Performance of two large language models for data extraction in evidence synthesis.

Accurate data extraction is a key component of evidence synthesis and critical to valid results. The advent of publicly available large language models (LLMs) has generated interest in these tools for evidence synthesis and created uncertainty about the choice of LLM. We compare the performance of two widely available LLMs (Claude 2 and GPT-4) for extracting pre-specified data elements from 10 published articles included in a previously completed systematic review. We use prompts and full study PDFs to compare the outputs from the browser versions of Claude 2 and GPT-4. GPT-4 required use of a third-party plugin to upload and parse PDFs. Accuracy was high for Claude 2 (96.3%). The accuracy of GPT-4 with the plug-in was lower (68.8%); however, most of the errors were due to the plug-in. Both LLMs correctly recognized when prespecified data elements were missing from the source PDF and generated correct information for data elements that were not reported explicitly in the articles. A secondary analysis demonstrated that, when provided selected text from the PDFs, Claude 2 and GPT-4 accurately extracted 98.7% and 100% of the data elements, respectively. Limitations include the narrow scope of the study PDFs used, that prompt development was completed using only Claude 2, and that we cannot guarantee the open-source articles were not used to train the LLMs. This study highlights the potential for LLMs to revolutionize data extraction but underscores the importance of accurate PDF parsing. For now, it remains essential for a human investigator to validate LLM extractions.

Data extraction for evidence synthesis using a large language model: A proof-of-concept study.

Data extraction is a crucial, yet labor-intensive and error-prone part of evidence synthesis. To date, efforts to harness machine learning for enhancing efficiency of the data extraction process have fallen short of achieving sufficient accuracy and usability. With the release of large language models (LLMs), new possibilities have emerged to increase efficiency and accuracy of data extraction for evidence synthesis. The objective of this proof-of-concept study was to assess the performance of an LLM (Claude 2) in extracting data elements from published studies, compared with human data extraction as employed in systematic reviews. Our analysis utilized a convenience sample of 10 English-language, open-access publications of randomized controlled trials included in a single systematic review. We selected 16 distinct types of data, posing varying degrees of difficulty (160 data elements across 10 studies). We used the browser version of Claude 2 to upload the portable document format of each publication and then prompted the model for each data element. Across 160 data elements, Claude 2 demonstrated an overall accuracy of 96.3% with a high test-retest reliability (replication 1: 96.9%; replication 2: 95.0% accuracy). Overall, Claude 2 made 6 errors on 160 data items. The most common errors (n = 4) were missed data items. Importantly, Claude 2's ease of use was high; it required no technical expertise or labeled training data for effective operation (i.e., zero-shot learning). Based on findings of our proof-of-concept study, leveraging LLMs has the potential to substantially enhance the efficiency and accuracy of data extraction for evidence syntheses.

Screening for Speech and Language Delay and Disorders in Children 5 Years or Younger: Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Children with speech and language difficulties are at risk for learning and behavioral problems. Objective: To review the evidence on screening for speech and language delay or disorders in children 5 years or younger to inform the US Preventive Services Task Force. Data Sources: PubMed/MEDLINE, Cochrane Library, PsycInfo, ERIC, Linguistic and Language Behavior Abstracts (ProQuest), and trial registries through January 17, 2023; surveillance through November 24, 2023. Study Selection: English-language studies of screening test accuracy, trials or cohort studies comparing screening vs no screening; randomized clinical trials (RCTs) of interventions. Data Extraction and Synthesis: Dual review of abstracts, full-text articles, study quality, and data extraction; results were narratively summarized. Main Outcomes and Measures: Screening test accuracy, speech and language outcomes, school performance, function, quality of life, and harms. Results: Thirty-eight studies in 41 articles were included (N = 9006). No study evaluated the direct benefits of screening vs no screening. Twenty-one studies (n = 7489) assessed the accuracy of 23 different screening tools that varied with regard to whether they were designed to be completed by parents vs trained examiners, and to screen for global (any) language problems vs specific skills (eg, expressive language). Three studies assessing parent-reported tools for expressive language skills found consistently high sensitivity (range, 88%-93%) and specificity (range, 88%-85%). The accuracy of other screening tools varied widely. Seventeen RCTs (n = 1517) evaluated interventions for speech and language delay or disorders, although none enrolled children identified by routine screening in primary care. Two RCTs evaluating relatively intensive parental group training interventions (11 sessions) found benefit for different measures of expressive language skills, and 1 evaluating a less intensive intervention (6 sessions) found no difference between groups for any outcome. Two RCTs (n = 76) evaluating the Lidcombe Program of Early Stuttering Intervention delivered by speech-language pathologists featuring parent training found a 2.3% to 3.0% lower proportion of syllables stuttered at 9 months compared with the control group when delivered in clinic and via telehealth, respectively. Evidence on other interventions was limited. No RCTs reported on the harms of interventions. Conclusions and Relevance: No studies directly assessed the benefits and harms of screening. Some parent-reported screening tools for expressive language skills had reasonable accuracy for detecting expressive language delay. Group parent training programs for speech delay that provided at least 11 parental training sessions improved expressive language skills, and a stuttering intervention delivered by speech-language pathologists reduced stuttering frequency.

Rapid reviews methods series: Guidance on assessing the certainty of evidence.

This paper is part of a series of methodological guidance from the Cochrane Rapid Reviews Methods Group. Rapid reviews (RRs) use modified systematic review methods to accelerate the review process while maintaining systematic, transparent and reproducible methods. This paper addresses considerations for rating the certainty of evidence (COE) in RRs. We recommend the full implementation of GRADE (Grading of Recommendations, Assessment, Development and Evaluation) for Cochrane RRs if time and resources allow.If time or other resources do not permit the full implementation of GRADE, the following recommendations can be considered: (1) limit rating COE to the main intervention and comparator and limit the number of outcomes to critical benefits and harms; (2) if a literature review or a Delphi approach to rate the importance of outcomes is not feasible, rely on informal judgements of knowledge users, topic experts or team members; (3) replace independent rating of the COE by two reviewers with single-reviewer rating and verification by a second reviewer and (4) if effect estimates of a well-conducted systematic review are incorporated into an RR, use existing COE grades from such a review. We advise against changing the definition of COE or the domains considered part of the GRADE approach for RRs.

Folic Acid Supplementation to Prevent Neural Tube Defects: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Neural tube defects are among the most common birth defects in the US. Objective: To review new evidence on the benefits and harms of folic acid supplementation for the prevention of neural tube defects to inform the US Preventive Services Task Force. Evidence Review: Sources included PubMed, Cochrane Library, Embase, and trial registries from July 1, 2015, through July 2, 2021; references; and experts, with surveillance through February 10, 2023. Two investigators independently reviewed English-language randomized studies and nonrandomized cohort studies in very highly developed countries that focused on the use of folic acid supplementation for the prevention of neural tube defect-affected pregnancies; methodological quality was dually and independently assessed. Findings: Twelve observational studies (reported in 13 publications) were eligible for this limited update (N = 1 244 072). Of these, 3 studies (n = 990 372) reported on the effect of folic acid supplementation on neural tube defects. For harms, 9 studies were eligible: 1 randomized clinical trial (n = 431) reported on variations in twin delivery, 7 observational studies (n = 761 125) reported on the incidence of autism spectrum disorder, and 1 observational study (n = 429 004) reported on maternal cancer. Two cohort studies and 1 case-control study newly identified in this update reported on the association between folic acid supplementation and neural tube defects (n = 990 372). One cohort study reported a statistically significant reduced risk of neural tube defects associated with folic acid supplementation taken before pregnancy (adjusted relative risk [aRR], 0.54 [95% CI, 0.31-0.91]), during pregnancy (aRR, 0.62 [95% CI, 0.39-0.97]), and before and during pregnancy (aRR, 0.49 [95% CI, 0.29-0.83]), but this association occurred for only the later of 2 periods studied (2006-2013 and not 1999-2005). No other statistically significant benefits were reported overall. No study reported statistically significant harms (multiple gestation, autism, and maternal cancer) associated with pregnancy-related folic acid exposure. Conclusions and Relevance: New evidence from observational studies provided additional evidence of the benefit of folic acid supplementation for preventing neural tube defects and no evidence of harms related to multiple gestation, autism, or maternal cancer. The new evidence was consistent with previously reviewed evidence on benefits and harms.

Screening for Latent Tuberculosis Infection in Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Latent tuberculosis infection (LTBI) can progress to active tuberculosis disease, causing morbidity and mortality. Objective: To review the evidence on benefits and harms of screening for and treatment of LTBI in adults to inform the US Preventive Services Task Force (USPSTF). Data Sources: PubMed/MEDLINE, Cochrane Library, and trial registries through December 3, 2021; references; experts; literature surveillance through January 20, 2023. Study Selection: English-language studies of LTBI screening, LTBI treatment, or accuracy of the tuberculin skin test (TST) or interferon-gamma release assays (IGRAs). Studies of LTBI screening and treatment for public health surveillance or disease management were excluded. Data Extraction and Synthesis: Dual review of abstracts, full-text articles, and study quality; qualitative synthesis of findings; meta-analyses conducted when a sufficient number of similar studies were available. Main Outcomes and Measures: Screening test accuracy; development of active tuberculosis disease, transmission, quality of life, mortality, and harms. Results: A total of 113 publications were included (112 studies; N = 69 009). No studies directly evaluated the benefits and harms of screening. Pooled estimates for sensitivity of the TST were 0.80 (95% CI, 0.74-0.87) at the 5-mm induration threshold, 0.81 (95% CI, 0.76-0.87) at the 10-mm threshold, and 0.60 (95% CI, 0.46-0.74) at the 15-mm threshold. Pooled estimates for sensitivity of IGRA tests ranged from 0.81 (95% CI, 0.79-0.84) to 0.90 (95% CI, 0.87-0.92). Pooled estimates for specificity of screening tests ranged from 0.95 to 0.99. For treatment of LTBI, a large (n = 27 830), good-quality randomized clinical trial found a relative risk (RR) for progression to active tuberculosis at 5 years of 0.35 (95% CI, 0.24-0.52) for 24 weeks of isoniazid compared with placebo (number needed to treat, 112) and an increase in hepatotoxicity (RR, 4.59 [95% CI, 2.03-10.39]; number needed to harm, 279). A previously published meta-analysis reported that multiple regimens were efficacious compared with placebo or no treatment. Meta-analysis found greater risk for hepatotoxicity with isoniazid than with rifampin (pooled RR, 4.22 [95% CI, 2.21-8.06]; n = 7339). Conclusions and Relevance: No studies directly evaluated the benefits and harms of screening for LTBI compared with no screening. TST and IGRAs were moderately sensitive and highly specific. Treatment of LTBI with recommended regimens reduced the risk of progression to active tuberculosis. Isoniazid was associated with higher rates of hepatotoxicity than placebo or rifampin.

End-user understanding of qualitative comparative analysis used within evidence synthesis: A mixed-methods study.

BACKGROUND: Enhanced uptake of systematic reviews that use qualitative comparative analyses (QCA) requires knowing how end-users interpret such findings. The study purpose was to identify effective approaches to communicating results from a QCA within a systematic review. METHODS: Sequential exploratory mixed methods design; thematic analysis of interviews with 11 end-users followed by a randomized experiment with 254 participants that provided QCA results for a hypothetical review presented through three formats (text, table, and figure). A survey administered after the experiment assessed subjective and objective comprehension of QCA results. RESULTS: Interview themes included use of jargon; appropriate use of appendices, tables, figures; and integration of QCA results within the systematic review. In the experiment, we observed a significant difference (p = 0.035) in subjective comprehension across the three presentation formats. Participants randomized to the figure and text formats scored higher compared to the table. No significant differences were observed for objective comprehension overall (p = 0.11). However, for parameter interpretation (a unique component of QCA results), scores among participants that received the figure format were significantly higher than scores for participants who received the text (p = 0.001) or table (p = 0.004). No significant differences (p = 0.09) were observed in objective comprehension for configuration interpretation. CONCLUSIONS: End-users of systematic reviews saw value in the use of QCA, but unfamiliar methods and terminology were barriers to full understanding of the findings. When presenting results, a figure format appears to be superior to text or table formats based on measures of subjective comprehension and some measures of objective comprehension.

Hormone Therapy for the Primary Prevention of Chronic Conditions in Postmenopausal Persons: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: It is uncertain whether hormone therapy should be used for the primary prevention of chronic conditions such as heart disease, osteoporosis, or some types of cancers. Objective: To update evidence for the US Preventive Services Task Force on the benefits and harms of hormone therapy in reducing risks for chronic conditions. Data Sources: PubMed/MEDLINE, Cochrane Library, EMBASE, and trial registries from January 1, 2016, through October 12, 2021; surveillance through July 2022. Study Selection: English-language randomized clinical trials and prospective cohort studies of fair or good quality. Data Extraction and Synthesis: Dual review of abstracts, full-text articles, and study quality; meta-analyses when at least 3 similar studies were available. Main Outcomes and Measures: Morbidity and mortality related to chronic conditions; health-related quality of life. Results: Twenty trials (N = 39 145) and 3 cohort studies (N = 1 155 410) were included. Participants using estrogen only compared with placebo had significantly lower risks for diabetes over 7.1 years (1050 vs 903 cases; 134 fewer [95% CI, 18-237]) and fractures over 7.2 years (1024 vs 1413 cases; 388 fewer [95% CI, 277-489]) per 10 000 persons. Risks per 10 000 persons were statistically significantly increased for gallbladder disease over 7.1 years (1113 vs 737 cases; 377 more [95% CI, 234-540]), stroke over 7.2 years (318 vs 239 cases; 79 more [95% CI, 15-159]), venous thromboembolism over 7.2 years (258 vs 181 cases; 77 more [95% CI, 19-153]), and urinary incontinence over 1 year (2331 vs 1446 cases; 885 more [95% CI, 659-1135]). Participants using estrogen plus progestin compared with placebo experienced significantly lower risks, per 10 000 persons, for colorectal cancer over 5.6 years (59 vs 93 cases; 34 fewer [95% CI, 9-51]), diabetes over 5.6 years (403 vs 482 cases; 78 fewer [95% CI, 15-133]), and fractures over 5 years (864 vs 1094 cases; 230 fewer [95% CI, 66-372]). Risks, per 10 000 persons, were significantly increased for invasive breast cancer (242 vs 191 cases; 51 more [95% CI, 6-106]), gallbladder disease (723 vs 463 cases; 260 more [95% CI, 169-364]), stroke (187 vs 135 cases; 52 more [95% CI, 12-104]), and venous thromboembolism (246 vs 126 cases; 120 more [95% CI, 68-185]) over 5.6 years; probable dementia (179 vs 91 cases; 88 more [95% CI, 15-212]) over 4.0 years; and urinary incontinence (1707 vs 1145 cases; 562 more [95% CI, 412-726]) over 1 year. Conclusions and Relevance: Use of hormone therapy in postmenopausal persons for the primary prevention of chronic conditions was associated with some benefits but also with an increased risk of harms.

Screening for Depression and Suicide Risk in Children and Adolescents: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Depression, suicidal ideation, and self-harm behaviors in youth are associated with functional impairment and suicide. Objective: To review the evidence on screening for depression or suicide risk in children and adolescents to inform the US Preventive Services Task Force (USPSTF). Data Sources: PubMed, Cochrane Library, PsycINFO, CINAHL, and trial registries through July 19, 2021; references, experts, and surveillance through June 1, 2022. Study Selection: English-language, randomized clinical trials (RCTs) of screening for depression or suicide risk; diagnostic test accuracy studies; RCTs of psychotherapy and first-line pharmacotherapy; RCTs, observational studies, and systematic reviews reporting harms. Data Extraction and Synthesis: Two reviewers assessed titles/abstracts, full-text articles, and study quality and extracted data; when at least 3 similar studies were available, meta-analyses were conducted. Main Outcomes and Measures: Test accuracy, symptoms, response, remission, loss of diagnosis, mortality, functioning, suicide-related events, and adverse events. Results: Twenty-one studies (N = 5433) were included for depression and 19 studies (N = 6290) for suicide risk. For depression, no studies reported on the direct effects of screening on health outcomes, and 7 studies (n = 3281) reported sensitivity of screening instruments ranging from 0.59 to 0.94 and specificity from 0.38 to 0.96. Depression treatment with psychotherapy was associated with improved symptoms (Beck Depression Inventory pooled standardized mean difference, -0.58 [95% CI, -0.83 to -0.34]; n = 471; 4 studies; and Hamilton Depression Scale pooled mean difference, -2.25 [95% CI, -4.09 to -0.41]; n = 262; 3 studies) clinical response (3 studies with statistically significant results using varying thresholds), and loss of diagnosis (relative risk, 1.73 [95% CI, 1.00 to 3.00; n = 395; 4 studies). Pharmacotherapy was associated with improvement on symptoms (Children's Depression Rating Scale-Revised mean difference, -3.76 [95% CI, -5.95 to -1.57; n = 793; 3 studies), remission (relative risk, 1.20 [95% CI, 1.00 to 1.45]; n = 793; 3 studies) and functioning (Children's Global Assessment Scale pooled mean difference, 2.60 (95% CI, 0.78 to 4.42; n = 793; 3 studies). Other outcomes were not statistically significantly different. Differences in suicide-related outcomes and adverse events for pharmacotherapy when compared with placebo were not statistically significant. For suicide risk, no studies reported on the direct benefits of screening on health outcomes, and 2 RCTs (n = 2675) reported no harms of screening. One study (n = 581) reported on sensitivity of screening, ranging from 0.87 to 0.91; specificity was 0.60. Sixteen RCTs (n = 3034) reported on suicide risk interventions. Interventions were associated with lower scores for the Beck Hopelessness Scale (pooled mean difference, -2.35 [95% CI, -4.06 to -0.65]; n = 644; 4 RCTs). Findings for other suicide-related outcomes were mixed or not statistically significantly different. Conclusion and Relevance: Indirect evidence suggested that some screening instruments were reasonably accurate for detecting depression. Psychotherapy and pharmacotherapy were associated with some benefits and no statistically significant harms for depression, but the evidence was limited for suicide risk screening instruments and interventions.

Screening for Anxiety in Children and Adolescents: Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Anxiety in children and adolescents is associated with impaired functioning, educational underachievement, and future mental health conditions. Objective: To review the evidence on screening for anxiety in children and adolescents to inform the US Preventive Services Task Force. Data Sources: PubMed, Cochrane Library, PsycINFO, CINAHL, and trial registries through July 19, 2021; references, experts, and surveillance through June 1, 2022. Study Selection: English-language, randomized clinical trials (RCTs) of screening; diagnostic test accuracy studies; RCTs of cognitive behavioral therapy (CBT) or US Food and Drug Administration-approved pharmacotherapy; RCTs, observational studies, and systematic reviews reporting harms. Data Extraction and Synthesis: Two reviewers assessed titles/abstracts, full-text articles, and study quality and extracted data; when at least 3 similar studies were available, meta-analyses were conducted. Main Outcomes and Measures: Test accuracy, symptoms, response, remission, loss of diagnosis, all-cause mortality, functioning, suicide-related symptoms or events, adverse events. Results: Thirty-nine studies (N = 6065) were included. No study reported on the direct benefits or harms of screening on health outcomes. Ten studies (n = 3260) reported the sensitivity of screening instruments, ranging from 0.34 to 1.00, with specificity ranging from 0.47 to 0.99. Twenty-nine RCTs (n = 2805) reported on treatment: 22 on CBT, 6 on pharmacotherapy, and 1 on CBT, sertraline, and CBT plus sertraline. CBT was associated with gains on several pooled measures of symptom improvement (magnitude of change varied by outcome measure), response (pooled relative risk [RR], 1.89 [95% CI, 1.17 to 3.05]; n = 606; 6 studies), remission (RR, 2.68 [95% CI, 1.48 to 4.88]; n = 321; 4 studies), and loss of diagnosis (RR range, 3.02-3.09) when compared with usual care or wait-list controls. The evidence on functioning for CBT was mixed. Pharmacotherapy, when compared with placebo, was associated with gains on 2 pooled measures of symptom improvement-mean difference (Pediatric Anxiety Rating Scale mean difference, -4.0 [95% CI, -5.5 to -2.5]; n = 726; 5 studies; and Clinical Global Impression-Severity scale mean difference, -0.84 [95% CI, -1.13 to -0.55]; n = 550; 4 studies) and response (RR, 2.11 [95% CI, 1.58 to 2.98]; n = 370; 5 studies)-but was mixed on measures of functioning. Eleven RCTs (n = 1293) reported harms of anxiety treatments. Suicide-related harms were rare, and the differences were not statistically significantly different. Conclusions and Relevance: Indirect evidence suggested that some screening instruments were reasonably accurate. CBT and pharmacotherapy were associated with benefits; no statistically significant association with harms was reported.

The impact of interactive advertising on consumer engagement, recall, and understanding: A scoping systematic review for informing regulatory science.

We conducted a scoping systematic review with respect to how consumer engagement with interactive advertising is evaluated and if interactive features influence consumer recall, awareness, or comprehension of product claims and risk disclosures for informing regulatory science. MEDLINE, PsycINFO, Business Source Corporate, and SCOPUS were searched for original research published from 1997 through February 2021. Two reviewers independently screened titles/abstracts and full-text articles for inclusion. Outcomes were abstracted into a structured abstraction form. We included 32 studies overall. The types of interactive ads evaluated included website banner and pop up ads, search engine ads, interactive TV ads, advergames, product websites, digital magazine ads, and ads on social network sites. Twenty-three studies reported objective measures of engagement using observational analyses or laboratory-based experiments. In nine studies evaluating the association between different interactivity features and outcomes, the evidence was mixed on whether more interactivity improves or worsens recall and comprehension. Studies vary with respect to populations, designs, ads evaluated, and outcomes assessed.

Minimally invasive sacroiliac joint fusion for chronic sacroiliac joint pain: a systematic review.

BACKGROUND CONTEXT: Sacroiliac (SI) joint pain causes significant disability and impairment to quality of life (QOL). Minimally invasive SI joint fusion is increasingly used to relieve chronic SI joint pain among patients who do not respond to nonsurgical treatment. PURPOSE: To systematically review the existing literature to assess the effectiveness and safety of minimally invasive SI joint fusion. STUDY DESIGN/SETTING: Systematic review. DATA SOURCES: PubMed, Embase, Cochrane, and a clinical trial registry from database inception to June 30, 2021. STUDY SELECTION: Eligible studies were primary research studies published in the English language, enrolled adults with SI joint pain, and compared SI joint fusion to nonsurgical interventions or alternative minimally invasive procedures. We included randomized controlled trials (RCTs) or controlled cohort studies (CCSs) that reported effectiveness (pain, physical function, QOL, opioid use) or safety outcomes (adverse events [AEs], revision surgeries) and uncontrolled studies that reported safety outcomes. DATA ABSTRACTION AND SYNTHESIS: Data were abstracted into structured forms; two independent reviewers assessed risk of bias using standard instruments; certainty of evidence was rated using GRADE. RESULTS: Forty studies (2 RCTs, 3 CCSs, and 35 uncontrolled studies) were included. Minimally invasive SI joint fusion with the iFuse Implant System appeared to result in larger improvements in pain (two RCTs: mean difference in visual analog scale -40.5 mm, 95% CI, -50.1 to -30.9; -38.1 mm, p<.0001) and larger improvements in physical function (mean difference in Oswestry Disability Index -25.4 points, 95% CI, -32.5 to -18.3; -19.8 points, p<.0001) compared to conservative management at 6 months. Improvements in pain and physical function for the RCTs appeared durable at 1- and 2-year follow-up. Findings were similar in one CCS. The two RCTs also found significant improvements in QOL at 6 months and 1 year. Opioid use may be improved at 6 months and 1 to 2 years. AEs appeared higher in the fusion group at 6 months. The incidence of revision surgery varied by study; the highest was 3.8% at 2 years. Two CCSs compared the effectiveness of alternative minimally invasive fusion procedures. One CCS compared iFuse to the Rialto SI Fusion System and reported no differences in pain, function, QOL, and revision surgeries from 6 months to 1 year. One CCS compared iFuse to percutaneous screw fixation and reported significantly fewer revisions among iFuse participants (mean difference -61.0%, 95% CI, -78.4% to -43.5%). The 35 uncontrolled studies had serious limitations and reported heterogeneous safety outcomes. Two of the larger studies reported a 13.2% incidence of complications from minimally invasive SI joint fusion at 90 days using an insurance claims database and a 3.1% incidence of revision surgery over 2.5 years using a postmarket surveillance database. CONCLUSIONS: Among patients meeting diagnostic criteria for SI joint pain and who have not responded to conservative care, minimally invasive SI joint fusion is probably more effective than conservative management for reducing pain and opioid use and improving physical function and QOL. Fusion with iFuse and Rialto appear to have similar effectiveness. AEs appear to be higher for minimally invasive SI joint fusion than conservative management through 6 months. Based on evidence from uncontrolled studies, serious AEs from minimally invasive SI joint fusion may be higher in usual practice compared to what is reported in trials. The incidence of revision surgery is likely no higher than 3.8% at 2 years. Limited evidence is available that compares different minimally invasive devices.

Screening for Atrial Fibrillation: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Atrial fibrillation (AF), the most common arrhythmia, increases the risk of stroke. Objective: To review the evidence on screening for AF in adults without prior stroke to inform the US Preventive Services Task Force. Data Sources: PubMed, Cochrane Library, and trial registries through October 5, 2020; references, experts, and literature surveillance through October 31, 2021. Study Selection: Randomized clinical trials (RCTs) of screening among asymptomatic persons without known AF or prior stroke; test accuracy studies; RCTs of anticoagulation among persons with AF; systematic reviews; and observational studies reporting harms. Data Extraction and Synthesis: Two reviewers assessed titles/abstracts, full-text articles, and study quality and extracted data; when at least 3 similar studies were available, meta-analyses were conducted. Main Outcomes and Measures: Detection of undiagnosed AF, test accuracy, mortality, stroke, stroke-related morbidity, and harms. Results: Twenty-six studies (N = 113 784) were included. In 1 RCT (n = 28 768) of twice-daily electrocardiography (ECG) screening for 2 weeks, the likelihood of a composite end point (ischemic stroke, hemorrhagic stroke, systemic embolism, all-cause mortality, and hospitalization for bleeding) was lower in the screened group over 6.9 years (hazard ratio, 0.96 [95% CI, 0.92-1.00]; P = .045), but that study had numerous limitations. In 4 RCTs (n = 32 491), significantly more AF was detected with intermittent and continuous ECG screening compared with no screening (risk difference range, 1.0%-4.8%). Treatment with warfarin over a mean of 1.5 years in populations with clinical, mostly persistent AF was associated with fewer ischemic strokes (pooled risk ratio [RR], 0.32 [95% CI, 0.20-0.51]; 5 RCTs; n = 2415) and lower all-cause mortality (pooled RR, 0.68 [95% CI, 0.50-0.93]) compared with placebo. Treatment with direct oral anticoagulants was also associated with lower incidence of stroke (adjusted odds ratios range, 0.32-0.44) in indirect comparisons with placebo. The pooled RR for major bleeding for warfarin compared with placebo was 1.8 (95% CI, 0.85-3.7; 5 RCTs; n = 2415), and the adjusted odds ratio for major bleeding for direct oral anticoagulants compared with placebo or no treatment ranged from 1.38 to 2.21, but CIs did not exclude a null effect. Conclusions and Relevance: Although screening can detect more cases of unknown AF, evidence regarding effects on health outcomes is limited. Anticoagulation was associated with lower risk of first stroke and mortality but with increased risk of major bleeding, although estimates for this harm are imprecise; no trials assessed benefits and harms of anticoagulation among screen-detected populations.

Screening for Vitamin D Deficiency in Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Low serum vitamin D levels have been associated with adverse clinical outcomes; identifying and treating deficiency may improve outcomes. Objective: To review the evidence about screening for vitamin D deficiency in adults. Data Sources: PubMed, EMBASE, the Cochrane Library, and trial registries through March 12, 2020; bibliographies from retrieved articles, outside experts, and surveillance of the literature through November 30, 2020. Study Selection: Fair- or good-quality, English-language randomized clinical trials (RCTs) of screening with serum 25-hydroxyvitamin D (25[OH]D) compared with no screening, or treatment with vitamin D (with or without calcium) compared with placebo or no treatment conducted in nonpregnant adults; nonrandomized controlled intervention studies for harms only. Treatment was limited to studies enrolling or analyzing participants with low serum vitamin D levels. Data Extraction and Synthesis: Two reviewers assessed titles/abstracts and full-text articles, extracted data, and assessed study quality; when at least 3 similar studies were available, meta-analyses were conducted. Main Outcomes and Measures: Mortality, incident fractures, falls, diabetes, cardiovascular events, cancer, depression, physical functioning, and infection. Results: Forty-six studies (N = 16 205) (77 publications) were included. No studies directly evaluated the health benefits or harms of screening. Among community-dwelling populations, treatment was not significantly associated with mortality (pooled absolute risk difference [ARD], 0.3% [95% CI, -0.6% to 1.1%]; 8 RCTs, n = 2006), any fractures (pooled ARD, -0.3% [95% CI, -2.1% to 1.6%]; 6 RCTs, n = 2186), incidence of diabetes (pooled ARD, 0.1% [95% CI, -1.3% to 1.6%]; 5 RCTs, n = 3356), incidence of cardiovascular disease (2 RCTs; hazard ratio, 1.00 [95% CI, 0.74 to 1.35] and 1.09 [95% CI, 0.68 to 1.76]), incidence of cancer (2 RCTs; hazard ratio, 0.97 [95% CI, 0.68 to 1.39] and 1.01 [95% CI, 0.65 to 1.58], or depression (3 RCTs, various measures reported). The pooled ARD for incidence of participants with 1 or more falls was -4.3% (95% CI, -11.6% to 2.9%; 6 RCTs). The evidence was mixed for the effect of treatment on physical functioning (2 RCTs) and limited for the effect on infection (1 RCT). The incidence of adverse events and kidney stones was similar between treatment and control groups. Conclusions and Relevance: No studies evaluated the direct benefits or harms of screening for vitamin D deficiency. Among asymptomatic, community-dwelling populations with low vitamin D levels, the evidence suggests that treatment with vitamin D has no effect on mortality or the incidence of fractures, falls, depression, diabetes, cardiovascular disease, cancer, or adverse events. The evidence is inconclusive about the effect of treatment on physical functioning and infection.

Making Narrative Statements to Describe Treatment Effects.

Accurately describing treatment effects using plain language and narrative statements is a critical step in communicating research findings to end users. However, the process of developing these narratives has not been historically guided by a specific framework. The Agency for Healthcare Research and Quality Evidence-based Practice Center Program developed guidance for narrative summaries of treatment effects that identifies five constructs. We explicitly identify these constructs to facilitate developing narrative statements: (1) direction of effect, (2) size of effect, (3) clinical importance, (4) statistical significance, and (5) strength or certainty of evidence. These constructs clearly overlap. It may not always be feasible to address all five constructs. Based on context and intended audience, investigators can determine which constructs will be most important to address in narrative statements.

Screening for Bacterial Vaginosis in Pregnant Adolescents and Women to Prevent Preterm Delivery: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.

Importance: Preterm delivery results in adverse outcomes; identifying and treating bacterial vaginosis may reduce its occurrence. Objective: To update the evidence on screening and treatment of asymptomatic bacterial vaginosis in pregnancy for the US Preventive Services Task Force. Data Sources: MEDLINE, Cochrane Library, and trial registries through May 29, 2019; bibliographies from retrieved articles, experts, and surveillance of the literature through December 31, 2019. Study Selection: Fair- or good-quality English-language studies evaluating diagnostic accuracy of tests feasible within primary care; randomized clinical trials (RCTs); nonrandomized controlled intervention studies (for harms only); or meta-analyses of metronidazole or clindamycin. Data Extraction and Synthesis: Two reviewers independently assessed titles/abstracts and full-text articles, extracted data, and assessed study quality; when at least 3 similar studies were available, meta-analyses were conducted. Main Outcomes and Measures: Sensitivity, specificity, preterm delivery, maternal adverse effects, congenital birth defects, childhood cancer. Results: Forty-four studies (48 publications) were included. No studies evaluated the benefits or harms of screening. Twenty-five studies (n = 15 785) evaluated the accuracy of screening tests; across individual studies and tests, sensitivity ranged from 0.36 to 1.0 and specificity ranged from 0.49 to 1.0. Among trials reporting findings from general obstetric populations (n = 7953), no significant association was observed between treatment and spontaneous delivery before 37 weeks (pooled absolute risk difference [ARD], -1.44% [95% CI, -3.31% to 0.43%]; 8 RCTs, n = 7571) or any delivery before 37 weeks (pooled ARD, 0.20% [95% CI, -1.13% to 1.53%]; 6 RCTs, n = 6307). Among 5 trials reporting findings among women with a prior preterm delivery, findings were inconsistent; 3 showed a significant beneficial effect, while 2 did not. Maternal adverse events from treatment were infrequent and minor (eg, candidiasis) but were slightly more common with active treatment compared with placebo across 8 RCTs. Two meta-analyses of observational studies reported no significant association between metronidazole exposure and congenital malformations (odds ratio, 0.96 [95% CI, 0.75 to 1.22]; odds ratio, 1.08 [95% CI, 0.90 to 1.29]). One cohort study reported no significantly increased incidence of childhood cancer among metronidazole-exposed children (adjusted relative risk, 0.81 [95% CI, 0.41 to 1.59]). However, studies of in utero exposure had important limitations. Conclusions and Relevance: Accuracy of screening tests for bacterial vaginosis varies. The evidence suggests no difference in the incidence of preterm delivery and related outcomes from treatment for asymptomatic bacterial vaginosis in a general obstetric population but was inconclusive for women with a prior preterm delivery. Maternal adverse events from treatment appear to be infrequent and minor, but the evidence about harms from in utero exposure was inconclusive.

Universal screening for SARS-CoV-2 infection: a rapid review.

BACKGROUND: Coronavirus disease 2019 (COVID-19) is caused by the novel betacoronavirus, severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). Most people infected with SARS-CoV-2 have mild disease with unspecific symptoms, but about 5% become critically ill with respiratory failure, septic shock and multiple organ failure. An unknown proportion of infected individuals never experience COVID-19 symptoms although they are infectious, that is, they remain asymptomatic. Those who develop the disease, go through a presymptomatic period during which they are infectious. Universal screening for SARS-CoV-2 infections to detect individuals who are infected before they present clinically, could therefore be an important measure to contain the spread of the disease. OBJECTIVES: We conducted a rapid review to assess (1) the effectiveness of universal screening for SARS-CoV-2 infection compared with no screening and (2) the accuracy of universal screening in people who have not presented to clinical care for symptoms of COVID-19. SEARCH METHODS: An information specialist searched Ovid MEDLINE and the Centers for Disease Control (CDC) COVID-19 Research Articles Downloadable Database up to 26 May 2020. We searched Embase.com, the CENTRAL, and the Cochrane Covid-19 Study Register on 14 April 2020. We searched LitCovid to 4 April 2020. The World Health Organization (WHO) provided records from daily searches in Chinese databases and in PubMed up to 15 April 2020. We also searched three model repositories (Covid-Analytics, Models of Infectious Disease Agent Study [MIDAS], and Society for Medical Decision Making) on 8 April 2020. SELECTION CRITERIA: Trials, observational studies, or mathematical modelling studies assessing screening effectiveness or screening accuracy among general populations in which the prevalence of SARS-CoV2 is unknown. DATA COLLECTION AND ANALYSIS: After pilot testing review forms, one review author screened titles and abstracts. Two review authors independently screened the full text of studies and resolved any disagreements by discussion with a third review author. Abstracts excluded by a first review author were dually reviewed by a second review author prior to exclusion. One review author independently extracted data, which was checked by a second review author for completeness and accuracy. Two review authors independently rated the quality of included studies using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool for diagnostic accuracy studies and a modified form designed originally for economic evaluations for modelling studies. We resolved differences by consensus. We synthesized the evidence in narrative and tabular formats. We rated the certainty of evidence for days to outbreak, transmission, cases missed and detected, diagnostic accuracy (i.e. true positives, false positives, true negatives, false negatives) using the GRADE approach. MAIN RESULTS: We included 22 publications. Two modelling studies reported on effectiveness of universal screening. Twenty studies (17 cohort studies and 3 modelling studies) reported on screening test accuracy. Effectiveness of screening We included two modelling studies. One study suggests that symptom screening at travel hubs, such as airports, may slightly slow but not stop the importation of infected cases (assuming 10 or 100 infected travellers per week reduced the delay in a local outbreak to 8 days or 1 day, respectively). We assessed risk of bias as minor or no concerns, and certainty of evidence was low, downgraded for very serious indirectness. The second modelling study provides very low-certainty evidence that screening of healthcare workers in emergency departments using laboratory tests may reduce transmission to patients and other healthcare workers (assuming a transmission constant of 1.2 new infections per 10,000 people, weekly screening reduced infections by 5.1% within 30 days). The certainty of evidence was very low, downgraded for high risk of bias (major concerns) and indirectness. No modelling studies reported on harms of screening. Screening test accuracy All 17 cohort studies compared an index screening strategy to a reference reverse transcriptase polymerase chain reaction (RT-PCR) test. All but one study reported on the accuracy of single point-in-time screening and varied widely in prevalence of SARS-CoV-2, settings, and methods of measurement. We assessed the overall risk of bias as unclear in 16 out of 17 studies, mainly due to limited information on the index test and reference standard. We rated one study as being at high risk of bias due to the inclusion of two separate populations with likely different prevalences. For several screening strategies, the estimates of sensitivity came from small samples. For single point-in-time strategies, for symptom assessment, the sensitivity from 12 cohorts (524 people) ranged from 0.00 to 0.60 (very low-certainty evidence) and the specificity from 12 cohorts (16,165 people) ranged from 0.66 to 1.00 (low-certainty evidence). For screening using direct temperature measurement (3 cohorts, 822 people), international travel history (2 cohorts, 13,080 people), or exposure to known infected people (3 cohorts, 13,205 people) or suspected infected people (2 cohorts, 954 people), sensitivity ranged from 0.00 to 0.23 (very low- to low-certainty evidence) and specificity ranged from 0.90 to 1.00 (low- to moderate-certainty evidence). For symptom assessment plus direct temperature measurement (2 cohorts, 779 people), sensitivity ranged from 0.12 to 0.69 (very low-certainty evidence) and specificity from 0.90 to 1.00 (low-certainty evidence). For rapid PCR test (1 cohort, 21 people), sensitivity was 0.80 (95% confidence interval (CI) 0.44 to 0.96; very low-certainty evidence) and specificity was 0.73 (95% CI 0.39 to 0.94; very low-certainty evidence). One cohort (76 people) reported on repeated screening with symptom assessment and demonstrates a sensitivity of 0.44 (95% CI 0.29 to 0.59; very low-certainty evidence) and specificity of 0.62 (95% CI 0.42 to 0.79; low-certainty evidence). Three modelling studies evaluated the accuracy of screening at airports. The main outcomes measured were cases missed or detected by entry or exit screening, or both, at airports. One study suggests very low sensitivity at 0.30 (95% CI 0.1 to 0.53), missing 70% of infected travellers. Another study described an unrealistic scenario to achieve a 90% detection rate, requiring 0% asymptomatic infections. The final study provides very uncertain evidence due to low methodological quality. AUTHORS' CONCLUSIONS: The evidence base for the effectiveness of screening comes from two mathematical modelling studies and is limited by their assumptions. Low-certainty evidence suggests that screening at travel hubs may slightly slow the importation of infected cases. This review highlights the uncertainty and variation in accuracy of screening strategies. A high proportion of infected individuals may be missed and go on to infect others, and some healthy individuals may be falsely identified as positive, requiring confirmatory testing and potentially leading to the unnecessary isolation of these individuals. Further studies need to evaluate the utility of rapid laboratory tests, combined screening, and repeated screening. More research is also needed on reference standards with greater accuracy than RT-PCR. Given the poor sensitivity of existing approaches, our findings point to the need for greater emphasis on other ways that may prevent transmission such as face coverings, physical distancing, quarantine, and adequate personal protective equipment for frontline workers.

Direct-to-Consumer Advertising of Prescription Drugs and the Patient-Prescriber Encounter: A Systematic Review.

We systematically reviewed the research on patients' and prescribers' perceptions of, and self-reported behaviors prompted by, exposure to direct-to-consumer advertising (DTCA) (For ease of reading we use the term "advertising" to encompass advertising and promotional labeling. Broad use of this term does not imply endorsement by FDA) of prescription drugs that occurs in the context of a clinical encounter. This research offers an important perspective on the broader goal of incorporating patient and prescriber voices in decision-making. Outcomes included patient information seeking, medication adherence, patient requests for DTCA-promoted prescription drugs, prescribing behaviors, and perceptions of the patient-prescriber relationship and interactions. We searched PubMed and other databases from 1982-2017 and identified 38 studies meeting our study criteria. Of these, 24 studies used patient-reported outcomes and 18 used prescriber-reported outcomes (four used both). Studies suggested some potential benefits of exposure to DTCA, including patients' enhanced information-seeking, increased patient requests for appropriate prescriptions (when addressing potential underuse) and patients' perceptions of higher-quality interactions with prescribers. Most prescribers perceived a neutral influence on the quality of their clinical interactions with patients regarding DTCA. Harms included patients receiving prescriptions for drugs that were not appropriate for them or that the patients did not need, and the potential for DTCA to interfere with medication adherence in some populations, such as those with mental illness. The potential benefits of DTCA on the patient-provider encounter must be balanced with the potential for harms.

Surgical Management of Lumbar Radiculopathy: a Systematic Review.

BACKGROUND: Lumbar radiculopathy is characterized by radiating pain with or without motor weakness or sensory disturbances; the point prevalence ranges from 1.6 to 13.4%. The objective of this review was to determine the efficacy, safety, and cost of surgical versus nonsurgical management of symptomatic lumbar radiculopathy in adults. METHODS: We searched PubMed from January 1, 2007, to April 10, 2019 with hand searches of systematic reviews for studies prior to 2007. One reviewer extracted data and a second checked for accuracy. Two reviewers completed independent risk of bias and strength of evidence ratings. RESULTS: We included seven RCTs (N = 1158) and three cost-effectiveness analysis. Surgery reduced leg pain by 6 to 26 points more than nonsurgical interventions as measured on a 0- to 100-point visual analog scale of pain at up to 26 weeks follow-up; differences between groups did not persist at 1 year or later. The evidence was somewhat mixed for function and disability in follow-up through 26 weeks (standardized mean difference [SMD] - 0.16 (95% CI, - 0.30 to - 0.03); minimal differences were observed at 2 years (SMD - 0.06 (95% CI, - 0.20 to 0.07). There were similar improvements in quality of life, neurologic symptoms, and return to work. No surgical deaths occurred and surgical morbidity was infrequent. The incidence of reoperations ranged from 0 to 10%. The average cost per quality-adjusted life year gained from a healthcare payor perspective ranged from $51,156 to $83,322 for surgery compared to nonsurgical interventions. DISCUSSION: Most findings are based on a body of RCT evidence graded as low to very low certainty. Compared with nonsurgical interventions, surgery probably reduces pain and improves function in the short- and medium-term, but this difference does not persist in the long-term. Although surgery appears to be safe, it may or may not be cost-effective depending on a decision maker's willingness to pay threshold.