Development and validation of a claims based index for adjusting for risk of mortality: the case of acute myocardial infarction

Large-scale evidence generation and evaluation across a network of databases for type 2 diabetes mellitus (LEGEND-T2DM): a protocol for a series of multinational, real-world comparative cardiovascular effectiveness and safety studies

INTRODUCTION

Therapeutic options for type 2 diabetes mellitus (T2DM) have expanded over the last decade with the emergence of cardioprotective novel agents, but without such data for older drugs, leaving a critical gap in our understanding of the relative effects of T2DM agents on cardiovascular risk.

METHODS AND ANALYSIS

The large-scale evidence generations across a network of databases for T2DM (LEGEND-T2DM) initiative is a series of systematic, large-scale, multinational, real-world comparative cardiovascular effectiveness and safety studies of all four major second-line anti-hyperglycaemic agents, including sodium-glucose co-transporter-2 inhibitor, glucagon-like peptide-1 receptor agonist, dipeptidyl peptidase-4 inhibitor and sulfonylureas. LEGEND-T2DM will leverage the Observational Health Data Sciences and Informatics (OHDSI) community that provides access to a global network of administrative claims and electronic health record data sources, representing 190 million patients in the USA and about 50 million internationally. LEGEND-T2DM will identify all adult, patients with T2DM who newly initiate a traditionally second-line T2DM agent. Using an active comparator, new-user cohort design, LEGEND-T2DM will execute all pairwise class-versus-class and drug-versus-drug comparisons in each data source, producing extensive study diagnostics that assess reliability and generalisability through cohort balance and equipoise to examine the relative risk of cardiovascular and safety outcomes. The primary cardiovascular outcomes include a composite of major adverse cardiovascular events and a series of safety outcomes. The study will pursue data-driven, large-scale propensity adjustment for measured confounding, a large set of negative control outcome experiments to address unmeasured and systematic bias.

ETHICS AND DISSEMINATION

The study ensures data safety through a federated analytic approach and follows research best practices, including prespecification and full disclosure of results. LEGEND-T2DM is dedicated to open science and transparency and will publicly share all analytic code from reproducible cohort definitions through turn-key software, enabling other research groups to leverage our methods, data and results to verify and extend our findings.

Prognostic impact of comorbidity measures on outcomes following acute coronary syndrome: A systematic review

AIM

To identify existing comorbidity measures and summarise their association with acute coronary syndrome (ACS) outcomes.

METHODS

We searched published studies from MEDLINE (OVIDSP) and EMBASE from inception to March 2021, studies of the pre-specified conference proceedings from Web of Science since May 2017, and studies included in any relevant systematic reviews. Studies that reported no comorbidity measures, no association of comorbid burden with ACS outcomes, or only used a comorbidity measure as a confounder without further information were excluded. After independent screening by three reviewers, data extraction and risk of bias assessment of each included study was undertaken. Results were narratively synthesised.

RESULTS

Of 4166 potentially eligible studies identified, 12 (combined n = 6 885 982 participants) were included. Most studies had a high risk of bias at quality assessment. Six different types of comorbidity measures were identified with the Charlson comorbidity index (CCI) the most widely used measure among studies. Overall, the greater the comorbid burden or the higher comorbidity scores recorded, the greater was the association with the risk of mortality.

CONCLUSION

The review summarised different comorbidity measures and reported that higher comorbidity scores were associated with worse ACS outcomes. The CCI is the most widely measure of comorbid burden and shows additive value to clinical risk scores in use.

The DANish Comorbidity Index for Acute Myocardial Infarction (DANCAMI): Development, Validation and Comparison with Existing Comorbidity Indices

Objective

To develop and validate the DANish Comorbidity index for Acute Myocardial Infarction (DANCAMI) for adjustment of comorbidity burden in studies of myocardial infarction prognosis.

Methods

Using medical registries, we identified patients with first-time myocardial infarction in Denmark during 2000–2013 (n=36,685). We developed comorbidity indices predicting 1-year all-cause mortality from all comorbidities (DANCAMI) and restricted to non-cardiovascular comorbidities (rDANCAMI). For variable selection, we eliminated comorbidities stepwise using hazard ratios from multivariable Cox models. We compared DANCAMI/rDANCAMI with Charlson and Elixhauser comorbidity indices using standard performance measures (Nagelkerke’s R², Harrell’s C-statistic, the Integrated Discrimination Improvement, and the continuous Net Reclassification Index). We assessed the significance of the novel DANCAMI variables not included in the Charlson Comorbidity Index. External validation was performed in patients with myocardial infarction in New Zealand during 2007–2016 (n=75,069).

Results

The DANCAMI included 24 comorbidities. The rDANCAMI included 17 non-cardiovascular comorbidities. In the Danish cohort, the DANCAMI indices outperformed both the Charlson and the Elixhauser comorbidity indices on all performance measures. The DANCAMI indices included multiple variables that were significant predictors of 1-year mortality even after controlling for all variables in the Charlson Comorbidity Index. These novel variables included valvular heart disease (hazard ratio for 1-year mortality=1.25, 95% CI: 1.14–1.35), coagulopathy (1.13, 95% CI: 1.05–1.22), alcohol and drug abuse (1.35, 95% CI: 1.15–1.58), schizophrenia (1.60, 95% CI: 1.46–1.76), affective disorder (1.29, 95% CI: 1.22–1.36), epilepsy (1.26, 95% CI: 1.05–1.50), neurodegenerative disorder (1.30, 95% CI: 1.10–1.54) and chronic pancreatitis (1.71, 95% CI: 1.14–2.56). The results were supported by the external validation in New Zealand.

Conclusion

DANCAMI assessed comorbidity burden of patients with first-time myocardial infarction, outperformed existing comorbidity indices, and was generalizable to patients outside Denmark. DANCAMI is recommended as a standard approach for comorbidity adjustment in studies of myocardial infarction prognosis.

A disease-specific comorbidity index for predicting mortality in patients admitted to hospital with a cardiac condition

BACKGROUND

Comorbidity indexes derived from administrative databases are essential tools of research in global health. We sought to develop and validate a novel cardiac-specific comorbidity index, and to compare its accuracy with the generic Charlson-Deyo and Elixhauser comorbidity indexes.

METHODS

We derived the cardiac-specific comorbidity index from consecutive patients who were admitted to hospital at a tertiary-care cardiology hospital in Quebec. We used logistic regression analysis and incorporated age, sex and 22 clinically relevant comorbidities to build the index. We compared the cardiac-specific comorbidity index with refitted Charlson-Deyo and Elixhauser comorbidity indexes using the C-statistic and net reclassification improvement to predict in-hospital death, and the Akaike information criterion to predict length of stay. We validated our findings externally in an independent cohort obtained from a provincial registry of coronary disease in Alberta.

RESULTS

The novel cardiac-specific comorbidity index outperformed the refitted generic Charlson-Deyo and Elixhauser comorbidity indexes for predicting in-hospital mortality in the derivation population (n = 10 137): C-statistic 0.95 (95% confidence interval [CI] 0.94-0.9) v. 0.81 (95% CI 0.77-0.84) and 0.86 (95% CI 0.82-0.89), respectively. In the validation population (n = 17 877), the cardiac-specific comorbidity index was similarly better: C-statistic 0.92 (95% CI 0.89-0.94) v. 0.76 (95% CI 0.71-0.81) and 0.82 (95% CI 0.78-0.86), respectively, and also numerically outperformed the Charlson-Deyo and Elixhauser comorbidity indexes for predicting 1-year mortality (C-statistic 0.78 [95% CI 0.76-0.80] v. 0.75 [95% CI 0.73-0.77] and 0.77 [95% CI 0.75-0.79], respectively). Similarly, the cardiac-specific comorbidity index showed better fit for the prediction of length of stay. The net reclassification improvement using the cardiac-specific comorbidity index for the prediction of death was 0.290 compared with the Charlson-Deyo comorbidity index and 0.192 compared with the Elixhauser comorbidity index.

INTERPRETATION

The cardiac-specific comorbidity index predicted in-hospital and 1-year death and length of stay in cardiovascular populations better than existing generic models. This novel index may be useful for research of cardiology outcomes performed with large administrative databases.

A new population-based risk stratification tool was developed and validated for predicting mortality, hospital admissions, and health care costs

OBJECTIVES

The aim of this study was to develop a new population-based risk stratification tool (Chronic Related Score [CReSc]) for predicting 5-year mortality and other outcomes.

STUDY DESIGN AND SETTING

The score included 31 conditions selected from a list of 65 candidates whose weights were assigned according to the Cox model coefficients. The model was built from a sample of 5.4 million National Health Service (NHS) beneficiaries from the Italian Lombardy Region and applied to the remaining 2.7 million NHS beneficiaries. Predictive performance was assessed by discrimination and calibration. CReSc ability in predicting secondary endpoints (i.e., hospital admissions and health care costs) was investigated. Finally, the relationship between CReSc and income was considered.

RESULTS

Among individuals aged 50-85 years, CReSc performance showed (1) an area under the receiver operating characteristic curve of 0.730, (2) an improved reclassification from 44% to 52% with respect to other scores, and (3) a remarkable calibration. A trend toward increasing rates of all the considered endpoints as CReSc increases was observed. Compared with individuals on low-intermediate income, NHS beneficiaries on high income showed better CReSc profile.

CONCLUSION

We developed a risk stratification tool able to predict mortality, costs, and hospital admissions. The application of CReSc may generate clinically and operationally important effects.

Importance of coding co-morbidities for APR-DRG assignment: Focus on cardiovascular and respiratory diseases

BACKGROUND

The All Patient-Refined Diagnosis-Related Groups (APR-DRGs) system has adjusted the basic DRG structure by incorporating four severity of illness (SOI) levels, which are used for determining hospital payment. A comprehensive report of all relevant diagnoses, namely the patient's underlying co-morbidities, is a key factor for ensuring that SOI determination will be adequate.

OBJECTIVE

In this study, we aimed to characterise the individual impact of co-morbidities on APR-DRG classification and hospital funding in the context of respiratory and cardiovascular diseases.

METHODS

Using 6 years of coded clinical data from a nationwide Portuguese inpatient database and support vector machine (SVM) models, we simulated and explored the APR-DRG classification to understand its response to individual removal of Charlson and Elixhauser co-morbidities. We also estimated the amount of hospital payments that could have been lost when co-morbidities are under-reported.

RESULTS

In our scenario, most Charlson and Elixhauser co-morbidities did considerably influence SOI determination but had little impact on base APR-DRG assignment. The degree of influence of each co-morbidity on SOI was, however, quite specific to the base APR-DRG. Under-coding of all studied co-morbidities led to losses in hospital payments. Furthermore, our results based on the SVM models were consistent with overall APR-DRG grouping logics.

CONCLUSION AND IMPLICATIONS

Comprehensive reporting of pre-existing or newly acquired co-morbidities should be encouraged in hospitals as they have an important influence on SOI assignment and thus on hospital funding. Furthermore, we recommend that future guidelines to be used by medical coders should include specific rules concerning coding of co-morbidities.

Developing and validating a novel multisource comorbidity score from administrative data: a large population-based cohort study from Italy

OBJECTIVE

To develop and validate a novel comorbidity score (multisource comorbidity score (MCS)) predictive of mortality, hospital admissions and healthcare costs using multiple source information from the administrative Italian National Health System (NHS) databases.

METHODS

An index of 34 variables (measured from inpatient diagnoses and outpatient drug prescriptions within 2 years before baseline) independently predicting 1-year mortality in a sample of 500 000 individuals aged 50 years or older randomly selected from the NHS beneficiaries of the Italian region of Lombardy (training set) was developed. The corresponding weights were assigned from the regression coefficients of a Weibull survival model. MCS performance was evaluated by using an internal (ie, another sample of 500 000 NHS beneficiaries from Lombardy) and three external (each consisting of 500 000 NHS beneficiaries from Emilia-Romagna, Lazio and Sicily) validation sets. Discriminant power and net reclassification improvement were used to compare MCS performance with that of other comorbidity scores. MCS ability to predict secondary health outcomes (ie, hospital admissions and costs) was also investigated.

RESULTS

Primary and secondary outcomes progressively increased with increasing MCS value. MCS improved the net 1-year mortality reclassification from 27% (with respect to the Chronic Disease Score) to 69% (with respect to the Elixhauser Index). MCS discrimination performance was similar in the four regions of Italy we tested, the area under the receiver operating characteristic curves (95% CI) being 0.78 (0.77 to 0.79) in Lombardy, 0.78 (0.77 to 0.79) in Emilia-Romagna, 0.77 (0.76 to 0.78) in Lazio and 0.78 (0.77 to 0.79) in Sicily.

CONCLUSION

MCS seems better than conventional scores for predicting health outcomes, at least in the general population from Italy. This may offer an improved tool for risk adjustment, policy planning and identifying patients in need of a focused treatment approach in the everyday medical practice.

Effect of Hand Antisepsis Agent Selection and Population Characteristics on Surgical Site Infection Pathogens

BACKGROUND

Selection of a pre-operative hand antisepsis agent has not been studied in relation to surgical site infection (SSI) culture data. In our hospital, we introduced an alcohol-based hand rub (ABR) in 2012 as an alternative to traditional aqueous surgical scrubs (TSS). It was the goal of this study to review any effect of this implementation on SSI pathogen characteristics. In addition, we sought to compare our SSI culture data with available National Healthcare Safety Network (NHSN) data. We hypothesized that SSI pathogens and resistant isolates are affected by surgical hand antisepsis technique.

METHODS

Data collected prospectively between 2007 and 2014 were retrospectively analyzed for two time periods at the Veterans Affairs Boston Healthcare System (VABHS): Before ABR implementation (TSS group) and after (ABR group). Pathogen distribution and pathogenic isolate resistance profiles were compared for TSS and ABR, and similar comparisons, along with procedure-associated SSI comparisons, were made between VABHS and NHSN. All VABHS data were interpreted and categorized according to NHSN definitions.

RESULTS

Compared with TSS (n = 4,051), ABR (n = 2,293) had a greater rate of Staphylococcus aureus (42.6% vs. 38.0%), Escherichia coli (12.8% vs. 9.9%), Pseudomonas aeruginosa (8.5% vs. 2.8%), and Enterobacter spp. (10.6% vs. 2.8%), and a lower rate of Klebsiella pneumoniae/K. oxytoca (4.3% vs. 8.5%) cultured from superficial and deep SSIs (p < 0.05). Of the S. aureus isolates, 35.0% and 44.4% were resistant to oxacillin/methicillin (MRSA) in ABR and TSS, respectively (p = 0.06). Looking at all SSIs, coagulase-negative staphylococci and K. pneumoniae/K. oxytoca at VABHS (4.0% and 10.4%, respectively) accounted for the biggest difference from NHSN (11.7% and 4.0%, respectively). Aside from MRSA, where there was no difference between VABHS and NHSN (42.9% vs. 43.7%, respectively; p = 0.87), statistically significant (p < 0.05) differences were observed among multi-drug-resistant K. pneumoniae/K. oxytoca (0% vs. 6.8%, respectively) and Escherichia coli (10.0% vs. 1.6%, respectively), as well as among extended-spectrum cephalosporin-resistant K. pneumoniae/K. oxytoca (4.8% vs. 13.2%, respectively) and Enterobacter (58.3% vs. 27.7%, respectively). VABHS had a greater proportion of SSIs in abdominal and vascular cases than did NHSN (48.6% vs. 22.5% and 13.2% vs. 1.5%, respectively). Overall, these differences were significant (p < 0.05).

CONCLUSIONS

The TSS and ABR groups differed in the distribution of pathogens recovered. Those differences, along with SSI pathogen distribution, pathogenic isolate resistance profiles, and procedure-associated SSIs between VABHS and NHSN, warrant further investigation.

The Use of Fixed-and Random-Effects Models for Classifying Hospitals as Mortality Outliers: A Monte Carlo Assessment

Background. There is an increasing movement towards the release of hospital “report-cards.” However, there is a paucity of research into the abilities of the different methods to correctly classify hospitals as performance outliers.Objective.To examine the ability of risk-adjusted mortality rates computed using conventional logistic regression and random-effects logistic regression models to correctly identify hospitals that have higher than acceptable mortality.Research Design.Monte Carlo simulations.Measures.Sensitivity, specificity, and positive predictive value of a classification as a high-outlier for identifying hospitals with higher than acceptable mortality rates.Results.When the distribution of hospital-specific log-odds of death was normal, random-effects models had greater specificity and positive predictive value than fixed-effects models. However, fixed-effects models had greater sensitivity than random-effects models.Conclusions.Researchers and policy makers need to carefully consider the balance between false positives and false negatives when choosing statistical models for determining which hospitals have higher than acceptablemortality in performance profiling.

Evaluation of Acute Myocardial Infarction In-hospital Mortality Using a Risk-adjustment Model Based on Japanese Administrative Data

This study aimed to develop a new risk-adjustment method to assess acute myocardial infarction (AMI) in-hospital mortality. Risk-adjustment was based on variables obtained from administrative data from Japanese hospitals, and included factors such as age, gender, primary diagnosis and co-morbidity. The infarct location was determined using the criteria of the International Classification of Diseases (10th version). Potential co-morbidity risk factors for mortality were selected based on previous studies and their critical influence analysed to identify major co-morbidities. The remaining minor co-morbidities were then divided into two groups based on their medical implications. The major co-morbidities included shock, pneumonia, cancer and chronic renal failure. The two minor co-morbidity groups also demonstrated a substantial impact on mortality. The model was then used to assess clinical performance in the participating hospitals. Our model reliably employed the available data for the risk-adjustment of AMI mortality and provides a new approach to evaluating clinical performance.

Can valid and practical risk-prediction or casemix adjustment models, including adjustment for comorbidity, be generated from English hospital administrative data (Hospital Episode Statistics)? A national observational study

Background

NHS hospitals collect a wealth of administrative data covering accident and emergency (A&E) department attendances, inpatient and day case activity, and outpatient appointments. Such data are increasingly being used to compare units and services, but adjusting for risk is difficult.

Objectives

To derive robust risk-adjustment models for various patient groups, including those admitted for heart failure (HF), acute myocardial infarction, colorectal and orthopaedic surgery, and outcomes adjusting for available patient factors such as comorbidity, using England’s Hospital Episode Statistics (HES) data. To assess if more sophisticated statistical methods based on machine learning such as artificial neural networks (ANNs) outperform traditional logistic regression (LR) for risk prediction. To update and assess for the NHS the Charlson index for comorbidity. To assess the usefulness of outpatient data for these models.

Main outcome measures

Mortality, readmission, return to theatre, outpatient non-attendance. For HF patients we considered various readmission measures such as diagnosis-specific and total within a year.

Methods

We systematically reviewed studies comparing two or more comorbidity indices. Logistic regression, ANNs, support vector machines and random forests were compared for mortality and readmission. Models were assessed using discrimination and calibration statistics. Competing risks proportional hazards regression and various count models were used for future admissions and bed-days.

Results

Our systematic review and empirical analysis suggested that for general purposes comorbidity is currently best described by the set of 30 Elixhauser comorbidities plus dementia. Model discrimination was often high for mortality and poor, or at best moderate, for other outcomes, for examplec = 0.62 for readmission andc = 0.73 for death following stroke. Calibration was often good for procedure groups but poorer for diagnosis groups, with overprediction of low risk a common cause. The machine learning methods we investigated offered little beyond LR for their greater complexity and implementation difficulties. For HF, some patient-level predictors differed by primary diagnosis of readmission but not by length of follow-up. Prior non-attendance at outpatient appointments was a useful, strong predictor of readmission. Hospital-level readmission rates for HF did not correlate with readmission rates for non-HF; hospital performance on national audit process measures largely correlated only with HF readmission rates.

Conclusions

Many practical risk-prediction or casemix adjustment models can be generated from HES data using LR, though an extra step is often required for accurate calibration. Including outpatient data in readmission models is useful. The three machine learning methods we assessed added little with these data. Readmission rates for HF patients should be divided by diagnosis on readmission when used for quality improvement.

Future work

As HES data continue to develop and improve in scope and accuracy, they can be used more, for instance A&E records. The return to theatre metric appears promising and could be extended to other index procedures and specialties. While our data did not warrant the testing of a larger number of machine learning methods, databases augmented with physiological and pathology information, for example, might benefit from methods such as boosted trees. Finally, one could apply the HF readmissions analysis to other chronic conditions.

Funding

The National Institute for Health Research Health Services and Delivery Research programme.

Co-morbidities in elderly patients with hip fracture: recommendations of the ISFR-IOF hip fracture outcomes working group

INTRODUCTION

Hip fractures are the second leading cause of hospitalization in the aged and by 2041, epidemiologists forecast an increase in economic cost to $2.4 billion. The hip patient population often presents with comorbidities causing these patients to receive less aggressive medical treatment and have a low quality of life. We believe that physical function is a patient-important outcome for many medical and surgical interventions. The functional co-morbidity index (FCI), unlike prior co-morbidity indices, was developed with physical function as an outcome instead of being designed for administrative purposes or to predict mortality. Our objective was to evaluate the perceptions of practitioners in hip fracture care about the impact of comorbidities on physical function as primary outcome.

METHODS

We piloted and then distributed a self-administered survey to members of the International Society for Fracture Repair hip fracture outcomes working group. For each of the 18 diagnoses included in the FCI index, we asked in our survey whether the presence of the co-morbidity and whether the severity of the co-morbidity was perceived to impact physical function in patients following a hip fracture.

RESULTS

Seventeen out of 20 respondents completed the questionnaire. The presence and severity of arthritis was 'strongly' believed to predict physical function in those with hip fracture (69 and 85.7 %, respectively). Respondents 'agreed' (range 53-73 %) that 10/18 diagnoses would predict changes in physical function following hip fracture treatment. Whereas, 63 % of the practitioners'strongly disagreed' that diabetes types I and II would change physical function scores. Furthermore, dementia was listed as an additional diagnosis that would affect physical function.

CONCLUSION

The FCI may provide a useful instrument to predict functional outcome after hip fracture; however, the index may need to be modified for this specific population.

Systematic review of comorbidity indices for administrative data

BACKGROUND

Adjustment for comorbidities is common in performance benchmarking and risk prediction. Despite the exponential upsurge in the number of articles citing or comparing Charlson, Elixhauser, and their variants, no systematic review has been conducted on studies comparing comorbidity measures in use with administrative data. We present a systematic review of these multiple comparison studies and introduce a new meta-analytical approach to identify the best performing comorbidity measures/indices for short-term (inpatient + ≤ 30 d) and long-term (outpatient+>30 d) mortality.

METHODS

Articles up to March 18, 2011 were searched based on our predefined terms. The bibliography of the chosen articles and the relevant reviews were also searched and reviewed. Multiple comparisons between comorbidity measures/indices were split into all possible pairs. We used the hypergeometric test and confidence intervals for proportions to identify the comparators with significantly superior/inferior performance for short-term and long-term mortality. In addition, useful information such as the influence of lookback periods was extracted and reported.

RESULTS

Out of 1312 retrieved articles, 54 articles were eligible. The Deyo variant of Charlson was the most commonly referred comparator followed by the Elixhauser measure. Compared with baseline variables such as age and sex, comorbidity adjustment methods seem to better predict long-term than short-term mortality and Elixhauser seems to be the best predictor for this outcome. For short-term mortality, however, recalibration giving empirical weights seems more important than the choice of comorbidity measure.

CONCLUSIONS

The performance of a given comorbidity measure depends on the patient group and outcome. In general, the Elixhauser index seems the best so far, particularly for mortality beyond 30 days, although several newer, more inclusive measures are promising.

A comparison of comorbidities obtained from hospital administrative data and medical charts in older patients with pneumonia

Background

The use of comorbidities in risk adjustment for health outcomes research is frequently necessary to explain some of the observed variations. Medical charts reviews to obtain information on comorbidities is laborious. Increasingly, electronic health care databases have provided an alternative for health services researchers to obtain comorbidity information. However, the rates obtained from databases may be either over- or under-reported. This study aims to (a) quantify the agreement between administrative data and medical charts review across a set of comorbidities; and (b) examine the factors associated with under- or over-reporting of comorbidities by administrative data.

Methods

This is a retrospective cross-sectional study of patients aged 55 years and above, hospitalized for pneumonia at 3 acute care hospitals. Information on comorbidities were obtained from an electronic administrative database and compared with information from medical charts review. Logistic regression was performed to identify factors that were associated with under- or over-reporting of comorbidities by administrative data.

Results

The prevalence of almost all comorbidities obtained from administrative data was lower than that obtained from medical charts review. Agreement between comorbidities obtained from medical charts and administrative data ranged from poor to very strong (kappa 0.01 to 0.78). Factors associated with over-reporting of comorbidities were increased length of hospital stay, disease severity, and death in hospital. In contrast, those associated with under-reporting were number of comorbidities, age, and hospital admission in the previous 90 days.

Conclusions

The validity of using secondary diagnoses from administrative data as an alternative to medical charts for identification of comorbidities varies with the specific condition in question, and is influenced by factors such as age, number of comorbidities, hospital admission in the previous 90 days, severity of illness, length of hospitalization, and whether inhospital death occurred. These factors need to be taken into account when relying on administrative data for comorbidity information.

Hospital quality: a PRIDIT approach

BACKGROUND

Access to high quality medical care is an important determinant of health outcomes, but the quality of care is difficult to determine.

OBJECTIVE

To apply the PRIDIT methodology to determine an aggregate relative measure of hospital quality using individual process measures.

DESIGN

Retrospective analysis of Medicare hospital data using the PRIDIT methodology.

SUBJECTS

Four-thousand-two-hundred-seventeen acute care and critical access hospitals that report data to CMS' Hospital Compare database.

MEASURES

Twenty quality measures reported in four categories: heart attack care, heart failure care, pneumonia care, and surgical infection prevention and five structural measures of hospital type.

RESULTS

Relative hospital quality is tightly distributed, with outliers of both very high and very low quality. The best indicators of hospital quality are patients given assessment of left ventricular function for heart failure and patients given beta-blocker at arrival and patients given beta-blocker at discharge for heart attack. Additionally, teaching status is an important indicator of higher quality of care.

CONCLUSIONS

PRIDIT allows us to rank hospitals with respect to quality of care using process measures and demographic attributes of the hospitals. This method is an alternative to the use of clinical outcome measures in measuring hospital quality. Hospital quality measures should take into account the differential value of different quality indicators, including hospital "demographic" variables.

In search of the perfect comorbidity measure for use with administrative claims data: does it exist?

BACKGROUND

Numerous measures of comorbidity have been developed for health services research with administrative claims.

OBJECTIVE

We sought to compare the performance of 4 claims-based comorbidity measures.

RESEARCH DESIGN AND SUBJECTS

We undertook a retrospective cohort study of 5777 Medicare beneficiaries ages 66 and older with stage III colon cancer reported to the Surveillance, Epidemiology, and End Results Program between January 1, 1992 and December 31, 1996.

MEASURES

Comorbidity measures included Elixhauser's set of 30 condition indicators, Klabunde's outpatient and inpatient indices weighted for colorectal cancer patients, Diagnostic Cost Groups, and the Adjusted Clinical Group (ACG) System. Outcomes included receipt of adjuvant chemotherapy and 2 year noncancer mortality.

RESULTS

For all measures, greater comorbidity significantly predicted lower receipt of chemotherapy and higher noncancer death. Nested logistic regression modeling suggests that using more claims sources to measure comorbidity generally improves the prediction of chemotherapy receipt and noncancer death, but depends on the measure type and outcome studied. All 4 comorbidity measures significantly improved the fit of baseline regression models for both chemotherapy receipt (baseline c-statistic 0.776; ranging from 0.779 after adding ACGs and Klabunde to 0.789 after Elixhauser) and noncancer death (baseline c-statistic 0.687; ranging from 0.717 after adding ACGs to 0.744 after Elixhauser).

CONCLUSIONS

Although some comorbidity measures demonstrate minor advantages over others, each is fairly robust in predicting both chemotherapy receipt and noncancer death. Investigators should choose among these measures based on their availability, comfort with the methodology, and outcomes of interest.

Assessment of Long-Term Physical Function in Acute Respiratory Distress Syndrome (ARDS) Patients

OBJECTIVE

It is often important to adjust for the effect of comorbid diseases on patient outcomes. This study compares the association between physical function in acute respiratory distress syndrome patients with scores on two comorbidity indices, the Charlson Comorbidity Index, designed to predict mortality, and the Functional Comorbidity Index (FCI), which was designed to predict physical function.

DESIGN

This is a prospective, longitudinal, observational study. A total of 73 survivors of acute respiratory distress syndrome were contacted at 3, 6, and 12 mos. Patient comorbidity was evaluated with the Charlson Comorbidity Index and the FCI. Physical function was measured using the Medical Outcomes Study 36-Item Short Form Health Survey Physical Function Subscale and the Physical Component Subscale scores.

RESULT

Mean FCI and Charlson Comorbidity Index scores correlated fairly strongly (Spearman rho = 0.62, P < 0.001). FCI, but not the Charlson Comorbidity Index, scores correlated with the Physical Function Subscale and Physical Component Subscale scores. After controlling for other potentially confounding variables such as age and severity of illness through regression analysis, the FCI score was still significantly associated with both Physical Function Subscale and Physical Component Subscale scores.

CONCLUSIONS

The FCI is a better method of measuring comorbidity with physical function as the outcome. This study illustrates the importance of choosing the most appropriate comorbidity index for the outcome of interest.

Is early- and late-onset depression after acute myocardial infarction associated with long-term survival in older adults? A population-based study

BACKGROUND

Early-onset depression after acute myocardial infarction (AMI) affects short-term survival in clinical samples of patients. There is no information on the impact of early-onset depression or late-onset depression on long-term survival.

OBJECTIVE

To investigate the impact of early- and late-onset depression on survival using administrative data.

METHODS

A historical inception cohort design was used, commencing in 1994 with up to eight years of follow-up. A province-wide administrative data set from British Columbia was used to select the cohort and construct the variables. Data regarding hospitalizations, physician visits and prescription drugs were available. All individuals 66 years of age and older who had an AMI in 1994 or 1995 were selected (n=4874). Individuals were categorized as depressed, possibly depressed or not depressed based on physician or hospital visits indicating depression as a diagnosis and/or prescriptions for antidepressants. Early-onset depression was assessed during the first six months post-AMI, and late-onset depression was assessed between six months and five years post-AMI. All-cause mortality up to eight years post-AMI was the outcome.

RESULTS

Both early- and late-onset depression were associated with long-term mortality. The hazard ratio was 1.34 (95% CI 1.04 to 1.73) for early-onset depression and 1.79 (95% CI 1.38 to 2.35) for late-onset depression.

CONCLUSIONS

Both early- and late-onset depression post-AMI were significantly associated with mortality up to eight years post-AMI. Depression is a strong independent predictor of post-AMI mortality in older adults.

Public reporting of hospital outcomes based on administrative data: risks and opportunities

In the wake of findings from the Bundaberg Hospital and Forster inquiries in Queensland, periodic public release of hospital performance reports has been recommended. A process for developing and releasing such reports is being established by Queensland Health, overseen by an independent expert panel. This recommendation presupposes that public reports based on routinely collected administrative data are accurate; that the public can access, correctly interpret and act upon report contents; that reports motivate hospital clinicians and managers to improve quality of care; and that there are no unintended adverse effects of public reporting. Available research suggests that primary data sources are often inaccurate and incomplete, that reports have low predictive value in detecting "outlier" hospitals, and that users experience difficulty in accessing and interpreting reports and tend to distrust their findings.

Specific comorbidity risk adjustment was a better predictor of 5-year acute myocardial infarction mortality than general methods

OBJECTIVE

To compare methods of risk adjustment in a population of individuals with acute myocardial infarction (AMI), in order to assist clinicians in assessing patient prognosis.

STUDY DESIGN AND SETTING

A historical inception cohort design was established, with follow-up of <or=5 years. A province-wide population-based administrative dataset from British Columbia, Canada, was used to select the cohort and construct variables. All individuals aged >or=66 years who had an AMI in 1994 or 1995 were selected (n = 4,874). The three risk-adjustment methods were the Ontario AMI prediction rule (OAMIPR), the D'Hoore adaptation of the Charlson Index, and the total number of distinct comorbidities. Logistic regression models were built including each of the adjustment methods, age, sex, socioeconomic status, previous AMI, and cardiac procedures at time of AMI.

RESULTS

The OAMIPR had the highest C-statistic and R(2).

CONCLUSION

Clinicians are advised to consider the specific comorbidities that are present, not merely their number, and those that emerge over time, not merely those present at the time of the infarct.

Comorbidity Indices to Predict Mortality From Medicare Data

BACKGROUND

By accounting for level of comorbidity, risk-adjustment models should quantify the risk of death. How accurately comorbidity indices predict risk of death in Medicare beneficiaries with atrial fibrillation is unclear.

OBJECTIVES

We sought to quantify how well 3 administrative-data based comorbidity indices (Deyo, Romano, and Elixhauser) predict mortality compared with a chart-review index.

DESIGN

We undertook a retrospective cohort study using Medicare claim data (1995-1999) and medical record review.

SUBJECTS

We studied Medicare beneficiaries (n = 2728; mean age = 77) with a common cardiac dysrhythmia, atrial fibrillation.

MEASURES

The outcome was time to death with the accuracy of the comorbidity indices measured by the c-statistic.

RESULTS

Correlation between Deyo and Romano indices was strong, but weak between them and the other indices. Prevalence of many comorbidity conditions varied with different indices. Compared with demographic data alone (c = 0.64), all comorbidity indices predicted death significantly (P < 0.001) better: the c index was 0.76 for Deyo, 0.78 for Romano, 0.76 for Elixhauser, and 0.75 for medical record review. The 95% confidence intervals of the c-statistic for the 4 indices overlapped with one another. Key comorbidity conditions for death included metastatic cancer, neuropsychiatric disease, heart failure, and liver disease.

CONCLUSION

The predictive accuracy of 3 administrative-data based indices was similar and comparable with chart-review.

A multipurpose comorbidity scoring system performed better than the Charlson index

BACKGROUND AND OBJECTIVES

To develop a comorbidity scoring system that out-performs the Charlson index.

METHODS

Population-based cohorts of medical (n=326,456), procedural (n=349,686), and psychiatric (n=16,895) inpatients in Western Australia were followed for 1-year mortality, 30-day readmissions, and length of stay (LOS) using data linkage. Conditions were identified at index admission and over the preceding 12 months. A Multipurpose Australian Comorbidity Scoring System (MACSS) was developed, based on the most frequent 102 comorbid conditions associated with a rate ratio (RR) > or = 1.1 of death or readmission or a LOS difference > or =0.5 days. The performance of MACSS and the Charlson index in predicting mortality, readmission, and LOS, and in controlling confounding by comorbidity, was compared in five test scenarios involving asthma, myocardial infarction, mastectomy, transurethral prostatectomy, and major depressive illness.

RESULTS

MACSS performed better than the Charlson index on all three outcomes in all five clinical groups. It reduced the failure of the Charlson index to discriminate on mortality and readmission outcomes by 5-40%, improved R(2) in LOS models by up to fourfold and often doubled the correction of originally confounded effect measures.

CONCLUSION

The use of the MACSS and similar alternatives to the Charlson index are a new methodologic standard for adjustment of comorbidity risk.

Care following acute myocardial infarction in the Veterans Administration Medical Centers: a comparison with Medicare

OBJECTIVE

To compare patients treated for acute myocardial infarction (AMI) in a Veterans Health Administration (VHA) facility to similar patients treated under Medicare.

DATA SOURCES

Administrative data on 13,129 elderly male veterans hospitalized for AMI in a VHA facility between October 1, 1996, and September 30, 1999, and a matched set of male Medicare beneficiaries with AMI treated in a non-VHA facility during the same time period.

STUDY DESIGN

We conducted a retrospective cohort study using propensity score methods to identify a matched set of male elderly AMI patients treated either in a VHA facility or in a non-VHA facility under Medicare. We compared the two groups of patients according to characteristics of the admitting hospital, distances traveled for care, the use of invasive procedures, and mortality. We assessed the robustness of our conclusions to biases arising from unmeasured confounders using sensitivity analyses.

PRINCIPAL FINDINGS

VHA patients were significantly less likely than Medicare beneficiaries to be admitted to high-volume facilities (for example, 25 percent versus 46 percent in 1999, p<0.001) or facilities with the capability to perform invasive cardiac procedures. Compared to Medicare patients, VHA patients traveled almost twice as far to their admitting hospital. The VHA patients were significantly less likely to undergo coronary angiography or revascularization in the 30 days following their AMI (p<0.001 for all comparisons). Veterans treated in the VHA had significantly higher mortality at one-year in all years studied (for example, 35.2 percent versus 30.6 percent in 1999). The proportion of elderly VHA patients admitted to high-volume facilities increased and 30-day mortality rates decreased between 1997 and 1999. Using sensitivity analyses to assess possible effects of unmeasured confounders, we could explain some but not all of the observed mortality differences.

CONCLUSIONS

We observed differences in the way care for AMI patients was structured, in the use of invasive therapies, and in long term mortality between patients treated in VHA hospitals and those treated in non-VHA facilities under Medicare. Future research should focus on explanations for the differences between the two systems and for the reduction in short-term mortality among VHA patients. Further study of these differences both between and within the systems of care may help identify cost-effective strategies to improve care in both sectors.

Comparing the Health Status of VA and Non-VA Ambulatory Patients

The purpose of this study was to compare health status and disease profiles of ambulatory patients in specific Veterans Administration (VA) and civilian healthcare settings. A random sample of 2425 male veterans seeking care at 4 Boston-area VA outpatient clinics, who took part in the Veterans Health Study (VHS) in 1993-1995, were compared to 1318 male patients seeking civilian outpatient care in 3 major metropolitan areas covered in the Medical Outcomes Study (MOS) in 1986. The MOS sampled patients who had 1 of 5 conditions--hypertension, noninsulin-dependent diabetes, recent myocardial infarction, congestive heart failure, or depression. These 2 samples were age adjusted and compared in terms of the SF-36 Health Status/Quality of Life measures, and a list of 100 clinical variables (diagnostic, symptom, and medical event reports) collected with comparable instruments by a trained clinical observer. Individual odds ratios (VHS to MOS) were calculated for each measure and clinical variables. SF-36 measures of patient health in the VHS were lower than those in the MOS by more than one half of a standard deviation (SD) on 4 of 8 scales, by more than one quarter of a SD on the other 4, by 58% of a SD on the physical health summary scale, and by 37% of a SD on the mental health summary scale (P < .0001 in all cases). The median odds ratio was 2.2 among the SF-36 scales and 1.9 among clinical variables. Outpatients in the 4 VA clinics had more than twice the illness burden than did patients in the MOS. Current economic condition and service-connected disability explain most, if not all, of the differences. The differences were clinically and socially meaningful and would be consistent with substantially higher expected healthcare use.

Assessing comorbidity using claims data: an overview

Comorbidity, additional disease beyond the condition under study that increases a patient's total burden of illness, is one dimension of health status. For investigators working with observational data obtained from administrative databases, comorbidity assessment may be a useful and important means of accounting for differences in patients' underlying health status. There are multiple ways of measuring comorbidity. This paper provides an overview of current approaches to and issues in assessing comorbidity using claims data, with a particular focus on established indices and the SEER-Medicare database. In addition, efforts to improve measurement of comorbidity using claims data are described, including augmentation of claims data with medical record, patient self-report, or health services utilization data; incorporation of claims data from sources other than inpatient claims; and exploration of alternative conditions, indices, or ways of grouping conditions. Finally, caveats about claims data and areas for future research in claims-based comorbidity assessment are discussed. Although the use of claims databases such as SEER-Medicare for health services and outcomes research has become increasingly common, investigators must be cognizant of the limitations of comorbidity measures derived from these data sources in capturing and controlling for differences in patient health status. The assessment of comorbidity using claims data is a complex and evolving area of investigation.

A comparison of Bayesian methods for profiling hospital performance

There is a growing interest in the use of Bayesian methods for profiling institutional performance. In the literature, several studies have compared different frequentist methods for classifying hospitals as performance outliers. The purpose of this study was to compare 4 different Bayesian methods for classifying hospitals as outcomes outliers, using 30-day hospital-level mortality rates for a cohort of acute myocardial infarction patients as a test case. The 1st Bayesian method involved determining the probability that a hospital's mortality rare for an average patient exceeded a specified threshold. The 2nd method involved ranking hospitals according to their mortality rate for an average patient. The 3rd method involved determining the probability that a hospital's standardized mortality ratio exceeded a specified threshold. The 4th method involved ranking hospitals according to their standardized mortality ratio. In most of the scenarios examined, there was only marginal agreement between the different methods. In only 4 of 19 comparisons, was there good agreement between the different methods (0.40 < or = kappa < or = 0.75). Methods based on ranking institutions were relatively insensitive to differences between hospitals. These inconsistencies raise questions about the choice of methods for classifying hospital performance, and they suggest a need for urgent research into which methods are best able to discriminate between institutions and which are most meaningful to decision makers.

Accuracy of administrative data to assess comorbidity in patients with heart disease. an Australian perspective

The objective of this study was to determine the accuracy of administrative data (by use of hospital discharge codes) for measuring comorbidity in patients with heart disease. One thousand seven hundred and sixty-five medical records of subjects admitted to hospital for AMI, unstable angina, angina pectoris, chronic IHD or heart failure were reviewed. The number and types of comorbidities were determined from the medical records (regarded as the "gold standard"). These were compared with the 10 discharge codes obtained from the hospital administrative records (referred to as the "administrative data"). The rate of false-negative and false-positive comorbidity diagnoses were determined. Twenty of the 21 comorbidities studied were underreported in the administrative data. For these 20 comorbidities, the median false-negative rate was 49.5% and ranged from 11% for diabetes to 100% for dementia. False-positive rates were low, less than 1.5%, except for chronic arrythmia (4.8%) and hypertension (4.2%). Mean percent agreement was high, ranging from 88% for hypertension to 100% for AIDS/HIV. Administrative data based on hospital discharge codes consistently underestimate the presence of comorbid conditions in our population. This has implications for administrators when estimating mortality, length of stay and disability. Researchers also need to be aware when using administrative data based on hospital discharge codes to assess subject's comorbidities that they may be widely underreported.

Development and validation of the Ontario acute myocardial infarction mortality prediction rules

OBJECTIVES

To develop and validate simple statistical models that can be used with hospital discharge administrative databases to predict 30-day and one-year mortality after an acute myocardial infarction (AMI).

BACKGROUND

There is increasing interest in developing AMI "report cards" using population-based hospital discharge databases. However, there is a lack of simple statistical models that can be used to adjust for regional and interinstitutional differences in patient case-mix.

METHODS

We used linked administrative databases on 52,616 patients having an AMI in Ontario, Canada, between 1994 and 1997 to develop logistic regression statistical models to predict 30-day and one-year mortality after an AMI. These models were subsequently validated in two external cohorts of AMI patients derived from administrative datasets from Manitoba, Canada, and California, U.S.

RESULTS

The 11-variable Ontario AMI mortality prediction rules accurately predicted mortality with an area under the receiver operating characteristic (ROC) curve of 0.78 for 30-day mortality and 0.79 for one-year mortality in the Ontario dataset from which they were derived. In an independent validation dataset of 4,836 AMI patients from Manitoba, the ROC areas were 0.77 and 0.78, respectively. In a second validation dataset of 112,234 AMI patients from California, the ROC areas were 0.77 and 0.78 respectively.

CONCLUSIONS

The Ontario AMI mortality prediction rules predict quite accurately 30-day and one-year mortality after an AMI in linked hospital discharge databases of AMI patients from Ontario, Manitoba and California. These models may also be useful to outcomes and quality measurement researchers in other jurisdictions.

A comparison of a Bayesian vs. a frequentist method for profiling hospital performance

The objective of this study was to compare the classification of hospitals as outcomes outliers using a commonly implemented frequentist statistical approach vs. an implementation of Bayesian hierarchical statistical models, using 30-day hospital-level mortality rates for a cohort of acute myocardial infarction patients as a test case. For the frequentist approach, a logistic regression model was constructed to predict mortality. For each hospital, a risk-adj usted mortality rate was computed. Those hospitals whose 95% confidence interval, around the risk-adjusted mortality rate, excludes the mean mortality rate were classified as outliers. With the Bayesian hierarchical models, three factors could vary: the profile of the typical patient (low, medium or high risk), the extent to which the mortality rate for the typical patient departed from average, and the probability that the mortality rate was indeed different by the specified amount. The agreement between the two methods was compared for different patient profiles, threshold differences from the average and probabilities. Only marginal agreement was shown between the Bayesian and frequentist approaches. In only five of the 27 comparisons was the kappa statistic at least 0.40. The remaining 22 comparisons demonstrated only marginal agreement between the two methods. Within the Bayesian framework, hospital classification clearly depended on patient profile, threshold and probability of exceeding the threshold. These inconsistencies raise questions about the validity of current methods for classifying hospital performance, and suggest a need for urgent research into which methods are most meaningful to clinicians, managers and the general public.

Death and readmission in the year after hospital admission with cardiovascular disease: the Hunter Area Heart and Stroke Register

OBJECTIVES

To compare outcomes one year after hospital admission for patients initially discharged with a diagnosis of acute myocardial infarction (AMI), other ischaemic heart disease (other IHD), congestive heart failure (CHF) or stroke.

DESIGN

Cohort study.

SETTING

Hunter Area Heart and Stroke Register, which registers all patients admitted with heart disease or stroke to any of the 22 hospitals in the Hunter Area Health Service in New South Wales.

PATIENTS

4981 patients with AMI, other IHD, CHF or stroke admitted to hospital as an emergency between 1 July 1995 and 30 June 1997 and followed for at least one year.

MAIN OUTCOME MEASURES

Death from any cause or emergency hospital readmission for cardiovascular disease.

RESULTS

In-hospital mortality varied from 1% of those with other IHD to 22% of those with stroke. Almost a third of all patients discharged alive (and 38% of those aged 70 or more) had died or been readmitted within one year. This varied from 22% of those with stroke to 49% of those with CHF. The causes of death and readmission were from a spectrum of cardiovascular disease, regardless of the cause of the original hospital admission.

CONCLUSIONS

Data from this population register show the poor outcome, especially with increasing age, among patients admitted to hospital with cardiovascular disease. This should alert us to determine whether optimal secondary prevention strategies are being adopted among such patients.

Impact of availability of hospital-based invasive cardiac services on racial differences in the use of these services

BACKGROUND

Reports indicate that black patients are less likely than white patients to receive invasive cardiac services after hospitalization for acute myocardial infarction (AMI). There is still uncertainty as to why racial differences exist and how they affect patient outcomes. This is the first study to focus on the availability of invasive cardiac services and racial differences in procedure use. Study objectives were to (1) document whether racial differences existed in the use of invasive cardiac procedures, (2) study whether these racial differences were related to availability of hospital-based invasive cardiac services at first admission for AMI, and (3) determine whether there were racial differences in long-term mortality rates.

METHODS

A historical cohort study was conducted with discharge records from all acute care hospitals in New Jersey for 1993 linked to death certificate records for 1993 and 1994. There were 13,690 black and white New Jersey residents hospitalized with primary diagnosis of AMI. Use of cardiac catheterization within 90 days, revascularization within 90 days (percutaneous transluminal coronary angioplasty [PTCA] or coronary artery bypass graft surgery [CABG]), and death within 1 year after admission for AMI were the main outcome measures. Patterns for PTCA and CABG as separate outcomes were also studied. Hospital-based cardiac services available were described as no invasive cardiac services, catheterization only, or PTCA/CABG. To account for payer status and comorbidity differences, patients 65 years and older with Medicare coverage were analyzed separately from those younger than 65 years.

RESULTS

Black patients aged 65 and older were generally less likely to receive catheterization and revascularization than white patients, regardless of facilities available at first admission. For patients younger than 65 years, the greatest differences between black and white patients in catheterization and PTCA/CABG use within 90 days after AMI occurred when no hospital-based invasive cardiac services were available. However, use of invasive cardiac procedures within 90 days after AMI was substantially increased if the first hospital offered catheterization only or PTCA/CABG services, among all patients, especially among blacks younger than age 65. No significant racial differences or interactions with available services were found in 1-year mortality rates.

CONCLUSIONS

Availability of invasive cardiac services at first hospitalization for AMI was associated with increased procedure use for both races. However, use of invasive cardiac procedures was generally lower for black patients than for white patients, regardless of services available. Long-term mortality rates after hospitalization for AMI did not differ between blacks and whites.

Misclassification and under-reporting of acute myocardial infarction by elderly persons: implications for community-based observational studies and clinical trials

We investigated the accuracy of self-report of hospitalization for acute myocardial infarction (MI) by elderly persons in a community-based prospective study. Among 3809 persons aged 65 years or older followed up for 6 years, self-reported hospitalization for MI was validated by review of primary records and Medicare diagnoses. Among 147 who self-reported MI and for whom hospital records were available, the diagnosis was confirmed in 79 (54%). Myocardial infarction was not a reason for hospitalization among the remaining 68 participants; misclassification with other cardiovascular diagnoses was common. Medicare diagnosis correlated well with primary hospital records. Using Medicare diagnoses as the standard, the diagnosis of MI was confirmed in 53% of self-reports; the sensitivity and specificity of self-report were 51% and 98%, respectively. False-negative reporting was common because only half of hospitalizations for MI were reported. Self-report of hospitalization for MI by elderly persons in the community may be unreliable for ascertaining trends in cardiovascular diseases.

Comparing AMI mortality among hospitals in patients 65 years of age and older: evaluating methods of risk adjustment

BACKGROUND

Interest in the reporting of risk-adjusted outcomes for patients with acute myocardial infarction is growing. A useful risk-adjustment model must balance parsimony and ease of data collection with predictive ability.

METHODS AND RESULTS

From our analysis of 82 359 patients >/=65 years of age admitted with acute myocardial infarction to 2401 hospitals, we derived a parsimonious model that predicts 30-day mortality. The model was validated on a similar group of 78 699 patients from 2386 hospitals. Of the 73 candidate predictor variables examined, 7 variables describing patient characteristics on arrival were selected for inclusion in the final model: age, cardiac arrest, anterior or lateral location of myocardial infarction, systolic blood pressure, white blood cell count, serum creatinine, and congestive heart failure. The area under the receiver-operating characteristic curve for the final model was 0.77 in the derivation cohort and 0.77 in the validation cohort. The rankings of hospitals by performance (in deciles) with this model were most similar to a comprehensive 27-variable model based on medical chart review and least similar to models based on administrative billing codes.

CONCLUSIONS

A simple 7-variable risk model performs as well as more complex models in comparing hospital outcomes for acute myocardial infarction. Although there is a continuing need to improve methods of risk adjustment, our results provide a basis for hospitals to develop a simple approach to compare outcomes.

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.

Use of cardiac procedures and outcomes in elderly patients with myocardial infarction in the United States and Canada

BACKGROUND

Acute myocardial infarction is a leading cause of morbidity and mortality in the United States and Canada. We performed a population-based study to compare the use of cardiac procedures and outcomes after acute myocardial infarction in elderly patients in the two countries.

METHODS

We compared the use of invasive cardiac procedures and the mortality rates among 224,258 elderly Medicare beneficiaries in the United States and 9444 elderly patients in Ontario, Canada, each of whom had a new acute myocardial infarction in 1991.

RESULTS

The U.S. patients were significantly more likely than the Canadian patients to undergo coronary angiography (34.9 percent vs. 6.7 percent, P< 0.001), percutaneous transluminal coronary angioplasty (11.7 percent vs. 1.5 percent, P<0.001), and coronary-artery bypass surgery (10.6 percent vs. 1.4 percent, P<0.001) during the first 30 days after the index infarction. These differences in the use of cardiac procedures narrowed but persisted through 180 days of follow-up. The 30-day mortality rates were slightly but significantly lower for the U.S. patients than for the Canadian patients (21.4 percent vs. 22.3 percent, P=0.03). However, the one-year mortality rates were virtually identical (34.3 percent in the United States vs. 34.4 percent in Ontario, P= 0.94).

CONCLUSIONS

Short-term mortality after an acute myocardial infarction was slightly lower in the United States than in Ontario, but these differences did not persist through one year of follow-up. The strikingly higher rates of use of cardiac procedures in the United States, as compared with Canada, do not appear to result in better long-term survival rates for elderly U.S. patients with acute myocardial infarction.

Outcomes of acute myocardial infarction in the Department of Veterans Affairs: does regionalization of health care work?

OBJECTIVES

This study examines the association between the regional availability of cardiac technology and outcomes of care for patients admitted to Department of Veterans Affairs (VA) hospitals. Patients using the VA regional medical system initially are admitted to a hospital with or without the on-site availability of technology-intensive cardiac services.

METHODS

The authors identified male veterans (n = 24,229) discharged from VA hospitals with a primary diagnosis of acute myocardial infarction (AMI) from January 1, 1988 through December 31, 1990. Analyses of mortality up to 2 years after AMI and the use of cardiac procedures were stratified by the type of VA hospitals to which patients initially were admitted. Logistic regression models adjusted for age, race, marital status, hospitalization in previous year, comorbidities, cardiac complications coded, and year of AMI.

RESULTS

Adjusted mortality was significantly higher for patients initially admitted to hospitals without on-site cardiac technology at: 2 days (odds ratio [OR] 0.70; 95% confidence interval [CI] 0.62-0.81), 90 days (OR 0.78; 95% CI 0.73-0.85); 1 year (OR 0.87, 95% CI 0.81-0.93); and 2 years (OR 0.86, 95% CI 0.81-0.92) compared with hospitals with on-site cardiac technology (ie, coronary angioplasty and cardiac surgery facilities). Patients initially admitted to hospitals without on-site cardiac technology also were less likely to undergo cardiac procedures than patients admitted to hospitals with on-site cardiac technology.

CONCLUSIONS

The regional distribution of cardiac technology may restrict patient access to technology-intensive services and to "equally good medical care." Policies that promote regionalization of medical services should consider carefully the distribution of benefits and burdens to patients.