Publications

PURPOSE: The objective of this study was to quantify the rate at which newly initiated antipsychotic therapy is continued on discharge from the intensive care unit (ICU) and describe risk factors for continuation post-ICU discharge.

MATERIALS AND METHODS: This is a retrospective cohort study of all patients receiving an antipsychotic in the ICUs of a large academic medical center from January 1, 2005, to October 31, 2011. Medical record review was conducted to ascertain whether a patient was newly started on antipsychotic therapy and whether therapy was continued post-ICU discharge.

RESULTS: A total of 39,248 ICU admissions over the 7-year period were evaluated. Of these, 4468 (11%) were exposed to antipsychotic therapy, of which 3119 (8%) were newly initiated. In the newly initiated cohort, 642 (21%) were continued on therapy on discharge from the hospital. Type of drug (use of quetiapine vs no use of quetiapine: odds ratio, 3.2; 95% confidence interval, 2.5-4.0; P < .0001 and use of olanzapine: odds ratio, 2.4, 95% confidence interval, 2.0-3.1; P ≤ .0001) was a significant risk factor for continuing antipsychotics on discharge despite adjustment for clinical factors.

CONCLUSIONS: Antipsychotic use is common in the ICU setting, and a significant number of newly initiated patients have therapy continued upon discharge from the hospital.

OBJECTIVES: To investigate patterns and predictors of use of antipsychotics in hospitalized adults.

DESIGN: Retrospective cohort study.

SETTING: Academic medical center.

PARTICIPANTS: Individuals aged 18 and older hospitalized from August 2012 to August 2013, excluding those admitted to obstetrics and gynecology or psychiatry or with a psychotic disorder.

MEASUREMENTS: Use was ascertained from pharmacy charges. Potentially excessive dosing was defined using guidelines for long-term care facilities. A review of 100 records was performed to determine reasons for use.

RESULTS: The cohort included 17,775 admissions with a median age 64; individuals could have been admitted more than once during the study period. Antipsychotics were used in 9%, 55% of which were initiations. The most common reasons for initiation were delirium (53%) and probable delirium (12%). Potentially excessive dosing occurred in 16% of admissions exposed to an antipsychotic. Of admissions with antipsychotic initiation, 26% were discharged on these medications. Characteristics associated with initiation included age 75 and older (relative risk (RR) = 1.4, 95% confidence interval (CI) = 1.2-1.7), male sex (RR = 1.2, 95% CI = 1.1-1.4), black race (RR = 0.8, 95% CI = 0.6-0.96), delirium (RR = 4.8, 95% CI = 4.2-5.7), dementia (RR = 2.1, 95% CI = 1.7-2.6), admission to a medical service (RR = 1.2, 95% CI = 1.1-1.4), intensive care unit stay (RR = 2.1, 95% CI = 1.8-2.4), and mechanical ventilation (RR = 2.0, 95% CI = 1.7-2.4). In individuals who were initiated on an antipsychotic, characteristics associated with discharge on antipsychotics were age 75 and older (RR = 0.6, 95% CI = 0.4-0.7), discharge to any location other than home (RR = 2.5, 95% CI = 1.8-3.3), and class of in-hospital antipsychotic exposure (RR = 1.6, 95% CI = 1.1-2.3 for atypical vs typical; RR = 2.7, 95% CI = 1.9-3.8 for both vs typical).

CONCLUSION: Antipsychotic initiation and use were common during hospitalization, most often for delirium, and individuals were frequently discharged on these medications. Several predictors of use on discharge were identified, suggesting potential targets for decision support tools that would be used to prompt consideration of ongoing necessity.

BACKGROUND: Dyspnea (breathing discomfort) can be as powerfully aversive as pain, yet is not routinely assessed and documented in the clinical environment. Routine identification and documentation of dyspnea is the first step to improved symptom management and it may also identify patients at risk of negative clinical outcomes.

OBJECTIVE: To estimate the prevalence of dyspnea and of dyspnea-associated risk among hospitalized patients.

DESIGN: Two pilot prospective cohort studies.

SETTING: Single academic medical center.

PATIENTS: Consecutive patients admitted to four inpatient units: cardiology, hematology/oncology, medicine, and bariatric surgery.

MEASUREMENTS: In Study 1, nurses documented current and recent patient-reported dyspnea at the time of the Initial Patient Assessment in 581 inpatients. In Study 2, nurses documented current dyspnea at least once every nursing shift in 367 patients. We describe the prevalence of burdensome dyspnea, and compare it to pain. We also compared dyspnea ratings with a composite of adverse outcomes: 1) receipt of care from the hospital's rapid response system, 2) transfer to the intensive care unit, or 3) death in hospital. We defined burdensome dyspnea as a rating of 4 or more on a 10-point scale.

RESULTS: Prevalence of burdensome current dyspnea upon admission (Study 1) was 13% (77 of 581, 95% CI 11%-16%). Prevalence of burdensome dyspnea at some time during the hospitalization (Study 2) was 16% (57 of 367, 95% CI 12%-20%). Dyspnea was associated with higher odds of a negative outcome.

CONCLUSIONS: In two pilot studies, we identified a significant symptom burden of dyspnea in hospitalized patients. Patients reporting dyspnea may benefit from a more careful focus on symptom management and may represent a population at greater risk for negative outcomes.

PURPOSE: Researchers have found mixed results about the risk to patient safety in July, when newly minted physicians enter U.S. hospitals to begin their clinical training, the so-called "July effect." However, patient and family satisfaction and perception of physician competence during summer months remain unknown.

METHOD: The authors conducted a retrospective observational cohort study of 815 family members of adult intensive care unit (ICU) patients who completed the Family Satisfaction with Care in the Intensive Care Unit instrument from eight ICUs at Beth Israel Deaconess Medical Center, Boston, Massachusetts, between April 2008 and June 2011. The association of ICU care in the summer months (July-September) versus other seasons and family perception of physician competence was examined in univariable and multivariable analyses.

RESULTS: A greater proportion of family members described physicians as competent in summer months as compared with winter months (odds ratio [OR] 1.9; 95% confidence interval [CI] 1.2-3.0; P = .003). After adjustment for patient and proxy demographics, severity of illness, comorbidities, and features of the admission in a multivariable model, seasonal variation of family perception of physician competence persisted (summer versus winter, OR of judging physicians competent 2.4; 95% CI 1.3-4.4; P = .004).

CONCLUSIONS: Seasonal variation exists in family perception of physician competence in the ICU, but opposite to the "July effect." The reasons for this variation are not well understood. Further research is necessary to explore the role of senior provider involvement, trainee factors, system factors such as handoffs, or other possible contributors.

STUDY DESIGN: Retrospective review of medical records and administrative data.

OBJECTIVE: Validate a claims-based algorithm for classifying surgical indication and operative features in lumbar surgery.

SUMMARY OF BACKGROUND DATA: Administrative data are valuable to study rates, safety, outcomes, and costs in spine surgery. Previous research evaluates outcomes by procedure, not indications and operative features. One previous study validated a coding algorithm for classifying surgical indication. Few studies examined claims data for classifying patients by operative features.

METHODS: Patients undergoing lumbar decompression or fusion at a single institution in 2009 for back pain, herniated disc, stenosis, spondylolisthesis, or scoliosis were included. Sensitivity and specificity of a claims-based algorithm for indication and operative features were examined versus medical record abstraction.

RESULTS: A total of 477 patients, including 246 (52%) undergoing fusion and 231 (48%) undergoing decompression were included in this study. Sensitivity of the claims-based coding algorithm for classifying the indication for the procedure was 71.9% for degenerative disc disease, 81.9% for disc herniation, 32.7% for spinal stenosis, 90.4% for degenerative spondylolisthesis, and 93.8% for scoliosis. Specificity was 87.9% for degenerative disc, 85.6% for disc herniation, 90.7% for spinal stenosis, 95.0% for degenerative spondylolisthesis, and 97.3% for scoliosis. Sensitivity and specificity of claims data for identifying the type of procedure for fusion cases was 97.6% and 99.1%, respectively. Sensitivity of claims data for characterizing key operative features was 81.7%, 96.4%, and 53.0% for use of instrumentation, combined (anterior and posterior) surgical approach, and 3 or more disc levels fused, respectively. Specificity was 57.1% for instrumentation, 94.5% for combined approaches, and 71.9% for 3 or more disc levels fused.

CONCLUSION: Claims data accurately reflected certain diagnoses and type of procedures, but were less accurate at characterizing operative features other than the surgical approach. This study highlights both the potential and current limitations of claims-based analysis for spine surgery.

BACKGROUND: Differences among hospitals in the use of inpatient consultation may contribute to variation in outcomes and costs for hospitalized patients, but basic epidemiologic data on consultations nationally are lacking.

OBJECTIVE: The purpose of the study was to identify physician, hospital, and geographic factors that explain variation in rates of inpatient consultation.

DESIGN: This was a retrospective observational study.

SETTING AND PARTICIPANTS: This work included 3,118,080 admissions of Medicare patients to 4,501 U.S. hospitals in 2009 and 2010.

MAIN MEASURES: The primary outcome measured was number of consultations conducted during the hospitalization, summarized at the hospital level as the number of consultations per 1,000 Medicare admissions, or "consultation density."

KEY RESULTS: Consultations occurred 2.6 times per admission on average. Among non-critical access hospitals, use of consultation varied 3.6-fold across quintiles of hospitals (933 versus 3,390 consultations per 1,000 admissions, lowest versus highest quintiles, p < 0.001). Sicker patients received greater intensity of consultation (rate ratio [RR] 1.18, 95% CI 1.17-1.18 for patients admitted to ICU; and RR 1.19, 95% CI 1.18-1.20 for patients who died). However, even after controlling for patient-level factors, hospital characteristics also predicted differences in rates of consultation. For example, hospital size (large versus small, RR 1.31, 95% CI 1.25-1.37), rural location (rural versus urban, RR 0.78, CI 95% 0.76-0.80), ownership status (public versus not-for-profit, RR 0.94, 95% CI 0.91-0.97), and geographic quadrant (Northeast versus West, RR 1.17, 95% CI 1.12-1.21) all influenced the intensity of consultation use.

CONCLUSIONS: Hospitals exhibit marked variation in the number of consultations per admission in ways not fully explained by patient characteristics. Hospital "consultation density" may constitute an important focus for monitoring resource use for hospitals or health systems.

OBJECTIVE: The Centers for Disease Control has recently proposed a major change in how ventilator-associated pneumonia is defined. This has profound implications for public reporting, reimbursement, and accountability measures for ICUs. We sought to provide evidence for or against this change by quantifying limitations of the national definition of ventilator-associated pneumonia that was in place until January 2013, particularly with regard to comparisons between, and ranking of, hospitals and ICUs.

DESIGN: A prospective survey of a nationally representative group of 43 hospitals, randomly selected from the American Hospital Association Guide (2009). Subjects classified six standardized vignettes of possible cases of ventilator-associated pneumonia as pneumonia or no pneumonia.

SUBJECTS: Individuals responsible for ventilator-associated pneumonia surveillance at 43 U.S. hospitals.

INTERVENTIONS: None.

MEASUREMENTS AND MAIN RESULTS: We measured the proportion of standardized cases classified as ventilator-associated pneumonia. Of 138 hospitals consented, 61 partially completed the survey and 43 fully completed the survey (response rate 44% and 31%, respectively). Agreement among hospitals about classification of cases as ventilator-associated pneumonia/not ventilator-associated pneumonia was nearly random (Fleiss κ 0.13). Some hospitals rated 0% of cases as having pneumonia; others classified 100% as having pneumonia (median, 50%; interquartile range, 33-66%). Although region of the country did not predict case assignment, respondents who described their region as "rural" were more likely to judge a case to be pneumonia than respondents elsewhere (relative risk, 1.25, Kruskal-Wallis chi-square, p = 0.03).

CONCLUSIONS: In this nationally representative study of hospitals, assignment of ventilator-associated pneumonia is extremely variable, enough to render comparisons between hospitals worthless, even when standardized cases eliminate variability in clinical data abstraction. The magnitude of this variability highlights the limitations of using poorly performing surveillance definitions as methods of hospital evaluation and comparison, and our study provides very strong support for moving to a more objective definition of ventilator-associated complications.

BACKGROUND: The US Centers for Disease Control and Prevention has implemented a new, multitiered definition for ventilator-associated events (VAEs) to replace their former definition of ventilator-associated pneumonia (VAP). We hypothesized that the new definition could be implemented in an automated, efficient, and reliable manner using the electronic health record and that the new definition would identify different patients than those identified under the previous definition.

METHODS: We conducted a retrospective cohort analysis using an automated algorithm to analyze all patients admitted to the ICU at a single urban, tertiary-care hospital from 2008 to 2013.

RESULTS: We identified 26,466 consecutive admissions to the ICU, 10,998 (42%) of whom were mechanically ventilated and 675 (3%) of whom were identified as having any VAE. Any VAE was associated with an adjusted increased risk of death (OR, 1.91; 95% CI, 1.53-2.37; P < .0001). The automated algorithm was reliable (sensitivity of 93.5%, 95% CI, 77.2%-98.8%; specificity of 100%, 95% CI, 98.8%-100% vs a human abstractor). Comparison of patients with a VAE and with the former VAP definition yielded little agreement (κ = 0.06).

CONCLUSIONS: A fully automated method of identifying VAEs is efficient and reliable within a single institution. Although VAEs are strongly associated with worse patient outcomes, additional research is required to evaluate whether and which interventions can successfully prevent VAEs.