Hello,
I am assessing the relationship between ethnicity (categorical variable) and suicide death (binary). In their association, I want to assess the effect mediated by self-harm and mental illness (indirect effect) for which I used the KHB model. I am finding difficulties interpreting the results. Following is a demonstration related to ethnicity X compared to White (odds ratio) (had to change details for confidentiality).
I used this syntax:
khb clogit suicide i.ethnicity || selfharm mentalill, group(matchedid) or dis summary
Reduced: 0.70
Full: 0.87
Diff: 0.86
Summary of confounding:
Conf_ratio: 2.04
Conf_pct: 41.2
Components of Difference:
self-harm:
coef: -0.08
Std_Err: .0019
P_Diff: 31.41
P_Reduced: 14.57
How would we interpret the results for effect (particularly indirect effect), Conf_pct and component of difference (coef and P_Diff)?
Would it be correct to say that:
The odds of suicide amongst patients with history of self-harm and mental illness in ethnicity X compared to those with a similar history in White was 23% [(1-0.87)*100] less (direct effect). Amongst ethnicity X compared to White, self-harm and mental illness contributed to reducing the risk by 24% (indirect effect). This means that 41.2% (Conf_pct) of the total risk in ethnicity X compared to White was related to self-harm and mental illness. Self-harm alone contributed to 31.41% (P_Diff) of the reduction in odds in ethnicity X compared to White.
Also the negative sign before the coef is puzzling.
Sorry for the long enquiry, but I am a yet jumbled and would very much appreciate your kind help.
Kindest regards
Danah
I am assessing the relationship between ethnicity (categorical variable) and suicide death (binary). In their association, I want to assess the effect mediated by self-harm and mental illness (indirect effect) for which I used the KHB model. I am finding difficulties interpreting the results. Following is a demonstration related to ethnicity X compared to White (odds ratio) (had to change details for confidentiality).
I used this syntax:
khb clogit suicide i.ethnicity || selfharm mentalill, group(matchedid) or dis summary
Reduced: 0.70
Full: 0.87
Diff: 0.86
Summary of confounding:
Conf_ratio: 2.04
Conf_pct: 41.2
Components of Difference:
self-harm:
coef: -0.08
Std_Err: .0019
P_Diff: 31.41
P_Reduced: 14.57
How would we interpret the results for effect (particularly indirect effect), Conf_pct and component of difference (coef and P_Diff)?
Would it be correct to say that:
The odds of suicide amongst patients with history of self-harm and mental illness in ethnicity X compared to those with a similar history in White was 23% [(1-0.87)*100] less (direct effect). Amongst ethnicity X compared to White, self-harm and mental illness contributed to reducing the risk by 24% (indirect effect). This means that 41.2% (Conf_pct) of the total risk in ethnicity X compared to White was related to self-harm and mental illness. Self-harm alone contributed to 31.41% (P_Diff) of the reduction in odds in ethnicity X compared to White.
Also the negative sign before the coef is puzzling.
Sorry for the long enquiry, but I am a yet jumbled and would very much appreciate your kind help.
Kindest regards
Danah
