J. Lance Lichtor, M.D., Editor
Automated identification of postoperative complications within an electronic medical record using natural language processing. JAMA 2011; 306:848–55
The promise of electronic records: Around the corner or down the road? JAMA 2011; 306:880–1
Although data are often difficult to extract, electronic medical records may provide insights into patient safety and quality of care. A cross-sectional study was conducted to determine the sensitivity and specificity of a natural language processing approach to identify complications compared with use of discharge codes. Electronic medical records from patients (N = 2,974) who underwent inpatient surgical procedures at Veterans Health Administration facilities between 1999 and 2006 were evaluated. In general, the use of natural language processing resulted in higher sensitivities than did use of patient safety indicators for all adverse events (fig. 1). Both methods showed high specificity.
Electronic medical records contain a wealth of information, although the information is not structured. The authors used the Department of Veterans Affairs electronic medical record to show that natural language processing could identify adverse events, such as myocardial infarction, venous thromboembolism, chronic renal failure, sepsis, or pneumonia, better than could the use of patient safety indicators. However, the sensitivity of finding the events is dependent on the algorithm used.
Postoperative myocardial injury after major head and neck cancer surgery. Head Neck 2011; 33:1085–91
Patients undergoing surgery for head and neck cancer are often at risk for coronary artery disease and myocardial infarction. This retrospective cohort study of patients (n = 378) who underwent major head and neck cancer surgery at a single institution evaluated the risk of postoperative myocardial infarction, as identified by peak postoperative troponin 1 (Tn1) concentrations. Fifteen percent of patients experienced increased Tn1 concentrations; the increased Tn1 concentrations occurred within the first 24 h after surgery in 90% of these patients. Preexisting comorbidities were associated with increased postoperative Tn1 concentrations (fig. 2). Hospital (8.5 vs. 10.1 day) and intensive care unit (3 vs. 4.5 days) length of stay were significantly longer in patients with increased Tn1 concentrations. Risk of death was also increased at 60 days (8-fold) and 1 yr (2-fold) in patients with increased Tn1 concentrations.
In this single-institution study, 15% of all patients undergoing major head and neck cancer surgery had increased postoperative Tn1 concentrations. As expected, the patients with increased Tn1 concentrations had longer length of hospital stay and increased risk of death. Although the limits of a retrospective study exist, the authors advise preventative measures and postoperative monitoring for these patients.
Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365:981–92
In a previous study, apixaban (an oral factor Xa inhibitor) reduced the rate of stroke or systemic embolism by 55% compared with aspirin in patients who were not candidates for vitamin K antagonists. The current large (N = 18,201) randomized, double-blind trial compared apixaban (5 mg twice daily) with warfarin (target international normalized ratio, 2.0–3.0) in patients with atrial fibrillation and at least one additional risk factor for stroke. The rate of the primary composite outcome of ischemic or hemorrhagic stroke or systemic embolism was significantly lower in the apixaban group than the warfarin group (P = 0.01; fig. 3). The rates of major bleeding (2.13%vs. 3.09%; P < 0.001) and death of any cause (3.52%vs. 3.94%; P = 0.047) were also significantly lower in the apixaban group compared with the warfarin group.
In this large randomized, double-blind study, apixaban showed efficacy similar to that of warfarin with fewer adverse events. In the future, it is likely that many patients treated with apixaban will present to the operating room. It will be important for anesthesiologists to gain perioperative experience with patients taking this drug for management in the operating room. Pain medicine physicians will also need to understand the management of patients taking apixaban.
Crisis checklists for the operating room: Development and pilot testing. J Am Coll Surg 2011; 213:212–17.e10
Strict adherence to critical management steps may improve quality of patient care during operating room crises such as hemorrhage or cardiac arrest. The current study used simulation scenarios to evaluate the utility of crisis-specific checklists. Operating room teams participated in eight simulations and were assigned randomly to either use the checklist or respond from memory. There was a 6-fold reduction in failure to adhere to critical steps when the checklists were used. Participants found the checklists straightforward, usable, and beneficial.
In this study, the authors developed a set of operating room crisis checklists for 10 specific crises and two scenarios for which a diagnosis was unclear. Testing was performed using an operating room simulator. Study participants found the checklists usable, adhered more to critical steps in management for simulated scenarios, and indicated they would want the checklists available in the case of an actual emergency. Validity testing of the checklists during actual patient care is required.
Critical Care Medicine
Jean Mantz, M.D., Ph.D., Editor
Early versus late rhythm analysis in patients with out-of-hospital cardiac arrest. N Engl J Med 2011; 365:787–97
The survival rate for patients experiencing an out-of-hospital cardiac arrest is low and highly variable. This prospective, cluster-randomized trial compared the survival rates of patients who received early analysis of cardiac rhythm (after 30–60 s of cardiopulmonary resuscitation [CPR]) with late analysis (after 180 s of CPR). Nearly 10,000 patients were included; 5,290 received early analysis, and 4,643 received late analysis. Survival was similar between the groups (5.9%vs. 5.9%; P = 0.59). Rates remained similar between groups even after adjustment for confounding factors and subgroup analyses.
This large prospective, cluster-randomized trial found no benefit in survival or any other outcome endpoint between strategies of early (30–60 s) versus late (3 min) analysis of cardiac rhythm and defibrillation after resuscitation maneuvers were started. Only 40% of the patients received cardioversion. Despite the early intervention of emergency medical services on the scene and state-of-the-art resuscitation performed according to international guidelines, mortality at hospital discharge was extremely high (92%).
Association of ICU or hospital admission with unintentional discontinuation of medications for chronic diseases. JAMA 2011; 306:840–7
Transition errors at hospital discharge may lead to serious complications, even death. This population-based cohort study sought to evaluate the rates of potentially unintentional discontinuation of medications after hospital or intensive care unit admission. Administrative records of hospitals and outpatient prescriptions over a 12-yr period were evaluated. Overall, patients admitted to the hospital were more likely to experience unintentional discontinuation of medications compared with controls across multiple medication types. Patients admitted to the intensive care unit had an increased risk. Patients who discontinued medications were at higher risk for death, emergency department visits, or emergent hospitalization at the 1-yr follow-up.
This large cohort study indicates that intensive care unit admission is a major risk factor for discontinuation of most important medications after hospital discharge, such as antiplatelet agents, anticoagulants, and statins. Although considered a secondary endpoint of the study, mortality was increased after discontinuation of medications after hospital and intensive care unit discharge. These findings point out the importance of improving transitions in health care to ensure medication continuity.
Increased blood glucose variability during therapeutic hypothermia and outcome after cardiac arrest. Crit Care Med 2011; 39:2225–31
Therapeutic hypothermia is recommended in some comatose patients after cardiac arrest. However, the effect of therapeutic hypothermia on blood glucose is not known. This two-site, prospective observational cohort study was conducted to examine the changes in blood glucose concentration during therapeutic hypothermia and its effects on outcomes in patients (N = 220) with coma after cardiac arrest. The blood glucose concentration and insulin dose were significantly higher during therapeutic hypothermia than during normothermia (P < 0.001; fig. 4). Higher mean blood glucose and increased blood glucose variability were associated with mortality. However, after adjusting for variables such as cardiac arrest etiology, increased blood glucose variability, but not mean concentrations, was an independent predictor of inhospital mortality (odds ratio 1.10).
This prospective observational cohort study emphasized that strategies that have demonstrated major benefits on outcome after out-of-hospital cardiac arrest may have limitations in part because of adverse events. Increased blood glucose variability during mild therapeutic hypothermia is a good example of this situation. This study may explain in part why improvement in survival and neurologic status after cardiac arrest is so difficult to achieve despite recent intense efforts.
Long-term use of selective decontamination of the digestive tract does not increase antibiotic resistance: A five-year prospective cohort study. Intensive Care Med 2011; 37:1458–65
Despite its proven efficacy in reducing ventilation-acquired pneumonias, selective digestive decontamination remains highly controversial because of the fear that it could promote the emergence of antibiotic-resistant strains of bacteria. This large (N = 1,588), 5-yr, prospective cohort study evaluated the incidence of intensive care unit-acquired carriage of antibiotic-resistant bacteria in patients receiving selective digestive decontamination. Overall the incidence of antibiotic-resistant bacteria remained stable at 18.91 carriers per 1,000 patient-days. In total, 101 infections in 97 patients occurred because of an antibiotic-resistant bacteria, including pneumonia and bloodstream and urinary tract infections.
This large, prospective single-center cohort study supports that long-term use of selective digestive decontamination is not associated with an increase in acquisition of resistant flora. A secondary endpoint, the increased incidence of infections caused by ceftazidime resistance, requires additional investigation.
Timothy J. Brennan, Ph.D., M.D., Editor
Occupational and other predictors of herniated lumbar disc disease: A 33-year follow-up in The Copenhagen Male Study. Spine 2011; 36:1541–6
A relationship between strenuous occupational activities and low back disorders has been suggested, but direct causality has not been proven. This prospective cohort study followed up 5,245 men for 33 yr to determine their risk of herniated lumbar disc disease. At baseline, 3,833 of the men had no history of back disease. Patients with frequent strenuous physical activity at work had a higher risk of herniated lumbar disc disease (hazard ratio = 3.90). Body height was also a significant predicator of herniated lumbar disc disease.
In this long-term prospective study, patients were followed up for admission into the hospital for herniated lumbar intervertebral disc. Hospitalization was associated with a history of heavy lifting at work and greater patient height. Even after a change from a strenuous to less strenuous occupation, the risk for admission with a herniated lumbar intervertebral disc remained.
Neurodegenerative properties of chronic pain: Cognitive decline in patients with chronic pancreatitis. PLoS One 2011; 6(8):e23363
Chronic pain is associated with reduced quality of life and is the main presenting symptom in patients with chronic pancreatitis. This study sought to determine whether chronic pancreatitis is associated with cognitive decline and what factors contribute to this decline. The neuropsychological profiles of 16 patients with chronic pancreatitis were compared with those of 16 matched healthy volunteers. Patients with chronic pancreatitis showed reductions in multiple cognitive domains, including psychomotor and executive function. Pain was the strongest predictor for cognitive decline. Depressive symptoms, sleep disturbance, opioid use, and a history of alcohol abuse also contributed to declines.
There is increased interest in changes in brain structure and function associated with chronic pain. This study examined functional aspects. Reductions in cognitive performance, executive function, and psychomotor domains were most influenced. The effects on cognitive function appeared to be best related to duration of pain; however, other factors associated with persistent pain could influence the decrease in performance.
The incidence of low back pain in active duty United States military service members. Spine 2011; 36:1492–500
Low back pain will affect as many as 85% of the U.S. population at some point in their lives. This epidemiologic study investigated the incidence and risk factors for developing low back pain in active duty military population using the U.S. Defense Medical Epidemiology Database. Overall, the unadjusted incidence rate of low back pain was 40.5 per 1,000 person-years.
This U.S. Department of Defense database was used to examine risk factors associated with low back pain in active duty personnel. Female sex, increasing age, enlisted rank, and service branch were associated factors (fig. 5).