Public hospital admissions for treating complications of clinical care: incidence, costs and funding strategy.

OBJECTIVE: To quantify the frequency of, and the costs and payments associated with, admissions for treatment of injuries and illnesses that are consequences of care. DATA SOURCES: Routinely-coded 2005/06 public hospital inpatient data from Victoria, Australia (1.25 million admissions) and corresponding patient-level cost data (1.04 million admissions). Payments reflected DRG-based prospective rates. STUDY DESIGN: Retrospective analysis of admissions with principal diagnoses that specify adverse events arising as a direct consequence of healthcare. RESULTS: 1.5% (15,336) of the costed admissions specifically treat an injury or illness arising from medical or surgical care, consuming 2.74% of hospital prospective payments and representing $89.3 m (2.84%) of total reported costs. 1.4% (17,429) of all public hospital admissions and 2.82% of hospital prospective payments (estimated cost-$101.5 m per year) are committed to treating complications of care. Private residences or aged care facilities are the source of 84.9% (14,804) of these admissions. Inpatient death was the outcome in 0.7% (118) of these admissions. IMPLICATIONS: Admissions for treating complications of care represent a small, relatively expensive, proportion of hospital admissions, which account for disproportionate levels of hospital costs and funding. A policy option providing incentives to reduce the incidence and costs of complications arising from care includes allocating all costs arising from transferred (re)admissions back to the original episode of care and developing a suite of specific DRGs to fund admissions for treatment of complications.

The US Medicare policy of not reimbursing hospital-acquired conditions: what impact would such a policy have in Victorian hospitals?

OBJECTIVE: To model the effect of excluding payment for eight hospital-acquired conditions (HACs) on hospital payments in Victoria, Australia. DESIGN, SETTING AND PARTICIPANTS: Retrospective ecological study using the Victorian Admitted Episodes Dataset. The analysis involved all acute inpatient admissions to Victorian public and private hospitals between 1 July 2007 and 30 June 2008. INTERVENTIONS: Each admission record includes up to 40 diagnosis and procedure codes from which payments are calculated. The model deleted diagnosis codes for eight HACs from all records, then recalculated payments to estimate the impact of a policy of non-payment for HACs. MAIN OUTCOME MEASURE: The effect on hospital payments of excluding diagnosis codes for eight HACs. RESULTS: 2,047,133 cases with total estimated payments of $4902 million were identified; 994 cases (0.05%) had one or more diagnoses meeting the code definition for a definable HAC, representing total payments of $24.1 million. In-hospital falls and pressure ulcers were the most commonly coded HACs. Applying a model that excluded HAC diagnosis codes changed the diagnosis-related group for 134 cases (13.5%), thereby generating a $448,630 reduction in payments. CONCLUSIONS: Introducing a non-payment for HACs policy similar to that introduced by Medicare in the United States would have little direct financial impact in the Australian context, although additional savings would accrue if HAC rates were reduced. Such a policy could add further incentive to current initiatives aimed at reducing HACs.

Prospective payment to encourage system wide quality improvement.

BACKGROUND: Casemix-based inpatient prospective payment systems allocate payments for acute care based on what is done within an episode of care without regard for the outcome. To date, they have provided little incentive to improve quality. The Centers for Medicare & Medicaid Services have recently excluded 8 avoidable complications from their payment system. OBJECTIVE: This study models an inpatient prospective payment system that comprehensively excludes not-present-on-admission and other complication diagnoses from the entire funding process, effectively adding a diagnosis-related group (DRG)-specific average complication payment across all discharges. RESEARCH DESIGN: Complication-averaged cost weights were estimated using the same patient level cost dataset used for estimating the relative resource weights for Victorian public hospitals in 2006-07. All codes with a "C" prefix (secondary diagnoses that are coded as having arisen after admission) and codes that define a condition that prima facie represent a specific complication of care were excluded from the code string. The episodes were then regrouped to DRGs and new complication-averaged cost weights were developed. RESULTS: When complication codes were excluded across 1.2 million discharges, 1.37% became ungroupable, 14.86% included at least one complication diagnosis code, and 1.56% grouped to another DRG. Modeled funding for individual metropolitan hospitals in Victoria, Australia, was redistributed by -2.5% to 1.8%. CONCLUSIONS: The cost weights reflect the average cost of preventable and unpreventable complications and have the potential to drive improvements in clinical care. This study is in contrast to previous studies estimating the funding impact of preventing all complications.

Risk adjustment policy options for casemix funding: international lessons in financing reform.

This paper explores modified hospital casemix payment formulae that would refine the diagnosis-related group (DRG) system in Victoria, Australia, which already makes adjustments for teaching, severity and demographics. We estimate alternative casemix funding methods using multiple regressions for individual hospital episodes from 2001 to 2003 on 70 high-deficit DRGs, focussing on teaching hospitals where the largest deficits have occurred. Our casemix variables are diagnosis- and procedure-based severity markers, counts of diagnoses and procedures, disease types, complexity, day outliers, emergency admission and "transfers in." The results are presented for four policy options that vary according to whether all of the dollars or only some are reallocated, whether all or some hospitals are used and whether the alternatives augment or replace existing payments. While our approach identifies variables that help explain patient cost variations, hospital-level simulations suggest that the approaches explored would only reduce teaching hospital underpayment by about 10%. The implications of various policy options are discussed.