Journal of Intensive Care Medicine http://jic.sagepub.com/
Comparison of Dexmedetomidine Versus Propofol on Hospital Costs and Length of Stay Asad E. Patanwala and Brian L. Erstad J Intensive Care Med published online 24 July 2014 DOI: 10.1177/0885066614544452 The online version of this article can be found at: http://jic.sagepub.com/content/early/2014/07/24/0885066614544452
Published by: http://www.sagepublications.com
Additional services and information for Journal of Intensive Care Medicine can be found at: Email Alerts: http://jic.sagepub.com/cgi/alerts Subscriptions: http://jic.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - Jul 24, 2014 What is This?
Downloaded from jic.sagepub.com at DREXEL UNIV LIBRARIES on October 7, 2014
Article
Comparison of Dexmedetomidine Versus Propofol on Hospital Costs and Length of Stay
Journal of Intensive Care Medicine 1-5 ª The Author(s) 2014 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0885066614544452 jic.sagepub.com
Asad E. Patanwala, PharmD1, and Brian L. Erstad, PharmD, FASHP, FCCP, MCCM1
Abstract Objective: The objective of this evaluation was to compare total hospital costs and length of stay of critically ill patients who received dexmedetomidine versus propofol for sedation in the intensive care unit (ICU). Methods: This was a retrospective quality improvement evaluation at a tertiary care, academic medical center in the United States. Data were retrieved for patients discharged between April 2012 and June 2013. Patients were included if they were admitted to the ICU, were 18 years of age or older, and received dexmedetomidine or propofol for sedation. Multivariate regression models were developed to determine the association between sedative type and hospital costs, ICU length of stay, and hospital length of stay. Results: The final cohort included 3294 patients. Of these, 2685 received propofol and 609 received dexmedetomidine. The median hospital cost was US$31 041 (interquartile range [IQR] US$17 963-US$57 826) in the propofol group and US$46 716 (IQR US$31 247 to US$85 490) in the dexmedetomidine group (P < .001). The median ICU length of stay was 2 days (IQR 1-6 days) and 4 days (IQR 2-9 days) in the propofol and dexmedetomidine groups, respectively (P < .001). Overall, hospital length of stay was 8 days (IQR 4-15 days) and 9 days (IQR 5-18 days) in the 2 groups, respectively (P < .001). In the multivariate analyses, dexmedetomidine use was associated with increased costs, ICU length of stay, and hospital length of stay (P < .001 for each outcome). Conclusions: In this academic medical center, dexmedetomidine use was associated with higher costs when compared to propofol for sedation in the ICU. Also, use of dexmedetomidine was associated with increased lengths of ICU and hospital stay. Future prospective trials are needed to confirm these findings. Keywords dexmedetomidine, propofol, costs and cost analysis, critical care
Introduction Dexmedetomidine is an a2-adrenergic agonist that has an indication for sedation of intubated and mechanically ventilated patients for a duration of up to 24 hours.1 It has theoretical appeal over more traditional sedative agents such as propofol that is associated with a potentially life-threatening infusion syndrome and benzodiazepines that may be a risk factor for delirium. Current guidelines suggest that dexmedetomidine (or propofol) may be preferred over benzodiazepines in mechanically ventilated patients, since it may improve outcomes such as duration of mechanical ventilation or the occurrence of delirium.2 However, at this time only 1 randomized controlled study has directly compared dexmedetomidine to propofol for prolonged sedation in the intensive care unit (ICU).3 Both sedatives were similar with respect to time at target sedation, duration of mechanical ventilation, mortality, length of ICU, and hospital stay. Dexmedetomidine has a relatively high cost compared to generic versions of propofol and benzodiazepines. Economic
evaluations have shown that dexmedetomidine is associated with decreased overall costs, by decreasing duration of mechanical ventilation and costs associated with episodes of delirium.4,5 Both of these investigations were based on efficacy data derived from a single randomized investigation comparing dexmedetomidine to midazolam.6 In particular, these published economic evaluations do not compare dexmedetomidine and propofol costs based on prescribing patterns outside well-
1
Department of Pharmacy Practice and Science, College of Pharmacy, The University of Arizona, Tucson, AZ, USA
Received February 20, 2014, and in revised form May 20, 2014. Accepted for publication May 21, 2014. Corresponding Author: Asad E. Patanwala, Department of Pharmacy Practice and Science, College of Pharmacy, The University of Arizona, 1295 N Martin Avenue, PO Box 210202, Tuscon, AZ 85721, USA. Email: [email protected]
Downloaded from jic.sagepub.com at DREXEL UNIV LIBRARIES on October 7, 2014
2
Journal of Intensive Care Medicine
controlled trials. The purpose of this investigation was to compare the total costs of dexmedetomidine and propofol regimens when used for sedation of critically ill patients in clinical practice.
missing values; thus, complete case analyses was conducted without the need for imputation.
End Points The primary outcome of interest was total hospital costs. The secondary outcomes were ICU, length of stay, and hospital length of stay. Outcomes were compared between patients who received propofol versus those who received dexmedetomidine.
Materials and Methods Study Population and Setting This was a retrospective quality improvement study conducted at an academic institution in the United States. The study was approved per the university’s process for approval of quality improvement projects involving humans. The study sample population was obtained from hospital administrative data submitted to University HealthSystem Consortium (UHC) Clinical Data Base/Resource Manager. Data were retrieved for patients discharged between April 2012 and June 2013 at our institution. Patients were included if they were admitted to the ICU, were 18 years of age or older, and received dexmedetomidine or propofol for sedation. Prior to April 2012, dexmedetomidine was considered to be a nonformulary agent at the institution and its use was not permitted. However, after April 2012, propofol was available in limited quantities due to a national shortage. Thus, the use of dexmedetomidine was permitted as an alternative agent for sedation in the ICU.
Database and Variables The UHC database contains data that are used for benchmarking and quality comparisons between hospitals. All patients who satisfied inclusion criteria were categorized as those receiving dexmedetomidine or propofol. For this study, the following fields are considered to be relevant to the analysis: age, sex, race, severity of illness, admission status, UHC service line, International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, total hospital cost, length of ICU stay, and length of hospital stay. The UHC severity of illness rating is an ordinal variable in which patients are classified as minor, moderate, severe, and extreme. Admission status is classified as elective, urgent, emergency, and trauma center. The UHC service line consists of up to 39 predefined groups. These groups are categorized to define similar clinical cohorts. These are based on Medicareseverity diagnosis-related groups that were created by the Centers for Medicare and Medicaid services to represent categories of patients differentiated by severity of illness and resource consumption. Given the large number of groups, service lines with less than 100 patients were grouped to form a single category named ‘‘other.’’ This resulted in 7 UHC service lines (trauma, general medicine, general surgery, cardiac surgery, neurosurgery, thoracic surgery, and other). In addition, ICD-9-CM diagnostic codes were used to calculate the Charlson comorbidity index for each case. Costs data were total hospital costs for the included cases (ie, labor and overhead costs, supplies and medication, and all other costs incurred by the institution). The data did not contain any
Statistical Analysis Propofol and dexmedetomidine patients were compared with regard to baseline demographic and outcome variables. Categorical variables were compared using the Fisher exact test. Continuous variables were analyzed using the Mann-Whitney U test. A nonparametric test was used because the data were not normally distributed. Data were reported as percentages or values expressed as medians and interquartile ranges (IQRs). Three linear multivariate regression analyses were conducted (1 for each outcome): (1) total costs, (2) ICU length of stay, and (3) hospital length of stay. The primary predictor of interest was sedative type (dexmedetomidine or propofol). All other predictor variables (age, sex, race, severity of illness, admission status, Charlson comorbidity index, and UHC service line) were added to each model. The intent was not to develop parsimonious models but rather to adjust for potential confounding. These variables were selected based on previous studies that show an association with hospital outcomes and based on the clinical experience of the investigators. The final models were evaluated for linearity, normality of the residuals, and homoscedasticity. To meet model assumptions, the outcome variables required log-transformation. Severity of illness and Charlson comorbidity index were added to each model as ordinal variables. Three sensitivity analyses were conducted for each final model. First, influential observations were identified using Cook distance. The models were evaluated after excluding these influential observations. Second, patients with early death (less than 48 hours) during hospitalization were excluded to determine whether this changed the results. Early death may result in decreased costs and shorter length of stay, although this is an avoidable outcome. Thus, it is important to conduct the analyses without these patients. Finally, because cardiac surgery patients accounted for a large proportion of patients who received dexmedetomidine, the analyses were also conducted without these patients. All analyses were conducted using STATA 13 (College Station, Texas), and a 2-tailed a of .05 was considered to be statistically significant.
Results Study Cohort The final cohort included 3294 patients. Of these, 2685 received propofol and 609 received dexmedetomidine. The median age was 59 years (IQR 45-70 years), 60.6% were male,
Downloaded from jic.sagepub.com at DREXEL UNIV LIBRARIES on October 7, 2014
Patanwala and Erstad
3 Table 2. Multivariate Analysis of Hospital Costs.a
Table 1. Baseline Comparisons Between Groups. Variable
Propofol Dexmedetomidine P (n ¼ 2685) (n ¼ 609) Value
Age, median (IQR), years 58 (43-69) Sex (male), % 59.5 Race (white), % 62.7 Admission status, % Elective 29.2 Emergency 42.4 Trauma 17.1 Urgent 11.4 Service category, %) Trauma 20.0 General medicine 15.2 General surgery 15.3 Cardiac surgery 1.2 Neurosurgery 8.2 Thoracic surgery 7.2 Other 33.0 Severity of illness, % Minor 9.8 Moderate 23.5 Major 31.8 Extreme 34.9 Charlson comorbidity index, 2 (0-3) Median (IQR)
63 (50-72) 65.2 71.1
Comparison of Dexmedetomidine Versus Propofol on Hospital Costs and Length of Stay Asad E. Patanwala and Brian L. Erstad J Intensive Care Med published online 24 July 2014 DOI: 10.1177/0885066614544452 The online version of this article can be found at: http://jic.sagepub.com/content/early/2014/07/24/0885066614544452
Published by: http://www.sagepublications.com
Additional services and information for Journal of Intensive Care Medicine can be found at: Email Alerts: http://jic.sagepub.com/cgi/alerts Subscriptions: http://jic.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> OnlineFirst Version of Record - Jul 24, 2014 What is This?
Downloaded from jic.sagepub.com at DREXEL UNIV LIBRARIES on October 7, 2014
Article
Comparison of Dexmedetomidine Versus Propofol on Hospital Costs and Length of Stay
Journal of Intensive Care Medicine 1-5 ª The Author(s) 2014 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0885066614544452 jic.sagepub.com
Asad E. Patanwala, PharmD1, and Brian L. Erstad, PharmD, FASHP, FCCP, MCCM1
Abstract Objective: The objective of this evaluation was to compare total hospital costs and length of stay of critically ill patients who received dexmedetomidine versus propofol for sedation in the intensive care unit (ICU). Methods: This was a retrospective quality improvement evaluation at a tertiary care, academic medical center in the United States. Data were retrieved for patients discharged between April 2012 and June 2013. Patients were included if they were admitted to the ICU, were 18 years of age or older, and received dexmedetomidine or propofol for sedation. Multivariate regression models were developed to determine the association between sedative type and hospital costs, ICU length of stay, and hospital length of stay. Results: The final cohort included 3294 patients. Of these, 2685 received propofol and 609 received dexmedetomidine. The median hospital cost was US$31 041 (interquartile range [IQR] US$17 963-US$57 826) in the propofol group and US$46 716 (IQR US$31 247 to US$85 490) in the dexmedetomidine group (P < .001). The median ICU length of stay was 2 days (IQR 1-6 days) and 4 days (IQR 2-9 days) in the propofol and dexmedetomidine groups, respectively (P < .001). Overall, hospital length of stay was 8 days (IQR 4-15 days) and 9 days (IQR 5-18 days) in the 2 groups, respectively (P < .001). In the multivariate analyses, dexmedetomidine use was associated with increased costs, ICU length of stay, and hospital length of stay (P < .001 for each outcome). Conclusions: In this academic medical center, dexmedetomidine use was associated with higher costs when compared to propofol for sedation in the ICU. Also, use of dexmedetomidine was associated with increased lengths of ICU and hospital stay. Future prospective trials are needed to confirm these findings. Keywords dexmedetomidine, propofol, costs and cost analysis, critical care
Introduction Dexmedetomidine is an a2-adrenergic agonist that has an indication for sedation of intubated and mechanically ventilated patients for a duration of up to 24 hours.1 It has theoretical appeal over more traditional sedative agents such as propofol that is associated with a potentially life-threatening infusion syndrome and benzodiazepines that may be a risk factor for delirium. Current guidelines suggest that dexmedetomidine (or propofol) may be preferred over benzodiazepines in mechanically ventilated patients, since it may improve outcomes such as duration of mechanical ventilation or the occurrence of delirium.2 However, at this time only 1 randomized controlled study has directly compared dexmedetomidine to propofol for prolonged sedation in the intensive care unit (ICU).3 Both sedatives were similar with respect to time at target sedation, duration of mechanical ventilation, mortality, length of ICU, and hospital stay. Dexmedetomidine has a relatively high cost compared to generic versions of propofol and benzodiazepines. Economic
evaluations have shown that dexmedetomidine is associated with decreased overall costs, by decreasing duration of mechanical ventilation and costs associated with episodes of delirium.4,5 Both of these investigations were based on efficacy data derived from a single randomized investigation comparing dexmedetomidine to midazolam.6 In particular, these published economic evaluations do not compare dexmedetomidine and propofol costs based on prescribing patterns outside well-
1
Department of Pharmacy Practice and Science, College of Pharmacy, The University of Arizona, Tucson, AZ, USA
Received February 20, 2014, and in revised form May 20, 2014. Accepted for publication May 21, 2014. Corresponding Author: Asad E. Patanwala, Department of Pharmacy Practice and Science, College of Pharmacy, The University of Arizona, 1295 N Martin Avenue, PO Box 210202, Tuscon, AZ 85721, USA. Email: [email protected]
Downloaded from jic.sagepub.com at DREXEL UNIV LIBRARIES on October 7, 2014
2
Journal of Intensive Care Medicine
controlled trials. The purpose of this investigation was to compare the total costs of dexmedetomidine and propofol regimens when used for sedation of critically ill patients in clinical practice.
missing values; thus, complete case analyses was conducted without the need for imputation.
End Points The primary outcome of interest was total hospital costs. The secondary outcomes were ICU, length of stay, and hospital length of stay. Outcomes were compared between patients who received propofol versus those who received dexmedetomidine.
Materials and Methods Study Population and Setting This was a retrospective quality improvement study conducted at an academic institution in the United States. The study was approved per the university’s process for approval of quality improvement projects involving humans. The study sample population was obtained from hospital administrative data submitted to University HealthSystem Consortium (UHC) Clinical Data Base/Resource Manager. Data were retrieved for patients discharged between April 2012 and June 2013 at our institution. Patients were included if they were admitted to the ICU, were 18 years of age or older, and received dexmedetomidine or propofol for sedation. Prior to April 2012, dexmedetomidine was considered to be a nonformulary agent at the institution and its use was not permitted. However, after April 2012, propofol was available in limited quantities due to a national shortage. Thus, the use of dexmedetomidine was permitted as an alternative agent for sedation in the ICU.
Database and Variables The UHC database contains data that are used for benchmarking and quality comparisons between hospitals. All patients who satisfied inclusion criteria were categorized as those receiving dexmedetomidine or propofol. For this study, the following fields are considered to be relevant to the analysis: age, sex, race, severity of illness, admission status, UHC service line, International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, total hospital cost, length of ICU stay, and length of hospital stay. The UHC severity of illness rating is an ordinal variable in which patients are classified as minor, moderate, severe, and extreme. Admission status is classified as elective, urgent, emergency, and trauma center. The UHC service line consists of up to 39 predefined groups. These groups are categorized to define similar clinical cohorts. These are based on Medicareseverity diagnosis-related groups that were created by the Centers for Medicare and Medicaid services to represent categories of patients differentiated by severity of illness and resource consumption. Given the large number of groups, service lines with less than 100 patients were grouped to form a single category named ‘‘other.’’ This resulted in 7 UHC service lines (trauma, general medicine, general surgery, cardiac surgery, neurosurgery, thoracic surgery, and other). In addition, ICD-9-CM diagnostic codes were used to calculate the Charlson comorbidity index for each case. Costs data were total hospital costs for the included cases (ie, labor and overhead costs, supplies and medication, and all other costs incurred by the institution). The data did not contain any
Statistical Analysis Propofol and dexmedetomidine patients were compared with regard to baseline demographic and outcome variables. Categorical variables were compared using the Fisher exact test. Continuous variables were analyzed using the Mann-Whitney U test. A nonparametric test was used because the data were not normally distributed. Data were reported as percentages or values expressed as medians and interquartile ranges (IQRs). Three linear multivariate regression analyses were conducted (1 for each outcome): (1) total costs, (2) ICU length of stay, and (3) hospital length of stay. The primary predictor of interest was sedative type (dexmedetomidine or propofol). All other predictor variables (age, sex, race, severity of illness, admission status, Charlson comorbidity index, and UHC service line) were added to each model. The intent was not to develop parsimonious models but rather to adjust for potential confounding. These variables were selected based on previous studies that show an association with hospital outcomes and based on the clinical experience of the investigators. The final models were evaluated for linearity, normality of the residuals, and homoscedasticity. To meet model assumptions, the outcome variables required log-transformation. Severity of illness and Charlson comorbidity index were added to each model as ordinal variables. Three sensitivity analyses were conducted for each final model. First, influential observations were identified using Cook distance. The models were evaluated after excluding these influential observations. Second, patients with early death (less than 48 hours) during hospitalization were excluded to determine whether this changed the results. Early death may result in decreased costs and shorter length of stay, although this is an avoidable outcome. Thus, it is important to conduct the analyses without these patients. Finally, because cardiac surgery patients accounted for a large proportion of patients who received dexmedetomidine, the analyses were also conducted without these patients. All analyses were conducted using STATA 13 (College Station, Texas), and a 2-tailed a of .05 was considered to be statistically significant.
Results Study Cohort The final cohort included 3294 patients. Of these, 2685 received propofol and 609 received dexmedetomidine. The median age was 59 years (IQR 45-70 years), 60.6% were male,
Downloaded from jic.sagepub.com at DREXEL UNIV LIBRARIES on October 7, 2014
Patanwala and Erstad
3 Table 2. Multivariate Analysis of Hospital Costs.a
Table 1. Baseline Comparisons Between Groups. Variable
Propofol Dexmedetomidine P (n ¼ 2685) (n ¼ 609) Value
Age, median (IQR), years 58 (43-69) Sex (male), % 59.5 Race (white), % 62.7 Admission status, % Elective 29.2 Emergency 42.4 Trauma 17.1 Urgent 11.4 Service category, %) Trauma 20.0 General medicine 15.2 General surgery 15.3 Cardiac surgery 1.2 Neurosurgery 8.2 Thoracic surgery 7.2 Other 33.0 Severity of illness, % Minor 9.8 Moderate 23.5 Major 31.8 Extreme 34.9 Charlson comorbidity index, 2 (0-3) Median (IQR)
63 (50-72) 65.2 71.1
