GENERAL COMMENTS

The paper reports the results of an interesting analysis of a pilot screening project. The interest is related both to the methodology adopted for assessing the impact of an intervention and to the results. Although the study was not designed as a randomized trial the authors can get valid and informative results by using propensity scores calculated at the individual level to identify a matched control cohort.

Also the analysis of the determinants of participation taking into account clustering by physician is appropriate. I have, however, some concerns about the interpretation of the results. As far as I can understand the approach using physican-linked invitations does not seem novel: personal invitation letters signed (i.e. with the printed signature ) by the GP have already been adopted in several pilots in Europe and are currently used in the context of organized screening programs in Italy, and probably also in several programs in France and Spain, based on previous experience with breast cancer screening. So the present study confirms (not really demonstrates) that this approach is feasible in the context of population based programs.

The finding of a higher response rate among patients of female physicians is interesting and novel in the context of CRC screening (some indication already existed in the context of breast cancer screening, as it could be expected). It might be related to a higher involvement of female practitioners in promoting preventive interventions, emerging from previous studies on smoking cessation, for example.

There are significant difference in the impact of the intervention by type of practice, the authors do not comment on this issue, but it deserves, in my opinion some comment, as this can represent a specific and novel contribution of this study. Indeed aspects related to the organization of practice are rarely investigated, as most studies have been focused on the type of invitation. I do not know in detail how the practice are actually organized, but if the practices adopting an interprofessional team model are basing their preventive activities on the collaboration with other health professionals (such as nurses), this would represent a useful indication on a possible way to enhance the impact of physician linked invitation for screening

GENERAL COMMENTS

The paper reports the results of an interesting analysis of a pilot screening project. The interest is related both to the methodology adopted for assessing the impact of an intervention and to the results. Although the study was not designed as a randomized trial the authors can get valid and informative results by using propensity scores calculated at the individual level to identify a matched control cohort. Also the analysis of the determinants of participation taking into account clustering by physician is appropriate.

I would only suggest to rephrase the initial statement of the strengths list: in fact the impact of physician's linked invitation has already been described in previous papers, some of them also mentioned by the authors. The EU guidelines for quality assurance in CRC screening and diagnosis are indeed already recommending to preferably use invitation letters signed by the GP – I A recommendation). Also as some large scale organized population bases European screening program are already routinely using invitation letters signed by the subject's GP, the feasibility of such approach has already been documented.

The present paper confirms both the effectiveness and the feasibility of such approach in a different jurisdiction; in addition it also documents the role of physician and patient related factors which influence the response rate

GENERAL COMMENTS

The revised version of the paper has addressed appropriately the issues raised by the reviewers. The methods, results and limitations of the interesting analysis of a well conducted intervention are adequately described . The paper offers also clear context related information which is useful to the reader to compare the results with different jurisdictions.