ORIGINAL ARTICLES: ADULT CARDIAC

ADULT CARDIAC SURGERY: The Annals of Thoracic Surgery CME Program is located online at http://www.annalsthoracicsurgery.org/cme/ home. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.

Edward L. Hannan, PhD, Feng Qian, MD, PhD, Michael Pine, MD, MBA, Donald E. Fry, MD, Kay Whitman, BS, and Barbara A. Dennison, MD Department of Health Policy, Management, and Behavior, University at Albany, State University of New York, Albany; Michael Pine Associates, Chicago, Illinois; and New York State Department of Health, Albany, New York

Background. Clinical databases are currently being used for calculating provider risk-adjusted mortality rates for coronary artery bypass grafting (CABG) in a few states and by the Society for Thoracic Surgeons. These databases contain very few laboratory data for purposes of risk adjustment. Methods. For 15 hospitals, New York’s CABG registry data from 2008 to 2010 were linked to laboratory data to develop statistical models comparing risk-adjusted mortality rates with and without supplementary laboratory data. Differences between these two models in discrimination, calibration, and outlier status were compared, and correlations in hospital risk-adjusted mortality rates were examined. Results. The discrimination of the statistical models was very similar (c [ 0.785 for the registry model and 0.797 for the registry/laboratory model, p [0.63). The correlation between hospital risk-adjusted mortality rates

by use of the two models was 0.90. The registry/laboratory model contained three additional laboratory variables: alkaline phosphatase (ALKP), aspartate aminotransferase (AST), and prothrombin time (PT). The registry model yielded one hospital with significantly higher mortality than the statewide average, and the registry/laboratory model yielded no outliers. Conclusions. The clinical models with and without laboratory data had similar discrimination. Hospital riskadjusted mortality rates were essentially unchanged, and hospital outlier status was identical. However, three laboratory variables, ALKP, AST, and PT, were significant independent predictors of mortality, and they deserve consideration of addition to CABG clinical databases.

C

The clinical databases used for reporting and risk-adjusting CABG outcomes contain numerous risk factors, including demographic variables, ventricular function, extent of cardiac disease, and numerous comorbidities. Typically, eight to 12 of these variables prove to be significant independent predictors of shortterm (in-hospital / 30-day) mortality in annual statistical models used for risk adjusting hospital mortality rates. However, with the exception of creatinine level and brain natriuretic peptide (BNP) information, clinical registry data do not include admission laboratory values that could provide more detail about the severity of the comorbidity. The purpose of this study was to determine whether the addition of laboratory data to the NYS clinical database for CABG would identify laboratory variables that are significant independent predictors of short-term (in-hospital / 30-day) mortality and if so, whether the statistical model containing these significant laboratory values would substantially change the hospital-level

oronary artery bypass grafting (CABG) is one of the most studied surgical procedures and is subject to more public reporting of quality of care and risk-adjusted mortality for hospitals and surgeons than any other procedure. CABG mortality outcomes are reported to the public in New York State (NYS), Massachusetts, California, Pennsylvania, New Jersey, and Washington State, and a few additional states are considering reporting CABG outcomes [1–5]. Also, the Society of Thoracic Surgeons (STS) has an extensive database that is used to report outcomes for individual hospitals [6]. All states except NYS and Pennsylvania use the STS clinical database for reporting (NYS has a similar clinical database that predates the STS database, and Pennsylvania uses a clinically enhanced administrative database). Accepted for publication Aug 29, 2014. Address correspondence to Dr Hannan, School of Public Health, State University of New York, University at Albany, One University Place, Rensselaer, NY 12144-3456; e-mail: [email protected].

Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier

(Ann Thorac Surg 2015;99:495–501) Ó 2015 by The Society of Thoracic Surgeons

0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.08.043

ADULT CARDIAC

The Value of Adding Laboratory Data to Coronary Artery Bypass Grafting Registry Data to Improve Models for Risk-Adjusting Provider Mortality Rates

496

HANNAN ET AL LABORATORY DATA IN PREDICTING CABG MORTALITY

ADULT CARDIAC

quality assessments (risk-adjusted mortality rates and outlier status) of hospitals in NYS where CABG is performed.

Material and Methods Endpoint The endpoint in the study was short-term mortality defined as mortality occurring during the index admission for isolated CABG, or within 30 days of the index procedure, regardless of whether or not the patient had been discharged. Deaths after discharge were identified by the use of NYS vital statistics data.

Databases The primary database used for the study was the NYS clinical registry for CABG, the Cardiac Surgery Reporting System (CSRS). This registry contains detailed information on patient demographics, preoperative surgical risk factors, hemodynamic state, left ventricular function, coronary vessels diseased and attempted, adverse events, procedure choices, provider identifiers, discharge status, and in-hospital adverse outcomes. Completeness of data reporting was monitored by matching CSRS to the acute care hospital discharge NYS database, the Statewide Planning and Research Cooperative System (SPARCS), and identifying any cases reported in SPARCS that were not reported in the cardiac registries. The SPARCS contains patient demographics, diagnoses and procedures, admission and discharge dates, and discharge disposition for all patients discharged from nonfederal acute care hospitals in NYS. The CSRS records were matched with SPARCS records by the use of unique hospital identifiers along with patient identifiers and admission, surgical procedures, and discharge dates. To ensure the accuracy of CSRS data, in-hospital outcomes in CSRS were matched to outcomes documented in SPARCS, and the NYS Department of Health utilization review agent audits samples of hospital medical records. Patient identifiers in CSRS were used to link the index procedure to the NYS vital statistics data to identify deaths that occurred after discharge. Laboratory data for 35 tests were obtained from hospitals from their computerized laboratory databases and were appended to the SPARCS data by use of patient identifiers. The appended SPARCS data were then matched to registry data in the same manner as the data completeness monitoring just described.

Hospitals and Patients All 38 NYS hospitals where CABG is performed were invited to participate in the study by extracting and submitting laboratory results for 30 laboratory tests to the Department of Health. Fifteen of the 38 eligible hospitals agreed to participate and submitted electronic laboratory data for patients who had CABG performed between January 1, 2008, and December 31, 2010. These laboratory data were appended on a patient-specific basis to CABG

Ann Thorac Surg 2015;99:495–501

registry data. Laboratory values were documented on the first day of hospitalization so as to reflect the condition of the patient at admission rather than adverse events occurring during the hospitalization. The study was limited to patients undergoing isolated CABG (CABG without other major open heart surgical procedures such as valve operations). Seventeen patients who were identified as being in shock before undergoing isolated CABG were removed from the database because patients with preoperative shock are not contained in public reports.

Laboratory Data Thirty admission laboratory test results were considered for inclusion in the risk-adjustment models. The results for each test were used to assign CABG cases to categories based on ranges that were internally homogeneous with regard to short-term mortality. Cases with missing data were identified, and their relative frequencies and mortality rates were examined for each laboratory variable. Six laboratory tests, wherein more than 90% of patients had a missing laboratory value, were eliminated from consideration as potential risk factors. These variables included amylase, neutrophil band, BNP, lactate dehydrogenase, pro brain natriuretic peptide (proBNP), and troponin T. Another six tests were eliminated from consideration because the mortality rate for missing values was higher than the overall mortality rate for all CABG patients (base units excess, calcium, glucose, hemoglobin A1C, arterial partial carbon dioxide, and arterial potential hydrogen). Four other laboratory variables (bilirubin, creatine phosphokinase, potassium, and troponin I) were eliminated because there was no significant difference in mortality rates across categories. Creatinine and BNP were used as reported in the registry because more complete data were available there. For the remaining tests, potential risk factors were categorized on the basis of on clinical judgment and empiric relationships between test results and shortterm mortality rates. Cases with missing test results were imputed as the median value of the categoric range corresponding to normal test results.

Statistical Analysis In-hospital / 30-day mortality for isolated CABG was examined on a bivariate basis for numerous potential risk factors, including demographics, comorbidities, ventricular function, preprocedural myocardial infarction, and hemodynamic state. Mean mortality rates were computed for ranges of laboratory tests being considered for inclusion in the risk-adjustment models: albumin, alkaline phosphatase (ALKP), aspartate aminotransferase (AST), bicarbonate, blood urea nitrogen (BUN), creatine kinase MB (CKMB), creatine kinase MB fraction (CKMBFR), international normalized ratio (INR), platelet count, partial thromboplastin time (PTT), sodium, white blood cell count, and prothrombin time. Creatinine and BNP three times normal were included as coded in the registry. Variables that were significant (c2

The value of adding laboratory data to coronary artery bypass grafting registry data to improve models for risk-adjusting provider mortality rates.

Clinical databases are currently being used for calculating provider risk-adjusted mortality rates for coronary artery bypass grafting (CABG) in a few...
194KB Sizes 2 Downloads 8 Views