A flushing schedule prevents nothing by itself. It moves water through outlets that would otherwise sit still — useful, but mechanical. The prevention happens later: when someone reads the log, notices a number drifting the wrong way, and acts before a risk turns into a case. This is the story of one of those moments.

Treat this flushing case study as a composite, assembled from the kinds of buildings and records UK facilities teams handle every week — not a single named incident. The site is illustrative. The decisions are the part worth keeping.

The building, and the part of it nobody used

Picture a four-storey office in a UK town centre, fully let until spring, when the third-floor tenant moved out. The space sat empty into the summer while the letting agents did their work. The water serving that floor — a run of cold-water outlets, two kitchenette taps, and a shower in the accessible WC — went from daily use to almost none.

On paper, water safety was in good order. There was a current risk assessment, a written scheme of control naming who did what, and a weekly flushing routine for low-use outlets that the on-site team logged as they went [1]. The plan was not the problem.

The blind spot was that the third floor had slid from “occupied” to “void” without anyone re-rating the outlets it contained. Pipework that barely sees a draw-off holds water still and lets it drift toward room temperature — the exact condition the HSE names as a driver of Legionella growth [5]. The flush list still covered those outlets, but it treated them as the low-priority taps they had been, not the stagnation risk they had become.

What the records actually caught

The flushing kept happening, more or less on time. The value, though, was in what sat next to each tick. The team recorded the temperature whenever they flushed, and across three weeks a pattern surfaced at the far end of the third-floor cold run: the cold outlet, which guidance expects to deliver genuinely cold water, was reading higher each week even after a full run-off of a couple of minutes [3]. Cold water that will not come up cold is telling you it is sitting warm somewhere behind the wall.

One elevated reading is noise. Three in a row, trending in the wrong direction, at an outlet nobody was using, is a signal. The responsible person treated it as one.

The decisions that mattered

A handful of choices turned a worrying trend into a non-event.

  • They raised the flushing frequency on the void floor straight away, rather than waiting for the next scheduled review to authorise it.
  • They walked the pipe run and found a capped branch — a leftover from a tea-point a previous tenant had removed — sitting as a dead leg, holding warm, still water hard against the live main.
  • They took a sample, not to replace the fix but to verify microbiological control while the remedial work was scheduled. The purpose and timing of the sample followed the risk assessment and the sampling code of practice, not habit [4][6].
  • They had the dead leg cut out, rather than adding it to the flush list to be babysat forever.

In my view, the first of those — escalating on a trend rather than waiting for a single hard breach — is the decision most teams miss. A reading that is “still within range, but creeping” rarely triggers anything, because nothing has technically failed yet. That creep is precisely when intervention is cheapest.

Why this counts as prevention, not luck

Nobody fell ill. No sample came back with a worrying count, and no outbreak was ever on the table. From the outside it looks like nothing happened, which is exactly what good prevention looks like.

What actually did the work was not the flush itself but the loop around it: flush, record, read, react. Flushing and temperature management are controls; the sample was verification — useful for confirming the picture, never a substitute for fixing a known fault [2]. Had the team logged the flushes and never looked at the trend, the dead leg would have stayed in place, the water would have kept warming, and the first sign of trouble might have been someone inhaling an aerosol from that accessible-WC shower the day it was finally switched on for a viewing.

What to take back to your own site

The transferable lessons are small and concrete.

Re-rate outlets when a space changes use, not just at the annual review — a floor going void is a material change. Record a result, not just a completed task; a flush with no temperature beside it cannot show a trend. Build an escalation step for “in range but drifting”, so a creeping number reaches a decision-maker instead of disappearing into the log. And when you find a dead leg, design it out where you can rather than committing to flush it indefinitely — the cheapest outlet to control is the one that no longer exists. For the mechanics of a regime that produces readings worth reading, see on building a flushing programme; for why stagnation sits behind cases like this, Neglected water systems: the danger of stagnation is the companion piece.

If you do one thing this week, pull the last month of flushing records for any low-use or void area and look — not at whether the flushes were done, but at whether the temperatures are holding steady or quietly drifting. The trend is the warning the tick-box hides.

A note on how to read this

This is general guidance illustrated through an invented scenario, not legal, engineering or microbiology advice, and not a template to drop onto your own building unaltered. The temperatures that count as acceptable, how often you flush, when you sample, and what a result obliges you to do all come from a competent, site-specific risk assessment for your system and the people it serves. Where a figure matters, confirm it against current HSE guidance rather than this account.

FAQ

What turns flushing from a tick-box task into something that actually prevents an incident?

Acting on what it records. A flush moves the water; the prevention comes from logging a temperature alongside it, watching for a trend, and escalating when an outlet starts to drift — before anything formally fails.

A flush was done but the temperature was out of range. What happens next?

The task is not “complete” just because water ran. An out-of-range reading is an exception that should be recorded, escalated to the responsible person, and investigated — stagnation, a dead leg, or a wider temperature fault are the usual culprits, and the response follows your written scheme [1].

Does flushing remove the need to take a Legionella sample?

No. Flushing is a control; sampling is verification, and they answer different questions. A sample can confirm or investigate microbiological control when your risk assessment calls for it, but it cannot make up for a known fault you have not fixed [4].

Sources

[1] HSE, “Legionnaires’ disease. The control of legionella bacteria in water systems — Approved Code of Practice and guidance (L8)”. https://www.hse.gov.uk/pubns/books/l8.htm [2] HSE, “Legionnaires’ disease: Technical guidance (HSG274)”. https://www.hse.gov.uk/pubns/books/hsg274.htm [3] HSE, “Hot and cold water systems”. https://www.hse.gov.uk/legionnaires/hot-and-cold.htm [4] HSE, “Testing and monitoring your water system for legionella”. https://www.hse.gov.uk/legionnaires/testing-monitoring-water-system.htm [5] HSE, “Systems most likely to create legionella risk”. https://www.hse.gov.uk/legionnaires/risk-systems.htm [6] BSI, “BS 7592:2022 — Sampling for Legionella bacteria in water systems. Code of practice”. https://knowledge.bsigroup.com/products/bs-7592-sampling-for-i-legionella-i-bacteria-in-water-systems-code-of-practice-1