Research Grant Proposal Writing Notes - HealthRex/CDSS GitHub Wiki

Notes from various workshops and mentors that have been helping my grant writing.

https://www.hhmi.org/science-education/programs/making-right-moves

Russ' R01 Writing Workshop

  • R01 - Classic, investigator initiated research (I have an idea, I think you should fund me to do it)

  • U - Cooperative agreement

  • P - Program project

  • Clearly must be a "sales pitch"

    • Mission of the Organization? Must be mission relevant

      • NIH: Health of the American people
      • NSF
      • Burroughs-Wellcome
      • Howard-Hughes
      • Proposal must be compatible with mission. Peer reviewers may not care as much, but someone in program will decide whether to fund if relevant or not
      • SF424 - 400 page document. Must obey all rules, otherwise may just reject without review

    • Scoring

      • 100 (best) - 900 (worst)
      • ~10% funding rate currently
      • Additional programmatic review on what project areas are needed not purely score based for borderline cases (e.g., sudden public interest in Ebola research. Score 30 usually hopeless, but at key time, may be pulled up to fill publicly important space)

    • Administrative Elements

      • Not the core, but opportunity to shine application further
      • Biosketch, make personal statement as relevant to project as possible
      • Budget justification, no page limitation, and can highlight key staff or equipment
      • Facilities, again no page limit, so opportunity to highlight great things about environment

    • Idea

      • Exciting, Impactful
      • Must Advance Field
      • Technologically Feasible
      • Relevant to the Field / Mission
      • Synthesis: Not necessary, but often works well. Multiple threads of research in field, both great, and obviously necessary idea is pulling them together

    • Overall Advice

      • Contact Program Officer to discuss rough Specific Aims
        • Is it relevant to their interest? They are paid to expect these calls and conversation.
        • Don't sales pitch here, you want their honest opinion on whether fits
        • They're looking to build a portfolio of research with good chance to produce key papers that will look good when report back to congress to justify funding
        • When submit, can steer toward specific study sections by identifying particular mission goals, and try quoting study section charge that fits well with proposal
        • Can ask about budget questions
          • <$250K modular budget, won't have to explain too much detail
          • >$500K, MUST discuss with program officer first before submitting app
      • Obey Rules
      • Bad mood reviewers
        • Assume they're rushed and no good will. Want to create good experience for reviewer that will get them excited again about why doing science. Otherwise if they find something bad, will just dismiss the rest cause know not going to get good score anyway
        • Beautiful, edifying, etc.
          • No walls of text. Mix in figures, bullet lists, etc.
          • Try having computer read text out loud. Will likely notice wording that sounds bad when said out loud. Opportunity to highlight and redo.
      • Well known rotating membership of review committees. Can know ~75% of reviewers, can drop specific plug citations as a bonus.

    • Title

      • Cram pile of keyword need to get in, then reslice into sentence format
      • Should be clear what trying to do. Expect to repeat what going to do several times:
        • Title, Description, Summary, Significance and some in Specific Aims

    • Project Description (abstract) (line by line)

      • 1-2 Big unsolved biomedical problem, mission critical to health of American people
        • Someone must be dead or dying within 2 sentences
      • 1 Narrow to the critical sub-problem that will focus on, necessary towards solving big problem (New grant writers often fail because propose too big / too ambitious a project that reviewers don't believe the writer can achieve)
      • 1 Make case for current sub-problem being unsolved
        • Reviewer should be outraged by now that such an important problem is unsolved
      • 2-4 Introduce unique opportunity. Key piece
        • Make sure it's unique, cause idea should be that only you (the writer) can take this opportunity. Nobody else can do it.
        • Review
        • Some reference to prelim work why believable opportunity. "Preliminary work not required," but really, won't trust your application without some
      • 1 Statement of your unique capabilities, again why only you can do this
      • 1 Overriding hypothesis, more when answering questions about nature. May not fit well for informatics / engineering, have to force somewhat:
        • "We can build technology that can solve problem..."
      • 1 The above combined lead to our Specific Aims to
        • 1, 2, 3... Best would be 3-5 aims.
      • Optional: Payoff summary of where world will be with success of the grant.
        • Optional because if prior so compelling, should not even really need to summarize

    • Public Health Relevance

      • 2-4 sentences for congressional staffers looking for research to mock.
      • Should be easy to understand for lay audience, easily agree that good use of tax money

    • Specific Aims (Russ doesn't like underlines, bold, etc. Should be tightly written to not need)

      • If bad specific aims, then the rest won't matter
      • Aim: Achievable goal or state of the world. (Informatics goal may be technical capability)
      • Specific: Evaluable/Verifiable. Stated in a way such that uninvolved person can determine whether the aim is achieved or not.
      • The "What" going to do. Not the "How" cause that can change as science adapts.
      • Don't bother worrying about interdependent specific aims, because worry later aims will not be possible if first fail. Nowadays, if have one aim that may fail, the grant is already not strong enough and will fail. All should inspire complete confidence that will be achieved.
        1. For example...
        2. Capability / Technology Development
        3. Small Confidence Building Applications
        4. Extension
        5. Evaluation against Gold Standards / Bigger Project
        6. Larger Application
      • Should recapitulate and expand on summary description, keep reviewer interested and excited. Otherwise they may already dismiss project at this point without reading the rest.

    • Significance ~3.5 pages (Important for new investigator to illustrate scholarly)

      • Justify choice of sub-problem and that is unsolved.
      • Demonstrate that know current science. Most citations here to demonstrate scholarship.
      • Write with eye toward setting up Specific Aims as the obvious, most reasonable thing to do.

    • Innovation ~ 2.5 pages

      • Summary of what is novel, spoonfeed it clearly
      • Preliminary results to demonstrate capabilities

    • Approach ~6 pages (For new investigator, more detail to demonstrate technical virtuosity / capability)

      • Exactly in same order as Aims (but can break up into sub-aims)
      • If you got money tomorrow, what would you do? The "How" would do things if had to do tomorrow, understanding that science may adapt. Best bet on what to do to achieve aims. Can change the "How" approach, but should not be changing Aims of goal.
      • Failure / back up plan (anticipated controls, etc.)
      • How to evaluate / confirm completion
      • Just strong scientific protocol. May be boring, but should be rock-solid plan to inspire confidence that you know how to and will achieve this.
      • Maybe timeline block format, though more relevant for U & P grants with multiple

    Get reviewer guidelines and go back and just put elements in bullet form so reviewer can easily checkbox that each item is fulfilled

Nigam's Version

  1. Important health/science problem … 1-2 sentences. By then end of these, there must be someone sick/dead. Scope it down to an important and defined sub problem … 1 sentence
  2. There exists/is an opportunity to address the problem … 1 sentence. Elaborate the opportunity in context of your prelim results.
  3. Summarize aims of this proposal as one overarching goal … 1-2 sentences
  4. if you did what you aimed for, how will you show a skeptic that you succeeded … 1 sentence.
  5. Closing sentence about impact of your [presumed successful] effort

Mark's DARPA Version

  1. There is a crisis.
  2. Lots of other people have tried to address this crisis.
  3. All these people have failed because they lack the “secret sauce” that is required.
  4. I have the secret sauce!
  5. I will apply the secret sauce to the problem and I will measure its effectiveness.
  6. My work will eliminate the crisis.

General Outline

  1. Significance
    What is the critical knowledge gap or clinical problem that needs to addressed? What potential patient safety issues exist as a result of this being left unsolved?
  2. Background
    What’s already known and already been done in this area, that leaves an unsatisfactory result?
  3. Innovation / Opportunity
    What’s the new idea, approach, method, etc. that makes it possible to address the gap now? Preliminary data?
  4. Objective / Central Hypothesis / Problem
    What’s the overall purpose or central hypothesis / idea for this proposal to achieve or answer?
  5. Specific Aims
    Verifiable, achievable goals towards overall objective
  6. Expected Outcomes
    How will the state of the world be different after the aims are completed? How will this directly resolve the overall objective?
  7. Research / Development Plan / Approach
    If you got the money tomorrow, how would you plan to complete each specific aim? What potential problems can you anticipate, and what is your backup plan?
  8. Budget Justification

NIH Career Development Award Writing - Mark Roltsch

  • Former NHLBI scientific review office, led 65 review sessions from K to R01 to Clinical Trials to Contract Reviews
  • Later Program Officer managing 200+ F, K, T grants

K Awards - Mostly for post-docs

  • Institution specific salaries, usually up to $90K/year for up to 5 years. Requires mentor. Not renewable.

  • K01 - Mentored Research Scientist Development

    • Subset designated only for minorities

  • K08 - Mentored Clinical Scientist for MDs, usually who want to do benchwork

  • K23 - Mentored Patient-Oriented usually for MD/PhDs, but must be patient facing. Patient database doesn't count

  • K25 - Mentored Quantitative Research for people with quantitative background, want to get new bio applications

  • K99/R00 - Best mechanism for a post-doc?

    • 2 years of training then R00 grant to take with you. Good for marketability. Usually as soon as get it, start applying for a job, usually move on within 9 months.
    • Harder, bigger pool of applicants, includes non-citizens
    • Eligibility: <5 years post-doc (usually after 2 years of post-doc)
    • Mentor: Yes
    • Past 4 years of MD or PhD degree, no longer qualify. But starts counting after residency.
    • Salary $50K during training then up to $250K per year as R00
    • Institutional letter needs to show commitment to faculty position for most other mechanisms. This might be better, because expected to transplant faculty stage?

  • Determining institute

    • Look for several options. See who funds mentor, which missions match well, what institute staff attend same meetings

  • What does it take to get funded?

    • Candidate
    • Idea
    • Mentor (Top-notch WITH funding and have prior K mentoring experience. If not, add co-mentors.)
    • Training Plan
    • Time (To put together a good grant. 4-6 months just to prepare. More if need to build prelim data.)

  • Developing Award

    • Match idea to right NIH institute
    • Select right K, study program announcement
    • Know section requirements
    • Know the Review Section most of all, even more than the top material
    • Logical, clear flow. Make it easy for reviewer to read
    • Work with mentor, make sure congruence on training plan, etc.
    • Other reviewers. Get non-scientists to review and comment if makes sense.
    • Grants.gov class for form completion processes
    • Plenty of time for final submission, so no last minute shenanigans. Lots of dependence on other people.
      • Plan for completion 1 month early, so time to clean up issues and get feedback

  • Success Rates (2014):

    • K01: 34%
    • K08: 40%
    • K23: 38%
    • K99: 22%
    • Depending on institution, single digit to dozens of applications. Actually not that many, so success rates not terrible. Even higher for patient oriented.
    • Funding score threshold ~20-22. K's rated by hard point threshold, unlike Rs, Fs picking top % of applicants.
    • Typically 9 months to find out about funding from application.

  • NIH Reporter to pull data out

    • Can do Freedom of Information Act (FOIA) requests for funded grants anyway. Can withhold some information (Letters of Rec, Salary information)
      • Takes ~6 weeks to get it

  • Reading the Program Announcement (PA)

    • Check to see what institute supports
    • Check due dates
    • Check purpose and eligibility
    • Dollar amount, but relatively fixed/low for K training awards
    • Commitment: Requires 75% of dedicated time. Don't lose too much time supporting other people's projects
    • Materials/Supplies/Consultants, usually $25K but up to $50K per year

  • Page Limits

    • 1 - Resubmission Introduction. After 1st resubmission, would have to apply under new title. No longer limited to just two applications.
    • 1 - Specific aims
    • 12 - Main body
      • Candidate Background
      • Career Goals and Objectives
      • Plan for Career Dev,
      • 5-6 - Research Strategy: Significance, Innovation, Approach
    • 1 - Training in the Responsible Conduct of Research (just pull from someone else, hitting 5 key points)
    • 6 - Plans and Statements of Mentor + Co-mentors
      • Will be written by you
    • 6 - Letters of support form collaborators, contributors, consultants
    • 1 - Institutional Environment description
    • 1 - Institutional Commitment to candidate's research career development
    • 5 - Biosketch (new format)
      • Probably have to edit mentor's biosketch to tailor contribution sections to match application purpose. Pull out key mentor publications that area related to application.

  • More documents

    • External letters of support, separate from any current mentor

  • Helpful Hints

    • Seek active input from mentor and senior investigators
    • Make friends with Academic Research Office on campus
    • Get to know Program Officer, network at conferences
    • Review entire application with mentors and others, catch typos, etc.
    • Plan ahead, expect 4-6 months
    • Easy to resubmit a grant (20 hours) (as long as not shredded in review process), hard to build one from scratch. By 3rd submission, 75% of grants successful.

  • Common problems

    • Lack of well though out research or training plains
    • Weak or absent hypotheses
    • Poor presentation. Figures too small. Writing errors.
    • Weak institutional support
    • Weak publication history or inexperienced PI.
      • How many publications needed to be competitive? 4-5 usually good.
      • 4 in grad school, 4 more in first 2 years of post-doc, should be pretty good.
      • MD only people might get away with 2 publications in a year or two after medical training
    • Distant mentor
    • Scholarship (knowledge of published work)
    • Unrealistically large amount of work proposed
    • Uncertainty regarding future career direction