Most water safety regimes are reactive by design, even the conscientious ones. The logbook fills up, the temperatures get logged, the showers get flushed, and a decision only gets made when something goes wrong: a reading drifts out of range, an outlet smells off, a sample comes back positive, an inspector asks a hard question. That approach keeps a building legal and, most of the time, safe. It just leaves you a step behind the system.
Proactive control is the shift from waiting for your water to tell you it has failed to reading the conditions for failure before they bite. For a team that already has a risk assessment, a written scheme and a monitoring routine, it is the obvious next move. It is also easy to get wrong, and easy to buy for show.
The useful way to weigh the two is not “which feels more diligent”. It is what each one actually buys you: how much risk it removes, how good the evidence is when someone challenges you, and how much standing effort it costs to keep up.
What a reactive regime actually looks like
Reactive is not lazy, and it is not non-compliant. A reactive programme can tick every box: a suitable and sufficient assessment, a control scheme, temperatures recorded at the right frequency, remedial action when a result fails. The defining feature is the trigger. Action waits for a defect to surface, usually at an outlet, usually after the bacteria have already had the conditions they needed.
The weakness shows up in the gaps. A reactive team treats a string of borderline temperatures as “in range, move on” rather than as a drifting calorifier. It treats a redundant spur as a flushing task forever, rather than a thing to remove. And it leans hard on sampling, because a clean result feels like proof — which is exactly the trap, since a negative sample describes one outlet at one moment and nothing about the system around it (see Lab testing limitations).
Proactive flips the trigger to the upstream signal. The drift gets investigated before it fails. The redundant spur gets capped or pulled. Sampling becomes verification of a system you already believe is controlled, not the thing you are hoping will reassure you.
The three questions that separate them
Strip away the language and the choice comes down to three questions a water safety lead can actually answer:
- How much real risk does it remove? Not tasks completed, but exposure prevented.
- How good is the evidence afterwards? Would your records explain why the system is safe, or only that jobs were done?
- What does it cost to sustain? Not the launch effort, but the weekly grind once it is running.
UK guidance frames the underlying duty the same way for both styles: understand the system, assess foreseeable risk, apply proportionate control, monitor, keep records, and review when things change [1][3]. The difference between reactive and proactive is not the duty. It is how early in that chain you intervene.
Reactive vs proactive, side by side
| Decision axis | Reactive (problem-led) | Proactive (system-led) |
|---|---|---|
| What sets off action | A failed reading, a complaint, a positive sample, an enforcement query | A trend, a design weakness, or a scheduled review of the evidence |
| Risk removed | Brings a known defect back under control after it has appeared | Designs the conditions out so fewer defects appear at all [5] |
| Quality of evidence | A pile of completed tasks; control is implied, not explained | A record of decisions and thresholds; the “why” is written down |
| Standing effort | Spiky — quiet, then a scramble around each incident | Higher to set up, lower to sustain once stagnation is engineered out |
| Where it tends to fail | Slow drifts and rarely-used outlets that never “fail” loudly enough | Tools and dashboards bought, then never reviewed or escalated |
Read the last row carefully, because it is the honest catch. Proactive done badly is just reactive with a subscription: sensors on the wall, a dashboard nobody opens, alerts that fire into an inbox and die. Predictive tools can genuinely move you forward — Machine learning for Legionella risk prediction looks at where machine learning helps and where it overpromises — but a model or a sensor is only proactive if a named person reviews what it says and acts. Otherwise you have bought the appearance of foresight.
Where to go proactive first
This is not all-or-nothing, and chasing the full system at once is how stretched teams stall. Pick the moves with the best ratio of risk removed to effort spent.
Start with the things you can design out permanently. Capping a genuine dead leg or removing a redundant outlet retires a weekly flushing task for good — proactive effort that pays back every week thereafter, not a recurring chore. Next, set escalation thresholds in writing: not just “log the temperature”, but “two consecutive borderline readings on this calorifier triggers an investigation, not another tick”. That single rule converts your monitoring from a diary into an early-warning system.
Then formalise ownership. A proactive programme has a clear answer to who reviews the evidence, how often, and who decides when a control gets escalated. In larger or higher-risk estates this is where a water safety group and a water safety plan earn their place, pulling the people who affect the water — estates, operations, clinical or catering leads — into one accountable forum rather than leaving it all on one responsible person’s desk [6]. Spreading that awareness is its own piece of work; Public awareness: making Legionella prevention everyone’s job covers making prevention everyone’s job.
What you do not need on day one is more sampling. Monitoring frequency is set by your risk assessment, not bought by the test, and a clean result never substitutes for keeping hot water hot, cold water cold, water moving and fittings clean [2][4]. Verification confirms control. It does not create it.
What this does and doesn’t change
Proactive and reactive are management styles, not legal categories. The law does not ask whether your approach feels forward-looking; it asks for a suitable and sufficient assessment, proportionate control, competent oversight and records that stand up [1]. A polished dashboard satisfies none of that on its own, and an old paper logbook kept by someone who genuinely understands the system can be more proactive than a sensor array nobody watches. Treat every threshold, frequency and escalation rule above as a prompt to check your own risk assessment with a competent person, not a figure to copy across.
FAQ
Is a reactive approach still legal?
Yes. The duty is risk-based, and a reactive programme that has a suitable assessment, proportionate control, monitoring and records can be fully compliant. The case for going proactive is not that reactive is illegal — it is that reactive tends to miss slow drifts and quiet, low-use outlets that never fail loudly enough to trigger a response until it is late.
Does going proactive mean buying sensors and dashboards?
No. Technology is optional. The shift is in how you use information, not how much of it you collect. Writing an escalation threshold into your scheme, removing a dead leg, or holding a regular review of the evidence are all proactive moves that cost nothing in hardware. A dashboard only counts as proactive if someone reviews it and acts on what it shows.
If we can only make one thing proactive, where should we start?
Engineer out your worst source of stagnation — cap or remove a confirmed dead leg or redundant outlet. It permanently removes both the risk and the recurring flushing task, which is the rare change that lowers effort and lowers exposure at the same time. After that, add written escalation thresholds so borderline readings get investigated rather than just recorded.
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, “Legionnaires’ disease — what you must do”. https://www.hse.gov.uk/legionnaires/what-you-must-do/index.htm [4] HSE, “Testing and monitoring your water system for legionella”. https://www.hse.gov.uk/legionnaires/testing-monitoring-water-system.htm [5] CDC, “Controlling Legionella”. https://www.cdc.gov/control-legionella/index.html [6] BSI, “BS 8680:2020 — Water quality. Water safety plans. Code of practice”. https://knowledge.bsigroup.com/products/water-quality-water-safety-plans-code-of-practice