Article type
Abstract
Background. The process of Health Technology assessment (HTA) typically involves a systematic review, including a bibliographic search performed by information specialists as a crucial aspect of information retrieval. With the increasing use of Technology-Based tools (TBTs) to support this process, it becomes important to identify their potential strengths and limitations.
Objective. To compare a TBT-assisted method to a conventional method (CM) used in searches aimed at retrieving randomized controlled trials (RCTs) for HTA projects.
Methods. The comparison between the two search methods to achieve information retrieval was carried out in the context of two HTA projects, one assessing outpatient cardiac rehabilitation and the other, genicular embolization for knee osteoarthritis. The CM involved the development of a general search strategy designed and adapted to CENTRAL-Wiley, MEDLINE-Ovid, Embase-Elsevier and CINAHL-EBSCOhost databases by an HTA information specialist. An specific report by database, and a file ready for the screening process were provided. The TBT-assisted method involved a technological solution developed by the Epistemonikos team, incorporating artificial intelligence, with an output consisted of a single set of search results, including CDSR, PubMed, MEDLINE, CINAHL, PsycINFO databases and Epistemonikos, and a single search report.
Results. The conventional method retrieved 1,720 RCTs for one of the HTA projects and 126 for the other. The TBT-assisted method retrieved 1,798 and 28 RCTs, respectively. All RCTs included after the screening process were identified by both methods. While the CM included full search strategies conducted in each database, the TBT-assisted method included overall search strategies.
Conclusion. The adequacy of documentation regarding search methods is a key factor in meeting the quality standards of HTA. While TBTs, such as the one compared, aid in identifying available evidence, it is essential that detailed reports of the specific searches performed are provided. Transparency ensures systematicity, and also allows their reuse in the development of future HTA projects. Reproducibility enables others to validate findings, build on previous work and update results. The conventional search method ensures transparency, emphasizing a commitment to the scientific principles and contributing to the overall integrity of HTA. TBT-assisted methods must enhance transparency to be deemed a viable alternative.
Objective. To compare a TBT-assisted method to a conventional method (CM) used in searches aimed at retrieving randomized controlled trials (RCTs) for HTA projects.
Methods. The comparison between the two search methods to achieve information retrieval was carried out in the context of two HTA projects, one assessing outpatient cardiac rehabilitation and the other, genicular embolization for knee osteoarthritis. The CM involved the development of a general search strategy designed and adapted to CENTRAL-Wiley, MEDLINE-Ovid, Embase-Elsevier and CINAHL-EBSCOhost databases by an HTA information specialist. An specific report by database, and a file ready for the screening process were provided. The TBT-assisted method involved a technological solution developed by the Epistemonikos team, incorporating artificial intelligence, with an output consisted of a single set of search results, including CDSR, PubMed, MEDLINE, CINAHL, PsycINFO databases and Epistemonikos, and a single search report.
Results. The conventional method retrieved 1,720 RCTs for one of the HTA projects and 126 for the other. The TBT-assisted method retrieved 1,798 and 28 RCTs, respectively. All RCTs included after the screening process were identified by both methods. While the CM included full search strategies conducted in each database, the TBT-assisted method included overall search strategies.
Conclusion. The adequacy of documentation regarding search methods is a key factor in meeting the quality standards of HTA. While TBTs, such as the one compared, aid in identifying available evidence, it is essential that detailed reports of the specific searches performed are provided. Transparency ensures systematicity, and also allows their reuse in the development of future HTA projects. Reproducibility enables others to validate findings, build on previous work and update results. The conventional search method ensures transparency, emphasizing a commitment to the scientific principles and contributing to the overall integrity of HTA. TBT-assisted methods must enhance transparency to be deemed a viable alternative.