What HR receives from a corporate programme

This question comes from two different people with opposite worries. The employee asks whether the employer will find out. The HR director asks whether there will be anything to defend the budget with. Both deserve a straight answer, and the two are not in conflict.

What goes into the report

Utilisation. What share of the purchased volume was used in the period. This is the number that shows whether the programme is alive or was bought and forgotten.

Time from request to first consultation. Measured automatically and usually part of the contract. If the contract sets a service level, this is the number that proves it.

Number of employees supported. As a total for the period, without names.

Broad categories of concern. Workplace stress, personal relationships, loss, organisational change. The categories are deliberately wide: the finer the categorisation, the more recognisable the person behind it.

Attendance at training and webinars. No confidentiality problem here, because attending a workshop is not therapy.

Aggregate satisfaction, if it was collected anonymously.

Emerging organisational themes and recommendations. This is the most valuable part and it comes from the therapist, not from a system. If three separate teams raise the same question about how change is managed, that is organisational information, not personal information.

What is deliberately left out

Names, emails, phone numbers. Dates and times. Which therapist. Topic, reason, diagnosis or notes. Number of visits for a specific person. Whether a named individual used the service at all.

That last one surprises people. "Did Ivan use his sessions?" has no answer, not because we decline to give it, but because the report is generated from a query that never touches the link between a person and an employer.

The cohort threshold

Below a minimum number of users, the report gives no figures at all — only that the threshold was not reached. At Integria the floor is five people and it cannot be contractually lowered below three.

The reason is arithmetic, not policy. "Average four sessions across two users" is information about two specific people. A department of six where five used the programme identifies the sixth by elimination. Aggregate numbers stop being aggregate somewhere, and the only safe place to draw that line is before the report is produced, not after someone asks a clever follow-up question.

Three questions worth asking a provider

"Show me a sample report." A provider who cannot show a redacted example has not thought about the boundary.

"What is your minimum cohort size and can it be lowered?" If the answer is that it can be lowered by agreement, the protection is a preference, not a rule.

"What happens if our CEO asks for a name?" The answer you want is that the request cannot be fulfilled, followed by an explanation of why the data path does not exist. The answer you do not want is a reassurance about professional ethics.

How Integria does it

The link between an employee and a contract exists in exactly one column in the database, and no report reads it. The usage report is a database function that refuses to return figures below the cohort threshold — so a new report written a year from now still cannot bypass it. The invoice to the company is a single line with a number of sessions: a list of dates in a small cohort is a list of people.