COMMENTS ON: PREDICTIVE VALUE OF HAVING POSITIVE FAMILY HISTORY OF CARDIOVASCULAR DISORDERS, DIABETES MELLITUS, DYSLIPIDEMIA, AND HYPERTENSION IN NON-ALCOHOLIC FATTY LIVER DISEASE PATIENTS (LETTER TO THE EDITOR)

摘要

A recent paper by Ghamar-Chehreh et al., about predictive value of positive family history of some common risk factors in nonalcoholic fatty liver disease patients (1) has a useful look at this issue; but, some methodological and statistical shortcomings present that are discussed. I believe that the quality of the analysis can considerably increase with the same data without the need for any additional work except the design of the analysis. Instead of only mentioning two percentages in two different groups and P-value (table 3), authors could calculate odds ratio (OR) and 95% confidence interval (CI) showing both the strength of each difference/ association in two groups (positiveegative family history) for each variable and significant difference/association. When 95% CI does not cover the number “one”, it means that the difference is significant. More advanced, they could also able to do logistic regression for estimating adjusted OR of the most important family histories for determining the status of different dependent variables such as HDL, TG, albumin, ALT, and so on.