Federal Reviewers Measured Patient Harm in LTACHs. The Findings Belong on the Leadership Agenda.

The HHS Office of Inspector General set out to measure how often Medicare patients in long-term acute care hospitals are harmed by their care. Its reviewers drew a national random sample of 587 beneficiaries admitted in March 2014. Nurses screened every record with a trigger tool adapted for the LTACH setting, and a physician panel that included an LTACH physician reviewed every record the nurses flagged. The report was published in November 2018. OIG uses the federal term long-term-care hospital, or LTCH. This article uses LTACH throughout.

Forty-six percent of patients experienced harm within the first 25 days of their stay. Twenty-one percent experienced an adverse event, which is OIG's term for the most serious harm: an event that prolonged the stay or forced a transfer, required life-sustaining intervention, caused permanent harm, or contributed to death. Another 25 percent experienced temporary harm that required intervention but left no lasting injury. Five percent of patients experienced an adverse event that contributed to or resulted in their death. That projects to an estimated 573 deaths among patients admitted in a single month.

Healthcare-acquired infections accounted for 45 percent of adverse events, and respiratory infections were the most common. Medication accounted for 31 percent and general patient care for 23 percent. Among temporary harm events, medication led at 49 percent, and hypoglycemia was the most frequent single event.

The LTACH harm rate is higher than the rates OIG found in acute-care hospitals (27 percent), skilled nursing facilities (33 percent), and rehabilitation hospitals (29 percent). The report is careful about why. LTACH stays are longer, averaging 26.3 days in the sample against 5.2 days in acute-care hospitals, and every added day brings more interventions and more opportunities for harm. Measured per 1,000 patient days, the LTACH rate of 38 events was well below the acute-care hospital rate of 69, though above the rates for rehabilitation hospitals (29) and skilled nursing facilities (24).

The patients matter as well. Reviewers judged 45 percent of harm events clearly or likely not preventable, and in 81 percent of those cases the patient was highly susceptible because of underlying health status. LTACH patients are among the sickest in American healthcare, and any honest reading of this report starts there.

Complexity accounts for the harm reviewers judged unavoidable. The other 54 percent still has to be explained.

For each preventable event, reviewers recorded why it was preventable, and they could cite more than one reason. Fifty-eight percent of preventable events involved appropriate treatment or preventive care delivered in a substandard way. Thirty-four percent involved errors in medical judgment, skill, or management. In 25 percent, the patient's progress was not adequately monitored. In 23 percent, the care plan was inadequate. In 18 percent, necessary treatment was never provided.

Read one at a time, these are clinical failures. Read together, they point to the systems around clinicians: the protocols a hospital adopts, the monitoring it requires, the way it builds and reviews care plans, and the staffing and supervision that determine whether the right thing happens reliably.

Hypoglycemia shows the pattern clearly. It was the most frequent type of preventable event in the sample, and reviewers judged 71 percent of hypoglycemic events preventable. Forty-two percent of patients had diabetes, and 18 percent of those patients experienced a hypoglycemic event. Reviewers traced several cases to reliance on sliding-scale insulin. That regimen corrects abnormal blood glucose after it occurs instead of preventing it, and current clinical guidelines discourage its exclusive use.

A nurse administers the dose. The hospital decides which insulin protocol is standard, whether pharmacy reviews it, how glucose trends are monitored, and who notices when the same event recurs on the same unit. Those decisions run through committees, policies, and accountability structures that hospital leadership builds and oversees.

Federal law treats every LTACH as a hospital. Each one must meet the same Conditions of Participation as a short-term acute-care hospital. One of those conditions requires a Quality Assessment and Performance Improvement program that tracks medical errors and adverse patient events, analyzes their causes, and puts preventive actions in place, with feedback and learning throughout the hospital. The obligation to find, study, and prevent this harm sits with the hospital, and therefore with the people who lead it.

OIG made three recommendations to CMS and AHRQ after the 2018 report, including a list of potential LTACH adverse events and outreach to LTACHs about patient harm. OIG records all three as implemented.

Federal attention has since moved to what hospitals do with harm once it happens. In July 2025, OIG traced the harm events from its national review of acute-care hospitals and found that hospitals had not captured half of them in their incident reporting or other surveillance systems, often because they applied narrow definitions of harm. Of the events hospitals did capture, few were investigated, and fewer still led to improvements. That review covered acute-care hospitals. The QAPI requirement it examined is the same one every LTACH operates under.

OIG recommended that CMS make sure surveyors prioritize the QAPI requirement. CMS neither concurred nor disagreed, and OIG closed that recommendation as unimplemented in November 2025.

Taken together, the two reports draw a clear line. The external standard exists, the federal list of LTACH harm events exists, and the regulatory requirement to track and prevent harm exists. What decides whether harm is defined broadly, captured, investigated, and acted on sits inside the hospital.

Every LTACH CEO posting in the country screens for years and setting. None screens for knowledge.

Postings screen for what can be audited after the fact. Budget variance, survey results, census. They cannot screen for the judgments that are invisible until they go wrong, and those are the expensive ones.

This report catalogs those judgments. Whether a hospital's definition of harm is broad enough to count a hypoglycemic event as harm. Whether a care plan is reviewed before the patient deteriorates or after. Whether respiratory infection prevention is built for patients who spend weeks on a ventilator. Whether the QAPI program finds the pattern before a surveyor does.