Multivariable logistic regression results*†: determinants of: (1) accessing‡ news media for health information, (2) accessing‡ social media for health information and (3) self-reporting perceived negative impacts from health information accessed‡ through news/social media
| Explanatory variables | News media (n=2111) | Social media (n=2111) | Negative impacts (n=1866) | |||
| Adjusted OR | 95% CI | Adjusted OR | 95% CI | Adjusted OR | 95% CI | |
| Region | ||||||
| Canada (ref) | 1.00 | – | 1.00 | – | 1.00 | – |
| Asia | 3.19 | (1.94 to 5.25) | 2.26 | (1.50 to 3.39) | 0.31 | (0.16 to 0.61) |
| Europe | 3.04 | (2.17 to 4.26) | 1.09 | (0.79 to 1.51) | 0.97 | (0.55 to 1.72) |
| Latin America | 2.32 | (1.34 to 4.01) | 2.26 | (1.35 to 3.79) | 0.69 | (0.26 to 1.83) |
| USA | 1.36 | (1.04 to 1.79) | 0.71 | (0.54 to 0.95) | 1.38 | (1.01 to 1.89) |
| Sociodemographics | ||||||
| Age | 1.01 | (1.001 to 1.02) | 0.97 | (0.96 to 0.98) | 0.98 | (0.97 to 0.98) |
| Female | 1.00 | (0.65 to 1.53) | 1.49 | (0.93 to 2.37) | 2.23 | (1.17 to 4.24) |
| Postsecondary education | 1.55 | (1.18 to 2.04) | 0.92 | (0.70 to 1.21) | 0.66 | (0.44 to 0.98) |
| Europe×postsecondary education§ | – | – | – | – | 3.05 | (1.57 to 5.91) |
| Latin America×postsecondary education§ | – | – | – | – | 4.05 | (1.34 to 12.25) |
| Access‡/Trust¶ in health information sources | ||||||
| Access—peers | 2.88 | (2.27 to 3.66) | 1.93 | (1.54 to 2.49) | – | – |
| Access—advocacy organisations | 1.61 | (1.25 to 2.07) | 1.96 | (1.55 to 2.49) | – | – |
| Access—news media | – | – | 6.93 | (5.30 to 9.05) | – | – |
| Access—social media | 8.65 | (6.51 to 11.50) | – | – | – | – |
| Access—alternative care providers | – | – | 1.43 | (1.03 to 1.97) | 1.59 | (1.14 to 2.22) |
| Trust—family physicians | 0.74 | (0.56 to 0.97) | – | – | – | – |
| Trust—news media** | 5.85 | (4.59 to 7.45) | – | – | 0.59 | (0.46 to 0.76) |
| Trust—social media†† | 0.62 | (0.46 to 0.84) | 4.68 | (3.65 to 6.00) | – | – |
Significant covariates in bold.
*Sociodemographics, disease characteristics, number of days to survey completion since 11 March 2020, and access to and trust in sources were considered in the models. Only significant variables are presented here.
†Variables with a ‘–’ were not included in the final model. Region, age, gender and postsecondary education were forced into the models, and other variables were chosen through backwards stepwise selection.
‡Respondents accessing source for health information sometimes/often/always post-11 March 2020.
§Interaction between region and postsecondary education was added to explore the relationship of postsecondary education in different regions. Only significant interactions included in the final model.
¶Respondents reporting source as somewhat/very trustworthy post-11 March 2020.
**Respondents were not asked to identify level of trust in news media overall. Value derived from grouping four news media sources (newspaper, online news media, radio, television) included in the survey to create a single trust in news media variable. Respondents were considered to trust news media if they reported any of the individual news media sources as somewhat/very trustworthy.
††Respondents were not asked to identify level of trust in social media overall. Value derived from grouping social media sources (Facebook, Instagram, internet blog, Twitter, YouTube, LinkedIn, Pinterest, Reddit, TikTok) included in the survey to create a single trust in social media variable. Respondents were considered to trust social media if they reported any of the individual social media sources as somewhat/very trustworthy.