Would like to hear all your other thoughts about K08s as a reviewer, e.g., with the research plan, career development plan, biosketches, letters of rec, etc.
Where I am at, if you don't get a K of some sort during your last year of residency/fellowship, then you get put on the clinician educator track and have to do a full time clinical job + a full time research job until you get a K or an R
I will preface this by saying that I think every reviewer has a different take. I had a K from an applicant who I thought was great, but the other reviewer saw flaws. Alternatively, there was an K that I and another reviewer had significant issues but the third reviewer "rescued" the application. So take all of it with a grain of salt.
My single best advice is to see what a funded K application looks like. Generally speaking, funding Ks (and well any grant) has typically gone through considerable polishing. This can be from colleagues, internal reviews, peer-reviewer mechanisms (ie unfunded attempts), whatever. But the more feedback one gets, generally the better the application gets. I did a number of mock study sections, external KL2s and other feedback prior to my K submission. But the baseline for that was seeing what a successful K looked like. If you've never written a grant, how can you know what a success looks like?
Generally speaking, the candidate >> research plan. That doesn't mean the research plan doesn't matter, it does, it just doesn't matter as much. The candidate like I mentioned above needs to be published, have a recent first author publication and one with the mentor. The mentor has to have 1) a track record of some K mentorship and hopefully a transition of someone from a K to an R, 2) must be an associate professor or higher and 3) must have active NIH at the time of the K award through a renewal mechanism (no R21). The mentoring team should be broad. This is one of the things were potentially the more, the better. It is a red flag if the mentoring team is 3 people or less. Ideally, one mentor should be external to the institution. There should be very clear roles of each mentor, what expertise they provide and how they will help the candidate in their career development. There should be VERY clear timelines of meetings, benchmarks of success. Career development plans should also include some didactics, but no more than 4 structured sessions in my opinion. If you have to few, it gives a sense you aren't learning anything new and if its too many, then when are you going to do the science? There should be conferences and presentations of course. There should be a plan of structured peer mentoring. These can be hard to find. Typically, they will be at your institution. Occasionally, there are national ones. The MOSAIC program comes to mind which has annual conferences and meeting of peer-mentoring, though that program is only open to certain individuals. But something structured. There should also be a plan to take a structured grant writing course or seminar. A lot of this is boilerplate stuff, so in that regard, every applicant writes a version of this. But the applicant who sets themselves apart is someone who writes WHY they are doing what they are doing, WHAT gaps it fills and HOW they see it helping them.
I talked about the candidate and publications before. The Candidate's Statement is very important. This is your chance to explain why you want this path, what obstacles you faced (and a place to explain training gaps or publication deficits) and how you overcame them, what training yu've had and what you think you need to become successful. Some of these items can and should also be embellish in the mentors' letters. You have 6 pages for the mentors' letter so use them all. It's especially helpful if the mentors' letter give very clear details about 1) what you have achieved thus far 2) who you sought as experts 3) how the mentor has been successful and 4) how you and the mentor have clearly discussed your plans for success. Along these lines, the Institutional Letter of Commitment must say that you will "be at the rank of assistant professor on the tenure track at the time of the award", that your "continued employment is not dependent on receipt of the award" and they will commit "at least 75% protected research time which takes into account clinical and administrative duties". If any of these are missing, its a red flag.
The biosketches, honestly I glance through those though not everyone will just glance. The personal statement is usually similar to the candidate's statement but you can write more details about experiences if you run out of room in the main grant. The only things I will definitely look at are 1) the Position/Honors to see if there is anything interesting or unique (awards for research, prior LRP awardee, etc.) and 2) the bibliography link (so make sure it is accurate and up to date). Nothing is more annoying to get the biosketch of applicant "joe blow" and the hyperlink doesn't work and so I have to go to PubMed and type in "joe blow" and have 50 hits come up and try to decipher which one is which.
The Research Strategy is going to be ~6 pages (this doesn't include the Aim page which you get 1 page for). Really, the most important parts to me is does it read well and make sense and are the tools, equipment, personnel you need to complete the aims readily accessible. Is the person with expertise and equipment 3 hours away? Red flag. And the clinical samples you plan to obtain dependent on a person who isn't nearby or there isn't at least a demonstration that you can get them? Red flag. Have you even done a proper estimate of the number of patient or animals you need? Are the numbers that you suggest even reasonable and obtainable? Do the numbers to state in the research proposal match the numbers you've 1) budgeted for and 2) suggested in the animal or human subjects documents. If all those are in place, it general comes down to grantsmanship. You need to explain to people (who may not be experts in your field) WHY you question matters, WHAT are the knowledge gaps and HOW you will complete them. While preliminary data is not technically required, it is. This data will be a demonstration of the HOW. Typically, you only need a couple of figures of preliminary data. There should be 1 figure of a diagram of all the aims and how they are related. Remember that aims should be independent of one another (ie the success of aim 2 can't depend on aim 1) but they are suppose to be connected to complete the story. In this regard, a figure that nicely sums up the WHY, WHAT and HOW can go a long way. Generally speaking, this part of the grant almost always get the worst score from reviewers, so that is typical. The best way to reduce the negative scoring is to make it readable and clear. If it is those things and the reviewer has a sense that you really understand what you are talking about with a clear knowledge of the gaps in research and how your research strategy will fill them, that always helps to defuse the scoring issues.
As for your last comment, that's institution and field dependent. And it's hard. The best you can do in that situation is use the people and resources around you (remember, it doesn't have to be your preliminary data exactly) to put together the most comprehensive grant together you can.