The Characteristics, Uses, and Biases of Studies Related to Malignancies Using Google Trends: Systematic Review (Preprint)

Abstract

BACKGROUND Google Trends (GT) is a freely-available tool presenting Google search statistics. GT is becoming a feasible tool to analyze the interest of Google users in different health phenomena. However, there are a few reviews on GT, and none has contributed specifically to oncology. OBJECTIVE We aimed to systematically characterize studies related to oncology using GT to describe its utilities and biases. METHODS We included all studies utilizing GT to analyze Google searches related to malignancies. We excluded studies written in non-English. The search was done on the PubMed engine on 1st August 2022. We used the following search input: "Google trends" AND ("oncology" OR "cancer" OR "malignancy" OR "tumor" OR "lymphoma" OR "multiple myeloma" OR "leukemia"). We analyzed the following sources of bias: 1) using search terms instead of topics, 2) lack of confrontation with real-world data, and 3) absence of a sensitivity analysis. We performed descriptive statistics. RESULTS A total of n=85 articles were included. We classified n=19 (22.4%) studies as related to prophylaxis, n=17 (20.0%) as an awareness event, n=9 (10.6%) as celebrity-related, n=11 (12.9%) related to COVID-19 and n=40 (47.1%) as others. The most frequently analyzed cancers were: breast (n=28), prostate (n=26), lung, and colorectal (both: n=18). n=79 (92.9%) of the studies provided all search input details to reproduce their results, and n=34 (40%.0) studies confronted GT statistics with real-world data. Authors of only n = 9 (10.6%) studies performed sensitivity analysis. CONCLUSIONS The studies in this systematic review varied regarding topics, search strategy, and statistical methods. Most researchers provided reproducible search inputs, but many studies lacked sensitivity analysis.Scientists using GT for medical research should ensure the quality of studies by 1) providing a search strategy to reproduce results, 2) preferring using topics instead of search terms, and 3) performing robust statistical calculations and sensitivity analysis.