Pharmaceutical Sciences ETDs

Publication Date

Summer 7-19-2017


Background: The International Classification of Diseases, Ninth-Revision-Clinical Modification (ICD-9-CM) for palliative care (V66.7) has been shown to affect risk-adjusted mortality rates. This code can increase the risk of mortality when included in billing data and incorporated into risk-adjustment models (1).

Objectives: The purpose of this study is to examine variations in coding between high-performing (low mortality indices (Observed/Expected)) and low-performing (high mortality indices) hospitals by examining the use of the ICD-9-CM code for palliative care; coding of severe sepsis; and assignment of higher-weighted Medicare-Severity Diagnosis Related Group (MS-DRG) codes.

Methods: Data were obtained from the Vizient™ Clinical Database/Resource Manager (CBD/RM) by permission of Vizient. (All rights reserved.) Adult patients with a present-on-admission diagnosis of severe sepsis and discharged from Vizient-member hospitals during calendar year 2014 were analyzed. Severe sepsis was defined as the presence of an ICD-9-CM code for severe sepsis (995.92); septic shock (785.52); or an infection with organ dysfunction. Hospitals were ranked on their mortality index and divided into quartiles; high-performing and low-performing hospitals were compared. Categorical variables were assessed using chi-square tests; continuous variables were compared using t-tests. The analyses of palliative care code usage and MS-DRG assignment were conducted using logistic regression models.

Results: A total of 352,275 patients representing 249 hospitals met inclusion criteria. There was no statistically significant difference in frequency between high- and low-performing hospitals with which patients were coded with an infection plus organ dysfunction (p = 0.4984) or assignment of higher-weighted MS-DRG codes. Patients with a code for severe sepsis in low-performing hospitals had 0.14 lower odds of utilizing the palliative care code (V66.7) (odds ratio 0.86, 95% CI: 0.78 to 0.94) when compared to high-performing hospitals, after adjusting for patient and hospitalization-related characteristics, to include discharge disposition.

Conclusion: Low-performing hospitals were less likely to have V66.7 code when compared to high-performing hospitals. Patients discharged to hospice were more likely to receive the V66.7 code when compared to those who died in-hospital. This suggests that coding of palliative care may be insufficient when a patient dies in-hospital and that there are opportunities for low-performing hospitals to improve their reported metrics.

Fourth Committee Member

Ning Yan Gu, Ph.D.

First Committee Member (Chair)

Dennis Raisch, Ph.D.

Degree Name

Pharmaceutical Sciences

Second Committee Member

Matthew Borrego, Ph.D.

Level of Degree


Third Committee Member

Melissa Roberts, Ph.D.

Department Name

College of Pharmacy



Document Type



risk adjusted mortality, severe sepsis, palliative care, medical coding, documentation, benchmarking