How Not to Write an Unsuccessful Grant Proposal

Daniel Clauw, M.D. Assistant Dean of Clinical and Translational Research Director, Center for the Advancement of Clinical Research The University of Michigan

What is the end result you are aiming for? Good science by a good investigator.
50% of applications: Bad idea, and poorly prepared investigator.
Table developed by James Ferrara, M.D.

Science

25% of applications: Good idea, but Investigator is unprepared to do the work. 15% of applications: Great idea, proposed by the perfect person to test it.

Investigator

10% of applications: Bad idea, but capable Investigator.

How to Get a Grant

Design a sound study

Write a good grant

Early Steps in Clinical Research

The idea
The hypothesis

The design
The study subjects

Can I do it?

The idea
Most clinicians can come up with

good ideas about clinical research based on gaps in existing knowledge

Has

question been asked before? If not, why? If so, how does my question or design add to the existing literature?

At the Beginning . . . Think of the End

At the inception of a clinical research

project, ask a series of questions about the end product:

Who

is the audience? Where would the article be published if it is successful? Is a negative study interesting or publishable? How would the results impact clinical care

i.e. Who would care?

The First in a Series of Compromises

Potential audience
Academics Peer review process Practicing clinicians Regulatory agencies (e.g. FDA) Industry (marketing bias)

Public

Early Steps in Clinical Research

The idea
The hypothesis

The design
The study subjects

Can I do it?

Developing the Study Hypothesis / Goal(s)

A surprising number of even established

clinical researchers begin designing the study before they define the hypothesis Keep It Simple (KISS) Unless you are in the position to do a hypothesis-generating study, have a single primary hypothesis and no more than 2 3 secondary objectives

Early Steps in Clinical Research

The idea
The hypothesis

The design
The study subjects

Can I do it?

Compromises in Study Design

Efficacy Long trial Single site Randomized, Blinded Effectiveness Short trial Multicenter Not

Active comparator

Placebo

Open

Early Steps in Clinical Research

The idea
The hypothesis

The design
The study subjects

Can I do it?

Inclusion and Exclusion Criteria

Might be most important aspect of study

design with respect to:

Ability

to recruit subjects

For clinical trials, this will largely

determine:
Efficacy

of treatment Placebo response rate

For mechanistic studies, this needs to

be considered for both patient (clearly define) and control groups

Selecting Participants
Heterogeneous
Tertiary Care
Easy to find / unresponsive

Homogeneous
Population
High placebo response

Extrapolatability of results Smaller n, more responsive

Selecting Controls
Should ideally have all of the

characteristics of the patient, except the condition being studied

Not

always possible Controlling for confounds

Sampling Statistics

What

is a healthy normal control?

Recruitment
Existing patients
Easy to find Pre-screened

Advertising
Apples vs. oranges > Screen failures Higher placebo and drop-out rates

Stay in study longer

Early Steps in Clinical Research

The idea
The hypothesis

The design
The study subjects

Can I do it?

Resources
Do I have enough?
Mentorship Individual Institutional (K-30, K-12) Patients / subjects Do I have enough? Do I have the right ones? Staff Money Federal grants Institutional resources (e.g. GCRC) Foundations Industry

How to Get a Grant

Design a sound study

Write a good grant

Get FREE Money from the Government!!!

OK, so maybe you dont need this guys book. But you need some of his skills . . . . .

Besides good research, you need

Communication Skills

Marketing Skills
Management Skills Flexibility and ingenuity (who can you sell

your research to?) Ability to follow directions, even the apparently meaningless ones Ability to plan ahead and commit significant time
Adapted from Christine Black, UM DRDA

When? Earlier than you think

Internal vs. external deadlines Should get science of grant done in plenty of time to have others read final grant, and the best people to read grant have the least time to do so Administrative shell of grant typically needs to

be turned in 10 days before grant is due

Study design issues, and especially sample size calculations, are necessary for administrative shell

Need for preliminary data

Do your homework:
What has the agency funded recently?

(CRISP, Community of Science, www.guidestar.org for foundations). Read recent successful applications by colleagues. The program officers are your resource, contact them early and often. Colleagues are essential for collaboration and consultation; senior colleagues may know who will likely review your grant. Consult a biostatistician early in the process.

Do your homework, part 2:

Check study section rosters!
Search

the literature to determine roster members expertise. there are content area experts on the study section, does your literature review cite their (appropriate) publications?

If

Do your homework, part 3:

Are their aspects of your project or methodology that may need extra justification because they will be unfamiliar to this roster of people? For foundation grants, trustees or boards may make funding decisions, so be sure that key sections like specific aims are in lay language. P.S. Do not attempt to contact reviewers!

Inouye, S.K., Fiellin, D.A. An Evidence-Based Guide to Writing Grant Proposals for Clinical Research, Ann Intern Med. 2005;142:274-282

Relationships are key:

Step 1: Get the right team together. Find

people with expertise in different, specific areas: working with this specific patient population, using this specific methodology or technique, statistically analyzing this type of data, etc.
Step 2: Buy your grants administrator a nice

gift. You will be depending on this person!

Why you should be nice to your grants administrator:

This is the easy part. The challenge will be figuring out the indirect cost rate.
Chronicle of Higher Education, Carol Cable

Care and feeding of the reviewer:

They are

Overworked (they will at most spend a few hours reviewing your grant) Mindful of the need to further their own careers Not interested in doing outside homework in a new topic area just so they can understand your proposal Mature

Translation: they dont want to strain their eyes deciphering figures and captions that youve shrunk down to 8-point font!

Adapted from Christine Black, UM DRDA

Care and feeding of the reviewer, part 2:

Assume reviewers are intelligent and savvy

about research ..but perhaps have little in-depth experience in your area of interest. Avoid jargon and topic-specific abbreviations. Many reviewers will read only the abstract and/or specific aims.

Inouye, S.K., Fiellin, D.A. An Evidence-Based Guide to Writing Grant Proposals for Clinical Research, Ann Intern Med. 2005;142:274-282

Care and feeding of the reviewer, part 3:

Make reviewer like you and your study in first

two pages so that you start with a 1.0

Teach them something new Convince them that you are the best person in the world to do this terrific study

Science Fiction Novel approach the world will be a better place after this study is done Disneyland approach make them smile while you are sucking money out of their wallets

If you can start with a 1.0 and eliminate

targets youll be fine

Targets
Controversial aspects of a grant

OK to do it, but dont write it Omissions Failure to follow rules

Its

Common Problems in Applications (from NIH and NIAID)

Lack of knowledge of published relevant work
Lack of experience in the proposed

methodology or techniques Uncertainty concerning future directions Unrealistically large amount of work Failure to discuss potential obstacles or alternative approaches

Adapted from Christine Black, UM DRDA

Major issues identified in NIH grant proposals:

Results from a review of pink sheets from 66

R01 applications; the authors categorized major problems found in unfunded grants. Findings presented for each major grant section.
Inouye, S.K., Fiellin, D.A. An Evidence-Based Guide to Writing Grant Proposals for Clinical Research, Ann Intern Med. 2005;142:274-282

Abstract

The broad, long-term objective is . . . . The Hypothesis is . . . . The Specific Aims are . . . . The Research Design is . . . . The health-related relevance of this project is .... Avoid

First person Excessive summary of past accomplishments Amounts of money Going outside the box

Specific Aims/Hypotheses
This (and abstract) are most important parts of a grant. Write this early in the process, send it around, and keep

re-writing to make it clearer, stronger, and more concise. This should be a two page executive summary of the following 23 pages of the grant. Writing tips: to one page of setting the stage Extremely clear, (one sentence if possible) hypothesis Specific aims (between 2 and 4) stated as action items To determine, evaluate, confirm, show . . . . Common critiques: Poorly Focused Too Ambitious

Comments from NIH Reviewer

In addition to proposing a research design that is a fishing expedition, the applicant also proposes to use every type of bait and piece of tackle known to mankind.

Adapted from Christine Black, UM DRDA

Special advice for Training (e.g. K-) awards

Safe study is more important than ground-breaking

science.

Should be a study that leads to other studies, no matter what the results.

Serves as a vehicle for your training. Should ideally encompass all of the elements of a research project, especially those that you have inadequate experience with. Environment Increasing preference for mentoring teams. Letters are extremely important, and read for any nuances. If your mentor cannot take the time to write a strong > 2 page letter, they may not have the time or commitment to mentor you.

Background and Significance

Begin by describing current state of science. Identify knowledge gaps. Justify the need for this specific study. Move science forward, not just laterally. Writing tips

Dont provide too much extraneous background information. Dont overstate the significance of your study. For each area covered in this section, explicitly state the relationship to your proposed project.

Comments from an NIH Reviewer

This application is characterized by ideas that are both original and scientifically important. Unfortunately the ideas that are scientifically important are not original and the ideas that are original are not scientifically important.

Adapted from Christine Black, UM DRDA

Preliminary Studies
Should demonstrate:

Feasibility You can recruit the required study participants. The team has successfully worked together. Promising data Youve partly done what you are asking for money to do. If there are novel methodologies that are being used, that these have been fully developed Writing tips: Abstract of each study is good start. Show how each study links to the proposed work.

Methods
At least 50% of the page allowance should go

to methods. Underdeveloped or vague methods are the most common reviewer complaint. Describe the study design in detail how will you randomize, blind, select controls, etc. Inclusion/exclusion criteria justify your criteria and address any biases they may cause.

Methods, part 2
Availability of participants

esp. with regards to women, minorities, and children. Data collection and procedures: Discuss each instrument you propose to use Describe quality assurance or staff training Consider a table of variables:

Variable Level of Function Instrument SF-36 Functional Sub-score Characteristics (when available) Sensitivity: Specificity: Reliability: Reference Ref.

Auto-antibody level

SuchandSuch Assay Sensitivity: Specificity:

Ref.

Interventions and Outcomes

Ensure adequate randomization and

blinding. What will happen if the blind is broken? Ensure standardization of your intervention. Will you provide training? Ensure your intervention is meaningful. Ensure that adherence is monitored. Clearly define all outcome measures.

Statistical Analysis
Work with your biostatistician often and early.
Address what you will do about missing data. Include realistic attrition rates in your power

calculations. Address what you will do with data from subjects who drop out. Present the plan clearly, referencing each specific aim when appropriate.

Writing style
Business not creative
Active voice

or expect to not would like to or may Anticipate questions Target analogy

will

Final Touches
Realistically assess weaknesses or potential

roadblocks and explain how you would overcome them. Dont blow off this section its your opportunity to think strategically. Provide a timetable or chart summarizing study activities. Proofread. A lot. Then give it to others. Follow all the directions with regards to fonts, margins, page limits, deadlines, etc. Finish with enough time for administrative approval, signatures, etc. check with your grants administrator. This process may take 10 days or more!

How to deal with rejection:

(Youre playing baseball, not shooting free throws)

1.

2.
3.

4.
5.

Write scathing letter of rebuttal. Throw it away. Start with a paragraph of gratitude for reviewers insight. List reviewers criticisms. Do not try to convince reviewers that they are idiots
Address each criticism objectively in the introduction to your resubmission.

Adapted from Christine Black, UM DRDA