hw2 Q6 clarification

hw2 Q6 clarification

by Laura Koth -

I don't understand from the article this week by Choe et al whether deaths outside of the 4 cycle follow up were included in the analysis. perhaps I am not understanding the follow up time points but I think the paper defines this as yearly follow up for 4 years after enrollment. Can death be recorded at any time point up until they did their analysis? This would indicate that there could have been up to ~10years of time between the last follow up time point and the time the authors collected their data for a subject enrolled in 1995? is that correct?

thank you

Laura

Re: hw2 Q6 clarification

by Monica Ospina Romero -

In CaPSURE study, data are collected from two sources: The urologist enrolls eligible patients into the study, complete medical history, records recurrent status, treatment, and laboratory results at each office visit, and the patients who complete a baseline questionnaire and follow-up questionnaires periodically (this varied on survey type, and could be every 3 to 12 months).

In the paper, the authors explain that data on the time-varying exposure (Anticoagulants, AC) was collected approximately every year. Then, the authors gave some statistics about this time-varying exposure saying that the median years that AC was reported was 4 years (that’s what they mean by 4 cycles) with an interquartile range from 2 years to 6 years. 

Can death be recorded at any time point up until they did their analysis?

Yes, death could have been documented at any time point before the authors started their analysis. Death was documented by other sources besides CaPSURE, so if a patient was “lost to follow-up" from CaPSURE and then died, this was captured by these other sources. The study was initiated in 1995, however, we don’t know exactly the earliest start of follow-up data or the end of censoring date from the paper. The study period is presumably some interval between 1995 and 2012, based on the info that the study started in 1995, and as this was published in 2012. This information is important when you want to reproduce studies. This information is important when you want to reproduce studies.

This would indicate that there could have been up to ~10years of time between the last follow up time point and the time the authors collected their data for a subject enrolled in 1995? is that correct?

Yes. Some patients have been followed up at least 10 years (see figure 1 and 2) and those patients that have been followed up for 10 years are more likely to have been enrolled earlier in the study, i.e towards 1995.

For question 6, you should answer as if a participant had follow-ups every year until the outcome (death, recurrence, or end of the study occurred)


Additional General Comment / Recommendation on q6

by June Chan -

Hi Laura and all, 

Monica has nicely summarized the input from me and François (the other TA) with her own responses.

I just wished to add a further general clarification on this question - on q6, we ask for 1 figure, though we recognize that there are actually several specific questions the authors investigated, each which may have a different diagram. In your answer(s), please be clear about which specific question you chose to diagram, by defining the exposure and outcome, and timing in your explanation.


thanks,

June