The first comprehensive analysis of inpatient hospital costs for COVID-19 patients in the United States from April to December 2020 was published in the peer-reviewed medical journal Advances in Therapy.
The study was published in October 2021 by a team of collaborators including Dr. Robert L. Ohsfeldt, a regents professor and Ph.D. program co-chair in the Department of Health Policy and Management at the Texas A&M University School of Public Health.
“The [study] was motivated by a need to have a source of credible evidence for hospital cost for use in economic evaluations of COVID treatments,” Ohsfeldt said.
Other collaborators on the six-to-eight-month project were Casey Kar-Chan Choong, Patrick L. McCollam and Hamed Abedtash, who are all with study sponsor Eli Lilly and Company of Indiana; Kari A. Kelton, with Medical Decision Modeling Inc. of Indiana; and Russel Burge with the University of Cincinnati, GPORWE-Bio-Medicines and Eli Lilly and Company.
The study examined the hospitalization costs for COVID-19 patients, factors associated with costs and length of stay, and the monthly trends of costs and length of stay from April to December 2020 in the United States, according to a Texas A&M Today article.
The independent variables included: age groups, sex, race, ethnicity, geographic region, comorbidity, hospital type, urban status, payer type, Adaptive COVID-19 Treatment Trial Ordinal Scale, Medicare Severity Diagnosis-Related Groups codes, discharge status and admission month; while dependent variables were cost and length of stay.
Results of this study showed that the median length of stay in hospitals was six days and the median hospital cost was $11,267. The median length of stay in intensive care units was five days and median ICU cost was $13,443. Older age, comorbidities and mechanical ventilation were major drivers of costs, hospital length of stay and risk of death, and a downward trend of cost and hospital stay was observed from April to December 2020, according to the study.
The team used admissions and discharge data from patient records in the Premier Healthcare Database to estimate the inpatient costs and length of stay for patients in the hospital overall, as well as those patients specifically in the ICU; regression analyses were used to examine patient and hospital characteristics for factors associated with changes in costs and length of stay, according to the A&M article.
“Premier Healthcare Database is a commercial database; what happens is a lot of hospitals will contribute their cost data, so this is data from their internal cost accounting systems that they will contribute to this data warehouse,” Ohsfeldt said.
He said not every hospital was included, but “a fair number of hospitals across the U.S. were.”
“The sample we had for the study was about 250,000 COVID admissions from April to December, so it is a fairly big set of hospitals that contributed data,” he said. “It is different than a lot of data that people use because it is not based on insurance claims, it comes from the cost data from the hospitals themselves; it is their internal cost accounting data that they use to track how much it costs them to produce the inpatient services that they produce.”
He also said he wasn’t sure how many people would be interested in the study because it was made up of a lot of numbers and variables. The main limitation the collaborators faced was working with a smaller amount of data, because of it was not readily available to them at the time of the study.
“The main limitation is that it takes time for data to become available, so at the time we started the study that was the most recent time frame for data availability,” he said.
Ohsfeldt said a somewhat unexpected pattern was that hospitalization costs declined over time.
“The costs over time were declining fairly substantially and I guess that makes sense, [because] keep in mind, this is adjusted for the changes in the composition of COVID cases that were hospitalized,” he said. “It is not because the COVID cases were getting less severe or other things about the COVID case mix [where] cases that happened to be hospitalized [from] April to December; so the analysis adjusted for that.”
He said the cost trends of hospitalizations had been “standardized in terms of their severity and the types of patients that were being hospitalized.”
Ohsfeldt said they gathered that the overall cost average of hospitalization’s declined 26 percent from April to December 2020.
“If you think about it logically, that makes sense because when COVID first presented itself, the health-care system didn’t know exactly how to handle it and were trying various things to treat it and some of those things didn’t work very well so they were not necessarily treating it as effectively as they could have,” he said. “Once they had time to learn about how to treat it more effectively, by December they learned what worked and what didn’t work and were treating it more effectively and less expensively.”
He said he didn’t necessarily expect to see a decline, but it was not “overwhelmingly surprising,” but the “magnitude” wasn’t something he expecting. He also said the study showed that older people had the higher COVID severity and higher hospital costs.
“That is also something that is pretty much what you would have expected to see in terms of the patterns across different types of patients,” he said.
The team observed a downward trend in length of stay, cost, discharge to post-acute care, and discharge due to death from April to December 2020; over this time period, there was almost a 50% reduction in median hospital costs, the study stated.
The results of the study suggest that reducing time to recovery by only one hospital day may save $2,118 per patient/day on average, and, if applied to the 198,806 patients in the cost analysis alone, the cost savings would exceed $421 million, the study stated. Reducing ICU length of stay by one day would save $3,586 per patient/day on average, and over $251 million based on the 70,054 ICU patients in the cost analysis, according to the study.
To read the study in detail, visit rdcu.be/cEIwu.