Grant writing is a fundamental skill for every clinical researcher who seeks funding to conduct trials and advance medical care. Most grant reviewers have a limited amount of time to sift through a large number of proposals, so grants that are poorly written, disorganized or missing important information are likely to garner only a brief glance before being passed over. One way to guarantee your grant stands out compared to the rest is to make sure the statistical component of your study is well designed, complete, and clearly stated.
Bruce Thompson, PhD, is president and chairman of the board of Clinical Trials and Surveys, a clinical trials solutions company. As a statistician who reviews grants and contracts, these are the issues Dr. Thompson likes to see addressed by every investigator.
Clearly articulate the hypothesis and limit the number of statistical tests. When it comes to developing a hypothesis and designing a test for that hypothesis, simple is better. Your hypothesis should be focused, fully conceptualized and clearly stated – not weighed down by jargon. It should be worded so that even someone with no knowledge of the topic is able to understand the potential significance of your study.
When it comes to testing that hypothesis, there should be a single primary test or a few tests with alpha-level control. Running multiple tests on a large number of overly ambitious hypotheses will diminish the feasibility of your study.
Fully describe the methods of data collection and analysis. Make sure your analytical design is fully developed and explained in detail. Identify any potential biases and how you will account for them. Many designs are discounted by reviewers because proper precautions have not been taken to avoid investigator bias. If running a clinical trial, the investigator collecting the data should preferably be unaware of which subjects receive therapy and which do not. If a laboratory apparatus will produce the primary endpoint measurement, it is important that you have steps in your proposal to identify that the machine is routinely calibrated so data are collected from each subject under the same conditions. If your endpoint requires evaluations from an expert evaluator, make sure the individual is blinded to who is in the treated and untreated groups.
Present a proper sample size and discuss how comparison groups will be established. Your sample size must be large enough to ensure there will be reasonable power to detect clinically significant differences and reach your primary endpoint. Make sure you clearly define your sample population and justify your criteria for inclusion or exclusion. Explain your system for randomly assigning subjects to the different comparison groups. Always include a discussion of informed consent – including which aspects of your data collection will require consent and provide a copy of the informed consent form you intend to use.
Be sure that you address how you will account for subjects that drop out during the study as well as subjects that change their treatment during the study. These two aspects of clinical trials are frequently overlooked and have a substantial impact on the power of your proposed analytical design. Thorough designs also have control measures in place to allow investigators to un-blind subjects in an emergency while still protecting the overall masking of the study.
Implement performance and quality monitoring plans. As data come in from your study, you must plan to generate periodic monitoring reports to assess the performance of your protocol and identify what adjustments, if any, are needed to improve the protocol from a performance perspective. Submitting duplicate samples from your subjects to laboratories will test the reliability of specimens being submitted to your study’s core laboratory (ies). Systematically reviewing documents will ensure data are being correctly recorded into your database. Plan to monitor your design’s safety and efficacy profile at pre-specified intervals throughout your study. Incorporate statistical methods and independent review boards to detect risks associated with your study and always be prepared to weigh the risks against the benefits of your study.
Bruce Thompson, PhD, is President and Chairman of the Board of Clinical Trials and Surveys Corp. (C-TASC), a clinical trials solutions company that supports best practices management of clinical trials, clinical cohort studies, case/control studies, clinical registries and laboratory studies. Dr. Thompson teaches young investigators statistical methods and techniques to write grants that will result in funded studies. He has taught clinical study grant writing with the Memorial Hospital in Fort Lauderdale and the American Society of Hematology.