Clinical Context

Hospital safety and the health risks for adverse events (AEs) among inpatients have been the focus of countless hours and dollars spent over the past several decades, and the authors of the current study credit the Harvard Medical Practice Study (HMPS) as critical research that led to programs to improve hospital safety. This study examined AEs among inpatients in New York hospitals in 1984. Results of the study were published in 2 articles published in The New England Journal of Medicine.^[1-2]

Researchers found an overall AE rate of 3.7 per 100 admissions; 28% of AEs were thought to be caused by negligence, and 16% of AEs led to death or disability.

Since then, multiple programs have been initiated to promote safe prescribing, reduce healthcare-acquired infections, and prevent errors related to surgery and procedures. These interventions are buttressed by technology that could hardly be imagined in the 1980s, but have we had influence in the rate of AEs in the hospital? The current study by Bates and colleagues uses a methodology similar to that employed in HMPS to find out.

Study Synopsis and Perspective

Nearly 25% of hospital admissions included at least 1 AE, as indicated from data from 2809 admissions at 11 hospitals.

The 1991 Harvard Medical Practice Study (HMPS),^[1] which focused on medical injury and litigation, documented an AE rate of 3.7 events per 100 admissions; 28% of those events were attributed to negligence, wrote David W. Bates, MD, of Brigham and Women's Hospital, Boston, Massachusetts, and colleagues.

Although patient safety has changed significantly since 1991, documenting improvements has been challenging, the researchers said. Several reports have shown a decrease in healthcare-associated infections; however, other aspects of safety -- notably, adverse drug events, defined as injuries resulting from drugs taken -- are not easily measured and tracked, the researchers said.

"We have not had good estimates of how much harm is being caused by care in hospitals in an ongoing way that looked across all types of [AEs]," and the current review is therefore important, Bates said in an interview.

In a study recently published in The New England Journal of Medicine,^[3] the researchers analyzed a random sample of 2809 hospital admissions from 11 hospitals in Massachusetts during the 2018 calendar year. The hospitals ranged in size from fewer than 100 beds to more than 700 beds; all patients were aged 18 years and older. A panel of 9 nurses reviewed the admissions records to identify potential AEs, and 8 physicians reviewed the AE summaries and either agreed or disagreed with the AE type. They ranked the severity of each event using a general severity scale into categories of significant, serious, life-threatening, or fatal.

Overall, at least 1 AE was identified in 23.6% of the hospital admissions. A total of 978 AEs were deemed to have occurred during the index admission, and 222 of these (22.7%) were deemed preventable. Among the preventable AEs, 19.7% were classified as serious, 3.3% as life-threatening, and 0.5% as fatal.

A total of 523 admissions (18.6%) involved at least 1 significant AE, defined as an event that caused unnecessary harm but from which recovery was rapid. A total of 211 admissions involved a serious AE, defined as harm resulting in substantial intervention or prolonged recovery; 34 included at least one life-threatening event; and 7 admissions involved a fatal AE.

A total of 191 admissions involved at least 1 AE deemed preventable. Of those, 29 involved at least 1 preventable AE that was serious, life-threatening, or fatal, the researchers wrote. Of the 7 deaths in the study population, 1 was deemed preventable.

The most common AEs were adverse drug events, which accounted for 39% of the AEs; surgical or other procedural events accounted for 30.4%; patient care events (including falls and pressure ulcers) accounted for 15%; and healthcare-associated infections accounted for 11.9%.

Overcoming Barriers to Better Safety

"The overall level of harm, with nearly 1 in 4 patients suffering an AE, was higher than I expected it might be," Bates told Medscape; however, techniques for identifying AEs have improved, and "it is easier to find them in electronic records than in paper records."

"Hospitals have many issues they are currently dealing with since COVID, and one issue is simply prioritization," Bates told Medscape, "but it is now possible to measure harm for all patients using electronic tools, and if hospitals know how much harm they are having in specific areas, they can make choices about which ones to focus on."

"We now have effective prevention strategies for most of the main kinds of harm," he said.

Generally, rates of harm are high because these strategies are not being used effectively, he said.

"In addition, there are new tools that can be used -- for example, to identify patients who are decompensating earlier," he noted.

As for additional research, some specific types of harm that have been resistant to interventions, such as pressure ulcers, deserve more attention, said Bates. "In addition, diagnostic errors appear to cause a great deal of harm, but we don't yet have good strategies for preventing these."

The study findings were limited by several factors, including the use of data from hospitals that might not represent hospitals at large and by the inclusion mainly of patients with private insurance, the researchers wrote. Other limitations include the likelihood that some AEs were missed, and the level of agreement on AEs between adjudicators was only fair.

Still, the findings serve as a reminder to healthcare professionals of the need for continued attention to improving patient safety, and measuring AEs remains a critical part of guiding these improvements, the researchers concluded.

Timely Reassessment and Opportunities to Improve

In the decades since the publication of the report, To Err Is Human, by the National Academies in 2000,^[4] significant attention has been paid to improving patient safety during hospitalizations, and healthcare systems have increased in both system and disease complexity, said Suman Pal, MBBS, a specialist in hospital medicine at the University of New Mexico, Albuquerque, New Mexico, in an interview.

"Therefore, this study is important in reassessing the safety of inpatient care at the current time," he said.

"The findings of this study showing preventable [AEs] in approximately 7% of all admissions, while concerning, is not surprising, as it is consistent with other studies over time, as the authors have also noted in their discussion," said Pal.

The current findings "underscore the importance of continuous quality improvement efforts to increase the safety of patient care for hospitalized patients," he noted.

"The increasing complexity of medical care, fragmentation of healthcare, structural inequities of health systems, and more recent widespread public health challenges such as the COVID-19 pandemic have been, in my opinion, barriers to improving patient safety," Pal said.

"The use of innovation and an interdisciplinary approach to patient safety and quality improvement in hospital-based care, such as the use of machine learning to monitor trends and predict the individualized risk of harm, could be a potential way out" to help reduce barriers and improve safety, he said.

"Additional research is needed to understand the key drivers of preventable harm for hospitalized patients in the United States," said Pal. "When planning for change, keen attention must be paid to understanding how these [drivers] may differ for patients who have been historically marginalized or are otherwise underserved so as to not exacerbate healthcare inequities," he added.

The study was funded by the Controlled Risk Insurance Company and the Risk Management Foundation of the Harvard Medical Institutions. Bates owns stock options with AESOP, Clew, FeelBetter, Guided Clinical Solutions, MDClone, and ValeraHealth and has grants/contracts from IBM Watson and EarlySense. He has also served as a consultant for CDI Negev. Pal has disclosed no relevant financial relationships.

Clinical Implications

• In HMPS, which gathered information on AEs in hospitals in 1984, there were 3.7 AEs per 100 admissions.
• The current study by Bates and colleagues found at least 1 AE during 23.6% of admissions; 22.7% of these events were judged to have been preventable. The most common categories of AEs were adverse drug events and surgical or procedural events.
• Implications for the healthcare team: To promote patient safety, the healthcare team should perform careful analyses of local rates of hospital AEs and the interventions used to prevent them.

Earn Credit