Predicting length of stay in an acute psychiatric hospital

OBJECTIVE: Multivariate statistical methods were used to identify patient-related variables that predicted length of stay in a single psychiatric facility. The study investigated whether these variables remained stable over time and could be used to provide individual physicians with data on length of stay adjusted for differences in clinical caseloads and to detect trends in the physicians' practice patterns. METHODS: Data on all patients discharged over two six-month periods were collected at an acute psychiatric inpatient facility. Stepwise multiple regression analyses were conducted on the two datasets. RESULTS: The results from both analyses revealed that five variables significantly predicted length of stay and were stable over time. They were a primary diagnosis of schizophrenia, the number of previous admissions, a primary diagnosis of a mood disorder, age, and a secondary diagnosis of an alcohol- or other drug-related disorder. For some physicians, the mean length of stay of their patients differed significantly from the length predicted by the regression model--generally, it was shorter. CONCLUSIONS: The results demonstrate that patient-related predictors of length of stay in a single psychiatric hospital can be identified using relatively simple statistical procedures and can be consistent across a large dataset and over time.     

Relation of direct hospitalization costs with length of stay

OBJECTIVE: Our main objective is to analyse to the relationship between the direct cost of a hospitalary discharge and the length of stay controlling for other care variables. METHODS: Analysis of the direct costs of pharmacy, laboratory, pathology and radiology tests of the 21,883 discharged patients in two Barcelona hospitals during 1993, in relationship to care variables contained in the basic minimum data set for discharged patient (BMDSDP). Using both hospital information systems in which are detailed the complete activity carried out and the assignment of unitary costs by means of different methods adapted to the available information, the direct cost is built up for patient and it is assembled by DRG. With the direct cost information and the care variables of the BMDSDP, a simple linear regression (least squared method) is carried out. RESULTS: The average direct cost is up to 31,533 pesetas. The regression by least squared method explains 70% of the variance (R2) and the variables with higher explanatory power are the length of stay and the relative weight of average DRG direct costs, that acts like variable of adjustment. CONCLUSIONS: The variability of the direct cost is explained principally by the length of stay. In addition, the length of stay is also very important on explaining the internal variability of DRG direct cost.     

Combined MMPI and demographic data in predicting length of neuropsychiatric hospital stay.

Data available on 316 psychiatric patients, soon after admission to a hospital, were used in a multiple regression formula and in less formal statistical techniques to predict length of stay. Demographic and MMPI predictors made possible early identification of long stay patients at greater than base rate frequency. The asymptote of prediction was reached when 6 of 24 variables had been included, suggesting that further search in the area would be redundant. Prediction with the 5 prediction schema was stable on a cross-validation sample of 352. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Psychiatric comorbidity and length of stay in the general hospital. A critical review of outcome studies

The authors critically reviewed 26 international and American outcome studies that assessed the effect of psychiatric comorbidity on length of stay for medical/surgical inpatients. Three generations of American studies were defined according to the rigorousness of the methodology. Eighty-nine percent of all studies with sample sizes greater than 110 and 75% of the prospective, rigorously controlled, American studies found a significant association between psychiatric comorbidity and increased length of stay. The findings lead to the conclusion that impaired cognition associated with delirium and dementia, depressed mood, and other personality variables contributes to prolonged hospital stays and greater utilization of hospitals and other health resources after discharge. Recommendations for future research are suggested.     

Patients admitted to psychiatric hospital: Presenting problems and resolution at discharge.

Examined a cohort of admissions (N?=?280) to a 150-bed public psychiatric hospital to investigate homogeneity in patient problems and patient subtypes. The interrelationships of presenting problems, treatment provided, length of stay, resolution of presenting problems, and several other variables were also studied. Results indicated that many problems experienced by patients were not psychiatric symptoms. Most presenting problems were addressed but few resolved. The primary form of treatment was psychotropic medication. Many patients were discharged before treatment effectiveness was determined. Little association was found between diagnosis, patients' problems or patient subgroup, and psychotropic drugs prescribed. (French abstract) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effects of Medicare's prospective payment system on service use by depressed elderly inpatients

OBJECTIVE: To determine the effects of Medicare's prospective payment system (PPS) on hospital care, changes in length of stay and intensity of clinical services received by 2,746 depressed elderly patients in 297 acute care general medical hospitals were studied. METHODS: A pre-post design was used, and differences in sickness at admission were controlled for. Data on length of stay and use of specific clinical services were obtained from the medical record using a medical record abstraction form. Care provided on units exempt from PPS was compared with care provided in nonexempt units. RESULTS: After implementation of PPS, the average length of stay fell by up to three days within the different types of acute care settings studied, but this decline was partially offset by proportionately more admissions to psychiatric units, which had longer lengths of stay. Intensity of clinical services increased after PPS implementation, especially in nonexempt psychiatric units. CONCLUSION: Despite financial incentives for hospitals to reduce clinical services under PPS, its implementation was not associated with a marked decline in length of stay, when averaged across all treatment settings, and was associated with an increase in the intensity of many clinical services used by depressed elderly patients in general hospitals.     

Variations in the postoperative length of stay according to patient characteristics in a general surgery service

With the purpose of identifying factors that explain variations in length of postoperative stay in a general surgery unit, information from 666 clinical records of discharged alive patients was collected. Factors related to the surgical procedure, the severity of the disease and the medical team together with sociodemographic characteristics of the patients were analyzed. Statistical analysis was based on analysis of variance and multiple linear regression. Variables with statistically significant regression coefficients (p < 0.001) were in this order: severity of complications, type of intervention, need for reintervention and severity of diagnosis. Sixty-seven percent of length of stay variation was explained by the included variables. Results allow to suggest an evaluation procedure which takes into account patient and operation characteristics.     

Cue reactivity as a predictor of drinking among male alcoholics.

Social learning theories suggest that conditioned responses may increase the risk for relapse. Responses to alcohol use cues (cue reactivity) are associated with variables suggestive of risk but little research exists on the relationship of cue reactivity to treatment outcome. Alcoholic men admitted for detoxification to a treatment program (n?=?45) underwent a cue reactivity assessment protocol, and 91% received 3-mo follow-up interviews. Greater salivary reactivity predicted greater frequency of drinking during follow-up. Attentional factors added independent variance to the prediction of drinking outcome, with greater attention to stimulus or to response predicting less drinking. Cue reactivity did not predict length of hospital stay or latency to first drink. Results are discussed in the context of information processing, social learning theories, and clinical implications for relapse prevention. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Longitudinal analysis of utilization of inpatient treatment based on insurance data of mandatory health insurance

AIM: To investigate which of the routinely collected claims data from the German "Legal sickness funds" on hospital utilisation may be used, in addition to that prescribed by the legislator. DESIGN: We used claims data to study a cohort of sickness fund beneficiaries who were insured during the complete year 1992 (n = 81,309). Six utilisation parameters, using the number of cases and in hospital days overall as well as diseases specific (i.e. readmission rates, in-hospital days per person with [at least] one hospital stay) were calculated. RESULTS: There are 88 persons with (at least) one hospital stay, 116 hospital cases and a total of 1306 in-hospital days per 1000 insured persons in the study cohort. The average hospital days per person (14.8 days) are ca. 30% higher than the average length of stay (11.2 days). Hospital utilisation increases with age. Hospital stays associated with ICD-239 (neoplasms of unknown origin) resulted in a higher than average number of hospital days in total although the mean length of stay is not above the average. This is due to a high readmission rate. Hospital stays associated with elective surgical procedures have a high prevalence rate but a low readmission ratio and short length of stay. CONCLUSION: The parameters related to insured persons, cases and specifically personal parameters of hospital utilisation allow a detailed analysis of hospital care; different utilisation and user patterns can be investigated and possible determinants of utilisation can be identified. After technical transformation, routine data of the sickness funds can be used to obtain information relevant for health care planners as well as for quality management.     

Prevalence and follow-up of psychiatric diseases in internal medicine departments

While several surveys have shown that psychiatric disorders are common in non-psychiatric hospital departments, information on the course of psychiatric morbidity during the hospital stay is still lacking. Therefore, in two medical hospital departments in Austria we investigated the course of illness in 265 patients staying in the hospital for more than four weeks. Patients were interviewed using the Clinical Interview Schedule. At admission, prevalence of all psychiatric disorders was 46.0%. During the four weeks after admission, mentally well were discharged from hospital more frequently than mentally ill. Among those staying in the hospital after four weeks, prevalence was 51.4%. In the intervening period, the incidence (9.5%) was slightly lower than the rate of remission (12.4%). At both assessments, organic mental illness was the most frequent diagnosis followed by neurotic and adjustment disorders. Thus, in spite of the slightly higher rate of remission than incidence, the overall psychiatric prevalence increased during the first four weeks after admission because of the earlier discharges of the mentally well. In conclusion, our results show that point-prevalence surveys not regarding disease course give an incomplete impression of psychiatric morbidity in non-psychiatric hospital departments.     

Predictors of return to a psychiatric hospital

Examined social factors as measured by a community adjustment scale and other known variables associated with recidivism for their ability to predict rehospitalization of 108 former psychiatric inpatients. One-third of this sample were readmitted to a psychiatric hospital during the 6-mo period. Multiple linear regression analyses indicated that 2 of the subscales of the Self-Assessment Guide, the number of previous hospital admissions, and whether the individual had been in the hospital during the year prior to admission accounted for a sizable portion of variance. Results suggest that social factors are important determinants of recidivism. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Use of hospital services by AIDS patients with psychiatric illness

The purpose of this study was to assess the effect of psychiatric illness on length of stay and patterns of admission among AIDS patients hospitalized for medical illnesses. Medical records were abstracted for AIDS patients admitted to hospitals in Washington State from 1990 through 1992. Psychiatric comorbidity was defined by the presence of an International Classification of Disease-9 code reflecting psychiatric illness. Medical morbidity was addressed using CD4 count and AIDS-defining illnesses as markers of disease severity. Of 2834 admissions, 15% included one or more psychiatric diagnoses. Psychiatric illness (F 39.1; df 1,2830; p < 0.001) and discharge disposition (F 81.2; df 2,2830; p < 0.001) contributed significantly to the model, explaining increased length of stay (F 67.2; df 3,2830; p < 0.001). Future research needs to address the possible etiology of psychiatric comorbidity's contribution to length of stay and the effect on quality and cost of care.     

Electrodermal activity and social network as predictors of outcome of episodes in schizophrenia.

The predictive value of electrodermal activity and social network was examined among 48 consecutively admitted schizophrenic patients. The patients were followed from an initial admission, through hospital stay, discharge, follow-up (M?=?31 months), and possible relapse. Outcome variables were the length of stay in the hospital at the key episode and time to relapse, defined as a marked exacerbation or return of schizophrenic symptoms requiring inpatient or expansion of outpatient treatment. Multivariate analyses showed that a psychosocial variable, the availability of attachment, was the only independent predictor of length of stay in the hospital. Age at admission was a strong predictor of time to relapse. Age interacted with both outcome and electrodermal activity, and young electrodermal nonresponders were found to have the shortest time to relapse. At the l-year followup, a main relapse effect was found for patients with a low skin conductance level. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Research without billing data. Econometric estimation of patient-specific costs

OBJECTIVES: This article describes a method for computing the cost of care provided to individual patients in health care systems that do not routinely generate billing data, but gather information on patient utilization and total facility costs. METHODS: Aggregate data on cost and utilization were used to estimate how costs vary with characteristics of patients and facilities of the US Department of Veterans Affairs. A set of cost functions was estimated, taking advantage of the department-level organization of the data. Casemix measures were used to determine the costs of acute hospital and long-term care. RESULTS: Hospitalization for medical conditions cost an average of $5,642 per US Health Care Financing Administration diagnosis-related group weight; surgical hospitalizations cost $11,836. Nursing home care cost $197.33 per day, intermediate care cost $280.66 per day, psychiatric care cost $307.33 per day, and domiciliary care cost $111.84 per day. Outpatient visits cost an average of $90.36. These estimates include the cost of physician services. CONCLUSIONS: The econometric method presented here accounts for variation in resource use caused by casemix that is not reflected in length of stay and for the effects of medical education, research, facility size, and wage rates. Data on non-Veteran's Affairs hospital stays suggest that the method accounts for 40% of the variation in acute hospital care costs and is superior to cost estimates based on length of stay or diagnosis-related group weight alone.     

Sociometric status and adjustment of hospitalized disturbed children.

Investigated the relationship between the social status and adjustment of 38 7-16 yr. old psychiatric inpatients. Sociometric information was obtained by comparing the friendship choices of each S, and adjustment was measured by staff ratings and S's length of stay in the hospital. No relationship was found between sociometric status and Ss' age, IQ, diagnosis, or popularity with staff members. There was a tendency, however, for long-term Ss to have fewer friends than short-term Ss. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Work locus of control: the intrinsic factor behind empowerment and work excitement

BACKGROUND: Mortality and length of stay are frequently used as performance measures for hospitals. If they are valid measures, they should be reproducible from year to year with attributable variation rather than random variation. METHODS: We compared hospitals on 2 outcomes, mortality and length of stay, in pneumonia in Medicare patients. The database was from 20 Illinois hospitals with the largest number of discharges for diagnosis-related group 89 (pneumonia with complications/comorbidities) for the years 1989 through 1992. This comprised 16,249 claims for hospitalization in patients 65 years of age or older. RESULTS: The distributions showed trends toward lower mortality and shorter stays over the 4 years. Correlation of performance from year to year at each hospital for mortality was low with none of the calculated correlation coefficients significant at p < .05. Correlations for length of stay were higher (all coefficients significant at p < .01). For length of stay, the correlation between 1991 and 1992 was .766 (p < .00005, r2 = .587), showing that nearly 60% of differences (variance) were caused by differences in performance. In contrast, for mortality in 1991 and 1992, the correlation was .301 (p = .0986, r2 = .091), showing that less than 10% of differences (variance) between hospitals were caused by differences in performance. Similar results were obtained when the 20 hospitals were ranked and their rank correlations calculated. CONCLUSION: For pneumonia in Medicare patients, differences in length of stay between hospitals are caused by differences in performance, while differences in mortality are random.     

Impact of an all-inclusive diagnosis-related group payment system on inpatient utilization

OBJECTIVES: Although case-based payment is one of the main reimbursement mechanisms for hospitals, little is known about its effects in the general population. Prior studies have focused on Medicare or on all-payer systems in particular states. This study estimates the effect of a prospective payment system based on diagnosis-related groups (DRGs) nationwide in the Department of Veterans Affairs. METHODS: Multiple regression analysis was used to estimate the effect of Department of Veterans Affairs's diagnosis-related group system separately for 22 diagnoses. The dependent variables were length of stay, inpatient days per patient, and discharges per patient. Covariates included patient, hospital, and area characteristics. RESULTS: Department of Veterans Affairs's diagnosis-related group system reduced lengths of stay and inpatient days per patient. The largest impacts were for the psychiatric diagnoses and several surgical procedures. The magnitudes of the effects were generally moderate. Department of Veterans Affairs's case-based system had a negligible effect on discharges per patient. CONCLUSIONS: Per case reimbursement is a potentially useful tool for improving the efficiency of inpatient care for all types of diagnoses and age groups. The effect may be larger than estimated here because of institutional barriers and caps on financial impact.     

Beliefs about mental illness as a functon of psychiatric status and psychiatric hospitalization.

Beliefs about mental illness were assessed among psychiatric inpatients at a VA hospital, the mental health staff responsible for their treatment, and a group of medical and surgical (control) patients. Results indicated that: (a) Psychiatric and nonpsychiatric patients generally hold similar opinions regarding mental illness. Severely disturbed psychiatric patients, however, view mental illness in more moralistic terms than do "normals." (b) Psychiatric hospitalization is generally accompanied by a change in the patient's beliefs concerning mental illness, toward those held by the staff. (c) Psychiatric patients whose beliefs about mental illness are most strikingly influenced by the staff tend to respond most favorably to treatment, as measured by length of hospital stay and gains in self-esteem during the 1st month of treatment. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Statistical versus clinical prediction: Length of neuropsychiatric hospital stay.

Statistical and clinical prediction methods are compared in a "field" situation, predicting patient stay in a psychiatric hospital. 5 statistical methods from a previous report averaged 71.92% accuracy on the cross-validation sample. 12 clinicians predicting independently in the same hospital over the following 18-mo period achieved an average accuracy of 71.94% on 499 patient predictions. Summaries of the clinicians' reasons for their judgments show some differences between the most and least accurate judges. Neither professional background nor years of professional experience of clinicians was related to predictive accuracy. Actuarial and clinical prediction were not demonstrably different in accuracy. Practical and economic aspects of this prediction problem favor clinical prediction. (PsycINFO Database Record (c) 2010 APA, all rights reserved)