Why Every Healthcare Technology has Trade Offs

Welcome to Episode 7 of the Ardexia Insights Series where we speak to leaders, clinicians and advocates for better healthcare. This week, we talk to Dr. Minal Shah, a hospitalist, physician informaticist, and health technology leader at CommonSpirit Health, one of the largest health systems in the US, where she supports more than 25,000 providers.

Technology doesn't have silver bullets. Every digital health solution involves trade-offs. Organizations that don't understand this create new problems while solving old ones.

Dr. Minal Shah designs and implements clinical workflows at scale across EHRs, virtual care, and applied AI. Her North Star is protecting clinical judgment and the patient-physician relationship as healthcare becomes more digital. She focuses on ensuring technology earns its place by reducing cognitive burden, strengthening trust, and enabling clinicians to practice medicine more effectively.

The conversation revealed what happens when organizations understand trade-offs versus when they chase innovation without understanding ecosystem impact.

Why Clinical Judgment Needs Protection

Over the last decade, physician autonomy has been slowly shrinking. Physicians get squeezed from all angles: insurance companies with prior authorizations, the sheer volume of patients they're expected to see, documentation burden. Technology also plays a role in eroding physician autonomy.

"In clinical informatics, I think of us as stewards of clinical judgment. I think we recognize that technology can not only shape behavior, but can affect burnout. And so for me, protecting clinical judgment, it really starts first with protecting a clinician's focus."

If clinicians can focus on their patients, if they have that attention, they're already enabled to use their best judgment by clearing out distractions.

Healthcare has become more complicated. More things pull at our time, more requirements to fulfill, more systems to navigate. The question isn't whether technology adds value but whether it adds more distraction than it removes.

Designing Workflows That Actually Work

Minal starts with a principle more organizations should follow: dig deep into the problem you're trying to solve.

"There is a tendency for us in the technology realm to get very excited about new things and to try to design solutions before we fully understand a problem. And that is really the first step in designing a workflow that is actually going to be accepted and adopted by your clinicians."

The goal: make clinicians feel supported in making clinical decisions rather than just checking boxes. Create workflows that support execution rather than adding barriers. Make the right thing to do the easy thing to do.

Organizations buy impressive technology, then discover nobody uses it because it solves a problem that doesn't exist or adds more friction than it removes.

Balancing Standardization and Clinical Judgment

Minal faces a familiar challenge: balancing standardization with individualized, judgment-based approaches. There are best practices worth supporting. But patients don't fit neatly into boxes.

"There are many situations where you may have to deviate from what the best evidence might show. You might have to deviate from what a guideline might show because patients are individual and there are many factors at play when making decisions."

Her approach focuses on three things:

Minimizing interruptions when presenting choices. Clinical decision support tools have options beyond interruptive alerts: setting defaults, having tiered choices based on clinical situation, using checklists appropriately, framing alternatives clearly, reserving just-in-time nudges for situations where clinicians genuinely need to stop and think.

Allowing off-ramps for providers who want to choose different paths. "We have to allow for some user-level personalization for them to choose a different path. We have to allow an easy path for them to click no and say, I want to be doing something else instead."

Learning from variability in how people use tools. When you deploy clinical decision support, watch what happens. If most providers aren't acting on an alert, that's a lesson. Either the problem isn't important to clinicians, or the tool is designed poorly.

Clinical Judgment vs. Medical Knowledge

Minal made a distinction that matters for AI discussions: clinical judgment and medical knowledge are not the same thing.

"Clinical judgment is so much broader than just the pure medical knowledge or the data synthesis piece. It is really about understanding the context of the patient. It's about other factors like readiness, values, ethics, judgment, all of those things are playing a role in the treatment plan that we create for our patients. So judgment lives in that gray zone and you can't really standardize every single approach to it."

AI can provide medical knowledge but can't replace clinical judgment. The knowledge component, the data synthesis, that's where AI excels. The contextual understanding, the patient readiness assessment, the values and ethics considerations, that's where human judgment remains essential.

Building Trust for AI Implementation

When implementing AI tools, Minal starts by understanding clinicians' thought processes. When designing for workflows she's not familiar with, she brings in subject matter experts.

"I ask them to explain to me what their thought process is as they are making decisions. So having that additional clinical context is huge. And it goes a long way towards building trust."

The other critical piece: pick problems clinicians actually want to solve. "You'd be surprised how many solutions are built for problems that nobody really cares about."

Speak the value language of clinicians. How is this going to give you time back? Make you more efficient? Help you provide better care? Help you achieve success in quality metrics you care about?

Start with early adopters willing to learn and experiment. Use them as champions. Physicians trust colleagues more than outsiders.

The Trade-Offs Nobody Wants to Discuss

"I can immediately tell when somebody is not very experienced in deploying technology at scale because they'll describe technology as like a silver bullet. Like this is gonna change everything immediately. And those of us that have been in health technology for some time, we know that there is almost no technology that doesn't have some trade-off that you're accepting."

Even ambient scribes, the most widely adopted AI technology in healthcare, have trade-offs. You review your note more and maybe edit things rather than typing it yourself and going through that sense-making process.

"There's really no technology that doesn't have some degree of trade-off. And those trade-offs are not necessarily very obvious at the beginning, but zooming out and thinking about the entire ecosystem of care rather than just the stakeholder that you're designing for can help you identify some of the potential implications of rolling out a technology."

The Inbox Crisis as Case Study

The patient portal inbox illustrates this perfectly. The tool was built for patients to easily connect with providers. Good intentions.

The unintended consequence? It dramatically increased burden on primary care providers in particular and all physicians generally. It's now a huge driver of burnout.

"Does that mean that we should not have rolled out patient portals? No, they are a huge patient empowerment tool. But if you can zoom out and understand the ecosystem, the players that are involved, all the stakeholders that are involved, you can design systems that better support people that may be absorbing a new burden."

When deploying inboxes to clinics without patient portals, the question becomes: how can we organize pools, triage messages, set up operational systems to help clinicians be effective and minimize additional burden?

The Operational Reality Nobody Wants to Hear

"That's not always a popular answer. When you tell someone the thing that's limiting this from scaling is not your technology, it's actually some operational piece, or there needs to be some operational standardization before it can scale, people don't like to hear that because changing the technology is the easy part, but changing the clinical and operational workflows and standardizing, that's the really tough part."

The last mile of implementation depends on structural pieces that have nothing to do with technology.

For Clinicians Who Want to Get Involved

Multiple paths exist: start at your local health system by joining committees or offering to pilot new tools. Pursue formal fellowships in clinical informatics. Take courses to become a more knowledgeable user. Explore vibe coding tools. Go the startup advisory route.

"I strongly encourage people to be exploratory with their career. I think that many of us hit mid-career. We feel very comfortable in our clinical practice and we start recognizing that we have a lot of interest in these other areas in healthcare. And I would say that your career is never going to look the same as it did as you thought it would when you were a first year medical student."

Be willing to learn something from scratch. Be willing to be humble and become a learner again.

The Bottom Line for Digital Health Implementation

Several principles emerge:

Technology without understanding trade-offs creates new problems while solving old ones. Every solution shifts burden somewhere. Account for that and design support systems accordingly.

Clinical judgment and medical knowledge are different things. AI can augment knowledge but can't replace contextual understanding, values assessment, and ethical considerations.

Problems worth solving are problems clinicians actually care about. Speak their value language: time, efficiency, better patient care, meaningful quality metrics.

Trust comes from clinical credibility and champion networks. Physicians trust colleagues more than outsiders. Early adopters create pathways for broader adoption.

The last mile of implementation is operational, not technological. Changing technology is easy. Changing clinical and operational workflows is hard.

Protecting clinical judgment means understanding that technology earns its place by reducing cognitive burden, strengthening trust, and enabling clinicians to practice medicine more effectively.

————————————————————————————————————————————————————————

Learn more about Dr. Minal Shah's work: LinkedIn | CommonSpirit Health

Watch the full conversation on Youtube.

Listen on Spotify or Apple Podcast.

Follow us to hear our next episode, drops every 3 weeks.

Ready to transform your healthcare innovation approach? Contact Ardexia to discuss how we can help you move from pilot to sustainable adoption.

Related Resources

Episode 6: Breaking Healthcare Silos: Lessons from Across Three Continents

Episode 5: Digital Empathy and Why AI Scribes Are the First Technology Doctors Actually Want

Dr. Aditi Joshi is the CEO of Ardexia and host of the Ardexia podcast. She's an emergency physician who has built multiple digital health programs across three continents and specializes in turning failed digital health implementations into measurable clinical and financial success.