Challenges in applying PLS guidance for diagnostic test accuracy reviews

Session: 

Oral session: Understanding and using evidence (1)

Date: 

Monday 17 September 2018 - 11:20 to 11:40

Location: 

Carrick 3

All authors in correct order:

Davenport C1, Dinnes J2, Williams H3, Matin R4, Ferrante di Ruffano L1, Deeks J1, Godfrey K5, O'Sullivan C5
1 University of Birmingham, UK
2 University of Birmngham, UK
3 University of Nottingham, UK
4 Oxford University NHS Foundation Trust, UK
5 Cochrane Skin Cancer Group, UK
Presenting author and contact person

Presenting author:

Clare Davenport

Contact person:

Clare Davenport
Abstract text
Background:
Plain language summaries (PLS) are standalone documents, which aim to communicate the scope, results and relevance of reviews for consumers. Guidance for writing PLS for diagnostic test accuracy (DTA) reviews has recently been developed and experience with its application in practice is now required.

Objectives:
- To identify problems arising from the application of current guidance for writing PLS for DTA reviews.
- Using experiential learning to recommend modifications to current PLS guidance and identify areas requiring further research.

Methods:
We applied the new DTA PLS guidance to a series of 11 reviews including over 200 studies concerning the diagnosis of the three commonest skin cancers with varying prognosis and implications for patients following a missed diagnosis. Eight diagnostic tests were evaluated when used alone or in combination in different healthcare settings. One research team (including methodologists, clinicians and patients) was involved in writing the PLS. A retrospective interrogation of the PLS development led to analysis of problems in applying the current DTA PLS guidance and suggestions for modifications.

Results:
Difficulties emerging as part of writing a standalone PLS document can be grouped as follows:

- Communicating the impact of review results for the same test being used in patients for three different types of skin cancer for whom the risk of a missed diagnosis can be very different.
- Communicating risk of bias and applicability of the evidence base.
- Communicating variation in test use (for example, a test used alone or in combination with another and in which healthcare setting).
- The potential to mislead readers by reporting results based on poor evidence.

We also observed a shift in writing style from 'lay language' (ensuring understanding of terminology) to 'lay style' (directing information at patients who might be offered a test in a specific healthcare setting).

Conclusions:
Accommodating multiple patient groups, test combinations and variable quality of evidence poses significant challenges when writing PLS. A standalone PLS for all potential users may not be feasible and at worst misleading.

Patient or healthcare consumer involvement:
Clinicians and patients were involved in the conduct and reporting of the 11 reviews including the PLS.

Relevance to patients and consumers: 

Whilst undertaking a series of 11 reviews concerned with diagnosis of the 3 commonest skin cancers, problems applying recently developed plain language summary (PLS) guidance were identified by review authors including patient representatives. Problems included communicating accuracy of the same test across 3 cancers with different implications for patients following a missed diagnosis and the impact of limitations of the available evidence on conclusions. Further piloting of the PLS Guidance in reviews with complex conclusions is necessary.A standalone PLS for all patient groups in a review may not be feasible and potentially misleading.