Kevin J. O'Leary

Fatal Pseudomembranous Colitis Associated with a Variant Clostridium difficile Strain Not Detected by Toxin A Immunoassay

Clostridium difficile infection should be a primary diagnostic consideration in any patient with diarrhea who has been hospitalized and has recently received antibiotics. Detection of toxin in patient stool specimens is the most important laboratory evidence for confirming a diagnosis of C. difficile diarrhea, although reliance on any one test may be insufficient to exclude the diagnosis (1). Clostridium difficile produces two large single-unit toxins (toxins A and B), which share significant functional domain homology and act by glycosylation of small guanosine triphosphatebinding proteins that are involved in cell cytoskeleton organization (2). Commercial immunoassays have been developed that detect toxin A by using the monoclonal antibody PCG-4, which recognizes epitopes encoded by a region of highly repetitive DNA sequences in the 3 end of the toxin A gene (3). In general, these immunoassays have a similar specificity but a somewhat decreased sensitivity compared with cell-culture cytotoxin assays, which primarily detect the effects of toxin B (1). In addition to using toxin testing, the clinical laboratory at our hospital in Chicago, Illinois, has increased the sensitivity of diagnostic testing for C. difficile disease by adding a culture for C. difficile. Toxin assays are also performed in vitro on the culture supernatants of recovered isolates to screen for nontoxigenic (nonpathogenic) strains. In 1998, the cell-culture cytotoxin assay that had been used at our laboratory was replaced with an immunoassay for toxin A (Clearview C. DIFF A, Wampole Laboratories, Division of Carter-Wallace, Inc., Cranbury, New Jersey) after a prospective study showed that this toxin A immunoassay had a sensitivity and specificity similar to those of the cytotoxin assay (4). We report on a patient who died of pseudomembranous colitis that was caused by a strain of C. difficile undetected by repeated toxin A immunoassays. We genetically characterize this strain of C. difficile and estimate the number of clinical laboratories in our region with test strategies that will not detect this variant strain. Case Report An 86-year-old man with multiple cardiac and pulmonary problems was admitted to the hospital for treatment of presumed inflammatory bowel disease. During a previous hospitalization 2 months earlier, the patient had developed diarrhea and had been treated for pneumonia with antibiotics. At that time, findings on stool specimen testing were negative for toxin A but the culture was positive for C. difficile. The recovered C. difficile isolate tested negative for toxin A production in vitro (using the same toxin A immunoassay used on the stool specimen) and was reported as nontoxigenic. The patient was subsequently evaluated on several occasions for lower abdominal cramps and recurrent diarrhea. Findings on colonoscopy and abdominal computed tomography were suspicious for C. difficileassociated pseudomembranous colitis, but this diagnosis was dismissed because of two additional stool specimens that were negative for toxin A and were culture positive for a presumed nontoxigenic C. difficile strain. Two months after the initial onset of symptoms, the patient returned to the hospital with diarrhea, fever (37.4 C [99.3 F]), abdominal tenderness, and an elevated leukocyte count (14.2 109 cells/L). The patient was treated with steroids for presumed inflammatory bowel disease, partly because a recent extensive gastrointestinal evaluation (which had included a stool culture for routine enteric pathogens and an ova and parasite examination) was unrevealing. The patients diarrhea improved, but the lower abdominal cramping persisted. On the 3rd day of hospitalization, the patient experienced cardiopulmonary arrest, and attempts to resuscitate him were unsuccessful. Autopsy showed diffuse pseudomembranous colitis (Figure, left panel). A stool specimen obtained at the time of autopsy was negative for toxin A but was culture positive for a presumed nontoxigenic strain of C. difficile. After the autopsy report was received, the four clinical isolates were tested in vitro by using a cell-culture cytotoxin assay. All isolates tested positive for cytotoxin. No specific treatment for C. difficile diarrhea had been administered at any time during the patients illness. (See Case Summary.) Figure. Gross findings at autopsy performed 2 months after symptom onset. Left. Right. Molecular Investigation All four clinical isolates were identical by using HindIII-restriction endonuclease analysis typing (5) (data not shown) and were designated as restriction endonuclease analysis type CF4. The restriction pattern of the four isolates matched that of a group of highly related isolates in our clinical isolate collection (5); this group of related isolates, designated as CF, is known to be toxin variant. Although strains in the CF group produce toxin B and are cytotoxic, these variant strains do not produce toxin A and are characterized by a 1.8-kb deletion in the 3 end of the toxin A gene (6) (Table). Thus, we amplified the 3 end of the toxin A gene by performing polymerase chain reaction. The amplification was done on DNA that had been isolated and purified from the C. difficile isolates of patients and controls, as previously described (6), by using the primers A-u2 and A-d1-b (the primers start at positions 5301 and 8136, respectively, on the toxin A gene). The amplified product from the four isolates recovered from our patient was 1.8 kb smaller than the product from amplification of the standard toxigenic strain VPI 10463; the size of the amplified product was identical in size to the amplified product from the toxin Bpositive, toxin Anegative variant strain, type CF2 (6) (data not shown). Table 1. Diagnostic and Clinical Characteristics of Clostridium difficile Strains with Different Toxigenic Potential Table 2. Case Summary To identify additional CF isolates, we screened our collection of more than 5000 clinical isolates (6). By using restriction endonuclease analysis, we identified 58 variant CF isolates (1.7% of the 3445 total isolates typed), including 3 additional CF4 isolates from patients at two Veterans Affairs hospitals (in Minneapolis, Minnesota, and in Chicago) and a county hospital (in Chicago). At that time, the clinical laboratory at each of the three hospitals had been using both cytotoxin assays and culture for C. difficile testing. In all three patients, results on stool cytotoxin assays were positive and C. difficile diarrhea was diagnosed. Telephone Survey In September 2000, we surveyed all hospital-based laboratories in the Chicago metropolitan area to determine the type of C. difficile testing being performed. Immunoassay for toxin A was the only test being used to detect C. difficile in 31 of the 67 (46%) laboratories surveyed. Two laboratories were using a latex agglutination test for nontoxin C. difficile antigen, and the remaining laboratories performed either immunoassay for both toxins A and B or a cell-culture cytotoxin assay. Discussion Reliance on a single laboratory test despite epidemiologic and clinical clues that were highly suggestive of C. difficile infection contributed to the death of this patient. The patient initially presented with abdominal pain and diarrhea in the hospital after receiving antibiotic therapy for pneumonia. Although findings on colonoscopy and abdominal computed tomography were highly suggestive of C. difficileassociated pseudo-membranous colitis, the diagnosis was not made. Laboratory testing of stool specimens for C. difficile toxin is an accepted method for the diagnosis of C. difficile diarrhea. However, these tests are not highly sensitive compared with stool culture, and a negative result on immunoassay or on cytotoxin assay should not be used to exclude the diagnosis of C. difficile diarrhea (8). Several C. difficile strains with variations in the pathogenicity locus of the toxin gene have been reported (9-11). Rupnik and colleagues classified these variations into 15 toxinotypes by using molecular techniques (12). Toxinotype VIII strains are characterized by a 1.8-kb deletion in the toxin A gene and altered restriction sites in the toxin B gene. The toxinotype VIII strains do not produce a functional toxin A and are undetectable by toxin A immunoassays (13). Until recently, reports had suggested that toxinotype VIII strains, which also correspond to serogroup F [14], were nonpathogenic (10, 14, 15). We recently characterized a toxin variant of a C. difficile strain that we designated as restriction endonuclease analysis type CF2. The CF2 strain was found in specimens recovered from seven patients at a Veterans Affairs hospital in Minneapolis; five of the seven patients had documented cases of C. difficile diarrhea (6). Despite producing no toxin A and having genotypic characteristics nearly identical to those of strain 1470, which is the prototype strain for toxinotype VIII, type CF2 is cytotoxic and can cause disease in hamsters (although with lower colonization efficiency and mortality compared with other, fully toxigenic strains of C. difficile [16]). Recently, a nosocomial outbreak of C. difficile diarrhea due to a toxinotype-VIII variant strain was reported at a hospital in Winnipeg, Manitoba, Canada (17). Toxin-variant strains of C. difficile that appear identical or similar to toxinotype VIII are more widespread than was previously recognized (6, 17-19). Moreover, these strains can cause the full spectrum of C. difficileassociated diseases. A recent survey conducted by the United Kingdom National External Quality Assessment Scheme reported that 108 of 243 (44%) laboratories surveyed had been performing only toxin A testing for diagnosis of C. difficile (United Kingdom National External Quality Assessment Scheme. Clostridium difficile Survey. Distribution no. 1305. 10 January 2000). The results of the U.K. survey and of our survey of hospital-based laboratories in the Chic

Where did the day go?—A time-motion study of hospitalists†

BACKGROUND Within the last decade hospitalists have become an integral part of inpatient care in the United States and now care for about half of all Medicare patients requiring hospitalization. However, little data exists describing hospitalist workflow and their activities in daily patient care. OBJECTIVE To clarify how hospitalists spend their time and how patient volumes affect their workflow. DESIGN Observers continuously shadowed each of 24 hospitalists for two complete shifts. Observations were recorded using a handheld computer device with customized data collection software. SETTING Urban, tertiary care, academic medical center. RESULTS : Hospitalists spent 17% of their time on direct patient contact, and 64% on indirect patient care. For 16% of all time recorded, more than one activity was occurring simultaneously (i.e., multitasking). Professional development, personal time, and travel each accounted for about 6% of their time. Communication and electronic medical record (EMR) use, two components of indirect care, occupied 25% and 34% of recorded time respectively. Hospitalists with above average patient loads spent less time per patient communicating with others and working with the EMR than those hospitalists with below average patient loads, but reported delaying documentation until later in the evening or next day. Patient load did not change the amount of time hospitalists spent with each patient. CONCLUSIONS Hospitalists spend more time reviewing the EMR and documenting in it, than directly with the patient. Multi-tasking occurred frequently and occupied a significant portion of each shift.

Improving Interunit Transitions of Care Between Emergency Physicians and Hospital Medicine Physicians: A Conceptual Approach

Patient care transitions across specialties involve more complexity than those within the same specialty, yet the unique social and technical features remain underexplored. Further, little consensus exists among researchers and practitioners about strategies to improve interspecialty communication. This concept article addresses these gaps by focusing on the hand-off process between emergency and hospital medicine physicians. Sensitivity to cultural and operational differences and a common set of expectations pertaining to hand-off content will more effectively prepare the next provider to act safely and efficiently when caring for the patient. Through a consensus decision-making process of experienced and published authorities in health care transitions, including physicians in both specialties as well as in communication studies, the authors propose content and style principles clinicians may use to improve transition communication. With representation from both community and academic settings, similarities and differences between emergency medicine and internal medicine are highlighted to heighten appreciation of the values, attitudes, and goals of each specialty, particularly pertaining to communication. The authors also examine different communication media, social and cultural behaviors, and tools that practitioners use to share patient care information. Quality measures are proposed within the structure, process, and outcome framework for institutions seeking to evaluate and monitor improvement strategies in hand-off performance. Validation studies to determine if these suggested improvements in transition communication will result in improved patient outcomes will be necessary. By exploring the dynamics of transition communication between specialties and suggesting best practices, the authors hope to strengthen hand-off skills and contribute to improved continuity of care.

Impact of hospitalist communication‐skills training on patient‐satisfaction scores

BACKGROUND Hospital patient-satisfaction scores now affect hospital payment, but little research addresses how hospitals or clinicians might improve performance. OBJECTIVE To assess the impact of a communication-skills training program on patient satisfaction with doctor communication and overall hospital care. DESIGN Preintervention vs postintervention comparison of patient-satisfaction scores. We designed a communication-skills training program for hospitalists consisting of three 90-minute sessions, based on a popular framework. SETTING Nonteaching hospitalist service in an urban academic hospital. MEASUREMENTS Doctor-communication items from the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) and Press Ganey surveys, and HCAHPS overall hospital rating. RESULTS Overall, 61 (97%) of 63 hospitalists completed the first session, 44 (70%) completed the second session, and 25 (40%) completed the third session of the program. Patient-satisfaction data was available for 278 patients during the preintervention period and 186 patients during the postintervention period. Two of the 3 HCAHPS and all 5 of the Press Ganey doctor-communication items were rated higher during the postintervention period, but no result was statistically significant. Similarly, the overall hospital rating was higher during the postintervention period, but the result was not significant. Analyses based on level of hospitalist participation did not show significant differences. CONCLUSIONS Patient satisfaction did not significantly improve after a communication-skills training program for hospitalists. Because of the small sample size, larger studies are needed to assess whether such a program might truly improve patient satisfaction.

Physician Report Cards and Implementing Standards of Practice Are Both Significantly Associated With Improved Screening Colonoscopy Quality

OBJECTIVES:Adenoma-detection rates (ADRs) are associated with decreased interval colorectal cancer (CRC) rates and CRC mortality; quality improvement strategies focus on improving physician ADRs. The objective of this study was to examine the sequential effect of physician report cards and implementing institutional standards of practice (SOP) on ADRs.METHODS:Colonoscopy metrics were prospectively evaluated at a single academic medical center over a 23-month period (November 2012 to October 2014). ADRs were evaluated over three time periods—Period 1: Before initial report card distribution or SOP (November 2012 to March 2013); Period 2: After individualized report card distribution detailing physician and institutional ADRs (April 2013 to March 2014); Period 3: After second report card and SOP implementation (April 2014 to October 2014). The SOP required physicians to have a minimum 5-min withdrawal time in normal colonoscopies (WT) and an ADR minimum of 20%; those who did not meet benchmarks would require further training or endoscopy block time alterations. Only endoscopists averaging >15 colonoscopies/month were included in this analysis.RESULTS:Twenty endoscopists met the inclusion criteria, performing 12,894 screening colonoscopies over the 23-month period. Following report card distribution, physician ADRs increased by 3% (P<0.001). SOP implementation resulted in a further significant increase in mean physician ADR of 8% (P<0.0001). Overall, mean ADR increased by 11% from Period 1 to Period 3 (P<0.0001). All physicians met the minimum 20% ADR benchmark during Period 3. Although ADRs significantly correlated with WT overall (r=0.45; 95% CI 0.01, 0.75; P=0.04), mean WT did not significantly increase from Period 1 to Period 3.CONCLUSIONS:Our data suggest that distributing colonoscopy quality report cards resulted in a significant ADR improvement. Further, we report evidence that implementing SOP significantly improved ADRs beyond report card distribution and resulted in all endoscopists meeting minimum benchmarks. This suggests that report cards and SOPs may have an additive effect in improving colonoscopy quality, and their implementation in endoscopy labs should be encouraged.

Comparison of traditional trigger tool to data warehouse based screening for identifying hospital adverse events

Background Research supports medical record review using screening triggers as the optimal method to detect hospital adverse events (AE), yet the method is labour-intensive. Method This study compared a traditional trigger tool with an enterprise data warehouse (EDW) based screening method to detect AEs. We created 51 automated queries based on 33 traditional triggers from prior research, and then applied them to 250 randomly selected medical patients hospitalised between 1 September 2009 and 31 August 2010. Two physicians each abstracted records from half the patients using a traditional trigger tool and then performed targeted abstractions for patients with positive EDW queries in the complementary half of the sample. A third physician confirmed presence of AEs and assessed preventability and severity. Results Traditional trigger tool and EDW based screening identified 54 (22%) and 53 (21%) patients with one or more AE. Overall, 140 (56%) patients had one or more positive EDW screens (total 366 positive screens). Of the 137 AEs detected by at least one method, 86 (63%) were detected by a traditional trigger tool, 97 (71%) by EDW based screening and 46 (34%) by both methods. Of the 11 total preventable AEs, 6 (55%) were detected by traditional trigger tool, 7 (64%) by EDW based screening and 2 (18%) by both methods. Of the 43 total serious AEs, 28 (65%) were detected by traditional trigger tool, 29 (67%) by EDW based screening and 14 (33%) by both. Conclusions We found relatively poor agreement between traditional trigger tool and EDW based screening with only approximately a third of all AEs detected by both methods. A combination of complementary methods is the optimal approach to detecting AEs among hospitalised patients.

Improving hospital care and collaborative communications for the 21st century: key recommendations for general internal medicine.

Background Communication and collaboration failures can have negative impacts on the efficiency of both individual clinicians and health care system delivery as well as on the quality of patient care. Recognizing the problems associated with clinical and collaboration communication, health care professionals and organizations alike have begun to look at alternative communication technologies to address some of these inefficiencies and to improve interprofessional collaboration. Objective To develop recommendations that assist health care organizations in improving communication and collaboration in order to develop effective methods for evaluation. Methods An interprofessional meeting was held in a large urban city in Canada with 19 nationally and internationally renowned experts to discuss suitable recommendations for an ideal communication and collaboration system as well as a research framework for general internal medicine (GIM) environments. Results In designing an ideal GIM communication and collaboration system, attendees believed that the new system should possess attributes that aim to: a) improve workflow through prioritization of information and detection of individuals’ contextual situations; b) promote stronger interprofessional relationships with adequate exchange of information; c) enhance patient-centered care by allowing greater patient autonomy over their health care information; d) enable interoperability and scalability between and within institutions; and e) function across different platforms. In terms of evaluating the effects of technology in GIM settings, participants championed the use of rigorous scientific methods that span multiple perspectives and disciplines. Specifically, participants recommended that consistent measures and definitions need to be established so that these impacts can be examined across individual, group, and organizational levels. Conclusions Discussions from our meeting demonstrated the complexities of technological implementations in GIM settings. Recommendations on the design principles and research paradigms for an improved communication system are described.

Effect of patient-centred bedside rounds on hospitalised patients’ decision control, activation and satisfaction with care

Importance Though interprofessional bedside rounds have been promoted to enhance patient-centred care for hospitalised patients, few studies have been conducted in adult hospital settings and evidence of impact is lacking. Objective To evaluate the effect of patient-centred bedside rounds (PCBRs) on measures of patient-centred care. Design and setting Cluster randomised controlled trial involving four similar non-teaching hospitalist service units in a large urban hospital. Participants Hospitalised general medical patients. Intervention We assembled working groups on two intervention units, consisting of professionals and patient/family members, to determine the optimal timing, duration and format for PCBR. Nurses and hospitalists rounded together in PCBR using a communication tool to provide a framework for discussion and unit leaders joined PCBR to provide coaching during initial weeks of implementation. Main outcomes Using patient interviews, we assessed preferred and experienced roles in medical decision-making using the Control Preferences Scale, activation using the Short Form of the Patient Activation Measure, and satisfaction. We also compared postdischarge patient satisfaction survey items related to teamwork, involvement in decisions and overall care. We assessed nurses’, physicians’ and advanced practice providers’ (APP) perceptions of PCBR using a survey developed for this study. Results Overall, 650 patients were approached for structured interview during hospitalisation: 284 were excluded because of disorientation, 54 were excluded because of non-English language, 72 declined to participate and 4 withdrew from the study after enrolment. Interview data were available for 236 (122 control and 114 intervention unit) patients, and postdischarge satisfaction survey data were available for 493 (274 control and 219 intervention unit) patients. We found no significant differences in patients’ perceptions of shared decision-making, activation or satisfaction with care. Results were similar in analyses based on whether PCBR had been performed (ie, per protocol). We also found no difference in postdischarge patient satisfaction items. Results were similar in multivariate analyses controlling for patient characteristics and clustering of patients within study units. A majority of nurses (78.6%), but only about half of hospitalist physicians and APPs felt that PCBR improved communication with patients (47.4%). A minority of nurses (46.4%) and physicians and APPs (36.8%) agreed that PCBR had improved the efficiency of their workday. Conclusions PCBR had no impact on patients’ perceptions of shared decision-making, activation or satisfaction with care. Additional research is needed to identify optimal approaches that can be reliably implemented in hospital settings to improve patient-centred care.

Acute vertebral fracture

Vertebral fractures resulting from osteoporosis are a significant cause of morbidity and mortality in the aging population and are commonly seen in the hospital setting. Appropriately assessing and treating these conditions continues to be a challenge. This article is an evidence-based clinical update on the evaluation and management options of acute vertebral fractures, ranging from conservative treatment to surgical intervention. Hospitalists can play an influential role in the management of osteoporosis.

Unit-based interprofessional leadership models in six US hospitals.

The landscape of hospital-based care has shifted to place greater emphasis on improving quality and delivering value. In response, hospitals and healthcare organizations must reassess their strategies to improve care delivery in their facilities and beyond. Although these institutional goals may be defined at the executive level, implementation takes place at local sites of care. To lead these efforts, hospitals need to appoint effective leaders at the frontlines. Hospitalists are well poised to take on the role of the local clinical care improvement leader based on their experiences as direct frontline caregivers and their integral roles in hospital-wide quality and safety initiatives. A unit-based leadership model consisting of a medical director paired with a nurse manager has been implemented in several hospitals to function as an effector arm in response to the changing landscape of inpatient care. We provide an overview of this new model of leadership and describe the experiences of 6 hospitals that have implemented it.