GENERAL COMMENTS

This interesting study has added useful knowledge to the important topic of the effectiveness of pulmonary rehabilitation in COPD, different formats, effects and cost effectiveness. It is well conducted and reported and my comments relate mainly to optimizing understanding for the clinician reader. The CHEERS Checklist of items to include when reporting economic evaluations of health interventions is endorsed by BMJ Open, yet the paper is accompanied by the EVEREST statement checklist for a Health economics paper instead, which was a little confusing. On balance I prefer the CHEERS checklist although relevant items in EVEREST were completed. Research question clear definition.

study design

The paper reports data on cost effectiveness from a pulmonary rehabilitation cluster RCT; this is referenced as ref 9, which is described as “forthcoming”. Does this mean it is accepted for publication or is still under review? Could you clarify please. [ 9. Murphy K, Casey D, Devane D, et al. The effectiveness of a structured education pulmonary rehabilitation programme for improving the health status of people with moderate and severe chronic obstructive pulmonary disease in primary care: The PRINCE cluster randomised trial.2013.Thorax(forthcoming)]

The introduction contained details of the PRINCE RCT that belong better in Methods section. I suggest moving this description to Methods under a subheading “details of PRINCE cluster RCT, i.e. from page 5 line 20 starting “Full details of the study methods….” To Page 6 line 11, ending “Notably however, concerns arose as the confidence intervals did not exclude differences in effect that were pre-specified as clinically insignificant.[9]”

The intervention when described in the main paper and the abstract is described as a “structured education pulmonary rehabilitation programme (SEPRP)” whereas the reader needs to understand its components relative to “traditional” pulmonary rehabilitation, against which it is being considered as an alternative. Thus it needs to be clear if it is a structured education and supervised exercise pulmonary rehabilitation programme, or what was the form of the exercise component.

The opening sentence in paragraph 2 page 5 should then contain a brief and accurate definition of the intervention, (line 18) such as, “The PRINCE study sought to examine the clinical and cost effectiveness of a structured education and supervised exercise pulmonary rehabilitation programme (SEPRP) for COPD delivered at the level of general practice in Ireland.” The closing sentence (Page 6 lines 38-43) of the Introduction should then define the outcome measure used, the Chronic Respiratory Questionnaire (CRQ)

Description of methods

I found it difficult to understand some details of the methods used in the cluster RCT that are relevant to understanding the cost- effectiveness analysis, but if the paper on the main results is published before these results, this will help the reader. The methods used in the cost-effectiveness analysis reported here are generally well described. Authors state (on page 7 lines 52-55), that they used imputation to estimate healthcare costs and patient costs in the case of missing values for some individuals at follow up. Could you clarify how many participants had missing data and whether any sensitivity analyses were done with varying estimates?

In “Effectiveness analysis”, page 8 lines 10-23, a description of the CRQ is given and referenced [Ref 10]. Could you clarify how the three CRQ aggregate scores have been calculated, as this is not included in the original measure as referenced. Similarly to the above comments, on page 8 lines 41-46, could you clarify how many participants had missing data and whether any sensitivity analyses were done with varying estimates?

In the Sub-Section “Overview”, in the Methods section, authors state the use of a time horizon of 22 weeks. Could you explain why this period was used and why it was appropriate. (Line 53 page 6) If using perspectives of both the healthcare provider and the patient, should “perspectives” not be plural? (line 58, page 6) I was puzzled initially by “practice note searches” and assume you mean practice records? Were these paper or electronic records?

Results:

These are clearly described, although placing the Table of “Characteristics of clusters (general practices) and baseline demographic and clinical characteristics of COPD patients” as an appendix seems a pity. I would prefer to have that data in the main paper. It is also referred to as “Appendix Table 3” in the results section (page 9 line 46), whereas there is a Table 3 on page 17 and the Table in the appendix is labeled “Appendix Table 1”.

Table 3 (page 17) add “per patient” to cost figures for healthcare resources and patient resources.

Discussion

The difference in incremental cost effectiveness ratios based on the CRQ and QALY gained from the health utility measure highlighted by the authors are very interesting. Although they discuss some explanations and compare the results with other studies, I think this section could be expanded; especially to highlight the difference in target population for the SEPRP programme (more moderate COPD than hospital recruited programmes even if actually delivered in a non-hospital setting). It would be useful to give more details on studies reporting reduced hospital utilization (mostly in severe, very severe COPD) and compare directly with any other ICER estimates. Discussion of cost effectiveness of other rehabilitation/self management programmes delivered in primary care in diabetes for example DESMOND (Gillett et al BMJ 2010).

Abstract In the objective here the intervention should be clarified, again using content description of the intervention, whether it was a structured education and supervised exercise pulmonary rehabilitation programme. Participants: useful to know here 69% with moderate COPD Intervention- use an accurate description Main outcome measure: Include the basis for QALYs, ie the generic EQ5D. Results: clarify the mean increases in healthcare and patient costs are per patient. Conclusions: Use an accurate description of the intervention. If you have defined use of the generic EQ5D in the outcome measures this can be deleted here.