How AI Can Address The Crisis In Healthcare Labor Costs

AI creates exciting possibilities in healthcare, and it’s tempting to fixate on the sexiest use cases: diagnosing dreaded diseases, communicating with inquisitive patients, and pinpointing who is at the highest risk of acute events. Unfortunately, in each of these cases the consequences of mistakes can be enormous, the demands of regulators will be high, and long-established workflows are deeply ingrained. These fields will change slowly.

A much nearer-term opportunity lies in addressing a critical pain point for health systems: labor costs. These costs are driving significant erosion in the industry’s finances. Premier, a large purchasing organization serving hospitals, calculates that hospital labor rates increased 6.5% last year, vs. a 2.7% increase received in Medicare reimbursement rates. The gap is even more severe in some fields that have had to turn increasingly to contract labor, where hours are estimated to have increased 91% from 2020 to 2023 even while contract nurses’ median hourly rates have risen from $64 to $132 an hour. As a result, health system spending on recruiting is up 27% year-over-year. With workers leaving healthcare at high rates due to factors such as burnout and aging, it’s imperative that the industry make better use of and retain the labor that remains. If they don’t, the consultancy McKinsey projects that clinical labor costs will continue to rise at more than an 11% compound annual growth rate to 2027. It’s alarming for the industry.

Changing How Nurses Spend Their Time

Given that nurses often comprise over half the headcount of health systems, one key place to look is the way that nurses spend their time. A study in the Permanente Journal showed that medical-surgical nurses spent 35% of their time on documentation and another 21% on care coordination. By comparison, patient care took 19% of time, medication administration 17%, and patient assessment 7%. That’s right – more than half of the nurses’ time was spent on paperwork and communication, not on direct patient care.

AI can tackle these tasks in numerous ways: recording patient encounters, charting patients’ vital signs, and drafting update notes for the health record, for example. Some leading health IT companies are already launching pilots to bring generative AI – similar to that found in ChatGPT – to these specific use cases. Success with these pilots will not only reduce the need for labor, but it will also focus more of nurses’ time away from tedium and more on the types of activities that drew them to the profession in the first place.

Improving Day-to-Day Management

A second key place to look is the way that frontline staff are managed. Laudio is one example of an AI start-up targeting this area. According to CEO and Co-Founder Russ Richmond, more effective front-line management can reduce health system staff turnover by 20-25% a year, while also generating an over 20% gain in employee engagement scores.

How? Richmond says, “There’s a lot of evidence that workers connect more to their primary manager than to the institution overall. It’s that affinity bond which impacts burnout and retention.” By nudging managers toward better behaviors, the company aims to strengthen those bonds.

AI is the mechanism used. Richmond explains, “We integrate with existing HR information systems, time and attendance systems, surveys about workers’ needs and interests, and data about managers’ actions. Then we can recommend what frontline managers can do in a very specific way. For instance, if we see that someone on the team was the only experienced nurse on a shift for the last five of seven shifts, we can suggest that the manager reach out to see how they’re doing and make necessary schedule adjustments.”

Customers First, AI Second

Laudio’s approach echoes that of successful tech start-ups, which focus on their IT only after having nailed the customer problem. Richmond notes, “We interviewed over 400 frontline managers, and they told us the most important features for them. Over time, we’ve built a dataset to see which actions are most effective in which scenarios, and these can vary a lot, for instance, between an ICU and the environmental services team.”

Through taking this type of user-centered approach, healthcare institutions can save on recruiting, contract labor, labor rates, and other labor costs. Healthcare work abounds with loosely managed work and labor that doesn’t touch a patient. AI can address these ills through helping to maximize both staff effectiveness and the time devoted to direct patient care.