Data Science as a Core Competency in Undergraduate Medical Education in the Age of Artificial Intelligence in Health Care.

The increasingly sophisticated and rapidly evolving application of artificial intelligence in medicine is transforming how health care is delivered, highlighting a need for current and future physicians to develop basic competency in the data science that underlies this topic. Medical educators must consider how to incorporate central concepts in data science into their core curricula to train physicians of the future. Similar to how the advent of diagnostic imaging required the physician to understand, interpret, and explain the relevant results to patients, physicians of the future should be able to explain to patients the benefits and limitations of management plans guided by artificial intelligence. We outline major content domains and associated learning outcomes in data science applicable to medical student curricula, suggest ways to incorporate these themes into existing curricula, and note potential implementation barriers and solutions to optimize the integration of this content.

Grains of Sand to Clinical Pearls: Realizing the Potential of Wearable Data.

Despite the rapid growth of wearables as a consumer technology sector and a growing evidence base supporting their use, they have been slow to be adopted by the health system into clinical care. As regulatory, reimbursement, and technical barriers recede, a persistent challenge remains how to make wearable data actionable for clinicians-transforming disconnected grains of wearable data into meaningful clinical "pearls". In order to bridge this adoption gap, wearable data must become visible, interpretable, and actionable for the clinician. We showcase emerging trends and best practices that illustrate these 3 pillars, and offer some recommendations on how the ecosystem can move forward.

Recent Advances in Systems and Network Medicine: Meeting Report from the First International Conference in Systems and Network Medicine.

The First International Conference in Systems and Network Medicine gathered together 200 global thought leaders, scientists, clinicians, academicians, industry and government experts, medical and graduate students, postdoctoral scholars and policymakers. Held at Georgetown University Conference Center in Washington D.C. on September 11-13, 2019, the event featured a day of pre-conference lectures and hands-on bioinformatic computational workshops followed by two days of deep and diverse scientific talks, panel discussions with eminent thought leaders, and scientific poster presentations. Topics ranged from: Systems and Network Medicine in Clinical Practice; the role of -omics technologies in Health Care; the role of Education and Ethics in Clinical Practice, Systems Thinking, and Rare Diseases; and the role of Artificial Intelligence in Medicine. The conference served as a unique nexus for interdisciplinary discovery and dialogue and fostered formation of new insights and possibilities for health care systems advances.

A randomized controlled trial of online symptom searching to inform patient generated differential diagnoses.

Patient online health searching is now commonplace, however, the accuracy of patient generated differentials for new symptoms and potential for patient anxiety are concerns. We aimed primarily to determine the accuracy of patient generated differentials for new symptoms with and without online searching, and secondarily, to evaluate the impact of searching on anxiety levels. In the waiting room prior to seeing a clinician, 300 patients with new symptoms were randomly assigned 1:1:1 to Google searching with health related features including a symptom search tool vs Google searching with health related features disabled vs no searching. Participants were 18 years or older and presenting to the emergency department of an urban academic medical center with new low-acuity symptoms that were not due to exacerbation of a chronic condition. Search groups received access on a tablet/smartphone to Google searching with or without health related features. Both search groups could access any websites; health related features led the patient to common diagnoses and physician-validated information. The primary outcome was accuracy of the patient generated differential assessed by matching at least two of the top three diagnoses on the clinician's differential. A secondary outcome was anxiety by a visual analogue scale. Patients were a median of 33.1 (IQI 26.2-45.9) years old, 60% women, 63% black, 82% had a high school education or less, and 45.7% reported having performed an online search prior to presentation. Search group patients spent a median of 3.82 (2.53-5.72) minutes searching online. Similar proportions of patients in each group matched at least two of three clinician diagnoses: 27.0% and 28.3% for Google searching with and without health related features vs 23.8% in the no search group. Patients in the search groups had a similar odds of matching ≥2/3 diagnoses as the no search group [OR (95% CI): 1.23 (0.70-2.13), p = 0.47]. Anxiety was unchanged with online searching. In conclusion, brief online searching in the waiting room did not improve accuracy of patient generated differential diagnoses for new symptoms. The absence of an increase in patient anxiety provides reassurance for subsequent work to refine and investigate online symptom search tools.

Online Health Searches and Their Perceived Effects on Patients and Patient-Clinician Relationships: ASystematic Review.

Online health searches are common and may be impacting patients and their relationships with their clinicians in ways that are not fully understood. We searched PubMed, Embase, Cochrane Reviews, Cochrane Trials, Scopus, and CINAHL from January 1, 1990 to January 29, 2016 for studies in which patients searched online for any aspect of health care and then visited their clinician. We extracted data pertaining to either patients' or clinicians' perceptions of the effects of these online searches on patients and the patient-clinician relationship. Searches seemed to induce patient anxiety but more often led to patient reassurance, clinical understanding, and empowerment. Patients tended to perceive that online health searches had a positive effect on the patient-clinician relationship, although the nature of the effect could depend on the clinician's response to patient queries about the information. Clinicians generally perceived neutral effects on patients and the patient-clinician relationship and commonly raised concerns about the accuracy of online content. Significant methodologic heterogeneity prevented quantitative synthesis. Accuracy of online health search content was not assessed, and randomized controlled trials were notably lacking.

Individual depressive symptoms and all-cause mortality In 6673 patients with myocardial infarction: Heterogeneity across age and sex subgroups.

BACKGROUND: Depression predicts poor prognosis in patients with myocardial infarction (MI). However, individual depressive symptoms may have different prognostic value, and age and sex could be important effect modifiers. This study compared the prognostic value of individual depressive symptoms across age and sex subgroups in post-MI patients. METHODS: Individual patient-data were compiled for 6673 post-MI patients from seven studies. Depressive symptoms were measured with 10 items of the Beck Depression Inventory (BDI10). The endpoint was all-cause mortality (mean=3.8 years). Multilevel multivariable Cox regression analysis was used to estimate the mortality risk across age groups (≤55, 56-69 and ≥70 years) and sex for symptoms that potentially interacted with age and sex. RESULTS: At follow-up, 995 (15%) post-MI patients had died. BDI10 depression scores were associated with an increased mortality risk (HR:1.20;95%CI:1.11-1.28,p<.001). Negative self-image (HR:1.53;1.06-2.21;p=.022) and indecisiveness (HR:1.53;1.15-2.04;p=.003) were associated with increased mortality in men <55. Dissatisfaction was associated with increased mortality in men aged 56-69 (HR:1.35;1.07-1.71;p=. 011), and dissatisfaction (HR:1.34;1.10-1.63;p=.003) and fatigue (HR:1.45;1.20-1.74;p<.001) in men >70. Fatigue was associated with mortality in women aged 56-69 (HR:1.54;1.09-2.15;p=.012), and suicidal ideation in women aged >70 (HR:1.58;1.03-2.43;p=.037). Left-ventricular ejection fraction (LVEF) accounted for much of the associations in men ≤55 years and women ≥70 years. LIMITATIONS: Findings are sample-specific and need replication in future research; BDI10 items were derived from the original BDI assessment. CONCLUSIONS: There is large heterogeneity in the prognostic value of individual depressive symptoms in post-MI patients across sex and age subgroups. LVEF partially explained the depression-prognosis association in specific subgroups.

A bifactor model of the Beck Depression Inventory and its association with medical prognosis after myocardial infarction.

OBJECTIVES: Evidence suggests that depression is associated with adverse outcomes in patients with myocardial infarction (MI). Some of the symptoms of depression may also be symptoms of somatic illness and these may confound the association between depression and prognosis. We investigated whether depression following MI is associated with medical prognosis independent of these somatic symptoms. METHOD: The database of an individual patient data meta-analysis was used. Endpoints were all-cause mortality and cardiovascular events. Nine studies were included. Bifactor factor analysis included 13,100 participants and 7,595 participants were included in survival models. Dimensions were generated from the Beck Depression Inventory using factor analyses. The prognostic association was assessed using mixed-effects Cox regression analysis. RESULTS: A bifactor model, consisting of a general factor and 2 general depression-free subgroup factors (a somatic/affective and a cognitive/affective), provided the best fit. There was a significant association between the general depression factor and all-cause mortality (hazard ratio [HR] = 1.25; 95% confidence interval [CI] [1.17, 1.34], p < .001) and cardiovascular events (HR = 1.18; 95% CI [1.13, 1.23], p < .001). After adjustment for demographics, measures of cardiac disease severity, and health-related variables, the association between the general depression factor and all-cause mortality (HR = 1.14; 95% CI [1.04, 1.25], p = .003) and cardiovascular events (HR = 1.16; 95% CI [1.10, 1.23], p = .014) attenuated. Additionally, the general depression-free somatic/affective factor was significantly associated with the endpoints, while the general depression-free cognitive/affective was not. CONCLUSIONS: A general depression factor is associated with adverse medical prognosis following MI independent of somatic/affective symptoms that may be partly attributable to somatic illness. (PsycINFO Database Record

Using Innovative Methodologies From Technology and Manufacturing Companies to Reduce Heart Failure Readmissions.

Heart failure (HF) patients have high 30-day readmission rates with high costs and poor quality of life. This study investigated the impact of a framework blending Lean Sigma, design thinking, and Lean Startup on 30-day all-cause readmissions among HF patients. This was a prospective study in an academic hospital in Baltimore, Maryland. Thirty-day all-cause readmission was assessed using the hospital's electronic medical record. The baseline readmission rate for HF was 28.4% in 2010 with 690 discharges. The framework was developed and interventions implemented in the second half of 2011. The impact of the interventions was evaluated through 2012. The rate declined to 18.9% among 703 discharges (P < .01). There was no significant change for non-HF readmissions. This study concluded that methodologies from technology and manufacturing companies can reduce 30-day readmissions in HF, demonstrating the potential of this innovations framework to improve chronic disease care.

A finger photoplethysmography waveform during the valsalva maneuver detects changes in left heart filling pressure after hemodialysis.

BACKGROUND: A noninvasive system for determining left ventricular (LV) filling pressure may help to improve personalized fluid removal goals in hemodialysis patients. We previously showed that the change in photoplethysmography (PPG) pulse amplitude measured by finger PPG during a Valsalva maneuver correlates with invasively measured left ventricular end-diastolic pressure (LVEDP). This key PPG change, the ratio of finger PPG pulse amplitude at end-Valsalva to baseline, is known as the Pulse Amplitude Ratio, PAR. The objective of this study was to determine how PAR changes after fluid removal in hemodialysis. METHODS: We tested subjects with end-stage renal disease, before and after hemodialysis. Each subject performed a Valsalva maneuver with an effort of 20 mmHg for 10 s, guided by the device display. Finger PPG was recorded continuously before and during the maneuver. PAR was calculated automatically. RESULTS: Twenty-seven subjects (21 Males) ages 25-75 years were tested. Access sites were AV-fistulas of the arm predominantly. Weight decreased from 99.7 ± 36.9 kg to 97.0 ± 6.0 kg (p < 0.0003) with an average fluid removal of 3.07 ± 1.08 l. Correspondingly, PAR decreased from 0.74 ± 0.24 to 0.62 ± 0.23 (p = 0.003). The change in PAR was correlated with baseline PAR (r = 0.48, p = 0.01). CONCLUSION: An index of left heart filling pressure obtained noninvasively using finger photoplethysmography during the Valsalva maneuver is sensitive enough to detect reductions in filling pressure after fluid removal with hemodialysis. Further studies are warranted to determine if this method can be used to guide fluid removal during hemodialysis.

Systematic Review and Individual Patient Data Meta-Analysis of Sex Differences in Depression and Prognosis in Persons With Myocardial Infarction: A MINDMAPS Study.

OBJECTIVE: Using combined individual patient data from prospective studies, we explored sex differences in depression and prognosis post-myocardial infarction (MI) and determined whether disease indices could account for found differences. METHODS: Individual patient data analysis of 10,175 MI patients who completed diagnostic interviews or depression questionnaires from 16 prospective studies from the MINDMAPS study was conducted. Multilevel logistic and Cox regression models were used to determine sex differences in prevalence of depression and sex-specific effects of depression on subsequent outcomes. RESULTS: Combined interview and questionnaire data from observational studies showed that 36% (635/1760) of women and 29% (1575/5526) of men reported elevated levels of depression (age-adjusted odds ratio = 0.68, 95% confidence interval [CI] = 0.60-0.77). The risk for all-cause mortality associated with depression was higher in men (hazard ratio = 1.38, 95% CI = 1.30-1.47) than in women (hazard ratio = 1.22, 95% CI = 1.14-1.31; sex by depression interaction: p < .001). Low left ventricular ejection fraction (LVEF) was associated with higher depression scores in men only (sex by LVEF interaction: B = 0.294, 95% CI = 0.090-0.498), which attenuated the sex difference in the association between depression and prognosis. CONCLUSIONS: The prevalence of depression post-MI was higher in women than in men, but the association between depression and cardiac prognosis was worse for men. LVEF was associated with depression in men only and accounted for the increased risk of all-cause mortality in depressed men versus women, suggesting that depression in men post-MI may, in part, reflect cardiovascular disease severity.

The diuresis clinic: a new paradigm for the treatment of mild decompensated heart failure.

BACKGROUND: Heart failure results in approximately 1 million hospital admissions annually in the United States and is the leading cause of 30-day readmissions. METHODS: This study explores the impact of a diuresis clinic on heart failure outcomes and cost. Data were collected prospectively on all consecutive patients who received intravenous diuretics and multidisciplinary care in the clinic from its establishment from October 2011 to December 2012, as well as a comparison cohort of patients with heart failure who were admitted to the hospital for <48 hours. The percentage of hospitalized days was calculated for both cohorts 180 days before and 180 days after each patient's index visit. RESULTS: In the diuresis clinic group, 106 patients (mean age, 68.2 ± 13 years; 48% were women) were treated over 328 visits (1-22 visits per person), with a mean intravenous furosemide dose of 100 mg, average urine output of 1460 ± 730 mL, and weight loss of 2.3 ± 1.8 kg. Days hospitalized decreased from 38.3 to 31.2 per 1000 patient-days after the index diuresis clinic visit (P < .01). In the comparison group, 143 patients (mean age, 69 ± 16 years; 54% were women) were admitted for <48 hours. Days hospitalized increased from 14.4 to 21.0 per 1000 patient-days after index admission (P < .01). On multivariate analysis, the diuresis clinic was associated with 3 fewer days in the hospital per 180 days per patient, with an estimated annual savings of $12,113 per patient. CONCLUSIONS: Compared with a brief hospital stay, treatment of mild decompensated heart failure in a diuresis clinic resulted in a substantial and cost-effective decline in the rate of subsequent hospitalization.

Sex dependent risk factors for mortality after myocardial infarction: individual patient data meta-analysis.

BACKGROUND: Although a number of risk factors are known to predict mortality within the first years after myocardial infarction, little is known about interactions between risk factors, whereas these could contribute to accurate differentiation of patients with higher and lower risk for mortality. This study explored the effect of interactions of risk factors on all-cause mortality in patients with myocardial infarction based on individual patient data meta-analysis. METHODS: Prospective data for 10,512 patients hospitalized for myocardial infarction were derived from 16 observational studies (MINDMAPS). Baseline measures included a broad set of risk factors for mortality such as age, sex, heart failure, diabetes, depression, and smoking. All two-way and three-way interactions of these risk factors were included in Lasso regression analyses to predict time-to-event related all-cause mortality. The effect of selected interactions was investigated with multilevel Cox regression models. RESULTS: Lasso regression selected five two-way interactions, of which four included sex. The addition of these interactions to multilevel Cox models suggested differential risk patterns for males and females. Younger women (age<50) had a higher risk for all-cause mortality than men in the same age group (HR 0.7 vs. 0.4), while men had a higher risk than women if they had depression (HR 1.4 vs. 1.1) or a low left ventricular ejection fraction (HR 1.7 vs. 1.3). Predictive accuracy of the Cox model was better for men than for women (area under the curves: 0.770 vs. 0.754). CONCLUSIONS: Interactions of well-known risk factors for all-cause mortality after myocardial infarction suggested important sex differences. This study gives rise to a further exploration of prediction models to improve risk assessment for men and women after myocardial infarction.

Frontiers of therapy for patients with heart failure.

This review broadly covers advances in heart failure, which is responsible for significant morbidity, mortality, and cost in the United States. It is a heterogeneous condition, and accurate classification helps ensure appropriate application of evidence-based therapies. Hemodynamics are important in acute heart failure syndromes and may help tailor therapy. Neurohormonal modulation forms the cornerstone of chronic systolic heart failure treatment but does not affect outcomes in diastolic heart failure where management goals emphasize optimization of central volume, blood pressure, and atrial rhythm, as well as the treatment of comorbidities. Frontiers of heart failure therapy range from advances in pharmacology (novel inotropic agents and neurohormonal modulators), to cell biology (nucleic acid-based drugs and cell therapy) to biomedical engineering (devices such as ultrafiltration, biventricular pacemakers, implantable cardiac defibrillators, remote monitoring systems, and left ventricular assist devices), and to health systems (risk stratification and integrated care of comorbidities). The ultimate frontier will be to integrate these data effectively to ensure that patients with heart failure consistently receive the best evidenced-based care possible.

The variable natural history of idiopathic dilated cardiomyopathy.

BACKGROUND: Determining the prognosis of patients with heart failure is essential for patient management and clinical trial conduct. The relative value of traditional prognostic criteria remains unclear and the assessment of long-term prognosis for individual patients is problematic. OBJECTIVES: To determine the ability of clinical, hemodynamic and echocardiographic parameters to predict the long-term prognosis of patients with idiopathic dilated cardiomyopathy. METHODS: We investigated the ability of clinical, hemodynamic and echocardiographic parameters to predict the long-term prognosis of individual patients in a large, representative, contemporary cohort of idiopathic dilated cardiomyopathy (IDCM) patients referred to Johns Hopkins from 1997 to 2004 for evaluation of cardiomyopathy. In all patients a baseline history was taken, and physical examination, laboratory studies, echocardiogram, right heart catheterization and endomyocardial biopsy were performed. RESULTS: In 171 IDCM patients followed for a median 3.5 years, there were 50 long-term event-free survivors (LTS) (median survival 6.4 years) and 34 patients died or underwent ventricular assist device placement or transplantation within 5 years (NLTS; non-long-term survivors) (median time to event 1.83 years. Established risk factors (gender, race, presence of diabetes, serum creatinine, sodium) and the use of accepted heart failure medications (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, beta blockers) were similar between the two groups. Although LTS had younger age, higher ejection fraction (EF) and lower New York Heart Association (NYHA) class at presentation, the positive predictive value of an EF < 25% was 64% (95% CI 41%-79%) and that of NYHA class > 2 was 53% (95% CI 36-69%). A logistic model incorporating these three variables incorrectly classified 29% of patients. CONCLUSIONS: IDCM exhibits a highly variable natural history and standard clinical predictors have limited ability to classify IDCM patients into broad prognostic categories. These findings suggest that there are important host-environmental factors still unappreciated in the biology of IDCM.

Depressive symptoms, physical inactivity and risk of cardiovascular mortality in older adults: the Cardiovascular Health Study.

BACKGROUND: Depressed older individuals have a higher mortality than older persons without depression. Depression is associated with physical inactivity, and low levels of physical activity have been shown in some cohorts to be a partial mediator of the relationship between depression and cardiovascular events and mortality. METHODS: A cohort of 5888 individuals (mean 72.8 ± 5.6 years, 58% female, 16% African-American) from four US communities was followed for an average of 10.3 years. Self-reported depressive symptoms (10-item Center for Epidemiological Studies Depression Scale) were assessed annually and self-reported physical activity was assessed at baseline and at 3 and 7 years. To estimate how much of the increased risk of cardiovascular mortality associated with depressive symptoms was due to physical inactivity, Cox regression with time-varying covariates was used to determine the percentage change in the log HR of depressive symptoms for cardiovascular mortality after adding physical activity variables. RESULTS: At baseline, 20% of participants scored above the cut-off for depressive symptoms. There were 2915 deaths (49.8%), of which 1176 (20.1%) were from cardiovascular causes. Depressive symptoms and physical inactivity each independently increased the risk of cardiovascular mortality and were strongly associated with each other (all p < 0.001). Individuals with both depressive symptoms and physical inactivity had greater cardiovascular mortality than those with either individually (p < 0.001, log rank test). Physical inactivity reduced the log HR of depressive symptoms for cardiovascular mortality by 26% after adjustment. This was similar for persons with (25%) and without (23%) established coronary heart disease. CONCLUSIONS: Physical inactivity accounted for a significant proportion of the risk of cardiovascular mortality due to depressive symptoms in older adults, regardless of coronary heart disease status.

Self-assessed physical health predicts 10-year mortality after myocardial infarction.

PURPOSE: In spite of their widespread use in other fields, global measures of health are not commonly used in determining the prognosis of patients with myocardial infarction (MI). The objective of the present study was to ascertain the relationship between self-assessed physical health at the time of the MI and long-term mortality. METHODS: This was a prospective cohort study of 284 patients with MI admitted to an academic community hospital between July 1995 and December 1996 who completed the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36). The physical component scale from the SF-36 was used as a self-assessment of physical health. All-cause mortality was assessed 10 years later by using the Social Security Death Index. RESULTS: Patients with lower self-reported physical health were significantly more likely to be women; older; depressed; have a history of coronary artery disease; have a family history of MI; have a non-Q wave MI; have a Killip class 3 or 4 MI; have hypertension, diabetes mellitus, renal insufficiency, and chronic obstructive pulmonary disease; and have a longer hospitalization period. Patients with higher physical component scores had significantly lower mortality in the 10 years after MI and this persisted after adjusting for confounders (hazard ratio = 0.97 [95% CI 0.96-0.99], P = .001). CONCLUSIONS: These data suggest that self-assessed physical health provides information on the long-term prognosis of patients with MI above and beyond that provided by traditional risk predictors.