I'm trying to help my wife with some statistics. She's looking to see if people with Removable Dentures (RD) have a higher prevalence of Combination Syndrome (CS). She did a cross-sectional study (survey) of 314 patients at her university. Of the 314 patients surveyed, 200 had Removable Dentures, and of those 200, 51 had Combination Syndrome. Of the 114 patients without Removable Dentures, 17 had Combination Syndrome. So, I think I set up the odds-ratio calculation correctly.

(CASES WITH POS. OUTCOME-> # in exposed group [a]51, # in control group [c]17: CASES WITH NEG. OUTCOME-> # in exposed group [b]149, # in control group [d]97)

An odds-ratio calculator gave me the following

Odds Ratio: 1.9530
95% Confidence Interval (CI): 1.0659 to 3.5784
z statistic: 2.167
Significance Level: P = 0.0303

My question: Does the CI show a statistically significant association? I believe it does because it isn't equal to 1.

Some info. I found on interpreting CI says "If the 95% confidence interval for the OR does not contain 1.0 we can conclude that there is a statistically significant* association between the exposure and the disease. (* at the 0 05 significance level)." My CI is 1.0659 to 3.5784, so it technically does contain "1.0", but is greater. So, I just want to be sure that I can correctly interpret the results and say YES, THERE IS A STATISTICALLY SIGNIFICANT ASSOCIATION.

I could say that the odds of having Combination Syndrome for patients with Removable Dentures is 1.9530 times greater than the odds of having Combination Syndrome if the patient does not have Removable Dentures, right?

With a 95% CI and a P value of 0.03, I would reject the null hypothesis and accept the alternative hypothesis, correct?

I also entered the data into a Relative Risk calculator and got the following:

Relative Risk: 1.7100
95% CI: 1.0388 to 2.8148
z statistic: 2.110
Significance level: P = 0.0349
NNT (Benefit): 9.445
95% CI: 5.003 (Benefit) to 84.310 (Benefit)

So I can say that patients who have Removable Dentures are 1.71 times more likely to have Combination Syndrome, right?

I really appreciate the help. I'm just trying to make sure we get the interpretation of the results right for her thesis. The rest of her classmates paid an expert (\$250) to do the work for them (which was encouraged by their director at the university here in Ecuador). Any other insights from this data are welcome and greatly appreciated

hi,
your interpretation of the OR is correct. I do not know the relative risk calculator, but I assume that it says that the risk is 1,7 times higher. I would not use the term"likely" because likelihood has a well defined meaning, which is probably not what is meant here.

regards and good luck to your wife

 Tweet