Results of a round-robin exercise on read-across.

A round-robin exercise was conducted within the CALEIDOS LIFE project. The participants were invited to assess the hazard posed by a substance, applying in silico methods and read-across approaches. The exercise was based on three endpoints: mutagenicity, bioconcentration factor and fish acute toxicity. Nine chemicals were assigned for each endpoint and the participants were invited to complete a specific questionnaire communicating their conclusions. The interesting aspect of this exercise is the justification behind the answers more than the final prediction in itself. Which tools were used? How did the approach selected affect the final answer?

A framework for validation of rule-based systems.

We describe a complete methodology for the validation of rule-based expert systems. This methodology is presented as a five-step process that has two central themes: 1) to create a minimal set of test inputs that adequately cover the domain represented in the knowledge base; and 2) a Turing Test-like methodology that evaluates the system's responses to the test inputs and compares them to the responses of human experts. The development of minimal set of test inputs takes into consideration various criteria, both user-defined, and domain-specific. These criteria are used to reduce the potentially very large set of test inputs to one that is practical, keeping in mind the nature and purpose of the developed system. The Turing Test-like evaluation methodology makes use of only one panel of experts to both evaluate each set of test cases and compare the results with those of the expert system, as well as with those of the other experts. The hypothesis being presented is that much can be learned about the experts themselves by having them anonymously evaluate each other's responses to the same test inputs. Thus, we are better able to determine the validity of an expert system. Depending on its purpose, we introduce various ways to express validity as well as a technique to use the validity assessment for the refinement of the rule base. Lastly, we describe a partial implementation of the test input minimalization process on a small but nontrivial expert system. The effectiveness of the technique was evaluated by seeding errors into the expert system, generating the appropriate set of test inputs and determining whether the errors could be detected by the suggested methodology.

Some impacts of nursing on acute care hospital outcomes.

Measuring nursing-sensitive patient outcomes using publicly available data provides exciting opportunities for the nursing profession to quantify the patient care impact of staffing changes at individual hospitals and to make comparisons among hospitals with differing staffing patterns. Using data from California and New York, this study tested the feasibility of measuring such outcomes in acute care hospitals and examining relationships between these outcomes and nurse staffing. Nursing intensity weights were used to acuity-adjust the patient data. Both higher nurse staffing and higher proportion of RNs were significantly related to shorter lengths of stay. Lower adverse outcome rates were more consistently related to a higher proportion of RNs.

Family practice graduate medical education and hospitals' patient care costs in New Jersey.

BACKGROUND: More generalists are needed for the American health care system. Will training these practitioners add to hospital costs? Although graduate medical education has been shown to add to hospital patient care costs, the authors questioned whether this were true for the hospital training of family physicians. METHOD: Based on data supplied by 12 participating New Jersey hospitals with family practice residencies, inpatients in 1991 were separated into three categories by the teaching status of their attending physicians: family practice, other teaching, non-teaching. The hospitals were stratified into two types for analysis: seven community and five multiresidency hospitals. Average cost (per case-mix--neutral case) was found for each category of patients within medical, surgical, pediatric, and obstetrical classes. RESULTS: Among community teaching hospitals, the mean case-mix--adjusted cost per case for inpatients associated with family practice training was 6.3% less than that for inpatients with an equivalent case mix not associated with family practice training. Among multiresidency teaching hospitals, there was no difference between mean costs for inpatients associated with family practice training and non-teaching patients. The mean cost for inpatients associated with graduate medical training other than family practice was 8% higher than that for non-teaching inpatients. CONCLUSION: These findings suggest that family practice residencies do not add to the direct inpatient costs of teaching hospitals, and in certain instances may even reduce hospital patient care costs. In times of increasing cost consciousness in health care and medical education, this provides a further rationale for institutions to sponsor graduate training in family practice.

Identification and categorization of and cost for care of trauma patients: a study of 12 trauma centers and 43,219 statewide patients.

Medical and demographic data for trauma patients (n = 7120) admitted to 12 trauma centers in 1 year were reviewed. Data from New York State on all discharges for the same year (n = 2,535,501) were obtained and analyzed. Patients were identified as trauma patients based on NYC EMS trauma center advisory committee criteria translated into ICD-9-CM codes, and a computer-based algorithm was developed that identified 43,219 trauma patients. A standard resource cost (SRC) was also developed to compare relative cost among trauma and non-trauma patients in the same diagnosis-related groups (DRGs). The mean age of the 43,219 trauma patients was 43.1 years, 61.8% were male, the mean LOS was 13.4 days, the mean ISS was 10.4, and 61% were discharged from community hospitals. Trauma centers treated the more severely injured patients: mean ISSs were 12.3, 10.9, and 9.2 for level I, level II, and community hospitals, respectively. Payor mix varied by category, with 71% of penetrating trauma victims covered by Medicaid or self pay compared with 21% of blunt trauma victims. Level I centers treated twice as many self-pay and Medicaid patients as community hospitals. A comparison of relative cost showed that trauma patients cost 27.5 million dollars more than non-trauma patients in the same DRGs.

Measuring variations in nursing care per DRG.

A state nurses' association, a state department of health and a consulting agency collaborate annually in a special nursing study to develop appropriate measures of the relative variation in nursing care associated with each DRG. This activity provides a mechanism for professional nurses and the nursing profession to have direct input into a state's reimbursement methodology. This methodology is unique in the nation and may serve as an important prototype for nurses in other states who are seeking recognition of nursing care and services in the DRG reimbursement system.

Revising diagnosis-related groups for neonates.

Groups of neonates who are usually treated at hospitals that provide specialized pediatric care are not adequately classified by the use of diagnosis-related groups (DRGs). Therefore, a set of revised DRGs, pediatric modified DRGs (PM-DRGs), have been developed. Use of PM-DRGs substantially improves the classification of neonates in the following ways: a single pediatric modified major diagnostic category has been defined to include only and all neonates (patients younger than 29 days of age when admitted to the hospital); deaths and transfers of newborns are no longer combined into a single group; birth weight (rather than diagnosis) is used as the primary variable to differentiate categories of neonates; and duration of mechanical ventilation, presence of major problems, and surgery are used to define specific PM-DRGs. A total of 46 PM-DRGs have been developed to replace the 7 DRGs for neonates. Based on a sample of discharged patients from 13 children's hospitals, the overall variance reduction in duration of stay for neonates using PM-DRGs was 38.7% compared with 20.4% for DRGs. Variance reduction for PM-DRGs was 45.9% compared with 16.3% for DRGs when operating cost per case was used instead of duration of stay. After removing outliers at 150 days, the duration of stay variance reduction was 53.3% vs 23.6%, respectively, and the operating cost variance reduction was 58.8% vs 17.8%, respectively.

Refining DRGs. The example of children's diagnosis-related groups.

Congress exempted certain types of specialty hospitals from the Medicare Prospective Payment System because it was recognized that Diagnosis-Related Groups (DRGs) may not properly define the case mixes in such institutions. This study is part of a larger investigation into case mix and payment in children's hospitals, one category of exempted institutions. A national sample of approximately 500,000 cases was developed with intentional oversampling of children's and university hospitals to allow detection of specialized types of cases. Five case-mix classification schemes--DRGs, Disease Staging, Patient Management Categories, Severity of Illness Score, and Pediatric Diagnosis System groups--were applied to the data set, and data items not included in the Uniform Hospital Discharge Data Set were collected. A set of Children's Diagnosis-Related Groups (CDRGs), based on modification of the current DRG system, resulted from the study. When CDRGs were applied to an independent sample of children's hospital data, length of stay variance was reduced by 47.6% compared with 32.3% for DRGs (length of stay outliers removed). These results suggest that incremental approaches to DRG refinement in other clinical areas where current definitions are inadequate may be better than rejecting the large amounts of statistical and clinical analyses existing in the DRG system. Similar methods can be used to correct problems brought on by changes in medical practice.