Forecasting COVID-19 Testing Load Using Google Trends: Experience from a Lower Middle-Income Country with over 10 Million Diagnostic Tests

Background: The ability to forecast changing trends of COVID-19 can help drive efforts to sustain the increasing burden on the healthcare system, specifically the clinical laboratories. We aimed to assess whether the trends of SARS-CoV-2 testing in Pakistan can be predicted using COVID-19 symptoms as search terms and analyzing the data from Google Trends. Methods: The number of weekly SARS-CoV-2 tests performed were retrieved from online COVID-19 data re-source. Google Trends data for the search terms with most common COVID-19 symptoms was analyzed for cross-correlation with the number of tests performed nationally. Results: A total of 10,066,255 SARS-CoV-2 diagnostic tests were analyzed. Search terms of fever, headache, and shortness of breath displayed a statistically significant correlation with total number of tests performed with a 1-week time lag. Conclusions: Google Trends data can be used to forecast the changing trends in COVID-19 testing. This informa-tion can be used for careful planning and arrangements to meet increased diagnostic and healthcare demands in difficult times.