Is Digital Epidemiology the Future of Clinical Epidemiology?

Trends in the Baidu Index in Search Activity Related to Mpox at Geographical and Economic Levels and Associated Factors in China: National Longitudinal Analysis

BACKGROUND

Research assessing trends in online search activity related to mpox in China is scarce.

OBJECTIVE

We aimed to provide evidence for an overview of online information searching during an infectious disease outbreak by analyzing trends in online search activity related to mpox at geographical and economic levels in China and explore influencing factors.

METHODS

We used the Baidu index to present online search activity related to mpox from May 19 to September 19, 2022. Segmented interrupted time-series analysis was used to estimate trends in online search activity. Factors influencing these trends were analyzed using a general linear regression (GLM) model. We calculated the concentration index to measure economic-related inequality in online search activity and related trends.

RESULTS

Online search activity was highest on the day the first imported case of mpox appeared in Chongqing compared to 3 other cutoff time points. After the day of the first imported mpox case in Taiwan, the declaration of a public health emergency of international concern, the first imported mpox case in Hong Kong, and the first imported mpox case in Chongqing, national online search activity increased by 0.642%, 1.035%, 1.199%, and 2.023%, respectively. The eastern regions had higher increases than the central and western regions. Across 31 provinces, municipalities, and autonomous regions, the top 3 areas with higher increases were Beijing, Shanghai, and Tianjin at 3 time points, with the exception of the day of the first imported mpox case in Chongqing (Chongqing replaced Tianjin on that day). When AIDS incidence increased by 1 per 100,000 people, there was an increase after the day of the first imported mpox case in Chongqing of 36.22% (95% CI 3.29%-69.15%; P=.04) after controlling for other covariates. Online search activity (concentration index=0.18; P<.001) was more concentrated among populations with a higher economic status. Unlike the central area, the eastern (concentration index=0.234; P<.001) and western areas (concentration index=0.047; P=.04) had significant economic-related disparities (P for difference <.001) in online search activity. The overall concentration index of changes in online search activity became lower over time.

CONCLUSIONS

Regions with a higher economic level showed more interest in mpox, especially Beijing and Shanghai. After the day of the first imported mpox case in Chongqing, changes in online search activity were affected by AIDS incidence rate. Economic-related disparities in changes in online search activity became lower over time. It would be desirable to construct a reliable information source in regions with a higher economic level and higher AIDS incidence rate and promote public knowledge in regions with a lower economic level in China, especially after important public events.

Trends in Online Search Activity and the Correlation with Daily New Cases of Monkeypox among 102 Countries or Territories

Research assessing the trend in online search activity on monkeypox (mpox) and the correlation with the mpox epidemic at the global and national level is scarce. The trend of online search activity and the time-lag correlations between it and daily new mpox cases were estimated by using segmented interrupted time-series analysis and Spearman correlation coefficient (rs), respectively. We found that after the declaration of a Public Health Emergency of International Concern (PHEIC), the proportion of countries or territories with increasing changes in online search activity was lowest in Africa (8.16%, 4/49), and a downward trend in online search activity was highest in North America (8/31, 25.81%). The time-lag effect of global online search activity on daily new cases was significant (rs = 0.24). There were eight countries or territories with significant time-lag effect; the top three countries or territories were Brazil (rs = 0.46), United States (rs = 0.24), and Canada (rs = 0.24). Interest behavior in mpox was insufficient, even after the declaration of PHEIC, especially in Africa and North America. Online search activity could be used as an early indicator of the outbreak of mpox at the global level and in epidemic countries.

The Global Impact of COVID-19 on Threat Appraisals

We planned an infodemiological analysis to estimate the global impact of coronavirus disease 2019 (COVID-19) on threat appraisals. We accessed Google Trends using the search terms “Anxiety”, “Distress”, “Fear”, “Rumination”, “Stress” and “Worry” within the “topic” domain, setting the geographical location to “worldwide”, between July 2017 and July 2022. The weekly Google Trends score for the six search terms, thus, mirroring Web popularity and probable prevalence, was compared between the two search periods, “pre-COVID” (between July 2017 and February 2020) and COVID (between March 2020 and July 2022), thus, reflecting the volume of searches before and during the ongoing COVID-19 pandemic. The median weekly Google Trends score of all these search terms significantly increased during the COVID-19 pandemic, i.e., anxiety by 22%, distress by 13%, fear by 9%, rumination by 18%, stress by 13% and worry by 20%. With variable strength, the weekly Google Trends scores of each search term were found to be significantly associated (all p < 0.001). We can, hence, conclude that the enhanced burden of threat appraisals observed after SARS-CoV-2 spread leads the way to establish preventive, diagnostic and therapeutic measures in order to limit the unfavorable mental health consequences caused by the ongoing COVID-19 pandemic.

Exploiting real-world data to monitor physical activity in patients with osteoarthritis: the opportunity of digital epidemiology

Osteoarthritis is a degenerative joint disease that affects millions of people worldwide. Current guidelines emphasize the importance of regular physical activity as a preventive measure against disease progression and as a valuable strategy for pain and functionality management. Despite this, most patients with osteoarthritis are inactive. Modern technological advances have led to the implementation of digital devices, such as wearables and smartphones, showing new opportunities for healthcare professionals and researchers to monitor physical activity and therefore engage patients in daily exercising. Additionally, digital devices have emerged as a promising tool for improving frequent health data collection, disease monitoring, and supporting public health surveillance. The leveraging of digital data has laid the foundation for developing a new concept of epidemiological study, known as "Digital Epidemiology". Analyzing real-world data can change the way we observe human behavior and suggest health interventions, as in the case of physical exercise and osteoarthritic patients. Furthermore, large-scale data could contribute to personalized and precision medicine in the future. Herein, an overview of recent clinical applications of wearables for monitoring physical activity in patients with osteoarthritis and the benefits of exploiting real-world data in the context of digital epidemiology are discussed.

Forecasting the future number of pertussis cases using data from Google Trends

Background

Alternative methods could be used to enhance the monitoring and forecasting of re-emerging conditions such as pertussis. Here, whether data on the volume of Internet searching on pertussis could complement traditional modeling based solely on reported case numbers was assessed.

Methods

SARIMA models were fitted to describe reported weekly pertussis case numbers over a four-year period in Germany. Pertussis-related Google Trends data (GTD) was added as an external regressor. Predictions were made by the models, both with and without GTD, and compared with values within the validation dataset over a one-year and for a two-weeks period.

Results

Predictions of the traditional model using solely reported case numbers resulted in an RMSE (residual mean squared error) of 192.65 and 207.8, a mean absolute percentage error (MAPE) of 58.59 and 72.1, and a mean absolute error (MAE) 169.53 and 190.53 for the one-year and for the two-weeks period, respectively. The GTD expanded model achieved better forecasting accuracy (RMSE: 144.22 and 201.78), a MAPE 43.86, and 68.54 and a MAE of 124.46 and 178.96. Corrected Akaike Information Criteria also favored the GTD expanded model (1750.98 vs. 1746.73). The difference between the predictive performances was significant when using a two-sided Diebold-Mariano test (DM value: 6.86, p < 0.001) for the one-year period.

Conclusion

Internet-based surveillance data enhanced the predictive ability of a traditionally based model and should be considered as a method to enhance future disease modeling.

•

Pertussis-related Google Trends Data (GTD) showed a weak but significant correlation with the reported weekly number of pertussis cases.

•

We fitted a SARIMA models to estimate reported weekly pertussis case numbers

•

The GTD-expanded models achieved significantly better predictive accuracy than the traditional model over a one-year-period.

•

Corrected Akaike Information Criteria also favored the GTD-Expanded SARIMA model.

•

The use of GTD should be considered as a method to enhance pertussis forecasting.

Public Interest in Refractive Diseases and Treatments During the COVID-19 Pandemic: A Google Trends Analysis

Purpose To assess national internet search trends/public interest in refractive diseases and treatments during the first year of the COVID-19 pandemic. Methods A Google Trends search for refractive terms was performed during the first year of the COVID-19 pandemic. Refractive terms were divided into disease and treatment terms. Relative search volume (RSV) indices were assessed in the United States from the initial lockdown period (March 1 - June 28), summer reopening period (July 5 - November 1), and winter case surge/vaccine rollout period (November 8 - February 28). A t-test of two independent samples assuming unequal variances was utilized in comparing the pandemic year to pooled data of overlapping weeks between 2016-2019.  Results The majority of disease and treatment terms showed a significant decrease in RSV during the initial lockdown period (p<0.05). There was a significant increase in RSV for cataract, astigmatism, cataract surgery, and photorefractive keratotomy (PRK) (p<0.05), accompanied by a significant decrease in RSV for contact lens during the summer reopening period. There was a significant increase in RSV for cataract, astigmatism, glasses, and PRK, accompanied by a significant decrease in RSV for hyperopia, keratoconus, contact lens, and LASIK during the winter case surge/vaccine rollout period.  Conclusion There was a significant decrease in the public interest in refractive diseases and treatments during the lockdown period, accompanied by an increase in interest later in the year. Decreased public interest can lead to delays in care, poorer health literacy, and potentially worse outcomes. Strategies to enhance public interest and care during the pandemic may prove to be beneficial.

Between Web search engines and artificial intelligence: what side is shown in laboratory tests?

BACKGROUND

The number of websites providing laboratory test information is increasing fast, although the accuracy of reported resources is sometimes questionable. The aim of this study was to assess the quality of online retrievable information by Google Search engine.

METHODS

Considering urinalysis, cholesterol and prostate-specific antigen (PSA) as keywords, the Google Search engine was queried. Using Google Trends, users' search trends (interest over time) were evaluated in a 5-year period. The first three or 10 retrieved hits were analysed in blind by two reviewers and classified according to the type of owner or publisher and for the quality of the reported Web content.

RESULTS

The interest over time constantly increased for all the three considered tests. Most of the Web content owners were editorial and/or publishing groups (mean percentage 35.5% and 30.0% for the first three and 10 hits, respectively). Public and health agencies and scientific societies are less represented. Among the first three and 10 hits, cited sources were found to vary from 26.0% to 46.7% of Web page results, whilst for cholesterol, 60% of the retrieved Web contents reported only authors' signatures.

CONCLUSIONS

Our findings confirm those obtained in other studies in the literature, demonstrating that online Web searches can lead patients to inadequately written or reviewed health information.

Accurate spatiotemporal mapping of drug overdose deaths by machine learning of drug-related web-searches

Persons who inject drugs (PWID) are at increased risk for overdose death (ODD), infections with HIV, hepatitis B (HBV) and hepatitis C virus (HCV), and noninfectious health conditions. Spatiotemporal identification of PWID communities is essential for developing efficient and cost-effective public health interventions for reducing morbidity and mortality associated with injection-drug use (IDU). Reported ODDs are a strong indicator of the extent of IDU in different geographic regions. However, ODD quantification can take time, with delays in ODD reporting occurring due to a range of factors including death investigation and drug testing. This delayed ODD reporting may affect efficient early interventions for infectious diseases. We present a novel model, Dynamic Overdose Vulnerability Estimator (DOVE), for assessment and spatiotemporal mapping of ODDs in different U.S. jurisdictions. Using Google^® Web-search volumes (i.e., the fraction of all searches that include certain words), we identified a strong association between the reported ODD rates and drug-related search terms for 2004–2017. A machine learning model (Extremely Random Forest) was developed to produce yearly ODD estimates at state and county levels, as well as monthly estimates at state level. Regarding the total number of ODDs per year, DOVE’s error was only 3.52% (Median Absolute Error, MAE) in the United States for 2005–2017. DOVE estimated 66,463 ODDs out of the reported 70,237 (94.48%) during 2017. For that year, the MAE of the individual ODD rates was 4.43%, 7.34%, and 12.75% among yearly estimates for states, yearly estimates for counties, and monthly estimates for states, respectively. These results indicate suitability of the DOVE ODD estimates for dynamic IDU assessment in most states, which may alert for possible increased morbidity and mortality associated with IDU. ODD estimates produced by DOVE offer an opportunity for a spatiotemporal ODD mapping. Timely identification of potential mortality trends among PWID might assist in developing efficient ODD prevention and HBV, HCV, and HIV infection elimination programs by targeting public health interventions to the most vulnerable PWID communities.

Google search volume predicts the emergence of COVID-19 outbreaks

Background and aim:

Digital epidemiology is increasingly used for supporting traditional epidemiology. This study was hence aimed to explore whether the Google search volume may have been useful to predict the trajectory of coronavirus disease 2019 (COVID-19) outbreak in Italy.

Materials and Methods:

We accessed Google Trends for collecting data on weekly Google searches for the keywords “tosse” (i.e., cough), “febbre” (i.e., fever) and “dispnea” (dyspnea) in Italy, between February and May 2020. The number of new weekly cases of COVID-19 in Italy was also obtained from the website of the National Institute of Health.

Results:

The peaks of Google searches for the three terms predicted by 3 weeks that of newly diagnosed COVID-19 cases. The peaks of weekly Google searches for “febbre” (fever), “tosse”( cough) and “dispnea” (dyspnea) were 1.7-, 2.2- and 7.7-fold higher compared to the week before the diagnosis of the first national case. No significant correlation was found between the number of newly diagnosed COVID-19 cases and Google search volumes of “tosse” (cough) and “febbre” (fever), whilst “dyspnea” (dyspnea) was significantly correlated (r= 0.50; p=0.034). The correlation between newly diagnosed COVID-19 cases and “tosse” (cough; r=0.65; p=0.008) or “febbre” (fever; 0.69; p=0.004) become statistically significant with a 3-week delay. All symptoms were also significantly inter-correlated.

Conclusions:

Continuously monitoring the volume of Google searches and mapping their origin can be a potentially valuable instrument to help predicting and identifying local recrudescence of COVID-19. (www.actabiomedica.it)

Current Cancer Epidemiology

In this brief report, we offer a concise overview on current cancer epidemiology garnered from the official databases of World Health Organization and American Cancer Society and provide recent information on frequency, mortality, and survival expectancy of the 15 leading types of cancers worldwide. Overall, cancer poses the highest clinical, social, and economic burden in terms of cause-specific Disability-Adjusted Life Years (DALYs) among all human diseases. The overall 0–74 years risk of developing cancer is 20.2% (22.4% in men and 18.2% in women, respectively). A total number of 18 million new cases have been diagnosed in 2018, the most frequent of which are lung (2.09 million cases), breast (2.09 million cases), and prostate (1.28 million cases) cancers. Beside sex-specific malignancies, the ratio of frequency between men and women is >1 for all cancers, except thyroid (i.e., 0.30). As concerns mortality, cancer is the second worldwide cause of death (8.97 million deaths) after ischemic heart disease, but will likely become the first in 2060 (~18.63 million deaths). Lung, liver, and stomach are the three most deadly cancers in the general population, while lung and breast cancers are the leading causes of cancer related-mortality in men and women, respectively. Prostate and thyroid cancers have the best prognosis, with 5-year survival ~100%, while esophagus, liver, and especially pancreas cancers have the worst prognosis, typically <20% at 5 years. We hope that this report will provide fertile ground for addressing health-care interventions aimed at preventing, diagnosing, and managing cancer around the world.