Inappropriate Medications in Elderly ICU Survivors: Where to Intervene?

Elderly patients are often prescribed potentially inappropriate medications (PIMs) during their hospital stay, which are still present at discharge.¹ It is, however, unknown where these PIM therapies are initiated (ie, before hospital admission or in the pre–intensive care unit [ICU] ward, ICU, or post-ICU ward) and if they are stopped or continued across care transitions within the hospital. Furthermore, it is unclear if these PIMs are actually inappropriate medications (AIMs), given the patients' underlying medical condition. We evaluated medication appropriateness in a cohort of critically ill elderly patients, assessing the number and types of PIMs and AIMs at hospital discharge and determining their source of initiation.

This prospective cohort study included patients 60 years or older at an 800-bed academic hospital who were discharged after receiving care in a medical, surgical, or cardiovascular ICU for shock or respiratory failure. We excluded patients who died during hospitalization or were discharged to hospice. Informed consent was obtained initially at enrollment from an available surrogate; patients provided consent when competent. The institutional review board of Vanderbilt Medical Center, Nashville, Tennessee, approved the study.

Demographics and severity of illness were recorded. Medications prescribed at up to 5 distinct time points—before admission (ie, outpatient medications recorded at the time of admission), ward admission (ie, outpatient medications that were continued at admission plus newly prescribed inpatient medications), ICU admission, ICU discharge, and hospital discharge—were abstracted from the electronic health record.

Potentially inappropriate medications were defined using the 2003 Beers criteria² and supplemented with additional medications from more recently published medication safety scales or lists.^3-6 The PIMs were evaluated to determine if they were actually inappropriate (ie, AIMs) based on the patient's medical status by a clinical panel that comprised a hospitalist, geriatrician, and clinical pharmacist. Via the hospital's integrated electronic health record, the clinical panel reviewed the patient's medical history, clinical course, hospital discharge medications, and laboratory data, including creatinine clearance, to determine if each PIM at hospital discharge was actually inappropriate.⁷ A PIM was considered an AIM when at least 2 of the 3 members of the clinical panel considered it inappropriate for the patient based on criteria specified in the Medication Appropriateness Index, including indication, dosage, and likely effectiveness, as well as drug-drug interactions and drug-disease interactions.⁸ This approach was designed to simulate multidisciplinary clinical decision making on rounds, and therefore we did not calculate κ statistics.

Patients' demographic and clinical variables were summarized using median and interquartile range (IQR) for continuous variables or percentages for categorical variables. The numbers of PIMs and AIMs prescribed per patient were described and analyzed as continuous variables. We used the Wilcoxon signed rank test to compare the distribution of preadmission PIMs per patient to the distribution at hospital discharge. We used R software (version 2.11.1; R Foundation for Statistical Computing, Vienna, Austria) for all statistical analyses. P < .05 was considered statistically significant.

We enrolled 120 patients, whose median age was 68 years (IQR, 66-74 years) and Acute Physiology and Chronic Health Evaluation II score at ICU admission was 27 (IQR, 20-32). The percentage of patients prescribed at least 1 PIM increased from 66% before admission to 85% at discharge. The number of patients with 0 PIMs dropped from 34% before admission to 14% at discharge, and the number of patients with 3 or more PIMs increased from 16% before admission to 37% at discharge. (Figure, A) Overall, the number of PIMs prescribed per patient at hospital discharge was significantly higher than before admission (median [IQR], 2 [1-3] vs 1 [0-2]; P < .001). Importantly, 50% of the PIMs at hospital discharge were first prescribed in the ICU; only 20% were first prescribed on the hospital wards and 30% were present before admission.

Among the 103 patients with 1 or more PIMs at hospital discharge, 59% had at least 1 AIM. The distribution of AIMs is shown in the Figure, B. Most (59%) of the AIMs at hospital discharge were first prescribed in the ICU; 20% were first prescribed on the wards and 21% were present before admission.

Advances in critical care medicine have led to a greater number of elderly ICU survivors. These survivors experience polypharmacy and numerous transitions in care during hospitalization and are at high risk of adverse drug effects after hospital discharge. In the present study, 85% of elderly ICU survivors were discharged from the hospital with 1 or more PIM. A unique aspect of our study, and a concerning finding, was that more than half of these patients were discharged with 1 or more AIM, medications deemed more harmful than beneficial to the patient. Similarly, Hajjar et al⁹ found that 44% of elderly veterans were prescribed at least 1 AIM at hospital discharge. In our cohort, half of PIMs and, more importantly, the majority of AIMs at hospital discharge were initiated in the ICU. While it is possible that these drug therapies may be appropriate when started during the course of an acute illness in the ICU (eg, stress ulcer prophylaxis with H₂ antagonists in mechanically ventilated patients), most should have been discontinued at ICU and/or hospital discharge.

Our study highlights the importance of performing an evaluation of medication appropriateness during reconciliation of medications at hospital admission, ICU discharge, and hospital discharge. The increase in PIMs and AIMs may reflect multiple unique challenges that include inadequate standardization of medication reconciliation procedures at transition of care, poor communication between ICU- and ward-based physicians, and inadequate understanding of the dangers of individual medications in the high-risk elderly population.

Given the high prevalence of PIMs and AIMs among elderly survivors of critical illness, physicians practicing in the ICU and hospital wards should be more attentive to the appropriateness of continuing medication therapies started in the ICU. Better coordination is needed at this transition point, with attention to the rationale for starting each medication therapy in the ICU and discussion of when it can be stopped. Potential solutions for reducing inappropriate medications may include electronic medical record surveillance, routine clinical evaluation (eg, geriatrician and clinical pharmacist), and/or improved hand-off communication between discharging and accepting health care providers. Incorporating this assessment of medication appropriateness into the medication reconciliation process when patients are discharged or transferred out of the ICU has the potential to enhance patient safety.

Correspondence: Dr Morandi, Pulmonary and Critical Care Medicine, Vanderbilt University, 1211 21st Ave S, Ste 6100, Nashville, TN 37212 (morandi.alessandro@gmail.com).

Author Contributions: The investigators had full access to the data and were responsible for the study protocol, statistical analysis plan, progress of the study, analysis, reporting of the study, and the decision to publish. Study concept and design: Morandi, Vasilevskis, Pandharipande, Girard, Shintani, Schnelle, Fick, Ely, and Kripalani. Acquisition of data: Morandi, Vasilevskis, Pandharipande, Solberg, Neal, and Koestner. Analysis and interpretation of data: Morandi, Vasilevskis, Pandharipande, Girard, Solberg, Neal, Torres, Thompson, Shintani, Han, Schnelle, Fick, Ely, and Kripalani. Drafting of the manuscript: Morandi. Critical revision of the manuscript for important intellectual content: Morandi, Vasilevskis, Pandharipande, Girard, Solberg, Neal, Koestner, Torres, Thompson, Shintani, Han, Schnelle, Fick, Ely, and Kripalani. Statistical analysis: Torres, Thompson, and Shintani. Obtained funding: Ely. Administrative, technical, and material support: Morandi, Neal, and Fick. Study supervision: Morandi, Vasilevskis, Pandharipande, Girard, Solberg, Han, Schnelle, Ely, and Kripalani.

Financial Disclosure: Dr Ely has received honoraria from GlaxoSmithKline, Pfizer, Lilly, Hospira, and Aspect. Dr Kripalani is a consultant to and holds equity in PictureRx LLC and has received honoraria from Pfizer Inc.

Funding/Support: Dr Pandharipande is supported by the VA Clinical Science Research and Development Service (VA Career Development Award). Dr Ely is supported by the VA Clinical Science Research and Development Service (VA Merit Review Award) and the National Institutes of Health (NIH) (grant AG027472). Dr Girard is supported by the NIH (grant AG034257). Drs Ely and Girard are both supported by the GRECC, VA Medical Center, Tennessee Valley Healthcare System. Dr Fick acknowledges partial support for this work by Award No. R01 NR011042 from the National Institute of Nursing Research (NINR). Dr Kripalani was supported by the NIH (grant K23 HL077597).

Role of the Sponsors: The funding sources did not participate in the planning, collection, analysis, or interpretation of data or in the decision to submit for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the NINR or NIH.