy being treated like a call center.)
/"AI can help predict gaps in outpatient access and the
supply-and-demand imbalance within a health system or clinic population
by provider type, time of day and acuity level. This predictive ability
can help health systems optimize staffing and scheduling to increase
productivity and patient satisfaction."
/
So... on Mondays I work 9am-5pm, but on Wednesdays I work the 12pm-8pm
shift. Every other weekend I'm needed Saturday nights 8pm-2am because
of a client demand surge? Perhaps I get called-in when demand surges
unexpectedly?
Yes -- AI could (and should?) be used to automate routine tasks and
assist in diagnostic assessments and treatment recommendations. Can you
imagine all the chart data required to be in the database on each
client? Can you imagine what goes wrong if the algorithms are tuned wrong?
/"We should focus on machine learning applications that use discrete,
anonymized data to improve care delivery without putting patient
information at risk."
/
Promising...
/
"AI-powered tools can streamline these processes, potentially using
natural language processing to generate clinical notes from recorded
sessions or automating insurance coding. This allows clinicians to focus
more of their energy on direct patient care, potentially increasing the
number of patients they can see without compromising quality."
/
I'm all for it. In other short essays I have already stated my opinions
that these AIs should be local, open source, and NOT connected to the
Internet or Cloud. It's becoming clear that I will lose this battle as
cloud services build market share and large hospitals incorporate AI
notes tools into EPIC and the like.
/"AI also can serve as a powerful decision support tool for
clinicians... But AI systems shouldn't replace clinical judgment...
For example, an AI system might flag potential drug interactions or
suggest alternative treatment approaches... However, it's always up to
the clinician to determine the appropriate level of care."
/
Yeah... and what happens to the employee therapist in a large health
system that goes against these recommendations? What happens in the
lawsuit in which the therapist did not follow the canned formulas
suggested by the AI? What diversity of approaches is lost in
standardization?
/"AI improves operational efficiency, optimizes resource allocation and
expands access to care – all of which affect a health system's bottom
line. AI algorithms can analyze patient data, historical patterns and
real-time factors to optimize appointment scheduling and clinician
workloads. This optimization can reduce no-show rates and improve
clinician efficiency."
/
I would love to be more efficient. I think this comes down to trust. I
do not trust that organizations with the money to invest in these tools
will make my work life better. Efficiency in what sense?
