BACKGROUND: According to the Health and Medicine Division of the National Academies of Sciences, Engineering, and Medicine, in order for health care systems to improve health quality, outcomes, cost, and equity there needs to be a process for transmitting new knowledge into everyday care. Systematic reviews are one potential source of knowledge. However, little is known about the types of evidence used by health-systems and how evidence reports produced by the Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Center (EPC) program could be used by learning health-systems. PURPOSE: To better understand how health-systems identify and use evidence and how EPC reports could help them with their decisionmaking in the future. METHODS: From February to September 2017, an AHRQ/EPC Methods Workgroup interviewed nine individuals with leadership roles in enhancing health-system quality, safety, and process improvement from academic, non-academic, and small community health-systems, as well as health-system collaborative organizations. Workgroup members reviewed interview notes and transcripts to identify key themes and exemplar quotations. The nine Key Informants (KI) reviewed the draft report for accuracy. RESULTS: Health-systems have various internal processes for getting and using evidence in their decisionmaking. All of the health-systems we spoke with have either centralized and standardized quality, safety, and process improvement functions within their health-system or have formed partnerships with other organizations to support these improvements. Health-systems recognize that evidence and improvement ideas can come from both the top down (from leadership to local level) as well as from the bottom up (local to leadership). One common process was to conduct searches themselves to obtain information and evidence from the literature. However, there was variation in how this information is obtained. Some of the larger health-systems have medical librarians and centralized committees to gather and disseminate this information. For smaller systems, it is more common that the local chief medical officers or clinical champions identify the information. Other processes for getting evidence include: using internal data to benchmark performance across the system and identify where improvements are needed; and forming subspecialty committees within the health-system to review information and make improvement recommendations within the subspecialty (e.g., cardiology, oncology, and orthopedics). Some of the challenges to the existing processes include: how to resolve conflicting information and whether the information applies to local needs; and how to know whether information is up-to-date. Health-systems feel the standard timeframe to generate systematic reviews is very long, which could hamper optimal utilization in decisionmaking. KIs generally feel that guidelines and systematic reviews are excellent sources of evidence and facilitate quality, safety, and process improvements. While some clinical experts in the health-systems prefer randomized clinical trials, most people in health-systems prefer the synthesized data in guidelines and systematic reviews. KIs generally prefer guidelines, especially those with treatment algorithms, because they are more actionable. KIs prefer evidence from systematic reviews to be summarized into short digestible summaries with the option to click on a link for the more detailed report. They also prefer systematic reviews from known entities and entities that do not have commercial bias. KIs suggest alerting people at multiple levels in a health-system when new evidence reports are available via email or listserv. Some KIs, especially those in small and rural health-systems, noted that they want to be able to obtain evidence in standard search engines, such as Google, rather than having to rely on literature databases. Some KIs felt that they could partner more closely with AHRQ to identify topics but felt that the prolonged turn-around time from topic generation until the report was available could be a barrier. The topics of greatest interest to health-systems varied, but some of the most common ones include: predictive analytics; high-value care; advance care planning, and care coordination. RECOMMENDATIONS AND CONCLUSIONS: Based on these interviews, we recommend the consideration of the following key actions to make EPC reports more useful to health-systems:1. Modify the dissemination emails that go out to health-systems to include not only the titles of the reports and hyperlinks to the full reports, but to include key messages (or hyperlinks to key messages) so that potential users can better assess the relevance of their report to their decisionmaking and better triage the findings internally within their system.2. Periodically construct an EPC newsletter -- perhaps building off the existing EHC listserv -- with a format similar to "The Medical Letter" or "Prescriber's Letter," which provides short three of four sentence summaries of report highlights with hyperlinks to reports.3. Ensure that EPC reports are searchable both within PubMed but also within common search engines such as Google, Bing, and Yahoo. This would include having medical librarians test search using these sites, strategize how to enhance the chances of finding reports for educational purposes, and to reach out to the common search engines and inquire about ways reports can be designed to enhance pick up in searches.4. Explore opportunities to deliver reports more quickly, perhaps through further limiting the scope of a project, writing more focused reports, or by dividing a larger topic into two or three subtopics so that more timely evidence can be available to health-system decision makers.5. EPC reports should not make clinical recommendations like guidelines. However, if the reports could more clearly identify what the current process of care is for a target disease or disorder and where the trials/studies being summarized or pooled will inform decisions that members of health-systems could make, it would make the results more actionable.6. Conduct broad outreach to health-systems to alert them of the topic nomination process so that these topics can be considered within the EPC program and discuss partnerships not unlike those AHRQ has with guideline groups and other Federal agencies. AHRQ and the EPCs can begin with the contacts made with health-systems from this project, contacts identified from the Spring 2017 EPC meeting, and contacts identified from the health-systems which have EPCs to form an advisory group and devise a strategy.7. Establish consistent and prominent branding of the EPC program and its products, ensure that the high-quality journal articles based on our reports are identified as part of the EPC program, and promote the EPC program as a source for timely and reliable reports to improve the quality, safety, and value of care.

BACKGROUND: The principles of evidence-based medicine increasingly govern healthcare policy and practice in the United States. The hallmark of the evidence-based approach is research data generated through clinical trials, and particularly through "gold standard" randomized controlled trials (RCTs). Healthcare policy increasingly relies on evidence furnished by RCTs. It is therefore of paramount importance that investigators are able to execute RCTs, and that those trials include a fair representation of the general population, as well as of the specific populations of relevance to topics being studied. Currently, valid concerns surround low rates of participation in clinical trials and disparities in clinical trial participation. Various factors might influence patients' interest and willingness to participate in clinical trials. Financial repercussions, which depend upon the payment policies of third-party insurers, may be an important element in patients' decisions regarding whether or not to participate in a clinical trial, as well as whether patients stay in the trial once initiated. Payment policies may exert an influence on clinical trials in other ways, impacting not only recruitment and retention, but also conduct of the trial and the subsequent quality of the evidence base. PURPOSE: The Duke Center for Clinical Health Policy Research and Duke Cancer Care Research Program conducted this study, supported by a contract with the Agency for Healthcare Research and Quality (AHRQ), to ascertain whether, and to what extent, payment policies may be influencing participant recruitment to clinical trials, rates of participation, and retention in clinical trials. A further objective was to gather input on the issue of whether or not payment policies, through influencing clinical trial participation, may have a deleterious effect on the resulting evidence base. This report will help to inform the Centers for Medicare and Medicaid Services (CMS) if there is causal relationship between the timing of initiating coverage for new therapeutic technologies and beneficiary participation in clinical trials to provide evidence of effectiveness of these new technologies in the elderly and disabled Medicare population. DESIGN: We employed a variety of strategies to gather data and experiences relevant to the topic. These strategies were: (1) a nationally selected Advisory Panel to provide expert input; (2) a systematic literature review of MEDLINE and ClinicalTrials.gov; (3) a Public Forum held on the CMS campus in Baltimore, MD, to gather public input; and, (4) a series of teleconferences with "key informants" representing diverse stakeholders including government, industry sponsors, third-party insurers, clinical trials investigators and staff, and patient advocates. Flexibility in the study design permitted iterative expansion of the inquiry based on information and insights gathered during the exploratory process. RESULTS: Published data are virtually non-existent to quantify the difficulties encountered by trials with recruitment and retention, as it pertains to third party payment policies. However, in practice several large-scale clinical trials have encountered substantial difficulties due to the deterrent effect of payment policy on participation. Medical device trials have been more affected by these issues than have drug trials. Lack of a common understanding of which costs should be assumed by which party (sponsor, site, third-party payer), and lack of common definitions of "standard of care" versus "research-related" costs complicate payment policy and likely impact enrollment. To ensure that trials get completed, investigators are developing creative solutions to assure participants' coverage on-trial. To ensure that payment policy does not result in a financial loss, sites are analyzing financial impact and may decide not to initiate trials if the financial prospects are negative. Poor coordination among government agencies, industry, third party payers, patients, and researchers is contributing to the difficulties. CONCLUSION: Payment policy does bear an impact on clinical trial participation, though this impact is difficult to quantify and unevenly felt across different types of studies, stage of trials, and study populations. The issue of payment policy is closely related to issues of access to care and disparities in care. Payment policies do affect evidence development, in that their impact on clinical trial enrollment results in slower accrual, longer time to completion of studies, and sometimes early termination of studies due to lack of sufficient sample size. Better coordination among government agencies, and between government, third-party payers, sponsors, and sites is necessary. Presuming that participation in clinical trials is a good thing for individual patients and the public at large, a coherent strategy that stipulates when coverage should be initiated, specifies which costs should be covered, and assigns responsibility for those costs to specific payers, coordinated to maximize clinical trials enrollment and retention, could help to (1) rationalize the process of reimbursement when patients are enrolled in clinical trials, (2) ensure equal access to clinical trials for patients interested in participating, and (3) facilitate the generation of high-quality evidence to support future policy-making and clinical practice.

The use of Bayesian statistical approaches has gained broader acceptance within the clinical trial community. The impact of these methods on CMS decisional contexts and policy-level decisionmaking however was uncertain. Our analyses explore the main proclaimed advantages of Bayesian statistics (namely, the use of prior information, sample size determination, borrowing strength from different trials, and sequential monitoring of trials) and provide examples of decisionmaking situations where the findings reached using these approaches both agree with and differ from findings reached using frequentist statistical techniques. Our findings confirm that, like classical techniques, Bayesian approaches are affected by the problems of model specification (i.e., the relationship between various factors - patient, provider, intervention, and other contextual features - and the outcome of interest). In addition, Bayesian approaches can be substantially affected by the "Bayesian prior" - the representation of beliefs about the quantity of interest (e.g., relative risk of events when a new device is used vs. a conventional device). Thus, when considering using or interpreting Bayesian analyses, the focus of attention and thoughtful ex ante agreement are the specification of the model and specification of the Bayesian prior. The case study of the use of ICD therapy in the prevention of sudden cardiac death demonstrates the application of these techniques and highlights some of the practical challenges. The use of Bayesian statistical approaches, and incorporation of our findings concerning their strengths and limitations into the CMS decisionmaking process will enable policymakers to harness the power of the available sources of clinical evidence, explore subgroup effects within a trial and across trials in a methodologically rigorous manner, assess the uncertainty in clinical trial findings, and - ideally - improve health outcomes for Medicare beneficiaries.

OBJECTIVES: To describe the benefits and harms of specific tools and strategies for screening for postpartum depression. DATA SOURCES: We searched PubMed(r), Embase(r), PsycINFO(r), and the Cochrane Database of Systematic Reviews for relevant English-language studies published from January 1, 2004, to July 24, 2012, that evaluated the performance of screening instruments for postpartum depression, potential benefits and harms of screening, and impact on appropriate postscreening actions. REVIEW METHODS: Two investigators screened each abstract and full-text article for inclusion; abstracted data; and performed quality ratings, applicability ratings, and evidence grading. A simulation model was used to estimate the effects of screening for postpartum depression on the overall balance of benefits and harms. RESULTS: Forty studies (represented by 45 articles) were identified as relevant to this review. Eighteen studies provided sensitivity and specificity data on 9 screening instruments: 11 on the Edinburgh Postnatal Depression Scale, 4 on the Postpartum Depression Screening Scale, 4 on different versions of the Beck Depression Inventory, 2 on a "two-question" screen, and 1 each on 5 other instruments. Heterogeneity in setting, patient population, and choice of threshold prevented formal synthesis. For most tests in most studies, sensitivity and specificity were in the 80-90 percent range, with higher sensitivity associated with lower specificity; the two-question screen had 100 percent sensitivity but specificities of 45-65 percent. Fifteen studies analyzed the association between risk factors and postpartum depression. Although adverse pregnancy outcomes and chronic medical conditions (low strength of evidence) and past history of depression, poor relationship quality, and poor social support (moderate strength of evidence) were all associated with an increased risk of postpartum depression, only two studies directly reported an effect on test results. (Sensitivity was nonsignificantly increased in primigravidas compared with multigravidas.) Based on two studies, there was insufficient evidence to evaluate whether timing relative to delivery, setting, or provider affected test characteristics of screening instruments. Based on five studies, there was low to moderate strength of evidence that screening resulted in decreased depressive symptoms and improved mental health; in four of these studies, improvement in depressive symptoms was not accompanied by improvement in measures of parenting stress. Rates of referral and treatment for women with positive screening results were substantially higher in two studies where screening, diagnosis, and treatment were provided in the same setting; referral rates in other studies were all 50 percent or less. Modeling suggests that serial testing with a two-question screen followed by a second more specific instrument for those who have a positive result may be a reasonable strategy to reduce false positives while minimizing false negatives. CONCLUSIONS: The potential effectiveness of screening for postpartum depression appears to be related to the availability of systems to ensure adequate followup of women with positive results. The ideal characteristics of a screening test for postpartum depression, including sensitivity, specificity, timing, and frequency, have not been defined. Because the balance of benefits and harms, at both the individual level and health system level, is highly dependent on these characteristics, broad consensus on these characteristics is needed.

OBJECTIVES: Pulmonary arterial hypertension (PAH) is a rare and progressive disease associated with increased pulmonary vascular resistance that, if unrelieved, progresses to right ventricular pressure overload, dysfunction, right heart failure, and premature death. PAH is more prevalent in some populations, thereby warranting screening of asymptomatic individuals. This review seeks to evaluate the comparative validity, reliability, and feasibility of echocardiography and biomarker testing for the screening, diagnosis, and management of PAH; to clarify whether the use of echocardiography or biomarkers affects decisionmaking and clinical outcomes; and to determine which medications are effective for treating PAH and whether combination therapy is more effective than monotherapy. DATA SOURCES: We searched PubMed(r), Embase(r), and the Cochrane Database of Systematic Reviews for relevant English-language comparative studies. REVIEW METHODS: Two investigators screened each abstract and full-text article for inclusion, abstracted data, rated quality and applicability, and graded the strength of evidence. Random-effects models were used to compute summary estimates of effect where several similar studies provided estimates. RESULTS: Sixty studies involving 7,096 patients evaluated biomarker tests, echocardiography, or both to screen for PAH. Symptom status of study populations consisted of asymptomatic (3 studies; 481 patients), symptomatic (41 studies; 4,394 patients), mixed (8 studies; 1,186 patients), and symptoms not described (8 studies; 1,035 patients). N-terminal pro-B-type natriuretic peptide (NT-proBNP) showed moderate correlation with right heart catheterization (RHC) hemodynamic measures and a great deal of variability between studies in its diagnostic accuracy and discrimination; however, one good-quality prospective cohort study suggested that biomarker testing with NT-proBNP might be useful in ruling out PAH in patients with symptoms suggestive of PAH who have elevated systolic pulmonary artery pressure (sPAP) by echocardiography. No data are available regarding combined echocardiography and biomarker screening in asymptomatic patients at high risk for PAH. Echocardiography estimates of pulmonary artery pressures (sPAP, tricuspid gradient [TG], and tricuspid regurgitant jet velocity [TRV]) and PVR (TRV/velocity-time integral of right ventricular outflow tract [VTIRVOT]) demonstrated good accuracy in screening for PAH, but accuracy varied with the prevalence of PAH in study populations. Ninety-nine studies involving 8,655 patients evaluated biomarker tests, echocardiography, or both to evaluate severity or prognosis and followed progression of disease or response to therapy. B-type natriuretic peptide (BNP) showed moderate correlation with most RHC measures (mean pulmonary artery pressure [mPAP], PVR, cardiac index, right atrial pressure [RAP]) and clinical measures of disease severity (6-minute walk distance [6MWD]) and showed weak correlation with pulmonary capillary wedge pressure (PCWP), indicating that BNP levels alone could not serve as an accurate surrogate marker for disease severity. Echocardiography-derived sPAP showed strong correlation with RHC-sPAP with a precise summary effect estimate, although there was a great deal of heterogeneity of results among individual studies. BNP level (summary hazard ratio [HR] 2.42; 95% confidence interval [CI], 1.72 to 3.41) and presence of pericardial effusion were strong predictors of mortality (summary HR 2.43; 95% CI, 1.57 to 3.77) RA size and uric acid were also predictive of mortality, but fractional area change (FAC) showed no significant ability to predict mortality, and data on TAPSE were insufficient. Thirty-seven studies involving 4,192 patients assessed the effectiveness of drug treatments for PAH in adults. Few deaths were observed in these limited duration studies, leading to wide CIs and lack of statistical power to detect a mortality difference associated with treatment. All drug classes demonstrated increases in 6WMD when compared with placebo, but comparisons between agents were inconclusive. Combination therapy also showed improved 6WMD compared with monotherapy, but the diversity of treatment regimens and the small number of combination therapy trials again make comparisons between specific regimens inconclusive. The odds ratio (OR) of hospitalization was lower in patients taking endothelin receptor antagonists or phosphodiesterase-5 inhibitors compared with placebo (OR 0.34 and 0.48, respectively), while the reduction in patients taking prostanoids compared with placebo was similar but not statistically significant. Each drug class showed a favorable impact on at least two of the three hemodynamic outcomes: cardiac index, mPAP, and PVR. The applicability of these findings is limited by the relative lack of diagnostic studies among asymptomatic patients and, in prognostic and diagnostic studies, inadequate description and apparent diversity of disease etiology and severity. CONCLUSIONS: Further confirmation is needed to determine if the combination of echocardiography and the biomarker NT-proBNP is sufficiently accurate to rule out PAH when testing symptomatic patients. In asymptomatic populations, more research is needed to permit conclusions regarding their effectiveness for screening. BNP, RA size, presence of pericardial effusion, and uric acid had prognostic value in patients with PAH, but other echocardiographic parameters and biomarkers either were not predictive or had insufficient data. Although no studies were powered to detect a mortality reduction, monotherapy was associated with improved 6MWD and reduced hospitalization rates. Comparisons of different drug combinations were inconclusive regarding a mortality reduction but suggested an improvement in 6MWD when a second drug was added to existing monotherapy.

OBJECTIVES: For patients with unstable angina or non-ST elevation myocardial infarction (UA/NSTEMI), antiplatelet and anticoagulant medications are prescribed to reduce and prevent ischemic events and mortality. There is uncertainty about the optimal dosing and timing of these medications to balance ischemic risk and bleeding risk across different treatment strategies (early invasive, initial conservative, and postdischarge). DATA SOURCES: We searched PubMed(r), Embase(r), and the Cochrane Database of Systematic Reviews for relevant English-language comparative studies. REVIEW METHODS: Two investigators screened each abstract and full-text article for inclusion, abstracted data, rated quality and applicability, and graded evidence. When possible, random-effects meta-analysis was used to compute summary estimates of effects. RESULTS: Our review included 174 studies (301 articles); 87 studies were relevant to early invasive management, 33 were relevant to initial conservative management, and 70 were relevant to the postdischarge setting. Patients undergoing an early invasive approach. Upstream (precatheterization) treatment using glycoprotein IIb/IIIa inhibitors (GPIs) was associated with lower rates of revascularization (odds ratio [OR] 0.77; 95% confidence interval [CI], 0.65 to 0.92) but higher risk of major bleeding events (OR 1.24; 95% CI, 1.08 to 1.43) at 30 days compared with deferred (periprocedural) GPI treatment (high strength of evidence [SOE]). This higher risk of bleeding from upstream GPI administration also occurred with either pretreatment (OR 1.49; 95% CI, 1.10 to 2.01; moderate SOE) or deferred clopidogrel administration (OR 1.27; 95% CI, 1.08 to 1.50; high SOE). Compared with clopidogrel, prasugrel reduced rates of cardiovascular death, myocardial infarction, or stroke at 30 days (5.7% prasugrel vs. 7.4% clopidogrel; moderate SOE). After 1 year, in a subgroup of patients who all had UA/NSTEMI, prasugrel reduced rates of the same composite endpoint compared with clopidogrel (9.9% prasugrel vs. 12.1% clopidogrel), as did ticagrelor (10.6% ticagrelor vs. 12.6% clopidogrel) (moderate SOE). Bivalirudin reduced major bleeding events at 30 days compared with heparin in several clinical scenarios: with planned GPI use (OR 0.52; 95% CI, 0.43 to 0.63); without planned GPI use (OR 0.63; 95% CI, 0.47 to 0.85; both high SOE); and in patients treated with clopidogrel before undergoing percutaneous coronary intervention (OR 0.64; 95% CI, 0.49 to 0.85; moderate SOE). Bivalirudin also reduced minor bleeding events at 30 days compared with heparin plus GPI (OR 0.49; 95% CI, 0.42 to 0.59; high SOE). Patients undergoing an initial conservative approach. In randomized trials, enoxaparin reduced composite ischemic events (OR 0.84; 95% CI, 0.76 to 0.93; high SOE) and myocardial infarction (OR 0.85; 95% CI, 0.76 to 0.95; moderate SOE) at around 30 days compared with unfractionated heparin. The addition of GPIs to unfractionated heparin reduced the rate of mortality up to 30 days (OR 0.80; 95% CI, 0.67 to 0.96), but minor bleeding rates were increased (OR 1.62; 95% CI, 1.20 to 2.19; both high SOE). Postdischarge treatment. Dual antiplatelet therapy (DAPT) reduced the rates of composite ischemic outcomes (ORs/relative risks ranging from 0.69 to 0.80; in-hospital, 9 months, and 1 year) and nonfatal myocardial infarction (DAPT 2.3% to 5.8% vs. aspirin 3.0% to 8.5%; 9 months and 1 year) compared with single antiplatelet therapy (high SOE). Meta-analyses using adjusted or propensity-scored hazard ratios from observational studies showed an association between proton pump inhibitor (PPI) use (any type with dual antiplatelet use) and increased rates of composite ischemic endpoints, death, nonfatal myocardial infarction, stroke, revascularization, stent thrombosis, and major bleeding. (Most outcomes were measured around 1 year and rated low SOE, and ratings were downgraded since the findings conflicted with the few randomized trials of omeprazole.) However PPIs with DAPT use reduced rates of upper gastrointestinal bleeding (moderate SOE). LIMITATIONS: This review was limited to comparative studies of antiplatelet and anticoagulant treatments, many of which did not separate findings by treatment approach (invasive, conservative, postdischarge) and included a mix of UA/NSTEMI and acute coronary syndrome populations. Also, different definitions of composite endpoints made quantitative analysis less feasible. Few trials of percutaneous coronary intervention reported long-term outcomes, and very few studies reported findings in the subpopulations of interest. CONCLUSIONS: The number of studies available for each comparison was relatively small, and the preponderance of observational studies made the findings for some comparisons inconclusive. Further study is needed to determine the effectiveness and safety of newer agents in combination with other antiplatelet and anticoagulant strategies. Uncertainty remains about the optimal dosing, timing, duration, and combinations of these options, especially in subpopulations of interest (e.g., the elderly, patients with diabetes, women, obese patients, and people with comorbid illness).

OBJECTIVES: Oral anticoagulation with vitamin K antagonists (VKAs) has long been the gold standard therapy for stroke prevention in nonvalvular atrial fibrillation (AF). Limitations in monitoring and compliance of VKAs have fueled the development of new antithrombotic strategies, devices, and oral anticoagulants, including oral direct thrombin inhibitors and factor Xa inhibitors. This review updates previous reviews, particularly with regard to these newer treatment options and the optimal risk stratification tools for stroke and bleeding prediction. DATA SOURCES: We searched PubMed(r), Embase(r), and the Cochrane Database of Systematic Reviews for relevant English-language comparative studies published from January 1, 2000, to August 14, 2012. REVIEW METHODS: Two investigators screened each abstract and full-text article for inclusion, abstracted data, rated quality and applicability, and graded evidence. When possible, random-effects models were used to compute summary estimates of effects. RESULTS: Our review included 122 articles (92 unique studies), comprising 37 studies relevant to predicting thromboembolic risk, 17 relevant to predicting bleeding risk, 43 relevant to interventions for preventing thromboembolic events, 13 relevant to anticoagulation strategies in patients undergoing invasive procedures, and no studies relevant to strategies for switching between warfarin and novel oral anticoagulants or to stroke prevention after a hemorrhagic event. Across the Key Questions addressing prediction of stroke and bleeding risk, evidence was limited by variability in reporting and in underlying treatment of AF. Data suggest that the continuous CHADS2 (Congestive heart failure, Hypertension, Age e75, Diabetes mellitus, prior Stroke/transient ischemic attack [2 points]) and continuous CHA2DS2-VASc (Congestive heart failure/left ventricular ejection fraction d40%, Hypertension, Age e75 [2 points], Diabetes mellitus, prior Stroke/transient ischemic attack/thromboembolism [2 points], Vascular disease, Age 65-74, Sex category female) scores have the greatest discrimination for stroke risk (c-statistic 0.71 [95% confidence interval (CI), 0.66 to 0.75], and c-statistic 0.70 [95% CI 0.66 to 0.75], respectively; low strength of evidence for both scores) and that the HAS-BLED (Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile international normalized ratio, Elderly [>65 years], Drugs/alcohol concomitantly) score has the greatest discrimination for bleeding risk (moderate strength of evidence). Evidence evaluating interventions for stroke prevention was limited by the small number of studies for specific comparisons and lack of direct comparisons of novel anticoagulants, although many included studies were good-quality randomized controlled trials involving more than 5,000 patients. We found that a factor IIa inhibitor (dabigatran 150 mg) was superior to warfarin in reducing the incidence of stroke (including hemorrhagic) or systemic embolism (relative risk [RR] 0.66; 95% CI 0.53 to 0.82), with no significant difference in the occurrence of major bleeding (RR 0.93; 95% CI 0.81 to 1.07) (high strength of evidence for both outcomes). The Xa inhibitor rivaroxaban was noninferior to warfarin in preventing stroke or systemic embolism (moderate strength of evidence), with similar rates of major bleeding and death (high strength of evidence). The Xa inhibitor apixaban was superior to warfarin in reducing the incidence of stroke or systemic embolism (hazard ratio [HR] 0.79; 95% CI 0.66 to 0.95; high strength of evidence); major bleeding (HR 0.69; 95% CI 0.60 to 0.80; high strength of evidence); and all-cause mortality (HR 0.89; 95% CI 0.80 to 0.998; moderate strength of evidence). Apixaban was also superior to aspirin in reducing the incidence of stroke or systemic embolism (HR 0.45; 95% CI 0.32 to 0.62), with similar hemorrhagic events, including major bleeding (HR 1.13; 95% CI 0.74 to 1.75), in patients who are not suitable for oral anticoagulation (high strength of evidence for both outcomes). However, no studies directly compared the new therapies. Evidence for patients undergoing invasive procedures, switching among anticoagulant therapies, and starting or restarting anticoagulant therapy after previous major bleeding events was insufficient. CONCLUSIONS: Overall, we found that CHADS2 and CHA2DS2-VASc scores have the best discrimination ability for stroke events in patients with AF among the risk scores we reviewed, whereas HAS-BLED provides the best discrimination of bleeding risk. Imaging tools require further evidence in regard to their appropriate use in clinical decisionmaking. Improved evidence of the use of these scores among patients on therapy is also required. Newer anticoagulants show early promise of reducing stroke and bleeding events when compared with warfarin, and apixaban shows safety and efficacy in patients who are not candidates for warfarin. However, further studies are required for key clinical scenarios involving anticoagulation use and procedures, switching or bridging therapies, and when to start anticoagulation after a hemorrhagic event.

This report describes an evaluation of the available scientific evidence on direct non-invasive imaging tests (NITs) for coronary artery disease. In particular, we focus on six key questions provided by the Agency for Healthcare Research and Quality (AHRQ) and the Centers for Medicare and Medicaid Services (CMS). The objective of this report is to provide background information to the Medicare Coverage Advisory Committee (MCAC) in their review of these questions during their May 2006 meeting. The six key questions examine the degree to which current evidence supports confident judgments about the use of NITs in the assessment of coronary anatomy in clinical practice. The two NITs that are examined in detail in this report are computed tomographic angiography (CTA) and magnetic resonance angiography (MRA) for evaluating native coronary arteries. In addition, we consider technologies on the horizon, as well as the general issue of establishing the value of NITs in specific clinical contexts in which coronary disease is being considered.

Selective reporting can bias estimates of effect, yet methods to detect such biases are limited. One method is to search ClinicalTrials.gov (CT.gov) and (a) compare studies identified there to published studies (to detect publication bias) and (b) compare planned analyses and outcomes reported in CT.gov to those reported in the final publication (to detect reporting bias). While conceptually sound, this approach may be labor-intensive, and its utility is uncertain. The overall goal of this project was to evaluate the utility of CT.gov for detecting selective reporting, and to determine the impact of selective reporting on the estimates of treatment effect. A secondary goal was to estimate the person-hours required to complete these analyses. To accomplish this goal, an ongoing systematic review entitled Management of Infertility was used to explore differences between information from published sources included in the review and CT.gov. The objectives of this systematic review are to evaluate the comparative safety and effectiveness of treatment strategies for: a) women of reproductive age (18-44) who are subfertile/infertile due to polycystic ovary syndrome (PCOS), endometriosis, unknown reasons, or tubal or peritoneal factors; or b) couples with male factor infertility; and evaluate short- and long-term health outcomes of gamete donors in infertility.

Section 641 of the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) calls for a demonstration that would pay for drugs and biologicals that are prescribed as replacements for drugs currently covered under Medicare Part B. The demonstration project will be national in scope and will be limited to 50,000 beneficiaries or $500,000,000 in funding, whichever comes first. Forty percent of the funding for this demonstration will be reserved for oral anti-neoplastic drugs. CMS has requested an assessment of the efficacy of selected oral cancer therapies included in the demonstration relative to drugs currently covered under Medicare Part B. This assessment will provide information that will be used to evaluate the likely effects of the demonstration on patient outcomes and may also provide underlying information to be used for cost-effectiveness analyses that will be completed by CMS. The scope of the assessment will be limited to the following demonstration drugs and conditions: ® Imatinib for treatment of chronic myeloid leukemia;® Imatinib for treatment of gastrointestinal stromal cancer;® Gefitinib for treatment of non-small cell lung cancer;® Thalidomide for treatment of multiple myeloma. This report is responsive to the third item: an assessment of gefitinib for the treatment of non-small cell lung cancer. After work on this report was begun, the parameters were modified to include the closely related orally administered epidermal growth factor tyrosine kinase inhibitor, erlotinib. This was done for three reasons: 1) pivotal trial data suggested that gefitinib had little clinical efficacy; 2) a large number of studies were forthcoming on erlotinib suggesting that this drug may have greater clinical efficacy than gefitinib; and 3) erlotinib was added to the demonstration project in January 2005.