Ethical considerations for artificial intelligence in dermatology: a scoping review.

The field of dermatology is experiencing the rapid deployment of artificial intelligence (AI), from mobile applications (apps) for skin cancer detection to large language models like ChatGPT that can answer generalist or specialist questions about skin diagnoses. With these new applications, ethical concerns have emerged. In this scoping review, we aimed to identify the applications of AI to the field of dermatology and to understand their ethical implications. We used a multifaceted search approach, searching PubMed, MEDLINE, Cochrane Library and Google Scholar for primary literature, following the PRISMA Extension for Scoping Reviews guidance. Our advanced query included terms related to dermatology, AI and ethical considerations. Our search yielded 202 papers. After initial screening, 68 studies were included. Thirty-two were related to clinical image analysis and raised ethical concerns for misdiagnosis, data security, privacy violations and replacement of dermatologist jobs. Seventeen discussed limited skin of colour representation in datasets leading to potential misdiagnosis in the general population. Nine articles about teledermatology raised ethical concerns, including the exacerbation of health disparities, lack of standardized regulations, informed consent for AI use and privacy challenges. Seven addressed inaccuracies in the responses of large language models. Seven examined attitudes toward and trust in AI, with most patients requesting supplemental assessment by a physician to ensure reliability and accountability. Benefits of AI integration into clinical practice include increased patient access, improved clinical decision-making, efficiency and many others. However, safeguards must be put in place to ensure the ethical application of AI.

The use of artificial intelligence (AI) in dermatology is rapidly increasing, with applications in dermatopathology, medical dermatology, cutaneous surgery, microscopy/spectroscopy and the identification of prognostic biomarkers (characteristics that provide information on likely patient health outcomes). However, with the rise of AI in dermatology, ethical concerns have emerged. We reviewed the existing literature to identify applications of AI in the field of dermatology and understand the ethical implications. Our search initially identified 202 papers, and after we went through them (screening), 68 were included in our review. We found that ethical concerns are related to the use of AI in the areas of clinical image analysis, teledermatology, natural language processing models, privacy, skin of colour representation, and patient and provider attitudes toward AI. We identified nine ethical principles to facilitate the safe use of AI in dermatology. These ethical principles include fairness, inclusivity, transparency, accountability, security, privacy, reliability, informed consent and conflict of interest. Although there are many benefits of integrating AI into clinical practice, our findings highlight how safeguards must be put in place to reduce rising ethical concerns.

Civil liability for clinical misdiagnosis of suicidal intention: procedure and guidelines to minimize fatal diagnostic error / La responsabilidad civil por el diagnóstico clínico erróneo de la intención suicida: procedimiento y pautas para minimizar el error diagnóstico mortal

Background/Objectives: A fatal diagnostic error of suicidal intention, i.e., an error implying death or serious injuries (i.e., incapacitating, chronic injury) to the patient, may have civil liability (punishable error) for the clinician. The Scale for Suicidal Ideation (SSI) is the reference psychometric instrument used to measure suicidal intention. A meta-analytical review was designed with the aim of estimating the true reliability of the SSI in general and in different settings (moderators) with the aim of correcting unreliability raw scores. Method: A total of 90 primary studies reporting SSI's reliability (internal consistency) was found in the literature, yielding a total of 92 effect sizes. Bare-bones meta-analysis of correlation coefficients correcting effect by sampling error were run. Results: The results showed an overall mean true internal consistency of .8904, 95% CI [.8878, .8930], meaning that 42.6% of the population standard deviation is error and 18.11% of an individual's measure is error. Additional estimations (moderators) of SSI's reliability for gender (men: .8873, women: .8808) adaptation version (English: .9212, Korean: .9052, Chinese: .8402, Italian: .9163, Persian: .8612), and population (subclinical: .8769, general: .9230, mental illness: .9040) were obtained. All mean true estimations were under the desirable standard for applied settings where critical decisions are made, .95. Furthermore, for populations with risk of suicide, such as prison inmates and militaries, mean true reliability could not be computed as k was insufficient. Conclusion: Implications of true reliabilities obtained for the estimation of individuals' true scores and population standard deviations are discussed. Examples of computation of true scores to minimize fatal diagnosis errors were performed for both known reliability and unknown reliability settings (e.g., risk populations). (AU)

Antecedentes/Objetivo: Un error diagnóstico mortal de intención suicida, es decir, un error que implique la muerte o lesiones graves (es decir, incapacitantes, crónicas) para el paciente, puede tener responsabilidad civil (error punible) para el clínico. La Escala de Ideación Suicida (SSI) es el instrumento psicométrico de referencia utilizado para medir la intención suicida. En este estudio se diseñó una revisión metaanalítica para medir la fiabilidad real de la SSI en general y en diferentes ámbitos (moderadores), con el fin de corregir las puntuaciones brutas por la falta de fiabilidad de la medida. Método: En la búsqueda de estudios se encontró un total de 90 estudios primarios que trataban de la fiabilidad (consistencia interna) del SSI, lo que arrojó un total de 92 tamaños del efecto. Se ejecutaron meta-análisis del tipo bare-bone corrigiendo el efecto por el error de muestreo. Resultados: Se obtuvo una consistencia interna verdadera media global de .8904, IC 95% [.8878, .8930], es decir, el 42.6% de la desviación estándar de la población es error y el 18.11% de la medida de un sujeto es error. Se obtuvieron estimaciones adicionales (moderadores) de la fiabilidad del SSI para el género (hombres: .8873, mujeres: .8808), versión de adaptación (inglés: .9212, coreano: .9052, chino: .8402, italiano: .9163, persa: .8612) y población (subclínica: .8769, general: .9230, enfermedad mental: .9040). Todas las estimaciones verdaderas medias estaban por debajo del estándar deseable para entornos aplicados donde se toman decisiones críticas, .95. Además, para las poblaciones de riesgo de suicidio, como reclusos y militares, no se pudo calcular la fiabilidad media verdadera, ya que k era insuficiente. Conclusiones: Se discuten las implicaciones de la fiabilidad verdadera obtenida para la estimación de las puntuaciones verdaderas de los individuos y las desviaciones típicas de las poblaciones. ... (AU)

Insuficiencias metodológicas lastran el diagnóstico de la enfermedad celiaca / Methodological shortage hampers the diagnostic of celiac disease

Introducción: La enfermedad celiaca es el resultado de una sensibilidad permanente al gluten. Puede conducir principalmente a trastornos intestinales. Cuatro criterios son utilizados para el diagnóstico de esta enfermedad: clínicos, histológicos, serológicos y moleculares. La insuficiente utilización de estos criterios conduce a falsos diagnósticos de dicha enfermedad. Objetivo: Demostrar la existencia de falsos diagnósticos de enfermedad celiaca cuando no se utilizan las herramientas necesarias para ello. Métodos: Se estudiaron 46 niños que fueron remitidos al Servicio de Genética Molecular del Hospital Hermanos Ameijeiras con diagnóstico de enfermedad celiaca basado en criterios clínicos e histopatológicos. Para completar los procederes diagnósticos, a cada paciente se le determinó anticuerpos antitransglutaminasa previa ingesta de gluten, y los alelos HLA DQ2 y HLA DQ8. Se consideraron pacientes con enfermedad celiaca aquellos casos que cumplieron los cuatro criterios. Resultados: De los 46 pacientes, trece (28,3 por ciento) fueron negativos a los alelos HLA DQ2/HLA DQ8, lo que niega estén padeciendo de enfermedad celiaca; ocho (17,39 por ciento) fueron positivos a los alelos HLA y negativos a la presencia de anticuerpos, lo que también niega la enfermedad. Es decir, 21 (45,7 por ciento) eran falsos diagnósticos de enfermedad celiaca. Los 25 (54,3 por ciento) restantes, además de los criterios con que fueron remitidos, cumplieron los serológicos (positividad a anticuerpos antitransglutaminasa) y moleculares (positividad para moléculas HLA DQ2/HLADQ8). Conclusiones: Para un diagnóstico de certeza de enfermedad celiaca es necesario, además de las herramientas clínicas e histopatológicas utilizadas en la red de hospitales pediátricos del país, el uso de procederes serológicos y moleculares(AU)

Introduction: Celiac disease is a caused by a permanent sensitivity to gluten, which results mainly in functional disorders of the small intestine. To successfully diagnose of celiac disease, it is necessary to properly convey four criteria: clinic, histological, serological and molecular. The insufficient utilization of them in the medical practice could conduce to false diagnosis of celiac disease. Objective: To demonstrate the occurrence of mistaken diagnoses of celiac disease when the four criteria are not properly addressed. Methods: Forty-six children were diagnosed with celiac disease based on clinical and histopathological criteria and remitted to the Hermanos Ameijeiras Hospital´s Molecular Genetics service. In order to complete the serological and molecular diagnosis procedure, there were detected antitransglutaminase antibodies after gluten ingestion, and HLA DQ2/HLA DQ8 alleles in every child. Individuals who met the four criteria were considered celiac disease patients. Results: The analysis of 46 patients showed that 13 (28.3 percent) where negative to the presence of both allele HLA DQ2/HLA DQ8, and hence negative for celiac disease diagnosis. Eight patients (17.39 percent) where HLA DQ2/HLA DQ8 positive and antitransglutaminase antibodies negative, so they were considered as negative for diagnosis of celiac disease. According to our results, 21 patients (45.7 percent) were mistakenly diagnosed. The remaining 25 patients (54.3 percent) where positive for all diagnosis criteria. Conclusions: In order to successfully diagnose of celiac disease, in addition to clinical and histopathological tools used in the network of pediatrics hospitals in the country, it is necessary to include the serological and molecular method(AU)

When No One Notices: Disorders of Consciousness and the Chronic Vegetative State.

On January 5, 2019, the Associated Press reported that a woman thought to have been in the vegetative state for over a decade gave birth at a Hacienda HealthCare facility. Until she delivered, the staff at the Phoenix center had not noticed that their patient was pregnant. The patient was also misdiagnosed. Misdiagnosis of patients with disorders of consciousness in institutional settings is more the norm than the exception. Misdiagnosis is also connected to a broad and extremely significant change in the understanding of the vegetative state-a change that the field of bioethics has not yet fully taken into account. In September 2018, the American Academy of Neurology, the American College of Rehabilitation Medicine, and the National Institute on Disability, Independent Living, and Rehabilitation Research issued a comprehensive evidence-based review on disorders of consciousness and an associated practice guideline on the care of these patients. These landmark publications update the 1994 Multi-Society Task Force Report on the Vegetative State, which subcategorized the persistent vegetative state as either persistent (once the vegetative state lasted one month) or permanent (once the vegetative state lasted three months after anoxic injury or twelve months after traumatic injury). Noting that 20 percent of patients thought to be permanently unconscious might regain some level of consciousness, the new guideline has eliminated the permanent vegetative state as a diagnostic category, replacing it with the chronic vegetative state.

Missed diagnosis of cancer in primary care: Insights from malpractice claims data.

BACKGROUND: In the ambulatory setting, missed cancer diagnoses are leading contributors to patient harm and malpractice risk; however, there are limited data on the malpractice case characteristics for these cases. OBJECTIVE: The aim of this study was to examine key features and factors identified in missed cancer diagnosis malpractice claims filed related to primary care and evaluate predictors of clinical and claim outcomes. METHODS: We analyzed 2155 diagnostic error closed malpractice claims in outpatient general medicine. We created multivariate models to determine factors that predicted case outcomes. RESULTS: Missed cancer diagnoses represented 980 (46%) cases of primary care diagnostic errors, most commonly from lung, colorectal, prostate, or breast cancer. The majority (76%) involved errors in clinical judgment, such as a failure or delay in ordering a diagnostic test (51%) or failure or delay in obtaining a consult or referral (37%). These factors were independently associated with higher-severity patient harm. The majority of these errors were of high severity (85%). CONCLUSIONS: Malpractice claims involving missed diagnoses of cancer in primary care most often involve routine screening examinations or delays in testing or referral. Our findings suggest that more reliable closed-loop systems for diagnostic testing and referrals are crucial for preventing diagnostic errors in the ambulatory setting.

Revisiting Death: Implicit Bias and the Case of Jahi McMath.

For nearly five years, bioethicists and neurologists debated whether Jahi McMath, an African American teenager, was alive or dead. While Jahi's condition provides a compelling study for analyzing brain death, circumscribing her life status to a question of brain death fails to acknowledge and respond to a chronic, if uncomfortable, bioethics problem in American health care-namely, racial bias and unequal treatment, both real and perceived. Bioethicists should examine the underlying, arguably broader social implications of what Jahi's medical treatment and experience represented. On any given day, disparities in the quality of health care and health outcomes for people of color in comparison to whites are evidenced in American hospitals and clinics. These disparities are not entirely explained by differences in patient education, insurance status, employment, income, expressed preference for treatments, and severity of disease. Instead, research indicates that, even for African Americans able to gain access to health care services and navigate institutional nuances, disparities persist across a broad range of services, including diagnostic screening and general medical care, mental health diagnosis and treatment, pain management, HIV-related care, and treatments for cancer, heart disease, diabetes, and kidney disease.

The Case of Jahi McMath: A Neurologist's View.

From the start, I followed the case of Jahi McMath with great interest. In December 2013, she clearly fulfilled the diagnostic criteria for brain death. As a neurologist with a special interest in chronic brain death, I was not surprised that, after she was flown to New Jersey, where she became statutorily resurrected and was treated as a comatose patient, Jahi's condition quickly improved. In 2014, her family reported that she sometimes responded to simple motor commands. I shared the general skepticism regarding these reports, assuming that the family was in denial and was misinterpreting spinal myoclonus (a rapid, involuntary twitch generated by the spinal cord) as volitional. The family had noticed that when Jahi's heart rate was above eighty beats per minute, she was more likely to respond, as though the heart rate reflected some sort of inner level of arousal. So they began to make video recordings. I have been privileged to be entrusted with copies of these recordings, forty-eight of which proved suitable for assessing alleged responsiveness. All have been certified by a forensic video expert as unaltered. The first thing that struck me was that the great majority of the alleged responses were not spinal myoclonus. In fact, they did not resemble any type of spontaneous, involuntary movement described in patients paralyzed from high spinal cord lesions.

Lessons from the Case of Jahi McMath.

Jahi McMath's case has raised challenging uncertainties about one of the most profound existential questions that we can ask: how do we know whether someone is alive or dead? The case is striking in at least two ways. First, how can it be that a person diagnosed as dead by qualified physicians continued to live, at least in a biological sense, more than four years after a death certificate was issued? Second, the diagnosis of brain death has been considered irreversible; in fact, there has never been a case of a person correctly diagnosed as brain-dead who improved to the point that the person no longer fulfilled the diagnostic criteria. If the neurologist Alan Shewmon is correct that, prior to her cardiac arrest in June 2018, McMath no longer met the criteria for brain death and was actually in a minimally conscious state, this case could have momentous implications for how we think about this diagnosis going forward. In this essay, I will offer a hypothesis that could, perhaps, explain both these aspects of the case. The hypothesis is based on differences in how we distinguish between biological and legal categories. The law tends to prefer to draw bright-line distinctions between categories, whereas biological categories tend to fall along a spectrum, without sharp distinctions.

Are Concerns About Irremediableness, Vulnerability, or Competence Sufficient to Justify Excluding All Psychiatric Patients from Medical Aid in Dying?

Some jurisdictions that have decriminalized assisted dying (like Canada) exclude psychiatric patients on the grounds that their condition cannot be determined to be irremediable, that they are vulnerable and in need of protection, or that they cannot be determined to be competent. We review each of these claims and find that none have been sufficiently well-supported to justify the differential treatment psychiatric patients experience with respect to assisted dying. We find bans on psychiatric patients' access to this service amount to arbitrary discrimination. Proponents of banning the practice ignore or overlook alternatives to their proposal, like an assisted dying regime with additional safeguards. Some authors have further criticized assisted dying for psychiatric patients by highlighting allegedly problematic practices in those countries which allow it. We address recent evidence from the Netherlands, showing that these problems are either misrepresented or have straightforward solutions. Even if one finds such evidence troubling despite our analysis, other jurisdictions need not adopt every feature of the Dutch system.

Commentary on duty of candour and cancer screening.

English law mandates a duty of candour (DOC) for all healthcare providers. They must be open and honest when something goes wrong with care causing harm. Providers must apologize to those affected and investigate what happened. Screening is not 100% accurate and false positive and false negative results are inevitable. Guidance on DOC assists providers to judge when something has gone wrong in screening and the DOC legislation applies. DOC guidance helps distinguish such incidents from harms that are an expected and inevitable consequence of the imperfections of screening tests. For breast cancer screening the classification of interval cancers has been updated to take account of DOC. This guidance on DOC and classification of prior films of those presenting with interval cancers has relevance to other areas of diagnostic imaging. Review of prior examinations after a significant diagnosis has been made may reveal a previously overlooked abnormality.

Sometimes Those Hoofbeats Are Zebras: A Narrative Analysis.

The case of BB, an 11-year-old girl who was hospitalized because of sudden odd seizure-like symptoms and catatonic affect, highlights several ethical issues and communication problems. The correct diagnosis was initially missed, partly because physicians are trained to think of the most common explanation for a patient's symptoms; the medical education truism "when you hear hoofbeats, think horses, not zebras" was not helpful in BB's case. The common habit of medical professionals to not revisit a diagnosis once one is established also led to missed opportunities to provide appropriate care for this young patient. The difficulty nurses and/or family members have in questioning a diagnosis and treatment plan are also discussed.

What If the Prenatal Diagnosis of a Lethal Anomaly Turns Out to Be Wrong?

Advances in prenatal diagnosis create a unique set of clinical ethics dilemmas. Doctors routinely obtain genetic screening, radiologic images, and biophysical profiling. These allow more accurate diagnosis and prognosis than has ever before been possible. However, they also reveal a wider range of disease manifestations than were apparent when prenatal diagnosis was less sophisticated. Sometimes, the best estimates of prognosis turn out to be wrong. The infant's symptoms may be less severe or more severe than anticipated based on prenatal assessment. We present a case in which a prenatal diagnosis was made of severe osteogenesis imperfecta, leading to a decision to induce delivery at 31 weeks. On postnatal evaluation, the infant's disease did not appear to be as bad as had been anticipated. We discuss the ethical implications of such diagnostic and prognostic errors.