At the risk of being confusing again, clinics selected patients at random from patients in their clinics who had a CFS diagnosis. As people have said, patients in their clinics are likely to have been at the more severe end.
Back to the CFI study: it had 1430 patients, 960 of whom returned surveys. 59 had died and of those 59, 19% (11) had suicided.
All 1430 patients had the opportunity to die, which is the core of the point that I'm trying to make.
We only know for sure that a minimum of 11 out of 1430 patients suicided, which is 0.8%. And, given that these are specialist clinics and (as argued) likely to be catering to the severer patients, that would be an overestimate of the rate in the ME population at large.
I realise that this is a lower limit, but it is the lower limit. So it would be true to say, 'As few as one in 120 CFS patients kill themselves'.
Unless I'm missing something, in which case I hope someone will correct me.
But we've got to be careful with these numbers, I think.
Yes, we confirmed that 19% of the 59 people whom we could confirm as deceased, had commited suicide. These people were not mailed surveys. They were mailed a notice that we would be contacting them (a presurvey mailing by a trusted doctor substantially increases participation). Then, a very strict contact protocol was followed for each patient. There were a minimum of seven calls made to each available contact number. Calls were required at various times of the day on various weekdays and at minimum intervals (the seventh call could not take place within 14 days of the first call in case someone was simply away and not checking messages, there we a minimun number of morning,mid-day and evening attempts as well as weekday and weekend attempts. Patients' "next of kin" contacts were called when patients had agreed to let the clinics contact these peoplefor follow-up (nearlyevery patient had affirmatively given that consenty). The phone contact protocol was restarted (7 required calls at specified intervals) for each new contact number provided by family, friends, former roommates, etc. And as I mentioned earlier, an extensive search of national death registries was conducted.
What I am getting at is that this study met very rigourous stnadards for population follow-yp studies, adopted from the current HHS and CDC standards. The people that were lost to follow-up were very lost to follow. As for your calculation of 11 in 1430, we contacted 1020 patients. There for, an absolute lowest possible figure would be 11 of 410. If everyone of the 410 patietns we were unable to reach had died (a frightening prospect in itself) and not one of them had committed suicide (again, highlyunlikey) the most optimitic percentage for deaths from suicide in this group of patients would be still be 11 in 469 (2.3%).
The odds of lost tocontact people all being dead are low (but many of them may have been). There is absolutely nothing to suggest thatthey had a lower suicdide rate than the rest of the group (I suspectthatit was likely equal as they probably had less stable housing, fewer close friends/relatives (weremoresocially isolated), etc.). Alll of that said,the Jason study and the Epi project definitely establish nothing on their own but they raise very disturbing questions.
People can interpret these project in any manner that suits their needs but I challenge you to find a single survey of any sort in any medical population with suicide rates that come anywhere close to 20% (and we have two such projects)! And methodologically, the Jason and EPI projects were close to "as good as you can possibly get" without millions of dollars to spend hunting down every last person.
