Safety

Last revised: 
2017-03-08

Introduction

Safety is an umbrella term for any unwanted or harmful effects caused by using a health technology. Safety information, balanced with data on effectiveness, forms the basis for further assessment of the technology. (1)

Safety issues can be 

  • direct or indirect (operator or setting dependent, or patient dependent);
  • classified according to fatality or intensity, dose-relatedness or time relatedness;
  • occurring not only in patients or individuals using the technology, but in families and close ones, foetuses, other patients, health care professionals, members of the public or the environment;
  • be previously known or unexpected. (1)

This chapter uses the term adverse effects to be consistent with the literature discussing information-seeking issues within this field. Most of the research findings included in this chapter are for adverse drug effects.

Sources to search

Relying solely on MEDLINE is not recommended, as it is unlikely to be a comprehensive source on adverse effects information (2,3).

A wide range of sources needs to be used for the search to be thorough and in order to provide the best results (4). In a systematic review (3) and a case study of a single drug (4) Golder and Loke identified a combination of sources and techniques that might be expected to provide comprehensive information on adverse effects (in alphabetical order):

  • AHFS First (American Hospital Formulary Service)
  • BIOSIS Previews
  • British Library Direct
  • Conference Papers Index
  • Derwent Drug File
  • Embase
  • Handsearching and reference checking
  • MEDLINE
  • Medscape DrugInfo
  • Science Citation Index
  • Thomson Reuters Integrity
  • Websites and registers of pharmaceutical companies

Golder et al. (5) and Wieseler et al. (6) found that unpublished data such as company clinical trials reports and drug approval information could be valuable sources of adverse effects information.

The HTA Core Model® recommends the following additional sources: product data sheets, national and international safety monitoring systems, disease and technology registers, routinely collected statistics from health care institutions and Internet discussion forums (1).

In a case study on spinal fusion, Golder et al. (7) found that multiple sources need to be searched in order to identify all the relevant studies with safety data for a medical device. The minimum combination of sources in the study was Science Citation Index, Embase, CENTRAL and either MEDLINE or PubMed, in addition to reference checking, contacting authors and using automated current awareness services.

Designing search strategies

In a study conducted in 2012, Golder and Loke found that adverse effects terms were increasingly prevalent in the title, abstract and indexing of adverse effects papers in MEDLINE and Embase (8). They concluded, therefore, that reviewers could, with some caution, choose to use more focused search filters or specific adverse effects terms in their search strategies, rather than run broad non-specific searches (without adverse effects terms), followed by evaluation of large numbers of full-text articles, at least for articles published more recently.

Even though no single published adverse effects search filter has been shown to capture all relevant records, such filters may still be useful in retrieving adverse effects data (9). The purpose of the search, topic under evaluation, resources available and anticipated gain in precision are factors one should take into consideration when applying such filters. Golder and Loke found that adverse effects search filters, when combined with specific adverse effects search terms, could be applied in MEDLINE with an increase in precision without major loss of sensitivity (10). They also found that adverse effects search filters should be applied with caution in Embase as there might be too high a loss of sensitivity without much improvement in precision (10).

Performance measurement of individual search terms included in search filters in MEDLINE and Embase has shown that:

  • Subheadings (‘floating’ subheadings and subheadings attached to an intervention term) provide the highest sensitivity in both databases, and can be particularly useful in identifying papers with adverse effects data in both MEDLINE and Embase (9).
  • Some free text terms for adverse effects in the title and abstract can also be useful, but they should be applied concurrently with other search terms such as subheadings (9).
  • The sensitivity of existing indexing terms for adverse effects seems to be low, particularly in MEDLINE (9). However, high precision of search strategies may be obtained in MEDLINE through use of adverse effects search filters which rely solely on MeSH terms (10).

Studies by Golder et al. (9, 10) provide an overview and comparisons of published search filters. Papers dealing with development of search filters are not included in this SuRe Info chapter, but these can be found at the InterTASC Information Specialists' Sub-Group (ISSG) Search Filter Resource.

However, the currently available adverse effects search filters may not necessarily be useful when searching for adverse effects data of medical devices (11). In a case study, Golder et al. (11) found that the most successful search terms in identifying adverse effects data of medical devices differed from the most successful terms used in search filters for adverse drug effects. The authors emphasize the need to create specific search fiters for adverse effects of medical devices.

Systematic reviews of adverse effects should not be restricted to specific study types (12). Golder et al. found that there was no difference on average between estimates of harm in meta-analyses of RCTs compared to observational studies (12).

Search approaches to identify systematic reviews of adverse effects should be similar to those used to identify primary studies of adverse effects. According to Golder et al. (13) ‘floating’ subheadings provided the highest sensitivity for searching the two major databases of systematic reviews: the Database of Abstracts of Reviews of Effects (DARE) and the Cochrane Database of Systematic Reviews (CDSR). In DARE, MeSH terms achieved the highest level of precision.

Acknowledgement

We acknowledge Carol Lefebvre and David Kaunelis for their work as co-authors of previous versions of the chapter.

Reference list

  1. (1) EUnetHTA Joint Action 2, Work Package 8.  HTA Core Model® version 3.0; 2016 (pdf).
    [Further reference details] [Publication appraisal] [Free Full text]
  2. (2) Golder S, Loke YK. Sources of information on adverse effects. Health Info Libr J 2010;27(3):176-190. [Further reference details] [Publication appraisal] [Free Full text]
  3. (3) Golder S. Optimising the retrieval of information on adverse drug effects. Health Info Libr J 2013;30(4);327-331. [Further reference details] [Publication appraisal] [Free full text]
  4. (4) Golder S, Loke YK. The contribution of different information sources for adverse effects data. Int J Technol Assess Health Care 2012;28(2):133-137. [Further reference details] [Publication appraisal] [Free Full text]
  5. (5) Golder S, Loke YK, Bland M. Unpublished data can be of value in systematic reviews of adverse effects: methodological overview. J Clin Epidemiol 2010;63(10):1071-1081. [Further reference details] [Publication appraisal] [Free full text]
  6. (6) Wieseler B, Wolfram N. McGauran N et al. Completeness of reporting of patient-relevant clinical trial outcomes: comparison of unpublished clinical study reports with publicly available data. PLoS Med 2013;10(10):e1001526. [Further reference details] [Publication appraisal] [Free full text]
  7. (7) Golder S, Wright K, Rodgers M. The contribution of different information sources to identify adverse effects of a medical device: a case study using a systematic review of spinal fusion. Int J Technol Assess Health Care 2014;(30)4:1-7. [Further reference details] [Publication appraisal] [Free full text]
  8. (8) Golder S, Loke YK. Failure or success of electronic search strategies to identify adverse effects data. J Med Libr Assoc 2012;100(2):130-134. [Further reference details] [Publication appraisal] [Free Full text]
  9. (9) Golder S, Loke Y. The performance of adverse effects search filters in MEDLINE and EMBASE. Health Info Libr J 2012;29(2):141-151. [Further reference details] [Publication appraisal] [Free Full text]
  10. (10) Golder S, Loke YK. Sensitivity and precision of adverse effects search filters in MEDLINE and EMBASE: a case study of fractures with thiazolidinediones. Health Info Libr J 2012;29(1):28-38.
    [Further reference details] [Publication appraisal] [Free Full text]
  11. (11) Golder S, Wright K, Rodgers M. Failure or success of search strategies to identify adverse effects of medical devices: a feasibility study using a systematic review. Syst Rev 2014;3:113.
    [Further reference details] [Publication appraisal] [Free full text]
  12. (12) Golder S, Loke YK, Bland M. Meta-analyses of adverse effects data derived from randomised controlled trials as compared to observational studies: methodological overview. PLoS Med 2011;8(5):e1001026.
    [Further reference details] [Publication appraisal] [Free Full text]
  13. (13) Golder S, McIntosh HM, Loke Y. Identifying systematic reviews of the adverse effects of health care interventions. BMC Med Res Methodol 2006;6:22. [Further reference details] [Publication appraisal] [Free Full text]