Apostasy!  Bombast! Truth!  I love it.  As a close-in orbiter of the Crimson Giant star of academic medicine, I have scoffed aloud but largely alone at several of the treasured paradigms of the reigning powers of medical teaching and academic medical center practice.  You single out one of the biggest of these fraudulent enterprises, the multicenter megabucks reductionist exercise in the trivial and irrelevant, but these enterprises and their spawn are many and are lately joined by the wholesale infliction of half-done social and economic theory wedded to junk science that has given us diluted postgraduate training, flash-mob "team" rounds, "productivity" instead of value in office medicine, third-party economic clout replacing professional ethics...  Good luck on your apparent efforts to turn things around, but I think we will see worse before better.  The academic enterprise, including the medical schools, is run by grant-funded researchers, the economics by government and insurance companies, and the policies by politicians who will not have to live by the rules they propound; none of these will relinquish status, power and wealth just because they are wrong.

I am so glad Dr.Topol has said it .He knows the issue of RCT very well .

N-of 1 trial design is the way of the real and enlightened.

We have done all our diabetic patients research with the N-of-1 trial design since 1986 and truly it is the way to all clinical research for efficacy in the future  ,for a number of reasons:

1)you avoid the heterogenity of patients found in every major "mass" trial.

3)you are able to compare the same patients to themselves on and off Rx .

2)Savings is immense : we have proven some interventions worth with just 10-20 patients with strength in P value of the range of 0.0001.

4)saving in time of reaching the answer  somtimes in less than 6 months.

5)you are testing one factor at the time with remarkable imitation of real life clinic.

For more information see examples :

 1)A. Tsapas . D. R. Matthews (2008) N of -1 trial.  Diabetologia 51:921–925.)

2). Cook DJ(1996).  Randomized trials in single subjects: the N of 1 study.  Psychopharmacol Bull. 1996;32(3):363-7).

3)Sakkal S.“A Case of Brittle Diabetes Treated with Pocket Insulin Dosage Computer”: American Diabetes Association.June 1986).

4).Sakkal S .Brittle Diabetes treated with intensified therapy and a pocket insulin dosage computer .Diabetes nutrition & metabolism .6: 97-101 .1993

5)Sakkal S. “Cost effectiveness of the use of telecommunication in Diabetes Management for daily diabetes control Abstract 77^th Ann.meet.of the Endoc.society  poster Number:P3-295- Washington DC. June  1995

Welcome back Topol-and-Teirstein.

Your informal but informative style is a pleasure to listen too during a busy day.

For honing in and finding things more diffuclut among the easy:

1. The way you look at data, its overall quality and existing weaknesses is meaningful. One question would be how to get that across to mass media interpreters?

2. Consider a slicing of the data into [a] the appropriate and [b] the inappropriate case sets. Then roughly sort a % of the cases within each group along a range or pole going from clear and easily seen moving into grey areas and more difficult to see. Finally focus on the grey areas and there raise the issues of quality of the data and where more than minimum symptoms are best applied. 

Those two thing can be best applied to controlling costs and achieving the most favorable media coverage in as clear, precise and concise a manner as possible.

Keep up the good work.