FROM SILOS TO SYNERGY
( continued from page 7)
even more challenging to be a good collaborator. Just as conducting is essential to performing a good half-time show, working together is essential for coordinating and delivering high quality medical care and developing a meaningful relationship between patient and provider. 1
Augmented Intelligence( especially medical AI programs) proport to connect fragmented medical systems into a unified network. AI has found a place in speeding up care, completing medical records and automating daily administrative tasks. 2 American medicine has become comfortable with ambient AI scribes and letter generators. However, the most advanced AI programs make mistakes, so we are cautioned to review all AI generated summaries and never send AI communications directly. 3 Patients love their doctors and want their human“ drum major” to coordinate their care. AI can do many tasks in a synthetic, mechanical manner, but physicians provide care in a holistic and personal way.
Has collaboration become easier with the integration of AI? I would submit that the music is more complicated and the skill level of many mid-level performers has not kept pace. Solo performances are brief and only a small piece of the show. Those providers with the least skills are the most dependent on artificial systems to fill in their knowledge gaps – much like having a synthesizer play the music for you. System failures can be reduced in mechanical systems by writing more rules for system management; not so in adaptive systems like healthcare. 4 All the various components respond to different stimuli and respond in unpredictable ways. Providers are free to make choices and advise based on their knowledge and experience. This is why it is frustrating when patients bring in their printouts from an influencer podcast or YouTube, insisting upon the care they should receive. Like a drummer playing the wrong beat, the team is out of step, and the care feels fragmented and disjointed.
Major hurdles have always existed in our medical hierarchies that have complicated care coordination:
1) Communication gaps were more common when private practice was the norm and different providers used disconnected electronic health records. PCPs and patients were dissatisfied with care delivery, especially mental healthcare. 5 Now in the day of EPIC, those barriers are less, but we now must try to avoid“ note bloat” and creating documentation that no one appreciates.
2) Financial disincentives rule in the“ fee for service” world as surgeries are historically paid at a much higher rate than cognitive skills. Many attempts have been piloted to reward cognitive work and not just procedures. 6
3) Time limitations have only worsened as many providers are paid based on productivity and there is no financial incentive to attend team meetings and interact with other professionals. 7
4) Adjusting care plans, especially as new data presents, remains an issue. Very few patients can reliably keep track of the complicated medical direction they receive and most assume their whole team is practicing together with the same goals in mind. Such is rarely the case and as a results mistakes are made, fixable problems are missed and polypharmacy abounds. The patient can suffer from this lack of coordination.
5) Mitigating legal and systemic risks: physicians hold the license and can be sued for wrong decisions or bad outcomes. Computer systems are agnostic and can’ t be sued( yet).
The American Medical Association has adopted guidelines for managing interprofessional relationships in their code of ethics. 8 Being an advocate, promoting values and fostering a team culture make sense on paper, but in practice, our time is constrained and true authority limited. We must continue to fight to maintain our leadership role in clinical care and not be overtaken by a synthetic intellect.
So much of what we do in medicine is subjective – blending objective data with a patient’ s personal experience. 9 We attempt to coordinate patient preference in decision making which ultimately impacts the outcomes achieved. AI attempts to eliminate such care inconsistencies but lacks the personal touch that most patients demand. Just like watching a CGI / AI generated marching band show or musical performance, the AI presentations seem too perfect and“ less than real” to an actual person anxious about their health and well-being. The goal is for machines and humans to work together. AI can do the heavy lifting with paperwork and data collection, but the human doctor brings empathy and compassion. For the music and medical performance to work, you must hang on as the leader of the band!
References
1
Int J Integr Care. 2016 Jul 18: 16( 3): doi
2
Artificial intelligence in healthcare: transforming the practice of medicine, Future Healthcr J. 2021 Jul: 8( 2)
3
AMA Code of Medical Ethics
4
Perm J, 2013; 17( 3): 91-93
5
Americal Journal of Primary Care, Aug 2022, Vol 28, Issue 8
6 https:// time. com
7
Challenges in Achieving Collaboration in Clinical Practice, Int J Integr Care. 2016 Jul 18; 16( 3)
8 https:// Code-medical-ethics. ama-assn. org / ethics-opinions / collaborative-care.
9
Subjectivity in decision making: common problems and limitations, NCBI. nlm. nih. gov
Dr. Wernert is an Executive Medical Director and practices with Norton Behavioral Medicine.
8 LOUISVILLE MEDICINE