key: cord-0798987-ton3wqcd authors: Piamonte, Bernadeth Lyn C.; Anlacan, Veeda Michelle M.; Jamora, Roland Dominic G.; Espiritu, Adrian I. title: Googling Alzheimer Disease: An Infodemiological and Ecological Study date: 2021-12-27 journal: Dement Geriatr Cogn Dis Extra DOI: 10.1159/000520692 sha: 43793a1f6ed54c3741ab2b299e450e32c5746f28 doc_id: 798987 cord_uid: ton3wqcd INTRODUCTION: Understanding the emergent role of the internet on the health-seeking behavior of people is critical not only in the areas of medicine and public health but also in the field of infodemiology. METHODS: Using Google Trends, data on global search queries for Alzheimer disease (AD) between January 2004 and April 2021 were analyzed. The relationship between online interest, as reflected by search volume index (SVI), and measures of disease burden, namely prevalence, deaths, and disability-adjusted life years, was evaluated. RESULTS: There was a reduction in the tendency to search for AD during the past two decades. SVI peaks corresponded to news of famous people with AD and awareness months. Symptoms, causes, and differences with the term dementia were central queries for persons interested in AD. No notable overall correlation between SVI and measures of disease burden was found due to competing results. Sub-group analyses, however, showed that these correlations may be influenced by socioeconomic development, with strong negative significant associations observed in lower middle-income countries. CONCLUSION: Online interest in AD may represent a more complex metric influenced by socioeconomic factors. Awareness of the impact of celebrity diagnosis and awareness months on online search behavior may prove useful in the planning of public health campaigns for AD. The internet has been an increasingly common source of online health-related information for millions of people. Public online interest in health and diseases can be reflected by trends in popular web search engines, such as Google [1] . Analysis of these trend data allows for realtime surveillance and ultimately serves to inform public health and policy [2] . This type of research constitutes the discipline of infodemiology: the science of determinants and distribution of information, especially on the internet [3] . DOI Dementia poses significant challenges to patients, caregivers, and healthcare systems due to its rising prevalence and socioeconomic burden [4] . Alzheimer disease (AD) is the prototype and accounts for approximately 50-60% of all dementias [5] . In helping meet these significant challenges, current research into better diagnostics and treatments continues to advance and has been very promising. As more and more research and newer findings about AD emerge, communication in the new age of information technology is vital and imperative in decision-making. Older adults are increasingly online, with 73% of Americans aged 65 years or older say they use the internet as of 2019 [6] . Analysis of Google searches about dementia has shown an increase over time in some countries and others have suggested that search volume data temporally coincided with the number of local new dementia cases [7, 8] . The aim of this study was to evaluate Google search trends worldwide over time for AD and assess the correlation between public online interest in AD and measures of the disease burden associated with AD and other dementias. Google Trends is a popular open tool that provides online traffic data on selected keywords and topics over time. Data analyses generated from Google Trends are being increasingly utilized in healthcare research [9] . Search activity is reported as search volume index (SVI), which is reflective of data that are normalized over a specific time period and geographic location to improve comparability between search terms. The numbers on the graphs generated reflect the number of searches completed using Google for a specific term relative to the total number of searches completed over time. They are scaled on a range of 0-100, where 100 is the maximum search interest for the time and location specified. A score of 0 means there were not enough data for this term. Online interest in AD was analyzed using Google Trends. The keyword "Alzheimer disease" or "Alzheimer's disease" was entered in the Google Trends main page (available at: http://www.google. com/trends, accessed 13 May 2021) [10] . The search phrase "Alzheimer's disease" yielded higher SVIs compared to "Alzheimer disease" and was hence the more commonly searched phrase. We opted to use "Alzheimer's disease (Disease)" than "Alzheimer's disease (search term)" as the former is inclusive of all search topics related to AD and not just search queries that include the selected text [1] . We set the following filters for the analyzed queries: "Worldwide," in "All categories" and for "Web searches." To reduce the noise in the data, we kept the default Google Trends option to exclude low search volume regions. We collected SVIs on a month-by-month basis dating from January 1, 2004, the earliest time with available data from Google Trends, to April 30, 2021. Data retrieval complied with a previously published Checklist for Documentation of Google Trends [9] . We descriptively analyzed the worldwide changes over time of search queries. We checked for seasonality in the time trend and possible correlations with major world events reported in Google Trends. We also reviewed top and rising queries related to AD. Information on the prevalence, deaths, and disability-adjusted life years (DALYs) associated with AD and other dementias reported per region were collected from the published article by the Global Burden of Disease 2016 Dementia Collaborators [11] . To evaluate the relationship between online interest in AD, as reflected by SVI, and these measures of disease burden, we performed a series of Spearman's rho correlation (ρ). Sub-group analyses were also performed to accommodate the different regions and economic profiles involved, as classified by the World Bank [12] . A value of p < 0.05 was considered statistically significant. Linear regression models and the creation of scattergram plots were also performed to determine the degree of variability contributed by the SVI in estimating disease burden. A trend line graph was created to determine the utility of SVI across the years. All statistical analyses were performed using the Stata 13 software. The infodemiology study does not include human subjects. Hence, the design of the study did not require Research Ethics Board approval. The study does not violate the terms of use of Google Trends. Overall, worldwide online interest in AD decreased by 29% over the entire period between January 2004 and April 2021 as shown in Figure 1 . A progressive decline in SVI over time was observed from January 2004 up until July 2013, with a reduction by 63% compared to its peak (SVI = 100) at June 2004. By April 2021, online interest has picked up by 62% compared to its nadir (SVI = 37) in July 2013. The SVI has remained to be within 49 and 69 worldwide and with an average of 59.57 (95% CI: 57.87-61.26). The time trend revealed a general seasonal variation: the interest was higher during September and lower during July and December. Main SVI peaks for AD corresponded with major news and current events as follows: "President Ronald Reagan (40th president of the USA) loses fierce battle with Alzheimer's …" (June 2004: SVI = 100) [13] ; "Robertson (religious broadcaster of The 700 Club) stirs passions with suggestion to divorce an Alzheimer's patient" (September 2011: SVI = 60) [14] ; "Biogen's Alzheimer's therapy (aducanumab) granted Food and Drug Administration Dement Geriatr Cogn Disord Extra 2021;11:333-339 DOI: 10.1159/000520692 fast-track status" (September 2016: SVI = 80) [15] ; "How Biogen's discontinued Alzheimer's drug (aducanumab) got a second life" (October 2019: SVI = 79) [16] ; and "B. Smith (American lifestyle guru) dies at 70 of early onset Alzheimer's" (February 2020: SVI = 73) [17] . On the other hand, a stark 23% drop was observed after the World Health Organization formally declared COVID-19 a global pandemic in March 2020. Top queries related to the Google search for AD included symptoms, causes, and differences with the term dementia. Search queries with the biggest increase over time included Russian, Chinese, Japanese, and Persian translations of terms related to dementia and AD [18] . As outlined in Table 1 , the greatest search volume for AD was in Puerto Rico (SVI = 100), followed by Spain (SVI = 60), Chile (SVI = 53), Portugal (SVI = 53), and the Netherlands (SVI = 51). The correlation between online interest in AD and the measures of disease burden associated with AD, and oth-er dementias are presented in Table 2 . There was no notable association as suggested by the correlation coefficients for the prevalence, deaths, and DALYs from AD. These correlation coefficients were also computed for the different regions and economic profiles included in the study -which showed variability across these locations and incomes. In the case of Europe and Central Asia -and to a certain extent Middle East and North Africa -the SVI increases along with the burden of disease. On the other hand, inverse correlation was notable for Latin America and the Caribbean and East Asia and Pacific -where the burden of disease was higher when the SVI was low. The correlation coefficients for the sub-Saharan African, South Asian, and North American regions could not be calculated because there were only one to two members in the regions listed. When the countries were classified based on their income, it was noted that only those with high incomes have a positive (nonsignificant) correlation between SVI and the burden of disease and the opposite for the lower and upper middle-income countries. There was a strong negative significant correlation between the SVI for the lower middle-income countries and the prev- DALYs, disability-adjusted life years. * Data from Google Trends: interest by region. † Information on the "prevalence," "deaths," and "DALYs" for "Alzheimer disease and other dementias" was obtained from the published article by the Global Burden of Disease 2016 Dementia Collaborators and was reported as counts. The data gathered from Google Trends were noted to account for a small proportion of variability in the study data. The computed R-squared values (i.e., the coefficients of variation) indicated that the SVI only accounted for 2.15%, 1.13%, and 2.22% of the variability in prevalence, deaths, and DALYs, respectively. There was also no strong and clear trend between SVI and these different measures of disease burden in any of the scattergram plots. In analyzing Google Trends data on AD from the period of January 2004 to April 2021, we found that worldwide online interest in AD has decreased during the past two decades. This may correspond to the decreasing temporal trends of dementia incidence -a 13% drop per decade -in Europe and North America, which encompasses most of the highest-ranked countries in terms of SVI for AD [19] . The rising queries related to the Google search for AD, in comparison, may reflect countries that have seen a recent increase in burden of disease, such as Japan and China [20, 21] . While this study failed to find a notable overall correlation between SVI and measures of disease burden likely due to the competing results across regions and incomes, sub-group analyses suggest that such correlations may be influenced partly by socioeconomics. We observe that the higher the income of a country, the more that public online interest aligns with and may be indicative of the disease burden. This could perhaps be attributed to more resources available to the region to invest in education, greater access to technology, and effective public awareness campaigns. It is noteworthy that the top 10 regions with the highest SVIs for AD in this study were all high-income economies and 7 out of these 10 had a National Dementia Plan [22] . The opposite trend, wherein a higher burden of disease is associated with lower online interest, is observed in lower income economies, significantly and strongly in lower middle-income economies and nonsignificantly in upper middle-income economies. These findings perhaps serve as a reflection of the challenge in educating and raising awareness for AD as well as providing greater online access to health-related information where resources are limited and priorities are competing to begin with. This puts lower income economies at an even more vulnerable position as they are projected to bear most of the brunt of the rise in dementia cases in future years [23] . Nine of the lowest 10 regions are also bereft of a National Dementia Plan [22] . There are sparse data existing about online information-seeking behaviors specific to AD, and few studies have documented variable observations of the impact of socioeconomic development on online search volume [24] [25] [26] . To our knowledge, this is the first study analyzing Google search trends worldwide over time for AD and its correlation with measures of disease burden. We also observed a relation between SVI peaks and news headlines, particularly about famous people with AD. This finding is in line with previous studies that have seen similar links between celebrity diagnosis and increased public interest in neurological disorders [24, 27] . Making use of this association in public health campaigns, either through photographs or other promotional multimedia, holds promise for more effective awareness programs on and stigma reduction related to AD. The seasonal pattern of SVI peaks in September also matches the celebration of World Alzheimer's Month in September [28] . This finding is consistent with past documentation of awareness months generating interest of Google users [29] . We surmise that World Alzheimer's Month further motivates healthcare professionals and institutions to initiate public health campaigns on AD. A slump in online search volume on AD during the month of December is compatible with previous reports citing that most people will prefer to spend this time celebrating the holidays with friends and family [29] . Likewise, the drop in interest observed after March 2020 can be attributed to hyperawareness and desire for more online information on CO-VID-19 and related novel concepts, such as social distancing and mask wearing, combined with a decline in health-seeking behavior by patients with underlying health risks made more vulnerable by a pandemic [30] . Commonly searched queries related to AD include information on its symptoms, causes, and difference with the term dementia, highlighting the level of basic knowledge of the general public and their priorities of concern. Roughly one in every three internet users has said that online health-related information has influenced their decision to visit a doctor [31] . In this context, necessary precautions must be taken to ensure that accurate, reliable, and updated information on AD is available online for both patients and their caregivers. The study has several limitations. While arguably the most popular, Google is not the only web search engine and its share of search traffic varies across regions [32] . The study data may be affected by the nonrepresentative sampling bias inherent in the Google Trends service. Us-DOI: 10.1159/000520692 ers were anonymous in keeping with data confidentiality, but this makes it impossible to subject the sample to statistical analyses by demographic characteristics. Information was collected from people with internet access, therefore possibly omitting lower income groups, novice computer users, or those in areas with lower scorings in freedom of speech [1] . Total search volume per region for AD was unavailable through Google Trends and would have added supplementary comparison metrics aside from SVI. Moreover, the study data gathered from Google Trends only accounted for a small proportion of variability in the estimates of disease burden likely because of outliers in the dataset and significant confounding brought about by the competing results in the correlation coefficients. This study did not include the incidence of AD in its correlational analyses. However, given that numerous studies have reported the growing use of Google Trends data in outbreak surveillance, a future study opportunity would be to investigate possible relationships between the incidence of AD and Google search trends for AD [9] . Lastly, this study excluded low search volume regions that may be susceptible to irregular variation. However, by doing so, low-income countries were not captured in the dataset. Studies including low search volume regions in similar correlational analyses may add further insight into the impact of socioeconomics on online search behavior. Given these limitations, caution must be given in processing the data and interpreting these results. Google Trends data analysis has shown that there was a reduction in the tendency to search for AD during the past two decades. While no notable overall correlation between SVI and measures of disease burden was found, sub-group analyses revealed that SVIs for AD may instead represent a more complex metric influenced by socioeconomic factors. News of celebrity diagnosis and awareness months seem to be major determinants in online search behavior for AD, and this knowledge may prove useful in planning public health campaigns for AD. Google trends in infodemiology and infoveillance: methodology framework Detecting influenza epidemics using search engine query data Infodemiology and infoveillance: framework for an emerging set of public health informatics methods to analyze search, communication and publication behavior on the internet The state of the art of dementia research: new frontiers. London: Alzheimer's Disease International New diagnostic concepts in Alzheimer's disease Pew Res Center An analysis of trends in internet searches about Dementia in Turkey. Dokuz Eylül Üniversitesi Hemşirelik Fakültesi Elektron Derg Forecasting the incidence of dementia and dementia-related outpatient visits with google trends: evidence from Taiwan The use of Google Trends in health care research: a systematic review Google trends a systematic analysis for the Global Burden of Disease Study World Bank list of economies President Ronald Reagan loses fierce battle with Alzheimer's at 93 in 2004 Robertson stirs passions with suggestion to divorce an Alzheimer's patient. The New York Times Biogen's Alzheimer's therapy granted FDA fast-track status How Biogen's discontinued Alzheimer's drug got a second life restaurateur and lifestyle icon, dies at 70 of early onset Alzheimer's. NPR Twenty-seven-year time trends in dementia incidence in Europe and the United States: the Alzheimer Cohorts Consortium Trends in dementia prevalence, incidence, and survival rate in a Japanese community Is the dementia rate increasing in Beijing? Prevalence and incidence of dementia 10 years later in an urban elderly population Alzheimer's Disease International. Dementia plans World Alzheimer report: the global impact of dementia. London: Alzheimer's Disease International Why do people google epilepsy? An infodemiological study of online behavior for epilepsy-related search terms Google search behavior for status epilepticus Levels of toothache-related interests of Google and YouTube users from developed and developing countries over time Web search behavior for multiple sclerosis: an infodemiological study London: Alzheimer's Disease International Googling on colonoscopy: a retrospective analysis of search engine statistics Information-seeking patterns during the COVID-19 pandemic across the United States: longitudinal analysis of Google Trends data The online health care revolution: how the web helps Americans take better care of themselves The Pew Internet & American Life Project Google: search engine market share in selected countries The authors would like to thank Alvin Duke R. Sy, RN, Department of Epidemiology and Biostatistics, College of Public Health, University of the Philippines Manila, Manila, Philippines, for statistical analysis. The authors have no conflicts of interest to declare. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Publicly available datasets were used in this study. These can be found in http://www.google.com/trends and http://dx.doi. org/10.1016/S1474-4422(18)30403-4.