Regression coefficients requirement for calibration plot

Dear all,

We are analyzing a dataset collected from over 20 hospitals trying to externally validate the qSOFA score to predict in-hospital mortality (PASSEM Study). After days of reading, including this nice dataform, I learned that the development cohort’s regression coefficient is a must to assess calibration and plot calibration plots. I tried for over two months to get the original coefficients from the authors of the development cohort, but I am still waiting for a response (not sure they will respond).

My question:
Are the development cohort’s regression coefficients required to assess calibration and plot calibration plots for external validation? If they so and didn’t get them, is it okay to mention in our manuscript that we could not get them.

I came across this excellent topic, but I need help finding an answer to my question.

If any other methods can help us, sharing them would be highly appreciated.

Thank you

1 Like

Failure to share the coefficients raises suspicion on the authors.

The only other approach is to give them your data and ask them to compute the predicted values and give back to you. Don’t share the outcome variable with them.

1 Like

Thank you so much Professor :hibiscus:

Hello! I created my own mortality prediction model for ICUs and need data for external validation. Would it be possible to get them from you? We can compare the results and outcomes after I calculate the risk of death.

1 Like

So happy to share data, let me first take approval from our colleagues :tulip:

1 Like