The Sociodemographic Biases in Machine Learning Algorithms: A Biomedical Informatics Perspective.

Artificial intelligence models represented in machine learning algorithms are promising tools for risk assessment used to guide clinical and other health care decisions. Machine learning algorithms, however, may house biases that propagate stereotypes, inequities, and discrimination that contribute to socioeconomic health care disparities. The biases include those related to some sociodemographic characteristics such as race, ethnicity, gender, age, insurance, and socioeconomic status from the use of erroneous electronic health record data. Additionally, there is concern that training data and algorithmic biases in large language models pose potential drawbacks. These biases affect the lives and livelihoods of a significant percentage of the population in the United States and globally. The social and economic consequences of the associated backlash cannot be underestimated. Here, we outline some of the sociodemographic, training data, and algorithmic biases that undermine sound health care risk assessment and medical decision-making that should be addressed in the health care system. We present a perspective and overview of these biases by gender, race, ethnicity, age, historically marginalized communities, algorithmic bias, biased evaluations, implicit bias, selection/sampling bias, socioeconomic status biases, biased data distributions, cultural biases and insurance status bias, conformation bias, information bias and anchoring biases and make recommendations to improve large language model training data, including de-biasing techniques such as counterfactual role-reversed sentences during knowledge distillation, fine-tuning, prefix attachment at training time, the use of toxicity classifiers, retrieval augmented generation and algorithmic modification to mitigate the biases moving forward.

Improving the Quality and Utility of Electronic Health Record Data through Ontologies.

The translational research community, in general, and the Clinical and Translational Science Awards (CTSA) community, in particular, share the vision of repurposing EHRs for research that will improve the quality of clinical practice. Many members of these communities are also aware that electronic health records (EHRs) suffer limitations of data becoming poorly structured, biased, and unusable out of original context. This creates obstacles to the continuity of care, utility, quality improvement, and translational research. Analogous limitations to sharing objective data in other areas of the natural sciences have been successfully overcome by developing and using common ontologies. This White Paper presents the authors' rationale for the use of ontologies with computable semantics for the improvement of clinical data quality and EHR usability formulated for researchers with a stake in clinical and translational science and who are advocates for the use of information technology in medicine but at the same time are concerned by current major shortfalls. This White Paper outlines pitfalls, opportunities, and solutions and recommends increased investment in research and development of ontologies with computable semantics for a new generation of EHRs.

Differential Perceptions of What Constitutes a Medical Error Associated with Electronic Medical Records.

Perceptions of errors associated with healthcare information technology (HIT) often depend on the context and position of the viewer. HIT vendors posit very different causes of errors than clinicians, implementation teams, or IT staff. Even within the same hospital, members of departments and services often implicate other departments. Organizations may attribute errors to external care partners that refer patients, such as nursing homes or outside clinics. Also, the various clinical roles within an organization (e.g., physicians, nurses, pharmacists) can conceptualize errors and their root causes differently. Overarching all these perceptual factors, the definitions, mechanisms, and incidence of HIT-related errors are remarkably conflictual. There is neither a universal standard for defining or counting these errors. This paper attempts to enumerate and clarify the issues related to differential perceptions of medical errors associated with HIT. It then suggests solutions.

Selecting venues for AMIA events and conferences: guiding ethical principles.

A discussion and debate on the American Medical Informatics Association's (AMIA) Ethical, Legal, and Social Issues (ELSI) Working Group listserv in 2021 raised important issues related to a forthcoming conference in Texas. Texas had recently enacted a restrictive abortion law and restricted voting rights. Several AMIA members advocated for a boycott of the state and the scheduled conference. The discussion led the AMIA Board of Directors to request that the organization's Ethics Committee provide general guidance for principle-based venue selection. This document recommends overarching principles for the venue selection for future AMIA events and conferences. Discussions by the AMIA Board, the Ethics Committee, and the ELSI Working Group informed these recommendations, and this document on guiding principles was approved by the AMIA Board of Directors in April 2022.

A Mobile, Electronic Health Record-Connected Application for Managing Team Workflows in Inpatient Care.

BACKGROUND: Clinical workflows require the ability to synthesize and act on existing and emerging patient information. While offering multiple benefits, in many circumstances electronic health records (EHRs) do not adequately support these needs. OBJECTIVES: We sought to design, build, and implement an EHR-connected rounding and handoff tool with real-time data that supports care plan organization and team-based care. This article first describes our process, from ideation and development through implementation; and second, the research findings of objective use, efficacy, and efficiency, along with qualitative assessments of user experience. METHODS: Guided by user-centered design and Agile development methodologies, our interdisciplinary team designed and built Carelign as a responsive web application, accessible from any mobile or desktop device, that gathers and integrates data from a health care institution's information systems. Implementation and iterative improvements spanned January to July 2016. We assessed acceptance via usage metrics, user observations, time-motion studies, and user surveys. RESULTS: By July 2016, Carelign was implemented on 152 of 169 total inpatient services across three hospitals staffing 1,616 hospital beds. Acceptance was near-immediate: in July 2016, 3,275 average unique weekly users generated 26,981 average weekly access sessions; these metrics remained steady over the following 4 years. In 2016 and 2018 surveys, users positively rated Carelign's workflow integration, support of clinical activities, and overall impact on work life. CONCLUSION: User-focused design, multidisciplinary development teams, and rapid iteration enabled creation, adoption, and sustained use of a patient-centered digital workflow tool that supports diverse users' and teams' evolving care plan organization needs.

Using Artificial Intelligence With Natural Language Processing to Combine Electronic Health Record's Structured and Free Text Data to Identify Nonvalvular Atrial Fibrillation to Decrease Strokes and Death: Evaluation and Case-Control Study.

BACKGROUND: Nonvalvular atrial fibrillation (NVAF) affects almost 6 million Americans and is a major contributor to stroke but is significantly undiagnosed and undertreated despite explicit guidelines for oral anticoagulation. OBJECTIVE: The aim of this study is to investigate whether the use of semisupervised natural language processing (NLP) of electronic health record's (EHR) free-text information combined with structured EHR data improves NVAF discovery and treatment and perhaps offers a method to prevent thousands of deaths and save billions of dollars. METHODS: We abstracted 96,681 participants from the University of Buffalo faculty practice's EHR. NLP was used to index the notes and compare the ability to identify NVAF, congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke or transient ischemic attack, vascular disease, age 65 to 74 years, sex category (CHA2DS2-VASc), and Hypertension, Abnormal liver/renal function, Stroke history, Bleeding history or predisposition, Labile INR, Elderly, Drug/alcohol usage (HAS-BLED) scores using unstructured data (International Classification of Diseases codes) versus structured and unstructured data from clinical notes. In addition, we analyzed data from 63,296,120 participants in the Optum and Truven databases to determine the NVAF frequency, rates of CHA2DS2­VASc ≥2, and no contraindications to oral anticoagulants, rates of stroke and death in the untreated population, and first year's costs after stroke. RESULTS: The structured-plus-unstructured method would have identified 3,976,056 additional true NVAF cases (P<.001) and improved sensitivity for CHA2DS2-VASc and HAS-BLED scores compared with the structured data alone (P=.002 and P<.001, respectively), causing a 32.1% improvement. For the United States, this method would prevent an estimated 176,537 strokes, save 10,575 lives, and save >US $13.5 billion. CONCLUSIONS: Artificial intelligence-informed bio-surveillance combining NLP of free-text information with structured EHR data improves data completeness, prevents thousands of strokes, and saves lives and funds. This method is applicable to many disorders with profound public health consequences.

Why Is the Electronic Health Record So Challenging for Research and Clinical Care?

BACKGROUND: The electronic health record (EHR) has become increasingly ubiquitous. At the same time, health professionals have been turning to this resource for access to data that is needed for the delivery of health care and for clinical research. There is little doubt that the EHR has made both of these functions easier than earlier days when we relied on paper-based clinical records. Coupled with modern database and data warehouse systems, high-speed networks, and the ability to share clinical data with others are large number of challenges that arguably limit the optimal use of the EHR OBJECTIVES: Our goal was to provide an exhaustive reference for those who use the EHR in clinical and research contexts, but also for health information systems professionals as they design, implement, and maintain EHR systems. METHODS: This study includes a panel of 24 biomedical informatics researchers, information technology professionals, and clinicians, all of whom have extensive experience in design, implementation, and maintenance of EHR systems, or in using the EHR as clinicians or researchers. All members of the panel are affiliated with Penn Medicine at the University of Pennsylvania and have experience with a variety of different EHR platforms and systems and how they have evolved over time. RESULTS: Each of the authors has shared their knowledge and experience in using the EHR in a suite of 20 short essays, each representing a specific challenge and classified according to a functional hierarchy of interlocking facets such as usability and usefulness, data quality, standards, governance, data integration, clinical care, and clinical research. CONCLUSION: We provide here a set of perspectives on the challenges posed by the EHR to clinical and research users.

A Simple Assessment of Information Security Awareness in Hospital Staff Across Five Danish Regions.

Information Security Awareness among employees in healthcare has become an essential part in safeguarding health information systems against cyber-attacks and data breaches. We present three simple security awareness questions that can be included in larger surveys gauging other aspects of information systems. The questions have been tested in a national Danish survey to evaluate correlations among medical profession, computer proficiency, experience, and place of employment. We find that dissatisfaction with system usability is strongly linked with reduced information security awareness, and that clinical professions have different responses to security concerns.

A retrospective look at the predictions and recommendations from the 2009 AMIA policy meeting: did we see EHR-related clinician burnout coming?

Clinicians often attribute much of their burnout experience to use of the electronic health record, the adoption of which was greatly accelerated by the Health Information Technology for Economic and Clinical Health Act of 2009. That same year, AMIA's Policy Meeting focused on possible unintended consequences associated with rapid implementation of electronic health records, generating 17 potential consequences and 15 recommendations to address them. At the 2020 annual meeting of the American College of Medical Informatics (ACMI), ACMI fellows participated in a modified Delphi process to assess the accuracy of the 2009 predictions and the response to the recommendations. Among the findings, the fellows concluded that the degree of clinician burnout and its contributing factors, such as increased documentation requirements, were significantly underestimated. Conversely, problems related to identify theft and fraud were overestimated. Only 3 of the 15 recommendations were adjudged more than half-addressed.

Ethics and informatics in the age of COVID-19: challenges and recommendations for public health organization and public policy.

The COVID-19 pandemic response in the United States has exposed significant gaps in information systems and processes that prevent timely clinical and public health decision-making. Specifically, the use of informatics to mitigate the spread of SARS-CoV-2, support COVID-19 care delivery, and accelerate knowledge discovery bring to the forefront issues of privacy, surveillance, limits of state powers, and interoperability between public health and clinical information systems. Using a consensus-building process, we critically analyze informatics-related ethical issues in light of the pandemic across 3 themes: (1) public health reporting and data sharing, (2) contact tracing and tracking, and (3) clinical scoring tools for critical care. We provide context and rationale for ethical considerations and recommendations that are actionable during the pandemic and conclude with recommendations calling for longer-term, broader change (beyond the pandemic) for public health organization and policy reform.

Transitions from One Electronic Health Record to Another: Challenges, Pitfalls, and Recommendations.

OBJECTIVE: We address the challenges of transitioning from one electronic health record (EHR) to another-a near ubiquitous phenomenon in health care. We offer mitigating strategies to reduce unintended consequences, maximize patient safety, and enhance health care delivery. METHODS: We searched PubMed and other sources to identify articles describing EHR-to-EHR transitions. We combined these references with the authors' extensive experience to construct a conceptual schema and to offer recommendations to facilitate transitions. RESULTS: Our PubMed query retrieved 1,351 citations: 43 were relevant for full paper review and 18 met the inclusion criterion of focus on EHR-to-EHR transitions. An additional PubMed search yielded 1,014 citations, for which we reviewed 74 full papers and included 5. We supplemented with additional citations for a total of 70 cited. We distinguished 10 domains in the literature that overlap yet present unique and salient opportunities for successful transitions and for problem mitigation. DISCUSSION: There is scant literature concerning EHR-to-EHR transitions. Identified challenges include financial burdens, personnel resources, patient safety threats from limited access to legacy records, data integrity during migration, cybersecurity, and semantic interoperability. Transition teams must overcome inadequate human infrastructure, technical challenges, security gaps, unrealistic providers' expectations, workflow changes, and insufficient training and support-all factors affecting potential clinician burnout. CONCLUSION: EHR transitions are remarkably expensive, laborious, personnel devouring, and time consuming. The paucity of references in comparison to the topic's salience reinforces the necessity for this type of review and analysis. Prudent planning may streamline EHR transitions and reduce expenses. Mitigating strategies, such as preservation of legacy data, managing expectations, and hiring short-term specialty consultants can overcome some of the greatest hurdles. A new medical subject headings (MeSH) term for EHR transitions would facilitate further research on this topic.

Physicians' gender and their use of electronic health records: findings from a mixed-methods usability study.

OBJECTIVE: Physician burnout associated with EHRs is a major concern in health care. A comprehensive assessment of differences among physicians in the areas of EHR performance, efficiency, and satisfaction has not been conducted. The study sought to study relationships among physicians' performance, efficiency, perceived workload, satisfaction, and usability in using the electronic health record (EHR) with comparisons by age, gender, professional role, and years of experience with the EHR. MATERIALS AND METHODS: Mixed-methods assessments of the medical intensivists' EHR use and perceptions. Using simulated cases, we employed standardized scales, performance measures, and extensive interviews. NASA Task Load Index (TLX), System Usability Scale (SUS), and Questionnaire on User Interface Satisfaction surveys were deployed. RESULTS: The study enrolled 25 intensive care unit (ICU) physicians (11 residents, 9 fellows, 5 attendings); 12 (48%) were men, with a mean age of 33 (range, 28-55) years and a mean of 4 (interquartile range, 2.0-5.5) years of Epic experience. Overall task performance scores were similar for men (90% ± 9.3%) and women (92% ± 4.4%), with no statistically significant differences (P = .374). However, female physicians demonstrated higher efficiency in completion time (difference = 7.1 minutes; P = .207) and mouse clicks (difference = 54; P = .13). Overall, men reported significantly higher perceived EHR workload stress compared with women (difference = 17.5; P < .001). Men reported significantly higher levels of frustration with the EHR compared with women (difference = 33.15; P < .001). Women reported significantly higher satisfaction with the ease of use of the EHR interface than men (difference = 0.66; P =.03). The women's perceived overall usability of the EHR is marginally higher than that of the men (difference = 10.31; P =.06). CONCLUSIONS: Among ICU physicians, we measured significant gender-based differences in perceived EHR workload stress, satisfaction, and usability-corresponding to objective patterns in EHR efficiency. Understanding the reasons for these differences may help reduce burnout and guide improvements to physician performance, efficiency, and satisfaction with EHR use. DESIGN: Mixed-methods assessments of the medical intensivists' EHR use and perceptions. Using simulated cases, we employed standardized scales, performance measures, and extensive interviews.

Usability Across Health Information Technology Systems: Searching for Commonalities and Consistency.

Usability of health information technology (HIT) remains a predominant concern - one often exacerbated by clinicians' need to access information created by many different professionals in different settings, often using very dissimilar EHRs or even different configurations of the same EHR. Because of these variations, we argue that we must no longer think of usability as anchored in one setting, one EHR, one data standard, or one type of clinician. Rather, usability must be understood as a collective and constantly evolving process. This paper seeks to address that reality by 1) substantially expanding our previously-developed conceptual matrix of the wide range of settings and interfaces comprising modern HIT and 2) presenting actual examples of EHR usability issues with similar data but very different displays or processes.

Context and Meaning in EHR Displays.

Many Electronic Health Record (EHRs) data displays are insensitive to their settings, contexts, and to clinicians' needs. Yet, the contexts in which the data are displayed critically affect EHR usability and patient safety. Medication prescribing is a complex task; especially sensitive to contextual variation in EHR displays as vast variations in formats and logic are often unnecessarily confusing, leading to unwanted cognitive burdens and medical errors. With examples of EHR screenshots, we illustrate contextual variations in medication and allergy displays across different EHR systems and implementations-noting often seemingly haphazard differences that can lead to misunderstandings and misinterpretations.

Effect of Restriction of the Number of Concurrently Open Records in an Electronic Health Record on Wrong-Patient Order Errors: A Randomized Clinical Trial.

Importance: Recommendations in the United States suggest limiting the number of patient records displayed in an electronic health record (EHR) to 1 at a time, although little evidence supports this recommendation. Objective: To assess the risk of wrong-patient orders in an EHR configuration limiting clinicians to 1 record vs allowing up to 4 records opened concurrently. Design, Setting, and Participants: This randomized clinical trial included 3356 clinicians at a large health system in New York and was conducted from October 2015 to April 2017 in emergency department, inpatient, and outpatient settings. Interventions: Clinicians were randomly assigned in a 1:1 ratio to an EHR configuration limiting to 1 patient record open at a time (restricted; n = 1669) or allowing up to 4 records open concurrently (unrestricted; n = 1687). Main Outcomes and Measures: The unit of analysis was the order session, a series of orders placed by a clinician for a single patient. The primary outcome was order sessions that included 1 or more wrong-patient orders identified by the Wrong-Patient Retract-and-Reorder measure (an electronic query that identifies orders placed for a patient, retracted, and then reordered shortly thereafter by the same clinician for a different patient). Results: Among the 3356 clinicians who were randomized (mean [SD] age, 43.1 [12.5] years; mean [SD] experience at study site, 6.5 [6.0] years; 1894 females [56.4%]), all provided order data and were included in the analysis. The study included 12 140 298 orders, in 4 486 631 order sessions, placed for 543 490 patients. There was no significant difference in wrong-patient order sessions per 100 000 in the restricted vs unrestricted group, respectively, overall (90.7 vs 88.0; odds ratio [OR], 1.03 [95% CI, 0.90-1.20]; P = .60) or in any setting (ED: 157.8 vs 161.3, OR, 1.00 [95% CI, 0.83-1.20], P = .96; inpatient: 185.6 vs 185.1, OR, 0.99 [95% CI, 0.89-1.11]; P = .86; or outpatient: 7.9 vs 8.2, OR, 0.94 [95% CI, 0.70-1.28], P = .71). The effect did not differ among settings (P for interaction = .99). In the unrestricted group overall, 66.2% of the order sessions were completed with 1 record open, including 34.5% of ED, 53.7% of inpatient, and 83.4% of outpatient order sessions. Conclusions and Relevance: A strategy that limited clinicians to 1 EHR patient record open compared with a strategy that allowed up to 4 records open concurrently did not reduce the proportion of wrong-patient order errors. However, clinicians in the unrestricted group placed most orders with a single record open, limiting the power of the study to determine whether reducing the number of records open when placing orders reduces the risk of wrong-patient order errors. Trial Registration: clinicaltrials.gov Identifier: NCT02876588.