advertisement

The Cochrane Collaboration Colloquium: Summing up: challenges and possible solutions when summarising the findings of Cochrane reviews

75 %
25 %
advertisement
Information about The Cochrane Collaboration Colloquium: Summing up: challenges and...

Published on November 9, 2007

Author: Cochrane.Collaboration

Source: slideshare.net

Description

Claire Glenton speaking at the XIV Cochrane Colloquium in Dublin, Ireland
advertisement

Plenary 1 Summing up: challenges and possible solutions when summarising the findings of Cochrane reviews Claire Glenton Pembroke

Summing up: challenges and possible solutions when summarising the findings of Cochrane reviews Claire Glenton, Nordic Cochrane Centre, Norwegian branch/ Norwegian Knowledge Centre for the Health Services

Past achievements, future challenges Cochrane Library: 4410 Cochrane reviews and protocols Reputation for reliability Our next challenge: to improve the accessibility of this information

Cochrane Library: 4410 Cochrane reviews and protocols

Reputation for reliability

Our next challenge: to improve the accessibility of this information

Improving access by summarising results Cochrane reviews already summarised in: journals and commercial websites HTAs and guidelines Cochrane products the media

Cochrane reviews already summarised in:

journals and commercial websites

HTAs and guidelines

Cochrane products

the media

The Cochrane Collaboration can improve understanding and avoid the possibility of misrepresentation.. ..by developing its own summaries that are: Precise Consistent Understandable

..by developing its own summaries that are:

Precise

Consistent

Understandable

Experiences when summarising Cochrane reviews Norwegian branch, Nordic Cochrane Centre The GRADE Working Group The Cochrane Back Group The Cochrane Musculoskeletal Group The Cochrane Consumer Network

Norwegian branch, Nordic Cochrane Centre

The GRADE Working Group

The Cochrane Back Group

The Cochrane Musculoskeletal Group

The Cochrane Consumer Network

We have tried to extract and consistently present what the reviews say about: what is known about benefits and harms what is not known about benefits and harms how sure we can be about this information

what is known about benefits and harms

what is not known about benefits and harms

how sure we can be about this information

Challenges we faced when trying to answer these questions Missing information about important outcomes Continuous outcomes that are difficult to interpret Variations in how effect and quality are presented

Missing information about important outcomes

Continuous outcomes that are difficult to interpret

Variations in how effect and quality are presented

Challenges 1. Missing information about important outcomes Often, data for important outcomes is missing or it is not significant 1 Data on adverse effects sometimes missing: not always looked for in reviews 1 not always found in individual trials 1 When data is missing: important to convey this lack of information rather than present secondary outcomes 1 Vist 2003

Often, data for important outcomes is missing or it is not significant 1

Data on adverse effects sometimes missing:

not always looked for in reviews 1

not always found in individual trials 1

When data is missing: important to convey this lack of information rather than present secondary outcomes

Challenges 2. Continuous outcomes often difficult to interpret ” 10 more points on the XY-Z scale ” Hard to understand unfamiliar scales Information is sometimes missing about scale range 1 1 Kho, Glenton, Pennick, Koy, Underland, Vist, Oxman, 2003

” 10 more points on the XY-Z scale ”

Hard to understand unfamiliar scales

Information is sometimes missing about scale range 1

Challenges 3. Effect presented in different ways Risk difference Odds ratio Numbers Needed to Treat Qualitative presentations Relative risk reduction Relative risk improvement Weighted Mean Difference Percentages Standard mean difference Risk ratio

Challenges 3 . Quality of evidence evaluated in different ways Quality of each individual study Quality of all studies together Quality of each individual outcome 1,2 1 Vist 2003 2 Kho, Glenton, Pennick, Koy, Underland, Vist, Oxman, 2003

Quality of each individual study

Quality of all studies together

Quality of each individual outcome 1,2

what is known what is not known how sure we can be How can we give a precise, consistent and understandable presentation of: ?

what is known

what is not known

how sure we can be

Solution: Summary of Findings Tables (SoF)

SoF Tables Focus on important outcomes only Authors encouraged to include: Outcomes with no data or non-significant data Outcomes tied to adverse effects

Authors encouraged to include:

Outcomes with no data or non-significant data

Outcomes tied to adverse effects

SoF Tables Present effect consistently OR RD RRR RRI Risk Ratio and Absolute Risk (# out of 100) RD % NNT

OR

RD

RRR

RRI Risk Ratio and Absolute Risk (# out of 100)

RD

%

NNT

SoF Tables Continuous outcomes still a challenge WMD Risk Ratio and Absolute Risk? SMD Information about scales for continuous outcomes

WMD

Risk Ratio and Absolute Risk?

SMD



Information about scales for continuous outcomes

SoF Tables Present quality consistently Quality for each outcome evaluated using the GRADE system

Quality for each outcome evaluated using the GRADE system

SoF Tables are precise and consistent overviews May still be difficult to interpret for many target groups, including: patients and consumers journalists other “non-statisticians” Improving accessibility to lay people

SoF Tables are precise and consistent overviews

May still be difficult to interpret for many target groups, including:

patients and consumers

journalists

other “non-statisticians”

While statistics may be precise ...they are often hard to understand Different statistical presentations also affect the decisions that people make 1 (risk ratios and absolute risk) 1 Moxey, O’Connell, McGettigan, Henry, 2003

...they are often hard to understand

Different statistical presentations also affect the decisions that people make 1 (risk ratios and absolute risk)

Qualitative presentations not the solution Current lay summaries often rely on qualitative statements 1 ” Medication is likely to lead to less pain” Qualitative presentations are interpreted differently by different people 2 1 Allen, Smith, Thornton 2001 2 Wills, Holmes-Rovner, 2003

Current lay summaries often rely on qualitative statements 1

” Medication is likely to lead to less pain”

Qualitative presentations are interpreted differently by different people 2

Event rates (“10 people out of 100”) More precise than qualitative statements Easier to understand than probabilities 3 Easy to extract from SoF Tables Have their own shortcomings imply high level of precision mask results that are not statistically significant 1 Schünemann, Herrin, Vist, Oxman, Akl, 2003

More precise than qualitative statements

Easier to understand than probabilities 3

Easy to extract from SoF Tables

Have their own shortcomings

imply high level of precision

mask results that are not statistically significant

Lay version of SoF Tables Event rates Standardised comments when data not statistically significant

Continuous outcomes: problems still remain For continuous outcomes that are hard to understand - standardised qualitative statements

For continuous outcomes that are hard to understand - standardised qualitative statements

Tables as a basis for Plain Language Summaries Description of the treatment and who the results apply to Qualitative description of effect Lay version of Summary of Findings Table

Description of the treatment and who the results apply to

Qualitative description of effect

Lay version of Summary of Findings Table

Future plans SoF Tables software available with RevMan 5 Continued testing and improvement of format and content of Tables

SoF Tables software available with RevMan 5

Continued testing and improvement of format and content of Tables

Conclusion: All summaries involve judgement and can lead to loss of precision, but... ... better to do this ourselves in a systematic and explicit fashion so that we can support decision makers in the best possible way

... better to do this ourselves in a systematic and explicit fashion so that we can support decision makers in the best possible way

Add a comment

Related pages

Videos: XIV Cochrane Colloquium, Plenary Sessions - Dublin ...

... Summing up: challenges and possible solutions when summarising the findings of Cochrane reviews ... between The Cochrane Collaboration and ...
Read more

Cochrane Reviews | LinkedIn

View 274 Cochrane Reviews posts, presentations, experts, and more. Get the professional knowledge you need on LinkedIn.
Read more

All book reviews from the Change Management Monitor ...

Change management blogs, ... that offered insightful reviews of books related to change management. ... Frame possible solutions;
Read more

ELearning Healthprof Report - scribd.com

augment capacities to scale up production, ... that will inform a series of Cochrane systematic reviews ... Cochrane Collaboration.1
Read more

Deriving Process-driven Collaborative Editing Pattern from ...

Sign Up; We're trying Google ... Deriving Process-driven Collaborative Editing Pattern from Collaborative Learning Flow Patterns. Authors. Pascal Molli + 6 ...
Read more