Veterans Affairs vs. Mayo Clinic on MRSA

Veterans Affairs: Testing every patient for MRSA is effective at reducing hospital acquired MRSA infections.

Mayo Clinic: Testing every patient for MRSA is NOT effective at reducing hospital acquired MRSA infections.

University of Rochester:  High-quality hospitals have death rates that are 34 percent lower, while spending nearly 22 percent less, on trauma patient care than average-quality hospitals.

Comments (5)

Trackback URL | Comments RSS Feed

  1. Devon Herrick says:

    I wonder why Veterans Affairs came to such a different conclusion at the Mayo Clinic? The answer may be because Veterans Affairs allowed the problem to become worse, with MRSA taking up residence in VA facilities until the VA had to take drastic action.

    It’s not surprising that providing the right treatment the first time improves outcomes and costs less for trauma patients.

  2. Linda Gorman says:

    Average mortality hospitals were the reference case in the hospital study which defined high-mortality hospitals as “low quality” and low mortality hospitals as “high quality.” Unadjusted mortality rates were 2.61% in high-quality hospitals and 4.24% in low-quality hospitals.

    The only adjustment variable that had obvious, knock your eyes out, relative cost impact was the Area Wage Index. This means that labor costs matter.

    The confidence intervals calculated for the estimated relative cost of high-mortality outlier hospitals included zero, meaning that most of the time high-mortality hospitals had costs that were indistinguishable from those of average mortality hospitals.

    In the hospital study, cost was “total inpatient cost derived by multiplying the total hospital charges by the group average cost-to charges (C/C) ratios (defined as a weighed [sic] average for the hospitals in a group based on state, urban/rural, hospital ownership, and hospital size.” So cost was determined using hospital characteristics that might have an effect on hospital cost and in-hospital mortality.

    The association between patient level cost and hospital performance was “explored” with the coefficients from low-mortality and high-mortality outliers by “exponentiating the model paramaters estimated using the GLM models.”

    Some sample unadjusted costs, in dollars, for low mortality, average mortality, and high mortality hospitals:

    Blunt trauma: $7870, 9009, 8646
    Motor vehicle: $9940, 12862, 11890
    Pedestrian trauma $11,734, 15,056, 13,839

    The outlier here was stab trauma. It is the only case given in which high mortality hospitals cost more than the average at $5,583, $6,588, and $7,177. In all the other cases, both the low mortality and high mortality hospitals spent less than the average mortality hospitals.

  3. Linda Gorman says:

    The VA/Mayo abstracts suggest that the studies used two different baselines. The VA compares ICU infections to non-ICU infections. Mayo compares ICU infections in ICU units with testing and ICU units without testing.

    VA saw decline in ICU infections per 1,000 patient days from 1.64 to 0.62 from October 2007 to June 2010. Problem is that non-ICU infections also fell from 0.47 to 0.26. The decline in ICUs was larger, so victory was declared.

    Mayo took cultures from all 5434 ICU patients over 6 months but reported results only to 10 “intervention” ICUs. The remaining 8 ICUs were used as controls.

    The intervention ICUs took more precautions with infected patients. Everyone else got standard ICU precautions. There was no significant difference in “events of colonization or infection” per 1000 patient-days at risk, adjusted for baseline incidence.”

  4. Joe S. says:

    Linda, thanks for this analysis.

  5. Paul H. says:

    @ Linda

    Looks like the Mayo study is the more reliable of the two.