Introduction to Research Methods and Data Analysis in the Health Sciences by Gareth Hagger-Johnson

Author:Gareth Hagger-Johnson [Hagger-Johnson, Gareth]
Language: eng
Format: azw3
ISBN: 9781317674412
Publisher: Taylor and Francis
Published: 2014-06-20T04:00:00+00:00

The authors present a p-value for the interaction which was not statistically significant (p = 0.87). This suggests that the combined effect of each behaviour is not significantly greater than their separate effects, and that there is no strong evidence supporting effect modification. To illustrate how the combined effect has been shown in Table 4 in the paper, Table 9.5 offers a simplified version of the results. Think about what you would expect to see if the risk of mortality was even greater when smoking was combined with heavy alcohol drinking. You would expect to see an even stronger hazard ratio for that group, shown in the box in Table 9.5. This is exactly what we see.

The results are slightly complicated by the fact that moderate alcohol drinking may be protective, so we do not see a linear increase from never, to moderate, to heavy alcohol consumer (as with the results focusing on alcohol only, in Table 2 in the paper). In both moderate smokers and heavy smokers, we see a decrease in the risk from never to moderate alcohol consumption, then an increase from moderate to heavy. This is consistent with a ‘U-shaped’ association. The effects are stronger among heavy smokers. In fact, what we do see is the strongest association in the combined, ‘super-exposed’ heavy alcohol and heavy smoker group (HR = 1.9). However, this combined effect is not significantly stronger overall, because the p-value for interaction was 0.87.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.