Rothschild JM, Landrigan CP, Cronin JW, et al. The critical care safety study: the incidence and nature of adverse events and serious medical errors in intensive care. Crit Care Med. 2005;33:1694-1700.
Background: Critically ill patients require complex, immediate, high-intensity care, potentially placing them at increased risk of iatrogenic injury. The frequency and nature of adverse events and errors in the modern ICU have not been clearly defined.
Methods: Harvard researchers conducted a prospective, one-year, observational study of a MICU and a CCU at a tertiary care medical center. Adverse events and medical errors were identified by a four-pronged approach: direct 24-hour observation of interns, voluntary incident reporting, a computerized adverse drug event monitoring system, and chart abstraction. Two physicians independently assessed the type, severity, and preventability of the incidents.
Results: A total of 391 patients comprising 1,490 patient-days were observed and included. Twenty percent of all patients suffered an adverse event, 45% of which were preventable and 13% of which were felt to be life-threatening. There were 223 serious errors (those that caused harm or had the potential to cause harm) observed of which 11% were life threatening. Medication adverse events and medication errors accounted for a large proportion of the incidents during the study. Slips and lapses in care were much more common than rule-based (such as using the wrong protocol) or knowledge-based mistakes.
Discussion: Since the Institute of Medicine report in 1999, there has been an increasing focus on patient safety in the inpatient setting. Based on the results of this study and others, it appears the high-intensity, fast-paced nature of critical care places patients at substantial risk for iatrogenic injury. Up to 20% of patients admitted to the ICU in this study suffered an adverse event or a medical error, which translates into 0.8 adverse events and 1.5 serious medical errors per day in a 10-bed ICU.
Because failure to carry out intended plans (usually secondary to slips and lapses on the part of healthcare providers) was the most common cause of adverse events and errors, the authors address possible solutions. They propose employing computerized-order entry, clinical pharmacists in the ICU, closed ICU staffing, “smart” intravenous pumps, and improved teamwork and communication among healthcare providers. Hospitalists often manage critically ill patients and should be aware of the high risk of medical errors and should consider implementing specific systems changes to mitigate the risk.
The Value of Obtaining Blood Cultures in Pneumonia Pts
Kennedy M, Bates DW, Wright SB, et al. Do emergency department blood cultures change practice in patients with pneumonia? Ann Emerg Med. 2005 Nov;46(5):393-400.
Background: Previous observational studies in patients hospitalized with community-acquired pneumonia (CAP) have shown obtaining blood cultures may have a mortality benefit. This practice has become expert guideline-recommended, the standard of care, as well as a quality marker in the management of CAP. Several recent studies have questioned the utility and cost-effectiveness of this practice.
Methods: Harvard researchers performed a prospective, observational, cohort study of adults admitted to an urban university medical center. Researchers identified patients who had all of the following: clinical CAP, radiographic CAP, and blood cultures at admission. Blood cultures were classified as positive, negative, or contaminated based on previously established criteria. Data were collected on antimicrobial sensitivities, empiric antibiotic choices, and antibiotic changes.
Results: In one year, 414 patients with clinical and radiographic CAP had blood cultures at the time of admission. Twenty-nine of 414 (7%) of patients had true bacteremia while 25 of 414 (6%) had contaminants. Antibiotic therapy was altered in response to blood culture results in 15 of 414 patients (3.6%), of which 11 (2.7%) had therapy narrowed and four (1.0%) had therapy broadened. Of the 11 patients with bacteremia whose therapy was not changed, culture results supported narrowing therapy in eight cases but this was not done.
Discussion: This well done prospective observational study adds to a growing body of evidence questioning the utility of routine blood cultures on all patients hospitalized with CAP. The argument traditionally has been made that blood cultures allow clinicians to narrow or broaden antibiotics based on sensitivities. Yet, empiric therapy was broadened in response to bacteremia in only a small fraction of patients (1%) and in only 11 of 19 patients was therapy appropriately narrowed based on the blood cultures. The study did not measure the impact of blood cultures on clinical outcomes, but these striking results reveal that routine blood cultures rarely alter our management of hospitalized patients with CAP.
Further, many have argued obtaining routine blood cultures in CAP can have negative consequences. Blood cultures are relatively costly and time intensive, contaminated blood cultures can lead to repeated testing and increased length of stay, and delays in obtaining blood cultures can delay antibiotic administration, another important quality marker in CAP. For now, it remains the standard of care to obtain blood cultures in these patients, but hospitalists should be aware of the limitations of this practice and consider focusing on other clinical interventions and quality measures in CAP.
A Review Study: A Dyspneic Emergency Patient
Wang CS, FitzGerald JM, Schulzer M, et al. Does this dyspneic patient in the emergency department have congestive heart failure? JAMA. 2005 Oct 19;294:944-1956.
Background: Distinguishing CHF from non-cardiac causes of dyspnea is a major challenge for hospitalists and emergency physicians, particularly in patients with a prior history of cardiac disease. Traditionally, clinicians have relied on the history, physical examination, and basic tests (chest X-ray and electrocardiogram) to diagnose CHF, but rapid B-type natriuretic peptide (BNP) testing is now widely incorporated as well.
A previous article in the Rational Clinical Examination series (Can the clinical examination diagnose left-sided heart failure in adults? JAMA. 1997;277(21):1712-1719) found that systolic dysfunction was moderately well predicted by an abnormal apical impulse on physical examination, radiographic cardiomegaly or venous redistribution, or electrocardiographic q waves or left bundle branch block.
Methods: In this review, the authors update and extend previous findings by also assessing the utility of serum BNP testing. The authors identified articles evaluating the diagnostic accuracy of the clinical exam and laboratory testing in diagnosing CHF in patients presenting to the emergency department with undifferentiated dyspnea. The “gold standard” was a clinical diagnosis of CHF made by the treating clinicians after an appropriate diagnostic workup. Summary likelihood ratios (LRs) were calculated using meta-analytic methodology.
Results/discussion: The authors determined that several findings increase the probability of CHF. A prior history of CHF (LR 5.8, CI 4.1-8.0) or myocardial infarction (LR 3.1, 95% CI 2.0-4.9), symptoms of paroxysmal nocturnal dyspnea (LR 2.6, 95% CI 1.5-4.5) and orthopnea (LR 2.2, 95% CI 1.2-3.9) were the most predictive historical factors. On physical examination, the presence of an S3 (LR 11, 95% CI 4.9-25), jugular venous distension (5.1, 95% CI 3.2-7.9), lung rales (LR 2.8, 95% CI 1.9-4.1), and peripheral edema (2.3, 95% CI 1.5-3.7) increased the probability of CHF. In interpreting these results, it is helpful to remember that a likelihood ratio of 2 increases the post-test probability by about 15%, and an LR of 5 increases the post-test probability by about 30%. Thus, a prior history of CHF and presence of an S3 or jugular venous distension are the most useful findings. Interestingly, clinician’s gestalt was equally predictive (LR 4.4, 95% CI 1.8-10.0.)
The most useful radiographic findings were venous congestion (LR 12.0, 95% CI 6.8-21) and the presence of cardiomegaly (LR 3.3; 95% CI 2.4-4.7). The single most predictive ECG finding was atrial fibrillation (LR 3.8; 95% CI 2.7-8.8); any abnormality on ECG had an LR of 2.2 (95% CI 1.6-3.1). Serum BNP levels were not more predictive of CHF than the history or physical examination; a BNP of >250 was associated with an LR of 4.6 (95% CI 2.6-8.0).
Few findings markedly decreased the probability of CHF. Here, it is helpful to remember that an LR of 0.5 decreases the post-test probability by about 15%, and an LR of 0.2 decreases the post-test probability by about 30%. With these in mind, the absence of cardiomegaly on CXR significantly changes the post-test probability (LR 0.33; 95% CI 0.23-0.48). A serum BNP level of less than 100pg/ml strongly argues against CHF, with an LR of 0.11 (95% CI 0.07-0.16); this finding lowers the post-test probability of CHF by about 45% compared to the pre-test probability.
In summary, the most useful findings for ruling in CHF in dyspneic emergency department patients were clinical gestalt, a prior history of CHF, findings of an S3 or jugular venous distension, and radiographic findings of venous congestion or cardiomegaly. Absence of radiographic cardiomegaly and a BNP of less than 100pg/ml argue against CHF. These must be interpreted in the context of the clinical pre-test probability of CHF, as none of the findings had likelihood ratios sufficient to be diagnostic of CHF when used individually.
What Should I Wear Today?
Rehman SU, Nietert PJ, Cope DW, Kilpatrick AO. What to wear today? Effect of doctor’s attire on the trust and confidence of patients. Am J Med. 2005 Nov; 118(11): 1279-1286.
Background: This study addresses the prototypical everyday clinical dilemma: What should I wear to work?
Methods: Patients and visitors to an outpatient Veterans Affairs internal medicine clinic in South Carolina were shown photographs of male and female physicians in four different styles of dress:
- Professional (male physician wearing white coat with tie, female physician wearing white coat with tailored skirt or trousers);
- Business (suit and tie for male, tailored trouser or skirt for female);
- Surgical (surgical scrubs for both male and female): and
- Casual (jeans and t-shirt or short skirt).
The study was randomized so that male and female respondents viewed photographs of either male or female physicians. Respondents were asked to report how strongly they felt about the importance of their physician’s appearance, and their preference for each style of dress; specifically, respondents were asked which physician was the most trustworthy, which physician they felt most comfortable with for routine examinations and emergencies, and which physician they felt most comfortable discussing psychological, sexual, and social problems with.
Results: Respondents overwhelmingly preferred professional attire for all questions: 76.3% felt most comfortable with a professionally dressed physician for all encounters, with surgical scrubs a distant second (10.2%), ahead of business dress (8.8%). Respondents were also significantly more willing to discuss psychological, sexual, and social problems with a professionally dressed physician. Even for care in an emergency situation, respondents still expressed a significant preference for professional attire over scrubs.
In a logistic regression model, patients who were older, African-American, and had less than a high school education were significantly more likely to prefer professional attire. Interestingly, female respondents who viewed photographs of female physicians placed significantly greater emphasis on physician’s attire than did male respondents.
Discussion: The study is clearly subject to caveats, chiefly that it was conducted at a single VA clinic and that only one aspect of the physician-patient encounter was addressed. Undoubtedly, patient’s preferences were influenced by the popular portrayal of physicians on TV shows. Nevertheless, given that hospitalists typically see older patients with whom they are not familiar, the initial clinical encounter may indeed by influenced by something as simple as wearing a white coat.
UA by Nephrologist Versus Hospital-Based Clinical Labs
Tsai JJ, Yeun JY, Kumar VA, Don BR. Comparison and interpretation of urinalysis performed by a nephrologist versus a hospital-based clinical laboratory. Am J Kidney Dis. 2005 Nov;46(5):820-829.
Background: Distinguishing the correct cause of acute renal failure is a frequent clinical dilemma for hospitalists, particularly diagnosing acute tubular necrosis (ATN), which is the most common cause of in-hospital acute renal failure. Although urinalysis with microscopy is the first test ordered on noting an abnormal serum creatinine, most hospitalists rely on the results generated by a laboratory technician. Anecdotally, many nephrologists have noted significant differences between urinalysis results performed by technicians and results found by nephrologists.
Methods: This study enrolled 26 patients hospitalized with acute renal failure on whom nephrology consultation was obtained. Urinalysis was performed both by laboratory personnel and a nephrologist (nephrologist A) who was blinded to the patient’s clinical information. Both sets of urinalysis results were independently used by nephrologist A and a second nephrologist (nephrologist B) to arrive at a clinical diagnosis for the patient, without having access to any other clinical information. These diagnoses were compared to the final diagnosis determined by the consulting nephrology service, who themselves did not have access to the diagnosis of either nephrologist A or B.
Results: The influence of having a nephrologist perform and interpret the urinalysis was striking. Nephrologist A was able to correctly diagnose 92.3% of cases based solely on his interpretation of the urinalysis. However, when given only the laboratory report of the urinalysis, both nephrologists were unable to diagnose most cases (23.1% for nephrologist A and 19.2% for nephrologist B). The major difference appeared to be in nephrologist A’s ability to find renal tubular epithelial (RTE) cells and RTE casts, which are pathognomonic of ATN. RTE cells and granular casts were frequently misinterpreted as squamous epithelial cells by laboratory personnel. This was particularly important as 81% of patients in the study had ATN as the primary cause of renal failure. Acanthocytes (dysmorphic red blood cells) were also missed by laboratory personnel in all six patients who were subsequently diagnosed with glomerulonephritis; nephrologist A correctly noted acanthocytes in five of these patients, and arrived at the correct diagnosis in all six patients.
Discussion: Microscopic evaluation of urine sediment has become a lost art among physicians, especially since passage of the Clinical Laboratory Improvement Amendments (CLIA) in 1988, which mandated that only CLIA-certified personnel could perform most laboratory tests. While it is probably unrealistic to call for training in microscopic urinalysis for all physicians, hospitalists in particular would benefit from such training, and at the very least should be mindful that laboratory urinalysis results may miss subtle findings that can be invaluable in diagnosing acute renal failure. This study points out the need for greater oversight and training of laboratory personnel, and serves as a reminder to clinicians that laboratory results should not be considered the gold standard. TH