Methods
The data for this project was obtained from the AHRQ (Agency for Healthcare Research and Quality) HCUP (Healthcare Cost and Utilization Project) Nationwide/National Inpatient Sample (NIS) for the years 2010-2015. This is the largest all-payer database in the United States covering about 20% of all hospitalizations in the acute care community hospitals that is up to 8 million for each year. The detailed information about the elements of the NIS database is available athttps://www.hcup-us.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp.
The study population included patients aged ≥70 years old who were hospitalized for coronary artery disease (CAD) and underwent CABG during seven days after hospitalization. To select these patients, we used the ICD-9-CM (International Classification of Diseases, Ninth Revision, and Clinical Modification) diagnosis codes 41001, 41011, 41021, 41031, 41041, 41051, 41061, 41071, 41081, 41091. The ICD code for CAD is 41401, and CCS code for CABG is 44. The ICD and CCS code was used for the principal diagnosis to limit the sample to patients aged 70-89 years old who were electively (ELECTIVE=1) or non-electively (ELECTIVE=0) hospitalized for MI (initial episode) or CAD (principal diagnosis for both) and underwent CABG (any procedure position) during 7 days after hospitalization.
Demographic information was collected for each patient. Pre-, intra-, and post-operative findings were identified through a thorough review of the data. Postoperative complications, morbidity, and survival information were also reviewed. Patients that are ≥70 years who underwent CABG were selected from the Nationwide/National Inpatient Sample from 2010 to 2015 using ICD-9-CM diagnosis and procedure codes. The patients who were ≥70–79 years old were compared to patients aged ≥80–89 years old to determine if the age difference of the patients had an impact on surgical outcomes. In addition, the gender of the patients 80-89 years old were compared to determine the influence of patient gender on surgical outcomes. The rates of postoperative complications, mortality, hospital length of stay (LOS) and cost were compared using the Chi-square test, multivariable logistic regression analysis, and Wilcoxon rank sum test.