To expand on previous works investigating the possibility of funding bias in clinical trials, and to document other factors contributing to biased results in Alzheimer’s research.

Funding institutions, particularly pharmaceutical companies, have been suspected of introducing bias into clinical trials through monetary sponsorship. However, sources of funding may not be the only biasing factor in Alzheimer’s literature.

For this study, we searched PubMed to compile the relevant Alzheimer’s literature. We applied novel clinical data extraction and aggregation technologies developed by MedAware Systems, Inc. This is a patent-pending process where intelligent software compares data extracted by two blinded scientists. Conflicts are checked by a third (senior) scientist. Where pre- and post-treatment outcome data are available, the standardized mean difference is calculated as the MedAware Standardized Index of Treatment Effect (MSI-E). This methodology is used to profile the possible difference in treatment effects in relation to funding sources, publication journal, and author affiliation.

Industry funded studies (0.78±1.21; n=503) reported significantly larger treatment effects when compared to government (0.36±0.49; n=715) and foundation/non-profit (0.36±0.50;n=460) funded publications. Industry funded studies were found to have both increased positive results for active treatments (p < 0.01) compared to all other tested funding sources, and increased negative results for control groups (p < 0.05) compared to federally funded studies. Compared to all other journal, significantly larger overall effect sizes were reported by the Journal of Alzheimer's Disease (p < 0.01) but, when factored by active vs. control group and MSI-E score, several more differences appeared among journals.

Sources of funding may be an important factor to consider when reviewing published literature. In addition to funding source, other factors such as publishing journal and author affiliation can significantly impact the magnitude of treatment effects, and may skew the interpretation of treatment efficacy.