Discussion

In this cross-sectional study of 172 RCTs evaluating chlorhexidine for HAI prevention from the last 20 years, we found that trials funded by chlorhexidine sponsors were more likely to report favorable primary results and conclusions compared to trials funded by other sources and unfunded trials. Investigator COI also showed an association with favorable primary results in the univariate analysis but was not statistically significant in multivariate analysis. Industry involvement was also associated with publication in a higher impact factor journal, which has been found previously.Reference Khan, Lombeida, Singh, Spencer and Torralba ²¹ Our findings suggest that industry bias may play a role in the evidence supporting chlorhexidine for HAI prevention. This is the first study, to our knowledge, that assesses the impact of industry involvement on clinical trial outcomes in the field of infection prevention and control.

There are several possible explanations for our findings. First, the chlorhexidine industry has substantial resources. ²² Thus, it is well positioned to fund higher-quality clinical trials with population sizes that provide the necessary statistical power to identify potential differences. Industry is also likely to invest in products with a greater likelihood of being successful, which may explain why industry funded trials in our study were published in journals with a higher impact factor. However, we controlled for sample size when investigating the association between industry funding and favorable outcomes, and the association held, suggesting this was not a contributing factor. Other possible explanations include that financial incentives may lead chlorhexidine sponsors to preferentially publish trials with statistical significant favorable results (‘publication bias’) ^{Reference de Benedictis23,Reference Sismondo24} or selectively sponsor trials they feel are more likely to lead to a statistically favorable result. ^{Reference Lundh, Lexchin, Mintzes, Schroll and Bero25,26} Industry bias due to selective reporting of outcomes ^{Reference Lundh, Lexchin, Mintzes, Schroll and Bero25,26} or other strategic approaches in study design (ie, choice or dosing of comparator drugs) is another possibility that seems likely. Supporting this theory is a previous Cochrane review, which found that industry funding is associated with outcomes favoring industry-sponsored products, suggesting an industry bias that cannot be explained by traditional risk of bias measures. ^{Reference Lundh, Lexchin, Mintzes, Schroll and Bero25}

Industry bias is not detected using traditional risk of bias tools, ^{Reference Bero27} which raises questions about if and how clinical practice guideline (CPG) authors assess for industry bias when evaluating chlorhexidine trials—many of which are cited as evidence in CPGs. Although the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework used by CPGs considers risk of bias and specific factors such as publication bias when assessing a study, ^{Reference Prasad28} GRADE does not incorporate a formal assessment of industry bias. To improve transparency, CPG authors could disclose which studies were industry-sponsored ^{Reference Djulbegovic, Kumar and Miladinovic29} and consider assessing industry bias as a separate risk of bias domain by incorporating specific funding elements of a study to inform this risk, such as if the sponsor had an active role in the study design. ^{Reference Stegeman, Schep, Kuijpers and Hofstede30} Barring these changes, clinicians and other healthcare decision makers should be aware of the challenges in detecting industry bias and the importance of critically reviewing the body of evidence when making decisions that impact clinical care.

Lastly, our finding that approximately one-quarter of studies did not include either an industry funding statement or investigator COI disclosure is concerning and warrants further study. That only 1% of trials lacked a funding statement in a recent study of highly cited clinical trials ^{Reference Siena, Papamanolis, Siebert, Bellomo and Ioannidis31} suggests there may be variations in journal disclosure requirements and enforcement of these policies. Although the International Committee of Medical Journal Editors requires that authors disclose funding sources and potential author COI, ³² our findings suggest that more work is needed to improve transparency.

Our study has several limitations. First, our sample size was small, leading to increased uncertainty in our outcomes. The lack of industry disclosures contributed to this as we excluded approximately one-third of studies in our primary analysis due to studies not disclosing the presence of industry funding or investigator COI. Second, we did not assess for risk of bias, which would allow us to summarize concerns regarding study design, execution, and analysis. Third, we relied on study authors’ self-report of COI, which may have led to underreporting of author COI.

In summary, our findings add to a growing body of literature demonstrating that industry involvement in pharmaceutical trials is associated with favorable outcomes for the sponsored product. Our study is the first to show an association between industry funding and favorable outcomes for chlorhexidine for HAI prevention, raising questions about the strength of evidence for chlorhexidine in healthcare. Further studies are needed to understand the factors underlying this association. Although industry collaboration is necessary to drive innovation, more work is needed to improve our ability to detect industry bias and ensure that clinicians are equipped to make informed decisions regarding the use of chlorhexidine to improve patient outcomes.