Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

The Centers for Medicare and Medicaid Services' decision to stop paying for certain preventable hospital complications. beginning in 2008, may eventually diffuse into the Medicaid program. However, because Medicaid is a joint responsibility of the state and federal governments, the implementation and effects of such a policy in the Medicaid program may vary across states. Moreover, recent studies suggest that existing state Medicaid policies, which affect approximately 40 percent of all pediatric inpatient discharges, may actually contribute to preventable complications among hospitalized children (Miller and Zhan 2004; Smith et al. 2007). If so, then a policy of withholding payment may only exacerbate the problem of preventable adverse events, particularly in states with less generous reimbursement or in hospitals with a high proportion of Medicaid beneficiaries.

Objective: Cases involving preventable complications among hospitalized children are associated with substantially longer length of visit (and higher total costs) than cases without complications. The purpose of this study is to examine the relationship between Medicaid characteristics of the patient, hospital, and state to the probability and length of a prolonged hospital stay, which may in part be due to the occurrence of a preventable adverse event for a hospitalized child. Results will provide insights on whether a nonpayment policy for preventable adverse events (such as that initiated for Medicare) would be advisable in the Medicaid program.

Data: The population for our study will consist of discharges from the Agency for Healthcare and Research's (AHRQ) 2003 Kids' Inpatient Database (KID), part of the Healthcare Cost and Utilization Project (HCUP) and the only all-payer inpatient care database for children in the United States. The 2003 KID contains data drawn from 36 states and 3,438 hospitals, representing over 7 million inpatient discharges for children 20 years of age and younger. Our analytic sample will consist of a subset of the 2003 KID, drawn for all patients, between 0 and 17, who are at risk of experiencing one of several carefully selected pediatric safety events, as defined by the most recent version of AHRQ's Pediatric Quality Indicators (PDI) software.

Methodology: We follow the method of Silber et al. (2003) to determine when a visit becomes prolonged, for each patient safety event and for each of the 36 states in our sample. Then, we employ two empirical models to test two main hypotheses. The first is that Medicaid reimbursement tends to fall below the cost of care, and therefore gives hospitals the incentive to reduce length of stay regardless of the presence, or consequent occurrence, of a complication. We use a logit model to estimate the effect of patient, hospital, and state Medicaid characteristics on the probability that a discharge becomes prolonged. The second hypothesis is that Medicaid influences the length of an extended hospital stay. Thus, in the second model, we estimate the prolonged length of stay, using only those discharges that experience a prolonged visit. Both models control for the occurrence of a patient safety event (PDI), illness severity, and other patient and hospital characteristics.