Table 2

Quality assessment criteria

Comparability of interview methods between time pointsGood: Identical
Fair: Change in mode
Poor: Change in disability, functioning or health outcomes
Quality of outcome measureGood: Detailed multiple item measure
Fair: Single item global measure
No criteria for poor
Uses more than two time points to assess trend*Good: Uses more than two time points
Fair: Uses only two time points
No criteria for poor
% Response in repeated cross-sectional studies*Good: >70% response rate and <10% drop in subsequent surveys
Fair: <70% or >10% drop in subsequent surveys
No criteria for poor
Loss to follow-upGood: NA or <5%
Fair: 5%–10%
Poor: >10%
Note: This only applies to longitudinal study designs (ie, not independent repeat cross-sections)
Proportion of proxy interviewsGood: <10%
Fair: 10%–20%
Poor: >20%
Proportion of missing dataGood: <5%
Fair: 5%–10%
Poor: >10%
  • *Added to the quality assessment following expert advice.