Table guidelines

A table should be clear on its own without context (self-explained, reduction of potential errors and confusion).

  • Reduce clutter whereever possible
  • Titles: [Important content] [outcomes (usually) ] [stratification] e.g.:
    • Number (%) of cUTI cases according to certain socio-economical and clinical characteristics.
    • Number and rates (per 1,000 admissions) for cUTI, cIAI and HAP by certain socio-economical and clinical characteristics.
    • Adjusted odds ratio (95% confidence interval) of the association between the receipt of inappropriate empiric antibacterial therapy and certain socio-economic and clinical characteristics by condition.
    • Adjusted odds ratio (95% confidence interval) of the association between the receipt of inappropriate empiric antibacterial therapy and admission to ICU after receiving antibiotics by condition and potential effect modifiers for 65+ patients
  • Use a total column when applicable
  • Avoid abbreviations (specify abbreviations in the footnote if needed)
  • Precision of at most 1 decimal for ratios, e.g. 7.5 [unless the estimates are so precise it warrants extra decimals; e.g. 0.77777 (0.77776-0.77779)]; EXCEPT p-values, use 3 decimals (0.075) or <0.001
  • Comma separate large numbers in groups of three, e.g. 123,456,789,000
    • SAS format family: comma32.3
    • Stata format family: %32.3fc
  • Only use parentheses as brackets, e.g. (years)
  • Use the most specific descriptions without duplications, eg:
    • Instead of: 1 week (1-7 days) , 2 weeks (8-14 days), >2 weeks (>14 days)
    • Use: 1-7 days, 8-14 days, >=15 days.

Specific epidemiology tables

  • Descriptive table 1
    • Column 1 = cases/exposed; column 2 = controls/unexposed
    • Start with sociodemographic vars (gender [is applicable], age [group], residential, income, etc.)
    • Balanced and sensible groups (based on context and distribition for study)
    • Order vars logically (e.g., metformin, insulin, other anti-diabetc instead of metformin, vaccine status, insulin)
    • Indent grouped variable under main group name (e.g. Age group, then indent 0-9, 10-19, etc. )
    • Include total row/total column where applicable
  • Descriptive table 2 in cohort studies
    • Include rate denominator in title.
    • Columns: # cases (outcomes), person-time, rates (95% CI) [, optionally rate ratio]
24/08/2018