The EAP Data That Doesn't Add Up

The presenting request (workshops) targets individual coping, but the data point to access and cultural barriers, not a skills deficit. A resilience workshop cannot plausibly move utilization when the structural problems are a single national intake queue that excludes 60% of the workforce and supervisory norms that punish help-seeking.

I would reframe in writing: 'Utilization is an access-and-safety signal, not an engagement metric.' Then propose a staged diagnostic with three parallel workstreams:

1. Access audit - quantify after-hours demand, wait times, and abandonment; model what an overnight-shift-specific intake pathway would cost vs. the workshop contract.
2. Supervisory norm assessment - anonymous interviews or a short validated climate measure across shifts, since the two informal reports are not yet evidence.
3. Utilization decomposition - request de-identified EAP data by shift, department, and issue type to test whether the 4.2% is uniform or concentrated.

Critically, I would not simply refuse the pre-committed budget. I'd propose running a small, evaluated workshop pilot alongside the diagnostic so the client keeps momentum while the data determine where the real spend goes. This respects the relationship without colluding with a premature solution.

Ethically, accepting an engagement framed around a solution the evidence doesn't support, with a satisfaction-only metric, risks colluding with a predetermined conclusion - the client pays for activity, not outcomes, and I lend professional credibility to a likely-ineffective intervention. Satisfaction ratings measure the workshop, not the problem.

Practically, I'd negotiate a middle path rather than deliver an ultimatum:

• Accept a scoped workshop pilot with a defined denominator (e.g., 120 employees across both shifts).
• Add two low-burden outcome measures agreed in advance: EAP utilization rate by shift at 90 and 180 days, and a 3-item psychological-safety pulse at baseline and follow-up. Satisfaction becomes a secondary, not primary, metric.
• State the falsifiable prediction in the proposal: if access and norms are the drivers, workshops should show high satisfaction but flat utilization. That protects the client from misreading a satisfied-room result as success.
• Document the disagreement in the engagement letter so the rationale for the design is transparent.

If the client insists on satisfaction-only evaluation with no baseline, I would decline the outcome claim in writing and limit my deliverable to 'training delivery,' not 'engagement improvement.'

This is the predicted pattern, so I'd frame it as informative, not failed: the pilot did what it was designed to do (high satisfaction) and also did what a skills-only intervention cannot do (move access-driven utilization). Presenting it defensively or apologetically would obscure the lesson.

Structure the readout:

1. Lead with the agreed prediction - restate that flat utilization alongside high satisfaction was the anticipated result if access and norms were the binding constraints.
2. Triangulate - show the access data (wait times, abandonment, overnight gap) and any climate findings next to the flat utilization line. The convergence is the argument.
3. Quantify the cost of inaction - presenteeism, turnover, and absence differentials by shift if available, to translate into business terms leadership already tracks.
4. Offer a decision menu, not a lecture: (a) overnight intake pathway, (b) supervisor training on help-seeking norms with a behavioral metric, (c) both, with a 12-month evaluation plan.

I'd explicitly credit the pilot: it built buy-in and gave us a credible baseline. The relationship is protected by making leadership the informed decision-maker rather than by softening the finding.