Archived Comments for:
A multifaceted feedback strategy alone does not improve the adherence to organizational guideline-based standards: a cluster randomized trial in intensive care
Observation regarding sampling and the discussion of findings
Anda Bayliss, Royal College of Nursing
13 July 2015
I really enjoyed reading this paper, being strong in its methodology and making a valuable contribution to knowledge and practice. Touching on both staffing levels and implementation of guidance domains, it is very topical, but also user-friendly. I would only like to make a couple of observations in support of the discussion of non-significant effects.
I noticed that the sampling size calculations were based on one of the outcomes of interest, bed occupancy. It is not reported whether the study had the power to detect an effect in the other outcome, nurse-patient ratio.This latter effect may have been small, given that on average both arms of the study had on average higher nurse-patient ratios (around 0.70) than the specified standard of 0.50 and also did not differ much at baseline. The wider question is what is the potential margin for improvement and/or have the settings reached their natural limits in terms of achievable staffing ratios?
An even wider question concerns the logic model of this study: (a) why and how would we expect an implementation of guidance intervention increase staffing ratios? The authors already mentioned contextual organisational and HR factors that may have a big influence; (b) would an effect (and of the same magnitude) be expected across the full range of compliance levels? (c) what was the purpose of implementation of guidance: to raise compliance for the units/shifts that were below the standard staffing level, maintain adherence to the standard or improve the level of the compliant units above the standard? While this may have been addressed by the multi-level multivariate statistical analysis, I am not sure we are clear conceptually.
Happy to have any oversight or misunderstanding of mine corrected!
Implementation Science
Observation regarding sampling and the discussion of findings
13 July 2015
I really enjoyed reading this paper, being strong in its methodology and making a valuable contribution to knowledge and practice. Touching on both staffing levels and implementation of guidance domains, it is very topical, but also user-friendly. I would only like to make a couple of observations in support of the discussion of non-significant effects.
I noticed that the sampling size calculations were based on one of the outcomes of interest, bed occupancy. It is not reported whether the study had the power to detect an effect in the other outcome, nurse-patient ratio.This latter effect may have been small, given that on average both arms of the study had on average higher nurse-patient ratios (around 0.70) than the specified standard of 0.50 and also did not differ much at baseline. The wider question is what is the potential margin for improvement and/or have the settings reached their natural limits in terms of achievable staffing ratios?
An even wider question concerns the logic model of this study: (a) why and how would we expect an implementation of guidance intervention increase staffing ratios? The authors already mentioned contextual organisational and HR factors that may have a big influence; (b) would an effect (and of the same magnitude) be expected across the full range of compliance levels? (c) what was the purpose of implementation of guidance: to raise compliance for the units/shifts that were below the standard staffing level, maintain adherence to the standard or improve the level of the compliant units above the standard? While this may have been addressed by the multi-level multivariate statistical analysis, I am not sure we are clear conceptually.
Happy to have any oversight or misunderstanding of mine corrected!
Competing interests
None