The framework you built in Lesson 1 is now an instrument. We will pick up that instrument and use it. A site profile is in front of you. The Director of Research has asked for an assessment before the next FDA inspection. You have one read of the site, and one chance to write a plan that the institution will actually act on. The work is not to identify every imperfection — sites have many imperfections, and a list of fifty findings is a list nobody acts on. The work is to identify which deficiencies matter, in what order, and how they connect.

I want to be candid about what makes this hard. The temptation is to evaluate each domain in isolation — submissions here, safety reporting there, training somewhere else — and to produce a domain-by-domain report. That report is useless. The deficiencies in real sites are almost never confined to one domain. A consent version finding from a monitor is rarely just a binder problem. It is usually three problems compounding: incomplete adoption of the electronic system that would have caught the version drift, an absent training program that would have taught the junior coordinator to verify version tags, and an undocumented process that lets each coordinator handle consent versions differently. The single visible finding is the symptom. The systemic risk is what produced it.

What I want from you in this lesson is the discipline of seeing through symptoms to causes, and the courage to recommend the fixes that address causes rather than the bandages that address symptoms. The first is harder, slower, and more politically delicate. The second is easier and useless. Capstone-level competence is the willingness to do the first.

The framework you built in Lesson 1 is now an instrument. We will pick up that instrument and use it. A site profile is in front of you. The Director of Research has asked for an assessment before the next FDA inspection. You have one read of the site, and one chance to write a plan that the institution will actually act on. The work is not to identify every imperfection — sites have many imperfections, and a list of fifty findings is a list nobody acts on. The work is to identify which deficiencies matter, in what order, and how they connect.

I want to be candid about what makes this hard. The temptation is to evaluate each domain in isolation — submissions here, safety reporting there, training somewhere else — and to produce a domain-by-domain report. That report is useless. The deficiencies in real sites are almost never confined to one domain. A consent version finding from a monitor is rarely just a binder problem. It is usually three problems compounding: incomplete adoption of the electronic system that would have caught the version drift, an absent training program that would have taught the junior coordinator to verify version tags, and an undocumented process that lets each coordinator handle consent versions differently. The single visible finding is the symptom. The systemic risk is what produced it.

What I want from you in this lesson is the discipline of seeing through symptoms to causes, and the courage to recommend the fixes that address causes rather than the bandages that address symptoms. The first is harder, slower, and more politically delicate. The second is easier and useless. Capstone-level competence is the willingness to do the first.