Examining the Correlation of Google Influenza Trend with Hospital Data: Retrospective Study

Predicting incidence of hepatitis E for thirteen cities in Jiangsu Province, China

Hepatitis E has placed a heavy burden on China, especially in Jiangsu Province, so accurately predicting the incidence of hepatitis E benefits to alleviate the medical burden. In this paper, we propose a new attentive bidirectional long short-term memory network (denoted as BiLSTM-Attention) to predict the incidence of hepatitis E for all 13 cities in Jiangsu Province, China. Besides, we also explore the performance of adding meteorological factors and the Baidu (the most widely used Chinese search engine) index as additional training data for the prediction of our BiLSTM-Attention model. SARIMAX, GBDT, LSTM, BiLSTM, and BiLSTM-Attention models are tested in this study, based on the monthly incidence rates of hepatitis E, meteorological factors, and the Baidu index collected from 2011 to 2019 for the 13 cities in Jiangsu province, China. From January 2011 to December 2019, a total of 29,339 cases of hepatitis E were detected in all cities in Jiangsu Province, and the average monthly incidence rate for each city is 0.359 per 100,000 persons. Root mean square error (RMSE) and mean absolute error (MAE) are used for model selection and performance evaluation. The BiLSTM-Attention model considering meteorological factors and the Baidu index has the best performance for hepatitis E prediction in all cities, and it gets at least 10% improvement in RMSE and MAE for all 13 cities in Jiangsu province, which means the model has significantly improved the learning ability, generalizability, and prediction accuracy when comparing with others.

Confounding Effect of Undergraduate Semester-Driven "Academic" Internet Searches on the Ability to Detect True Disease Seasonality in Google Trends Data: Fourier Filter Method Development and Demonstration

Background

Internet search volume for medical information, as tracked by Google Trends, has been used to demonstrate unexpected seasonality in the symptom burden of a variety of medical conditions. However, when more technical medical language is used (eg, diagnoses), we believe that this technique is confounded by the cyclic, school year-driven internet search patterns of health care students.

Objective

This study aimed to (1) demonstrate that artificial "academic cycling" of Google Trends' search volume is present in many health care terms, (2) demonstrate how signal processing techniques can be used to filter academic cycling out of Google Trends data, and (3) apply this filtering technique to some clinically relevant examples.

Methods

We obtained the Google Trends search volume data for a variety of academic terms demonstrating strong academic cycling and used a Fourier analysis technique to (1) identify the frequency domain fingerprint of this modulating pattern in one particularly strong example, and (2) filter that pattern out of the original data. After this illustrative example, we then applied the same filtering technique to internet searches for information on 3 medical conditions believed to have true seasonal modulation (myocardial infarction, hypertension, and depression), and all bacterial genus terms within a common medical microbiology textbook.

Results

Academic cycling explains much of the seasonal variation in internet search volume for many technically oriented search terms, including the bacterial genus term ["Staphylococcus"], for which academic cycling explained 73.8% of the variability in search volume (using the squared Spearman rank correlation coefficient, P<.001). Of the 56 bacterial genus terms examined, 6 displayed sufficiently strong seasonality to warrant further examination post filtering. This included (1) ["Aeromonas" + "Plesiomonas"] (nosocomial infections that were searched for more frequently during the summer), (2) ["Ehrlichia"] (a tick-borne pathogen that was searched for more frequently during late spring), (3) ["Moraxella"] and ["Haemophilus"] (respiratory infections that were searched for more frequently during late winter), (4) ["Legionella"] (searched for more frequently during midsummer), and (5) ["Vibrio"] (which spiked for 2 months during midsummer). The terms ["myocardial infarction"] and ["hypertension"] lacked any obvious seasonal cycling after filtering, whereas ["depression"] maintained an annual cycling pattern.

Conclusions

Although it is reasonable to search for seasonal modulation of medical conditions using Google Trends' internet search volume and lay-appropriate search terms, the variation in more technical search terms may be driven by health care students whose search frequency varies with the academic school year. When this is the case, using Fourier analysis to filter out academic cycling is a potential means to establish whether additional seasonality is present.