Chest Radiograph Interpretation with Deep Learning Models: Assessment with Radiologist-adjudicated Reference Standards and Population-adjusted Evaluation.

BackgroundDeep learning has the potential to augment the use of chest radiography in clinical radiology, but challenges include poor generalizability, spectrum bias, and difficulty comparing across studies.PurposeTo develop and evaluate deep learning models for chest radiograph interpretation by using radiologist-adjudicated reference standards.Materials and MethodsDeep learning models were developed to detect four findings (pneumothorax, opacity, nodule or mass, and fracture) on frontal chest radiographs. This retrospective study used two data sets. Data set 1 (DS1) consisted of 759 611 images from a multicity hospital network and ChestX-ray14 is a publicly available data set with 112 120 images. Natural language processing and expert review of a subset of images provided labels for 657 954 training images. Test sets consisted of 1818 and 1962 images from DS1 and ChestX-ray14, respectively. Reference standards were defined by radiologist-adjudicated image review. Performance was evaluated by area under the receiver operating characteristic curve analysis, sensitivity, specificity, and positive predictive value. Four radiologists reviewed test set images for performance comparison. Inverse probability weighting was applied to DS1 to account for positive radiograph enrichment and estimate population-level performance.ResultsIn DS1, population-adjusted areas under the receiver operating characteristic curve for pneumothorax, nodule or mass, airspace opacity, and fracture were, respectively, 0.95 (95% confidence interval [CI]: 0.91, 0.99), 0.72 (95% CI: 0.66, 0.77), 0.91 (95% CI: 0.88, 0.93), and 0.86 (95% CI: 0.79, 0.92). With ChestX-ray14, areas under the receiver operating characteristic curve were 0.94 (95% CI: 0.93, 0.96), 0.91 (95% CI: 0.89, 0.93), 0.94 (95% CI: 0.93, 0.95), and 0.81 (95% CI: 0.75, 0.86), respectively.ConclusionExpert-level models for detecting clinically relevant chest radiograph findings were developed for this study by using adjudicated reference standards and with population-level performance estimation. Radiologist-adjudicated labels for 2412 ChestX-ray14 validation set images and 1962 test set images are provided.© RSNA, 2019Online supplemental material is available for this article.See also the editorial by Chang in this issue.

Willingness to adopt wearable devices with behavioral and economic incentives by health insurance wellness programs: results of a US cross-sectional survey with multiple consumer health vignettes.

BACKGROUND: The number of health-related wearable devices is growing but it is not clear if Americans are willing to adopt health insurance wellness programs based on wearables and the incentives with which they would be more willing to adopt. METHODS: In this cross-sectional study we used a survey methodology, usage vignettes and a dichotomous scale to examine U.S. residents' willingness to adopt wearables (WTAW) in six use-cases where it was mandatory to use a wearable device and share the resulting data with a health insurance company. Each use-case was tested also for the influence of additional economic incentives on WTAW. RESULTS: A total of 997 Americans across 46 states participated in the study. Most of them were 25 to 34 years old (40.22%), 57.27% were female, and 74.52% were white. On average, 69.5% of the respondents were willing to adopt health-insurance use-cases based on wearable devices, though 77.8% of them were concerned about issues related to economic benefits, data privacy and to a lesser extent, technological accuracy. WTAW was 11-18% higher among consumers in use-cases involving health promotion and disease prevention. Furthermore, additional economic incentives combined with wearables increased WTAW overall. Notably, financial incentives involving providing healthcare credits, insurance premium discount, and/or wellness product discounts had particularly greater effectiveness for increasing WTAW in the consumer use-cases involving participation: for health promotion (RR = 1.06 for financial incentive, 95% CI: 1.01-1.11; P = 0.018); for personalized products and services (RR = 1.11 for financial incentive, 95% CI: 1.01-1.21; P = 0.018); and for automated underwriting discount at annual renewal (RR = 1.28 for financial incentive, 95% CI: 1.20-1.37; P < 0.001). CONCLUSIONS: Under the adequate economic, data privacy and technical conditions, 2 out of 3 Americans would be willing to adopt health insurance wellness programs based on wearable devices, particularly if they have benefits related to health promotion and disease prevention, and particularly with financial incentives.

The role of integrated medical and prescription drug plans in addressing racial and ethnic disparities in medication adherence.

Medication nonadherence in the United States contributes to 125,000 deaths and 10% of hospitalizations annually. The pain of preventable deaths and the personal costs of nonadherence are borne disproportionately by Black, Latino, and other minority groups because nonadherence is higher in these groups due to a variety of factors. These factors include socioeconomic challenges, issues with prescription affordability and convenience of filling and refilling them, lack of access to pharmacies and primary care services, difficulty taking advantage of patient engagement opportunities, health literacy limitations, and lack of trust due to historical and structural discrimination outside of and within the medical system. Solutions to address the drivers of lower medication adherence, specifically in minority populations, are needed to improve population outcomes and reduce inequities. While various solutions have shown some traction, these solutions have tended to be challenging to scale for wider impact. We propose that integrated medical and pharmacy plans are well positioned to address racial and ethnic health disparities related to medication adherence. DISCLOSURES: This study was not supported by any funding sources other than employment of all authors by Humana Inc. Humana products and programs are referred to in this article.

Diagnostic imaging rates for head injury in the ED and states' medical malpractice tort reforms.

OBJECTIVE: Physicians' fears of being sued may lead to defensive medical practices, such as ordering nonindicated medical imaging. We investigated the association between states' medical malpractice tort reforms and neurologic imaging rates for patients seen in the emergency department with mild head trauma. METHODS: We assessed neurologic imaging among a national sample of 8588 women residing in 10 US states evaluated in an emergency setting for head injury between January 1, 1992, and December 31, 2001. We assessed the odds of imaging as it varied by the enactment of medical liability reform laws. RESULTS: The medical liability reform laws were significantly associated with the likelihood of imaging. States with laws that limited monetary damages (odds ratio [OR], 0.63; 95% confidence interval [CI], 0.40-0.99), mandated periodic award payments (OR, 0.64; 95% CI, 0.43-0.97), or specified collateral source offset rules (OR, 0.62; 95% CI, 0.40-0.96) had an approximately 40% lower odds of imaging, whereas states that had laws that limited attorney's contingency fees had significantly higher odds of imaging (OR, 1.5; 95% CI, 0.99-2.4), compared to states without these laws. When we used a summation of the number of laws in place, the greater the number of laws, the lower the odds of imaging. In the multivariate analysis, after adjusting for individual and community factors, the total number of laws remained significantly associated with the odds of imaging, and the effect of the individual laws was attenuated, but not eliminated. CONCLUSION: The tort reforms we examined were associated with the propensity to obtain neurologic imaging. If these results are confirmed in larger studies, tort reform might mitigate defensive medical practices.

Diagnosing cervical spine instability: role of the post-computed tomography scan out-of-collar lateral radiograph.

BACKGROUND: Cervical spine injuries may have devastating neurological consequences, which makes accurate diagnosis of injury a key part of the trauma evaluation. Imaging plays a significant role in making the diagnosis and guiding management. Current American College of Radiology Appropriateness Criteria guidelines recommend computed tomography (CT) of the cervical spine with multi-planar reformats as the highest-rated imaging examination for patients meeting clinical criteria, without subsequent need for further imaging with a negative scan in a neurologically intact and non-obtunded patient. Although CT is fast and accurate for evaluation of bony injury, it may overlook ligamentous injury. OBJECTIVE AND CASE REPORT: We report a case in which ligamentous instability was demonstrated as subluxation with an out-of-collar lateral radiograph after a CT scan showed no subluxation or fracture in a patient without neurological deficits. Our Radiology Department routinely performs an out-of-collar lateral radiograph after a negative CT scan, and this case suggests that there may be an important role for this practice. CONCLUSION: Magnetic resonance is the optimal study for soft tissue and ligamentous evaluation; however, a simple lateral out-of-collar radiograph after CT clearance, in an otherwise neurologically intact and non-obtunded patient, may be a quick and cost-effective means to assess for instability of the cervical spine.

Profile of English salaried GPs: labour mobility and practice performance.

BACKGROUND: Recent national policy changes have provided greater flexibility in GPs' contracts. One such policy is salaried employment, which offers reduced hours and freedom from out-of-hours and administrative responsibilities, aimed at improving recruitment and retention in a labour market facing regional shortages. AIM: To profile salaried GPs and assess their mobility within the labour market. DESIGN OF STUDY: Serial cross-sectional study. SETTING: All GPs practising in England during the years 1996/1997, 2000/2001, and 2004/2005. METHOD: Descriptive analyses, logistic regression. RESULTS: Salaried GPs tended to be either younger (<35 years) or older (> or =65 years), female, or overseas-qualified; they favoured part-time working and personal medical services contracts. Salaried GPs were more mobile than GP principals, and have become increasingly so, despite a trend towards reduced overall mobility in the GP workforce. Practices with salaried GPs scored more Quality and Outcomes Framework points and were located in slightly more affluent areas. CONCLUSION: Salaried status appears to have reduced limitations in the labour market, leading to better workforce deployment from a GP's perspective. However, there is no evidence to suggest it has relieved inequalities in GP distribution.

Curbing adolescent smoking: a review of the effectiveness of various policies.

Tobacco-related mortality is one of the biggest killers in American medicine. Evidence suggests that if adolescents can be kept tobacco-free, most will never start using tobacco. Therefore, tobacco control policies directed at the youth population could provide an effective method for sustaining long-term reductions in smoking in all segments of the population. Many forms of tobacco control policies have been implemented including restrictive laws, public campaigns, and taxation duties; there has been disagreement over which is most effective. We investigate the efficacy of various methods of tobacco control in youth and present a review of the published evidence. Econometric data for both youth access restrictions and environmental tobacco smoke restrictions afford ambiguous results. Results vary in a continuum from a moderate negative effect toward, ironically, a marginal positive effect on smoking. While information dissemination policies may be somewhat effective on the onset, they are limited in their effect and eventually diminish over time. We conclude that increases in price affect teen smoking to a great degree. Most estimates show that for a 10 percent increase in prices, which could be implemented by a tax per pack, a 15 percent decrease in cigarettes consumed could be accomplished. Taxation policies are an effective means of preventative medicine.

Youth are more sensitive to price changes in cigarettes than adults.

Virtually all smoking begins in our population's youth and remains as a habit into those smokers' elder years. If we desire to halt smoking in its infancy, we should seek to deter and induce cessation in the youth years. It has been cited that taxation is an effective means to deter smoking at all ages, particularly efficacious in the youth population. This paper explores the merits of this method of preventative medicine, and intends to investigate differences between the price elasticity of cigarette demand between various cohorts, particularly the adult versus the youth population. We use a two-variable log-log, ordinary least-squares econometric regression to determine the extent that price alterations have on participation rates and quantity smoked. Our results show that youth are quite responsive to price increases showing a decrease of 14 percent prevalence in smoking for a 10 percent increase in price; whereas, the adult population is relatively less responsive to such price changes, exhibiting nearly a 2 percent decrease in prevalence for a 10 percent increase in price. We conclude that taxation is an effective means of socially-enacted preventative medicine in deterring youth smoking.

Expert witness blinding strategies to mitigate bias in radiology malpractice cases: a comprehensive review of the literature.

Like all physicians, radiologists in the United States are subject to frequent and costly medical malpractice claims. Legal scholars and physicians concur that the US civil justice system is neither precise nor accurate in determining whether malpractice has truly occurred in cases in which claims are made. Sometimes, this inaccuracy is driven by biases inherent in medical expert-witness opinions. For example, expert-witness testimony involving "missed" radiology findings can be negatively affected by several cognitive biases, such as contextual bias, hindsight bias, and outcome bias. Biases inherent in the US legal system, such as selection bias, compensation bias, and affiliation bias, also play important roles. Fortunately, many of these biases can be significantly mitigated or eliminated through the use of appropriate blinding techniques. This paper reviews the major works on expert-witness blinding in the legal scholarship and the radiology professional literature. Its purpose is to acquaint the reader with the evidence that unblinded expert-witness testimony is tainted by multiple sources of bias and to examine proposed strategies for addressing these biases through blinding.

Sustainable growth rate 2013: time for definitive intervention.

Federal healthcare spending has been a subject of intense concern as the US Congress continues to search for ways to reduce the budget deficit. The Congressional Budget Office (CBO) estimated that, even though it is growing more slowly than previously projected, federal spending on Medicare, Medicaid and the State Children's Health Insurance Program (SCHIP) will reach nearly $900 billion in 2013. In 2011 the Medicare program paid $68 billion for physicians and other health professional services, 12% of total Medicare spending. Since 2002 the sustainable growth rate (SGR) correction has called for reductions to physician reimbursements; however, Congress has typically staved off these reductions, although the situation remains precarious for physicians who accept Medicare. The fiscal cliff agreement that came into focus at the end of 2012 averted a 26.5% reduction to physician reimbursements related to the SGR correction. Nonetheless, the threat of these devastating cuts continues to loom. The Administration, Congress and others have devised many options to fix this unsustainable situation. This review explores the historical development of the SGR, touches on elements of the formula itself and outlines current proposals for fixing the SGR problem. A recent CBO estimate reduces the potential cost of a 10-year fix of SGR system to $138 billion. This has provided new hope for resolution of this long-standing issue.

The economic burden of incidentally detected findings.

Practitioners in all medical disciplines recognize the high frequency of incidentally detected findings. Although some findings are discovered on physical examination, an increasing majority are detected at imaging performed for another indication. With increasing federal scrutiny on the net value of imaging services, the costs and benefits of incidental findings need to be more rigorously quantified. In this article, the authors examine current related work on imaging expenditures for incidental findings and provide a framework for future investigations that will efficiently and substantially advance the knowledge in this field.

ACR Appropriateness Criteria® right lower quadrant pain--suspected appendicitis.

The diagnostic imaging of patients presenting with right lower quadrant pain and suspected appendicitis may be organized according to age and gender and to the presence or absence of "classic" signs and symptoms of acute appendicitis. Among adult patients presenting with clinical signs of acute appendicitis, the sensitivity and specificity of CT are greater than those of ultrasound, with improved performance when CT is performed with intravenous contrast. The use of rectal contrast has been associated with decreased time in the emergency department. Computed tomography has also been shown to reduce cost and negative appendectomy rates. Both CT and ultrasound are also effective in the identification of causes of right lower quadrant pain unrelated to appendicitis. Among pediatric patients, the sensitivity and specificity of graded-compression ultrasound can approach those of CT, without the use of ionizing radiation. Performing MRI after inconclusive ultrasound in pregnant patients has been associated with sensitivity and specificity of 80% to 86% and 97% to 99%, respectively. The ACR Appropriateness Criteria(®) are evidence-based guidelines for specific clinical conditions that are reviewed every 2 years by a multidisciplinary expert panel. The guideline development and review include an extensive analysis of current medical literature from peer-reviewed journals and the application of a well-established consensus methodology (modified Delphi) to rate the appropriateness of imaging and treatment procedures by the panel. In those instances in which evidence is lacking or not definitive, expert opinion may be used to recommend imaging or treatment.

The breadth of teaching commitment in radiology departments: a national survey.

PURPOSE: The aim of this study was to investigate the teaching effort and associated resources across variously sized academic and nonacademic hospitals across the country. METHODS: A simple survey was constructed and distributed nationally to academic and nonacademic radiology departments of attending-level radiologists with residency programs. The survey queried human and financial resources used for education in the department or division in comparison with compensated human and financial resources for this effort. RESULTS: Sixty institutions across the country were surveyed, with 40 surveys returned from 35 unique institutions, representing a 67% response rate. All respondents reported performing teaching duties. The average teaching in departments was 30 hours per week, including teaching a wide scope of constituents, including trainees, colleagues, and allied staff members. The majority of teaching was spent on teaching medical students, residents, and fellows as well as referring physicians at multidisciplinary rounds. The commitment of dedicated full-time equivalent staff members and funding to teaching was limited, with 45% of respondents without dedicated full-time equivalents for teaching and just over 7% of the teaching funded. Eighty percent of institutions surveyed provided some amount of weekly academic time during which many staff members carried out their teaching preparations. CONCLUSIONS: Respondent radiology departments reported significant resource and manpower allocations to teaching efforts, in particular to trainees and referring physicians. However, limited compensation or incentives are allocated to this effort with regard to both human and financial resources. Compensated, allotted staff academic time may help offset the lack of dedicated educational resources to the teaching effort of radiology departments.