Economic Outcomes of Enteral and Parenteral Proton Pump Inhibitor (PPI) Use in the Intensive Care Unit (ICU)

BACKGROUND AND SIGNIFICANCE:

Detail: Proton pump inhibitors (PPIs) are commonly used in intensive care unit (ICU) patients for various clinical conditions, including acute stress related gastrointestinal bleeding (ASRGIB) and clinically significant upper gastrointestinal bleeding (UGIB). Although randomized prospective clinical trials are limited with PPIs and other agents that suppress gastric acid, the Joint Commission on the Accreditation of Healthcare Organizations (JCAHO) requires patients receive acid-suppressive therapy on admission to any ICU. Physicians and pharmacists tend to choose the parenteral route for PPI administration rather than the enteral route despite several potential advantages for administration into the gastrointestinal tract. Factors supporting enteral PPI administration include: the ease with which PPIs can be formulated as a liquid suspension, their high bioavailability in healthy patients, an efficacy that has been shown to be greater than intravenous histamine-2 receptor antagonists (H2RA) therapy, and a low acquisition cost. In fact, studies completed in healthy patients demonstrate that enterally administered PPIs may lead to greater acid suppression than parenterally administered PPIs on a mg-for-mg basis. A potential limitation associated with the use of the gut to administer medication to the critically ill patient relates to whether the drug will be adequately absorbed. We have repeatedly showed at our institution (650 bed teaching hospital with 54 ICU beds) that seriously ill patients, adequately absorbed enterally administered PPIs and provide equal or superior acid suppression profiles when compared to equivalent doses of PPIs administered parenterally. Limited information is available on the utilization trends and associated potential cost savings with the use of enteral therapy compared with parenteral PPI administration in the ICU setting. Lansoprazole oral solutab use in the ICU and hospital setting may lead to further cost savings by reducing the lengthy preparation ease in administration compared with enterally administered PPIs like omeprazole, pantoprazole and esomeprazole. Therefore, to inform formulary decisions and cost-effective quality of care, we propose a real-world retrospective study of the correlates of and economic outcomes in terms of potential savings in drug costs with use of enteral PPI including lansoprazole solutab compared with parenteral PPI use in and out of the ICU.

RESEARCH DESIGN AND METHODS:

Overview: Study Design: This study will be a retrospective electronic medical chart review of all patients admitted to adult intensive care medicine units (ICU) at The Nebraska Medical Center over a three-year period from January 2006 through December 2008.

Following investigational review board (IRB) approval, a computerized list of all patients admitted to the ICU over the designated period will be generated. Using laboratory personal, all patients will be reviewed and data recorded into an access data base. Data will be collected in our outcomes laboratory which is a 400 square feet room housing 5 computers equipped with related software and connected to PHAMIS, the hospital electronic medical record of our institution.

Subjects: Inclusion Criteria:All patients greater than 18 years admitted to the ICU during 2006-2008 (January 2006 through December 2008; 3-year period) Exclusion Criteria Patients ineligible to receive a PPI (e.g. hypersensitivity to any PPI) Sample size: We average 2,500 eligible patients per year admitted to the ICU. So we anticipate a final sample size of nearly 5000 patients which would give us a meaningful indication of real-world use and cost estimates.

Procedures: Data Collection: All data will be collected into an access file database for ease of analysis. Demographic data collected will include: APACHE II score, age, gender, underlying illness, ICU admission diagnosis and gut function, enteral and parenteral nutrition, including tubes in place. In addition length of hospital, ICU stay, and death during hospitalization will be recorded. We will also record the type, route, duration and dose of acid-suppressive therapy before, during and after ICU admission and details of all medications administered during the ICU stay.To aid in drug cost calculations, PPI associated preparation and administration time will also be recorded by observation of 30 ICU admissions each for enteral and parenteral PPI group. Drug costs will be based on 2008 Red book average wholesale price of drugs and resource utilization costs will be based on hospital billing data.

Data analysis and Interpretation: Descriptive Summary: We will report descriptive statistics in terms of proportions of enteral and parenteral PPI use (dose and duration) in ICU and after discharge from ICU. We will also report patient, disease, and resource utilization (continuous measures of PPI dose and duration, and PPI associated preparation and administration time, length of stay (LOS) in ICU, out of ICU, and hospital) characteristics by the two PPI groups (enteral including Lansoprazole solutab versus parenteral). Further, using micro-costing procedures we will estimate economic outcomes i.e. costs in dollars associated with drug resource utilization (PPI dose and duration, and PPI associated preparation and administration time) for the two groups. Savings in drug costs will be also be computed in dollars by calculating difference between parenteral and enteral PPI groups. Death during hospitalization will also be reported.

Analyses for primary aims will be as follows - Aim 1: The proportion of PPI patients initiated on parenteral PPI and enteral PPI will be compared for statistically significant differences on ICU patient and disease characteristics using suitable (chi-square or Fischer's exact) tests. The use of the two PPI groups may be confounded by patient and disease characteristics. To adjust for these confounding factors, multivariate regression modeling will be used. We will carry out multivariate logistic regression modeling to compare the odds of enteral versus parenteral PPI use. Aim 2: Number of patients (mean +/- SD) initiated on parenteral PPIs that could receive enteral therapy as first choice in ICU and after discharge from ICU will be estimated. Aim 3: Savings in drug costs associated with conversion of eligible patients from parenteral to enteral PPI therapy in the ICU and conversion of eligible patients requiring continued AST from parenteral to enteral PPI therapy after discharge from ICU will be reported as mean for normal distribution or median for skewed distribution.

Analyses for secondary aims will be as follows - For the secondary aim 1, we will compare the proportion of patients initiated on enteral and parenteral PPIs using chi-square or fishers exact test. For the secondary aims 2, and 3, we will compare the mean or median for resource and economic outcome variables between the two PPI groups using an independent sample t-test or a Wilcoxon rank sum test for skewed distributions, respectively.