Andrew Fong

Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.

Objectives:Implementation of the International Statistical Classification of Disease and Related Health Problems, 10th Revision (ICD-10) coding system presents challenges for using administrative data. Recognizing this, we conducted a multistep process to develop ICD-10 coding algorithms to define Charlson and Elixhauser comorbidities in administrative data and assess the performance of the resulting algorithms. Methods:ICD-10 coding algorithms were developed by “translation” of the ICD-9-CM codes constituting Deyos (for Charlson comorbidities) and Elixhausers coding algorithms and by physicians’ assessment of the face-validity of selected ICD-10 codes. The process of carefully developing ICD-10 algorithms also produced modified and enhanced ICD-9-CM coding algorithms for the Charlson and Elixhauser comorbidities. We then used data on in-patients aged 18 years and older in ICD-9-CM and ICD-10 administrative hospital discharge data from a Canadian health region to assess the comorbidity frequencies and mortality prediction achieved by the original ICD-9-CM algorithms, the enhanced ICD-9-CM algorithms, and the new ICD-10 coding algorithms. Results:Among 56,585 patients in the ICD-9-CM data and 58,805 patients in the ICD-10 data, frequencies of the 17 Charlson comorbidities and the 30 Elixhauser comorbidities remained generally similar across algorithms. The new ICD-10 and enhanced ICD-9-CM coding algorithms either matched or outperformed the original Deyo and Elixhauser ICD-9-CM coding algorithms in predicting in-hospital mortality. The C-statistic was 0.842 for Deyos ICD-9-CM coding algorithm, 0.860 for the ICD-10 coding algorithm, and 0.859 for the enhanced ICD-9-CM coding algorithm, 0.868 for the original Elixhauser ICD-9-CM coding algorithm, 0.870 for the ICD-10 coding algorithm and 0.878 for the enhanced ICD-9-CM coding algorithm. Conclusions:These newly developed ICD-10 and ICD-9-CM comorbidity coding algorithms produce similar estimates of comorbidity prevalence in administrative data, and may outperform existing ICD-9-CM coding algorithms.

Variation in health services utilization among ethnic populations

Background: Although racial and ethnic disparities in health services utilization and outcomes have been extensively studied in several countries, this issue has received little attention in Canada. We therefore analyzed data from the 2001 Canadian Community Health Survey to compare the use of health services by members of visible minority groups and nonmembers (white people) in Canada. Methods: Logistic regression was used to compare physician contacts and hospital admissions during the 12 months before the survey and recent cancer screening tests. Explanatory variables recorded from the survey included visible minority status, sociodemographic factors and health measures. Results: Respondents included 7057 members of visible minorities and 114 255 white people for analysis. After adjustments for sociodemographic and health characteristics, we found that minority members were more likely than white people to have had contact with a general practitioner (adjusted odds ratio [OR] 1.28, 95% confidence interval [CI] 1.14–1.42), but not specialist physicians (OR 1.01, 95% CI 0.93–1.10). Members of visible minorities were less likely to have been admitted to hospital (OR 0.83, 95% CI 0.70– 0.98), tested for prostate-specific antigen (OR 0.64, 95% CI 0.52–0.79), administered a mammogram (OR 0.68, 95% CI 0.59–0.80) or given a Pap test (OR 0.47, 95% CI 0.39–0.56). Interpretation: Use of health services in Canada varies considerably by ethnicity according to type of service. Although there is no evidence that members of visible minorities use general physician and specialist services less often than white people, their utilization of hospital and cancer screening services is significantly less.

Epidemiology and natural history of primary biliary cirrhosis in a Canadian health region: A population‐based study

The recent epidemiology and outcomes of primary biliary cirrhosis (PBC) in North America are incompletely described, partly due to difficulties in case ascertainment. In light of their availability, broad coverage, and limited expense, administrative databases may facilitate such investigations. We used population‐based administrative data (inpatient, ambulatory care, and physician billing databases) and a validated International Classification of Diseases coding algorithm to describe the epidemiology and natural history of PBC in the Calgary Health Region (population ≈1.1 million). Between 1996 and 2002, the overall age/sex‐adjusted annual incidence of PBC was 30.3 cases per million (48.4 per million in women, 10.4 per million in men). Although the incidence remained stable, the prevalence increased from 100 per million in 1996 to 227 per million in 2002 (P < 0.0005). Among 137 incident cases with a total follow‐up of 801 person‐years from diagnosis (median 5.8 years), 27 patients (20%) died and six (4.4%) underwent liver transplantation. The estimated 10‐year probabilities of survival, liver transplantation, and transplant‐free survival were 73% (95% confidence interval [CI] 60%–83%), 6% (95% CI 2.5%–12.6%), and 68% (95% CI 55%–78%), respectively. Survival in PBC patients was significantly lower than that of the age/sex‐matched Canadian population (standardized mortality ratio 2.87; 95% CI 1.89–4.17); male sex (hazard ratio [HR] 3.80; 95% CI 1.85–7.82) and an older age at diagnosis (HR per additional year, 1.06; 95% CI 1.03–1.10) were independent predictors of mortality. Conclusion: This population‐based study demonstrates that the burden of PBC in Canada is high and growing. Survival of PBC patients is significantly lower than that of the general population, emphasizing the importance of developing new therapies for this condition. (HEPATOLOGY 2009.)

Cost-effectiveness of computerized tomographic colonography versus colonoscopy for colorectal cancer screening

Background: Computerized tomographic (CT) colonography is a potential alternative to colonoscopy for colorectal cancer screening. Its main advantage, a better safety profile, may be offset by its limitations: lower sensitivity, need for colonoscopy in cases where results are positive, and expense. Methods: We performed an economic evaluation, using decision analysis, to compare CT colonography with colonoscopy for colorectal cancer screening in patients over 50 years of age. Three-year outcomes included number of colonoscopies, perforations and adenomas removed; deaths from perforation and from colorectal cancer from missed adenomas; and direct health care costs. The expected prevalence of adenomas, test performance characteristics of CT colonography and colonoscopy, and probability of colonoscopy complications and cancer from missed adenomas were derived from the literature. Costs were determined in detail locally. Results: Using the base-case assumptions, a strategy of CT colonography for colorectal cancer screening would cost

A comparison between the APACHE II and Charlson Index Score for predicting hospital mortality in critically ill patients

2.27 million extra per 100 000 patients screened; 3.78 perforation-related deaths would be avoided, but 4.11 extra deaths would occur from missed adenomas. Because screening with CT colonography would cost more and result in more deaths overall compared with colonoscopy, the latter remained the dominant strategy. Our results were sensitive to CT colonographys test performance characteristics, the malignant risk of missed adenomas, the risk of perforation and related death, the procedural costs and differences in screening adherence. Interpretation: At present, CT colonography cannot be recommended as a primary means of population-based colorectal cancer screening in Canada.

Assessing record linkage between health care and Vital Statistics databases using deterministic methods

BackgroundRisk adjustment and mortality prediction in studies of critical care are usually performed using acuity of illness scores, such as Acute Physiology and Chronic Health Evaluation II (APACHE II), which emphasize physiological derangement. Common risk adjustment systems used in administrative datasets, like the Charlson index, are entirely based on the presence of co-morbid illnesses. The purpose of this study was to compare the discriminative ability of the Charlson index to the APACHE II in predicting hospital mortality in adult multisystem ICU patients.MethodsThis was a population-based cohort design. The study sample consisted of adult (>17 years of age) residents of the Calgary Health Region admitted to a multisystem ICU between April 2002 and March 2004. Clinical data were collected prospectively and linked to hospital outcome data. Multiple regression analyses were used to compare the performance of APACHE II and the Charlson index.ResultsThe Charlson index was a poor predictor of mortality (C = 0.626). There was minimal difference between a baseline model containing age, sex and acute physiology score (C = 0.74) and models containing either chronic health points (C = 0.76) or Charlson index variations (C = 0.75, 0.76, 0.77). No important improvement in prediction occurred when the Charlson index was added to the full APACHE II model (C = 0.808 to C = 0.813).ConclusionThe Charlson index does not perform as well as the APACHE II in predicting hospital mortality in ICU patients. However, when acuity of illness scores are unavailable or are not recorded in a standard way, the Charlson index might be considered as an alternative method of risk adjustment and therefore facilitate comparisons between intensive care units.

Health resource use in epilepsy: Significant disparities by age, gender, and aboriginal status.

BackgroundWe assessed the linkage and correct linkage rate using deterministic record linkage among three commonly used Canadian databases, namely, the population registry, hospital discharge data and Vital Statistics registry.MethodsThree combinations of four personal identifiers (surname, first name, sex and date of birth) were used to determine the optimal combination. The correct linkage rate was assessed using a unique personal health number available in all three databases.ResultsAmong the three combinations, the combination of surname, sex, and date of birth had the highest linkage rate of 88.0% and 93.1%, and the second highest correct linkage rate of 96.9% and 98.9% between the population registry and Vital Statistics registry, and between the hospital discharge data and Vital Statistics registry in 2001, respectively. Adding the first name to the combination of the three identifiers above increased correct linkage by less than 1%, but at the cost of lowering the linkage rate almost by 10%.ConclusionOur findings suggest that the combination of surname, sex and date of birth appears to be optimal using deterministic linkage. The linkage and correct linkage rates appear to vary by age and the type of database, but not by sex.

Hospitalizations for acetaminophen overdose: a Canadian population-based study from 1995 to 2004

Purpose: Epilepsy imposes a significant burden on society. The objective of this study was to estimate health resource utilization (HRU) over a 1‐year period in epilepsy patients, using administrative databases.

Effect of Invasive Coronary Revascularization in Acute Myocardial Infarction on Subsequent Death Rate and Frequency of Chronic Heart Failure

BackgroundAcetaminophen overdose (AO) is the most common cause of acute liver failure. We examined temporal trends and sociodemographic risk factors for AO in a large Canadian health region.Methods1,543 patients hospitalized for AO in the Calgary Health Region (population ~1.1 million) between 1995 and 2004 were identified using administrative data.ResultsThe age/sex-adjusted hospitalization rate decreased by 41% from 19.6 per 100,000 population in 1995 to 12.1 per 100,000 in 2004 (P < 0.0005). This decline was greater in females than males (46% vs. 29%). Whereas rates fell 46% in individuals under 50 years, a 50% increase was seen in those ≥ 50 years. Hospitalization rates for intentional overdoses fell from 16.6 per 100,000 in 1995 to 8.6 per 100,000 in 2004 (2004 vs. 1995: rate ratio [RR] 0.49; P < 0.0005). Accidental overdoses decreased between 1995 and 2002, but increased to above baseline levels by 2004 (2004 vs. 1995: RR 1.24;P < 0.0005). Risk factors for AO included female sex (RR 2.19; P < 0.0005), Aboriginal status (RR 4.04; P < 0.0005), and receipt of social assistance (RR 5.15; P < 0.0005).ConclusionHospitalization rates for AO, particularly intentional ingestions, have fallen in our Canadian health region between 1995 and 2004. Young patients, especially females, Aboriginals, and recipients of social assistance, are at highest risk.

Validation of Coding Algorithms for the Identification of Patients with Primary Biliary Cirrhosis Using Administrative Data

There is debate about whether therapies that reduce mortality in acute myocardial infarction (AMI) will increase the risk for heart failure. In this study, an inception cohort of patients hospitalized with AMIs from April 1, 1994, to March 31, 1999 (without previous diagnoses of heart failure or myocardial infarction), were followed for a mean of 32 months to explore whether invasive coronary revascularization during the index AMI hospitalization was associated with a trade-off between reduced mortality in the short term and increased heart failure in the intermediate term. Of 13,472 patients (mean age 65 +/- 13 years, 70% men), 3,278 (24%) underwent invasive coronary revascularization during their index AMI hospitalizations. Patients who underwent invasive revascularization during their index AMI hospitalizations were less likely to die (171 of 3,278 [5%] vs 1,688 of 10,194 [17%], p <0.0001) and were less likely to develop heart failure, either during the AMI hospitalization (571 of 3,278 [17%] vs 2,422 of 10,194 [24%], p <0.0001) or after discharge (144 of 3,278 [4%] vs 754 of 10,194 [7%], p <0.0001). These associations persisted after covariate adjustment (for heart failure, hazard ratio 0.68, 95% confidence interval 0.56 to 0.81; for death or heart failure, hazard ratio 0.60, 95% confidence interval 0.51 to 0.70). In conclusion, invasive coronary revascularization during AMI hospitalization is associated with lower rates of death and subsequent heart failure; there is no trade-off of 1 outcome for the other.