I've been going on about the issue of top-down spread / scaling up initiatives which overstate the benefits. For example, if the pilot projects gets a 50% improvement, then that is, of course, relevant to their baseline. And we can only translate this benefit across a system f we know the baseline of all the individual potential beneficiaries. The worse we can do is take a national average and then assume everyone can get the same 50% improvement that the pilot site obtained.
I've written about this in "Undressing the Elephant: Why good practice doesn't spread in healthcare" and it's great to see a systematic review based around diabetes care that supports the notion that "interventions solely targeted at healthcare professionals seem to be beneficial only is baseline HbA(1c) control is poor". [This research and paper was funded by Ontario Ministry for Health]
Lancet. 2012 Jun 16;379(9833):2252-61. Epub 2012 Jun 9
Effectiveness of quality improvement strategies on the management of diabetes: a systematic review and meta-analysis.
Tricco AC, Ivers NM, Grimshaw JM, Moher D, Turner L, Galipeau J, Halperin I, Vachon B, Ramsay T, Manns B, Tonelli M, Shojania K.