Social Data: An Underutilized Metric for Determining Participation in COVID-19 Vaccinations

Abstract

Many measures have been taken since late 2019 to combat the coronavirus disease (COVID-19) pandemic. National, state, and local governments employed precautions, including mask mandates, stay-at-home orders, and social distancing policies, to alleviate the burden on healthcare workers and slow the spread of the severe acute respiratory syndrome coronavirus 2(SARS-CoV-2) virus until an efficacious vaccine was made widely available. By early spring of 2021, three effective and well-tolerated SARS-CoV-2 vaccines emerged and underwent broad distribution. Throughout the course of the COVID-19 vaccination campaign, several key logistical and psychological issues surfaced. Of these, access to vaccines and vaccination hesitancy are cited as two substantial hindrances towards vaccination. Noting the demand for the SARS-CoV-2 vaccine and its highly sensitive storage requirements, accurate dose allocation is critical for vaccinating the population quickly and successfully. Here, we propose the use of social data as a tool to predict vaccination participation by correlating Google searches with state-level daily vaccination. We identified a temporal and regionally-ubiquitous Google search syntax that broadly captures daily vaccination trends. By correlating trends in the search syntax with daily vaccination rates, we were able to quantify the correlation and identify optimal lag periods between Google searches and daily vaccination. This work highlights the importance of analyzing social data as a metric to effectively arrange vaccination roll-outs, identify voluntary vaccination participation, and identify inflection points in vaccination participation. In addition, social data assessments can help direct dose allocation, identify geographic areas that may seek, but lack, access to the vaccines, and actively prepare for fluctuations in vaccination demands.

Keywords: covid-19 vaccination; google trends; johnson and johnson vaccine; moderna vaccine; pfizer vaccine; social data; vaccination hesitancy; vaccine; vaccine hesitancy; vaccine participation.

Figure 1. Temporal and regional determination of highest sensitivity syntax

(A) Temporal RSV of searches relating to participation in COVID-19 vaccination. (B) Average RSV of each search term over the given time period. (C) US map illustrating the highest state average RSV for each term. (D) Individual state RSV for each search phrase (A = covid vaccine appointment, B = where to get covid vaccine, C = covid vaccine near me, D = how to get covid vaccine). RSV: relative search volume Data source: Google Trends

Figure 2. Public interest and sentiment in vaccination by manufacturers

(A) Temporal and (B) regional interest in the three major COVID-19 vaccine manufacturers. (C) State interest in each of the three major COVID-19 manufacturers. (D) Temporal and (E) regional interest in each vaccine manufacturer’s side effects. (F) State interest in each vaccine manufacturer’s side effects Data source: Google Trends

Figure 3. State-level vaccine dose utilization

(A) Percent of received COVID-19 vaccine doses administered. (B) Geographic heat map illustrating regional differences in dose utilization (dark = low utilization (50%); light = high utilization (95%)) Source: CDC VTrckS

Figure 4. Optimal lag correlation (individual dots) and population (aqua blue bars) of each state

White graph area = no correlation/negative correlation; yellow graph area = moderate correlation; green graph area = strong correlation) Data source: Google Trends and United States Census Bureau

Figure 5. States with highest correlation values between RSV for “covid vaccine near me” and daily vaccinations, both expressed relative to their maximum over the given time period

RSV: relative search volume Data source: Google Trends and CDC VTrckS

Figure 6. State-level temporal correlation between RSV for “covid vaccine near me” and daily vaccination numbers in states AL-HI

RSV: relative search volume Data source: Google Trends and CDC VTrCK

Figure 7. State-level temporal correlation between RSV for “covid vaccine near me” and daily vaccination numbers in states ID-MN

RSV: relative search volume Data source: Google Trends and CDC VTrCK

Figure 8. State-level temporal correlation between RSV for “covid vaccine near me” and daily vaccination numbers in states MS-OH

RSV: relative search volume Data source: Google Trends and CDC VTrCK

Figure 9. State-level temporal correlation between RSV for “covid vaccine near me” and daily vaccination numbers in states OK-WA

RSV: relative search volume Data source: Google Trends and CDC VTrCK

Figure 10. State-level temporal correlation between RSV for “covid vaccine near me” and daily vaccination numbers in states WI-WY

RSV: relative search volume Data source: Google Trends and CDC VTrCK