Increasing utility of Google Trends in monitoring cardiovascular disease.

INTRODUCTION: Cardiovascular disease is the most common cause of morbidity and mortality in the United States. Patients are increasingly using internet search to find health-related information, including searches for cardiovascular diseases and risk factors. We sought to evaluate the change in the state by state correlation of cardiovascular disease and risk factors with Google Trends search volumes. METHODS: Data on cardiovascular disease hospitalizations and risk factor prevalence were obtained from the publically available Centers for Disease Control and Prevention website from 2006 to 2018. Google Trends data were obtained for matching conditions and time periods. Simple linear regression was performed to evaluate for an increase in correlation over time. RESULTS: Hospitalizations for six separate cardiovascular disease conditions showed moderate to strong correlation with online search data in the last period studied (heart failure (0.58, p < .001), atrial fibrillation (0.57, p < .001), coronary heart disease (0.58, p < .001), myocardial infarction (0.70, p < .001), stroke (0.62, p < .001), cardiac dysrhythmia (0.46, p < .001)) in the United States. All diseases studied showed a positive increase in correlation throughout the time period studied (p < .05). All five of the cardiovascular risk factors studied showed strong correlation with online search data; diabetes (R = 0.78, p < .001), cigarette use (R = 0.79, p < .001), hypertension (R = 0.81, p < .001), high cholesterol (R = 0.59, p < .001), and obesity (p = 0.80, p < .001) in the United States. Three of the five risk factors showed an increasing correlation over time. CONCLUSION: The prevalence of and hospitalizations for cardiovascular conditions in the United States strongly correlate with online search volumes in the United States when analyzed by state. This relationship has progressively strengthened or been strong and stable over recent years for these conditions. Google Trends represents an increasingly valuable tool for evaluating the burden of cardiovascular disease and risk factors in the United States.

Effectiveness of a Weight Loss Program Using Digital Health in Adolescents and Preadolescents.

Objective: To identify an efficacious intervention on treating adolescents with overweight and obesity, this might result in health benefits. Methods: Adolescents with overweight or obesity aged 10-17 years with BMI percentile ≥85th were included in this historical observational analysis. Subjects used an entirely remote weight loss program combining mobile applications, frequent self-weighing, and calorie restriction with meal replacement. Body weight changes were evaluated at 42, 60, 90, and 120 days using different metrics including absolute body weight, BMI, and BMI z-score. Chi-square or Fisher exact tests (categorical variables) and Student's t-test (continuous variables) were used to compare subjects. Results: In total, 2,825 participants, mean age 14.4 ± 2.2 years, (54.8% girls), were included from October 27, 2016, to December 31, 2017, in mainland China; 1355 (48.0%) had a baseline BMI percentile ≥97th. Mean BMI and BMI z-score were 29.20 ± 4.44 kg/m2 and 1.89 ± 0.42, respectively. At day 120, mean reduction in body weight, BMI, and BMI z-score was 8.6 ± 0.63 kg, 3.13 ± 0.21 kg/m2, and 0.42 ± 0.03; 71.4% had lost ≥5% body weight, 69.4% of boys and 73.2% of girls, respectively. Compared with boys, girls achieved greater reduction on BMI z-score at all intervals (p < 0.004 for all comparisons). Higher BMI percentile at baseline and increased frequency of use of the mobile application were directly associated with more significant weight loss. Conclusions: An entirely remote digital weight loss program is effective in facilitating weight loss in adolescents with overweight or obesity in the short term and mid term.

Google Trends Insights Into Reduced Acute Coronary Syndrome Admissions During the COVID-19 Pandemic: Infodemiology Study.

BACKGROUND: During the coronavirus disease (COVID-19) pandemic, a reduction in the presentation of acute coronary syndrome (ACS) has been noted in several countries. However, whether these trends reflect a reduction in ACS incidence or a decrease in emergency room visits is unknown. Using Google Trends, queries for chest pain that have previously been shown to closely correlate with coronary heart disease were compared with searches for myocardial infarction and COVID-19 symptoms. OBJECTIVE: The current study evaluates if search terms (or topics) pertaining to chest pain symptoms correlate with the reported decrease in presentations of ACS. METHODS: Google Trends data for search terms "chest pain," "myocardial infarction," "cough," and "fever" were obtained from June 1, 2019, to May 31, 2020. Related queries were evaluated for a relationship to coronary heart disease. RESULTS: Following the onset of the COVID-19 pandemic, chest pain searches increased in all countries studied by at least 34% (USA P=.003, Spain P=.007, UK P=.001, Italy P=.002), while searches for myocardial infarction dropped or remained unchanged. Rising searches for chest pain included "coronavirus chest pain," "home remedies for chest pain," and "natural remedies for chest pain." Searches on COVID-19 symptoms (eg, cough, fever) rose initially but returned to baseline while chest pain-related searches remained elevated throughout May. CONCLUSIONS: Search engine queries for chest pain have risen during the pandemic as have related searches with alternative attribution for chest pain or home care for chest pain, suggesting that recent drops in ACS presentations may be due to patients avoiding the emergency room and potential treatment in the midst of the COVID-19 pandemic.

Artificial Intelligence in Cardiology: Present and Future.

Artificial intelligence (AI) is a nontechnical, popular term that refers to machine learning of various types but most often to deep neural networks. Cardiology is at the forefront of AI in medicine. For this review, we searched PubMed and MEDLINE databases with no date restriction using search terms related to AI and cardiology. Articles were selected for inclusion on the basis of relevance. We highlight the major achievements in recent years in nearly all areas of cardiology and underscore the mounting evidence suggesting how AI will take center stage in the field. Artificial intelligence requires a close collaboration among computer scientists, clinical investigators, clinicians, and other users in order to identify the most relevant problems to be solved. Best practices in the generation and implementation of AI include the selection of ideal data sources, taking into account common challenges during the interpretation, validation, and generalizability of findings, and addressing safety and ethical concerns before final implementation. The future of AI in cardiology and in medicine in general is bright as the collaboration between investigators and clinicians continues to excel.

A Digital Health Weight Loss Program in 250,000 Individuals.

Importance: Obesity is a worsening epidemic worldwide. Effective and accessible weight loss programs to combat obesity on a large scale are warranted, but a need for frequent face-to-face care might impose a limitation. Objective: To evaluate whether individuals following a weight loss program based on a mobile application, wireless scale, and nutritional program but no face-to-face care can achieve clinically significant weight loss in a large cohort. Design: Retrospective observational analysis. Setting. China from October 2016 to December 2017. Participants. Mobile application users with a minimum of 2 weights (baseline and ≥35 days). Intervention. A commercial (Weijian Technologies) weight loss program consisting of a dietary replacement, self-monitoring using a wireless home scale, and frequent guidance via mobile application. Main Outcome. Mean weight change around 42, 60, 90, and 120 days after program initiation with subgroup analysis by gender, age, and frequency of use. Results: 251,718 individuals, with a mean age of 37.3 years (SD: 9.86) (79% female), were included with a mean weight loss of 4.3 kg (CI: ±0.02) and a mean follow-up of 120 days (SD: 76.8 days). Mean weight loss at 42, 60, 90, and 120 d was 4.1 kg (CI: ±0.02), 4.9 kg (CI: ±0.02), 5.6 kg (CI: ±0.03), and 5.4 kg (CI: ±0.04), respectively. At 120 d, 62.7% of participants had lost at least 5% of their initial weight. Both genders and all usage frequency tertiles showed statistically significant weight loss from baseline at each interval (P < 0.001), and this loss was greater in men than in women (120 d: 6.5 vs. 5.2 kg; P < 0.001). The frequency of recording (categorized as high-, medium-, or low-frequency users) was associated with greater weight loss when comparing high, medium, and low tertile use groups at all time intervals investigated (e.g., 120 d: -8.6, -5.6, and -2.2 kg, respectively; P < 0.001). Conclusions: People following a commercially available hybrid weight loss program using a mobile application, wireless scale, and nutritional program without face-to-face interaction on average achieved clinically significant short- and midterm weight loss. These results support the implementation of comparable technologies for weight control in a large population.

A digital health weight-loss intervention in severe obesity.

INTRODUCTION: Severe obesity is a growing epidemic that causes significant morbidity and mortality, and is particularly difficult to reverse. Efficacious and cost-effective interventions are needed to combat this epidemic. This study hypothesized that obese people (body mass index (BMI) ≥35 kg/m2) using a remote weight-loss program combining a mobile application, wireless scales, and low-calorie meal replacement would experience clinically significant weight loss. METHODS: This study was a retrospective observational analysis of 8275 individuals with a baseline BMI ≥35 kg/m2 who used a remote weight-loss program combining mobile applications, frequent self-weighing, and calorie restriction via meal replacement for a minimum of 35 days. Weight changes were evaluated at multiple intervals (42, 60, 90, and 120 days), and weight loss was evaluated for all and for pre-specified subgroups based on demographic features and frequency of self-weighing. RESULTS: Mean weight loss at 42 days (N = 6781) was 8.1 kg (margin of error (MOE) = 0.126 kg) with 73.6% of users experiencing >5% total body weight loss. Both men (9.1 kg; MOE = 0.172 kg; 7.9% from baseline) and women (7.1 kg; MOE = 0.179 kg; 7.2% from baseline) experienced significant weight loss. At the 120-day interval (N = 2914), mean weight loss was 14 kg (MOE = 0.340 kg), 13% total body weight loss from baseline, and 82.3% of participants had lost >5% of their initial body weight. The decrease in body-fat percent correlated well with weight loss (R = 0.92; p < 0.001). CONCLUSIONS: In a large cohort of individuals with class II or III obesity, a remote weight-loss program combining mobile applications, daily self-weighing, and calorie restriction via meal replacement resulted in dramatic weight loss among subjects who were active users when evaluated through a retrospective observational analysis.

Dose-Response Effect of a Digital Health Intervention During Cardiac Rehabilitation: Subanalysis of Randomized Controlled Trial.

BACKGROUND: Previous data have validated the benefit of digital health interventions (DHIs) on weight loss in patients following acute coronary syndrome entering cardiac rehabilitation (CR). OBJECTIVE: The primary purpose of this study was to test the hypothesis that increased DHI use, as measured by individual log-ins, is associated with improved weight loss. Secondary analyses evaluated the association between log-ins and activity within the platform and exercise, dietary, and medication adherence. METHODS: We obtained DHI data including active days, total log-ins, tasks completed, educational modules reviewed, medication adherence, and nonmonetary incentive points earned in patients undergoing standard CR following acute coronary syndrome. Linear regression followed by multivariable models were used to evaluate associations between DHI log-ins and weight loss or dietary adherence. RESULTS: Participants (n=61) were 79% male (48/61) with mean age of 61.0 (SD 9.7) years. We found a significant positive association of total log-ins during CR with weight loss (r2=.10, P=.03). Educational modules viewed (r2=.11, P=.009) and tasks completed (r2=.10, P=.01) were positively significantly associated with weight loss, yet total log-ins were not significantly associated with differences in dietary adherence (r2=.05, P=.12) or improvements in minutes of exercise per week (r2=.03, P=.36). CONCLUSIONS: These data extend our previous findings and demonstrate increased DHI log-ins portend improved weight loss in patients undergoing CR after acute coronary syndrome. DHI adherence can potentially be monitored and used as a tool to selectively encourage patients to adhere to secondary prevention lifestyle modifications. TRIAL REGISTRATION: ClinicalTrials.gov (NCT01883050); https://clinicaltrials.gov/ct2/show/NCT01883050.

Leveraging Machine Learning Techniques to Forecast Patient Prognosis After Percutaneous Coronary Intervention.

OBJECTIVES: This study sought to determine whether machine learning can be used to better identify patients at risk for death or congestive heart failure (CHF) rehospitalization after percutaneous coronary intervention (PCI). BACKGROUND: Contemporary risk models for event prediction after PCI have limited predictive ability. Machine learning has the potential to identify complex nonlinear patterns within datasets, improving the predictive power of models. METHODS: We evaluated 11,709 distinct patients who underwent 14,349 PCIs between January 2004 and December 2013 in the Mayo Clinic PCI registry. Fifty-two demographic and clinical parameters known at the time of admission were used to predict in-hospital mortality and 358 additional variables available at discharge were examined to identify patients at risk for CHF readmission. For each event, we trained a random forest regression model (i.e., machine learning) to estimate the time-to-event. Eight-fold cross-validation was used to estimate model performance. We used the predicted time-to-event as a score, generated a receiver-operating characteristic curve, and calculated the area under the curve (AUC). Model performance was then compared with a logistic regression model using pairwise comparisons of AUCs and calculation of net reclassification indices. RESULTS: The predictive algorithm identified a high-risk cohort representing 2% of all patients who had an in-hospital mortality of 45.5% (95% confidence interval: 43.5% to 47.5%) compared with a risk of 2.1% for the general population (AUC: 0.925; 95% confidence interval: 0.92 to 0.93). Advancing age, CHF, and shock on presentation were the leading predictors for the outcome. A high-risk group representing 1% of all patients was identified with 30-day CHF rehospitalization of 8.1% (95% confidence interval: 6.3% to 10.2%). Random forest regression outperformed logistic regression for predicting 30-day CHF readmission (AUC: 0.90 vs. 0.85; p = 0.003; net reclassification improvement: 5.14%) and 180-day cardiovascular death (AUC: 0.88 vs. 0.81; p = 0.02; net reclassification improvement: 0.02%). CONCLUSIONS: Random forest regression models (machine learning) were more predictive and discriminative than standard regression methods at identifying patients at risk for 180-day cardiovascular mortality and 30-day CHF rehospitalization, but not in-hospital mortality. Machine learning was effective at identifying subgroups at high risk for post-procedure mortality and readmission.

Association of Search Engine Queries for Chest Pain With Coronary Heart Disease Epidemiology.

Importance: Online search for symptoms is common and may be useful in early identification of patients experiencing coronary heart disease (CHD) and in epidemiologically studying the disease. Objective: To investigate the correlation of online symptom search for chest pain with disease prevalence of CHD. Design, Setting, and Participants: This retrospective study used Google Trends, a publicly available tool that provides relative search frequency for queried terms, to find searches for chest pain from January 2010 to June 2017 in the United States, the United Kingdom, and Australia. For the United States, results were obtained by state. These data were compared with publicly available prevalence data from the US Centers for Disease Control and Prevention of CHD hospitalizations by state for the same period. The same terms were used to evaluate seasonal and diurnal variation. Data were analyzed from July 2017 to October 2017. Main Outcomes and Measures: Correlation of search engine query for chest pain symptoms with temporal and geographic epidemiology. Results: State-by-state comparisons with reported CHD hospitalization were correlated (R = 0.81; P < .001). Significant monthly variation was appreciated in all countries studied, with the United States, United Kingdom, and Australia showing an 11% to 39% increase in search frequency in winter months compared with summer months. Diurnal variation showed a morning peak for search between local time 6 am and 8 am, with a greater than 100% increase seen in peak searching hours, which was consistent among the 3 countries studied. Conclusions and Relevance: Relative search frequency closely correlated with CHD epidemiology. This may have important implications for search engines as a resource for patients and a potential early-detection mechanism for physicians moving forward.

Natural history and predictors of mortality of patients with Takotsubo syndrome.

BACKGROUND: Takotsubo syndrome is a unique transient cardiomyopathy. The pathogenesis, management, and long-term prognosis of Takotsubo syndrome are incompletely understood. The study was designed to evaluate the natural history and determinants of outcomes in patients with Takotsubo syndrome patients. METHODS: We analyzed 265 patients in the Mayo Clinic Takotsubo syndrome registry for clinical presentation, treatment, and long-term outcomes with a focus on identifying prognostic factors for mortality and recurrence. RESULTS: 95% of patients were women with a mean age of 70 ±â€¯11.8 years. Among 257 patients discharged alive, there were 89 (34.6%) deaths, 18 (6.8%) non-fatal myocardial infarction, 12 (4.7%) cerebrovascular accidents and 23 (8.9%) re-hospitalization for heart failure over a mean follow-up of 5.8 ±â€¯3.6 years. Only 4 (5%) patients died from cardiac causes. Cancer was the single leading cause of death. Overall 1-year survival rate was 94.2%. Independent prognostic predictors of mortality were a history of cancer (HR 2.004, 1.334-3.012, p = 0.004), physical stress as precipitating factors (HR 1.882, 1.256-2.822, p = 0.012), history of depression (HR 1.622, 1.085-2.425, p = 0.009) and increased age (HR 1.059, 1.037-1.081, p < 0.001) after multivariate analysis. Beta-blockers and ACE inhibitors at discharge were not significant predictors. There were 24 (9.1%) recurrences during follow-up, but there were no significant differences in medical therapy compared to patients without recurrence. CONCLUSION: The high mortality rate is related to non-cardiac co-morbidities such as cancer. Additional determinants include physical stressors, increased age, and history of depression. Use of beta-blockers and ACE inhibitors did not affect development, prognosis or recurrence.

Digital health intervention as an adjunct to a workplace health program in hypertension.

BACKGROUND: Hypertension is a common and difficult-to-treat condition; digital health tools may serve as adjuncts to traditional pharmaceutical and lifestyle-based interventions. Using a retrospective observational study we sought to evaluate the effect of a desktop and mobile digital health intervention (DHI) as an adjunct to a workplace health program in those previously diagnosed with hypertension. METHODS: As part of a workplace health program, 3330 patients were identified as previously diagnosed with hypertension. A DHI was made available to participants providing motivational and educational materials assisting in the management of hypertension. We evaluated changes in blood pressure, weight, and body mass index (BMI) between users and nonusers based on login frequency to the DHI using multivariate regression through the five visits over the course of 1 year. RESULTS: One thousand six hundred twenty-two (49%) participants logged into the application at least once. DHI users had significant greater improvements in systolic blood pressure (SBP; -2.79 mm Hg), diastolic blood pressure (-2.12 mm Hg), and BMI (-0.23 kg/m2) at 1 year. Increased login frequency was significantly correlated with reductions in SBP, diastolic blood pressure, weight, and BMI (P ≤ .014). DISCUSSION: This large, observational study provides evidence that a DHI as an adjunct to a workplace health program is associated with greater improvement in blood pressure and BMI at 1 year. This study adds to the growing body of evidence that DHIs may be useful in augmenting the treatment of hypertension in addition to traditional management with pharmaceuticals and lifestyle changes.

Usage of a Digital Health Workplace Intervention Based on Socioeconomic Environment and Race: Retrospective Secondary Cross-Sectional Study.

BACKGROUND: Digital health tools have been associated with improvement of cardiovascular disease (CVD) risk factors and outcomes; however, the differential use of these technologies among various ethnic and economic classes is not well known. OBJECTIVE: To identify the effect of socioeconomic environment on usage of a digital health intervention. METHODS: A retrospective secondary cross-sectional analysis of a workplace digital health tool use, in association with a change in intermediate markers of CVD, was undertaken over the course of one year in 26,188 participants in a work health program across 81 organizations in 42 American states between 2011 and 2014. Baseline demographic data for participants included age, sex, race, home zip code, weight, height, blood pressure, glucose, lipids, and hemoglobin A1c. Follow-up data was then obtained in 90-day increments for up to one year. Using publicly available data from the American Community Survey, we obtained the median income for each zip code as a marker for socioeconomic status via median household income. Digital health intervention usage was analyzed based on socioeconomic status as well as age, gender, and race. RESULTS: The cohort was found to represent a wide sample of socioeconomic environments from a median income of US $11,000 to $171,000. As a whole, doubling of income was associated with 7.6% increase in log-in frequency. However, there were marked differences between races. Black participants showed a 40.5% increase and Hispanic participants showed a 57.8% increase in use with a doubling of income, compared to 3% for Caucasian participants. CONCLUSIONS: The current study demonstrated that socioeconomic data confirms no relevant relationship between socioeconomic environment and digital health intervention usage for Caucasian users. However, a strong relationship is present for black and Hispanic users. Thus, socioeconomic environment plays a prominent role only in minority groups that represent a high-risk group for CVD. This finding identifies a need for digital health apps that are effective in these high-risk groups.