Results (
Thai) 1:
[Copy]Copied!
Study Data And MethodsIn an earlier study we analyzed data from799 nonfederal acute care general hospitals ineleven states. Discharge abstracts and nursestaffing data were obtained from the states;data on hospital size, location, teaching status,from the American Hospital Association(AHA) annual survey; and cost-to-charge ratios,from Medicare cost reports.In regression analyses we found an associationof nurse staffing and (1) lengths-of-stay,urinary tract infections, upper gastrointestinalbleeding, hospital-acquired pneumonia,shock, or cardiac arrest among medical patientsand (2) “failure to rescue,” defined as thedeath of a patient with one of five life-threateningcomplications—pneumonia, shock orcardiac arrest, upper gastrointestinal bleeding,sepsis, or deep vein thrombosis—among surgicalpatients. Details of that study are describedelsewhere.3 Exhibit 1 presents rates of theseoutcomes and descriptive statistics for the799-hospital sample.In this study we simulated the effect ofthree options to increase nurse staffing: raisethe proportion of hours provided by registerednurses (RNs) to the seventy-fifth percentile forhospitals below this level; raise the number oflicensed (that is, RNs and licensed practical/vocational nurses, or LPNs) nursing hours perday to the seventy-fifth percentile; and raisestaffing to each of these levels in hospitalswhere each is below the seventy-fifth percentile.This percentile was chosen based on ourjudgment that attaining this level of staffing isfeasible for most hospitals (Exhibit 2).The required number of additional nursehours to meet the seventy-fifth-percentile levelswas estimated from the original sample. Estimatesof avoided adverse outcomes and daysof care were simulated from the regressionmodels from the earlier study, and estimates ofavoided costs and deaths were made with additionalregression modeling in the originaldata. Costs of avoided adverse outcomes wereestimated from patient-level regressions ofcosts per case on patient diagnosis and othercharacteristics and variables for each adverseoutcome. Costs of avoided days were estimatedby multiplying average costs per day byregression-based estimates of reduced days netof the days associated with adverse outcomes.Because many hospital costs are fixed in theshort run, hospitals might not fully recover theaverage costs of avoided days or avoided complications.Based on a review of studies of hospitalfixed and variable costs, we estimatedvariable costs of hospitals to be 40 percent ofaverage costs, and we multiplied calculatedcosts by this amount to estimate the shorttermcost impact of reduced hospital patientdays and avoided adverse outcomes.4 Overtime, hospitals should be able to adjust theirfixed costs to reflect the change in volume. Wepresent estimates of cost savings assumingshort-term savings of 40 percent of averagecosts and with full recovery of fixed costs.We projected the results from the sample toall nonfederal U.S. acute care hospitals and updatedthe estimates of needed staffing, avoidedadverse outcomes and days, and costs to reflecthospital costs, admissions, and lengthsof-stayin 2002. Specifically, our sample had 26percent of the discharges from U.S. nonfederalacute care hospitals in 1997. We constructednational estimates of adverse outcomes, nursingfull-time equivalents (FTEs), and costs bymultiplying estimates from the sample by 100divided by 26. We used data on RN wagesfrom the 1997 and 2002 Current PopulationSurveys (CPS) and the change in admissions,lengths-of-stay, spending per admission, andspending per day between 2002 and 1997 fromthe AHA annual survey to update the estimatesof avoided adverse outcomes, avoideddays, deaths, and costs. In aggregate, between1997 and 2002, licensed hours per day and theproportion of licensed hours provided by RNsreported to the AHA, and average case-mix,measured by the Medicare case-mix index, didnot change substantially; thus, no adjustmentswere made to the staffing variables.5Because neither our prior work nor otherstudies capture all of the effects of nurse staffingon patient care, and because we do nothave direct measures of patient-reported quality,we do not attempt a cost-effectivenessanalysis of the impact of raising nurse staffing.We do present estimates of the cost peravoided death.
Being translated, please wait..
