Table A10: Study Design Strengths and Limitations by Severity of Hypoxemia for Relevant Studies From the Cochrane Review^*.

Study	COPD Study Population	Adequate Sample Size	Exclusions Detailed	Randomization Achieved	Blinding	Adequately Measured Adherence	All-cause Mortality	Survival Analysis	Intent-to- Treat Analysis†	Minimal Attrition
Severe Hypoxemia
MRC (10), ‡,§, ║	✓		✓	✓			✓		✓	✓
Mild-to-moderate Hypoxemia
Gorecka et al (24), ‡	✓			✓		✓	✓	✓	✓	✓
Haidl et al (25), ‡	✓			✓		✓	✓		✓	✓

^*Abbreviations: COPD, chronic obstructive pulmonary disease; MRC, Medical Research Council; checkmark (•) refers to the presence of study design strengths.

^†Considering mortality/survival as the main comparison of interest.

^‡Allocation concealment was adequate for none of the above studies and the process of generating randomized schedules was adequate for MRC (10) and Gorecka et al (24).

^§Survival analysis was not shown for the primary comparison of interest.

^║A larger proportion of patients in the treatment group smoked; however a post hoc analysis showed that this difference was not statistically significant (P > 0.05).

Source: Cranston et al, 2005 (23)