Mayo Clinic Proceedings. Digital health最新文献

Long QT syndrome (LQTS) is a potentially life-threatening cardiac repolarization disorder characterized by an increased risk of fatal arrhythmias. Accurate and timely diagnosis is essential for risk stratification and appropriate management. However, traditional diagnostic approaches have limitations, necessitating more objective and efficient tools. Artificial intelligence (AI) offers promising solutions by enhancing the accuracy and efficiency of electrocardiography (ECG) interpretation. The AI algorithms can process ECG data more rapidly than human experts, providing real-time analysis and prompt identification of individuals at risk, and reducing interobserver variability. By analyzing large volumes of ECG data, AI algorithms can extract meaningful features that may not be apparent to the human eye. Advancements in AI-driven corrected QT interval monitoring using mobile ECG devices, such as smartwatches, offer a valuable and convenient tool for identifying individuals at risk of LQTS-related complications, which is particularly applicable during pandemic conditions, such as COVID-19. Integration of AI into clinical practice poses a number of challenges. Bias in data gathering and patient privacy concerns are important considerations that must be addressed. Safeguarding patient privacy and ensuring data protection are crucial for maintaining trust in AI-driven systems. In addition, the interpretability of AI algorithms is a concern because understanding the decision-making process is essential for clinicians to trust and confidently use these tools. Future perspectives in this field may involve the integration of AI into diagnostic protocols, through genetic subtype classifications on the basis of ECG data. Moreover, explainable AI techniques aim to uncover ECG features associated with LQTS diagnosis, suggesting new insights into LQTS pathophysiology.

Objective

To compare patient satisfaction between telemedicine and in-person visits for historically vulnerable groups at risk of worse experiences with telemedicine.

Patients and Methods

Individuals seen at Mount Sinai Beth Israel Department of Medicine ambulatory practices from April 23, 2020, to March 7, 2023, who completed a post-video or in-person appointment survey. Primary outcomes were: satisfaction with ability to get appointments, quality of time with doctor, explanations from care team, and likelihood to recommend practice. Patients were subdivided by age, gender, English proficiency, and clinician type.

Results

Among 8948 in-person and 1101 telemedicine visits, telemedicine scored lower in how the clinical team explained care to patients in the first year, but differences diminished thereafter. Within subgroups, those who were older than 65 years, non-English speakers, and seen by a faculty physician had a lower satisfaction with telemedicine that improved after the first year. Lack of English proficiency was a predictor of lower satisfaction in both types of visits, whereas older age and faculty physician were predictors of higher in-person visit satisfaction, and medicine subspecialties were linked to better telehealth visit satisfaction.

Conclusion

These findings suggest improved patient satisfaction with time after the initial COVID-19 expansion, both broadly and within subgroups, but overall differences between in-person and telehealth visits do not appear to be clinically significant. There appear to be differences among certain populations that warrant further study and may require targeted intervention to maintain quality of care.

Medication adherence has long been viewed as a patient issue, but what if we shift this perspective? What if medications could adjust to the needs and context of patients, instead of the other way around? We used design thinking to create a contextually sensitive digital health mobile application to improve medication adherence in older adults. We define contextual sensitivity as sensitivity to the context of patient needs. Through persona and scenario ideation, interviews, evaluations of existing solutions, prototypes, and consultations with subject matter experts, we uncovered key barriers to medication adherence. We outline 4 key challenges: alert fatigue, poor health literacy, lack of social support, and lack of behavior change and motivation, which are specific to older adults. The resulting application features reminders and alerts, a dashboard and calendar, educational resources, social sharing, and reward features. These 5 elements emphasize the significance of design thinking, contextual sensitivity, trimodal alerts, and co-interventions in developing effective digital health solutions for medication adherence among older adults.

Objective

To develop and validate a machine learning model that predicts the most successful antihypertensive treatment for an individual.

Patients and Methods

The causal, deep neural network-based model was trained on data from 16,917 newly diagnosed hypertensive patients attending Mayo Clinic’s primary care practices from January 1, 2005, to December 31, 2021. Eligibility criteria included a diagnosis of primary hypertension, blood pressure and creatinine measurements before antihypertensive treatment, treatment within 9 months of diagnosis, and at least 1 year of follow up. The primary outcome was model performance in predicting the likelihood of a successful antihypertensive treatment 1 year from the start of treatment. Treatment success was defined as achieving blood pressure control with no moderate or severe adverse effects. Model validation and guideline agreement was assessed on 1000 patients.

Results

In the training set of 16,917 participants (60.8±14.7 years; 8344 [49.3%] women), 33.8% achieved blood pressure control without moderate or severe adverse effects for at least a year with initial treatment. The most common treatment was angiotensin-converting enzyme inhibitor (39.1% average success), and the most successful was angiotensin-converting enzyme inhibitor-thiazide combination (44.4% average success). Our custom-built causal, deep neural network-based model exhibited the highest accuracy in predicting individualized treatment success with a precision of 51.7%, recall of 44.4%, and F1 score of 47.8%. Compared with actual physician practice on the validation set (77.9% agreement), the algorithm aligned with the Eighth Joint National Committee hypertension guidelines 95.7% of the time.

Conclusion

A machine learning algorithm can accurately predict the likelihood of antihypertensive treatment success and help personalize hypertension management.

Objective

To investigate the feasibility and accuracy of artificial intelligence (AI) methods of facial deidentification in hospital-recorded epileptic seizure videos, for improved patient privacy protection while preserving clinically important features of seizure semiology.

Patients and Methods

Videos of epileptic seizures displaying seizure-related involuntary facial changes were selected from recordings at Taipei Veterans General Hospital Epilepsy Unit (between August 1, 2020 and February 28, 2023), and a single representative video frame was prepared per seizure. We tested 3 AI transformation models: (1) morphing the original facial image with a different male face; (2) substitution with a female face; and (3) cartoonization. Facial deidentification and preservation of clinically relevant facial detail were calculated based on: (1) scoring by 5 independent expert clinicians and (2) objective computation.

Results

According to the clinician scoring of 26 facial frames in 16 patients, the best compromise between deidentification and preservation of facial semiology was the cartoonization model. A male facial morphing model was superior to the cartoonization model for deidentification, but clinical detail was sacrificed. Objective similarity testing of video data reported deidentification scores in agreement with the clinicians’ scores; however, preservation of semiology gave mixed results likely due to inadequate existing comparative databases.

Conclusion

Artificial intelligence-based face transformation of medical seizure videos is feasible and may be useful for patient privacy protection. In our study, the cartoonization approach provided the best compromise between deidentification and preservation of seizure semiology.