December 08, 2015
1 min read
Save

National spine surgery databases may have large amounts of missing data

SAN DIEGO — A sample of spine surgery patients from a national database showed large amounts of missing patient data, which can significantly affect the results of spine studies performed using the database study set, according to the results of a study presented here.

Nathaniel T. Ondeck, BS, and colleagues identified a significant amount of missing data from a subset of spinal surgery patients in the American College of Surgeons National Quality Improvement Program (ACS-NSQIP). Ondeck said researchers should be aware of the missing data when using the database for research.

“There is a large amount of missing data for some of the variables, which we later found after talking from NSQIP was due to smaller institutions not being required to report all the data. In addition, we found that there can be drastically different results with different methods of managing the missing data. In conclusion, we suggest that clinicians be aware of this and researchers be forward with what approach they use,” Ondeck said at the meeting.

Ondeck and colleagues identified 88,471 patients in the ACS-NSQIP database who underwent spine surgery from 2005 to 2013. Researchers analyzed the missing demographics, perioperative lab values and comorbidity data for each patient.

Nineteen comorbidity variables had 65% missing data. Among these comorbidities were alcohol use, previous cardiac surgery, angina, current pneumonia status, coma, history of myocardial infarction and several others, according to Ondeck.  

Missing data for lab values ranged from 8.85% for hematocrit to 80.13% for prothrombin time, Ondeck said. – by Robert Linnehan

Reference:

Basques BA, et al. Presentation #88. Presented at: Cervical Spine Research Society Annual Meeting; Dec. 3-5, 2015; San Diego.

Disclosure: Ondeck reports no relevant financial disclosures.