In This Edition
Literature at a Glance
A guide to this month’s studies
- Continuous insulin infusion in non-ICU patients
- How hospitalists spend their day
- Outcomes of patients leaving against medical advice
- Prediction rule for readmission
- Effects of high- vs. low-dose PPIs for peptic ulcer
- Hospital utilization by generalists before hospitalists
- Effect of hospitalist fragmentation on length of stay
- Medication errors at admissions in older patients
Continuous Insulin Infusion Provides Effective Glycemic Control in Non-ICU Patients
Clinical question: Is continuous insulin infusion (CII) a safe and effective option in the management of hyperglycemia in non-ICU patients?
Background: Hyperglycemia has been associated with worse outcomes in hospitalized patients. Prior research has demonstrated the benefit of CII in managing hyperglycemia in the ICU setting. However, outcomes have not been evaluated in the general medical (non-ICU) setting, where hyperglycemia is often inadequately addressed.
Study design: Retrospective chart review.
Setting: Urban tertiary-care medical center.
Synopsis: Charts of 200 adult patients treated with CII in non-ICU areas were reviewed with the primary outcomes including mean daily blood glucose (BG) levels and number of hyper- and hypoglycemic events occurring on CII. Mean BG dropped from 323 mg/dL to 182 mg/dL by day one, with a BG≤of 150 achieved in 67% of patients by day two of therapy. Twenty-two percent of patients suffered a hypoglycemic event (BG<60), reportedly similar to prior studies of insulin use in ICU and non-ICU settings. Eighty-two percent of patients received some form of nutritional support while on CII. In multivariate analyses, receiving oral nutrition (either a solid or liquid diet) was the only factor associated with increased risk of hyperglycemia and hypoglycemia.
This study was limited by its retrospective analysis in a single center. No comparison was made with basal-bolus or sliding-scale insulin therapy regarding efficacy or safety.
Bottom line: Non-ICU patients with hyperglycemia who received CII were able to achieve effective glycemic control within 48 hours of initiation, with rates of hypoglycemia comparable to those observed in ICU settings.
Citation: Smiley D, Rhee M, Peng L, et al. Safety and efficacy of continuous insulin infusion in noncritical care settings. J Hosp Med. 2010;5(4):212-217.
Hospitalists Spend More Time on Indirect, Rather Than Direct, Patient Care
Clinical question: What are the components of the daily workflow of hospitalists working on a non-housestaff service?
Background: The use of hospitalists is associated with increased efficiency in the hospital setting. However, it is not known how this efficiency is achieved. Prior literature has attempted to address this question, but with increasing demands and patient census, the representativeness of existing data is unclear.
Study design: Observational time-motion study.
Setting: Urban tertiary-care academic medical center.
Synopsis: Twenty-four hospitalists were directly observed for two weekday shifts. An electronic collection tool was developed using initial data on hospitalist activities and piloted prior to formal study data collection. Direct patient care was defined as involving face-to-face interaction between hospitalist and patient, while indirect patient care involved activities relevant to patient care but not performed in the patient’s presence.
Approximately 500 hours of observation were collected. Direct patient care comprised only a mean of 17.4% of the hospitalists’ daily workflow, while more was spent on indirect care, mainly electronic health record (EHR) documentation (mean 34.1%) and communication activities (mean 25.9%). Multitasking occurred 16% of the time, typically during communication or “critical documentation activities” (e.g. writing prescriptions). As patient volume increased, less time was spent in communication, and documentation was deferred to after hours.
These results were consistent with prior observational studies but were limited to a single center and might not represent the workflow of hospitalists in other settings, such as community hospitals, or nocturnists.
Bottom line: Hospitalists on non-housestaff services spend most of their time on indirect patient care and, with increasing patient census, communication is sacrificed. Multitasking is common during periods of communication and critical documentation.
Citation: Tipping MD, Forth VE, O’Leary KJ, et al. Where did the day go?—a time-motion study of hospitalists. J Hosp Med. 2010;5(6):323-328.
Patients Who Leave Against Medical Advice Have Higher Readmission, Mortality Rates
Clinical question: What are the 30-day hospital readmission and mortality rates for Veterans Administration (VA) patients discharged against medical advice (AMA) compared with those appropriately discharged from the hospital?
Background: Patients discharged AMA might be at increased risk of experiencing worse outcomes. Small studies have demonstrated that patients with asthma and acute myocardial infarction (MI) discharged AMA have increased risk of readmission and death. However, it is unclear whether these risks are generalizable to a wider medical population.
Study design: Five-year retrospective cohort study.
Setting: One hundred twenty-nine VA acute-care hospitals.
Synopsis: Of the nearly 2 million patients admitted to the VA from 2004 to 2008, 1.7% were discharged AMA. Patients discharged AMA generally were younger, had lower incomes, and were more likely to be black. Furthermore, patients discharged AMA had statistically significant higher rates of 30-day readmission (17.7% vs. 11%, P<0.001) and higher 30-day mortality rates (0.75% vs. 0.61%, P=0.001) compared with those who had been appropriately discharged. In hazard models, discharge AMA was a significant predictor of 30-day readmission and conferred a nonstatistically significant increase in 30-day mortality.
Because all patients were seen in VA facilities, the results might not be generalizable to other acute-care settings. Although VA patients differ from the general medical population, the characteristics of patients discharged AMA are similar to those in previously published studies. The study utilized administrative data, which are very reliable but limited by little information on clinical factors that could contribute to AMA discharges.
Bottom line: Patients discharged AMA are at increased risk of worse post-hospitalization outcomes, including hospital readmission and death.
Citation: Glasgow JM, Vaughn-Sarrazin M, Kaboli PJ. Leaving against medical advice (AMA): risk of 30-day mortality and hospital readmission. J Gen Intern Med. 2010;25(9): 926-929.
Simple Model Predicts Hospital Readmission
Clinical question: Which patient-level factors can be used in a simple model to predict hospital readmission of medicine patients?
Background: Hospital readmissions are common and costly. Previously published readmission prediction models have had limited utility because they focused on a specific condition, setting, or population, or were too cumbersome for practical use.
Study design: Prospective observational cohort study.
Setting: Six academic medical centers.
Synopsis: Data from nearly 11,000 general medicine patients were included in the analysis. Overall, almost 18% of patients were readmitted within 30 days of discharge.
In the prediction model derived and validated from the data, seven factors were significant predictors of readmission within 30 days of discharge: insurance status, marital status, having a regular healthcare provider, Charlson comorbidity index, SF 12 physical component score, one or more admissions within the last year, and current length of stay greater than two days. Points assigned from each significant predictor were used to create a risk score. The 5% of patients with risk scores of 25 and higher had 30-day readmission rates of approximately 30%, compared to readmission rates of approximately 16% in patients with scores of less than 25.
These results might not be generalizable to small, rural, nonacademic settings. Planned admissions could not be excluded from the data, and readmissions to nonstudy hospitals could not be ascertained. Despite these limitations, this model is easier to use than prior models and relevant to a broad population of patients.
Bottom line: A simple prediction model using patient-level factors can be used to identify patients at higher risk of readmission within 30 days of discharge to home.
Citation: Hasan O, Meltzer DO, Shaykevich SA, et al. Hospital readmission in general medicine patients: a prediction model. J Gen Intern Med. 2010;25(3):211-219.
No Difference in Outcomes Between High- and Non-High-Dose Proton Pump Inhibitors in Bleeding Peptic Ulcers
Clinical question: Do high-dose proton pump inhibitors (PPIs) decrease the rate of rebleeding, surgical intervention, or mortality in patients with bleeding peptic ulcers who have undergone endoscopic treatment?
Background: Previous studies have demonstrated superiority of both high- and low-dose PPIs to H2 receptor antagonists and placebo in reducing rebleeding rates in patients with peptic ulcers. However, no clear evidence is available to suggest that high-dose PPIs are more effective than non-high-dose PPIs for treatment of bleeding peptic ulcers.
Study design: Systematic review and meta-analysis.
Setting: Multicenter and single-site studies conducted in several countries.
Synopsis: Studies were included if they were randomized controlled trials, compared high- versus non-high-dose PPIs, evaluated endoscopically confirmed bleeding ulcers, gave PPIs after endoscopy, and documented outcomes regarding rates of rebleeding, surgical intervention, or mortality. High-dose PPIs were defined as equivalent to omeprazole 80 mg intravenous bolus followed by continuous intravenous infusion at 8 mg/hr for 72 hours.
Seven studies met inclusion criteria. The pooled odds ratios for rebleeding, surgical intervention, and mortality were 1.30 (95% CI, 0.88-1.91), 1.49 (95% CI, 0.66-3.37), and 0.89 (95% CI, 0.37-2.13), respectively, for high-dose versus non-high-dose PPIs. The authors concluded that high-dose PPIs were not superior to non-high-dose PPIs in reducing the rates of these adverse outcomes after endoscopic treatment of bleeding ulcers. Considering the cost of high-dose PPIs, further studies are indicated to help guide PPI dosing for patients with peptic ulcers.
Major limitations of this study were the small number of studies (1,157 patients in total) and their heterogeneity, and the lack of intention-to-treat analysis. The studies also included a high Asian predominance, and it has been shown that Asian populations have an enhanced PPI effect.
Bottom line: High-dose PPIs did not demonstrate reduced rates of ulcer rebleeding, surgical intervention, or mortality compared with non-high-dose PPIs in this meta-analysis, which included a small number of studies and patients.
Citation: Wang CH, Ma MH, Chou HC, et al. High-dose vs. non-high-dose proton pump inhibitors after endoscopic treatment in patients with bleeding peptic ulcer: a systematic review and meta-analysis of randomized controlled trials. Arch Intern Med. 2010;170(9):751-758.
Hospital Utilization by Practicing Generalists Declined before the Emergence of Hospitalists
Clinical question: What has been the pattern of hospital utilization by generalists before and after the emergence of hospitalists?
Background: It has been proposed that the emergence of hospitalists has “crowded out” generalist physicians from the U.S. hospital setting. This study evaluated the trends of inpatient practice by generalists both before and after the emergence of hospitalists.
Study design: Retrospective analysis of national databases.
Setting: U.S. data from 1980 to 2005.
Synopsis: Utilizing the National Hospital Discharge Survey and the American Medical Association’s Physician Characteristics and Distribution in the U.S., information was extracted to evaluate the average number of annual inpatient encounters relative to generalist workforce from 1980 to 2005. Total inpatient encounters for each year were calculated by multiplying the total number of hospital admissions by the average hospital length of stay. The emergence of hospitalists was defined as beginning in 1994.
Total inpatient encounters by generalists declined by 35% in the pre-hospitalist era but remained essentially unchanged in the hospitalist era. During the pre-hospitalist period, the number of generalists doubled, to more than 200,000 from approximately 100,000, while the number of hospital discharges remained relatively stable and the length of stay declined by a third. The decrease in average inpatient encounters in the pre-hospitalist era is thought to have been secondary to decreased length of stay and increased workforce.
Bottom line: Hospital utilization relative to generalist physician workforce was decreasing prior to the emergence of hospitalists mainly due to decreased length of hospital stay and increased numbers of physicians.
Citation: Meltzer DO, Chung JW. U.S. trends in hospitalization and generalist physician workforce and the emergence of hospitalists. J Gen Intern Med. 2010;25(5):453-459.
Fragmentation of Hospitalist Care Is Associated with Increased Length of Stay
Clinical question: Does fragmentation of care (FOC) by hospitalists affect length of stay (LOS)?
Background: Previous investigations have explored the impact of FOC provided by residency programs on LOS and quality of care. Results of these studies have been mixed. However, there have been no prior studies on the impact of the fragmentation of hospitalist care on LOS.
Study design: Concurrent control study.
Setting: Hospitalist practices all over the country managed by IPC: The Hospitalist Company.
Synopsis: Investigators looked at 10,977 patients admitted with diagnoses of pneumonia or heart failure. The primary endpoint was LOS. The independent variable of interest was a measure of FOC. The FOC was calculated as a quantitative index, by determining the percentage of hospitalist care delivered by a physician other than the primary hospitalist.
Multivariable analyses revealed a statistically significant increase in LOS of 0.39 days for each 10% increase in fragmentation level for pneumonia admissions. Similarly, for patients with heart failure, there was a significant increase in LOS of 0.30 days for each 10% increase in fragmentation level.
The study is a concurrent control study, so conclusions cannot be drawn about causality. Additionally, there are likely unmeasured differences between every hospital and hospitalist practice, which could further confound the relationships between hospitalist care and LOS.
Bottom line: Fragmentation of care provided by hospitalists is associated with an increased LOS in patients hospitalized for pneumonia or heart failure.
Citation: Epstein K, Juarez E, Epstein A, Loya K, Singer A. The impact of fragmentation of hospitalist care on length of stay. J Hosp Med. 2010;5(6):335-338.
Admission Medication Errors Are Common and Most Harmful in Older Patients Taking Many Medications
Clinical question: What are the risk factors and potential harm associated with medication errors at hospital admission?
Background: Obtaining a medication history from a hospitalized patient is an error-prone process. Several variables can affect the completeness and quality of medication histories, but existing data are limited regarding patient or medication risk factors associated with medication errors at admission.
Study design: Prospective cohort study.
Setting: Academic hospital in Chicago.
Synopsis: Pharmacist and admitting physician medication histories were compared with admission medication orders to identify any unexplained discrepancies. Discrepancies resulting in order changes were defined as medication errors.
Of the 651 adult medical inpatients studied, 234 (35.9%) had medication order errors identified at admission. Errors originated in the medication histories for 85% of these patients. The most frequent type of error was an omission (48.9%). An age of 65 or older (odds ratio [OR]=2.17, 95% confidence interval [CI], 1.09-4.30) and increased number of medications (OR=1.21, 95% CI, 1.14-1.29) were the only risk factors identified by multivariate analysis to be independently associated with increased risk of medication order errors potentially causing harm or requiring monitoring or intervention. Presenting a medication list upon admission was a significant protective factor (OR=0.35, 95% CI, 0.19-0.63).
Though this is the largest study to date evaluating admission medication errors in hospitalized medical patients, it remains limited by its single hospital site. Because the authors were unable to interview patients who were too ill or unwilling to participate and had no caregiver present, they might have underestimated the number of admission errors. Further, the harm assessment was based on potential and not actual harm.
Bottom line: Admission medication order errors are frequent, most commonly originate in the medication histories, and have increased potential to cause adverse outcomes in older patients and those taking higher numbers of medications.
Citation: Gleason KM, McDaniel MR, Feinglass J, et al. Results of the Medications At Transitions and Clinical Handoffs (MATCH) study: an analysis of medication reconciliation errors and risk factors at hospital admission. J Gen Intern Med. 2010;25(5):441-447.