ASHRAE 188: what the standard asks of a building, in plain terms

ANSI/ASHRAE Standard 188 is the document most often named when somebody asks what a building is supposed to do about legionella. It is searched 260 times a month, that number has risen by about a third over six months, and almost everybody arriving at it wants the same thing: not the standard itself, which is sold by ASHRAE, but an answer to whether their building is doing what it is supposed to.

This article does not reproduce the standard and cannot substitute for reading it. What it can usefully do is describe the shape of what the standard asks for, because that shape is what tells an owner whether the work has actually been done or whether somebody has sold them a binder.

The one idea underneath it

The standard treats a building water system the way food safety treats a kitchen: not by testing the output and hoping, but by identifying where in the process things can go wrong, deciding in advance what will be controlled and within what range, watching those points, and having a written response ready for when a reading falls outside the range.

That is the whole of it, and it explains the thing operators find most counter-intuitive: sampling is not the center of the standard. A program built the right way round decides what it controls first. Sampling then has a specific and limited job, which is to provide evidence about whether the program is working. Nothing else in this field causes as much wasted money as getting that order backwards.

What a program contains, element by element

The seven elements below are the structure the CDC toolkit uses to walk a building through the same work, and they are the structure of the builder on this site.

A named team

Not a document owner, and not a contractor. People who between them know the plumbing, the operations and the occupants, named individually, with somebody accountable. This is the element most often skipped and the one that makes the rest possible, because every later step ends in a person deciding something. A program owned by the binder is owned by nobody.

A description of the systems, and a flow diagram

Every system that holds or moves water, and the path water takes from the point it enters the building to the point somebody uses it. The standard expects a diagram. Drawing it is the step at which most buildings discover something nobody had recorded: a capped branch, a tank taken out of service, a wing whose outlets nobody opens.

Where control measures apply

The places in those systems where conditions can favor growth. Warm water sitting still is the general case, and the specific cases follow from the building: storage vessels, recirculation returns, long runs to distal outlets, tower basins out of the flow, anything rarely used. Naming a location is not a finding that something is present there. It is a commitment to watch it.

Control limits and monitoring

For each location, what is measured, against what limit, how often, by whom, and where the reading is written down. The limit is the team’s decision for that building. This site publishes no figure, and treats anyone who hands you one without having seen the building as guessing, however confidently.

Corrective actions

What happens when a reading falls outside the limit, written before it happens. Decided in advance, it is a procedure. Decided afterwards, it is an explanation, and the difference is obvious to anyone who later reads the file.

Verification, and validation

Two questions that get collapsed into one constantly. Verification asks whether the program is being carried out as written: are the readings being taken, by the person named, at the stated interval. Validation asks whether the program is working, and this is where environmental sampling belongs, as evidence about the program rather than a verdict on the building. What a result can and cannot carry is in what a legionella test tells you.

Documentation and communication

Where the records live, how long they are kept, who is told what. Unglamorous, and the element that decides whether the previous six can be shown to have happened at all.

Whether it applies to you is a separate question

A standard is not a law. ASHRAE 188 becomes binding on a particular building through something else: a state or local rule that adopts it, a condition of participation attached to funding, an accreditation body, an insurer, or a clause in a contract or a lease. Which of those reaches your building is a question about your building, and this page cannot answer it. What a written program has to contain once one of them does reach you is on the water management programs page.

It is worth knowing that many buildings write a program without being required to. The reason is not virtue. It is that the alternative to a written program is not the absence of scrutiny; it is scrutiny arriving at a bad moment, with nothing on file.

How to tell whether yours is real

Four questions, and they sort a real program from a purchased document faster than reading it does.

Who is on the team, by name? If the answer is a company, the program has no team.

Where is the flow diagram? If there is none, the systems were never walked.

Show me last quarter’s monitoring records. Not the plan for monitoring: the readings, signed, with dates.

Show me a corrective action that was closed. One instance where a reading went out of limit, something was done, and a follow-up reading closed it. A program that has never recorded an exception in two years is usually not being run.