HW4 q6

HW4 q6

by June Chan -

Questions received by email:

Question 6.  "urologists enrolled eligible patients into the study (after diagnosis of prostate cancer and before treatment)"
-> I am wondering if the determination of eligibility is decided between the diagnosis and the treatment of RT or RP or after the treatment. I assumed that patients who completed RT or RP would be eligible to participate in the study. This is not critical to answering the question, but I am not 100% clear about determination of eligibility in general. Maybe it might be up to the researchers? 

Response:  Yes, determination of eligibility is up to the researchers.  In this case, they considered participants eligible after diagnosis and before treatment for prostate cancer, as described at beginning of the Methods. The main criteria mentioned in the paper is that the men had to have biopsy proven cancer, and be being seen at one of the enrolling sites.  What is not specified, and really more of a FYI, is also that I believe men probably had to read/speak English to complete consent forms and surveys. I don't think they are bilingual.

Q6. "You may assume that the investigators initiated follow-up at the time of treatment for exposures that were not updated over time; and that for analyses using Cox proportional hazards regression, with updating of exposure at each year, that they started follow-up at the return of the first survey after primary treatment.”

-> The first sentence describes that T0 would be the time of treatment for exposure (=AC or non-AC?), and there is an additional description of the time, saying that they started follow-up at the return of the first survey. 

-> I am wondering if starting of follow-up is different from T0. 
Response:  T0 is the start of follow-up. 


Re: HW4 q6

by Kate Chirikova -
I have a follow-up question regarding Q6.
It is not entirely clear when the baseline questionnaire was completed for this study. According to the CaPSURE design it seems like the baseline data was collected at the time of defining eligibility status and before the treatment (RT or RP). But in the Q6 text T0 is defined in relation to the time of the treatment. So, should we assume that the baseline data that is used in this study is available at the time of eligibility or at the time of the study start (T0)?

Re: HW4 q6

by June Chan -
Hi All,
I just conferred with one of the TA's after Office Hours and heard about a few points of ongoing confusion with regards to this question. I will aim to clarify here.

First, answers may vary for this question, and you should state your assumptions and understanding of when T0, E, and A occurred for your sample subject that you are drawing, based on the information given but also based on the "story" you create, for that person. T0 is "time zero", E is "eligibility", and A is "assignment of exposure" (we know that the Hernan paper refers to this as "assignment of treatment", but for reasons given below, you may find it more helpful to think of this as "assignment of the exposure of interest").

The person-time drawings for this question in an open cohort, should include common events of the individual's "story", like "diagnosis with prostate cancer", "consent into the CaPSURE cohort", "Treatment for Prostate cancer with XRT or RP", "recurrence", and "Death due to prostate cancer". You do not necessarily need to label calendar time or years; however, if you have already done so, that is fine.

Note that for this question, we are only asking for an example drawing for 1-2 participants (you can choose to use the same person to depict 6a and 6b, or you can make two different stories of sample subjects). We also recognize that there is what is planned in a study and then there is also what happens for a single individual's case. Thus, answers may vary and you should provide some text explanation for the events that occurred for your participant(s) depicted.

In answer to Ekaterina's specific question, the "baseline questionnaire" refers to the FIRST questionnaire the participants are asked to fill out for the CaPSURE study in general, which was ideally given/returned shortly after consent. What we tell you to assume is that participants provided data on AC use on this baseline questionnaire, and roughly annually thereafter. For 6a (analysis with time-varying updating of exposure / Cox model), we tell you that T0 starts at the return of the first survey after treatment. For 6b, we tell you that T0 starts at time of treatment. It is up to you to decide which survey data informs the exposure categorization, for each question, based on the assumptions you make about this sample subject's story. For example, consider the difference in which survey is used to define exposure, if your sample subject got treated 6 months after diagnosis, or 3 yrs after diagnosis? What about if a subject never turned in any survey after baseline? or never turned in the baseline, but turns in a survey 3 yrs later?  All of these things actually happen in cohort studies. Thus, please make sure to label and describe your assumptions for your sample subject's story. Part of the point of this question is to consider the natural messiness that happens with observational cohort studies (compared to RCT's), and how this relates to the potential for bias.

For further clarification on terms used, see below:

"Eligibility" for CaPSURE cohort- There are really two types of eligibility going on in this paper, and in my response to one of the questions fielded on email (posted at the beginning of this thread), I only answered about the first one, not recognizing the confusion. There is "eligibility" to be in the CaPSURE study in general (which was being diagnosed with biopsy proven prostate cancer), and theoretically being seen at one of the enrolling sites, and speaking/reading English, then there is "Eligibility" for this one analysis within the CaPSURE cohort. To be "eligible" for this one paper - in addition to being eligible for the study in general, participants had to have localized disease and be treated with XRT or RP (2nd paragraph of the Methods).

"Treatment" - please note that in the question stem, for the part in quotations about the CaPSURE study, the term "treatment" in that paragraph is referring to "treatment for prostate cancer". This is not referring to "treatment" with anti-coagulants as the "exposure of interest", as in the causal inference framework. This is generic text about the study, and the investigators use "treatment" here to refer to the participants having XRT or RP. To help keep things separate, we recommend that for this question you use the word "exposure" to refer to anti-coagulants or just "AC".

I hope that is helpful!
Best,
JMC

Re: HW4 q6

by Hyelee Kim -
Thank you for your additional clarification. Since it is mentioned in the question that 'You may assume the following methods were applied in this study, to help develop the diagrams,' I thought that maybe we were supposed to apply the first eligibility to the study. So, according to your second posting, we can think that the eligibility for the Choe et al. study is established after the treatment of RT or RP.