Dr. Christian Rose is a dual-boarded emergency physician and clinical informaticist specializing in the broad intersection of clinical medicine, informatics and innovation - specifically in machine learning, decision support, user-centered design and global health. He is particularly interested in the role of information systems to help to improve patient outcomes while allowing space for the human experience in medicine.
Dr. Rose began studying the effect of technology on the practice of medicine as part of his undergraduate degree in both Physics and Science, Technology and Society. As a medical student at Columbia University, with fantastic mentorship, he pursued numerous informatics projects including identifying alert fatigue in electronic ordering systems, gene discovery using big data and human-centered design for breast cancer decision aids and was awarded a Doris Duke Research Fellowship to pursue these interests as well as awards for his research in neoplastic disease and informatics.
He completed residency training at the University of California, San Francisco (UCSF), where he continued to broaden his scope of informatics interventions with projects ranging from radiology interface design to the development and deployment of a point-of-care decision aid to support the WHO’s Basic Emergency Care initiatives. He was selected as a chief resident in his final year leading to foundational experiences with data acquisition and analysis for continuous quality improvement initiatives.
Dr. Rose has since completed his informatics training at Stanford University where he had the opportunity to study the burgeoning field of deep learning and AI, exploring new methodologies for various clinical use cases and how they may be used to innovate clinical practice. However, it became clear that just because technologies are powerful and continually growing does not mean that they are the right solutions for every problem. Finding product fit and designing for the people that use these systems is ultimately necessary for their successful deployment.
In pursuing his goal of developing and implementing human-centered informatics solutions, Dr. Rose continues his innovative work here a Stanford where he works with an interdisciplinary team to develop and support the advancement of clinical practice through information technologies.
- Emergency Medicine
- Medical Informatics
- Machine Learning
- Decision Support Systems, Clinical
- Global Health
Assistant Professor - University Medical Line, Emergency Medicine
Honors & Awards
Donald A.B. Lindberg, MD Award for Excellence in Biomedical Informatics, Columbia University, College of Physicians & Surgeons (2013)
Miriam Berkman Spotnitz Award for Excellence in Neoplastic Disease Research, Columbia University, College of Physicians & Surgeons (2013)
Student Research Day Award, Columbia University, College of Physicians & Surgeons (2012)
Lucy Kellogg English Prize for Excellence in Physics, Vassar College (2007)
Boards, Advisory Committees, Professional Organizations
Member, Society for Academic Emergency Medicine (2013 - Present)
Member, American College of Emergency Physicians (2013 - Present)
Member, American Medical Informatics Association (2010 - Present)
Board Certification: American Board of Emergency Medicine, Emergency Medicine (2018)
Fellowship, Stanford/VA, Medical Informatics (2020)
Board Certification, American Board of Preventative Medicine, Clinical Informatics (2021)
Residency, University of California, San Francisco, Emergency Medicine (2017)
MD, Columbia University, College of Physicians and Surgeons (2013)
Community and International Work
Basic Emergency Care (BEC)
Global Health Emergency Medicine
Opportunities for Student Involvement
Technology and Education
Current Research and Scholarly Interests
Uncertainty permeates the practice of emergency medicine. There can be uncertainty in diagnosis: what causes particular symptoms, will they get worse, or what is the risk of a bad outcome? There can also be uncertainty in how to manage that diagnosis: should we watch and wait, attempt treatment A or B, and how do I decide which is best?
Attempting to answer these questions can help bring closure to patients and physicians alike, but at what cost? Testing can be expensive or even dangerous in the case of radiation exposure or stress testing. We all struggle to know more, to be more certain or less ambiguous, but little is known about the impact of things we cannot be certain about.
Ultimately, I want to answer the question: what do you do when you don't know what to do?
Signal from the Noise: A Mixed Graphical and Quantitative Process Mining Approach to Evaluate Care Pathways Applied to Emergency Stroke Care.
Journal of biomedical informatics
OBJECTIVE: Mapping real-world practice patterns vs. deviations from intended guidelines and protocols is necessary to identify and improve the quality of care for emergent medical conditions like acute ischemic stroke. Most status-quo process identification relies on expert opinion or direct observation, which can be biased or limited in scalability. We propose a mixed graphical and quantitative process mining approach to Electronic Health Record (EHR) event log data as a unique opportunity not only to more easily identify practice patterns, but also to compare real-world care processes and measure their conformance or variability.MATERIALS: Data was obtained from the event log of a major EHR vendor (Epic) for Stanford Health Care Hospital patients aged 18 years and older presenting to the ED from January 1, 2010 through December 31, 2018 and receiving tPA (tissue plasminogen activator) within 4.5 hours of presentation.METHODS: We developed an unsupervised process-mining algorithm to create a process map from clinical event logs. The method first identifies the most common events across the cohort. Then, all possible ordered events are recorded, and a summarized vector of nodes (events) and edges (events occurring in series) are mapped by their timing and probability. The highest probability ordered pairs are used to identify the most common path. We define measures for individual pathways conformity and average conformity across all encounters.RESULTS: Automatically generated process mining graphs, and specifically it's the most common path, mimicked our institutions recommended "code stroke" clinical pathway. The average conformity score for our cohort was 0.36 (i.e. paths had an average of 36% overlap with all possible paths), with a range from high of 0.64 and low of 0.20.DISCUSSION: This method allows for unsupervised visualization of the current state of common processes as well as their most common path, which can then be used to calculate the conformity of individual pathways through this process. These results may be used to evaluate the consistency of quality care at a given institution. It may also be extended to other common processes like sepsis or myocardial infarction care or even those which currently lack standardized clinical pathways.CONCLUSION: Our mixed graphical and quantitative process mining approach represents an essential data analysis step to improve complex care processes by automatically generating qualitative and quantitative process measures from existing event log data which can then be used to target quality improvement initiatives.
View details for DOI 10.1016/j.jbi.2022.104004
View details for PubMedID 35085813
Trial by fire: How physicians responding to the COVID-19 pandemic illuminated the need for digital credentials.
2022; 8: 20552076221084462
The current credentialing process for physicians struggled to accommodate fluctuating regional demands for providers during the severe acute respiratory syndrome coronavirus 2 pandemic. This hurdle highlighted existing inefficiencies and difficulties facing healthcare systems across the world and led us to explore how credentialing can be improved using digital technologies. We explain how this is a critical moment to make the shift from physical to digital credentials by specifying how a digital credentialing system could simplify onboarding for providers, enable secure expansion of telehealth services, and enhance information exchange.
View details for DOI 10.1177/20552076221084462
View details for PubMedID 35309389
View details for PubMedCentralID PMC8922044
Developing machine learning models to personalize care levels among emergency room patients for hospital admission.
Journal of the American Medical Informatics Association : JAMIA
OBJECTIVE: To develop prediction models for intensive care unit (ICU) vs non-ICU level-of-care need within 24 hours of inpatient admission for emergency department (ED) patients using electronic health record data.MATERIALS AND METHODS: Using records of 41 654 ED visits to a tertiary academic center from 2015 to 2019, we tested 4 algorithms-feed-forward neural networks, regularized regression, random forests, and gradient-boosted trees-to predict ICU vs non-ICU level-of-care within 24 hours and at the 24th hour following admission. Simple-feature models included patient demographics, Emergency Severity Index (ESI), and vital sign summary. Complex-feature models added all vital signs, lab results, and counts of diagnosis, imaging, procedures, medications, and lab orders.RESULTS: The best-performing model, a gradient-boosted tree using a full feature set, achieved an AUROC of 0.88 (95%CI: 0.87-0.89) and AUPRC of 0.65 (95%CI: 0.63-0.68) for predicting ICU care need within 24 hours of admission. The logistic regression model using ESI achieved an AUROC of 0.67 (95%CI: 0.65-0.70) and AUPRC of 0.37 (95%CI: 0.35-0.40). Using a discrimination threshold, such as 0.6, the positive predictive value, negative predictive value, sensitivity, and specificity were 85%, 89%, 30%, and 99%, respectively. Vital signs were the most important predictors.DISCUSSION AND CONCLUSIONS: Undertriaging admitted ED patients who subsequently require ICU care is common and associated with poorer outcomes. Machine learning models using readily available electronic health record data predict subsequent need for ICU admission with good discrimination, substantially better than the benchmarking ESI system. The results could be used in a multitiered clinical decision-support system to improve ED triage.
View details for DOI 10.1093/jamia/ocab118
View details for PubMedID 34402507
Professional development during a pandemic: a live virtual conference for emergency medicine chief residents.
Limited professional development training exists for chief residents. The available training uses in-person lectures and workshops at annual national conferences. The COVID-19 pandemic prevented most in-person gatherings in 2020, including pivotal onboarding and training events for new chief residents. However, for the last five years, Academic Life in Emergency Medicine's Chief Resident Incubator conducted year-long remote training programs, creating virtual communities of practice for chief residents in emergency medicine (EM). As prior leaders and alumni from the Incubator, we sought to respond to the limitations presented by the pandemic and create an onboarding event to provide foundational knowledge for incoming chief residents. We developed a half-day virtual conference, whereupon 219 EM chief residents enrolled. An effective professional development experience is feasible and scalable using online videoconferencing technologies, especially if constructed with content expertise, psychological safety, and production design in mind.
View details for DOI 10.1007/s43678-021-00146-3
View details for PubMedID 34264507
Reaching further: Lessons from the implementation of the WHO Basic Emergency Care Course Companion App in Tanzania.
African journal of emergency medicine : Revue africaine de la medecine d'urgence
2021; 11 (2): 325-330
Introduction: The World Health Organization's (WHO's) Basic Emergency Care (BEC) course was developed to address training gaps in low- and middle-income countries (LMICs). Simultaneously, LMICs have experienced an unprecedented increase in the number of cell phone and internet users. We developed a mobile application adjunct to the BEC course (BEC app) and sought to assess the reach of the BEC app.Methods: Forty-six BEC course participants, made up of doctors and nurses from three hospital sites in Tanzania, were given access to the BEC app with download instructions. Moderators tracked mobile access characteristics and barriers. This is a descriptive study outlining the implementation of the BEC app and associated findings from the process.Results: Fewer than 10% of participants were able to independently download and use the application. The download process revealed three key barrier areas: accessibility (no smartphone, smartphone without charge, no access to data/WiFi to download app, increased cost of data), technical (outdated operating system, inconsistent access to data/WiFi to run the app, insufficient phone storage), and participant-related characteristics (variability in smartphone literary, language discordance, smartphone turnover). Of the 46 participants, 29 (63%) were able to download and use the BEC app successfully with moderator support.Conclusions: There is potential utility of mobile health in LMICs. However, barriers still exist to reaching the largest possible audience for these initiatives. The importance of app compatibility with a broad range of operating systems and limitation of the amount of data needed to download and use the app was underscored by our study. Moreover, creative solutions are needed to facilitate large-scale roll-outs of mobile health interventions, such as a distribution model that relies on super user and peer support rather than an individual moderator. Additional local perspectives on the download process and the utilisation and acceptance of the application post-implementation are needed.
View details for DOI 10.1016/j.afjem.2021.04.001
View details for PubMedID 34012767
ALiEM Connect: Large-Scale, Interactive Virtual Residency Programming in Response to COVID-19.
Academic medicine : journal of the Association of American Medical Colleges
PROBLEM: The COVID-19 pandemic restricted in-person gatherings, including residency conferences. The pressure to quickly reorganize educational conferences and convert content to a remote format overwhelmed many programs. This article describes the pilot event of a large-scale, interactive virtual educational conference model designed and implemented by Academic Life in Emergency Medicine (ALiEM), called ALiEM Connect.APPROACH: The pilot ALiEM Connect event was conceptualized and implemented within a 2-week period in March 2020. The pilot was livestreamed via a combination of Zoom and YouTube and was archived by YouTube. Slack was used as a backchannel to allow interaction with other participants and engagement with the speakers (via moderators who posed questions from the backchannel to the speakers live during the videoconference).OUTCOMES: The RE-AIM (Reach, Effectiveness, Adoption, Implementation, Maintenance) framework was used for program evaluation, showing that 64 U.S. Accreditation Council for Graduate Medical Education-accredited emergency medicine residency programs participated in the pilot event, with 1,178 unique users during the event (reach). For effectiveness, 93% (139/149) of trainees reported the pilot as enjoyable and 85% (126/149) reported it was equivalent to or better than their usual academic proceedings. Adoption for ALiEM Connect was fairly good with 64/237 (27%) of invited residency programs registering and participating in the pilot event. Implementation was demonstrated by nearly half of the livestream viewers (47%, 553/1,178) interacting in the backchannel discussion, sending a total of 4,128 messages in the first 4 hours.NEXT STEPS: The final component of the RE-AIM framework, maintenance, will take more time to evaluate. Further study is required to measure the educational impact of events like the ALiEM Connect pilot. The ALiEM Connect model could potentially be used to replace educational conferences that have been cancelled or to implement and/or augment a large-scale, shared curriculum among residency programs in the future.
View details for DOI 10.1097/ACM.0000000000004122
View details for PubMedID 33883400
Utilizing Lean software development strategies to improve global eHealth initiatives: viewpoint from a basic emergency care app.
JMIR formative research
BACKGROUND: Health systems in low- and middle-income countries (LMICs) face considerable challenges in providing high-quality, accessible care. eHealth has had mounting interest as a possible solution given the unprecedented growth in cell phone and internet technologies in these locations. However, few apps or software programs have as of yet gone beyond the testing phase, most downloads are never opened, and consistent use is extremely rare. This is believed to be due to a failure to engage and meet local stakeholder needs as well as high costs of software development.OBJECTIVE: Feedback from World Health Organization (WHO) Basic Emergency Care (BEC) course participants requested a mobile, point-of-care adjunct to learning the complexities of the primary survey. Our team undertook the task of developing this solution through a community-based participatory model in an effort to meet trainees' reported needs and avoid some of the above failings. e aimed to use the well-described Lean software development strategy - owing to familiarity with its elements and ubiquitous use in medicine, global health and software development - to complete this task efficiently and with maximal stakeholder involvement.METHODS: From September 2016 through January 2017, the BEC App was roadmapped and developed at UCSF in California. When a prototype was complete, it was piloted in Cape Town, South Africa and Dar es Salaam, Tanzania - WHO BEC partner sites. Feedback from this pilot shaped continuous amendments to the app before subsequent user testing and study of the effect of use of the app on trainee retention of BEC course material.RESULTS: Our user-centered mobile app was developed relatively quickly and with high acceptance - 95% of BEC Course participants felt it was useful. Our solution had minimal direct costs and resulted in a robust infrastructure for subsequent assessment and maintenance which is familiar to the global health and medical communities and allows for efficient feedback and expansion.CONCLUSIONS: We believe that utilizing Lean software development strategies may help global health advocates and researchers build eHealth solutions utilizing a familiar process with buy-in across stakeholders that is responsive, rapid to deploy and sustainable.CLINICALTRIAL:
View details for DOI 10.2196/14851
View details for PubMedID 33882013
- Am I Part of the Cure or Am I Part of the Disease? Keeping Coronavirus Out When a Doctor Comes Home. The New England journal of medicine 2020
Novel educational adjuncts for the World Health Organization Basic Emergency Care Course: A prospective cohort study
AFRICAN JOURNAL OF EMERGENCY MEDICINE
2020; 10 (1): 30–34
The World Health Organization's (WHO) Basic Emergency Care Course (BEC) is a five day, in-person course covering basic assessment and life-saving interventions. We developed two novel adjuncts for the WHO BEC: a suite of clinical cases (BEC-Cases) to simulate patient care and a mobile phone application (BEC-App) for reference. The purpose was to determine whether the use of these educational adjuncts in a flipped classroom approach improves knowledge acquisition and retention among healthcare workers in a low-resource setting.We conducted a prospective, cohort study from October 2017 through February 2018 at two district hospitals in the Pwani Region of Tanzania. Descriptive statistics, Fisher's exact t-tests, and Wilcoxon ranked-sum tests were used to examine whether the use of these adjuncts resulted in improved learner knowledge. Participants were enrolled based on location into two arms; Arm 1 received the BEC course and Arm 2 received the BEC-Cases and BEC-App in addition to the BEC course. Both Arms were tested before and after the BEC course, as well as a 7-month follow-up exam. All participants were invited to focus groups on the course and adjuncts.A total of 24 participants were included, 12 (50%) of whom were followed to completion. Mean pre-test scores in Arm 1 (50%) were similar to Arm 2 (53%) (p=0.52). Both arms had improved test scores after the BEC Course Arm 1 (74%) and Arm 2 (87%), (p=0.03). At 7-month follow-up, though with significant participant loss to follow up, Arm 1 had a mean follow-up exam score of 66%, and Arm 2, 74%.Implementation of flipped classroom educational adjuncts for the WHO BEC course is feasible and may improve healthcare worker learning in low resource settings. Our focus- group feedback suggest that the course and adjuncts are user friendly and culturally appropriate.
View details for DOI 10.1016/j.afjem.2019.11.003
View details for Web of Science ID 000519198800007
View details for PubMedID 32161709
View details for PubMedCentralID PMC7058880
Physically Distant, Educationally Connected: Interactive Conferencing in the Era of COVID-19.
During the coronavirus outbreak, physical distancing restrictions led to the cancellation of live, large-group events worldwide. This included weekly educational conferences required of Emergency Medicine (EM) residency programs in the United States. Specifically, the Residency Review Committee in EM under the Accreditation Council for Graduate Medical Education has mandated that there be at least four hours per week of synchronous conference didactics.
View details for DOI 10.1111/medu.14192
View details for PubMedID 32324933
Context is Key: Using the Audit Log to Capture Contextual Factors Affecting Stroke Care Processes.
AMIA ... Annual Symposium proceedings. AMIA Symposium
2020; 2020: 953–62
High quality patient care through timely, precise and efficacious management depends not only on the clinical presentation of a patient, but the context of the care environment to which they present. Understanding and improving factors that affect streamlined workflow, such as provider or department busyness or experience, are essential to improving these care processes, but have been difficult to measure with traditional approaches and clinical data sources. In this exploratory data analysis, we aim to determine whether such contextual factors can be captured for important clinical processes by taking advantage of non-traditional data sources like EHR audit logs which passively track the electronic behavior of clinical teams. Our results illustrate the potential of defining multiple measures of contextual factors and their correlation with key care processes. We illustrate this using thrombolytic (tPA) treatment for ischemic stroke as an example process, but the measurement approaches can be generalized to multiple scenarios.
View details for PubMedID 33936471
Spokes for Our Folks: Public Health Bike Tour.
AEM education and training
2019; 3 (4): 393–95
Nearly half of medical care in the United States is managed through the emergency department, a large portion of which could be managed by "lateral" health services provided by public health facilities like human immunodeficiency virus (HIV) prophylaxis, alcohol and drug treatment programs, emergency psychiatric resources, and medical respite or rehabilitation centers. These options may be underutilized due to lack of knowledge of their services and demographics by patients and health care workers alike. We aimed to educate all levels of emergency medicine trainees and staff to citywide services via bike tour. Participants reported an improved understanding of health services as well as a sense of "camaraderie" toward lateral health services and other providers on the rides.
View details for DOI 10.1002/aet2.10371
View details for PubMedID 31637357
- Strategies to Enhance Wellness in Emergency Medicine Residency Training Programs ANNALS OF EMERGENCY MEDICINE 2017; 70 (6): 891–97
- Toward Precision Diagnostics ACADEMIC EMERGENCY MEDICINE 2017; 24 (5): 644–46