Artificial Intelligence (AI) is defined by Wikipedia as intelligence exhibited by machines or software. In the article 10 Ways Artificial Intelligence Could Make Me a Better Doctor, the author lists 10 ways to take advantage of AI support. Number 6 on the list is: “Help me make hard decisions rational.”
The others are worth reading but are more related to time management aspects of practice. Being guided toward trustworthy, intelligent, rational decisions based on up-to-data, practice guidelines, and cost consciousness would help most any clinician. After all, there is already doing this processing in their gray matter.
Here are some practical areas in which a well-functioning EHR-AI (Electronic Health Record Artificial Intelligence) system could effectively support providers in making timely good clinical decisions
1. Creating an easily understood differential diagnosis hierarchy
a. It would be quite helpful if the AI used incoming clinical data to create a hierarchical differential diagnosis tree (perhaps graphic) based on likelihood and potential risk. The AI could supplement the clinician’s differential diagnosis, and create a high-risk profile, with alerts, individualized for each patient.
2. Fully probe the Differential Diagnosis
a. For example, if a pulmonary embolism is considered, the AI could risk stratify the patient using clinical scoring schemes. It may ask the provider for more data, and could give the provider instant access to relevant clinical articles and best practice guidelines.
b. It could then make recommendations regarding the work-up needed.
c. It could provide real-time cost data for any testing.
3. Developing a treatment plan
a. The AI system would most commonly offer established treatment protocols
b. It would show time frames, when treatment is time critical
c. It might offer alternatives if patient refuses
d. And it might also give guidance on shared-decision making data
4. Selecting appropriate medications
a. AI could use tailored real-time displays to signal:
i. Significant clinical side effects
iii. Drug-drug interactions
iv. Cost analysis
v. Data supporting alternative (non-pharmaceutical) treatment
5. Disposition Analysis:
a. At time of disposition, the high-level AI system could address key areas to remind about needed attention:
b. Scan the final-stage record to be sure no key oversights exist
c. Flag and display any unaddressed high-risk warnings
d. Verify prescribed follow-up timese. Recommend consultants, primary physicians and/or others
Most clinicians already do all these things with each patient. The AI, if functioning expediently, would function as another “eye” on the case. Quality of care should improve, and perhaps come closer to its higher potential.
No, an AI system will never replace flesh-and-bones clinicians, but it has the possibility of greatly augmenting their performance capacity. If these clinical decision support systems appear on the market, and are well build, acquiring one for a busy practice would be a no-brainer.
Xpress Technologies Electronic Health Record is an industry leader in creating an EHR that is “Provider Friendly”.