Bots are breaking into the most human industry on earth: health care. Just look at Microsoft — the company recently launched a new division to address the intersection between health care and AI.
That’s not to say AI will be replacing doctors anytime soon. But the integration of AI into regular health care tasks is upon us. From surgical assistance to patient education to imaging analysis, robots are already part of the process — and their roles will only increase over time.
Breaking barriers to achieve smarter health care
Of course, not everything is smooth sailing — AI faces challenges in health care just as it does in other fields.
Medical providers often lack the sheer volume of information necessary to train an AI tool, and when the data does exist, it’s usually behind silos. For AI to make greater strides in health care, the industry needs to create common data exchange standards. Tangentially related is privacy: When large batches of data move between these silos, security becomes a challenge.
June 5th: The AI Audit in NYC
Join us next week in NYC to engage with top executive leaders, delving into strategies for auditing AI models to ensure fairness, optimal performance, and ethical compliance across diverse organizations. Secure your attendance for this exclusive invite-only event.
Health care providers take their responsibility for patients’ safety seriously, which means new technologies take time to gain acceptance. Even when the data supports AI, bots still face an uphill battle to win over an industry where the stakes are literally life and death.
But innovation in health care cannot be stopped; it’s only being kept for observation until it’s stable.
If you’re a developer eager to make your mark on the medical world, heed the following best practices:
Cut back on the grunt work
If you want to gain approval from the health care industry quickly, create technology to automate redundant, low-cognition medical tasks. Those small tasks add up — many doctors spend a whopping 49.2 percent of their time on menial tasks like paperwork instead of focusing on patients. Any tool that reduces that grunt work will be valuable.
For instance, create tools that let AI check for interactions between medications or recommend specific tests based on doctor-discovered symptoms, thereby reducing human error. Consider customized patient follow-ups — a task absolutely within AI’s power — to reduce the administrative load on doctors, freeing them to pursue tasks that require their specialized skills.
This concept is already popular in farming. Blue River Technology uses cameras on crop sprayers to identify plants, killing weeds and fertilizing crops. Farmers can customize their sprayers to save on chemicals and time. Medical technology should aim to do the same for doctors — eliminate simple tasks to let them focus on more pressing matters.
Don’t play workflow whack-a-mole
It’s easy to get so wrapped up in creating a new technology that you gloss over the extra work it will require. Remember, health care providers don’t rush to change, so you’ve got to make your tool even more appealing and user-friendly than you would for any other audience.
For example, getting patients to adhere to their discharge plans can be difficult. A chatbot can help them stay on schedule, but if that chatbot begs for daily updates, it’s just creating more work for doctors. A better system would only flag the provider when the patient failed to refill a medication or missed a critical appointment. It’s these kinds of situations you need to consider as you analyze every user touchpoint and determine whether your solution is a relief or just a new chore.
As a parallel example, look to the legal industry: Kira Systems is a lawyer’s AI assistant specializing in contract review. The beauty of Kira is its simplicity — it doesn’t require lawyers to waste hours feeding it information. Kira simply checks databases faster than any human could and presents its findings for review, saving time without adding new steps. Similarly, to satisfy medical providers, reduce their workload instead of replacing it with a higher-tech version.
Bottom line: When workflows go awry in any other industry, things get messy. But when that happens in health care, individuals’ lives are at risk. User-friendliness matters more here than in any other sector.
Put security — not blind innovation — in the driver’s seat
Technical challenges in health care are second to privacy. The most effective tool will be tossed aside if it can’t secure patient information at every turn.
To prepare an AI for the medical field, you need a staff with the right expertise: a high-level security officer who understands the risks, an attorney who is familiar with the laws in patient privacy and security, and a team of developers and engineers committed to keeping technology updated and hack-proof. That means patching software regularly, establishing clear lines of ownership, encrypting everything, and educating staff on security.
It sounds like a lot of overhead work, but again, when it comes to health care, you’re dealing with individuals’ most personal information. Satisfy those concerns, and your new tool will be one step closer to adoption.
The AI revolution is coming to health care, and if you want to get involved, now is the time. Just consider these tips before you dive in — you’ll be much more equipped to confidently develop something useful for medical providers.
Kevin Yamazaki is the founder and chief executive officer of Sidebench, a leading digital product and venture studio.
