Premium
Do‐Not‐Resuscitate Orders and Resource Use in Patients with Pancreatic Cancer
Author(s) -
Hao Q.,
Segel J.,
Gusani N.,
Hollenbeak C.
Publication year - 2020
Publication title -
health services research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.706
H-Index - 121
eISSN - 1475-6773
pISSN - 0017-9124
DOI - 10.1111/1475-6773.13348
Subject(s) - medicine , healthcare cost and utilization project , emergency medicine , pancreatic cancer , do not resuscitate order , logistic regression , observational study , health care , cancer , intensive care medicine , economics , economic growth
Research Objective A do‐not‐resuscitate (DNR) order informs health care providers of patients’ preference not to receive cardiopulmonary resuscitation if their breathing stops or their heart stops beating. The use of DNR orders has risen over the past few decades among older people and patients with terminal illnesses such as pancreatic cancer. However, evidence about whether DNR orders are associated with changes in utilization or cost savings is mixed. Here, we focus on pancreatic cancer due to its low survival rate resulting from late stage of diagnosis. The objective of this study was to evaluate whether DNR status had a significant association with mortality, length of stay (LOS), and costs of an inpatient stay among pancreatic cancer patients. Study Design This was an observational study using 2011‐2013 inpatient data from the Healthcare Cost and Utilization Project’s (HCUP) National Inpatient Sample (NIS). Patients with pancreatic cancer were identified using a primary ICD‐9‐CM diagnosis code of 52.0‐52.9. In addition, DNR status was identified using ICD‐9‐CM code of V49.86 (DNR status). This study compared three outcomes: mortality, LOS, and costs between patients with and without a DNR. Total costs were estimated using the cost‐to‐charge ratio and then inflated to 2015 USD using the Consumer Price Index. Univariate statistical analysis included t tests for continuous variables and chi‐squared tests for binary and categorical variables, respectively. We used logistic regression to estimate the association between of DNR and mortality, after controlling for other covariates. Total costs and LOS were modeled using generalized linear regression models and assumed gamma family of distributions and a log link function. To better control for covariate imbalance between patients with and without a DNR, we used propensity score matching. Population Studied A total of 7850 patients were admitted to a hospital with pancreatic cancer who were over 18 years old between 2011 and 2013, and DNR status was identified in 2.48% (n = 195). Principal Findings Patients with older age, female, higher number of comorbidities, had Medicare as primary insurance, were admitted on an urgent or emergent basis, and treated in rural or urban nonteaching hospitals were more likely to have a DNR order. After controlling for all covariates, patients with a DNR had more than 8 times greater odds of mortality compared to patients without a DNR (OR = 8.14; P < 0.001); however, patients with a DNR had a shorter expected LOS relative to patients without a DNR, but the effect was not statistically significant (OR = −0.89; P = 0.150). Total costs for patients with a DNR were $1075 less than patients without a DNR, but this difference was not statistically significant ( P = 0.62). We obtained qualitatively similar results using propensity score matching. Conclusions The presence of a DNR order among pancreatic cancer patients was significantly associated with increased mortality risk, but not significantly associated with shorter LOS or lower cost of stay. Implications for Policy or Practice Among pancreatic cancer patients, while DNR orders can help patients receive the care they prefer and may also increase quality of end‐of‐life care, they do not appear to lower health care costs for health care providers.