VP31 Searching Non-English Literature For HTA Reports May Be Unnecessary

Abstract

IntroductionCurrently, the Institute for Quality and Efficiency in Health Care (IQWiG) does not restrict literature searches by language. Given limited resources, it is unclear whether the effort put into screening and translating studies published in non-English and non-German (nEnG) languages yields much new information when compared to including only English and German literature. Therefore, we aimed to analyze the impact of nEnG literature on the conclusion of IQWiG's health technology assessments (HTAs).MethodsWe checked for seventy-two IQWiG HTAs (all non-drug intervention HTAs published until August 2018 and three additional HTAs on drugs) whether they included nEnG studies. For all HTAs including at least one nEnG study, we analyzed whether the statistical significance would have changed for any endpoint without the respective nEnG study(ies). If no endpoint was impacted by a nEnG study, we classified the study as non-relevant to the HTA's conclusion and specified a reason for this.ResultsOf seventy-two HTAs, twenty-nine (40 percent) included a total of eighty-three nEnG publications). Three HTAs were impacted by the inclusion of altogether seven Chinese publications. For one HTA on systemic therapy, five endpoints’ conclusions were changed; for the other two HTAs, the statistical significance would have changed for one endpoint each. The remaining seventy-six publications (included in sixty-nine HTAs) were judged as non-relevant to the HTA's conclusion, the most prominent reason being “meta-analysis would have had the same result without respective study” (44 percent of nEnG publications).ConclusionsOnly three of seventy-two HTAs (4 percent) were impacted by nEnG publications, the changes being minimal for two of these. When faced with limited time or personnel resources, searching only for English and German publications may be sufficient, especially when generalizability issues are a possible concern.