The Importance of Identifying Potential Inpatient Complications

By Richard Bankowitz, MD

This November marks the 15-year anniversary of the Institute of Medicine report, “To Err is Human: Building a Safer Health System.” This groundbreaking analysis of our health system estimated that preventable medical errors cause 98,000 inpatient hospital deaths each year in the United States. Since it was published, hospitals and other care providers have led impressive efforts to better measure, report and prevent harm events. To give just a few examples, hospitals participating in Premier’s QUEST collaborative have been improving in 23 distinct measures of potential harm, preventing an estimated 17,991 potential patient safety events. Similarly, hospitals participating in the federal Partnership for Patients program report thousands of avoided harm events and dramatic savings as a result.

However, there remain areas of unjustified variation in harm events that deserve attention as an opportunity for cost and quality improvement.

To help hospitals better identify harm events that have the most significant impact on cost and quality, Premier undertook a study of nearly 6 million patient records in 2013 to identify, analyze and evaluate a wider set of potential inpatient complications – or PICs. The methodology used to identify these PICs was peer-reviewed and published in the American Journal of Healthcare Quality.

The study yielded 86 “high-impact” complications associated with nearly 50,000 deaths, $4.3 billion in added costs and 1.7 million added hospital days.

The analysis also found that while some of these high-impact complications are covered in federal payment policies, many are not. Combined, only 22 of the 86 high impact conditions are measured through current policies, capturing just 9 percent of the patients experiencing a harmful complication.

Similarly, payment measures have a far lesser impact on mortality, cost and length of stay. Of the 49,827 potential deaths associated with PICs, harms covered in the Medicare reimbursement policy accounted for 1,071 of the total, or 2 percent. Of the more than $4 billion in added costs associated with PICs, Medicare-measured harm made up just 13 percent of the total. And in length of stay, the Medicare measures accounted for just 12 percent of the total added days associated with the PICs. This suggests that although the Medicare-measured harm is important to prevent, an exclusive focus on these conditions may lead to missed opportunities to identify and improve care processes for other complications whose elimination could substantially lower mortality, length of stay and costs. Moreover, such a narrow list of measures fails to capture the broader picture of unjustified variation, and the quality of patient care in a particular institution.

To help providers prioritize and focus their quality improvement efforts, Premier developed a list of 10 PICs that had the highest frequency, as well as a significant association with mortality, cost and length of stay. As part of the process, we found that respiratory failure was the most expensive complication, associated with additional costs to health systems of $940 million. In addition, to the costs it was associated with more than two additional days of stay per patient and more than 16,000 deaths for the 119,000 patients with this PIC. Similarly, acute renal failure affected about 144,000 cases, and was associated with $490 million in additional costs, 1.78 extra days per case and about 2,700 deaths. Hypotension also affected approximately119,000 cases, and was associated with an additional $200 million in costs, .58 extra days to the length of stay and an additional 1,500 deaths.

Of the top 10, just two of the conditions measured by the Medicare payment programs were found to have a statistical effect on length of stay, costs and mortality. Sepsis, for instance, was a PIC in about 51,000 patients, and was associated with $330 million in added costs, adding 2.34 days to the length of stay and 669 deaths. Similarly, pulmonary embolism was a PIC in about 8,700 patients, and associated with 2.67 additional days in the length of stay and $44 million in added costs.

Although not all of these complications may be preventable, using a broader measure for identifying harm could assist hospital leaders with prioritization and help them focus their quality improvement efforts by broadening hospital ability to identify, and then address care processes that may lead to or contribute to complications that have a high impact on other outcomes (mortality, length of stay and cost). Rather than taking a myopic view and just evaluating performance against Medicare measures of harm, hospitals should examine their own data to find unjustified variation in at least the top 10 conditions and identify whether processes could be improved to prevent harm events in the areas of acute renal failure, hypotension, respiratory failure, sepsis, aspiration pneumonia, acute myocardial infarction, gastrointestinal ulceration and hemorrhage, cerebral infarction, pulmonary embolism and ventilator-associated pneumonia.

Ramon Meguiar, MD, senior vice president and chief medical officer at Memorial University Medical Center (Savannah, Ga.) summed it up well when he said, “CMS HACs are serious and important events to prevent, but providers need to focus on a wider set of higher-impact complications to avoid missing opportunities for quality and cost improvement.”

The safety measures tracked through federal programs should be of significant focus for hospitals. But identifying improvement areas across a wider set of complications is just as important to deliver more efficient, higher-quality care overall.

Richard Bankowitz, MD, is chief medical officer of Premier, Inc.