Literature at a Glance
- Discharge innovation and readmission rates.
- Electronic medical records and outcomes.
- CXR findings predict outcomes.
- NSAIDs and congestive heart failure morbidity.
- Outcomes of interpreter use.
- Predictors and outcomes of postoperative delirium.
- Perioperative beta-blockers.
- Perioperative stroke risk.
Clinical question: Does a standardized discharge intervention lead to a decrease in ED visits and readmission rates following hospital discharge?
Background: Hospital discharge is a complex process that is not standardized at many institutions. Deficiencies in the process can lead to poor outcomes, unnecessary rehospitalizations, and increased costs. Previous studies of peridischarge interventions have yielded mixed results and typically focus on specific patient populations.
Study design: Randomized trial.
Setting: Boston Medical Center, a large, urban, academic medical center.
Synopsis: In this single-institution study, 749 English-speaking hospitalized adults were randomly assigned one of two discharge plans: a multidisciplinary package of discharge services or the usual discharge process. Patients in the intervention group were assigned a nurse discharge advocate who performed patient education, medication reconciliation, discharge coordination, and scheduled follow-up appointments. A pharmacist also telephoned participants two to four days after discharge to reinforce the discharge plan and review medications.
Participants in the intervention group had a 30% relative reduction in hospital utilization (defined as ED visit or hospital readmission) at 30 days. Overall, 21.6% of intervention patients and 26.9% of usual-discharge patients had at least one hospital utilization within 30 days of discharge.
This study was limited to a single center, and 27% of the patients did not meet eligibility criteria. The applicability also is limited by the resource utilization required for the intervention. The authors estimated that 0.5 full-time-equivalent (FTE) nursing time and 0.15 FTE pharmacist time was required to maintain 14 patients per week.
Bottom line: A systematic, intensive approach to discharges can reduce ED return visits and readmission rates.
Citation: Jack B, Chetty V, Anthony D, et al. A re-engineered hospital discharge program to decrease re-hospitalization. Ann Intern Med. 2009:150(3):178-187.
Clinical question: Is improved automation of hospital information associated with reduced rates of inpatient mortality, complications, cost, and length of stay (LOS)?
Background: Clinical information technologies, including electronic medical records (EMR), are touted as an antidote for the fragmented, unsafe, and expensive American healthcare system. Most studies on the effect of such technologies are limited to a single site, and few involve commercially available information systems.
Study design: Cross-sectional study.
Setting: Urban hospitals in Texas.
Synopsis: Researchers used the previously validated Clinical Information Technology Assessment Tool to survey physicians providing inpatient care in 72 Texas hospitals. This tool measures the degree to which clinical information processes are computerized. Automation is divided into four subdomains: test results, notes and records, order entry, and decision support. To achieve a high score, a process must be fully computerized, the physician must know how to activate it, and the physician must choose the computerized process over alternatives. The authors examined the association between a hospital’s degree of automation and mortality, costs, and LOS among patients with myocardial infarction, congestive heart failure, coronary artery bypass grafting, and pneumonia.
Overall, greater automation was associated with lower mortality, fewer complications, and lower costs. No clear impact on LOS was found. Higher scores in the notes and records subdomains were most associated with lower mortality. Higher decision-support scores were most associated with lower complication rates and costs.
This study is one of the first to demonstrate the benefits of clinical information technologies across a variety of institutions using different information systems.
Bottom line: Hospitals with EMR, order entry, and clinical decision support have lower mortality rates, fewer complications, and lower costs.
Citation: Amarasingham R, Plantinga L, Diener-West M, et al. Clinical information technologies and inpatient outcomes. Arch Intern Med. 2009;169(2):108-114.
Clinical question: In patients admitted to the ICU with severe community-acquired pneumonia (CAP), do bacteremia and rapid radiologic progression of pulmonary infiltrates increase the risk of shock and mortality?
Background: Severe CAP is associated with considerable morbidity and mortality; however, data focusing on short-term outcomes is limited. The role of the chest radiograph is established in diagnosis but is unclear as a prognostic tool. Bacteremia is associated with higher mortality risk but also is more common in patients with comorbid illnesses.
Study design: Retrospective cohort.
Setting: 33 hospitals in Spain.
Synopsis: This study retrospectively analyzed 457 patients with severe CAP admitted to the ICU between Dec. 1, 2000, and Feb. 28, 2002. Patients were classified into four groups according to the presence or absence of rapid radiographic spread of pulmonary infiltrates and CAP-associated bacteremia. Patients demonstrating significant worsening by chest radiography within the first 48 hours after admission had a threefold increase in the risk of death. Bacteremia was not associated with increased mortality.
The retrospective nature of this study is its major limitation. Other limitations are the probable inclusion of unrecognized bacteremia in the nonbacteremic groups, the fact that repeat chest radiographs were obtained only once (at 48 hours), and that the cause of radiographic deterioration was not examined.
This study contributes to the literature by identifying a subset of patients (those with worsening chest radiographs at 48 hours) who may benefit from further study and targeted interventions.
Bottom line: In severe CAP patients, radiographic worsening at 48 hours is a negative prognostic factor, while bacteremia is not associated with worse outcomes.
Citation: Lisboa T, Blot S, Waterer G, et al. Radiologic progression of pulmonary infiltrates predicts a worse prognosis in severe community-acquired pneumonia than bacteremia. Chest. 2009;135(1):165-172.
Clinical question: Is NSAID use by patients with congestive heart failure (CHF) associated with a higher risk of death or hospitalization due to acute myocardial infarction (MI) or heart failure?
Background: NSAID use is widespread and generally perceived to be low-risk given their over-the-counter availability. However, clinical guidelines discourage the use of NSAIDs in patients with chronic heart failure due to the risk of fluid retention and worsening heart failure.
Study design: Retrospective cohort.
Setting: All hospitals in Denmark.
Synopsis: This study identified 107,092 patients who survived their first hospitalizations due to heart failure between 1995 and 2004. Subsequent use of NSAIDs was determined from a national prescription registry. Patient records were retrospectively analyzed to assess mortality and hospitalization due to MI or heart failure.
At least one NSAID prescription was claimed by 33.9% of the patients after discharge. All NSAIDs were associated with higher death rates, and there was a dose-dependent increase in the risk of death. Ibuprofen and naproxen demonstrated increased mortality only at high doses. All NSAIDs that were studied increased the risk of hospitalization for MI or heart failure.
The observational design is the study’s major limitation. Other important limitations include lack of detailed information about heart failure diagnoses and indication for starting NSAID therapy.
This study intensifies the debate regarding the increased risk of cardiovascular events in NSAID patients, which has been ongoing since the publication of the VIGOR Study in 2000.
Bottom line: In patients with a history of heart failure, NSAIDs are associated with an increased risk of death and cardiovascular morbidity.
Citation: Gislason G, Rasmussen J, Abildstrom S, et al. Increased mortality and cardiovascular morbidity associated with use of non-steroidal anti-inflammatory drugs in chronic heart failure. Arch Intern Med. 2009;169(2):141-149.
Clinical question: Do family meetings that require the use of an interpreter have different characteristics than those in which an interpreter is not needed?
Background: Communication about end-of-life care is essential in the ICU, yet limited English proficiency (LEP) can be a barrier to effective discussions. Overall, outcomes and satisfaction are improved when interpreters are used, but specific effects on ICU family conferences are unknown.
Study design: Cross-sectional evaluation of family meetings.
Setting: Four hospitals in Seattle.
Synopsis: Fifty-one noninterpreted (English-speaking members only) and 10 interpreted (non-English-speaking members present) ICU family meetings were recorded and analyzed for the amount of speaking time and content. The total duration was similar for interpreted versus noninterpreted conferences (26.3 minutes vs. 32.0 minutes, P=0.25), but clinician speech was significantly less in the interpreted group (10.9 minutes vs. 19.6 minutes, P=0.001). Family speaking time was similar in both interpreted and noninterpreted conversations (7.1 minutes vs. 8.2 minutes, P=0.66). Clinicians used more emotional support for families in noninterpreted meetings, including active listening and pausing for questions.
This study is limited by the use of an audio recorder; a video recorder would have provided researchers with participants’ physical interaction and expressions. Additionally, the study does not differentiate between cultural and linguistic difficulties, or provide a reference for the degree of complexity of each conference.
Bottom line: Family meetings in the ICU that require the use of an interpreter provide less information and emotional support to family members than those in which an interpreter is not required.
Citation: Thornton J, Pham K, Engelberg R, et al. Families with limited English proficiency receive less information and support in interpreted intensive care unit family conferences. Crit Care Med. 2009;37(1):89-95.
Clinical question: In patients 50 and older, what are the risk factors for the development of postoperative delirium, and how are outcomes affected by delirium?
Background: Delirium in the elderly postoperative patient is common. It results in increased costs, morbidity, and mortality. As the population ages, more elderly patients will undergo surgical procedures, so identification of delirium risk factors is essential.
Study design: Prospective, observational, cohort study.
Setting: Veterans Affairs Medical Center, Denver.
Synopsis: Researchers assessed 144 patients 50 and older who were scheduled to undergo surgical procedures with a planned, postoperative ICU stay for cognitive function, overall functional status, and comorbidities. Postoperatively, patients were assessed daily for the development of dementia using the cognitive assessment method-ICU instrument (CAM-ICU). Additionally, a validated chart review method for diagnosing delirium was used. The overall delirium incidence was 44%, and only 12% of cases had an identifiable etiology. The mean onset of delirium was 2.4 days; duration was 4.5 days. The incidence of delirium increased with age, reaching 92% in the 80- to 89-year-old group. In multivariate analysis, preoperative cognitive dysfunction was the strongest predictor of delirium.
Delirium development in patients 50 and older was associated with marked increases in costs ($50,000 vs. $32,000), length of stay (16 days vs. eight days), discharge to a facility (33% vs. 1%), and mortality (9% vs. 1%).
Limitations of this study included the patient population studied (97% men) and lack of data regarding medication use during hospitalization.
Bottom line: In older patients undergoing surgery requiring postoperative ICU care, delirium is common, is associated with prior cognitive dysfunction, and results in significant increases in LOS and mortality.
Citation: Robinson T, Raeburn C, Tran Z, et al. Postoperative delirium in the elderly: risk factors and outcomes. Ann Surg. 2009;249(1):173-178.
Clinical question: Are perioperative beta-blockers effective in preventing cardiac events in noncardiac surgery?
Background: The American College of Cardiology (ACC) and American Heart Association (AHA) guidelines for noncardiac surgery recommend beta-blockers for high-risk patients undergoing intermediate or high-risk surgery. The results of previous randomized controlled trials have been inconsistent.
Study design: Meta-analysis.
Setting: Literature search of PubMed, Embase, and the Cochrane Library.
Synopsis: Thirty-three randomized, controlled trials—which included 12,306 patients—were selected for statistical analysis. In these trials, beta-blockers were initiated in the perioperative period with 30-day followup for outcomes of interest. Overall, perioperative beta-blocker use was associated with a 35% risk reduction in nonfatal MI, a 116% increased risk of nonfatal stroke, and no significant difference in all-cause mortality. Trials with higher levels of bias showed greater statistical benefit of beta-blockers. Of note, the POISE trial carried the largest amount of weight, accounting for nearly two-thirds of the total number of patients included in treatment arms.
The application of this data is challenging, as studies differ in several important variables, such as timing of beta-blocker initiation, dosing regimen, and duration of treatment. The POISE trial, in particular, employed a very large dose of metoprolol, compared with doses of beta-blockers used in other studies. There is less data regarding the perioperative efficacy of beta-blockers when therapy is started well before surgery.
Bottom line: Current evidence does not support the routine use of beta-blockers started immediately prior to noncardiac surgery to prevent perioperative cardiac events.
Citation: Bangalore S, Wetterslev J, Pranesh S, et al. Perioperative beta-blockers in patients having non-cardiac surgery: a meta-analysis. Lancet. 2008;372:1962-1976.
Clinical question: What are the incidence, risk factors, and outcomes of acute ischemic stroke in nonvascular surgery?
Background: Ischemic stroke is a well-understood complication of cardiovascular surgery. However, little data exist in medical literature regarding the frequency, associations, and outcomes of perioperative ischemic stroke in noncardiac surgery.
Study design: Observational chart review using administrative data.
Setting: The Nationwide Inpatient Sample, a public-use database, in which approximately 1,000 hospitals submit data from nonfederal acute-care hospitals.
Synopsis: Three common surgeries were sampled from an administrative database to characterize the epidemiology of perioperative ischemic stroke in noncardiac surgery. This outcome occurred in 0.7% of hemicolectomy patients; 0.2% of total hip replacement patients; and 0.6% of lobectomy/segmental lung resection patients. Studying the rate of perioperative ischemic stroke in coronary artery bypass graft (CABG) patients validated the authors’ method of data extraction. This rate was consistent with prior studies of this outcome in cardiovascular surgery patients.
Multivariate analysis showed that renal disease (odds ratio 3.0), atrial fibrillation (OR 2.0), prior stroke (OR 1.6), and valvular disease (OR 1.5) are statistically associated with an increased risk of perioperative ischemic stroke. The primary outcome was associated with a marked increase in the odds of in-hospital mortality or need for chronic care upon hospital discharge.
This study is limited by its use of administrative coding data, as well as potential bias introduced by the use of only three major types of surgery.
Bottom line: Ischemic stroke is a serious complication of intermediate and major noncardiovascular surgery. It is associated with poor patient outcomes; more evidence is needed to confirm associations with this outcome and to discover strategies to reduce risk.
Citation: Bateman B, Schum-acher H, Wang S, et al. Perioperative acute ischemic stroke in non-cardiac and nonvascular surgery. Anesthesiology. 2009;110:231-238.