. 2021 Feb 15;21:35. doi: 10.1186/s12874-021-01220-1

Table 1.

Assessment of applicability of causal inference concepts to vaccine studies

Reference	Objective	Design / statistical model	Results	Hill’s Criteria	Counterfactual Thinking
Guignard et al., 2014 [8]	To assess whether diabetes is a risk factor for HZ	Matched cohort study in a US insurance claim database. Cox Proportional hazards regression Analysis adjusted for age was applied to estimate the HR of developing HZ, including age, gender and co-morbidities as covariates	• Type II diabetes in subjects ≥65 years HR 3.12; 95% CI 2.77–3.52 • Type II diabetes in subjects 40–64 years HR 1.51; 95% CI 1.42–1.61 Cardiac disease and chronic pulmonary disease were also risk factors in non-diabetic subjects: • Cardiac disease HR 1.92; 95% CI 1.73–2.13 • Chronic pulmonary disease HR 1.52; 95% CI 1.38–1.67	Str: N Cons: N Spe: N Temp: Y Biogr: N Plau: Y Coh: Y Exp: N Ana: N	Exch: N Pos: Y Cons: N
Willame et al., 2016 [9]	Risk of new onset autoimmune disease in 9–25-year-old women exposed to AS04-HPV-16/18 vaccine in the United Kingdom	Retrospective cohort study to assess the risk of autoimmune diseases in vaccinated and unvaccinated individuals (frequency matched for age and practice region) using a Poisson regression model	• IRR for “other autoimmune diseases” in the vaccinated and unvaccinated cohorts 1.41; 95% CI 0.86–2.31	Str: N Cons: N Spe: N Temp: Y Biogr: N Plau: Y Coh: N Exp: N Ana: N	Exch: N Pos: Y Cons: Y
Braeckman et al., 2012 [10]	Effectiveness of two doses of HRV in preventing hospitalization for rotavirus gastroenteritis in Belgium among children born after October 2006 and aged at least 14 weeks	Prospective, hospital-based, multicenter, matched case-control study. Percent VE = (1- OR)*100. OR calculated from a conditional logistic regression with 95% CI	• Unadjusted VE for two doses 90% (95% CI: 81–95). • Unadjusted VE of at least one dose 91% (95% CI: 82–95)	Str: Y Cons: Y Spe: Y Temp: Y Biogr: N Plau: Y Coh: Y Exp: Y Ana: Y	Conditional Exch: likely Pos: Y Cons: Y

Reference

Objective

Design / statistical model

Results

Hill’s Criteria

Counterfactual Thinking

Guignard et al., 2014 [8]

To assess whether diabetes is a risk factor for HZ

Matched cohort study in a US insurance claim database. Cox Proportional hazards regression

Analysis adjusted for age was applied to estimate the HR of developing HZ, including age, gender and co-morbidities as covariates

• Type II diabetes in subjects ≥65 years HR 3.12; 95% CI 2.77–3.52

• Type II diabetes in subjects 40–64 years HR 1.51; 95% CI 1.42–1.61

Cardiac disease and chronic pulmonary disease were also risk factors in non-diabetic subjects:

• Cardiac disease HR 1.92; 95% CI 1.73–2.13

• Chronic pulmonary disease HR 1.52; 95% CI 1.38–1.67

Str: N

Cons: N

Spe: N

Temp: Y

Biogr: N

Plau: Y

Coh: Y

Exp: N

Ana: N

Exch: N

Pos: Y

Cons: N

Willame et al., 2016 [9]

Risk of new onset autoimmune disease in 9–25-year-old women exposed to AS04-HPV-16/18 vaccine in the United Kingdom

Retrospective cohort study to assess the risk of autoimmune diseases in vaccinated and unvaccinated individuals (frequency matched for age and practice region) using a Poisson regression model

• IRR for “other autoimmune diseases” in the vaccinated and unvaccinated cohorts 1.41; 95% CI 0.86–2.31

Str: N

Cons: N

Spe: N

Temp: Y

Biogr: N

Plau: Y

Coh: N

Exp: N

Ana: N

Exch: N

Pos: Y

Cons: Y

Braeckman et al., 2012 [10]

Effectiveness of two doses of HRV in preventing hospitalization for rotavirus gastroenteritis in Belgium among children born after October 2006 and aged at least 14 weeks

Prospective, hospital-based, multicenter, matched case-control study. Percent VE = (1- OR)*100. OR calculated from a conditional logistic regression with 95% CI

• Unadjusted VE for two doses 90% (95% CI: 81–95).

• Unadjusted VE of at least one dose 91% (95% CI: 82–95)

Str: Y

Cons: Y

Spe: Y

Temp: Y

Biogr: N

Plau: Y

Coh: Y

Exp: Y

Ana: Y

Conditional Exch: likely

Pos: Y

Cons: Y

AS04-HPV-16/18 vaccine, Cervarix, GSK; CI Confidence Interval, HZ Herpes zoster, HR Hazard ratio, HRV Oral human live-attenuated rotavirus vaccine, IRR Incidence rate ratio, OR Odds ratio, VE Vaccine effectiveness, Str Strength, Cons Consistency, Spe Specificity, Temp Temporality, Biogr Biological Gradient, Plau Plausibility, Coh Coherence, EXP Experiment, Ana Analogy, Exch Exchangeability, Pos Positivity