Microsoft word - v - papas summary of findings document

Cochrane Pain, Palliative & Supportive Care Review Group
for the Pain, Palliative, and Supportive Care (PaPaS) Review Group A ‘Summary of findings’ (SoF) table is intended to provide key information concerning the quality of evidence, the magnitude of effect of interventions examined, and the sum of available data on all the important outcomes for a given comparison in a Cochrane (or other) review. SoF tables are typically constructed to satisfy interventions with a limited effect in a large population, predominantly in prophylactic therapy settings in which there is a generalised initial risk for the population, but whereby no means all would suffer without treatment (aspirin or statin to prevent heart attack would be examples). Pain, palliative, and supportive care are different, because in almost all circumstances every single patient has a problem initially (pain, for example), and therapy removes or at least decreases the pain in some who are treated; the initial risk is one in one, or 100%. Moreover, developments in understanding what patients want from treatment, and what clinical trials tell us, has moved decisively towards a description of response – where response is variably described, but usually a major change in an outcome like pain brings with it concomitant improvements in other areas, like sleep, fatigue, depression, functioning, and quality of life. The handbook advises that we consider the GRADE system, and GRADEpro for the production of Summary of Findings Tables. We are working with the GRADE working Group on an innovations fund project in 2012 to 2013 to consider how GRADE can evolve to take account of many of the issues pertinent to PaPaS reviews. We recognize that the Methodological expectations of Cochrane Intervention Reviews (MECIR) process is also reviewing all aspects of methods guidance, and is likely to influence future PaPaS standards. In the meantime, we recognize the importance of SoF and have developed outline tables that capture the essence of communicating complex results readily by focussing on the important outcomes of benefit and harm. Version Control
This is version 1.0 of this document and was agreed as correct on September 1st 2011. It is scheduled for review in August 2012. Version 1.1 - Last updated January 2012 Patients or population:_________________________________________ Setting:_____________________________________________________ Intervention:_________________________________________________ Comparison:_________________________________________________ Note that not all these outcomes may be available in every review. For some reviews, where an intervention may have been tested in different painful conditions, it may be appropriate for one or more efficacy outcomes to be presented for each condition. Version 1.1 - Last updated January 2012 Assessing the quality of evidence across studies for an outcome (notes) Systematic reviews typically assess quality and validity according to individual studies. For the summary of findings table, an assessment of overall quality is required for each outcome chosen for inclusion in the summary of findings table. This aide memoir is designed to help in making those assessments. The table is limited to seven outcomes for one comparison, but can be changed in terms of the number of outcomes and the number of comparisons, consistent with requirements of the Cochrane Review Group (CRG). The following notes are intended to help guide the completion of the form. 1. Number of studies and participants: the number reporting particular outcomes is likely to vary. Results from only a limited number of studies and participants might tend to be more subject to bias. 2. Number of events: an event is a beneficial or harmful outcome in one patient; with fewer than 50 events, random-effects can be very high, and with over 200 events random-effects will be limited. 3. Quality score: studies with an Oxford Quality Score of 3/5 or above have been shown to be unbiased. The score includes items describing randomisation, blinding and withdrawal, and their appropriateness. Where all studies with the outcome have scores of three or more, the likelihood of bias is minimal. 4. Appropriate study duration: chronic pain studies are known to demonstrate less efficacy over longer periods. Studies with the outcome of two weeks or less are likely to confer bias; this will be less in those of two to six weeks, and minimal in those of eight weeks or more. In other circumstances authors will need to be sure that outcomes are reported appropriately – for acute migraine, for example, are adverse events collected over 10 days for an acute attack relevant and appropriate? 5. Imputation method appropriate: this section includes issues over intention to treat (ITT), and how missing data points are dealt with. Completer analyses are not ITT, and tend to overestimate treatment effects. Last observation carried forward (LOCF) produces more favourable results than baseline observation carried forward (BOCF). Might these be factors for any outcome? 6. Potential threat from publication bias: this is a function of size and magnitude of the result for any outcome. Have you performed a publication bias threat for the outcome (see Guidelines on Reviews in Pain document)? 7. Other limitations: are there any other limitations affecting the evidence for the outcome? 8. Overall assessment of the quality of evidence: here you have to make an overall assessment, based on the questions already answered. For simplicity please use high, medium, or low. Version 1.1 - Last updated January 2012 Assessing the quality of evidence across studies for an outcome (table) Comparison: __________________________________________________ Quality assessment (please complete for each individual outcome) Summary of findings table – pregabalin chronic pain example Patients or population: Postherpetic neuralgia – moderate or severe pain much or very much improved Quality of life Note that not all these outcomes may be available in every review. For some reviews, where an intervention may have been tested in different painful conditions, it may be appropriate for one or more efficacy outcomes to be presented for each condition. Summary of findings table – aspirin for acute migraine example Patients or population: Migraine (IHS), moderate to severe pain intensity, adults Summary of findings table – ibuprofen in acute pain example Patients or population: Acute postoperative pain formulations. NNTs significantly better for dental pain and soluble formulations formulations. NNT significantly better for soluble formulations and formulations. Data mostly over 4 to 8 h Note that not all these outcomes may be available in every review. For some reviews, where an intervention may have been tested in different painful conditions, it may be appropriate for one or more efficacy outcomes to be presented for each condition.

Source: http://papas.cochrane.org/sites/papas.cochrane.org/files/uploads/V%20-%20PaPaS%20Summary%20of%20Findings%20document.pdf