Practical considerations on the use of the charlson comorbidity index with administrative data bases

doi:10.1016/S0895-4356(96)00271-5

Journal of Clinical Epidemiology

Volume 49, Issue 12, December 1996, Pages 1429-1433

https://doi.org/10.1016/S0895-4356(96)00271-5 Get rights and content

Abstract

To develop a measure of the burden of comorbid disease from the MED-ECHO data base (Québec), the so-called Charlson index was adapted to International Classification of Disease (ICD-9) codes. The resulting comorbidity index was applied to the study of inpatient death in 33,940 patients with ischemic heart disease. Multiple logistic regression was used to relate inpatient death to its predictors, including gender, principal diagnosis, age, and the comorbidity index. Various transformations of the comorbidity score were performed, and their effect on the predictive accuracy was assessed. The comorbidity index was constantly and strongly associated with death. From a statistical viewpoint, the best results were obtained when the index was transformed into four dummy independent variables (the area under the receiver-operating curve is then 0.87). In a validation analysis performed on 1990–1991 MED-ECHO data (36,012 admissions with ischemic heart disease), the comorbidity index has the same statistical properties. We conclude that the Charlson index may be an efficient approach to risk adjustment from administrative data bases, although it should be tested on other conditions.

References (32)

ME Charlson et al.
A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation
J Chron Dis
(1987)
LL Roos et al.
Risk adjustment in claimsbased research: The search for efficient approaches
J Clin Epidemiol
(1989)
PS Romano et al.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: Differing perspectives
J Clin Epidemiol
(1993)
PS Romano et al.
A comparison of administrative versus clinical data: Coronary artery bypass surgery as an example
J Clin Epidemiol
(1994)
DJ Malenka et al.
Using administrative data to describe casemix: A comparison with the medical record
J Clin Epidemiol
(1994)
C Melfi et al.
Selecting a patient characteristics index for the prediction of medical outcomes using administrative data
J Clin Epidemiol
(1995)
NP Roos et al.
Mortality and reoperation after open and transurethral resection of the prostate for benign prostatic hyperplasia
N Engl J Med
(1989)
LL Roos et al.
Postsurgical mortality in Manitoba and New England
JAMA
(1990)
GT O'Connor et al.
A regional prospective study of in-hospital mortality associated with coronary artery bypass grafting
JAMA
(1991)
RA Deyo et al.
Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases
J Clin Epidemiol
(1992)

Gouvernement du Québec. Ministère de la Santé et des Services sociaux. Direction générale du budget et de...

DW Hosmer et al.

G Anderson et al.

Development of clinical and economic prognosis from Medicare claims data

JAMA

(1990)

S DesHarnais et al.

Measuring outcomes of hospital care using multiple risk-adjusted indexes

Health Serv Res

(1991)

A Fink et al.

The condition of the literature on differences in hospital mortality

Med Care

(1989)

KL Kahn et al.

Interpreting hospital mortality data: How can we proceed?

JAMA

(1988)

Cited by (784)

Multimorbidity in atrial fibrillation for clinical implications using the Charlson Comorbidity Index
2024, International Journal of Cardiology
Predicting survival in atrial fibrillation (AF) patients with comorbidities is challenging. This study aimed to assess multimorbidity in AF patients using the Charlson Comorbidity Index (CCI) and its clinical implications.
We analyzed 451,368 participants from the Korea National Health Insurance Service-Health Screening cohort (2002−2013) without prior AF diagnoses. Patients were categorized into new-onset AF and non-AF groups, with a high CCI defined as ≥4 points. Antithrombotic treatment and outcomes (all-cause death, stroke, major bleeding, and heart failure [HF] hospitalization) were evaluated over 9 years.
In total, 9.5% of the enrolled patients had high CCI. During follow-up, 12,241 patients developed new-onset AF. Among AF patients, antiplatelet drug use increased significantly in those with high CCI (adjusted odds ratio [OR] 1.05, 95%confidence interval [CI] 1.02–1.08, P < .001). However, anticoagulants were significantly less prescribed in patients with high CCI (OR 0.97, 95%CI 0.95–0.99, P = .012). Incidence of adverse events (all-cause death, stroke, major bleeding, HF hospitalization) progressively increased in this order: low CCI without AF, high CCI without AF, low CCI with AF, and high CCI with AF (all P < .001). Furthermore, high CCI with AF had a significantly higher risk compared to low CCI without AF (all-cause death, adjusted hazard ratio [aHR] 2.52, 95% CI 2.37–2.68, P < .001; stroke, aHR 1.43, 95% CI 1.29–1.58, P < .001; major bleeding, aHR 1.14, 95% CI 1.04–1.26, P = .007; HF hospitalization, aHR 4.75, 95% CI 4.03–5.59, P < .001).
High CCI predicted increased antiplatelet use and reduced oral anticoagulant prescription. AF was associated with higher risks of all-cause death, stroke, major bleeding, and HF hospitalization compared to high CCI.
Omission of Axillary Dissection in Node Positive Breast Cancer After Neoadjuvant Systemic Therapy
2023, Journal of Surgical Research
Guidelines recommend axillary lymph node dissection (ALND) for ypN + positive patients as patients receiving neoadjuvant systemic therapy (NST) were excluded from trials omitting ALND in pN + patients. We sought to characterize trends in omission of ALND in patients with ypN + disease.
Adult women with invasive breast carcinoma in the National Cancer Database between 2012 and 2019 who received NST (chemotherapy or endocrine) and had ypN + disease were included. Patients were excluded if they did not have definitive surgery within eight months of diagnosis. The primary study outcome was completion of ALND versus omission. Differences in demographics, tumor characteristics, and treatment were identified using bivariate and multivariate logistic regression models.
In total, 103,121 women were included. Most had cT1 (26%) or cT2 (45%) tumors, cN + disease (71%), and ductal histology (83%). 69% of patients received neoadjuvant chemotherapy and 31% neoadjuvant endocrine without chemotherapy (30% both). ALND was performed in 77% of patients. Omission of ALND became more prevalent each year from 2012 (14%) to 2019 (34%). On multivariate modeling, year of diagnosis, black race, cN status, higher grade, estrogen receptor+/HER2-receptor subtype, and mastectomy were associated with increased prevalence of ALND. Age, Charlson/Deyo comorbidity index score, endocrine versus chemotherapy, and adjuvant radiation were not associated with receipt of ALND.
Despite guidelines recommending ALND, omission is common in patients with ypN + breast cancer after NST. Omission of ALND increased significantly over time and is associated with clinical and demographic factors. Future study is needed to determine the oncologic safety of this approach.
Early signaling of inflammation in patients following traumatic injury with accurately estimated time of injury by profiling C-reactive protein levels
2023, Clinica Chimica Acta
Despite its widespread use, the precise dynamics of CRP response in clinical practice remain poorly defined. We employed a novel quadratic model to explore the time-course analysis of CRP values in trauma patients with known precise time of injury.
Relevant data on all adult patients admitted to our hospital following traumatic incidents between January 1st 2010 to December 31, 2020 were retrospectively collected. Those with a documented time of injury and who underwent CRP evaluation within the first 24 h since injury were studied.
Based on the findings from our annual health check-up center, we established a reference upper normal CRP value of 12.99 mg/L. Within the first 7 h after injury, the CRP levels of 8–9% of the 1545 study patients exceeded the reference threshold. The proportion of patients with CRP levels > 12.99 mg/L increased to 18.5% at 8–9 h later and rose sharply to 91.6% at 22–24 h later. Our quadratic model yielded the equation: CRP = 5.122–0.528xTime + 0.139xTime ². It accounted for > 40% of the variance in CRP levels (R² = 42.4%).
Clear and prominent CRP elevations following atraumatic event are detected only 9–12 h following the insult. This novel finding has crucial implications for accurate CRP assessment of inflammatory responses to physical injuries.
Assessment of a close respiratory follow-up schedule at 3 and 6 months after acute COVID-19 and its related investigations
2023, Respiratory Medicine
Respiratory sequela after acute COVID-19 is common and requires medical follow-up. Considering its vast economic impact, there is still no consensus regarding the mid-term follow-up plan after recovery.
To evaluate the necessity of a close pulmonary follow-up schedule after acute COVID-19 and its related investigations.
A prospective cohort study including adult patients after acute COVID-19 pneumonia. Patients were invited or referred to a 3- and 6-month follow-up visits at a large pulmonary institute in a tertiary center. Before each visit, patients completed demographic and clinical questionnaires, pulmonary function tests (PFTs), and chest CT scans.
168 patients were included after completing both visits (medians of 80 and 177 days). Their mean age was 58 ± 15 and 52% recovered from severe or critical COVID-19. Between the two visits, there was no change in DLCOc (mean 73 ± 18 %predicted in both visits) and FVC (mean 90 ± 16 vs. 89 ± 16 %predicted). The COPD assessment tool and modified Medical Research Council scale had inverse correlations with the DLCOc, and similarly did not change between the visits. Occupational exposures were the only factor associated with a change in DLCOc during follow-up (3% decrease, p = 0.04). An improvement in chest CT findings at the second visit was not associated with a change in PFTs.
Most clinical variables did not change during a close follow-up schedule in the first six months after acute COVID-19. Such a follow-up plan does not appear necessary and should be personalized to limit excessive costs and resources.
Prescribing trends of glucose-lowering drugs in older adults from 2010 to 2021: A population-based study of Northern Italy
2023, Diabetes Research and Clinical Practice
To describe glucose-lowering drugs prescribing pattern in a large population of older diabetics from 2010 to 2021.
Using linkable administrative health databases, we included patients aged 65–90 years treated with glucose-lowering drugs. Prevalence rate of drugs was collected within each study year. A stratified analysis by gender, age and coexistence of cardiovascular disease (CVD) was conducted.
A total of 251 737 and 308 372 patients were identified in 2010 and 2021, respectively. Use of metformin (68.4% to 76.6%), DPP-4i (1.6% to 18.4%), GLP-1-RA (0.4% to 10.2%), SGLT2i (0.6% to 11.1%) increased, while sulfonylureas (53.6% to 20.7%) and glinides (10.5% to 3.5%) decreased over time. Metformin, glitazones, GLP1-RA, SGLT2i and DPP4i (except for 2021) usage decreased with aging, in contrast to sulfonylureas, glinides and insulin. The coexistence of CVD was associated with a higher prescription of glinides, insulin, DPP-4i, GLP1-RA and SGLT2i, particularly in 2021.
We found a significant increase in the prescriptions of GLP-1 RA and SGLT2i in older diabetics, mainly in those with CVD. However, drugs without CV benefits including sulfonylureas and DPP-4i continued to be highly prescribed in older patients. There is still room to improve the management in this population according to recommendations.
The role of ECMO in COVID-19 acute respiratory failure: Defining risk factors for mortality
2023, American Journal of Surgery
Veno-venous extracorporeal membrane oxygenation (VV ECMO) utilization increased substantially during the COVID-19 pandemic, but without patient selection criteria.
We conducted a retrospective review of all adult patients with COVID-19-associated ARDS placed on VV ECMO at our institution from April 2020 through June 2022.
162 patients were included (n = 95 Pre-Delta; n = 58 Delta; n = 9 Omicron). The frequency of ECMO duration greater than three weeks was variable by pandemic period (17% pre-Delta, 41% Delta, 22% Omicron, p = 0.003). In-hospital mortality was 60.5%. Age ≥50 years (RR 1.28, 95% CI 1.01, 1.62), ≥7 days of respiratory support (1.39, 95% CI 1.05, 1.83) and pre-cannulation renal failure requiring dialysis (RR 1.42, 95% CI 1.13, 1.78) were associated with mortality.
In this cohort of VV ECMO patients with COVID-19, older age, a longer duration of pre-ECMO respiratory support, and pre-ECMO renal failure all increased the risk of mortality by approximately 30%.

View all citing articles on Scopus

View full text

Original articlePractical considerations on the use of the charlson comorbidity index with administrative data bases

Abstract

J Chron Dis

J Clin Epidemiol

J Clin Epidemiol

J Clin Epidemiol

J Clin Epidemiol

J Clin Epidemiol

Mortality and reoperation after open and transurethral resection of the prostate for benign prostatic hyperplasia

N Engl J Med

Postsurgical mortality in Manitoba and New England

JAMA

A regional prospective study of in-hospital mortality associated with coronary artery bypass grafting

JAMA

Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases

J Clin Epidemiol

Development of clinical and economic prognosis from Medicare claims data

JAMA

Measuring outcomes of hospital care using multiple risk-adjusted indexes

Health Serv Res

The condition of the literature on differences in hospital mortality

Med Care

Interpreting hospital mortality data: How can we proceed?

JAMA

Original article
Practical considerations on the use of the charlson comorbidity index with administrative data bases