Adverse Events in Hospitals: A Quarter of Medicare Patients Experienced Harm in October 2018

WHY WE DID THIS STUDY

In 2010, OIG reported the first national incidence rate of patient harm events in hospitals—27 percent of hospitalized Medicare patients experienced harm in October 2008. During that month, hospital care associated with these events cost Medicare and patients an estimated $324 million in reimbursement, coinsurance, and deductible payments. Nearly half of these events were preventable.

OIG conducted a new study to update the national incidence rate of patient harm events among hospitalized Medicare patients in October 2018. This work included calculating a new rate of preventable events and updating the cost of patient harm to the Medicare program.

The Department of Health and Human Services (HHS) leads national efforts to promote quality health care and prevent patient harm. Several agencies share this responsibility, including AHRQ, which leads HHS's efforts to improve health care quality, and CMS, which is the Nation's largest health care payer and oversight entity.

Although HHS agencies have reported progress during the past decade toward improving patient safety, protecting the health and safety of HHS beneficiaries remains one of HHS's top management and performance challenges. An increased understanding of the prevalence and nature of patient harm will further assist efforts to reduce patient harm events and the factors contributing to these events.

HOW WE DID THIS STUDY

We reviewed medical records for a random sample of 770 Medicare patients who were discharged from acute-care hospitals during October 2018. We conducted a two-stage medical record review to estimate a national incidence rate of adverse events and temporary harm events. Our review included all causes of patient harm regardless of whether the harm was preventable.

Stage 1: Nurses screened the records for possible patient harm events using a "trigger tool" method. A "trigger" is a clinical clue (e.g., documentation of a fall) that may indicate harm. From the Medicare claims data, nurses also reviewed present-on-admission indicators to identify harm that developed after the patient was admitted. We automatically referred records to Stage 2 when patients were readmitted within 30 days of discharge, regardless of whether the nurse identified harm (these include readmissions in October and November).

Stage 2: Physicians reviewed the records flagged during Stage 1 as containing possible harm events. Physician-reviewers identified harm events and assessed the severity of events, whether events were preventable, and factors that contributed to events.

We calculated the potential cost incurred by Medicare and patients as a result of these events. We also determined whether events were on CMS's lists of hospital-acquired conditions. Finally, we compared the results of this report to our 2010 report and explained the limitations of this comparison.

WHAT WE FOUND

Twenty-five percent of Medicare patients experienced patient harm during their hospital stays in October 2018. Patient harm includes adverse events and temporary harm events.

Twelve percent of patients experienced adverse events, which are events that led to longer hospital stays, permanent harm, life-saving intervention, or death. In addition to the patients who experienced adverse events, 13 percent of patients experienced temporary harm events, which required intervention but did not cause lasting harm, prolong hospital stays, or require life-sustaining measures. Temporary harm events were sometimes serious and could have caused further harm if providers had not promptly treated patients.

• Categories of Harm Events. The most common type of harm event was related to medication (43 percent), such as patients experiencing delirium or other changes in mental status. The remaining events related to patient care (23 percent), such as pressure injuries; to procedures and surgeries (22 percent), such as intraoperative hypotension; and to infections (11 percent), such as hospital-acquired respiratory infections.
• Preventability of Harm Events. Physician-reviewers determined that 43 percent of harm events were preventable, with preventable events commonly linked to substandard or inadequate care provided to the patient. (The overall harm rate would be 13 percent if we were to include only events that our physician-reviewers determined were preventable.) Reviewers determined that 56 percent of harm events were not preventable and occurred even though providers followed proper procedures. Events were determined not preventable for several reasons, including that the patients were found to be highly susceptible to the events because of their poor health status.
• CMS's Lists of Hospital-Acquired Conditions. CMS's two policies on hospital-acquired conditions (HACs) create payment incentives for harm prevention by reducing payment for certain HACs. However, because the policies use narrowly scoped lists of HACs and employ specific criteria for counting harm events, they have limited effectiveness in broadly promoting patient safety. The lists did not cover most of the harm events that patients in our study experienced. Of the harm events we identified, only 5 percent were on CMS's HAC Reduction Program list and only 2 percent were on CMS's Deficit Reduction Act HAC list.
• Harm Events Resulting in Costs to Medicare. Nearly a quarter of Medicare patients who experienced harm events (23 percent), either preventable or non-preventable, required treatment that led to additional Medicare costs. These events also potentially increased patient costs in the form of coinsurance and deductible payments. Costs were incurred during the sample hospital stay or for an additional hospital stay necessary to ameliorate the harm. Combined, we estimated the costs for all events to be in the hundreds of millions of dollars for October 2018.

WHAT WE RECOMMEND AND HOW AGENCIES RESPONDED

Given the scale and persistence of patient harm in hospitals in the decade since our last report, HHS leadership and agencies must work with urgency to reduce patient harm in hospitals. Although HHS agencies took steps to improve patient safety in hospitals, including implementing many of our prior recommendations, substantial efforts are still needed. We made seven recommendations and received concurrences from CMS and AHRQ on all:

• We made the following three recommendations to CMS: (1) update and broaden its lists of HACs to capture common, preventable, and high-cost harm events; (2) explore expanding the use of patient safety metrics in pilots and demonstrations for health care payment and service delivery, as appropriate; and (3) develop and release interpretive guidance to surveyors for assessing hospital compliance with requirements to track and monitor patient harm. In its response to our draft report, CMS provided details about ongoing and planned efforts to improve patient safety.
• We made the following four recommendations to AHRQ: (1) with support from HHS leadership, coordinate agency efforts to update agency-specific Quality Strategic Plans; (2) optimize use of the Quality and Safety Review System, including assessing the feasibility of automating data capture for national measurement and to facilitate local use; (3) develop an effective model to disseminate information on national clinical practice guidelines or best practices to improve patient safety; and (4) continue efforts to identify and develop new strategies to prevent common patient harm events in hospitals. After receiving AHRQ's comments, we revised our first recommendation to make the recommended actions clearer and we revised the second recommendation to acknowledge recent progress by AHRQ regarding its event surveillance system.