I am working with medical data and specifically determining whether Treatment_X is safe in certain prognostic groups of patients with chronic liver disease.

To summarize: a common prognostic score for liver disease patients is the Child Pugh Score, which takes on integer score 5-15, and is classified into categories A, B, and C depending on the numeric score (A, 5-6; B, 7-9; C 10-15). Please note that lower is better

We are interested in the change in numeric CP Score from pre-treatment to post-treatment after treatment with Treatment_X.

If we define a clinically meaningful increase in CP score as 2, we are interested in testing whether the rate of clinically meaningful increase in CP score is greater among patients who are "Child Pugh B" category before treatment compared to those who are in the more favorable "Child Pugh A" category.

What would be the best statistical method to analyze this data? Thank you so much for your help.


This raises a large number of worrisome issues. Foremost, a pre-post design is ill-suited for this type of question. It is too sensitive to time trends, Hawthorne effects, and regression to the mean. Second, an increase of 2 units has absolutely nothing to do with the way you analyze the data. It is used only at the final interpretation, or if you compute the Bayesian posterior probability of a difference of at least 2. Third, the classification into categories A, B, C is statistically invalid and clinically nonsensical.

Redesign the study as a parallel-group randomized trial. If you are unable to do that, then just ask the most general question you can. For a patient who begins at level X=x (y=5-15) what is the probability that she will end up at level Y=y. To do that I'd suggest a proportional odds model to estimate the P(Y >= y at follow-up given a baseline value of x) and model x as a quadratic expression in 5-15. You are likely to see a nonlinear effect (which also demonstrates that computing the change from baseline is a mistake). You can also use the prop. odds model to plot the mean of Y as a function of x.

Doing anything that loses information in the 5-15 levels is a crime against data. And using a pre-post design will not land the research in the best medical journals.

  • $\begingroup$ Thank you for your thorough answer, Dr. Harrell. I didn't describe our clinical question nearly well enough in my post so here's it is: "Are patients with Child Pugh B chronic liver disease significantly more likely to experience hepatic decompensation as a toxicity of Treatment_X as compared to patients with Child Pugh A chronic liver disease?" We and others in the literature have defined hepatic decompensation as an increase in the Child Pugh score of 2 or greater. Would this clarification affect your answer, or does all of that still stand? $\endgroup$ – JJM Feb 11 at 23:19
  • $\begingroup$ I'll also add that the A, B, and C categories of Child Pugh score, wisely or not, are standard "stages" of chronic liver disease and defined by the numeric prognostic score. We didn't create the a,b,c categories. $\endgroup$ – JJM Feb 11 at 23:26
  • $\begingroup$ Resist the temptation to use A,B,C - these are arbitrary, information losing, and have never been validated. To validate them you'd need to demonstrate that patient prognosis is the same for all levels of the ordinal scale that make up each group. Using A,B,C is declaring to the world that the underlying ordinal scale is defective. Similarly the definition of decompensation has not been validated nor will it be. To validate it you'd need to demonstrate a tie-in to patient reports and no difference with an increase of 4 or greater. $\endgroup$ – Frank Harrell Feb 12 at 13:14

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.