Adapting the Elixhauser comorbidity index for cancer patients

Mehta, Hemalkumar B.; Sura, Sneha; Adhikari, Deepak; Andersen, Clark R.; Williams, Stephen B.; Kuo, Yong Fang

doi:10.1002/cncr.31269

Cited by 61 publications

(38 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The majority of physicians included in the analysis were male (95.9% [n ¼ 5430]) and practiced in urban locations (92.6% [n ¼ 5243]; Table II). The median number of years in practice was 23 years (IQR, 16-30 years), and the median number of patients who received a new diagnosis of claudication during the study period was 22 (IQR, [15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] Patient and physician characteristics associated with early PVI. Based on univariable logistic regression analysis, patient characteristics associated with higher early PVI rates included age 65 to 74 years, male sex, hypertension, and smoking (all P < .001; Table III).…”

Section: Resultsmentioning

confidence: 99%

“…To define a comorbidity, we required at least one diagnosis from the inpatient claims or at least two diagnoses recorded >30 days apart from the outpatient or carrier claims. 15,16 To define an ever smoker, we required at least one diagnosis of smoking from any type of claim. 17 We also determined whether patients had undergone noninvasive testing with ankle-brachial indices within 3 months before or after the first claudication diagnosis using the codes listed in the Supplementary Table (online only).…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Overuse of early peripheral vascular interventions for claudication

Hicks

Holscher

Wang

et al. 2020

Journal of Vascular Surgery

View full text Add to dashboard Cite

Objective: Guidelines from the Society for Vascular Surgery and the Choosing Wisely campaign recommend that peripheral vascular interventions (PVIs) be limited to claudication patients with lifestyle-limiting symptoms only after a failed trial of medical and exercise therapy. We sought to explore practice patterns and physician characteristics associated with early PVI after a new claudication diagnosis to evaluate adherence to these guidelines. Methods:We used 100% Medicare fee-for-service claims to identify patients diagnosed with claudication for the first time between 2015 and 2017. Early PVI was defined as an aortoiliac or femoropopliteal PVI performed within 6 months of initial claudication diagnosis. A physician-level PVI utilization rate was calculated for physicians who diagnosed >10 claudication patients and performed at least one PVI (regardless of indication) during the study period. Hierarchical multivariable logistic regression was used to identify physician-level factors associated with early PVI.Results: Of 194,974 patients who had a first-time diagnosis of claudication during the study period, 6286 (3.2%) underwent early PVI. Among the 5664 physicians included in the analysis, the median physician-level early PVI rate was low at 0% (range, 0%-58.3%). However, there were 320 physicians (5.6%) who had an early PVI rate $14% ($2 standard deviations above the mean). After accounting for patient characteristics, a higher percentage of services delivered in ambulatory surgery center or office settings was associated with higher PVI utilization (vs 0%-22%; 23%-47%: adjusted odds ratio [aOR], 1.23; 48%-68%: aOR, 1.49; 69%-100%: aOR, 1.72; all P < .05). Other risk-adjusted physician factors independently associated with high PVI utilization included male sex (aOR, 2.04), fewer years in practice (vs $31 years; 11-20 years: aOR, 1.23; 21-30 years: aOR, 1.13), rural location (aOR, 1.25), and lower volume claudication practice (vs $30 patients diagnosed during study period; #17 patients: aOR, 1.30; 18-29 patients: aOR, 1.35; all P < .05). Conclusions:Outlier physicians with a high early PVI rate for patients newly diagnosed with claudication are identifiable using a claims-based practice pattern measure. Given the shared Society for Vascular Surgery and Choosing Wisely initiative goal to avoid interventions for first-line treatment of claudication, confidential data-sharing programs using national benchmarks and educational guidance may be useful to address high utilization in the management of claudication. (J Vasc Surg 2020;71:121-30.)

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Overuse of early peripheral vascular interventions for claudication

Hicks

Holscher

Wang

et al. 2020

Journal of Vascular Surgery

View full text Add to dashboard Cite

show abstract

“…With the assessment of comorbidi- Table 2. Optimal proportional hazard models predicting overall and non-bladder cancer mortality analyzing all individual conditions contributing to the modified self-administrable comorbidity index recommended in the ICHOM standard sets [5][6][7][8][9] ty, more complex instruments may be expected to provide better predictive performance [13]. However, particularly when no detailed documentation of concomitant diseases is available, or when data are obtained via questionnaire from the patients themselves, complexity may hinder data collection.…”

Section: Discussionmentioning

confidence: 99%

Validation of a Questionnaire-Suitable Comorbidity Index in Patients Undergoing Radical Cystectomy

et al. 2020

View full text Add to dashboard Cite

Objective: To investigate the capability of a modified selfadministrable comorbidity index recommended in the standard sets for neoplastic diseases published by the International Consortium for Health Outcomes Measurement (ICHOM) to predict 90-day and long-term mortality after radical cystectomy. Methods: A single-center series of 1,337 consecutive patients who underwent radical cystectomy for muscle-invasive or high-risk non-muscle-invasive urothelial or undifferentiated bladder cancer were stratified by the modified self-administrable comorbidity index and Charlson score, respectively. Multivariate logit models (for 90-day mortality) and proportional-hazards models (for overall and non-bladder cancer mortality) were used for statistical workup. Results: Considering 90-day mortality, both comorbidity indexes contributed independent information when analyzed together with age (p < 0.0001). The Charlson score performed slightly better (area under the curve [AUC] 0.74 vs. 0.72 for the ICHOM-recommended comorbidity index). Considering 5-year overall mortality in 727 patients with complete observation, the performance of both measures was similar (AUC 0.63 vs. 0.62, including age AUC 0.66 for both indexes). With 6-sided stratifications, the modified self-administrable comorbidity index separated the risk groups slightly better (p values for directly neighboring curves: 0.0068-0.1043 vs. 0.0001-0.8100). Conclusion: The ICHOMrecommended modified self-administrable comorbidity index is capable of predicting 90-day mortality and long-term non-bladder cancer mortality after radical cystectomy similarly to the commonly used Charlson score.

show abstract

“…Moreover, the prognostic impact of comorbidities may differ across subpopulations . For example, the prognostic impact of congestive heart failure on 1‐year mortality was greater in breast cancer patients (hazard ratio = 2.03) compared to colorectal cancer patients (hazard ratio = 1.41) . Therefore, comorbidity scores developed for a target population perform better than scores developed using broader populations .…”

Section: Discussionmentioning

confidence: 99%

“…30 For example, the prognostic impact of congestive heart failure on 1-year mortality was greater in breast cancer patients (hazard ratio = 2.03) compared to colorectal cancer patients (hazard ratio = 1.41). 31 Therefore, comorbidity scores developed for a target population perform better than scores developed using broader populations. 15,[32][33][34] The surgery-specific comorbidity score has 63 comorbidities in comparison to the 79 hierarchal condition categories in the CMS-HCC.…”

Section: Discussionmentioning

confidence: 99%

Development of comorbidity score for patients undergoing major surgery

Mehta

Shan

Sura

et al. 2019

Health Services Research

Self Cite

View full text Add to dashboard Cite

Objective To develop and validate a claims‐based comorbidity score for patients undergoing major surgery, and compare its performance with established comorbidity scores. Data Source Five percent Medicare data from 2007 to 2014. Study Design Retrospective cohort study of patients aged ≥65 years undergoing six major operations (N = 99 250). Data Collection One‐year mortality was the primary outcome. Secondary outcomes were hospital mortality, 30‐day mortality, 30‐day readmission, and length of stay. The comorbidity score was developed in the derivation cohort (70 percent sample) using logistic regression model. The comorbidity score was calibrated and validated in the validation cohort (30 percent sample), and compared against the Charlson, Elixhauser, and Centers for Medicare and Medicaid Services Hierarchical Condition Categories (CMS‐HCC) comorbidity scores using c‐statistic, net reclassification improvement, and integrated discrimination improvement. Principal Findings In the validation cohort, the surgery‐specific comorbidity score was well calibrated and performed better than the Charlson, Elixhauser, and CMS‐HCC comorbidity scores for all outcomes; the performance was comparable to the CMS‐HCC for 30‐day readmission. For example, the surgery‐specific comorbidity score (c‐statistic = 0.792; 95% CI, 0.785‐0.799) had greater discrimination than the Charlson (c‐statistic = 0.747; 95% CI, 0.739‐0.755), Elixhauser (c‐statistic = 0.747; 95% CI, 0.735‐0.755), or CMS‐HCC (c‐statistic = 0.755; 95% CI, 0.747‐0.763) scores in predicting 1‐year mortality. The net reclassification improvement and integrated discrimination improvement were greater for surgery‐specific comorbidity score compared to the Charlson, Elixhauser, and CMS‐HCC scores. Conclusions Compared to commonly used comorbidity measures, a surgery‐specific comorbidity score better predicted outcomes in the surgical population.

show abstract

Adapting the Elixhauser comorbidity index for cancer patients

Cited by 61 publications

References 19 publications

Overuse of early peripheral vascular interventions for claudication

Overuse of early peripheral vascular interventions for claudication

Validation of a Questionnaire-Suitable Comorbidity Index in Patients Undergoing Radical Cystectomy

Development of comorbidity score for patients undergoing major surgery

Contact Info

Product

Resources

About