Is Reverse Osmosis Water Good for You? Treating Water at Home

Most people buy water treatment before they know what is in their water, which is how you end up with an expensive system that removes the wrong thing. Every method here solves a specific problem and none solves all of them. Filtration, disinfection and mineral content are three separate questions.

Start by finding out what you actually have. Your utility publishes an annual water quality report, and under federal rules it must also tell you what your service line is made of. If the answer involves lead, read our guide to reducing lead in your tap water first, because that changes which option below is right for you.

Reverse osmosis: what it does, and the catch

Reverse osmosis forces water through a semi-permeable membrane and strips out almost everything — dissolved solids, metals, nitrates, fluoride, most contaminants. It is the most thorough point-of-use treatment available to a household, and for a genuine contamination problem it is the right tool.

The catch is that the membrane has no opinion about which minerals are useful. It removes calcium and magnesium along with everything else, and that matters more than the internet’s favourite objection.

The mineral question is real. The pH scare is not.

You will read that RO water is dangerously acidic. That part is overstated: RO water is mildly acidic mostly because dissolved carbon dioxide forms carbonic acid, and your stomach is orders of magnitude more acidic than any drinking water. pH is not the problem.

Demineralisation is. A WHO review of the evidence, Health Risks From Drinking Demineralised Water, documents findings worth knowing:

• Human studies found low-mineral water increased urine output by around 20% on average, with higher excretion of sodium, potassium, chloride, calcium and magnesium, and lower serum potassium.
• Reported early symptoms include tiredness, weakness and headache, with muscle cramps and impaired heart rate in more severe cases.
• Soft, low-calcium and low-magnesium water is associated with higher cardiovascular morbidity and mortality, and low magnesium specifically with sudden death.
• Cooking in very soft water can lose up to 60% of the magnesium and calcium in the food itself.

The suggested targets from that review: calcium a minimum of 20 mg/L and optimally around 50 mg/L, magnesium a minimum of 10 mg/L and optimally 20 to 30 mg/L, and total hardness between 2 and 4 mmol/L. The older 1980 WHO desalination guidance put minimum total dissolved solids at 100 mg/L with an optimum of 200 to 400 mg/L.

Honest caveat: the review itself notes much of this evidence comes from older studies that would not meet current methodological standards, and the WHO drinking water guidelines set no minimum for calcium, magnesium or hardness. This is a reasonable concern supported by consistent evidence, not a settled finding.

The part that connects to lead

This is the detail almost nobody mentions, and it is the strongest practical argument for remineralisation. Low-mineral water is chemically aggressive and dissolves metal out of plumbing more readily. The same review documents three US infants with blood lead levels of 15, 37 and 42 µg/dL — against a concern level of 10 — whose water systems used low-mineral water that intensified leaching from brass fittings and lead-soldered tanks. First-draw lead reached 495 to 1,050 µg/L in the two worst cases.

Calcium also appears to reduce intestinal absorption of lead and cadmium, so stripping it out may work against you twice. If your home has brass fittings, lead solder or a lead service line, unremineralised RO is the opposite of a safety upgrade.

The fix is straightforward: buy an RO system with a remineralisation stage as the final step, or add one. Do not improvise with additive drops, and do not over-mineralise — the same review treats hardness above 5 mmol/L as the upper bound, where stone risk rises.

One more thing worth knowing about plastics

Researchers from Columbia and Rutgers, publishing in PNAS, found an average of around 240,000 plastic particles per litre in bottled water, ranging from 110,000 to 400,000 across three brands. The notable part for this article: most of that plastic appeared to come from the bottle itself and from the reverse osmosis membrane used in purification.

That is a finding about commercial bottling rather than a measurement of your under-sink unit, and it does not undo the case for RO where you have a contaminant to remove. But it is an honest counterweight, and it is a further argument for filtered tap water in a glass or steel vessel over bottled water as a default. On the health side of this, see our call on microplastics.

UV disinfection: for biology, not chemistry

UV is the method most households have never considered, and it does something the others cannot.

Water passes a low-pressure mercury lamp inside a quartz sleeve, and UV light inactivates microorganisms. Dose is intensity multiplied by contact time, measured in mJ/cm², and per the UK’s Drinking Water Inspectorate a minimum of roughly 40 mJ/cm² is generally taken as sufficient.

Its real advantage: UV inactivates Cryptosporidium, which resists chlorination. If you are on a private well or a supply with a biological risk, that is the gap UV fills.

What it does not do, and the limits that matter:

• It removes no chemicals — no metals, no nitrates, no hardness, no lead. Pair it with filtration.
• The water must be clear. Particles shield microbes from the light. Where Cryptosporidium is a concern, pre-filter to below 1 NTU. Iron and manganese should be reduced too.
• There is no residual. Chlorine keeps working downstream; UV protects only at the lamp, so storage and pipework hygiene still matter.
• The lamp degrades invisibly. Output falls to about 70% after 10 to 12 months while looking exactly the same, and scale on the sleeve cuts transmittance further. Replace on schedule, not on appearance.

Carbon filtration

Activated carbon is what most pitcher and faucet filters use. It is good at chlorine taste and odour and some organic compounds, and it is cheap.

The trap is assuming it does more. A carbon filter certified only for chlorine taste and odour is not a lead filter. For lead you need certification against NSF/ANSI Standard 53, and the mark should be on the box. Carbon also does nothing for nitrates, most dissolved metals or hardness.

Distillation

Boiling and condensing removes essentially everything, including minerals, which puts it in the same position as RO without remineralisation. It is slow and energy-hungry, and the WHO review’s infant cases specifically involved formula mixed with distilled or low-mineral water. It has niche uses; it is not a good default for drinking water.

Spring water

Genuine spring water from a tested source does arrive with calcium and magnesium already in it, which is a real advantage over demineralised water and the reason many people prefer it. Regulation varies by country, so check which named spring it comes from and look for that source’s water test report.

What it is not: structurally different water, or a source of vitamins. Water contains minerals, not vitamins, and claims about altered molecular structure are marketing. Judge it on the mineral analysis, the source and the price per litre, and factor in the plastic — glass or refillable containers from a source you have verified are the better version of this choice.

Which one do you need

• Lead, or an unknown service line: an NSF/ANSI 53 certified filter on the drinking tap. RO works too, with remineralisation.
• Private well or biological risk: UV, plus pre-filtration to keep the water clear.
• Nitrates, high dissolved solids, a specific chemical contaminant: RO with a remineralisation stage.
• Chlorine taste only: carbon. Do not overbuy.
• Nothing known to be wrong: read your utility report before spending anything.

Treat it like it will get more expensive, because it will

Everything above assumes clean water arrives when you open a tap. That assumption is getting weaker, and the sensible posture is to treat water as a resource with a price rather than a utility that is effectively free.

The cheap moves are the old ones: fix leaks promptly, since a dripping tap wastes continuously; run dishwashers and washing machines full; keep drinking water in the fridge instead of running the tap until it cools; and water the garden early or late so less of it evaporates before it reaches a root.

The interesting development is that pulling drinking water out of the air is becoming real. The principle is familiar from a dehumidifier — condense moisture out of air — but the newer approach uses porous materials called metal-organic frameworks that capture water even where humidity is too low for condensation to work. The underlying chemistry won the 2025 Nobel Prize in Chemistry, shared by Omar Yaghi, Susumu Kitagawa and Richard Robson.

Yaghi’s company Atoco is now moving to commercial rollout, in 20-foot containerised units aimed at agriculture: around 2,000 to 4,000 litres of high-purity water per day on-grid, and up to 1,000 litres per day off-grid, using ambient energy and able to run on low-grade waste heat. The material has been tested in Death Valley and produced water there.

It is industrial equipment, not a household appliance, and nobody is drinking air-harvested water at home this year. But the direction is clear, and it is a reasonable bet that water scarcity ends up being solved partly by machines that work on the same principle as the one in your basement.

This is general information rather than medical or engineering advice. For a specific contamination problem, start with your utility’s water quality report and your service line inventory, and get certified testing before buying equipment.