Amendments and corrections

Scand J Work Environ Health 2012;38(6):604    pdf


Re: Young AE, Wasiak R, Webster BS, Shayne RGF. Urban–rural differences in work disability after an occupational injury. Scand J Work Environ Health. 2008;34(2):158–164. doi:10.5271/sjweh.1217

by Young AE, Wasiak R, Webster BS, Shayne RGF

Thanks to the eyes of a keen reader, some inconsistency between the Wald Statistic P-values and the confidence intervals presented in Table 3 was observed. Specifically, it was noted that in our original Table 3, the age category 36–35 versus >35 years had a P-value of 0.778, but that the odds ratio (OR) 95% confidence intervals (95% CI) were 0.79–0.96. In addition, for the State comparison New York versus Texas, the P-value was 0.001, but the OR 95% CI were 0.99-1.39.

Looking back over the analysis, we realized that, for categorical variables with more than two levels, SAS requires a reference category be defined in the syntax. This was not included in original analysis. This omission influenced the calculation of the Wald Statistic and associated P-values, but had no impact on the OR, 95% CI, or other previously reported statistics. We have prepared this addendum to Table 3 with updated Wald and P-values for independent variables with more than two levels (ie, age, industry and State) and note that the change to values does not require any change to the body of the document or the conclusions drawn.

Thank you to the reader who point out this inconsistency and to the editors of SJWEH for giving us the opportunity to publish this addendum.

See also the article: 2008;34(2):158-164
This article refers to the following text of the Journal: 2008;34(2):158-164