Validation of a combined comorbidity index

Prediction of hospital readmission for heart failure: development of a simple risk score based on administrative data

OBJECTIVES

The purpose of this study was to develop a convenient and inexpensive method for identifying an individual's risk for hospital readmission for congestive heart failure (CHF) using information derived exclusively from administrative data sources and available at the time of an index hospital discharge.

BACKGROUND

Rates of readmission are high after hospitalization for CHF. The significant determinants of rehospitalization are debated.

METHODS

Administrative information on all 1995 hospital discharges in New York State which were assigned International Classification of Diseases-9-Clinical Modification codes indicative of CHF in the principal diagnosis position were obtained. The following were compared among hospital survivors who did and did not experience readmission: demographics, comorbid illness, hospital type and location, processes of care, length of stay and hospital charges.

RESULTS

A total of 42,731 black or white patients were identified. The subgroup of 9,112 patients (21.3%) who were readmitted were distinguished by a greater proportion of blacks, a higher prevalence of Medicare and Medicaid insurance, more comorbid illnesses and the use of telemetry monitoring during their index hospitalization. Patients treated at rural hospitals, those discharged to skilled nursing facilities and those having echocardiograms or cardiac catheterization were less likely to be readmitted. Using multiple regression methods, a simple methodology was devised that segregated patients into low, intermediate and high risk for readmission.

CONCLUSIONS

Patient characteristics, hospital features, processes of care and clinical outcomes may be used to estimate the risk of hospital readmission for CHF. However, some of the variation in rehospitalization risk remains unexplained and may be the result of discretionary behavior by physicians and patients.

Chronic comorbidity and outcomes of hospital care: length of stay, mortality, and readmission at 30 and 365 days

This article evaluates the behavior of an adaptation of the Charlson Index (CHI) applied to administrative databases to measure the relationship between chronic comorbidity and the hospital care outcomes of length of stay (LOS), in-hospital mortality, and emergency readmissions at 30 and 365 days. These outcomes were analyzed in 106,673 hospitalization episodes whose records are registered in a minimum basic data set maintained by the public health authorities of the community of Valencia, Spain. The highest comorbidity measured by the CHI was associated with greater LOS and in-hospital mortality and increased readmission at 30 and 365 days. The rate of readmissions at 1 year dropped, however, in the group with the greatest comorbidity, probably owing to an increase in mortality after hospitalization. While comorbidity does appear to increase the risk of adverse outcomes in general and mortality and readmission specifically, the second outcome is only possible if the first has not occurred. For this reason, information and selection biases derived from administrative databases, or from the CHI itself, should be taken into account when using and interpreting the index.

A comparison of the Charlson comorbidity index derived from medical record data and administrative billing data

The objective of this article is to compare the Charlson comorbidity index derived from medical record data (Chart Index) with the same index derived from billing data (ICD-9 Index) to determine how well each predicted inpatient and 30-day mortality, length of stay, and complications among Medicare beneficiaries hospitalized for carotid endarterectomy. Economic and time constraints have increased the need for risk adjusters derived from administrative data, yet few studies have compared these measures with those derived from chart review. Using logistic regression, the Chart Index was found to be a significant predictor of inpatient mortality, 30-day mortality, length of stay, and complications, after controlling for age, gender, and neurologic and medical risk factors (P values = 0.004, 0.056, 0.0001, and 0.042, respectively). The ICD-9 Index approached significance as a predictor of the outcomes (P values = 0.092, 0.100, 0.093, and 0.080, respectively). The Chart Index was shown to be superior to the ICD-9 Index within this patient sample.

A new casemix adjustment index for hospital mortality among patients with congestive heart failure

OBJECTIVES

Comparative analysis of hospital outcomes requires reliable adjustment for casemix. Although congestive heart failure is one of the most common indications for hospitalization, congestive heart failure casemix adjustment has not been widely studied. The purposes of this study were (1) to describe and validate a new congestive heart failure-specific casemix adjustment index to predict in-hospital mortality and (2) to compare its performance to the Charlson comorbidity index.

METHODS

Data from all 4,608 admissions to the Massachusetts General Hospital from January 1990 to July 1996 with a principal ICD-9-CM discharge diagnosis of congestive heart failure were evaluated. Massachusetts General Hospital patients were randomly divided in a derivation and a validation set. By logistic regression, odds ratios for in-hospital death were computed and weights were assigned to construct a new predictive index in the derivation set. The performance of the index was tested in an internal Massachusetts General Hospital validation set and in a non-Massachusetts General Hospital external validation set incorporating data from all 1995 New York state hospital discharges with a primary discharge diagnosis of congestive heart failure.

RESULTS

Overall in-hospital mortality was 6.4%. Based on the new index, patients were assigned to six categories with incrementally increasing hospital mortality rates ranging from 0.5% to 31%. By logistic regression, "c" statistics of the congestive heart failure-specific index (0.83 and 0.78, derivation and validation set) were significantly superior to the Charlson index (0.66). Similar incrementally increasing hospital mortality rates were observed in the New York database with the congestive heart failure-specific index ("c" statistics 0.75).

CONCLUSION

In an administrative database, this congestive heart failure-specific index may be a more adequate casemix adjustment tool to predict hospital mortality in patients hospitalized for congestive heart failure.

Diagnostic groups and depressed mood as predictors of 22-month mortality in medical inpatients

OBJECTIVE

While depression has been found to predict mortality in acute myocardial infarction, results from many other groups of medical patients are inconclusive. It is, therefore, unclear whether depression also predicts mortality in the typical mixed patient populations treated on medical hospital wards and whether an increased risk can be identified by means of patients' self ratings of depression.

METHOD

The Hospital Anxiety and Depression scale was used as a routine screening tool in consecutive admissions to the general medical wards of a university hospital. The official survival data were obtained 22 months later. For all 454 patients who completed the screening questionnaire, complete survival data were available.

RESULTS

High depression scores significantly predicted mortality in univariate comparisons (odds ratio 3.2; 95% CI 1.9-5.5) and in multivariate Cox regression analyses controlling for demographic and medical baseline variables (multivariate odds ratio 1.9; 95% CI 1.2-3.1; p < .01). Other significant predictors in the multivariate model were having a principal diagnosis of hematological disease or cancer, and older age. Disability, as assessed by nurses' ratings, and gender were not related to mortality. Subgroup analyses showed that the effect of depression scores was greatest in cardiopulmonary patients, but there was also a uniform trend toward higher mortality in depressed patients with other diagnoses.

CONCLUSION

Depressed mood is an independent risk factor for all-cause mortality in medical inpatients. Identifying patients at risk does not require formal psychiatric diagnoses, but can be achieved by means of a short, routinely administered self-rating questionnaire.

Managed care for congestive heart failure: influence of payer status on process of care, resource utilization, and short-term outcomes

BACKGROUND

Although health maintenance organizations (HMO) are insuring an increasing number of Americans, there are concerns that cost-reduction strategies may limit access to medical care or jeopardize its quality. This study was conducted to examine the influence of insurance payer status on the process of care and resource utilization among patients hospitalized for congestive heart failure (CHF).

METHODS AND RESULTS

Administrative information on all 1995 New York State hospital discharges assigned ICD-9-CM codes indicative of CHF in the principal diagnosis position were obtained from the Statewide Planning and Research Cooperative System database. The following were compared among patients with HMO, indemnity, Medicaid fee-for-service, and Medicare fee-for-service insurance coverage: demographics, comorbid illness, process of care, length of stay, hospital charges, mortality rate, and CHF readmission rate. A total of 43,157 patients were identified (HMO, 1322; indemnity, 4350; Medicaid, 3878; Medicare, 33 607). Noninvasive procedures were used with similar frequency, whereas greater use of invasive techniques was observed among HMO and indemnity patients. After adjustment for patient characteristics and hospital type and location, HMO care was associated with shorter length of stay and lower hospital charges, the latter partially explained by fewer hospital days. Medicaid patients had the longest length of stay, greatest hospital charges, and highest CHF readmission rate. The adjusted risk of death during the index hospitalization did not vary among insurance groups.

CONCLUSIONS

Though insuring only a small proportion of New Yorkers hospitalized for CHF, managed care plans provide similar access to clinical services while generating fewer charges. Whether these observed differences in short-term outcomes derive from patient mix or quality of care is uncertain and deserves wider prospective study.

The correlation of comorbidity with function of the shoulder and health status of patients who have glenohumeral degenerative joint disease

We studied the effect of comorbidities on function of the shoulder and health status in a group of eighty-five consecutive patients who had glenohumeral degenerative joint disease of sufficient severity to meet one surgeon's criteria for the performance of shoulder arthroplasty. A questionnaire was used to identify the comorbidities, such as other diseases, social factors, or a work-related injury, for each patient. The number of functions on the Simple Shoulder Test that the patient could perform had a significant negative correlation with the number of comorbidities (r = -0.32, intercept = 4.6 per cent, slope = -0.6, and p = 0.0031). Each parameter on the Short Form-36 (except for physical role function) had a significant negative correlation with the number of comorbidities (p < 0.05). This negative relationship was strongest for general health perception (r = -0.42) and vitality (r = -0.35). We concluded that the number of comorbidities has a quantitative effect on function of the shoulder. In the evaluation of the functional status of patients and the effectiveness of treatment, the effects of comorbidity must be controlled. The results of the present study demonstrate that the scores on the Short Form-36 are quantitatively related to the number of comorbidities. The six parameters that are unrelated to function of the shoulder (physical function, social function, emotional role function, mental health, vitality, and general health perception) may provide a practical way to integrate the effects of all potential comorbidities on individual patients. Future clinical research will be strengthened by efforts to measure the impact of comorbidities and by strategies to control for their effects.

Influence of race and gender on care process, resource use, and hospital-based outcomes in congestive heart failure

Race and gender are important determinants of certain clinical outcomes in cardiovascular disease. To examine the influence of race and gender on care process, resource use, and hospital-based case outcomes for patients with congestive heart failure (CHF), we obtained administrative records on all 1995 New York State hospital discharges assigned ICD-9-CM codes indicative of this diagnosis. The following were compared among black and white women and men: demographics, comorbid illness, care processes, length of stay (LOS), hospital charges, mortality rate, and CHF readmission rate. We identified 45,894 patients (black women, 4,750; black men, 3,370; white women, 21,165; white men, 16,609). Blacks underwent noninvasive cardiac procedures more often than whites; procedure and specialty use rates were lower among women than among men. After adjusting for other patient characteristics and hospital type and location, we found race to be an important determinant of LOS (black, 10.4 days; white, 9.3 days; p = 0.0001), hospital charges (black, $13,711; white, $11,074; p = 0.0001), mortality (black-to-white odds ratio = 0.832; p = 0.003), and readmission (black-to-white odds ratio = 1.301; p = 0.0001). Gender was an important determinant of LOS (women, 9.8 days; men, 9.2 days; p = 0.0001), hospital charges (women, $11,690; men, $11,348; p = 0.02), and mortality (women-to-men odds ratio = 0.878; p = 0.0008). We conclude that race and gender influence care process and hospital-based case outcomes for patients with CHF.

Chronic disease as a barrier to breast and cervical cancer screening

OBJECTIVE

To assess whether chronic disease is a barrier to screening for breast and cervical cancer.

DESIGN

Structured medical record review of a retrospectively defined cohort.

SETTING

Two primary care clinics of one academic medical center.

PATIENTS

All eligible women at least 43 years of age seen during a 6-month period in each of the two study clinics (n = 1,764).

MEASUREMENTS AND MAIN RESULTS

Study outcomes were whether women had been screened: for mammogram, every 2 years for ages 50-74; for clinical breast examinations (CBEs), every year for all ages; and for Pap smears, every 3 years for ages under 65. An index of comorbidity, adapted from Charlson (0 for no disease, maximum index of 8 among our patients), and specific chronic diseases were the main independent variables. Demographics, clinic use, insurance, and clinical data were covariates. In the appropriate age groups for each test, 58% of women had a mammogram, 43% had a CBE, and 66% had a Pap smear. As comorbidity increased, screening rates decreased (p < .05 for linear trend). After adjustment, each unit increase in the comorbidity index corresponded to a 17% decrease in the likelihood of mammography (p = .005), 13% decrease in CBE (p = .006), and 20% decrease in Pap smears (p = .002). The rate of mammography in women with stable angina was only two fifths of that in women without.

CONCLUSIONS

Among women who sought outpatient care, screening rates decreased as comorbidity increased. Whether clinicians and patients are making appropriate decisions about screening is not known.

Predicting cognitive decline in patients with hypoxaemic COPD

The objective was to identify predictors of cognitive decline in patients with hypoxaemic COPD on continuous oxygen therapy. Eighty-four consecutive ambulatory hypoxaemic COPD patients in stable clinical conditions were prospectively studied over the course of 2 yr. Baseline multidimensional assessment included respiratory function tests, blood gas analysis, Mini Mental Status (MMS) test, Geriatric Depression Scale (GDS), Activities of Daily Living (ADLs) and Charlson's index of comorbidity. Reassessments were made 1 yr and 2 yr thereafter. Sequential changes in MMS, GDS and ADLs were assessed by Friedman's ANOVA by rank test. Forty patients completed the study (group A), while 44 died or were lost to follow-up (group B). Group B was characterized by more severe respiratory function impairment and worse performances on ADLs and GDS. In group A, MMS deteriorated from baseline to the 1 yr and 2 yr reassessments (27 +/- 2.9 vs. 25.8 +/- 4.1 and 25.4 +/- 4, P < 0.005), whereas GDS and ADLs did not change significantly; the 23 patients experiencing a decline of MMS had baseline lower percentage predicted FVC (52.3 +/- 17.1 vs. 66.9 +/- 13.4, P < 0.03) and FEV1 (27.2 +/- 8.6 vs. 44 +/- 26.8, P < 0.02) values and better affective status (GDS score: 11.9 +/- 7.7 vs. 16.5 +/- 5.6, P < 0.04). Two-year changes in MMS and in GDS scores were inversely correlated (Spearman's p = -0.32, P = 0.04). Cognitive decline is faster in the presence of severe bronchial obstruction and parallels the worsening of the affective status in COPD patients on oxygen therapy. The onset of depression rather than baseline depressive symptoms seems to be a risk factor for cognitive decline.

The clinical benefit of in-hospital observation in 'low-risk' pneumonia patients after conversion from parenteral to oral antimicrobial therapy

OBJECTIVE

To assess the benefit of in-hospital observation in "low-risk" patients with community-acquired pneumonia.

DESIGN

Retrospective review of data from a prospective study.

SETTING

Teaching community hospital.

PATIENTS

We studied 717 consecutive, adult patients admitted to the hospital for pneumonia.

MEASUREMENTS AND RESULTS

One hundred forty-five patients were classified at low-risk for complications using previously studied criteria; 144 (99%) charts were available for review. Two patients had "obvious reasons for continued hospitalization" on the day of antibiotic conversion and were excluded. One hundred two patients were observed, and 40 were not observed in-hospital after switch to oral antibiotics. No patient from either group required medical intervention within 24 h after hospital discharge. Five "observed" patients (5%, 95% confidence interval [CI], 2 to 11%) returned to the emergency department, three (3%; 95% CI, 0 to 9%) with respiratory complaints. Two (2%; 95% CI, 0 to 7%) "observed" patients were admitted to the hospital with recurrent pneumonia. One (3%; 95% CI, 0 to 13%) "not observed" patient returned to the emergency department with a nonrespiratory complaint and was not admitted. No patient from either group died within 30-day clinical follow-up. The length of stay for the "observed" and "not observed" groups was 98+/-33 h and 83+/-49 h, respectively. The difference in length of stay was 15 h (95% CI, 3 to 27).

CONCLUSIONS

In-hospital observation for low-risk patients admitted with community-acquired pneumonia after switch from parenteral to oral antibiotics is of limited benefit, and elimination of this practice could potentially reduce length of stay by almost 1 day per patient. This could translate into a cost savings of $57,200 for the 22-month study period. These results require prospective validation in a larger study.

Do disease specific characteristics add to the explanation of mobility limitations in patients with different chronic diseases? A study in The Netherlands

STUDY OBJECTIVES

To determine whether disease specific characteristics, reflecting clinical disease severity, add to the explanation of mobility limitations in patients with specific chronic diseases.

DESIGN AND SETTING

Cross sectional study of survey data from community dwelling elderly people, aged 55-85 years, in the Netherlands.

PARTICIPANTS AND METHODS

The additional explanation of mobility limitations by disease specific characteristics was examined by logistic regression analyses on data from 2830 community dwelling elderly people.

MAIN RESULTS

In the total sample, chronic non-specific lung disease, cardiac disease, peripheral atherosclerosis, diabetes mellitus, stroke, arthritis and cancer (the index diseases), were all independently associated with mobility limitations. Adjusted for age, sex, comorbidity, and medical treatment disease specific characteristics that explain the association between disease and mobility mostly reflect decreased endurance capacity (shortness of breath and disturbed night rest in chronic non-specific lung disease, angina pectoris and congestive heart failure in cardiac disease), or are directly related to mobility function (stiffness and lower body complaints in arthritis). For atherosclerosis and diabetes mellitus, disease specific characteristics did not add to the explanation of mobility limitations.

CONCLUSIONS

The results provide evidence that, to obtain more detailed information about the differential impact of chronic diseases on mobility, disease specific characteristics are important to take into account.

Case-mix adjustment using administrative databases: a paradigm to guide future research

One of the most persistent problems in the field of quality assessment remains how to remove the confounding effect of different institutions providing care to patients with dissimilar severity of illness and case complexity. The authors review the literature to determine whether risk adjustment systems based on administrative data are inherently inferior to systems that depend on primary data collection and conclude that they are not. In light of the potential competence of risk adjustment systems based on administrative data, the authors identify those systems that are best supported by theory and evidence. Data elements that have been found most explanatory of medical outcomes are also identified. On the basis of an evaluation of the performance of various risk adjustment approaches, the authors propose a paradigm that could serve to unify and direct future studies.

Survival estimates for patients with abnormal swallowing studies

OBJECTIVE

To better understand the life expectancy of patients who have an abnormal videofluoroscopic swallowing study.

DESIGN

Retrospective cohort study. The common starting point was the time of the severely abnormal swallowing study. Hospital charts were reviewed for clinical variables of potential prognostic significance by reviewers blinded to the outcome of interest, survival time.

SETTING

A university-affiliated, community teaching hospital.

PATIENTS

One hundred forty-nine hospitalized patients who were deemed nonoral feeders based on their swallowing study. Patients excluded were those with head, neck, or esophageal cancer, or those undergoing a thoracotomy procedure.

MEASUREMENTS AND MAIN RESULTS

Clinical and demographic variables and time until death or censoring were measured. Overall 1-year mortality was 62%. Multivariable Cox proportional hazards analyses identified four variables that independently predicted death: advanced age, reduced serum albumin concentration, disorientation to person, and higher Charlson comorbidity score. Eighty patients (54%) subsequently underwent placement of a percutaneous endoscopic gastrostomy (PEG) tube after their swallowing study.

CONCLUSIONS

Mortality is high in patients with severely abnormal swallowing studies. Common clinical variables can be used to identify groups of patients with particularly poor prognoses. This information may help guide discussions regarding possible PEG placement.

The effect of aging on the utilization of chemotherapy for metastatic breast cancer: a population-based study

Older women (i.e., > or = 65 years of age) receive less adjuvant chemotherapy than younger women, in part because chemotherapy has been less effective in postmenopausal than premenopausal women in clinical trials. Metastatic breast cancer, however, does not respond differently to chemotherapy by age. Therefore, to evaluate further the effect of age on chemotherapy utilization, we conducted a population-based study of the treatment of metastatic breast cancer. Patients (n = 132) were identified by cross-tabulating death certificates from 1984 to 1991 with breast cancer cases in the Washington County Cancer Registry. Treatment information was obtained from the Tumor Registry of the Washington Country Hospital and Hospital medical records. Forty patients (74%) < 65 years old received chemotherapy compared to 11 (42%) 65-74 and 6 (12%) > or = 75 (p < 0.0001). Adjusting for other medical conditions and whether or not the patient saw a medical oncologist, there was still a significant effect of age in patients > or = 75 (p < 0.001) and a trend (p = 0.17) for patients 65-74. The different patterns of chemotherapy utilization were not associated with survival differences. Radiation therapy was also utilized significantly less frequently in older than younger patients, but the age effect was not as pronounced as with chemotherapy. There was no age effect on the utilization of hormonal therapy. Less frequent utilization of chemotherapy in older patients is probably caused by a combination of patient and physician factors and may result in less effective palliation for older patients.

Patterns of comorbidity and mortality risk in blacks and whites

Comorbidity, the co-existence of multiple chronic conditions in a single individual, has been shown to modify the prognosis of disease states. To estimate disease burdens within and among racial subpopulations of the United States, we examined cross-sectional patterns of comorbidity and their impact on survival using data from the NHANES-1 Epidemiologic Follow-up Study (NHEFS). We considered the occurrence of four cardiovascular conditions: stroke, coronary heart disease, hypertension and diabetes. We summarize the joint occurrence of these four conditions using these different methodologies: the number of conditions occurring in each individual and two summaries that weight the conditions according to their prognostic significance. Using all three methodologies, we found an excess burden of chronic disease in black women as compared with white women. Black men had an excess burden compared to white men for the first two methodologies. However, when we model the relationship of the joint occurrence of the conditions to subsequent mortality, black men and white men are seen to have a similar burden. This similarity of black and white men is due to an interaction between race and prevalent stroke in men that we hypothesize may be due to the small number of black men available for study. Given the apparent conditioning effect of co-existing diseases, it is evident that estimation of disease burdens among groups that differ in terms of health status, in particular among U.S. blacks and whites, requires accounting for the occurrence of multiple chronic diseases. Using either the number of conditions or the prognosis weighted summary, we demonstrated a higher burden of the conditions considered in blacks that in whites in a sample of the U.S. population.