A Vendor Meta-Analysis and a Two-Week Protocol Deadline
You are the lead clinical dietitian on a 32-bed cardiology unit. A supplement vendor sends your medical director a meta-analysis of 12 randomized trials (n=3,842) reporting that a high-dose omega-3 capsule reduces 6-month readmission after myocardial infarction (RR 0.78, 95% CI 0.65–0.94). The forest plot shows I²=71%, and nine of the 12 trials were manufacturer-funded. Baseline omega-3 intake was not measured in any trial; participants were mostly men aged 55–70 from three high-income countries. Your local population is older, 45% female, and many patients already take fish oil purchased over the counter. The medical director wants a unit protocol within two weeks and has asked you to present at the next journal club. Two nurses have already started offering the capsule to patients. The pharmacy budget can cover 40 patients per month; the unit admits about 90 post-MI patients monthly.
1. Appraise whether this meta-analysis is sufficient to change unit practice. Identify the specific methodological and applicability concerns, and state what you would and would not conclude.
Read the worked answer
What the estimate supports. The pooled RR of 0.78 is statistically significant, but the 95% CI (0.65–0.94) is compatible with a trivial 6% relative reduction as well as a 35% reduction. With I²=71%, the prediction interval for a new setting would likely include 1.0, so the result is not a stable estimate for my unit.
Threats to validity. Nine of 12 trials were manufacturer-funded; small industry-funded trials tend to overestimate effects. Funnel-plot asymmetry suggests possible publication bias. The outcome is 6-month readmission, not mortality or quality of life, and readmission is sensitive to local discharge practices.
Applicability. Trial participants were mostly men aged 55–70 from high-income countries, and baseline omega-3 intake was unmeasured. My patients are older and 45% female; many already take over-the-counter fish oil, so the incremental benefit of a high-dose capsule is likely smaller. I would conclude that there is low-certainty evidence of a possible small reduction in readmission in a population unlike ours-not enough to mandate a protocol.
2. How would you handle the two nurses who have already started offering the capsule and the medical director's two-week deadline?
Read the worked answer
First, protect patients and staff: ask nursing to pause new starts while allowing anyone already on the capsule to continue if the prescriber agrees; check for bleeding risk (dual antiplatelet therapy, anticoagulation) and drug interactions. Do not frame this as disobedience-say the pause is to avoid inconsistent practice while we set criteria.
Second, renegotiate the deadline with the medical director. Two weeks is enough for a rapid review and interim guidance, not a full protocol. Propose a one-page interim document with explicit eligibility (for example, post-MI, no anticoagulation, tolerated oral intake), a stop date, and a planned review at 8 weeks. Bring pharmacy into the budget discussion: 40 capsules per month versus 90 admissions means explicit prioritization is unavoidable.
Third, use the journal club to show the appraisal and invite the nurses to describe what they observed. Shared ownership reduces informal practice. Document the decision, rationale, and review date, and tell the director what evidence would justify expansion.
3. Design the minimum local evidence step that would give a defensible answer for your population.
Read the worked answer
The minimum defensible step is a rapid systematic search plus a local practice audit, not a new RCT. Search PubMed, Embase, and CENTRAL for post-MI omega-3 trials, screen for trials in adults over 65 and in women, and grade the body of evidence with GRADE. In parallel, audit 30 consecutive post-MI charts for baseline fish oil use, bleeding risk, and 30-day readmission to see whether our patients resemble the trial population.
If the team still wants to proceed, implement it as a registered quality-improvement project with a pre-specified primary outcome (30-day readmission or unplanned visit), a comparison group (pre-intervention cohort), and a stopping rule for bleeding or adverse events. With 90 admissions per month, a 3-month pilot gives about 270 patients-enough for audit and safety, not for a definitive efficacy answer. Pre-register the analysis and report it internally. This gives a defensible local answer without pretending the meta-analysis settled the question.