Search

Problem : U.S. medical schools are searching for ways to address issues of health justice in undergraduate medical education. Physicians have not typically received training in how to be effective advocates for systemic change and individuals in policy fields are not usually equipped to understand the complex issues of health science and their intersection with the health system and society. To address this gap, medical school faculty partnered with school of public policy faculty on a collaborative learning model that engaged MD and Master of Public Policy students together to strengthen their collective knowledge of the healthcare landscape, and to build skills to work for health justice. Intervention: We hypothesized that pairing medical students with public policy students to learn about the intersections of health justice and advocacy could enhance the efficacy of each group and provide a new model of collaboration between medical and policy professionals. The students collaborated on a health justice advocacy project through which they provided consultation to an established community organization. Context : The 8-week course took place in the spring of 2021 in Los Angeles, California. Due to Covid-19 the course was taught online and included asynchronous learning modules and live Zoom sessions. The project also served as a pilot for the post-clerkship phase of a new longitudinal health justice curriculum for MD students that launched in August 2021. Impact : Analysis of student work products, course evaluations, partner interviews, and student focus groups showed that students valued learning through their interdisciplinary collaborative work which gave them new perspectives on health justice issues. The community partners indicated that the students consultative work products were useful for their initiatives, and that they found working with MD and MPP students to be a valuable way to think about how to build stronger and more inclusive coalitions to advocate for health justice. This project has the potential for national impact as it aligns with the Association of American Medical Colleges' renewed focus on the responsibility of academic medicine to partner with communities for health justice. The project also contributed to the national conversation on how to align health systems science education with the aims of health justice through our participation in the American Medical Association Accelerating Change in Medical Education Consortium. Lessons Learned : Leveraging faculty relationships with community partners was crucial for developing meaningful projects for students. Cultivating and expanding community partner networks is necessary to sustain and scale up this type of intervention. Centering the needs of communities and supporting their on-going work for health justice is essential for becoming an effective advocate. Learning communities that bring interdisciplinary students, healthcare providers, policy professionals, and community partners together to learn from one another can create key opportunities for ameliorating health inequities.

Diffusion MRI studies consistently report group differences in white matter between individuals diagnosed with schizophrenia and healthy controls. Nevertheless, the abnormalities found at the group-level are often not observed at the individual level. Among the different approaches aiming to study white matter abnormalities at the subject level, normative modeling analysis takes a step towards subject-level predictions by identifying affected brain locations in individual subjects based on extreme deviations from a normative range. Here, we leveraged a large harmonized diffusion MRI dataset from 512 healthy controls and 601 individuals diagnosed with schizophrenia, to study whether normative modeling can improve subject-level predictions from a binary classifier. To this aim, individual deviations from a normative model of standard (fractional anisotropy) and advanced (free-water) dMRI measures, were calculated by means of age and sex-adjusted z-scores relative to control data, in 18 white matter regions. Even though larger effect sizes are found when testing for group differences in z-scores than are found with raw values (p < .001), predictions based on summary z-score measures achieved low predictive power (AUC < 0.63). Instead, we find that combining information from the different white matter tracts, while using multiple imaging measures simultaneously, improves prediction performance (the best predictor achieved AUC = 0.726). Our findings suggest that extreme deviations from a normative model are not optimal features for prediction. However, including the complete distribution of deviations across multiple imaging measures improves prediction, and could aid in subject-level classification., (© 2021 The Authors. Human Brain Mapping published by Wiley Periodicals LLC.)

Importance: Mammography screening currently relies on subjective human interpretation. Artificial intelligence (AI) advances could be used to increase mammography screening accuracy by reducing missed cancers and false positives., Objective: To evaluate whether AI can overcome human mammography interpretation limitations with a rigorous, unbiased evaluation of machine learning algorithms., Design, Setting, and Participants: In this diagnostic accuracy study conducted between September 2016 and November 2017, an international, crowdsourced challenge was hosted to foster AI algorithm development focused on interpreting screening mammography. More than 1100 participants comprising 126 teams from 44 countries participated. Analysis began November 18, 2016., Main Outcomes and Measurements: Algorithms used images alone (challenge 1) or combined images, previous examinations (if available), and clinical and demographic risk factor data (challenge 2) and output a score that translated to cancer yes/no within 12 months. Algorithm accuracy for breast cancer detection was evaluated using area under the curve and algorithm specificity compared with radiologists' specificity with radiologists' sensitivity set at 85.9% (United States) and 83.9% (Sweden). An ensemble method aggregating top-performing AI algorithms and radiologists' recall assessment was developed and evaluated., Results: Overall, 144 231 screening mammograms from 85 580 US women (952 cancer positive ≤12 months from screening) were used for algorithm training and validation. A second independent validation cohort included 166 578 examinations from 68 008 Swedish women (780 cancer positive). The top-performing algorithm achieved an area under the curve of 0.858 (United States) and 0.903 (Sweden) and 66.2% (United States) and 81.2% (Sweden) specificity at the radiologists' sensitivity, lower than community-practice radiologists' specificity of 90.5% (United States) and 98.5% (Sweden). Combining top-performing algorithms and US radiologist assessments resulted in a higher area under the curve of 0.942 and achieved a significantly improved specificity (92.0%) at the same sensitivity., Conclusions and Relevance: While no single AI algorithm outperformed radiologists, an ensemble of AI algorithms combined with radiologist assessment in a single-reader screening environment improved overall accuracy. This study underscores the potential of using machine learning methods for enhancing mammography screening interpretation.

Purpose: Since 2010, medical schools across the United States have engaged in a new cycle of curricular revision and renewal for their undergraduate medical curricula. But what structures, features, and trends have emerged in U.S. medical schools as a result of deliberate curricular redesign efforts? An analysis of the ways that medical schools have approached the reorganization of their curricula to prepare their students for the growing complexity of medical practice is presented., Method: This study drew a total pool of 40 U.S. MD-granting programs, of which 25 met the inclusion criteria for the study. The authors used a qualitative coding approach to materials from the undergraduate medical education (UME) program websites to identify 4 dimensions of strategies that these programs used to renew their curricula., Results: The analysis of the curricular maps and website content of the UME programs provided evidence for a continuum approach to the description of innovation strategies: 96% of schools employed a cohort-based linear pathway, 80% of schools used thematic basic science blocks, 47% placed their Step 1 exams outside of the second year, and 68% moved their clerkships to the second year., Conclusions: The Continuum of Innovation strategies will enable programs to renew their curricula in ways that promote deliberate curricular changes that are consistent with emerging needs in the field. This study and future research may be useful for UME programs with limited resources by providing consensus practices that enable them to plan curricular changes in ways that best serve their institutions.

Depth from Defocus (DFD) suggests a simple optical set-up to recover the shape of a scene through imaging with shallow depth of field. Although numerous methods have been proposed for DFD, less attention has been paid to the particular problem of alignment between the captured images. The inherent shift-variant defocus often prevents standard registration techniques from achieving the accuracy needed for successful shape reconstruction. In this paper, we address the DFD and registration problem in a unified framework, exploiting their mutual relation to reach a better solution for both cues. We draw a formal connection between registration and defocus blur, find its limitations and reveal the weakness of the standard isolated approaches of registration and depth estimation. The solution is approached by energy minimization. The efficiency of the associated numerical scheme is justified by showing its equivalence to the celebrated Newton-Raphson method and proof of convergence of the emerged linear system. The computationally intensive approach of DFD, newly combined with simultaneous registration, is handled by GPU computing. Experimental results demonstrate the high sensitivity of the recovered shapes to slight errors in registration and validate the superior performance of the suggested approach over two, separately applying registration and DFD alternatives.

Importance: The patient-centered medical home (PCMH) model holds promise for improving primary care delivery, but it has not been adequately tested in teaching settings., Design, Setting, and Participants: We implemented an intervention guided by PCMH principles at a safety-net teaching clinic with resident physician providers. Two similar clinics served as controls., Main Outcomes and Measures: Using a cross-sectional design, we measured the effect on patient and resident satisfaction using the Consumer Assessment of Healthcare Providers and Systems survey and a validated teaching clinic survey, respectively. Both surveys were conducted at baseline and 1 year after the intervention. We also measured the effect on emergency department and hospital utilization., Results: Following implementation of our intervention, the clinic’s score on the National Committee for Quality Assurance’s PCMH certification tool improved from 35 to 53 of 100 possible points, although our clinic did not achieve all must-pass elements to qualify as a PCMH. During the 1-year study period, 4676 patients were exposed to the intervention; 39.9% of these used at least 1 program component. Compared with baseline, patient-reported access and overall satisfaction improved to a greater extent in the intervention clinic, and the composite satisfaction rating increased from 48% to 65% in the intervention clinic vs from 50% to 59% in the control sites (P = .04). The improvements were particularly notable for questions relating to access. For example, satisfaction with urgent appointment scheduling increased from 12% to 53% in the intervention clinic vs from 14% to 18% in the control clinics (P < .001). Resident satisfaction also improved in the intervention clinic: the composite satisfaction score increased from 39% to 51% in the intervention clinic vs a decrease from 46% to 42% in the control clinics (P = .01). Emergency department utilization did not differ significantly between the intervention and control clinics, and hospitalizations increased from 26 to 27 visits per 1000 patients per month in the intervention clinic vs a decrease from 28 to 25 in the control clinics (P = .02)., Conclusions and Relevance: Our PCMH-guided intervention, which represented a modest but substantive step toward the PCMH vision, had favorable effects on patient and resident satisfaction at a safety-net teaching clinic but did not reduce emergency department or hospital utilization in the first year. Our experience may provide lessons for other teaching clinics in safety-net settings hoping to implement PCMH-guided reforms.