Fixed factors that modify the effects of time-varying factors: applying the case-only approach.

摘要: BACKGROUND: The effects of air pollution or weather on mortality may be stronger in susceptible groups. Conventional investigation such effect modification through interaction terms time-series regression analysis depends hard-to-verify modeling assumptions, and can computationally unwieldy. As an alternative, we investigate the use case-only approaches originally proposed for studying gene-environment interactions. METHODS: We consider whether persons low socio-economic status (SES) are more to high outside temperatures mortality. If SES prevalent among deaths hot days than with moderate temperatures, then this suggests group is susceptible. Extending theory developed interactions allows described quantitatively. RESULTS: Conventionally based estimated that Sao Paulo rose by 2.3% (SE 0.3%) each degree increase temperature above 20 degrees C. This was greater 1.11% 0.72) lowest compared highest quartile SES. Case-only difference 1.14% 0.72). CONCLUSION: simplicity reduced assumptions approach provide advantage over conventional analysis, although gives information only modification, not main effects.