Grant Writing: How To Get a Grant or Fellowship

• Scientific merit, based on peer review
• Program priorities of the NIH and individual Institutes and Centers (ICs)
• Current research being supported
• Availability of funds/NIH budget

General principles

• Think
• Communicate
• Take Time

Outline

• Developing an Application: Resources & General Guidelines
• The Research Plan
• The Budget and Other Common Pitfalls
• The Summary Statement
• Resubmitted Applications
• Tips
• The NIH Budget

Develop an application:

NIH resources

• Understand the funding priorities of NIH and the relevant ICs
• Check the relevant IC websites for program priorities
• Call the appropriate program director(s) for advice and guidance
• Search the NIH Guide (and have it automatically emailed to you every week) for special funding opportunities
• Check what is already being funded in your area of interest:
• Check the Computer Retrieval of Information on Scientific Projects (CRISP) website: projectreporter.nih.gov/reporter.cfm
• Check the NIH Office of Extramural Research Award Information website grants.nih.gov/grants/award/award.htm

Develop an application: Guidelines

• Good idea
• Clear, testable, potentially important and focused set of hypotheses to test
• Team with all the expertise you need
• Sound Design
• Feasible
• Links proposed questions and analysis plan
• Ensure you have the appropriate resources, facilities and patient population available

Research plan:

Specific aims and hypotheses

• Start with a good idea FOR A GOOD REASON
• Present testable and compelling aims/hypotheses
• Be clear, concise, complete
• Work from a plausible model

Background and preliminary data

• Make the case that this is exactly the right study, exactly the right time, with the right outcomes
• Don't fudge
• If the data are not there or are not strong, do not try to make the case
• Enthusiasm and your own conviction are nice but will not make the case for you
• Empirical evidence is essential

Outcomes

• Feasible (i.e., measurable)
• Relevant to the questions you are asking
• Linked to the disease/condition you are interested in AND the intervention you are testing
• Not only relevant to health but also important to health
• Specific outcome variable
• How is the outcome variable measured?
• Reliability and validity of measure

⇒ While a shortened patient reported outcome instrument may reduce participant burden, the psychometric properties of the instrument may not have been tested.

Blinding and random allocation

• Clarify the blinded personnel
• Endpoint assessment
• Who allocates people to a study group?
• Exactly what mechanism are they using?
• Stratifying or balancing or matching?
• Data verified before assignment
• Lots of programming and oversight time
• When exactly are study participants randomized/allocated?

Sample size

• Specific analyses to answer each hypothesis
• Appropriate to study design
• Multiple comparisons
• Independent outcomes?
• Clarity
• Why not use a 2-sided test?
• Attrition/drop-outs/missing data
• Longitudinal/repeated measures
• Interim analyses of the outcome data

A few study design issues

• Choosing the right dose
• Designing the right control group(s)

What is a dose?

• Number of sessions/pills/treatments
• Frequency
• Length of sessions (each, total treatment time)
• Amount of practice
• Practitioner or teacher
• Contact time
• Who that person is
• Many different combinations

What is the right control group?

• Waitlist control
• Placebo control
• Remove 'active' components
• May still have an impact on outcomes
• May be called Attention control
• Active controls
• Standard of care
• Alternative intervention
• Fewer parts of multi-part intervention
• Different study population (eg, case-control)
• No universally-appropriate control group

Analysis plan

• Appropriate analyses for the study design
• If it needs to be a two sample t-test, t-test
• If it needs to be fancy, fancy
• Reconsider repeated measures ANOVA
• Consider linear mixed models or generalized estimating equations
• Drop outs/missing data in the analyses
• Detail the interim analysis plan if interim analyses are anticipated (they may not be)
• Work with a statistician and/or epidemiologist

Analysis follows design

Questions → Hypotheses →
Experimental Design → Samples →
Data → Analyses → Conclusions

Protection of human subjects

• Data quality
• Safety
• Adverse events
• Does data and safety monitoring plan match the intervention and study population?

General guidelines:
Make sure your application is…

• Clear and precise
• Easy to read
• Free of typos
• In compliance with font guidelines and all other requirements detailed in the announcement
• Consistent throughout
• Makes critical points more than once in slightly different words
• Does not require the reviewers to:
• Flip between sections of the application
• Make assumptions about what you will do

The Budget and Other Common Pitfalls

Problems and common pitfalls

• Formulating the question
• Making the case
• Design/analysis problems
• Scope/ambitious nature of the research
• Timing
• Proposed research is right for the state of the science
• Feasibility
• Access to patient populations, technology, expertise, other resources, time

Formulating the question

• Good
• Logical next step in research
• Preliminary health services or epidemiologic research
• Clear, biologically plausible model
• Bad
• Does not fit current knowledge
• Specific aims ≠ Public Health Importance ≠ Study Design ≠ Analysis ≠ Endpoints

Making the case

• Good
• 5 million people with this condition turn to this CAM therapy. It costs approximately $500 out of pocket.
• Bad
• No case made

Design and analysis problems

• Would XYZ be a more appropriate study design to answer the study aims?
• Toxicity, tolerability, some efficacy
• Pharmacodynamics, pharmacokinetics
• Phase I, II, or an observational study
• Which adaptive design?
• Not everything needs to be on the Phase III design framework
• Teach and implement a wide variety of question-appropriate designs

Timing

• If ambitious, show it can be done
• Show you have done it before, successfully
• Back up plans
• Preparation of documents (IRBs, SRCs, IACUC, FDA, NIH, …)
• Training staff
• Recruitment is slow
• Analyzing samples and data

Feasibility

• Part of the timing
• Access to Resources
• Facilities
• Staff and expertise
• Study participants
• Laboratory needs, sample storage
• Computer databases, electronic records
• Technology

Developing a budget

• Do you really need to think about a budget before you get the grant?
• How does a budget help you?
• What NOT to do (and include)
• What NOT to forget to do (and include)
• How to get help

The Summary Statement
Resubmitted Applications

The summary statement

• Read it carefully
• Discuss with the assigned NIH Program official
• If not likely to be funded, are the identified flaws readily correctable?
• Seek advice from mentors & colleagues
• If you really believe that what you proposed to do was optimal, can you make your case (more) clearly?
• Respect the reviewers' comments even if you disagree

NIH peer review:

Priority scores

• 100-150: Outstanding
• 150-200: Excellent
• 200-250: Very Good
• 250-350: Good
• 350-500: Acceptable

If you don't receive a fundable score the first time…

• Two resubmissions are allowed for funding announcements with multiple receipt dates
• There is no time limit
• If the funding announcement expires and is not replaced, no resubmission is possible.
• New applications must be substantially (at least 50%) different

Special issues when resubmitting applications

• Should you resubmit?
• Deadlines
• What is expected
• Impact on score
• How to respond to the Summary Statement

The resubmitted application:

Introduction

• Introduction is for re-submitted (amended) applications ONLY
• Thank reviewers for their helpful comments
• Respond briefly and tactfully where necessary
• Summarize changes VERY briefly

Application

• Signify changes UNOBTRUSIVELY in the text of the application

Tips

Mock study sections

• Set up a mock study section with investigators from your institution. Send them your draft and any prior feedback you have received
• Include on your mock study section people
• Who have been on NIH study sections
• Familiar with the area of study and good mentors
• A clinical researcher familiar with the biological/medical aspects of the condition you are studying
• A biostatistician, preferably someone who has been on a study section
• Get feedback from your mock study section by meeting as a group, one-on-one, telephone, or email

Look at the CSR Web site

• Become familiar with the Center for Scientific Review (CSR) website www.csr.nih.gov
• Read the Guidelines for Reviewers www.csr.nih.gov/Guidelines/revguide.htm and others, available at CSR
• Read CSR's "What happens to your grant application: A primer for new applicants"
• Watch the video on CSR review
• Research where your application is likely to be reviewed www.csr.nih.gov/committees/rosterindex.asp

Three fundamental drivers of the NIH budget

• Large capacity-building throughout U.S. research institutions and dramatic increase in number of tenure-track faculty
• Large increase in applicants and applications occurring after 2003
• Budgets:
• After 2003, appropriations below inflation (inflation in 2004 was ~ 5.5%)
• 4-year budget cycling phenomenon

The NIH Budget

Availability of funds across the NIH

Why is funding tight now?

• Demand for grants "took off" (2002–2006) just as the NIH budget was "landing" (1999–2003)
• Greater increase in demand for grants in the last few years than in the full 5 years of the doubling
• NIH budget increases below inflation
• Budget cycle effects — commitments from the doubling years "tie up" funds for new grants
• 2003 commitments became available in 2007, allowing for a very small (3%) increase in competing grants (even with a flat budget)

More resources

• Grants and Funding
• How Your Application Is Reviewed
• Ten Ways to Write a Better Grant (TheScientist)
• Grant Application Basics (NIH)
• NIH Institutes, Centers & Offices (NIH)
• Principles and Practice of Clinical Research (NIH Clinical Center)

Note: PDF files requires a viewer such as the free Adobe Reader.