Objectives:Systematic evidence reviews (SERs) produced by the U.S. Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Center (EPC) Program use contextual questions to provide context and background information on the topic. There is currently no standardized approach to address contextual questions in systematic reviews. This study explored the use of publicly available large language models (LLMs) in addressing contextual questions. Study Design:Using a set of 20 published and 5 yet to be published SERs, we selected one contextual question per report and used it as a prompt to elicit answers from an LLM (ChatGPT, Bard, Claude, or Perplexity). Two independent reviewers rated the results using a priori established evaluation criteria (https://osf.io/4k3cu/), comparing the response in the SER to LLM-generated responses. The study was guided by six research questions addressing feasibility, validity of content, validity of structure, mistakes, congruence between responses, and incremental validity of using LLMs to address contextual questions. Results:Using minimal prompt engineering produced relevant responses and documented the feasibility of LLM-generated answers to contextual questions. Responses differed in content and format and are not reproducible (e.g., LLMs update regularly), but LLMs were able to produce articulate, clinically plausible, and well-structured responses. We detected few factual errors, contradictions, and no instance of suspected bias, but citations supporting LLM-generated responses could often not be produced or could not be verified ('confabulations'). Congruence with human generated responses varied, with LLM-generated responses providing more background on the topic and SERs providing more nuanced answers in response to the contextual question. Results regarding incremental validity were mixed and may depend on the tool. Conclusion:LLMs are potentially helpful in addressing contextual questions in systematic reviews but human expertise remains essential for using the generated information in a meaningful way.

Objectives The US Agency for Healthcare Research and Quality (AHRQ), through the Evidence-based Practice Center (EPC) Program, aims to provide health system decision makers with the highest-quality evidence to inform clinical decisions. However, limitations in the literature may lead to inconclusive findings in EPC systematic reviews (SRs). The EPC Program conducted pilot projects to understand the feasibility, benefits, and challenges of utilizing health system data to augment SR findings to support confidence in healthcare decision-making based on real-world experiences. Study Design and Setting Three contractors (each an EPC located at a different health system) selected a recently completed systematic review conducted by their center and identified an evidence gap that electronic health record (EHR) data might address. All pilot project topics addressed clinical questions as opposed to care delivery, care organization, or care disparities topics that are common in EPC reports. Topic areas addressed by each EPC included infantile epilepsy, migraine, and hip fracture. EPCs also tracked additional resources needed to conduct supplemental analyses. The workgroup met monthly in 2022-2023 to discuss challenges and lessons learned from the pilot projects. Results Two supplemental data analyses filled an evidence gap identified in the systematic reviews (raised certainty of evidence, improved applicability) and the third filled a health system knowledge gap. Project challenges fell under three themes: regulatory and logistical issues, data collection and analysis, and interpretation and presentation of findings. Limited ability to capture key clinical variables given inconsistent or missing data within the EHR was a major limitation. The workgroup found that conducting supplemental data analysis alongside an SR was feasible but adds considerable time and resources to the review process (estimated total hours to complete pilot projects ranged from 283-595 across EPCs), and that the increased effort and resources added limited incremental value. Conclusion Supplementing existing systematic reviews with analyses of EHR data is resource intensive and requires specialized skillsets throughout the process. While using EHR data for research has immense potential to generate real-world evidence and fill knowledge gaps, these data may not yet be ready for routine use alongside systematic reviews.

Background: Beta-lactam antibiotics are first-line agents for most patients receiving antimicrobial prophylaxis in surgical procedures. Despite evidence showing low cross-reactivity between penicillins and cephalosporins, patients with beta-lactam allergies commonly receive vancomycin as an alternative to avoid allergic reaction. Methods: Adult patients receiving vancomycin for surgical prophylaxis with a reported beta-lactam allergy at our institution between August 2017 to July 2018 were retrospectively evaluated for potential eligibility for penicillin allergy testing and/or receipt of standard prophylaxis. Results: Among 830 patients who received vancomycin for surgical prophylaxis, 196 reported beta-lactam allergy and were included in the analysis. Approximately 40 % of surgeries were orthopedic. Of patients receiving vancomycin as first-line therapy, 189 (96.4 %) were potentially eligible for beta-lactam prophylaxis. Conclusions: Patients with beta-lactam allergies often qualify for receipt of a first-line antibiotic. An opportunity exists for improved allergy assessment as an antimicrobial stewardship intervention in surgical prophylaxis.

This study aimed to assess understanding of antibiotic resistance and evaluate antibiotic use themes among the general public. In March 2018, respondents that were ≥21 years old and residing in the United States were recruited from ResearchMatch.org and surveyed to collect data on respondent expectations, knowledge, and opinions regarding prescribing antibiotics and antibiotic resistance. Content analysis was used to code open-ended definitions of antibiotic resistance into central themes. Chi-square tests were used to assess differences between the definitions of antibiotic resistance and antibiotic use. Among the 657 respondents, nearly all (99%) had taken an antibiotic previously. When asked to define antibiotic resistance, the definitions provided were inductively coded into six central themes: 35% bacteria adaptation, 22% misuse/overuse, 22% resistant bacteria, 10% antibiotic ineffectiveness, 7% body immunity, and 3% provided an incorrect definition with no consistent theme. Themes that were identified in respondent definitions of resistance significantly differed between those who reported having shared an antibiotic versus those who had not (p = 0.03). Public health campaigns remain a central component in the fight to combat antibiotic resistance. Future campaigns should address the public's understanding of antibiotic resistance and modifiable behaviors that may contribute to resistance.

BACKGROUND:The durability of the antibody response after SARS-CoV-2 infection and the role of antibodies in protection against reinfection are unclear. PURPOSE:To synthesize evidence on the SARS-CoV-2 antibody response and reinfection risk with a focus on gaps identified in our prior reports. DATA SOURCES:MEDLINE (Ovid), EMBASE, CINAHL, World Health Organization Research Database, and reference lists from 16 December 2021 through 8 July 2022, with surveillance through 22 August 2022. STUDY SELECTION:English-language, cohort studies evaluating IgG antibody duration at least 12 months after SARS-CoV-2 infection, the antibody response among immunocompromised adults, predictors of nonseroconversion, and reinfection risk. DATA EXTRACTION:Two investigators sequentially extracted study data and rated quality. DATA SYNTHESIS:Most adults had IgG antibodies after SARS-CoV-2 infection at time points greater than 12 months (low strength of evidence [SoE]). Although most immunocompromised adults develop antibodies, the overall proportion with antibodies is lower compared with immunocompetent adults (moderate SoE for organ transplant patients and low SoE for patients with cancer or HIV). Prior infection provided substantial, sustained protection against symptomatic reinfection with the Delta variant (high SoE) and reduced the risk for severe disease due to Omicron variants (moderate SoE). Prior infection was less protective against reinfection with Omicron overall (moderate SoE), but protection from earlier variants waned rapidly (low SoE). LIMITATION:Single review for abstract screening and sequential review for study selection, data abstraction, and quality assessment. CONCLUSION:Evidence for a sustained antibody response to SARS-CoV-2 infection is considerable for both Delta and Omicron variants. Prior infection protected against reinfection with both variants, but, for Omicron, protection was weaker and waned rapidly. This information may have limited clinical applicability as new variants emerge. PRIMARY FUNDING SOURCE:Agency for Healthcare Research and Quality. (PROSPERO: CRD42020207098).

Objective Improving dementia care is a policy priority nationally and internationally; there is a 'diagnosis gap' with less than half of the cases of dementia ever diagnosed. The English Health Department9s Quality and Outcomes Framework (QOF) encourages primary care recognition and recording of dementia. The codes for dementia are complex with the possibility of underidentification through miscoding. We developed guidance on coding of dementia; we report the impact of applying this to 'clean up' dementia coding and records at a practice level. Design The guidance had five elements: (1) identify Read Codes for dementia; (2) access QOF dementia register; (3) generate lists of patients who may have dementia; (4) compare search with QOF data and (5) review cases. In each practice, one general practitioner conducted the exercise. The number of dementia QOF registers before and after the exercise was recorded with the hours taken to complete the exercise. Setting London primary care. Participants 23 (85%) of 27 practices participated, covering 79 312 (19 562 over 65 s) participants. Outcomes The number on dementia QOF registers; time taken. Results The number of people with dementia on QOF registers increased from 1007 to 1139 (χ2=8.17, p=0.004), raising identification rates by 8.8%. It took 4.7 h per practice, on an average. Conclusions These data demonstrate the potential of a simple primary care coding exercise, requiring no specific training, to increase the dementia identification rate. An improvement of 8.8% between 2011 and 2012 is equivalent to that of the fourth most improved primary care trust in the UK. In absolute terms, if this effects were mirrored across the UK primary care, the number of cases with dementia identified would rise by over 70 000 from 364 329 to 434 488 raising the recognition rate from 46% to 54.8%. Implementing this exercise appears to be a simple and effective way to improve recognition rates in primary care.

Objective: Many post-9/11 Veterans have received Department of Veterans Affairs (VA) healthcare for traumatic brain injury (TBI). Pain conditions are prevalent among these patients and are often managed with opioid analgesics. Opioids may impose unique risks to Veterans with a history of TBI, especially when combined with other psychotropic medications. We examined receipt of opioid and sedative-hypnotic prescriptions among post-9/11 Veterans with TBI who received VA care nationally between 2012 and 2020. Setting: Nationwide VA outpatient care. Participants: Veterans with, versus without, clinician-confirmed TBI based on the VA's Comprehensive TBI Evaluation (CTBIE) were followed up for subsequent years in which they received VA care. Design: A historical cohort study. Main Measures: Proportions of Veterans who received opioid, benzodiazepine, and nonbenzodiazepine sedative-hypnotic prescriptions were compared by CTBIE outcome (TBI yes/no) and by year; overlaps between medication classes, long-term opioid therapy, and high-dose opioid therapy were also examined. Among those with confirmed TBI, logistic regression was used to examine associations between veteran characteristics and likelihood of these potentially high-risk opioid use outcomes. Results: Among 69 752 Veterans with clinician-confirmed TBI, 26.9% subsequently received opioids. The prevalence receiving opioids each year increased from 2012 (16.7%) to 2014 (17.7%), and then decreased each of the following years through 2020 (5.8%). Among Veterans with TBI who received opioids, large proportions also received benzodiazepine (30.1%) and nonbenzodiazepine (36.0%) sedative-hypnotic prescriptions; these proportions also decreased in recent years. In both bivariable and multivariable regression models, Veterans' demographic, TBI, and clinical characteristics were associated with likelihood of potentially high-risk opioid use. Conclusions: VA opioid prescribing to Veterans with TBI has decreased in recent years but remains an important source of risk, particularly when considering coprescriptions of sedative-hypnotic medication. Understanding patterns of psychotropic prescription use among Veterans with TBI can highlight important healthcare and rehabilitation needs in this large patient cohort.

Objective: Receipt of concurrent psychotropic prescription medications from both US Department of Veterans Affairs (VA) and non-VA healthcare providers may increase risk of adverse opioid-related outcomes among veterans with traumatic brain injury (TBI). Little is known about patterns of dual-system opioid or sedative-hypnotic prescription receipt in this population. We estimated the prevalence and patterns of, and risk factors for, VA/non-VA prescription overlap among post-9/11 veterans with TBI receiving opioids from VA providers in Oregon. Setting: Oregon VA and non-VA outpatient care. Participants: Post-9/11 veterans in Oregon with TBI who received an opioid prescription from VA providers between the years of 2014 and 2019. Design: Historical cohort study. Main Measures: Prescription overlap of VA opioids and non-VA opioids or sedative-hypnotics; proportions of veterans who received VA or non-VA opioid, benzodiazepine, and nonbenzodiazepine sedative-hypnotic prescriptions were also examined by year and by veteran characteristics. Results: Among 1036 veterans with TBI receiving opioids from the VA, 210 (20.3%) received an overlapping opioid prescription from a non-VA provider; 5.3% received overlapping benzodiazepines; and none received overlapping nonbenzodiazepine sedative-hypnotics. Proportions of veterans with prescription overlap tended to decrease over time. Veterans with other than urban versus urban addresses (OR = 1.4; 95% CI, 1.0-1.8), high versus medium average annual VA visits (OR = 1.7; 95% CI, 1.1-2.6), and VA service connection of 50% or more versus none/0% to 40% (OR = 4.3; 95% CI, 1.3-14.0) were more likely to have concurrent VA/non-VA prescriptions in bivariable analyses; other than urban remained associated with overlap in multivariable models. Similarly, veterans with comorbid posttraumatic stress disorder diagnoses were more likely to have concurrent VA/non-VA prescriptions in both bivariable and multivariable (OR = 2.1; 95% CI, 1.0-4.1) models. Conclusion: Among post-9/11 veterans with TBI receiving VA opioids, a considerable proportion had overlapping non-VA prescription medications. Providers and healthcare systems should consider all sources of psychotropic prescriptions, and risk factors for overlapping medications, to help mitigate potentially unsafe medication use among veterans with TBI.

BACKGROUND:The wearing of medical and non-medical masks by the general public in community settings is one intervention that is important for the reduction of SARS-CoV-2 transmission, and has been the subject of considerable research, policy, advocacy and debate. Several observational studies have used ecological (population-level) data to assess the effect of masks on transmission, hospitalization, and mortality at the region or community level. METHODS:We undertook this systematic review to summarize the study designs, outcomes, and key quality indicators of using ecological data to evaluate the association between mask wearing and COVID-19 outcomes. We searched the World Health Organization (WHO) COVID-19 global literature database up to 5 March 2021 for studies reporting the impact of mask use in community settings on outcomes related to SARS-CoV-2 transmission using ecological data. FINDINGS:Twenty one articles were identified that analysed ecological data to assess the protective effect of policies mandating community mask wearing. All studies reported SARS-CoV-2 benefits in terms of reductions in either the incidence, hospitalization, or mortality, or a combination of these outcomes. Few studies assessed compliance to mask wearing policies or controlled for the possible influence of other preventive measures such as hand hygiene and physical distancing, and information about compliance to these policies was lacking. INTERPRETATION:Ecological studies have been cited as evidence to advocate for the adoption of universal masking policies. The studies summarized by this review suggest that community mask policies may reduce the population-level burden of SARS-CoV-2. Methodological limitations, in particular controlling for the actual practice of mask wearing and other preventive policies make it difficult to determine causality. There are several important limitations to consider for improving the validity of ecological data.

Abstract Background Beta-lactam (BL) antibiotics are first-line agents for most patients receiving antimicrobial prophylaxis in surgical procedures. Despite evidence showing low cross-reactivity between classes of BLs, patients with allergies commonly receive vancomycin as an alternative to avoid allergic reaction. The objective of this study was to identify potentially inappropriate use of vancomycin surgical prophylaxis among patients with reported BL allergies. Methods Adult patients (≥18 years) receiving vancomycin for surgical prophylaxis with a reported penicillin and/or cephalosporin allergy at our institution between August 2017 to July 2018 were retrospectively evaluated for potential eligibility for penicillin allergy testing and/or receipt of standard prophylaxis. Surgery type and allergy history were extracted from the electronic medical record. Per our institution’s penicillin-testing protocol, patients with IgE-mediated reactions < 10 years ago were eligible for penicillin skin testing (PST), mild reactions or IgE-mediated reaction > 10 years ago were eligible for direct oral amoxicillin challenge, and severe non-IgE mediated allergies were ineligible for penicillin allergy evaluation or BL prophylaxis. Results Among 830 patients who received vancomycin for surgical prophylaxis, 196 reported BL allergy and were included in the analysis (155 with penicillin allergy alone; 21 with cephalosporin allergy; 20 with both cephalosporin and penicillin allergy). Approximately 40% of surgeries were orthopedic. Six patients were ineligible for BL prophylaxis. Per institutional protocol, 73 of 155 patients (48%) may have qualified for PST; 81 of 155 (52%) patients may have received a direct oral amoxicillin challenge. Only 3 of 22 patients with history of methicillin-resistant Staphylococcus aureus appropriately received additional prophylaxis with vancomycin and a BL. Conclusion Patients with BL allergies often qualify for receipt of a first-line BL antibiotic. An opportunity exists for improved BL allergy assessment as an antimicrobial stewardship intervention. Future studies should evaluate outcomes associated with BL allergy evaluation and delabeling in patients receiving surgical prophylaxis. Disclosures All Authors: No reported disclosures

Abstract Background Current Infectious Disease Society of America guidelines recommend anti-methicillin-resistant Staphylococcus aureus(MRSA) agents for treatment of community-acquired pneumonia (CAP) only in specific high-risk patients. There are limited data on duration of vancomycin use that is appropriate in hospitalized patients with CAP. The objective of this study was to evaluate the use of vancomycin for CAP among inpatients. Methods We conducted a retrospective cohort study of inpatients at Oregon Health and Science University Hospital from August 1st, 2017 to July 31st, 2018 who received IV vancomycin and had a pneumonia encounter ICD-9 diagnosis code. Patients with hospital or ventilator-associated pneumonia were excluded. Appropriate therapy was defined as empiric therapy with known risk-factors, concordant therapy with no de-escalation option, or concurrent sepsis or febrile neutropenia. Vancomycin appropriateness was assessed based on medical history and microbiology for both empiric and definitive therapy. We characterized patients receiving inappropriate therapy and calculated the proportion of inappropriate days of therapy (DOT). Results We identified 52 patients with CAP who were treated with vancomycin for a median of 2 DOT (Interquartile Range (IQR): 1–3). Approximately 21% (11/52) of patients had risk factors warranting vancomycin empiric therapy and 42% (22/52) had concurrent sepsis. Nine CAP patients received inappropriate courses of vancomycin, median of 1 day (IQR: 1–2.25) of inappropriate therapy. The most common reason for classifying use as inappropriate was a positive culture for organisms other than MRSA. Patients receiving inappropriate therapy were more frequently transferred from another hospital (44% vs. 30%, P = 0.22). Overall, 16% (20/125) of vancomycin DOT were inappropriate. Conclusion In our study,CAP patients accounted for a small number of pneumonia patients who received vancomycin. The median inappropriate DOT was relatively short, possibly indicating that identification and de-escalation was performed quickly. Further work is required to determine the impact of these findings on patients. Disclosures All authors: No reported disclosures.

Abstract Background Assessing appropriateness of hospital antibiotic use is typically a labor-intensive task for antimicrobial stewardship teams and relies heavily on clinician judgement rather than a systematic process. Vancomycin is a frequently used agent that is a common stewardship target. We developed an algorithm to automatically classify the appropriateness of vancomycin days of therapy (DOTs) based upon electronic health record data. Methods We constructed a retrospective cohort of Oregon Health and Science University (OHSU) Hospital and Doernbecher Children’s Hospital patients admitted August 1, 2017 to July 31, 2018 receiving vancomycin. Data were collected on demographic, encounter, pharmacy, microbiology, and surgery data. An electronic algorithm was applied to classify vancomycin DOTs as appropriate, inappropriate, or indeterminate. Inappropriate use was defined as any case in which there was an opportunity for de-escalation as identified using microbiology data, ICD-10 codes, and procedure codes. Results We included 4,231 encounters; 493 (12%) were pediatric patients. Our algorithm automatically classified 59%, 3%, and 38% of encounters as having either appropriate, inappropriate, or indeterminate DOTs, respectively. Forty-four percent of all encounters received no more than a 24-hour course of vancomycin and were considered appropriate empiric therapy; half of these were attributed to surgical prophylaxis. Nine percent of all encounters had vancomycin administered within 3 days of a blood, sputum or tissue culture in which either a methicillin-resistant Staphylococcus species or an ampicillin-resistant, vancomycin-susceptible Enterococcus species was isolated and were classified as appropriate. Six percent of all encounters had cultures in which only Gram-negatives, fungi, or yeast were isolated and were therefore considered appropriate in the empiric period (≤48 hours) but inappropriate thereafter. Conclusion Automated assessments of antibiotic appropriateness could facilitate more informed antimicrobial stewardship initiatives and serve as a valuable stewardship metric. Characterization of indeterminate vancomycin use may inform increased automated classification. Further effort is needed to validate these assessments. Disclosures All authors: No reported disclosures.

Antibiotic use metrics are utilized by antimicrobial stewardship programs to benchmark performance against peer institutions and inform stewardship efforts. Benchmarking requires risk adjustment for patient- and facility-level factors so that remaining differences are attributable only to prescribing practices. Antibiotics for the treatment of methicillin-resistant Staphylococcus aureus (MRSA) are one of the most frequently used drug classes. Our objective was to identify predictors of anti-MRSA antibiotic use in a nationwide network of hospitals. We used data from inpatient encounters at facilities participating in the Vizient data repository between January 1, 2016 and December 31, 2016. The outcome, anti-MRSA antibiotic use, was calculated as days of therapy per patient-days present for each encounter. We constructed a multivariable negative binomial regression model and assessed the following predictors for inclusion: age, sex, race, ethnicity, diagnosis related groups (DRGs), ICU days, admit month, facility bed size, facility teaching status, and region. A clinical framework was used to categorize DRGs based on risk of anti-MRSA antibiotic use. A backwards stepwise approach was used to identify the final model. We evaluated predictor effect size and significance, and assessed model fit using a deviance-based pseudo-R2. One hundred forty-five facilities representing 3,608,711 encounters met inclusion criteria. All predictors considered in our model were significant. Predictors with the greatest magnitude of association included DRG categories and patient age. The DRG categories with the strongest associations were DRGs for infections likely due to Staphylococcus aureus (RR = 1.66, P < 0.0001) or for diagnoses likely to receive long-term MRSA coverage (RR = 1.49, P < 0.0001). The age group with the strongest association was age 2–10 years (RR = 1.64; P < 0.001). The deviance-based pseudo-R2 of the final model was 0.19, indicating good model fit. DRGs and patient-level characteristics can be utilized to account for variability in anti-MRSA antibiotic use beyond what is explained through facility-level characteristics. Incorporation of the significant predictors identified in this study may aid in more meaningful interhospital comparisons of anti-MRSA antibiotic use in both adults and pediatrics. J. C. McGregor, Merck: Grant Investigator, Research grant.

STUDY DESIGN:Systematic review. OBJECTIVES:The objective of this study was to answer 5 key questions: What is the comparative effectiveness and safety of (1a) anticoagulant thromboprophylaxis compared to no prophylaxis, placebo, or another anticoagulant strategy for preventing deep vein thrombosis (DVT) and pulmonary embolism (PE) after acute spinal cord injury (SCI)? (1b) Mechanical prophylaxis strategies alone or in combination with other strategies for preventing DVT and PE after acute SCI? (1c) Prophylactic inferior vena cava filter insertion alone or in combination with other strategies for preventing DVT and PE after acute SCI? (2) What is the optimal timing to initiate and/or discontinue anticoagulant, mechanical, and/or prophylactic inferior vena cava filter following acute SCI? (3) What is the cost-effectiveness of these treatment options? METHODS:A systematic literature search was conducted to identify studies published through February 28, 2015. We sought randomized controlled trials evaluating efficacy and safety of antithrombotic strategies. Strength of evidence was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system. RESULTS:Nine studies satisfied inclusion criteria. We found a trend toward lower risk of DVT in patients treated with enoxaparin. There were no significant differences in rates of DVT, PE, bleeding, and mortality between patients treated with different types of low-molecular-weight heparin or between low-molecular-weight heparin and unfractionated heparin. Combined anticoagulant and mechanical prophylaxis initiated within 72 hours of SCI resulted in lower risk of DVT than treatment commenced after 72 hours of injury. CONCLUSION:Prophylactic treatments can be used to lower the risk of venous thromboembolic events in patients with acute SCI, without significant increase in risk of bleeding and mortality and should be initiated within 72 hours.

The Centers for Disease Control and Prevention National Healthcare Safety Network (NHSN) provides hospitals a mechanism to report antibiotic use (AU) data to benchmark against peer institutions and direct antibiotic stewardship efforts. Differences in patient populations need to be adjusted for to ensure unbiased comparisons across hospitals. Our objective was to identify predictors of total AU across a nationwide network of hospitals. Data from 126 academic hospitals were extracted from the Vizient Clinical Data Base Resource Manager for adult inpatients (age ≥ 18 years) in 2015. AU was expressed as total antibiotic days of therapy/patient-days. We constructed a negative binomial regression model to explore potential predictors of AU including age, race, sex, case mix index, hospital bed size, length of stay, geographic region, transfer cases, service line, and illness severity. A backwards stepwise approach based on likelihood ratio test was used to identify significant (P < 0.05) predictors and construct the final, parsimonious model. We calculated dispersion-based R2 to assess the percent variability explained by the full and final models. A total of 3,076,394 total admissions, representing 17,544,763 patient days, were included. Factors identified as significant predictors in the final model are shown in the Table. The percent variance explained by the full and final models was 90.3% and 89.6%, respectively. Table: Independent predictors of total facility antibiotic use per patient days Table: Independent predictors of total facility antibiotic use per patient days The current NHSN AU risk adjustment metric, the standardized antimicrobial administration ratio (SAAR), has been developed separately for different antibiotic groupings and adjusts for a limited set of facility characteristics. Further work is needed to assess if the independent predictors identified in this model can improve upon the performance of existing SAAR metrics and aid in directing stewardship strategies. All authors: No reported disclosures.

Study Design: Systematic review. Objective: To conduct a systematic review and synthesis of the literature to assess the comparative effectiveness, safety, and cost-effectiveness of early (≤24 hours) versus late decompression (>24 hours) in adults with acute spinal cord injury (SCI). Methods: A systematic search was conducted of Medline, EMBASE, the Cochrane Collaboration Library, and Google Scholar to identify studies published through November 6, 2014. Studies published in any language, in humans, and with an abstract were considered for inclusion. Included studies were critically appraised and the overall strength of evidence was determined using methods proposed by the Grading of Recommendation Assessment, Development and Evaluation working group. Results: The search yielded 449 potentially relevant citations. Sixteen additional primary studies were identified through other sources. Six studies met inclusion criteria. All but 2 studies were considered to have moderately high risk of bias. Across studies and injury levels, the impact of early surgical decompression (≤24 hours) on clinically important improvement in neurological status was variable. Isolated studies reported statistically significant and clinically important improvements at 6 months (cervical injury, low strength of evidence) and following discharge from inpatient rehabilitation (all levels, very low strength of evidence) but not at other time points; another study observed a statistically significant 6 point improvement in ASIA Impairment Scale (AIS) among patients with AIS B, C, or D, but not for those with AIS A (very low strength of evidence). In one study of acute central cord syndrome without instability, a clinically and statistically meaningful improvement in total motor scores was reported at 6 and 12 months in patients treated early (versus late). There were, however, no significant differences in AIS improvement between early and late surgical groups at 6- or 12-months (very low strength of evidence). One of 3 studies found a shorter length of hospital stay associated with early surgical decompression. Of 3 studies reporting on safety, no significant differences in rates of complications (including mortality, neurologic deterioration, pneumonia or pressure ulcers) were noted between early and late decompression groups. Conclusions: Results surrounding the efficacy of early versus late decompressive surgery, as well as the quality of evidence available, were variable depending on the level of SCI, timing of follow-up, and specific outcome considered. Existing evidence supports improved neurological recovery among cervical SCI patients undergoing early surgery; however, evidence regarding remaining SCI populations and clinical outcomes was inconsistent.

OBJECTIVES:In randomized controlled clinical trials, continuous outcomes are typically measured at both baseline and follow-up, and mean difference could be estimated using the change scores from baseline or the follow-up scores. This study assesses the impact of using change score vs. follow-up score on the conclusions of meta-analyses. STUDY DESIGN AND SETTING:A total of 63 meta-analyses from six comparative effectiveness reviews were included. The combined mean difference was estimated using a random-effects model, and we also evaluated whether the impact qualitatively varied by alternative random-effects estimates. RESULTS:Based on the Dersimonian-Laird (DL) method, using the change vs. the follow-up score led to five meta-analyses (7.9%) showing discrepancy in conclusions. Based on the profile likelihood (PL) method, nine (14.3%) showed discrepancy in conclusions. Using change score was more likely to show a significant difference in effects between interventions (DL method: 4 of 5; PL method: 7 of 9). A significant difference in baseline scores did not necessarily lead to discrepancies in conclusions. CONCLUSIONS:Using the change vs. the follow-up score could lead to important discrepancies in conclusions. Sensitivity analyses should be conducted to check the robustness of results to the choice of mean difference estimates.

A Measurement Tool to Assess Systematic Reviews (AMSTAR) is a commonly used tool to assess the quality of systematic reviews; however, modifications are needed to improve its usability, reliability, and validity. In this commentary, we summarize our experience and the experiences of others who have used AMSTAR and provide suggestions for its improvement. We propose that AMSTAR should modify a number of individual items and their instructions and responses to make them more congruent with an assessment of the methodologic quality of systematic reviews. We recommend adding new items and modifying existing items to assess the quality of the body of evidence and to address subgroup and sensitivity analyses. More detailed instructions are needed for scoring individual items across multiple reviewers, and we recommend that a total score should not be calculated. These suggestions need to be empirically tested prior to implementation.

Background In randomized controlled clinical trials, continuous outcomes are typically measured at both baseline and followup time points, and mean difference is analyzed as the effect measure. There are multiple ways to estimate the mean difference: using the change score from the baseline, using the followup scores, or estimating the mean difference using the analysis of covariance (ANCOVA) model. Use of the ANCOVA model is generally preferable to using the change scores from the baseline or using the followup scores. When the baseline scores are imbalanced, using either the change score from the baseline or the followup scores would produce biased effect estimates of mean difference, while the ANCOVA model provides the least biased estimates. Nonetheless, individual studies often report results incompletely, and investigators have to summarize results across studies that are not optimally reported. The impact of using the change versus the followup score on meta-analysis has not been well studied. Methods We selected six comparative effectiveness reviews published by the Agency for Healthcare Research and Quality that included at least one meta-analysis for continuous outcomes using mean difference. Data were abstracted from a total of 63 meta-analyses (156 trials) to evaluate differences in baseline scores and how the choice of using the change score or the followup score impacted the combined mean difference using a random effects model and discrepancy in conclusions. Discrepancy in conclusion occurs when one estimate (e.g., change score) shows significant difference and the other estimate (e.g., followup score) does not. We also evaluated whether the impact qualitatively varied by the comparator and alternative random effect estimates. Results Based on the Dersimonian-Laird (DL) method, using the change score versus the followup score led to 5 out of the 63 meta-analyses (7.9%) showing discrepancy in conclusions; based on the profile likelihood (PL) method, 9 (14.3%) showed discrepancy in conclusions. Using the change score was more likely to show a significant difference in effects between interventions (4 out of 5 using the DL method, and 7 out of 9 using the PL method). The impact of the change score versus the followup score using the maximum likelihood was similar to using the DL method, and the impact using the restricted maximum likelihood method was similar to using the PL method. Using the Knapp-Hartung modification of random effect estimate led to most (10) meta-analyses showing discrepancy in conclusions. A significant difference in baseline scores did not necessarily lead to discrepancy in conclusion. Finally, among the 10 meta-analyses that compared active intervention versus control or usual care, using the change score versus the followup score led to one discrepancy in conclusion using the DL or the PL method (10%) but using change scores consistently produced larger intervention effects in nine meta-analyses. Among the other 53 meta-analyses comparing different interventions, there were 4 discrepancies using the DL method (7.5%) and 8 discrepancies using the PL method (15.1%). Conclusions This study of 63 meta-analysis indicated that using the change score versus the followup score to estimate mean difference could lead to important discrepancies in conclusions. Using the change score is more likely to produce significant results when there are discrepancies in conclusions; using the followup score is more likely to produce more conservative results. Sensitivity analyses using both change scores and final values should be conducted to check the robustness of results to the choice of mean difference estimates.