AI is making Healthcare Domain Experts more valuable
Audience: Specifically for Healthcare domain experts – How you are even more relevant in the age of AI.
Let’s dive right in to understand as of now what AI Can Do vs. What Only Healthcare Experts Know!
Relevance of Healthcare Operations/Clinical Experts
AI can manage appointments and triage patients.But only the healthcare operations experts truly understand:
- What symptoms are urgent vs. nonurgent (e.g., chest pain and shortness of breath = immediate ER visit, not appointment in a week)
- The impact of insurance plans on appointment setting (whether or not certain providers are covered, or if authorisations are required within insurance plans)
- Reasons why back-to-back scheduling is ineffective (need buffer for emergencies, late patients, room cleanup, conversations)
- When to override optimal schedule: There is a regular patient with anxiety, and they require the same physician and time slot.
Relevance of Healthcare Compliance/HIPAA experts
AI can point to possible HIPAA violations.But only a healthcare compliance expert appreciates that:
- What data transfer situations require BAA agreements, and what do not
- When de-identified data is still identifiable (small town, rare disease, and age range may lead to patient identification)
- Which state laws on privacy are tougher on HIPAA (California, Texas have follow-along rules)
- When the “minimum necessary” rule is applicable, rather than full access permission
Relevance for Medical Billing/Revenue Cycle experts
AI systems can code procedures and also produce health insurance claims.But only a medical billing specialist knows:
- Which procedure code bundles vs. bills separately, where billing both is a rejection of the claim
- When to apply modifier codes to avoid denied claims (Same Procedure, Different Sites: Require Modifier)
- Which insurance companies require pre-authorisation for which procedures (variable with respect to plans, procedures, and diagnostics)
- Reasons why the claim was denied, how to appeal it successfully (Denial Code says one thing, but the reason is different)
Relevance for Clinical Decision Support experts
AI may use the symptoms to make an assisted diagnosis.However, only a clinician would know:
- Which of these red flag symptoms completely turns your diagnosis around (patient complaint of worst headache ever = possible aneurysm, rather than migraines
- When lab results are normal but really indicate a concern (creatinine level of 1.2 is normal, but doubling from the patient’s baseline indicates kidney issues)
- What interactions the AI system has missed (technically possible, but in elderly patients, this = dangerous)
- When to trust clinical judgment rather than a recommendation from AI (patient presentation doesn’t match the algorithm)
Relevance for Healthcare IT/EHR Implementation experts
AI is capable of transferring patient information from one system to another.But only a healthcare IT expert will be able to answer the following:
- List data fields that are ‘critical’ and ‘nice to have’ (for example: allergy information is critical, while preferred language is nice to have.
- How different EHRs organise the same data in different ways (e.g., medication dosage in one system = three separate data fields in another system)
- When data migration will impact clinical flows (nurses look at the data in a specific manner – new system may cause chaos)
- What integrations are mandated vs. what is optional (ADT feeds into public health = mandated in certain states)
Relevance for Population Health/Care Coordination experts
AI systems have the capacity to detect patients who may require intervention.However, only a population health expert can know:
- What risk factors are adjustable versus non-adjustable (e.g., diabetes – adjustable, genetic disorder – non-adjustable)
- Whether Frequent ER usage points to issues of a non-medical nature, such as patients going to the ER for food, shelter, social interaction, etc.
- What worked in theory vs. actual interventions (looks good on the care plan, but the patient does not have transportation)
- Social Determinants of Health influence medical outcomes (patient non-compliant with meds because of unaffordability, and not because of lack of understanding of what to do
Relevance for Healthcare Quality/Patient Safety experts
AI can monitor quality factors and point out incidents.Only a quality/safety expert can tell:
- What near misses point to larger system problems vs isolated problems? (Same type of problem in multiple units means a system problem)
- Good metrics conceal issues (Low readmission rate because patients are sent to a different hospital)
- Root cause that AI is unable to identify (listed in incident report as medication error, actual cause is understaffing + bad handoff)
- How to make changes that stick and changes that are worked around (people will find a way to work around changes if changes do not meet processes)
These are only a few examples in healthcare where subject matter or on-ground experts are becoming invaluable when the underlying systems use AI.
We are AI advocates ourselves; we work and build AI systems. However, do domain experts need to fear AI, probably not; rather, realise that they are the ones driving AI.
Share this post
About the Author
Shailendra Gupta
(Co-Founder and CEO of Mind IT Systems)
Shailendra is Co-Founder and CEO of Mind IT Systems and is responsible for strategy and business relations.
With around two decades of experience in getting things done in marketing, sales, strategy, delivery, or technology, he has a successful track record of leading startups and mid-size companies and being a prime contributor to stakeholder management, growth, and value creation. A thought leader in the geo-social space, he is highly respected for realizing new paradigms in marketing, solutions, and approaches.