How Inspections Shape Life and Death in U.S. Nursing Homes

A new paper by Assistant Professor Maggie Shi, with Ashvin Gandhi of the University of California, Los Angeles and Andrew Olenski of Lehigh University, offers new evidence on a low-cost policy solution that could save lives: making the timing of nursing home inspections less predictable.

Inspections are a cornerstone of regulation across sectors, from food safety and environmental protection to health care quality. While inspections are typically unannounced, they often follow regular cycles, making their timing somewhat predictable. In “Predictably Unpredictable Inspections,” the authors show that this predictability in the nursing home industry has real consequences, not only for how those firms behave, but with life-and-death outcomes for one of the most vulnerable populations.

Nursing homes provide both short-term rehabilitative care and long-term custodial care to some of the most medically vulnerable Americans. Persistent understaffing has long been identified as a key driver of poor quality, including preventable falls, bed sores, and premature death. Using detailed administrative data from U.S. nursing homes, the authors document a striking pattern throughout the nursing home inspection cycle. After an inspection occurs, facilities face a very low risk of being inspected again for many months. During this “low-risk” period, nursing homes reduce staffing and other care inputs. As time passes and the likelihood of another inspection rises, facilities ramp effort back up. These cycles are not merely administrative: patient outcomes move in lockstep with facility behavior. Mortality rates are highest when inspection risk is lowest and staffing effort is weakest and fall as effort increases later in the inspection cycle.

“Our results show that simply changing the timing of inspections—making them harder to anticipate—can save lives at a scale comparable to increasing inspection frequency, while remaining essentially budget-neutral,” Shi said. “Predictability turns inspections into a countdown. Unpredictability turns them into a constant incentive, and that difference has real consequences for patient health.”

Beyond documenting these empirical patterns, the paper develops a dynamic model of how nursing homes choose effort over time. Facilities weigh the costs of higher staffing today against the expected benefits if an inspection occurs in the near future.

Crucially, inspections play two roles. They incentivize effort, but they also generate information. Regulators such as the federal Centers for Medicare & Medicaid Services (CMS) rely on inspections to update their assessments of facility quality and to inform the public. Making inspections less predictable could weaken this informational function by increasing uncertainty about true quality.

The authors’ model allows them to quantify this tradeoff. Using the model, the authors simulate alternative inspection regimes. Compared with a world without inspections, the current system saves an estimated 850.6 lives per year by inducing higher effort. Increasing inspection frequency by 25 percent would save even more lives, but at significant administrative cost, especially given documented inspection backlogs. By contrast, changing when inspections occur costs virtually nothing.

The paper finds that making inspection timing unpredictable would increase lives saved by about 12 percent, an effect comparable to increasing inspection frequency by the same amount. While unpredictability slightly reduces how informative inspections are about underlying quality, this informational cost is modest. Moreover, unpredictability and frequency reinforce one another: more frequent inspections are substantially more effective when their timing is unpredictable.

The findings suggest that regulators may be able to improve compliance and save lives—not by doing more inspections, but by redesigning inspection schedules to be less predictable. In high-stakes environments like health care, small changes in regulatory design can have outsized effects.

For families with loved ones in nursing homes, these findings may be unsettling, but not entirely surprising. Many families already worry about whether care quality fluctuates when oversight is less immediate, and some report trying when possible to vary the timing of their own visits for similar reasons. The results underscore a broader challenge families face: much of what determines care quality happens out of view, shaped by institutional incentives rather than constant monitoring. While inspections remain a critical safeguard, the study suggests that families’ concerns about consistency of high-quality care are rooted in structural features of the system, and are not just isolated failures at individual facilities.

More broadly, the paper points to a lesson that extends well beyond nursing homes. A large empirical literature shows that inspections work better when they are more frequent, better targeted, or more salient. This paper adds a new and policy-relevant dimension to that conversation: predictability. Wherever inspections are used to enforce standards—from health care and environmental regulation to labor and safety oversight—how predictable those inspections are may matter just as much as how often they occur.

Full paper can be accessed here.