A set of 14 hospital-acquired conditions that the CMS considers avoidable accounted for 48,771 adverse patient outcomes, 3,219 deaths and more than $2 billion in excess hospital costs in 2016, according to a new research brief.
Patients experiencing these hospital-acquired conditions (HACs) were in the hospital about 8 days longer on average than other patients, with an average excess hospital cost of $41,917, not including physician fees, according to the analysis by IBM Watson Health. Their increased mortality risk was 72.3%.
One expert said the IBM Watson data show that the CMS is justified in reducing Medicare payments to hospitals that rank in the worst-performing 25% of all hospitals on these 14 quality measures. Facilities in that group are hit with a 1% payment cut in the next fiscal year.
"This paper indicates that this payment policy identifies real problems and makes sense," said Noel Eldridge, a patient-safety specialist with the federal Agency for Healthcare Research and Quality. "If these patients have a cost of $42,000 more and a much-increased chance of dying, these HACs are pretty consequential."
The 14 HACs covered in the IBM Watson analysis include air embolism, blood incompatibility, stage 3 and 4 pressure ulcers, falls and trauma, catheter-associated urinary tract infection, and deep vein thrombosis/pulmonary embolism with total knee or hip replacement. These include so-called never events for which the CMS does not pay for additional care associated with the complication.
The American Hospital Association has argued that the CMS' HAC Reduction Program has a flawed methodology because it doesn't recognize improvement each year. The program, mandated under the Affordable Care Act, also has been criticized for lacking risk adjustment for the socioeconomic status of a hospital's patient population.
The CMS had proposed to stop releasing data on hospital-acquired infections but withdrew that proposal earlier this year.
Dr. Anil Jain, chief health informatics officer at IBM Watson, said the new analysis shows that the CMS' HAC reporting and penalty program may not be enough by itself to reduce the number of these dangerous events.
"One would expect that if you penalize hospitals that have HACs they shouldn't have, and you put quality indicators out there, hospitals will find a way to reduce them," said Jain, who also practices as an internist at the Cleveland Clinic.
"But sometimes you need to examine how you go from A to B to C to D, from admission to discharge, and make data-driven process improvements," he said. "Just tweaking won't get sustainable results."
A report issued in June by the AHRQ found that the rate of hospital-acquired conditions overall declined by 8% from 2014 to 2016, saving about $2.9 billion and preventing about 8,000 deaths.
The trend line for that much larger set of HACs than those studied by IBM Watson has been dropping consistently since 2010, Eldridge said. He attributed that decline to providers increasingly using evidence-based practices and root-cause analysis to improve clinical care processes.
IBM Watson's analysis drew on an inpatient database of nearly 20 million CMS and private insurance discharges in 2016.
