The development of search filters for the adverse effects of medical devices in MEDLINE and Embase

Session: 

Oral session: Searching and information retrieval (3)

Date: 

Tuesday 18 September 2018 - 14:20 to 14:30

Location: 

All authors in correct order:

Golder S1, Farrah K2, Mierzwinski-Urban M3, Wright K4, Loke Y1
1 Cochrane Adverse Effects Methods Group, UK
2 CADTH, Canada
3 CADTH, Canada
4 CRD, UK
Presenting author and contact person

Presenting author:

Su Golder

Contact person:

Abstract text
Background:
Systematic reviews should include search strategies that aim to identify as many relevant papers as possible. However, searching for information on adverse effects is challenging, not least because adverse effects are often secondary or even tertiary outcomes and inconsistent terminology used. Research indicates that authors and indexers are increasingly including terms for adverse drug effects in their titles, abstracts or indexing of database records such as MEDLINE and Embase. However, it is not clear if this is the same for studies with adverse effects data for medical devices.

Objective:
To develop and validate search filters for MEDLINE and Embase for the adverse effects of medical devices.

Methods:
We identified systematic reviews, published from January 2016 to June 2017 in Epistemonikos and the Health Technology Assessment (HTA) database, which evaluated the adverse effects of a medical device (either as a primary or secondary outcome). We employed the WHO definition of a medical device. The included studies within these reviews that reported on adverse effects of medical devices formed our cohort of papers for analysis. We randomly divided the database records in MEDLINE into three test sets and one validation set of records and then repeated this process in Embase. Using word frequency analysis from one test set, we constructed a sensitivity-maximising search strategy. We tested this strategy and refined it using the other two test sets, then validated it using the final set.

Results:
Two researchers independently screened 6433 systematic reviews from which we included 186. From these, 1984 unique included studies were available from MEDLINE and 1986 unique included studies were available from Embase. MEDLINE test sets and the validation set each contained 496 records. The three test sets in Embase contained 496 records and the validation set 498.

Conclusions:
We will present the search terms that maximised the sensitivity of retrieval for articles on the adverse effects of medical devices. Further, we will discuss the frequency of indexing and author keyword use in the literature on the adverse effects of devices in comparison to studies on drug and surgical interventions.

Patient or healthcare consumer involvement:
The methodological nature of this research means that Information Specialists are our consumers.

Relevance to patients and consumers: 

Adverse effects are important patient outcomes. Whilst much research has been undertaken on the adverse effects of drug interventions, other types of interventions, such as medical devices, can cause harm. Identifying information on the adverse effects of medical devices can be challenging. We therefore undertook to devise and validate a search filter (combination of search terms) that would maximise sensitivity in identifying the adverse effects of medical devices. This will help searchers identify important papers on the harms of medical devices, enabling less biased systematic reviews and thus provide more balanced information to patients.