DETROIT – Despite a decreasing incidence of inpatients with Clostridium difficile colitis, hospital costs associated with caring for these patients continues to rise.
A large propensity score matched analysis found that mean Clostridium difficile colitis (CDC) admissions decreased by 13% from 2005 to 2008 in Pennsylvania, while the mean per admission costs for CDC patients rose by 9%.
Mean hospital costs per admission were $22,094 for patients with CDC vs. $10,865 for those without the infection.
"For each study year, CDC costs were consistently 50% higher than for non-CDC patients," Dr. David B. Stewart said at the annual meeting of the Central Surgical Association.
Dr. David B. Stewart
Notably, only 450 of the CDC patients, or less than 1%, underwent total colectomy for their infection.
Dr. Stewart and his colleagues at Pennsylvania State Hershey Medical Center in Hershey used the Pennsylvania Health Care Cost Containment Council (PHC4) database to identify CDC cases from 2005 to 2008 and study factors associated with CDC costs. Cost-to-charge ratios were used to derive per admission costs from hospital charges and were inflation adjusted to 2008 dollar values.
For 2005-2008, there were 7,227,788 hospital admissions in Pennsylvania, of which 78,273 patients, or 1.08%, had primary or secondary CDC. More than 75% of all patients were at least 65 years of age and had Medicaid as a payer. More than half, or 57%, were female and 87% were white.
Most of the hospitals were large (68%) and located in an urban setting (90%), while a minority (27%) were teaching hospitals, said Dr. Stewart, who specializes in colon and rectal surgery. Urban vs. rural setting was based on U.S. census definitions, and hospital size was based on Healthcare Cost and Utilization Program designation.
The analysis revealed that costs varied by hospital type. Small facilities had the highest costs overall at a mean of $25,465 per case compared with $20,160 for medium hospitals and $22,068 for large hospitals, he said.
Based on teaching status alone, teaching hospitals had significantly greater costs per CDC case at $33,528 vs. $17,854 for nonteaching hospitals .
Based on setting, urban hospitals had significantly higher costs at $23,194 vs. $11,969 for rural hospitals. In part, this is explained by the fact that urban and teaching hospitals had significantly higher Charlson Comorbidity Index scores among their patients than did rural and nonteaching facilities, Dr. Stewart said.
Overall, the heavy rural makeup of Pennsylvania, patient demographics, and the distribution of hospitals also lend themselves to higher costs for treating CDC. Out of 200 hospitals in the database, only 25 were designated as teaching facilities. Teaching hospitals cared for a disproportionate percentage, or 27%, of CDC admissions, he noted. None of the rural hospitals were designated as teaching facilities. Subsequently, rural patients with CDC had to travel roughly twice as far as urban residents for their care.
In linear regression and multivariable analysis, factors significantly associated with higher attributable excess costs for CDC included age less than 18 years, male gender, ethnicity other than white or African American, elective hospital admission, Medicaid payer type, a Charlson comorbidity index of 3 or more, urban hospital setting, and small hospital setting.
The biggest detriment to CDC research is the complete dearth of information regarding such prognostic factors as bacterial virulence factors, toxin type, and sensitivities to antibiotics that would help to explain the disease manifestations clinicians are observing in their patients, Dr. Stewart said. This is a reflection of the rapid change the disease has undergone over the last 2 decades, creating a great deal of uncertainty about the disease.
"If you had to describe the state of CDC management at this time, unfortunately, you would have to characterize it as being roughly empiric treatment," he said. "We’re using the same antibiotics we did 20 years ago when they worked better, and we have no diagnostic tools with which to differentiate any one patient’s infection from anyone else’s.
"This creates a great deal of uncertainty into the disease process and disease course, and as such, it might be wise if we began to look at CDC like we do other potentially surgical diseases like adhesive small bowel obstructions or diverticulitis, where the preferential treatment is nonoperative, but it mandates the early or even principal involvement of a surgeon."
This might also have implications as to where CDC patients are admitted and how early in their course they are transferred to other facilities, he said.
During a discussion of the study, it was asked whether the presence of particularly virulent strains such as B1/NAP1 may be driving up hospital costs overall. Dr. Stewart said that according to preliminary data from 25 tissue samples he’s analyzed and typed at his institution, 82% were NAP1 and had such virulence characteristics as the presence of binary toxin, toxins A&B, and very low minimal inhibitory concentrations for metronidazole (Flagyl) and vancomycin (Vancocin). Still, patients with these strains were no more likely to require ICU care, to undergo surgery or to have longer hospital stays, he said, noting that his institution does only about 10 colectomies per year.