Search

The NIH Pragmatic Trials Collaboratory supports the design and conduct of 32 embedded pragmatic clinical trials, and many of these trials rely on data from the electronic health record (EHR) to monitor outcomes and/or use functionality provided by the EHR platform to deliver the intervention. Given the complexity and dynamic nature of EHR systems, study teams have encountered challenges in use of the EHR for these purposes, including challenges related to local implementation of trial interventions, rapid technology evolution, EHR updates, and transitions in EHR systems. In this article, we share case examples and lessons learned, and suggest that teams need to be aware of-and perhaps proactively investigate- possible changes to EHR systems and data that will affect the delivery of interventions and the integrity and safety of pragmatic clinical trials., Competing Interests: Declaration of competing interest KM has no disclosures relevant to the manuscript and reports that he is a Co-inventor of Hive Networks, Inc. and his salary is funded in part through research contracts from Bayer, BMS, BI, Amgen, Novartis, Seqirus, and Genentech. JGJ: Dr. Jarvik was supported by the University of Washington Clinical Learning, Evidence, And Research (CLEAR) Center for Musculoskeletal Disorders, Administrative, Methodologic Cores and NIAMS/NIH grant P30AR072572. Dr. Jarvik r has no disclosures relevant to the manuscript but reports royalties as a book co-editor from Springer Publishing and travel reimbursement for Faculty Board of Review from GE-Association of University Radiologists Radiology Research Academic Fellowship (GERRAF) and royalties as a chapter author from Wolters Kluwer/UpToDate., (Copyright © 2024 The Authors. Published by Elsevier Inc. All rights reserved.)

Purpose: To examine associations between lumbar intervertebral disc degeneration (LDD) and type II Modic changes (MC) when retaining information at each interspace ("interspace-level analysis"), as compared to aggregating information across interspaces as is typically done in spine research ("person-level analysis"). The study compared results from (1) interspace-level analyses assuming a common relationship across interspaces (the "interspace-level, common-relationship" approach), (2) interspace-level analyses allowing for interspace-specific associations (an "interspace-level, interspace-specific" approach), and (3) a conventional person-level analytic approach., Methods: Adults in primary care (n=147) received lumbar spine magnetic resonance imaging (MRI) and neuroradiologist-evaluated assessments of prevalent disc height narrowing (DHN), type II MC, and other LDD parameters. Analyses examined associations between DHN and type II MC in interspace-level, commonrelationship analyses, interspace-level, interspace-specific analyses, and conventional person-level analyses., Results: Cross-sectional, interspace-level, common-relationship analyses found large-magnitude DHN-type II MC associations (adjusted OR [aOR]=6.5, 95% confidence intervals (CIs) 3.3-13; p<0.001). The magnitude of this association was larger and more precise than that yielded by person-level analyses (aOR=2.9 [95% CI 1.2-7]), and substantially more precise than interspace-level, interspace-specific analyses which allowed the association between DHN and type II MC to vary across levels. Across exploratory analyses of disc signal intensity and other MC types, interspace-level, common-relationship analyses produced larger-magnitude and more precise associations than person-level analyses in most situations, and were more precise than interspace-level, interspace-specific analyses., Conclusions: Interspace-level analytic approaches offer some advantages to person-level analyses that may be useful in understanding relationships between spinal degeneration findings.

Importance: The Oral Rheumatoid Arthritis Trial Surveillance demonstrated an increased cancer risk among patients with rheumatoid arthritis (RA) taking tofacitinib compared with those taking tumor necrosis factor inhibitors (TNFis). Although international cohort studies have compared cancer outcomes between TNFis, non-TNFi drugs, and Janus kinase inhibitor (JAKis), their generalizability to US patients with RA is limited., Objective: To assess the comparative safety of TNFis, non-TNFi drugs, and JAKis among US patients with RA (ie, the cancer risk associated with the use of these drugs among these patients)., Design, Setting, and Participants: This retrospective cohort study used US administrative claims data from Merative Marketscan Research Databases from November 1, 2012, to December 31, 2021. Follow-up occurred up to 2 years after initiation of biologic or targeted synthetic disease-modifying antirheumatic drugs (DMARDs). Participants included individuals aged 18 to 64 years with RA, identified using at least 2 RA International Classification of Diseases, Ninth Revision or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision diagnostic codes on or before the date of TNFi, non-TNFi, or JAKi initiation ("index date"). Statistical analysis took place from June 2022 to September 2024., Exposures: New initiations of TNFis, abatacept, interleukin 6 inhibitors (IL-6is), rituximab, or JAKis. Individuals could contribute person-time to more than 1 treatment exposure if treatment escalation mimicked typical clinical practice but were censored if they switched to a previously trialed medication class., Main Outcomes and Measures: Incident cancer, excluding nonmelanoma skin cancer, after at least 90 days and within 2 years of initiation of biologic or targeted synthetic DMARDs. Outcomes were associated with the most recent drug exposure., Results: Of the 25 305 individuals who initiated treatment and who met the inclusion criteria, most were female (19 869 [79%]), had a median age of 50 years (IQR, 42-56 years), and were from the South US (12 516 [49%]). Of a total 27 661 drug exposures, drug initiations consisted of 20 586 TNFi exposures (74%), 2570 JAKi exposures (9%), 2255 abatacept exposures (8%), 1182 rituximab exposures (4%), and 1068 IL-6i exposures (4%). Multivariable Cox proportional hazards regression analysis showed that rituximab was associated with a higher risk of incident cancer compared with TNFis (hazard ratio [HR], 1.91; 95% CI, 1.17-3.14), followed by abatacept (HR, 1.47; 95% CI, 1.03-2.11), and JAKis (HR, 1.36; 95% CI, 0.94-1.96)., Conclusions and Relevance: In this cohort study of individuals with RA and new biologic or targeted synthetic DMARD exposures, individuals initiating rituximab, abatacept, and JAKis demonstrated higher incidence rates and statistically significantly increased risks of incident cancers compared with those initiating TNFis in the first 2 years after initiation of biologic or targeted synthetic DMARDs. Given the limitations of administrative claims data and confounding by indication, it is likely that these patients may have a higher disease burden, resulting in channeling bias. To better understand these associations, larger studies with longer follow-up time are needed.

Abstract: Given the negative impact of opioid use on population health, prescriptions for alternative pain-relieving medications, including gabapentin, have increased. We wanted to determine whether people who filled gabapentin and opioid prescriptions concurrently ("gabapentin + opioids") had greater mortality than those who filled an active control medication (tricyclic antidepressants [TCAs] or duloxetine) and opioids concurrently ("TCAs/duloxetine + opioids"). In this population-based, propensity score-matched cohort study, we identified Medicare beneficiaries with spine-related diagnoses from 2017 to 2019. We compared people treated with gabapentin + opioids (n = 67,133) to people treated with TCAs/duloxetine + opioids (n = 67,133) who were matched on demographic and clinical factors. The primary outcome was mortality at any time, and a secondary outcome was occurrence of a major medical complication at any time. Among 134,266 participants (median age 73.4 years; 66.7% female), 2360 died before the end of follow-up. No difference in mortality was observed between groups (adjusted hazard ratio and 95% confidence interval for gabapentin + opioids 0.98 [0.90-1.06]; P = 0.63). However, people treated with gabapentin + opioids were at slightly increased risk of a major medical complication (1.02 [1.00-1.04]; P = 0.03) compared to those treated with TCAs/duloxetine + opioids. Results were similar in analyses (1) restricted to ≤30-day follow-up and (2) that required ≥2 fills of each prescription. When treating pain in older adults taking opioids, the addition of gabapentin did not increase mortality risk relative to addition of TCAs or duloxetine., (Copyright © 2024 International Association for the Study of Pain.)

Background and Purpose: Vertebral compression fractures may indicate osteoporosis but are underdiagnosed and underreported by radiologists. We have developed an ensemble of vertebral body (VB) segmentation models for lateral radiographs as a critical component of an automated, opportunistic screening tool. Our goal is to detect the approximate location of thoracic and lumbar VBs, including fractured vertebra, on lateral radiographs., Materials and Methods: The Osteoporotic Fractures in Men Study (MrOS) data set includes spine radiographs of 5994 men aged ≥65 years from 6 clinical centers. Two segmentation models, U-Net and Mask-RCNN (Region-based Convolutional Neural Network), were independently trained on the MrOS data set retrospectively, and an ensemble was created by combining them. Primary performance metrics for VB detection success included precision, recall, and F1 score for object detection on a held-out test set. Intersection over union (IoU) and Dice coefficient were also calculated as secondary metrics of performance for the test set. A separate external data set from a quaternary health care enterprise was acquired to test generalizability, comprising diagnostic clinical radiographs from men and women aged ≥65 years., Results: The trained models achieved F1 score of U-Net = 83.42%, Mask-RCNN = 86.30%, and ensemble = 88.34% in detecting all VBs, and F1 score of U-Net = 87.88%, Mask-RCNN = 92.31%, and ensemble = 97.14% in detecting severely fractured vertebrae. The trained models achieved an average IoU per VB of 0.759 for U-Net and 0.709 for Mask-RCNN. The trained models achieved F1 score of U-Net = 81.11%, Mask-RCNN = 79.24%, and ensemble = 87.72% in detecting all VBs in the external data set., Conclusions: An ensemble model combining predictions from U-Net and Mask-RCNN resulted in the best performance in detecting VBs on lateral radiographs and generalized well to an external data set. This model could be a key component of a pipeline to detect fractures on all vertebrae in a radiograph in an automated, opportunistic screening tool under development., (© 2024 by American Journal of Neuroradiology.)

Background Context: Recent work indicates no increased mortality risk with concurrent gabapentin and opioid use when using an active comparator control design. However, concurrent gabapentin and opioid prescriptions have been associated with greater risk of respiratory depression in some studies., Purpose: To compare the risk of respiratory events among Medicare enrollees with spine-related diagnoses treated with gabapentin + opioids vs those treated with tricyclic antidepressants (TCA) or duloxetine + opioids. We hypothesized that enrollees treated with gabapentin + opioids would have increased risk of adverse respiratory events compared to those treated with an active control + opioids., Study Design/setting: Propensity score-matched cohort study with an incident user, active comparator (TCA/duloxetine) control design. The primary analysis included those who concurrently (within 30 days) filled ≥1 incident gabapentin + ≥1 opioid or ≥1 incident TCA/duloxetine + ≥1 opioid prescription., Patient Sample: U.S. Medicare beneficiaries with spine-related diagnoses 2017-2019. People treated with gabapentin + opioids (n=66,580) were matched on demographic and clinical factors to people treated with TCAs/duloxetine + opioids (n=66,580)., Outcome Measures: Time to a composite respiratory outcome consisting of mechanical ventilation, intubation, respiratory failure, pneumonia, or acute respiratory distress syndrome., Methods: Cox proportional hazard regression was used to estimate adjusted hazard ratios (aHRs) and 95% confidence intervals (95% CIs)., Results: Among 133,160 Medicare enrollees (median age 73.3 years; 66.7% female), 6089 (4.6%) experienced respiratory events before the end of follow-up. A total of 3297 (5.0%) of people who were treated with gabapentin + opioids (median initial dose/day of gabapentin was 338 mg) had respiratory events compared to 2792 (4.2%) of those treated with an active control + opioids. The increased risk in those treated with gabapentin + opioids was statistically significant after adjustment (HR 1.14; 95% CI 1.09, 1.20; p<0.0001). The most common respiratory events were pneumonia (3.5% of people in the gabapentin + opioids group versus 3.0% of people in the TCA/duloxetine + opioids group) and respiratory failure (2.2% in the gabapentin + opioids group versus 1.8% in the TCA/duloxetine + opioids group). Results were similar in analyses (a) restricted to ≤30-day follow-up and (b) that required ≥2 fills of each prescription., Conclusions: While recent work has indicated no increased mortality risk with concurrent gabapentin and opioid use, the current findings suggest clinicians should exercise caution in prescribing gabapentin to people experiencing pain who are also being treated with opioids, due to the potentially increased risk of respiratory events.

Automated screening for vertebral fractures could improve outcomes. We achieved an AUC-ROC = 0.968 for the prediction of moderate to severe fracture using a GAM with age and three maximal vertebral body scores of fracture from a convolutional neural network. Maximal fracture scores resulted in a performant model for subject-level fracture prediction. Combining individual deep learning vertebral body fracture scores and demographic covariates for subject-level classification of osteoporotic fracture achieved excellent performance (AUC-ROC of 0.968) on a large dataset of radiographs with basic demographic data., Purpose: Osteoporotic vertebral fractures are common and morbid. Automated opportunistic screening for incidental vertebral fractures from radiographs, the highest volume imaging modality, could improve osteoporosis detection and management. We consider how to form patient-level fracture predictions and summarization to guide management, using our previously developed vertebral fracture classifier on segmented radiographs from a prospective cohort study of US men (MrOS). We compare the performance of logistic regression (LR) and generalized additive models (GAM) with combinations of individual vertebral scores and basic demographic covariates., Methods: Subject-level LR and GAM models were created retrospectively using all fracture predictions or summary variables such as order statistics, adjacent vertebral interactions, and demographic covariates (age, race/ethnicity). The classifier outputs for 8663 vertebrae from 1176 thoracic and lumbar radiographs in 669 subjects were divided by subject to perform stratified fivefold cross-validation. Models were assessed using multiple metrics, including receiver operating characteristic (ROC) and precision-recall (PR) curves., Results: The best model (AUC-ROC = 0.968) was a GAM using the top three maximum vertebral fracture scores and age. Using top-ranked scores only, rather than all vertebral scores, improved performance for both model classes. Adding age, but not ethnicity, to the GAMs improved performance slightly., Conclusion: Maximal vertebral fracture scores resulted in the highest-performing models. While combining multiple vertebral body predictions risks decreasing specificity, our results demonstrate that subject-level models maintain good predictive performance. Thresholding strategies can be used to control sensitivity and specificity as clinically appropriate., (© 2024. International Osteoporosis Foundation and Bone Health and Osteoporosis Foundation.)

Purpose: Osteoporotic vertebral compression fractures (OVCFs) are a highly prevalent source of morbidity and mortality, and preventive treatment has been demonstrated to be both effective and cost effective. To take advantage of the information available on existing chest and abdominal radiographs, the authors' study group has developed software to access these radiographs for OVCFs with high sensitivity and specificity using an established artificial intelligence deep learning algorithm. The aim of this analysis was to assess the potential cost-effectiveness of implementing this software., Methods: A deterministic expected-value cost-utility model was created, combining a tree model and a Markov model, to compare the strategies of opportunistic screening for OVCFs against usual care. Total costs and total quality-adjusted life-years were calculated for each strategy. Screening and treatment costs were considered from a limited societal perspective, at 2022 prices., Results: In the base case, assuming a cost of software implantation of $10 per patient screened, the screening strategy dominated the nonscreening strategy: it resulted in lower cost and increased quality-adjusted life-years. The lower cost was due primarily to the decreased costs associated with fracture treatment and decreased probability of requiring long-term care in patients who received preventive treatment. The screening strategy was dominant up to a cost of $46 per patient screened., Conclusions: Artificial intelligence-based opportunistic screening for OVCFs on existing radiographs can be cost effective from a societal perspective., (Published by Elsevier Inc.)

In contrast to traditional randomized controlled trials, embedded pragmatic clinical trials (ePCTs) are conducted within healthcare settings with real-world patient populations. ePCTs are intentionally designed to align with health system priorities leveraging existing healthcare system infrastructure and resources to ease intervention implementation and increase the likelihood that effective interventions translate into routine practice following the trial. The NIH Pragmatic Trials Collaboratory, funded by the National Institutes of Health (NIH), supports the conduct of large-scale ePCT Demonstration Projects that address major public health issues within healthcare systems. The Collaboratory has a unique opportunity to draw on the Demonstration Project experiences to generate lessons learned related to ePCTs and the dissemination and implementation of interventions tested in ePCTs. In this article, we use case studies from six completed Demonstration Projects to summarize the Collaboratory's experience with post-trial interpretation of results, and implications for sustainment (or de-implementation) of tested interventions. We highlight three key lessons learned. First, ineffective interventions (i.e., ePCT is null for the primary outcome) may be sustained if they have other measured benefits (e.g., secondary outcome or subgroup) or even perceived benefits (e.g., staff like the intervention). Second, effective interventions-even those solicited by the health system and/or designed with significant health system partner buy-in-may not be sustained if they require significant resources. Third, alignment with policy incentives is essential for achieving sustainment and scale-up of effective interventions. Our experiences point to several recommendations to aid in considering post-trial sustainment or de-implementation of interventions tested in ePCTs: (1) include secondary outcome measures that are salient to health system partners; (2) collect all appropriate data to allow for post hoc analysis of subgroups; (3) collect experience data from clinicians and staff; (4) engage policy-makers before starting the trial., (© 2024. The Author(s), under exclusive licence to Society of General Internal Medicine.)

Importance: Given the negative impact of opioid use on population health, prescriptions for alternative pain-relieving medications, including gabapentin, have increased. Concurrent gabapentin and opioid prescriptions are commonly reported in retrospective studies of opioid-related overdose deaths., Objective: To determine whether people who filled gabapentin and opioid prescriptions concurrently ('gabapentin + opioids') had greater mortality than those who filled an active control medication (tricyclic antidepressants [TCAs] or duloxetine) and opioids concurrently ('TCAs/duloxetine + opioids'). We hypothesized that people treated with gabapentin + opioids would have higher mortality rates compared to people treated with TCAs/duloxetine + opioids., Design: Propensity score-matched cohort study with an incident user, active control design. The median (maximum) follow-up was 45 (1093) days., Setting: Population-based., Participants: Medicare beneficiaries with spine-related diagnoses 2017-2019. The primary analysis included those who concurrently (within 30 days) filled at least 1 incident gabapentin + at least 1 opioid or at least 1 incident TCA/duloxetine + at least 1 opioid., Exposures: People treated with gabapentin + opioids (n=67,133) were matched on demographic and clinical factors in a 1:1 ratio to people treated with TCAs/duloxetine + opioids (n=67,133)., Main Outcomes and Measures: The primary outcome was mortality at any time. A secondary outcome was occurrence of a major medical complication at any time., Results: Among 134,266 participants (median age 73.4 years; 66.7% female), 2360 died before the end of follow-up. No difference in mortality was observed between groups (adjusted hazard ratio (HR) and 95% confidence interval (CI) for gabapentin + opioids was 0.98 (0.90, 1.06); p=0.63). However, people treated with gabapentin + opioids were at slightly increased risk of a major medical complication (1.02 (1.00, 1.04); p=0.03) compared to those treated with TCAs/duloxetine + opioids. Results were similar in analyses (a) restricted to less than or = 30-day follow-up and (b) that required at least 2 fills of each prescription., Conclusions and Relevance: When treating pain in older adults taking opioids, the addition of gabapentin did not increase mortality risk relative to addition of TCAs or duloxetine. However, providers should be cognizant of a small increased risk of major medical complications among opioid users initiating gabapentin compared to those initiating TCAs or duloxetine.

Rationale and Objectives: Spinal osteoporotic compression fractures (OCFs) can be an early biomarker for osteoporosis but are often subtle, incidental, and underreported. To ensure early diagnosis and treatment of osteoporosis, we aimed to build a deep learning vertebral body classifier for OCFs as a critical component of our future automated opportunistic screening tool., Materials and Methods: We retrospectively assembled a local dataset, including 1790 subjects and 15,050 vertebral bodies (thoracic and lumbar). Each vertebral body was annotated using an adaption of the modified-2 algorithm-based qualitative criteria. The Osteoporotic Fractures in Men (MrOS) Study dataset provided thoracic and lumbar spine radiographs of 5994 men from six clinical centers. Using both datasets, five deep learning algorithms were trained to classify each individual vertebral body of the spine radiographs. Classification performance was compared for these models using multiple metrics, including the area under the receiver operating characteristic curve (AUC-ROC), sensitivity, specificity, and positive predictive value (PPV)., Results: Our best model, built with ensemble averaging, achieved an AUC-ROC of 0.948 and 0.936 on the local dataset's test set and the MrOS dataset's test set, respectively. After setting the cutoff threshold to prioritize PPV, this model achieved a sensitivity of 54.5% and 47.8%, a specificity of 99.7% and 99.6%, and a PPV of 89.8% and 94.8%., Conclusion: Our model achieved an AUC-ROC>0.90 on both datasets. This testing shows some generalizability to real-world clinical datasets and a suitable performance for a future opportunistic osteoporosis screening tool., Competing Interests: Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Nathan Cross, Qifei Dong, Sandra Johnston, Jessica Perry, and David Haynor report financial support was provided by the National Institute of Arthritis and Musculoskeletal and Skin Diseases. Gang Luo reports financial support was provided by the National Institute of Arthritis and Musculoskeletal and Skin Diseases and National Heart Lung and Blood Institute. Jeffrey Jarvik reports financial support was provided by the National Institute of Arthritis and Musculoskeletal and Skin Diseases and reports a relationship with the General Electric-Association of University Radiologists Radiology Research Academic Fellowship (GERRAF) that includes travel reimbursement. Nathan Cross reports financial support was provided by the General Electric-Association of University Radiologists Radiology Research Academic Fellowship (GERRAF). Nathan M. Cross, Jeffrey Jarvik, David R. Haynor, Gang Luo, Sandra Johnston, Qifei Dong, Jonathan Renslo, Brian Chang, and Jessica Perry have patent #63/463823 pending to the University of Washington., (Copyright © 2023 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)

We evaluated whether older adults who received kyphoplasty had reduced risk of mortality compared to those who did not. In unmatched analyses, those receiving kyphoplasty were at reduced risk of death but after matching on age and medical complications, patients who received kyphoplasty were at increased risk of death., Purpose: In previous observational studies, kyphoplasty for treatment of osteoporotic vertebral fractures has been associated with decreased mortality compared to conservative management. The purpose of this research was to determine whether older adults who received kyphoplasty had reduced risk of mortality compared to matched patients who did not., Methods: Retrospective cohort study of US Medicare enrollees with osteoporotic vertebral fractures between 2017-2019 comparing patients who underwent kyphoplasty to those who did not. We identified 2 control groups a priori: 1) non-augmented patients who met inclusion criteria (group 1); 2) propensity-matched patients on demographic and clinical variables (group 2). We then identified additional control groups using matching for medical complications (group 3) and age + comorbidities (group 4). We calculated hazard ratios (HRs) and 95% confidence intervals (95% CIs) associated with mortality., Results: A total of 235,317 patients (mean (± standard deviation) age 81.1 ± 8.3 years; 85.8% female) were analyzed. In the primary analyses, those who received kyphoplasty were at reduced risk of death compared to those who did not: adjusted HR (95% CI) in group 1 = 0.84 (0.82, 0.87); and in group 2 = 0.88 (0.85, 0.91). However, in post hoc analyses, patients who received kyphoplasty were at increased risk of death: adjusted HR (95% CI) in group 3 = 1.32 (1.25, 1.41) and 1.81 (1.58, 2.09) in group 4., Conclusion: An apparent benefit of kyphoplasty on mortality among patients with vertebral fractures was not present after rigorous propensity matching, illustrating the importance of comparing similar individuals when evaluating observational data., (© 2023. International Osteoporosis Foundation and Bone Health and Osteoporosis Foundation.)

Background: Imaging tests are one of the most frequently used diagnostic modalities in healthcare, but the benefits of their direct impacts on clinical decision-making have been countered by concerns that they can be overused. Assessing the relative value of imaging tests has largely focused on measures of test accuracy, which overlooks more comprehensive benefits and risks of imaging tests, particularly their impact on patient-centred outcomes (PCOs). We present the findings of the Patient Reported Outcomes of Diagnostics (PROD) research study in response to a methodological gap in the area of diagnostic test comparative effectiveness research., Methods: Over a 3-year period, the PROD Study engaged with multiple stakeholders to identify existing conceptual models related to PCOs for imaging testing, conducted primary research and evidence synthesis, and developed consensus recommendations to describe and categorise PCOs related to imaging testing., Results: The PROD framework categorises PCOs from imaging studies within four main domains: information or knowledge yielded, physical impact, emotional outcomes and test burden. PCOs interact with each other and influence effects across domains, and can be modified by factors related to the patient, clinical situation, healthcare team and the testing environment., Conclusions: Using PCOs to inform healthcare decision-making will require ways of collating and presenting information on PCOs in ways that can inform patient-provider decision-making, and developing methods to determine the relative importance of outcomes (including test accuracy) to one another., Competing Interests: Competing interests: JGJ reports grants from NIH/NIAMS, during the conduct of the study, in particular he was supported by the UW Clinical Learning, Evidence And Research (CLEAR) Center for Musculoskeletal Disorders, NIH/NIAMS P30AR072572 and the content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. JGJ also resports grants from Springer Publishing: royalties as a book co-editor; GE-Association of University Radiologists Radiology Research Academic Fellowship (GERRAF): travel reimbursement for Faculty Board of Review; Wolters Kluwer/UpToDate: royalties as a chapter author., (© Author(s) (or their employer(s)) 2023. Re-use permitted under CC BY. Published by BMJ.)

Background: Low back pain is a common presentation across different healthcare settings. Clinicians need to confidently be able to screen and identify people presenting with low back pain with a high suspicion of serious or specific pathology (e.g. vertebral fracture). Patients identified with an increased likelihood of having a serious pathology will likely require additional investigations and specific treatment. Guidelines recommend a thorough history and clinical assessment to screen for serious pathology as a cause of low back pain. However, the diagnostic accuracy of recommended red flags (e.g. older age, trauma, corticosteroid use) remains unclear, particularly those used to screen for vertebral fracture., Objectives: To assess the diagnostic accuracy of red flags used to screen for vertebral fracture in people presenting with low back pain. Where possible, we reported results of red flags separately for different types of vertebral fracture (i.e. acute osteoporotic vertebral compression fracture, vertebral traumatic fracture, vertebral stress fracture, unspecified vertebral fracture)., Search Methods: We used standard, extensive Cochrane search methods. The latest search date was 26 July 2022., Selection Criteria: We considered primary diagnostic studies if they compared results of history taking or physical examination (or both) findings (index test) with a reference standard test (e.g. X-ray, magnetic resonance imaging (MRI), computed tomography (CT), single-photon emission computerised tomography (SPECT)) for the identification of vertebral fracture in people presenting with low back pain. We included index tests that were presented individually or as part of a combination of tests., Data Collection and Analysis: Two review authors independently extracted data for diagnostic two-by-two tables from the publications or reconstructed them using information from relevant parameters to calculate sensitivity, specificity, and positive (+LR) and negative (-LR) likelihood ratios with 95% confidence intervals (CIs). We extracted aspects of study design, characteristics of the population, index test, reference standard, and type of vertebral fracture. Meta-analysis was not possible due to heterogeneity of studies and index tests, therefore the analysis was descriptive. We calculated sensitivity, specificity, and LRs for each test and used these as an indication of clinical usefulness. Two review authors independently conducted risk of bias and applicability assessment using the QUADAS-2 tool., Main Results: This review is an update of a previous Cochrane Review of red flags to screen for vertebral fracture in people with low back pain. We included 14 studies in this review, six based in primary care, five in secondary care, and three in tertiary care. Four studies reported on 'osteoporotic vertebral fractures', two studies reported on 'vertebral compression fracture', one study reported on 'osteoporotic and traumatic vertebral fracture', two studies reported on 'vertebral stress fracture', and five studies reported on 'unspecified vertebral fracture'. Risk of bias was only rated as low in one study for the domains reference standard and flow and timing. The domain patient selection had three studies and the domain index test had six studies rated at low risk of bias. Meta-analysis was not possible due to heterogeneity of the data. Results from single studies suggest only a small number of the red flags investigated may be informative. In the primary healthcare setting, results from single studies suggest 'trauma' demonstrated informative +LRs (range: 1.93 to 12.85) for 'unspecified vertebral fracture' and 'osteoporotic vertebral fracture' (+LR: 6.42, 95% CI 2.94 to 14.02). Results from single studies suggest 'older age' demonstrated informative +LRs for studies in primary care for 'unspecified vertebral fracture' (older age greater than 70 years: 11.19, 95% CI 5.33 to 23.51). Results from single studies suggest 'corticosteroid use' may be an informative red flag in primary care for 'unspecified vertebral fracture' (+LR range: 3.97, 95% CI 0.20 to 79.15 to 48.50, 95% CI 11.48 to 204.98) and 'osteoporotic vertebral fracture' (+LR: 2.46, 95% CI 1.13 to 5.34); however, diagnostic values varied and CIs were imprecise. Results from a single study suggest red flags as part of a combination of index tests such as 'older age and female gender' in primary care demonstrated informative +LRs for 'unspecified vertebral fracture' (16.17, 95% CI 4.47 to 58.43). In the secondary healthcare setting, results from a single study suggest 'trauma' demonstrated informative +LRs for 'unspecified vertebral fracture' (+LR: 2.18, 95% CI 1.86 to 2.54) and 'older age' demonstrated informative +LRs for 'osteoporotic vertebral fracture' (older age greater than 75 years: 2.51, 95% CI 1.48 to 4.27). Results from a single study suggest red flags as part of a combination of index tests such as 'older age and trauma' in secondary care demonstrated informative +LRs for 'unspecified vertebral fracture' (+LR: 4.35, 95% CI 2.92 to 6.48). Results from a single study suggest when '4 of 5 tests' were positive in secondary care, they demonstrated informative +LRs for 'osteoporotic vertebral fracture' (+LR: 9.62, 95% CI 5.88 to 15.73). In the tertiary care setting, results from a single study suggest 'presence of contusion/abrasion' was informative for 'vertebral compression fracture' (+LR: 31.09, 95% CI 18.25 to 52.96)., Authors' Conclusions: The available evidence suggests that only a few red flags are potentially useful in guiding clinical decisions to further investigate people suspected to have a vertebral fracture. Most red flags were not useful as screening tools to identify vertebral fracture in people with low back pain. In primary care, 'older age' was informative for 'unspecified vertebral fracture', and 'trauma' and 'corticosteroid use' were both informative for 'unspecified vertebral fracture' and 'osteoporotic vertebral fracture'. In secondary care, 'older age' was informative for 'osteoporotic vertebral fracture' and 'trauma' was informative for 'unspecified vertebral fracture'. In tertiary care, 'presence of contusion/abrasion' was informative for 'vertebral compression fracture'. Combinations of red flags were also informative and may be more useful than individual tests alone. Unfortunately, the challenge to provide clear guidance on which red flags should be used routinely in clinical practice remains. Further research with primary studies is needed to improve and consolidate our current recommendations for screening for vertebral fractures to guide clinical care., (Copyright © 2023 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.)

Study Design: Secondary analysis of a randomized controlled trial., Objective: To assess how baseline treatment with opioids is associated with pain and function in older adults with lumbar spinal stenosis who receive epidural injections., Summary of Background Data: Data were obtained from the Lumbar Epidural Steroid injections for Spinal Stenosis trial, a double-blind, multisite, randomized controlled trial., Methods: Baseline treatment with opioids was assessed from electronic medical record prescription pharmacy data or from health utilization records collected from patients. We calculated adjusted changes in back pain numerical rating scale, leg pain numerical rating scale, and back-related disability (Roland Morris Disability Questionnaire scores) from baseline to three weeks and to six weeks among patients treated and not treated with opioids at baseline using generalized linear regression., Results: Baseline treatment with opioids was not significantly associated with back pain intensity (adjusted difference in means at three weeks of follow-up between patients treated with opioids at baseline versus not [±95% CI, 0.1 (-0.7, 0.7)], leg pain intensity [-0.2 (-0.9, 0.4)], or back-related function [-0.8 (-2.1, 0.4)]. We found similar results at six weeks of follow-up., Conclusions: Among older adults with lumbar spinal stenosis who are receiving epidural injections, those treated with opioids at baseline had similar outcomes to those who were not., Competing Interests: The authors report no conflicts of interest., (Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.)

Abstract: Conventional "1-variable-at-a-time" analyses to identify treatment effect modifiers are often underpowered and prone to false-positive results. This study used a "risk-modeling" approach guided by the Predictive Approaches to Treatment effect Heterogeneity (PATH) Statement framework: (1) developing and validating a multivariable model to estimate predicted future back-related functional limitations as measured by the Roland-Morris Disability Questionnaire (RMDQ) and (2) stratifying patients from a randomized controlled trial (RCT) of lumbar epidural steroid injections (LESIs) for the treatment of lumbar spinal stenosis into subgroups with different individualized treatment effects on RMDQ scores at the 3-week follow-up. Model development and validation were conducted in a cohort (n = 3259) randomly split into training and testing sets in a 4:1 ratio. The model was developed in the testing set using linear regression with least absolute shrinkage and selection regularization and 5-fold cross-validation. The model was then applied in the testing set and subsequently in patients receiving the control treatment in the RCT of LESI. R2 values in the training set, testing set, and RCT were 0.38, 0.32, and 0.34, respectively. There was statistically significant modification ( P = 0.03) of the LESI treatment effect according to predicted risk quartile, with clinically relevant LESI treatment effect point estimates in the 2 quartiles with greatest predicted risk (-3.7 and -3.3 RMDQ points) and no effect in the lowest 2 quartiles. A multivariable risk-modeling approach identified subgroups of patients with lumbar spinal stenosis with a clinically relevant treatment effect of LESI on back-related functional limitations., (Copyright © 2022 International Association for the Study of Pain.)

Background: Conducting an embedded pragmatic clinical trial in the workflow of a healthcare system is a complex endeavor. The complexity of the intervention delivery can have implications for study planning, ability to maintain fidelity to the intervention during the trial, and/or ability to detect meaningful differences in outcomes., Methods: We conducted a literature review, developed a tool, and conducted two rounds of phone calls with NIH Pragmatic Trials Collaboratory Demonstration Project principal investigators to develop the Intervention Delivery Complexity Tool. After refining the tool, we piloted it with Collaboratory demonstration projects and developed an online version of the tool using the R Shiny application (https://duke-som.shinyapps.io/ICT-ePCT/)., Results: The 6-item tool consists of internal and external factors. Internal factors pertain to the intervention itself and include workflow, training, and the number of intervention components. External factors are related to intervention delivery at the system level including differences in healthcare systems, the dependency on setting for implementation, and the number of steps between the intervention and the outcome., Conclusion: The Intervention Delivery Complexity Tool was developed as a standard way to overcome communication challenges of intervention delivery within an embedded pragmatic trial. This version of the tool is most likely to be useful to the trial team and its health system partners during trial planning and conduct. We expect further evolution of the tool as more pragmatic trials are conducted and feedback is received on its performance outside of the NIH Pragmatic Trials Collaboratory., (Copyright © 2023 The Authors. Published by Elsevier Inc. All rights reserved.)

The 0 to 10 numeric rating scale of pain intensity is a standard outcome in randomized controlled trials (RCTs) of pain treatments. For individuals taking analgesics, there may be a disparity between "observed" pain intensity (pain intensity with concurrent analgesic use) and pain intensity without concurrent analgesic use (what the numeric rating scale would be had analgesics not been taken). Using a contemporary causal inference framework, we compare analytic methods that can potentially account for concurrent analgesic use, first in statistical simulations, and second in analyses of real (non-simulated) data from an RCT of lumbar epidural steroid injections. The default analytic method was ignoring analgesic use, which is the most common approach in pain RCTs. Compared to ignoring analgesic use and other analytic methods, simulations showed that a quantitative pain and analgesia composite outcome based on adding 1.5 points to pain intensity for those who were taking an analgesic (the QPAC 1.5 ) optimized power and minimized bias. Analyses of real RCT data supported the results of the simulations, showing greater power with analysis of the QPAC 1.5 as compared to ignoring analgesic use and most other methods examined. We propose alternative methods that should be considered in the analysis of pain RCTs. PERSPECTIVE: This article presents the conceptual framework behind a new quantitative pain and analgesia composite outcome, the QPAC 1.5 , and the results of statistical simulations and analyses of trial data supporting improvements in power and bias using the QPAC 1.5 . Methods of this type should be considered in the analysis of pain RCTs., (Published by Elsevier Inc.)

Background: Low back pain (LBP) is a common condition made up of a variety of anatomic and clinical subtypes. Lumbar disc herniation (LDH) and lumbar spinal stenosis (LSS) are two subtypes highly associated with LBP. Patients with LDH/LSS are often started with non-surgical treatments and if those are not effective then go on to have decompression surgery. However, recommendation of surgery is complicated as the outcome may depend on the patient's health characteristics. We developed a deep learning (DL) model to predict decompression surgery for patients with LDH/LSS., Materials and Method: We used datasets of 8387 and 8620 patients from a prospective study that collected data from four healthcare systems to predict early (within 2 months) and late surgery (within 12 months after a 2 month gap), respectively. We developed a DL model to use patients' demographics, diagnosis and procedure codes, drug names, and diagnostic imaging reports to predict surgery. For each prediction task, we evaluated the model's performance using classical and generalizability evaluation. For classical evaluation, we split the data into training (80%) and testing (20%). For generalizability evaluation, we split the data based on the healthcare system. We used the area under the curve (AUC) to assess performance for each evaluation. We compared results to a benchmark model (i.e. LASSO logistic regression)., Results: For classical performance, the DL model outperformed the benchmark model for early surgery with an AUC of 0.725 compared to 0.597. For late surgery, the DL model outperformed the benchmark model with an AUC of 0.655 compared to 0.635. For generalizability performance, the DL model outperformed the benchmark model for early surgery. For late surgery, the benchmark model outperformed the DL model., Conclusions: For early surgery, the DL model was preferred for classical and generalizability evaluation. However, for late surgery, the benchmark and DL model had comparable performance. Depending on the prediction task, the balance of performance may shift between DL and a conventional ML method. As a result, thorough assessment is needed to quantify the value of DL, a relatively computationally expensive, time-consuming and less interpretable method., (© 2023. The Author(s).)

Currently, there is no reproducible, widely accepted gold standard to classify osteoporotic vertebral body fractures (OVFs). The purpose of this study is to refine a method with clear rules to classify OVFs for machine learning purposes. The method was found to have moderate interobserver agreement that improved with training., Introduction: The current methods to classify osteoporotic vertebral body fractures are considered ambiguous; there is no reproducible, accepted gold standard. The purpose of this study is to refine classification methodology by introducing clear, unambiguous rules and a refined flowchart to allow consistent classification of osteoporotic vertebral body fractures., Methods: We developed a set of rules and refinements that we called m2ABQ to classify vertebrae into five categories. A fracture-enriched database of thoracic and lumbar spine radiographs of patients 65 years of age and older was retrospectively obtained from clinical institutional radiology records using natural language processing. Five raters independently classified each vertebral body using the m2ABQ system. After each annotation round, consensus sessions that included all raters were held to discuss and finalize a consensus annotation for each vertebral body where individual raters' evaluations differed. This process led to further refinement and development of the rules., Results: Each annotation round showed increase in Fleiss kappa both for presence vs absence of fracture 0.62 (0.56-0.68) to 0.70 (0.65-0.75), as well as for the whole m2ABQ scale 0.29 (0.25-0.33) to 0.54 (0.51-0.58)., Conclusion: The m2ABQ system demonstrates moderate interobserver agreement and practical feasibility for classifying osteoporotic vertebral body fractures. Future studies to compare the method to existing studies are warranted, as well as further development of its use in machine learning purposes., (© 2022. International Osteoporosis Foundation and Bone Health and Osteoporosis Foundation.)

Rationale and Objectives: Osteoporosis affects 9% of individuals over 50 in the United States and 200 million women globally. Spinal osteoporotic compression fractures (OCFs), an osteoporosis biomarker, are often incidental and under-reported. Accurate automated opportunistic OCF screening can increase the diagnosis rate and ensure adequate treatment. We aimed to develop a deep learning classifier for OCFs, a critical component of our future automated opportunistic screening tool., Materials and Methods: The dataset from the Osteoporotic Fractures in Men Study comprised 4461 subjects and 15,524 spine radiographs. This dataset was split by subject: 76.5% training, 8.5% validation, and 15% testing. From the radiographs, 100,409 vertebral bodies were extracted, each assigned one of two labels adapted from the Genant semiquantitative system: moderate to severe fracture vs. normal/trace/mild fracture. GoogLeNet, a deep learning model, was trained to classify the vertebral bodies. The classification threshold on the predicted probability of OCF outputted by GoogLeNet was set to prioritize the positive predictive value (PPV) while balancing it with the sensitivity. Vertebral bodies with the top 0.75% predicted probabilities were classified as moderate to severe fracture., Results: Our model yielded a sensitivity of 59.8%, a PPV of 91.2%, and an F 1 score of 0.72. The areas under the receiver operating characteristic curve (AUC-ROC) and the precision-recall curve were 0.99 and 0.82, respectively., Conclusion: Our model classified vertebral bodies with an AUC-ROC of 0.99, providing a critical component for our future automated opportunistic screening tool. This could lead to earlier detection and treatment of OCFs., (Copyright © 2022 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)

Background: The importance of lumbar findings on magnetic resonance imaging (MRI) remains controversial. Changes in lumbar MRI findings over time may provide important insights into the causes of low back pain. However, the reliability and validity of temporal changes are unknown., Objective: To (1) investigate the interrater reliability of subjective radiologist reporting of temporal changes in lumbar spine MRI findings and (2) determine how commonly temporal changes are reported when two scans are conducted 30 minutes apart (considered false positives)., Design: Cross-sectional study., Setting: Radiology clinic., Participants: Forty volunteers (mean age 40; 53% female) with current (n = 31) or previous (n = 9) low back pain underwent initial lumbar MRI on a single 3T scanner. Participants then lay on a bed for 30 minutes before undergoing an identical MRI. In addition, we purposely selected five participants from a previous study with repeat lumbar MRI scans where temporal changes were reported in at least one MRI finding (1-12 weeks after initial scan) and another five participants where no temporal change was reported. The 10 participants were included in analyses for aim 1 only., Interventions: Not applicable., Main Outcome Measures: Two blinded radiologists reported on temporal changes between the baseline and repeat scan for 12 different MRI findings (eg, disk herniation, annular fissure) at five levels., Results: The interrater reliability of subjective reporting of temporal changes was poor for all MRI findings based on Kappa values (≤ 0.24), but agreement was relatively high (≥ 90.8%). This is explained by the low prevalence of temporal changes as demonstrated by high values for Prevalence and Bias Adjusted Kappa (≥ 0.82). "False positive" temporal changes were reported by at least one radiologist for most MRI findings, but the rate was generally low., Conclusions: Caution is required when interpreting temporal changes in lumbar MRI findings owing to low reliability and some false positive reporting., (© 2021 American Academy of Physical Medicine and Rehabilitation.)

Because osteoporosis is under-recognized in patients with vertebral fractures, we evaluated characteristics associated with osteoporosis identification. Most patients with vertebral fractures did not receive evaluation or treatment for osteoporosis. Black, younger, and male participants were particularly unlikely to have had recognized osteoporosis, which could increase their risk of negative outcomes., Introduction: Vertebral fractures may be identified on imaging but fail to prompt evaluation for osteoporosis. Our objective was to evaluate characteristics associated with clinical osteoporosis recognition in patients who had vertebral fractures detected on their thoracolumbar spine imaging reports., Methods: We prospectively identified individuals who received imaging of the lower spine at primary care clinics in 4 large healthcare systems who were eligible for osteoporosis screening and lacked indications of osteoporosis diagnoses or treatments in the prior year. We evaluated characteristics of participants with identified vertebral fractures that were associated with recognition of osteoporosis (diagnosis code in the health record; receipt of bone mineral density scans; and/or prescriptions for anti-osteoporotic medications). We used mixed models to estimate adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs)., Results: A total of 114,005 participants (47% female; mean age 65 (interquartile range: 57-72) years) were evaluated. Of the 8579 (7%) participants with vertebral fractures identified, 3784 (44%) had recognition of osteoporosis within the subsequent year. In adjusted regressions, Black participants (OR (95% CI): 0.74 (0.57, 0.97)), younger participants (age 50-60: 0.48 (0.42, 0.54); age 61-64: 0.70 (0.60, 0.81)), and males (0.39 (0.35, 0.43)) were less likely to have recognized osteoporosis compared to white participants, adults aged 65 + years, or females., Conclusion: Individuals with identified vertebral fractures commonly did not have recognition of osteoporosis within a year, particularly those who were younger, Black, or male. Providers and healthcare systems should consider efforts to improve evaluation of osteoporosis in patients with vertebral fractures., (© 2022. International Osteoporosis Foundation and National Osteoporosis Foundation.)

Background: There is limited research on the long-term effectiveness of epidural steroid injections (ESI) in older adults despite the high prevalence of back and leg pain in this age group. We tested the hypotheses that older adults undergoing ESI, compared to patients not receiving ESI: (1) have worse pain, disability and quality of life ('outcomes') pre-ESI, (2) have improved outcomes after ESI and (3) have improved outcomes due to a specific ESI effect., Methods: We prospectively studied patients ≥65 years old presenting to primary care with new episodes of back pain in three US healthcare systems (BOLD registry). Outcomes were leg and back pain intensity, disability and quality of life, assessed at baseline and 3-, 6-, 12- and 24-month follow-ups. We categorized participants as: (1) ESI within 6 months from the index visit (n = 295); (2) no ESI within 6 months (n = 4809); (3) no ESI within 6 months, propensity-score matched to group 1 (n = 483). We analysed the data using linear regression and Generalized Estimating Equations., Results: Pain intensity, disability and quality of life at baseline were significantly worse at baseline in ESI patients (group 1) than in group 2. The improvement from baseline to 24 months in all outcomes was statistically significant for group 1. However, no statistically significant differences were observed between outcome trajectories for the propensity-score matched groups 1 and 3., Conclusions: Older adults treated with ESI have long-term improvement. However, the improvement is unlikely the result of a specific ESI effect., Significance: In this large, two-year, prospective study in older adults with a new episode of low back pain, back pain, leg pain, disability and quality of life improved after epidural steroid injections; however, propensity-score matching revealed that the improvement was unlikely the result of a specific effect of the injections, indicating that epidural steroids are unlikely to provide long-term benefits in older adults with new episodes of back and leg pain., (© 2022 European Pain Federation - EFIC®.)

Background: Lumbar spinal stenosis (LSS) is a common degenerative condition that contributes to back and back-related leg pain in older adults. Most patients with symptomatic LSS initially receive non-operative care before surgical consultation. However, there is a scarcity of data regarding prognosis for patients seeking non-surgical care. The overall goal of this project is to develop and evaluate a clinically useful model to predict long-term physical function of patients initiating non-surgical care for symptomatic LSS., Methods: This is a protocol for an inception cohort study of adults 50 years and older who are initiating non-surgical care for symptomatic LSS in a secondary care setting. We plan to recruit up to 625 patients at two study sites. We exclude patients with prior lumbar spine surgeries or those who are planning on lumbar spine surgery. We also exclude patients with serious medical conditions that have back pain as a symptom or limit walking. We are using weekly, automated data pulls from the electronic health records to identify potential participants. We then contact patients by email and telephone within 21 days of a new visit to determine eligibility, obtain consent, and enroll participants. We collect data using telephone interviews, web-based surveys, and queries of electronic health records. Participants are followed for 12 months, with surveys completed at baseline, 3, 6, and 12 months. The primary outcome measure is the 8-item PROMIS Physical Function (PF) Short Form. We will identify distinct phenotypes using PROMIS PF scores at baseline and 3, 6, and 12 months using group-based trajectory modeling. We will develop and evaluate the performance of a multivariable prognostic model to predict 12-month physical function using the least absolute shrinkage and selection operator and will compare performance to other machine learning methods. Internal validation will be conducted using k-folds cross-validation., Discussion: This study will be one of the largest cohorts of individuals with symptomatic LSS initiating new episodes of non-surgical care. The successful completion of this project will produce a cross-validated prognostic model for LSS that can be used to tailor treatment approaches for patient care and clinical trials., (© 2022. The Author(s).)

Rationale and Objectives: The use of natural language processing (NLP) in radiology provides an opportunity to assist clinicians with phenotyping patients. However, the performance and generalizability of NLP across healthcare systems is uncertain. We assessed the performance within and generalizability across four healthcare systems of different NLP representational methods, coupled with elastic-net logistic regression to classify lower back pain-related findings from lumbar spine imaging reports., Materials and Methods: We used a dataset of 871 X-ray and magnetic resonance imaging reports sampled from a prospective study across four healthcare systems between October 2013 and September 2016. We annotated each report for 26 findings potentially related to lower back pain. Our framework applied four different NLP methods to convert text into feature sets (representations). For each representation, our framework used an elastic-net logistic regression model for each finding (i.e., 26 binary or "one-vs.-rest" classification models). For performance evaluation, we split data into training (80%, 697/871) and testing (20%, 174/871). In the training set, we used cross validation to identify the optimal hyperparameter value and then retrained on the full training set. We then assessed performance based on area under the curve (AUC) for the test set. We repeated this process 25 times with each repeat using a different random train/test split of the data, so that we could estimate 95% confidence intervals, and assess significant difference in performance between representations. For generalizability evaluation, we trained models on data from three healthcare systems with cross validation and then tested on the fourth. We repeated this process for each system, then calculated mean and standard deviation (SD) of AUC across the systems., Results: For individual representations, n-grams had the best average performance across all 26 findings (AUC: 0.960). For generalizability, document embeddings had the most consistent average performance across systems (SD: 0.010). Out of these 26 findings, we considered eight as potentially clinically important (any stenosis, central stenosis, lateral stenosis, foraminal stenosis, disc extrusion, nerve root displacement compression, endplate edema, and listhesis grade 2) since they have a relatively greater association with a history of lower back pain compared to the remaining 18 classes. We found a similar pattern for these eight in which n-grams and document embeddings had the best average performance (AUC: 0.954) and generalizability (SD: 0.007), respectively., Conclusion: Based on performance assessment, we found that n-grams is the preferred method if classifier development and deployment occur at the same system. However, for deployment at multiple systems outside of the development system, or potentially if physician behavior changes within a system, one should consider document embeddings since embeddings appear to have the most consistent performance across systems., (Copyright © 2021 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)

While the embedded nature of pragmatic clinical trials (PCTs) can improve the efficiency and relevance of research for multiple stakeholders, embedding research into ongoing clinical care can also involve ethical and regulatory challenges. An emergent challenge is the management of pragmatic clinical trial collateral findings (PCT-CFs). While PCT-CFs share some features with incidental or secondary findings that are encountered in conventional clinical trials and clinical care, the PCT context differs in ethically relevant ways that complicate PCT-CF identification and management. We report on the results of a two-year multi-method investigation of PCT-CFs. Overall, five core themes emerged: 1) the liminal nature of PCTs and the implications of this for PCT-CFs; 2) the context-specific nature of PCT-CF management; 3) the centrality of institutions; 4) the importance of prospective planning; and 5) patient expectations. Among the central lessons of this work are that prior ethics guidance from other settings cannot easily be adapted to address PCT-CFs, nor can a single approach readily accommodate all PCT-CFs. Moving forward, stakeholders, including researchers, institutions, ethics oversight bodies, and funders, should anticipate and plan for PCT-CFs in the design, conduct, and analysis of PCTs. Future scholarship is needed to examine experiences with PCT-CFs, and the practical and conceptual issues they raise for the future conduct of PCTs., (Copyright © 2021 Elsevier Inc. All rights reserved.)

Background: Modifying physician behavior to more closely align with guideline-based care can be challenging. Few effective strategies resulting in appropriate spine-related health care have been reported. The Lumbar Imaging With Reporting of Epidemiology (LIRE) intervention did not result in reductions in spine care but did in opioid prescriptions written., Objectives: To estimate organizational resource needs and costs associated with implementing a pragmatic, decision support-type intervention that inserted age- and modality-matched prevalence information for common lumbar spine imaging findings, using site-based resource use data from the LIRE trial., Research Design: Time and cost estimation associated with implementing the LIRE intervention in a health organization., Subjects: Providers and patients assessed in the LIRE trial., Measures: Expected personnel costs required to implement the LIRE intervention., Results: Annual salaries were converted to daily average per person costs, ranging from $400 to $2,200 per day (base case) for personnel (range: $300-$2,600). Estimated total average cost for implementing LIRE was $5,009 (range: $2,651-$12,020), including conducting pilot testing with providers. Costs associated with a small amount of time for a radiologist (6-12 hours) and imaging-ordering providers (1-8 hours each) account for approximately 75% of the estimated total cost., Conclusions: The process of implementing an intervention for lumbar spine imaging reports containing age- and modality-appropriate epidemiological benchmarks for common imaging findings required radiologists, imaging-ordering providers, information technology specialists, and limited testing and monitoring. The LIRE intervention seems to be a relatively low-cost, evidence-based, complementary tool that can be easily integrated into the reporting of spine imaging., (Copyright © 2021. Published by Elsevier Inc.)

Importance: Machine learning could be used to predict the likelihood of diagnosis and severity of illness. Lack of COVID-19 patient data has hindered the data science community in developing models to aid in the response to the pandemic., Objectives: To describe the rapid development and evaluation of clinical algorithms to predict COVID-19 diagnosis and hospitalization using patient data by citizen scientists, provide an unbiased assessment of model performance, and benchmark model performance on subgroups., Design, Setting, and Participants: This diagnostic and prognostic study operated a continuous, crowdsourced challenge using a model-to-data approach to securely enable the use of regularly updated COVID-19 patient data from the University of Washington by participants from May 6 to December 23, 2020. A postchallenge analysis was conducted from December 24, 2020, to April 7, 2021, to assess the generalizability of models on the cumulative data set as well as subgroups stratified by age, sex, race, and time of COVID-19 test. By December 23, 2020, this challenge engaged 482 participants from 90 teams and 7 countries., Main Outcomes and Measures: Machine learning algorithms used patient data and output a score that represented the probability of patients receiving a positive COVID-19 test result or being hospitalized within 21 days after receiving a positive COVID-19 test result. Algorithms were evaluated using area under the receiver operating characteristic curve (AUROC) and area under the precision recall curve (AUPRC) scores. Ensemble models aggregating models from the top challenge teams were developed and evaluated., Results: In the analysis using the cumulative data set, the best performance for COVID-19 diagnosis prediction was an AUROC of 0.776 (95% CI, 0.775-0.777) and an AUPRC of 0.297, and for hospitalization prediction, an AUROC of 0.796 (95% CI, 0.794-0.798) and an AUPRC of 0.188. Analysis on top models submitting to the challenge showed consistently better model performance on the female group than the male group. Among all age groups, the best performance was obtained for the 25- to 49-year age group, and the worst performance was obtained for the group aged 17 years or younger., Conclusions and Relevance: In this diagnostic and prognostic study, models submitted by citizen scientists achieved high performance for the prediction of COVID-19 testing and hospitalization outcomes. Evaluation of challenge models on demographic subgroups and prospective data revealed performance discrepancies, providing insights into the potential bias and limitations in the models.

Background: To describe characteristics of patients, providers, and clinics associated with opioid or non-opioid pain medication prescribing patterns for patients who received lower spine imaging in primary care clinics., Methods: In these secondary analyses of the Lumbar Imaging with Reporting of Epidemiology (LIRE) study, a randomized controlled trial conducted in 4 health systems in the United States, we evaluated characteristics associated with receipt of pain medication prescriptions. The outcomes were receipt of prescriptions for opioid or, separately, non-opioid pain medications within 90 days after imaging. Among patients who received opioid or non-opioid prescriptions, we evaluated receipt of multiple prescriptions in the year following imaging. Mixed models were used to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs)., Results: Compared with whites, patients identified as Asian (OR, 0.53; 95% CI, 0.51-0.56), Native Hawaiian/Pacific Islander (OR, 0.73; 95% CI, 0.64-0.83), multiracial (OR, 0.84; 95% CI, 0.71-0.98) or Black (OR, 0.92; 95% CI, 0.89-0.96) had significantly reduced odds for receiving prescriptions for opioids within 90 days. Patients identified as Native American/Alaska Native had greater odds for receiving prescriptions for non-opioid pain medications within 90 days (OR, 1.12; 95% CI, 1.01-1.24). Receipt of pain prescriptions 120 days before imaging was strongly predictive of subsequent receipt of pain prescriptions across all categories., Conclusions: After adjusting for factors that could affect prescribing, the strongest differences observed in pain-medication prescribing were across racial categories and for patients with previous pain prescriptions. Further research is needed to understand these differences and to optimize prescribing., Competing Interests: Conflict of interest: In the spirit of full disclosure, we have several potential conflicts of interest to report. JAT reports receiving royalty payments from Psychological Assessment Resources, Inc (PAR). DFK holds patients related to spine augmentation and has licensed intellectual property related to spine augmentation from the Mayo Clinic. RAD reports receiving royalties from Wolters Kluwer/UpToDate for authoring topics on low back pain and consulting fees from Kaiser Permanente Washington Health Research Institute, as well as partial salary support from an endowment from Kaiser Permanente. JGJ reports royalties as a book co-editor from Springer Publishing and travel reimbursement for Faculty Board of Review from GE-Association of University Radiologists Radiology Research Academic Fellowship (GERRAF) and royalties as a chapter author from Wolters Kluwer/UpToDate., (© Copyright 2021 by the American Board of Family Medicine.)

Background: Information on the prevalence of common imaging findings among patients without back pain in spine imaging reports might affect pain medication prescribing for patients with back pain. Prior research on inserting this text suggested a small reduction in opioid prescribing., Objective: To evaluate the effect of epidemiologic information in spine imaging reports on non-opioid pain medication prescribing for primary care patients with back pain., Design: Post hoc analysis of the Lumbar Imaging with Reporting of Epidemiology cluster-randomized trial., Participants: A total of 170,680 patients aged ≥ 18 years from four healthcare systems who received thoracolumbar, lumbar, or lumbosacral spine imaging from 2013 to 2016 and had not received a prescription for non-opioid pain medication in the preceding 120 days., Intervention: Text of age- and modality-specific epidemiologic benchmarks indicating the prevalence of common findings in people without back pain inserted into thoracolumbar, lumbar, or lumbosacral spine imaging reports at intervention clinics., Main Measures: Primary outcomes: any non-opioid prescription within 90 days after index imaging, overall, and by sub-class (skeletal muscle relaxants, NSAIDs, gabapentinoids, tricyclic antidepressants, benzodiazepines, duloxetine)., Secondary Outcomes: count of non-opioid prescriptions within 90 days, overall, and by sub-class., Key Results: The intervention was not associated with the likelihood of patients receiving at least one prescription for new non-opioid pain-related medications, overall (adjusted OR, 1.02; 95% CI, 0.97-1.08) or by sub-class. The intervention was not associated with the number of prescriptions for any non-opioid medication (adjusted incidence rate ratio [IRR], 1.02; 95% CI, 0.99-1.04). However, the intervention was associated with more new prescriptions for NSAIDs (IRR, 1.12) and tricyclic antidepressants (IRR, 1.11)., Conclusions: Inserting epidemiologic text in spine imaging reports had no effect on whether new non-opioid pain-related medications were prescribed but was associated with the number of new prescriptions for certain non-opioid sub-classes., Trial Registration: ClinicalTrials.gov identifier: NCT02015455., (© 2021. Society of General Internal Medicine.)

Background Context: Osteoporotic vertebral fractures (OVFs) can lead to severe pain and reduced function and quality-of-life, but the strength of evidence for treatments remains low, particularly in younger populations., Purpose: To determine whether patients with OVFs who received kyphoplasty had different patterns of healthcare utilization compared to propensity-matched patients who did not receive vertebral augmentation., Design: Observational cohort study., Patient Sample: We identified patients with OVFs from 2007 to 2018 in the IBM MarketScan Commercial Claims and Encounters Databases who received kyphoplasty and compared them to propensity-matched controls who did not receive vertebral augmentation (either kyphoplasty or vertebroplasty)., Outcome Measures: Major medical complications within 30 days, fills of opioids from 1-week through 1-month postaugmentation, and spine-related gross covered payments from 3-days postkyphoplasty through 1-year post-OVF., Methods: We used logistic regression to obtain odds ratios (ORs) and 95% confidence intervals (95% CIs) to compare binary outcomes and median analysis to compare continuous outcomes., Results: Of the 15,197 OVF patients identified, 1,969 (13%) received kyphoplasty and 1,928 (98%) of these were propensity matched to nonaugmented controls. We did not observe differences in the odds of major medical complications within 30 days (adjusted OR [95% CI]: 1.0 [0.6, 1.8]) but patients who received kyphoplasty were more likely to have filled opioid medications within 30 days (adjusted OR [95% CI]: 1.3 [1.1, 1.5]) and had greater spine-related gross covered payments (kyphoplasty median [95% CI]: $1,340 [$240, $4,850]; nonaugmented: $7,870 [$7,480, $8,270]; adjusted difference in medians [95% CI]: $260 [$190, $2,050])., Conclusions: In this cohort of patients <65 years, receipt of kyphoplasty was associated with greater likelihood of opioid fills and somewhat greater spine-related gross covered payments, but no difference in major medical complications. In this retrospective study of administrative data, we did not detect advantages of treatment with kyphoplasty compared with nonaugmentation for any of our outcomes., (Copyright © 2021 Elsevier Inc. All rights reserved.)