Predict Resource Utilization Research Articles

The validity of the canadian triage and acuity scale in predicting resource utilization and the need for immediate life-saving interventions in elderly emergency department patients

BackgroundWe evaluated the validity of the Canadian Triage and Acuity Scale (CTAS) in elderly emergency department (ED) patients. In particular, we examined the sensitivity and specificity of the CTAS for identifying elderly patients who received an immediate life-saving intervention in the ED.MethodsWe reviewed the medical records of consecutive patients who were 65 years of age or older and presented to a single academic ED within a three-month period. The CTAS triage scores were compared to actual patient course, including disposition, discharge outcome and resource utilization. We calculated the sensitivity and specificity of the CTAS triage for identifying patients who received an immediate intervention.ResultsOf the 1903 consecutive patients who were ≥ 65 years of age, 113 (5.9%) had a CTAS level of 1, 174 (9.1%) had a CTAS level of 2, 1154 (60.6%) had a CTAS level of 3, 347 (18.2%) had a CTAS level of 4, and 115 (6.0%) had a CTAS level of 5. As a patient's triage score increased, the severity (such as mortality and intensive care unit admission) and resource utilization increased significantly. Ninety-four of the patients received a life-saving intervention within an hour following their arrival to the ED. The CTAS scores for these patients were 1, 2 and 3 for 46, 46 and 2 patients, respectively. The sensitivity and specificity of a CTAS score of ≤ 2 for identifying patients for receiving an immediate intervention were 97.9% and 89.2%, respectively.ConclusionsThe CTAS is a triage tool with high validity for elderly patients, and it is an especially useful tool for categorizing severity and for recognizing elderly patients who require immediate life-saving intervention.

289 A Comparison Between the 4-level and 5-level Taiwan Pediatric Triage System in Predicting Resource Utilization in the Emergency Department

289 A Comparison Between the 4-level and 5-level Taiwan Pediatric Triage System in Predicting Resource Utilization in the Emergency Department

The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients

Objective(1) To determine chronic illness costs for large cohort of primary care patients, (2) to develop prospective model predicting total costs over one year, using demographic and clinical information including widely used comorbidity index. Study Design and SettingData including diagnostic, medication, and resource utilization were obtained for 5,861 patients from practice-based computer system over a 1-year period beginning December 1, 1993, for retrospective analysis. Hospital cost data were obtained from hospital cost accounting system. ResultsAverage annual per patient cost was $2,655. Older patients and those with Medicare or Medicaid had higher costs. Hospital costs were $1,558, accounting for 58.7% of total costs. In the predictive model, individuals with higher comorbidity incurred exponentially higher annual costs, from $4,317 with comorbidity score of two, to $5,986 with score of three, to $13,326 with scores greater than seven. To use an adapted comorbidity index to predict total yearly costs, four conditions should be added to the index: hypertension, depression, and use of warfarin with a weight of one, skin ulcers/cellulitis, a weight of two. ConclusionThe adapted comorbidity index can be used to predict resource utilization. Predictive models may help to identify targets for reducing high costs, by prospectively identifying those at high risk.

Cardiac magnetic resonance findings predict increased resource utilization in elective coronary artery bypass grafting

Morbidity following CABG (coronary artery bypass grafting) is difficult to predict and leads to increased healthcare costs. We hypothesized that pre-operative CMR (cardiac magnetic resonance) findings would predict resource utilization in elective CABG. Over a 12-month period, patients requiring elective CABG were invited to undergo CMR 1 day prior to CABG. Gadolinium-enhanced CMR was performed using a trueFISP inversion recovery sequence on a 1.5 tesla scanner (Sonata; Siemens). Clinical data were collected prospectively. Admission costs were quantified based on standardized actual cost/day. Admission cost greater than the median was defined as 'increased'. Of 458 elective CABG cases, 45 (10%) underwent pre-operative CMR. Pre-operative characteristics [mean (S.D.) age, 64 (9) years, mortality (1%) and median (interquartile range) admission duration, 7 (6-8) days] were similar in patients who did or did not undergo CMR. In the patients undergoing CMR, eight (18%) and 11 (24%) patients had reduced LV (left ventricular) systolic function by CMR [LVEF (LV ejection fraction) <55%] and echocardiography respectively. LE (late enhancement) with gadolinium was detected in 17 (38%) patients. The average cost/day was $2723. The median (interquartile range) admission cost was $19059 ($10891-157917). CMR LVEF {OR (odds ratio), 0.93 [95% CI (confidence interval), 0.87-0.99]; P=0.03} and SV (stroke volume) index [OR 1.07 (95% CI, 1.00-1.14); P=0.02] predicted increased admission cost. CMR LVEF (P=0.08) and EuroScore tended to predict actual admission cost (P=0.09), but SV by CMR (P=0.16) and LV function by echocardiography (P=0.95) did not. In conclusion, in this exploratory investigation, pre-operative CMR findings predicted admission duration and increased admission cost in elective CABG surgery. The cost-effectiveness of CMR in risk stratification in elective CABG surgery merits prospective assessment.

A revised method to assess intensive care unit clinical performance and resource utilization*

In 1994, Rapoport et al. published a two-dimensional graphical tool for benchmarking intensive care units (ICUs) using a Mortality Probability Model (MPM0-II) to assess clinical performance and a Weighted Hospital Days scale (WHD-94) to assess resource utilization. MPM0-II and WHD-94 do not calibrate on contemporary data, giving users of the graph an inflated assessment of their ICU's performance. MPM0-II was recently updated (MPM0-III) but not the model for predicting resource utilization. The objective was to develop a new WHD model and revised Rapoport-Teres graph. Multicenter cohort study. One hundred thirty-five ICUs in 98 hospitals participating in Project IMPACT. Patients were 124,855 MPM0-II eligible Project IMPACT patients treated between March 2001 and June 2004. None. WHD was redefined as 4 units for the first day of each ICU stay, 2.5 units for each additional ICU day, and 1 unit for each non-ICU day after the first ICU discharge. Stepwise linear regression was used to construct a model to predict ICU-specific log average WHD from 39 candidate variables available in Project IMPACT. The updated WHD model has four independent variables: percent of patients dying in the hospital, percent of unscheduled surgical patients, percent of patients on mechanical ventilation within 1 hr of ICU admission, and percent discharged from the ICU to an external post-acute care facility. The first three variables increase average WHD and the last decreases it. The new model has good performance (R = 0.47) and, when combined with MPM0-II, provides a well-calibrated Rapoport-Teres graph. A new WHD model has been derived from a large, contemporary critical care database and, when used with MPM0-III, updates a popular method for benchmarking ICUs. Project IMPACT participants will likely perceive a decline in their ICU performance coordinates due to the recalibrated graph and should instead focus on their unit's performance relative to their peers.

Admission Physiology Criteria After Injury on the Battlefield Predict Medical Resource Utilization and Patient Mortality

Medical resources and resource allocation including operating room and blood utilization are of prime importance in the modern combat environment. We hypothesized that easily measurable admission physiologic criteria and injury site as well as injury severity calculated after diagnostic evaluation or surgical intervention, would be strongly correlated with resource utilization and in theater mortality outcomes. We retrospectively reviewed the Joint Theater Trauma Registry for all battlefield casualties presenting to surgical component facilities during Operation Iraqi Freedom from January to July 2004. Data were collected from the composite population of 1,127 battlefield casualty patients with respect to demographics, mechanism, presentation physiology (blood pressure, heart rate, temperature), base deficit, admission hematocrit, Glasgow Coma Score (GCS), Injury Severity Score (ISS), operating room utilization, blood transfusion, and mortality. Univariate and multivariate analyses were conducted to determine the degree to which admission physiology and injury severity correlated with blood utilization, necessity for operation, and acute mortality. Univariate analysis demonstrated a significant (p < 0.05) association between hypothermia (T < 34 degrees C) and the subsequent requirement for operation and mortality. In addition, the outcome variable total blood product utilization was significantly correlated with base deficit (r = 0.61), admission hematocrit (r = 0.51), temperature (r = 0.47), and ISS (r = 0.54). Using multiple logistic regression techniques, blood pressure, GCS, and ISS together demonstrated a significant association (p < 0.05) with mortality (area under ROC curve = 95%). Multiple linear regression established that blood pressure, heart rate, temperature, hematocrit, and ISS had a collective significant effect (p < 0.05) on total blood product utilization explaining 67% of the variance in this outcome variable. Admission physiology and injury characteristics demonstrate a strong capacity to predict resource utilization in the contemporary battlefield environment. In the future, such predictive yield could potentially have significant implications for triage and medical logistics in the resource constrained environment of war and potentially in mass casualty and disaster incidents in the civilian trauma setting which will likely have mechanistic similarity with war related injury.

AB17-5: Model for predicting resource utilization in clinics treating atrial fibrillation

AB17-5: Model for predicting resource utilization in clinics treating atrial fibrillation

Comparison of the Emergency Severity Index (ESI) and the Taiwan Triage System in Predicting Resource Utilization

The importance of accurate triage in Taiwan is becoming more apparent with the increasing number of emergency department (ED) patients, and resources for the National Health Insurance becoming constrained. This study compared the ability of the Taiwan triage system (TTS) and the standardized 5-level Emergency Severity Index (ESI) triage system to predict ED resource utilization. Patients arriving at the ED were triaged by both TTS and by using a two-page checklist of ESI criteria during the 3-month study period. The ESI triage level was calculated independently to avoid bias. Disease category (trauma vs. nontrauma), length of stay (LOS) and hospitalization data were evaluated. A total of 3172 patients with both ESI and TWN evaluation were included. The distributions of ESI ratings within TTS level 1 were: ESI 1, 21.1%; ESI 2, 68.1%; ESI 3, 7.4%; ESI 4, 3.4%; ESI 5, 0%. For TTS level 3, they were: ESI 1, 0.1%; ESI 2, 26.2%; ESI 3, 39.5%; ESI 4, 27.5%; ESI 5, 6.8%. Hospitalization rates were 74.5%, 40.9% and 22.2% in TTS levels 1, 2 and 3, respectively; and were 96.2%, 47.0%, 30.9%, 6.7%and 6.6% in ESI levels 1, 2, 3, 4 and 5, respectively. TTS triaged more trauma patients as life-threatening/emergent condition than nontrauma patients (68.8% vs. 48.4%, p < 0.001). Triage by ESI, however, showed no significant difference in the percentage of trauma and nontrauma patients with highly acute conditions (44.2% vs. 46.6%, p = 0.230). Patients with ESI level 4 or 5 have significantly shorter ED LOS than those with ESI level 3. ESI produces more accurate discriminating patient acuity, ED LOS and hospitalization rate than TTS. Adopting a standardized 5-level triage tool might improve resource utilization planning of ED practice.

Rehabilitation service utilization and determinants among people with an intellectual disability: Preliminary findings in Taiwan

Improving rehabilitation services for people with intellectual disabilities (ID) remains an ongoing challenge in the public health system. The purpose of this article was to investigate the types of rehabilitation services used by people with ID and determine what factors predict resource utilization in Taiwan. Samples of 957 people with ID were recruited from the Taiwan National Disability Register in a cross-sectional study in 2001. The findings indicated that 24.5% of individuals with ID had received rehabilitation services in the past 7 months. The main types of services used were speech and communication therapy (50%), psychotherapy (32.1%), occupational therapy (30.3%) and physiotherapy (25.2%). Stepwise logistic regression was carried out for the utilization of rehabilitation services (yes/no). The model revealed that the following factors: (i) Major Illness Card holder, (ii) time spent in medical care, (iii) having a family physician, (iv) having illnesses, (v) age of ID individual, and (vi) gender of the main carer, were all significantly associated with the utilization of rehabilitation services. We should reorient the healthcare system to respond adequately to the health needs of rehabilitation service users and its determinants, and further research should focus on the effectiveness and efficiency of rehabilitation for people with ID in Taiwan.

An incremental EM-based learning approach for on-line prediction of hospital resource utilization

Inpatient length of stay (LOS) is an important measure of hospital activity, health care resource consumption, and patient acuity. This research work aims at developing an incremental expectation maximization (EM) based learning approach on mixture of experts (ME) system for on-line prediction of LOS. The use of a batch-mode learning process in most existing artificial neural networks to predict LOS is unrealistic, as the data become available over time and their pattern change dynamically. In contrast, an on-line process is capable of providing an output whenever a new datum becomes available. This on-the-spot information is therefore more useful and practical for making decisions, especially when one deals with a tremendous amount of data. The proposed approach is illustrated using a real example of gastroenteritis LOS data. The data set was extracted from a retrospective cohort study on all infants born in 1995-1997 and their subsequent admissions for gastroenteritis. The total number of admissions in this data set was n = 692. Linked hospitalization records of the cohort were retrieved retrospectively to derive the outcome measure, patient demographics, and associated co-morbidities information. A comparative study of the incremental learning and the batch-mode learning algorithms is considered. The performances of the learning algorithms are compared based on the mean absolute difference (MAD) between the predictions and the actual LOS, and the proportion of predictions with MAD < or = 1 day (Prop(MAD < or = 1)). The significance of the comparison is assessed through a regression analysis. The incremental learning algorithm provides better on-line prediction of LOS when the system has gained sufficient training from more examples (MAD = 1.77 days and Prop(MAD < or = 1) = 54.3%), compared to that using the batch-mode learning. The regression analysis indicates a significant decrease of MAD (p-value = 0.063) and a significant (p-value = 0.044) increase of Prop(MAD < or = 1) with the incremental learning algorithm. The incremental learning feature and the self-adaptive model-selection ability of the ME network enhance its effective adaptation to non-stationary LOS data. It is demonstrated that the incremental learning algorithm outperforms the batch-mode algorithm in the on-line prediction of LOS.

INVITED COMMENTARY

The article by Nilsson and coworkers claims that the EuroSCORE (a popular mortality risk scoring system used extensively in Europe) is also useful for predicting resource utilization in cardiac surgical patients. This work deserves comment in three areas.

Predictive validity comparison of two five-level triage acuity scales

Study objectives: Triage is an initial assessment of acuity and sorting process applied to patients on presentation to an emergency department (ED). It is believed that patients triaged as having a higher level of acuity require a greater number of ED resources during their ED visit. This study seeks to prospectively measure and compare the ability of the 2 most commonly used, 5-level triage tools in North America, the Emergency Severity Index and the Canadian Triage and Acuity Scale, to predict the ED resources utilized by each patient during his or her visit. Methods: In this multicenter, prospective, observational cohort study of a population-based random sample of patients triaged at 2 EDs during a 4-month period, all patients presenting to the ED within the study periods were simultaneously and independently triaged by 2 nurses each using 1 of the 2 triage tools and blind to each other's assessment. The patients were followed up during their stay in the ED by a third research nurse blind to the triage assessments who recorded the resources utilized for that patient using a standardized data collection form until disposition. Results: Correlational analyses of the 486 cases revealed no significant difference between the ability of the Emergency Severity Index ( r =0.524) and the Canadian Triage and Acuity Scale ( r =0.485) to predict resource utilization. Conclusion: Triage scales, although valid and reliable measures of acuity, are only moderately accurate measures of ED resource utilization.

Predicting Resource Utilization by Children with Serious Emotional Disturbance and Their Families

Our study represents a rural case management agency's use of its data to aid in its transition from being funded by federal grant funds to self-sustainment. We wanted to learn if some characteristics of those children served could predict the resources needed to serve them. We analyzed data on 90 children and adolescents living in a rural community. Using multiple regression analysis we found that CAFAS total and CBCL total problem scores predicted Medicaid reimbursements but not flexible funds spending or case management hours. History of psychiatric hospitalization, however, predicted flexible funds spending and case management hours. Using one-way ANOVA and post hoc analysis we found that the staff members' assessments of each child and family's level of service needs were reflected in differences in mean CAFAS total scores, Medicaid reimbursements and case management hours, but not CBCL total problem scores or flexible funds spending.

A decision support tool for allocating hospital bed resources and determining required acuity of care

Limitations in health care funding require physicians and hospitals to find effective ways to utilize resources. Neural networks provide a method for predicting resource utilization of costly resources used for prolonged periods of time. Injury severity knowledge is used to determine the acuity of care required for each patient and length of stay is used to determine duration of inpatient hospitalization. Neural networks perform well on these medical domain problems, predicting total length of stay within 3 days for pediatric trauma (population mean and S.D. 4.37±45.12) and within 4 days for acute pancreatitis patients (7.75±79.19).

Association between heart rate variability recorded on postoperative day 1 and length of stay in abdominal aortic surgery patients.

To determine whether heart rate variability (HRV) measured in the surgical intensive care unit (ICU) on the first postoperative day predicts clinical outcome in patients undergoing abdominal aortic surgery. Prospective study. Eighteen-bed surgical ICU of a 1,442-bed university hospital. One hundred and six patients admitted to the ICU after abdominal aortic surgery. Twenty-four-hour Holter recordings were analyzed for standard time and frequency domain indices and one nonlinear index (slope) of HRV. Clinical and demographic data were collected from medical records. Patients were dichotomized into short (< or = 7 days) and long (> 7 days) length of stay (LOS) by median split. Patients with long LOS had increased heart rates and decreased short- and intermediate-term HRV but no difference in overall HRV, which primarily reflects circadian rhythm. Independent predictors of LOS were increased age, insulin-dependent diabetes, and decreased HRV. Increased heart rates and decreased intermediate-term HRV indices measured on postoperative day 1 were independent predictors of complicated recovery. The strongest HRV predictors of outcome were natural logarithm very-low-frequency power measured over 24 hrs and during the daytime. Results support the potential use of HRV for the prediction of postsurgical resource utilization.

Need for cardiac output monitoring in elderly trauma patients determined by base deficit

Elderly patients tolerate trauma less well than younger patients. Early monitoring of the cardiac output to improve peripheral perfusion can diminish mortality and complications in elderly patients. Base deficit can be used as a marker of significant injury and to predict resource utilization and mortality. The purpose of this study is to determine which patients will require intervention.

Correlation of Risk Adjustment Measures Based on Diagnoses and Patient Self-Reported Health Status

Case-mix adjustments have traditionally used diagnosis-based models such as Diagnostic Cost Groups (DCGs). The recent development and availability of reliable and valid patient self-reported health status measures such as the Veterans SF-36 (Short Form Health Survey) may be useful in complementing existing diagnostic information in describing patients' health status for purposes of risk adjustment. However, the correlation between these two approaches has not been explored. We collected SF-36 data from 31,419 veterans nationwide based on a national probability sample of veterans receiving ambulatory care to assess the physical (PCS) and mental (MCS) component of patient self-reported health status. In addition, we used inpatient and outpatient diagnoses from one year (1/1/97 to 1/1/98) to calculate DCG relative risk scores, with the 1991 Medicare beneficiary population as the benchmark. We found that higher DCG relative risk scores were associated with worse PCS (r=−0.223, p<0.05) and MCS (r=−0.174, p<0.05) scores. Further examination of the distribution of MCS categories (MCS≤40) across the five psychiatric hierarchical condition categories (HCCs) in the DCG/HCC model showed a small association between MCS category and disease severity level. These results suggest that risk adjustment approaches based on patient self-reported health status and diagnoses convey different case-mix information, specifically for patients with psychiatric conditions. These two approaches can be used as the basis for the development of a more comprehensive risk adjustment model which incorporates both the providers' and the patients' perspectives in predicting resource utilization.

Orthopedic Subacute Rehabilitation—Predictors of Functional Outcome and Resource Utilization

The purpose of this study was to determine what factors predict resource utilization and outcomes in patients recovering from an acute care orthopedic treatment episode in facilities offering subacute rehabilitation. We obtained data from a subacute orthopedic rehabilitation program offered at a continuing care facility. We obtained administrative and follow-up data for a case series of 779 patients undergoing subacute rehabilitation following an acute care hospital stay for an orthopedic condition. We administered the Functional Independence Measure (FIM) at admission, discharge, and 12 weeks postsubacute discharge. We estimated resource utilization through the calculation of subacute length of stay. A regression model explaining 30% of the variance in subacute length of stay (SLOS) contained admission motor FIM score, acute care length of stay (LOS), age, and lower leg/ankle fracture. A regression model explaining 85% of the variance in motor functional gain contained admission motor FIM score, lower/leg ankle fracture, and knee replacement. These results indicate that admission functional status is the key predictor of both SLOS and functional change. Prediction of 12-week functional change was more successful than prediction of SLOS.

Prediction of resource utilization and case cost for acute nonvariceal upper gastrointestinal hemorrhage at a canadian community hospital

OBJECTIVE: Upper gastrointestinal hemorrhage (UGIH) is common, and thus imposes a substantial burden on health care resources. We describe resource utilization and cost for management of acute nonvariceal UGIH, and studied their variation among population subgroups. METHODS: Resource utilization and direct medical case costs were extracted for consecutive admissions for nonvariceal UGIH at a large community hospital in southern Ontario through chart review and adaptation of an administrative case cost database. Univariate and multiple regression models were then developed to identify independent demographic predictors of case cost and length of stay. RESULTS: Among 116 eligible admissions the average length of stay and case cost were 4.26 days and Can$2690, respectively (Can$1 = US$0.70). Both cost and length of stay demonstrated significant univariate relationships with age, comorbid illness, prior peptic ulcer disease (PUD), and prior UGIH. Age and prior PUD persisted as independent predictors in multiple regression models. An inverse transformation of total case cost allowed these variables to explain 26% of the total variance. CONCLUSIONS: Resource utilization for management of acute nonvariceal UGIH at a Canadian community hospital varies substantially among population subgroups, but correlates independently with age and prior ulcer history. Careful attention must be paid to practice environments and demographic profiles before economic models of strategies to prevent or treat UGIH are applied to specific subpopulations.

Prediction of resource utilization and case cost for acute nonvariceal upper gastrointestinal hemorrhage at a Canadian community hospital.

Upper gastrointestinal hemorrhage (UGIH) is common, and thus imposes a substantial burden on health care resources. We describe resource utilization and cost for management of acute nonvariceal UGIH, and studied their variation among population subgroups. Resource utilization and direct medical case costs were extracted for consecutive admissions for nonvariceal UGIH at a large community hospital in southern Ontario through chart review and adaptation of an administrative case cost database. Univariate and multiple regression models were then developed to identify independent demographic predictors of case cost and length of stay. Among 116 eligible admissions the average length of stay and case cost were 4.26 days and Can$2690, respectively (Can$1 = US$0.70). Both cost and length of stay demonstrated significant univariate relationships with age, comorbid illness, prior peptic ulcer disease (PUD), and prior UGIH. Age and prior PUD persisted as independent predictors in multiple regression models. An inverse transformation of total case cost allowed these variables to explain 26% of the total variance. Resource utilization for management of acute nonvariceal UGIH at a Canadian community hospital varies substantially among population subgroups, but correlates independently with age and prior ulcer history. Careful attention must be paid to practice environments and demographic profiles before economic models of strategies to prevent or treat UGIH are applied to specific subpopulations.