Writing about food and health is one of the easiest ways to get traffic. It’s also one of the easiest ways to get it wrong.
Say a new study is in the news: “Coffee cuts liver cancer risk in half!” You write it up, it gets shared, and it turns out the study looked at one small group of people and never proved coffee caused anything. Now your site has your name on a claim that isn’t true, and Google notices.
I’m Damon Nelson. I have built marketing software for more than twenty years, including software that writes news articles. This guide gives you a simple, repeatable way to write health and food articles that are accurate, easy to trust, and built to hold up to Google’s stricter rules for health content.
First, what is YMYL?
YMYL stands for “Your Money or Your Life.” It’s Google’s name for topics where bad information could hurt someone’s health, safety or finances.
Food and nutrition land right in the middle of it. A wrong claim about a supplement or a diet isn’t just a weak blog post. It could lead someone to make a bad choice about their body. That’s why Google holds these pages to a higher standard than, say, a review of coffee grinders.
Google also hires people called quality raters to review search results and grade them. For health topics, they look closely at who wrote the page, what knowledge the page shows, and whether it was written to help readers or just to get clicks. Google lays this out in its advice on creating helpful, people-first content.
Here’s a quick way to tell when the higher standard kicks in:
| Usually low risk | YMYL: be careful |
|---|---|
| A recipe for banana bread | “This food boosts your immune system” |
| Which apples are in season | “This diet lowers cholesterol” |
| A review of a kitchen gadget | Supplement claims or doses |
| Restaurant news | Food safety, allergies or a new nutrition study |
The rest of this guide is a step-by-step process for the right-hand column.
Step 1: Decide your one claim before you write
Most bad health articles fail before the first sentence, because nobody decided what the article actually says.
Before you write, finish this sentence: “This article will tell readers that ______, according to ______.” For example: “Eating fermented foods may improve gut health, according to a 2024 trial.”
Then write the question a reader would type into Google that your claim answers. If you can’t say your claim in one sentence, you don’t have an article yet.
Step 2: Write down your source before you draft
Find the original study, health agency guidance or named expert behind the claim. Then make a quick note with three things:
- The link to the original source
- The date it was published
- How many people were in the study (this is called the sample size)
If you can’t find the original source, skip the story. Doing this first means your fact check later is a quick confirmation, not a treasure hunt.
Step 3: Use sources you can trust
Not every story about a study deserves space on your site. Start at the top of this list and work down. Your article is only as strong as the weakest source you lean on.
| Rank | Source | Use it for |
|---|---|---|
| 1 | Peer-reviewed journals, meaning studies checked by other experts before publishing (search them free on PubMed) | The study itself, its results and its size |
| 2 | Reviews that combine many studies (called systematic reviews or meta-analyses) | Whether a finding holds up across lots of research |
| 3 | Government and health agencies like the NIH, FDA, NHS and WHO | Official advice and safety limits |
| 4 | University news offices | Background, quotes from the researchers and limits of the study |
| 5 | Early research that hasn’t been checked yet (preprints and conference talks) | Label it as early, or skip it |
| 6 | News roundups and other blogs | Finding stories only. Never cite them as proof. |
Drop the story if:
- The write-up doesn’t name a study at all.
- The study was done on animals, but the headline talks about people.
- The “journal” will publish anything if you pay it.
- The only source is a company that sells the product.
If a story rests only on rank-5 research and you can’t confirm it anywhere else, skipping it beats publishing a correction later.
Step 4: Run the five-check fact check
Fact-checking is a checklist you run before you hit publish, not a feeling you get after skimming the study. Once you’ve done it a few times, it takes about fifteen minutes.
- Find the original source. Trace the claim back to the actual study, review or guideline.
- Read the study, not just the summary. Summaries tend to oversell. Read the parts that explain how the study was done and what its limits were.
- Check who paid for it. A chocolate study paid for by chocolate makers isn’t neutral.
- Check who was studied. A study on mice, or on 12 people, can’t support advice for everyone.
- Check the date. Nutrition advice changes. Old advice needs current guidance to back it up.
Step 5: Match your words to the strength of the proof
This is where most AI tools go wrong. A study finds a weak link in one small group, and the article says “X boosts your immune system.” Fix it with this simple rule:
| What the research is | Words you can use | Example |
|---|---|---|
| An observational study, meaning researchers watched what people already do, without testing anything | “may help” or “is linked to” | “Eating more fiber is linked to better heart health.” |
| One controlled trial | “In one study of [who], [what happened]” | “In one study of [number] adults, a daily walk lowered blood pressure.” |
| A review of many human studies | “has been shown to” | “Regular exercise has been shown to improve sleep.” |
The examples above show the wording, not real study results. Use the real numbers from your own source.
Keep every claim at or below the level its proof supports. Do that, and you can hand this rule to a writer or a tool and get the same careful results every time.
Step 6: Run the headline test
Put the headline you want to write next to the type of study behind it. If the headline promises more than the study can deliver, rewrite the headline, not the warning.
- Mismatch: a study on mice, with the headline “Scientists say eat less sugar.”
- Mismatch: an observational study, with the headline “This food prevents cancer.”
- Match: “In a study of [number] adults, people who ate more beans had lower cholesterol.”
Step 7: Turn scary percentages into real numbers
Headlines love percentages like “50% lower risk.” That’s called relative risk, and it can make a tiny change sound huge. Absolute risk tells you what actually happened to real people.
| How it’s said | What it means |
|---|---|
| “Cuts your risk in half!” (relative risk) | The risk went from 2 in 10,000 people to 1 in 10,000 people. |
| The honest version (absolute risk) | “About 1 fewer case for every 10,000 people.” |
The numbers above are an example of how the math works, not a real study. In your article, always give the starting risk and then the change. If the study doesn’t share the raw numbers, say so.
Also name who was studied in your first three paragraphs. A study of middle-aged men in one country doesn’t tell much about women over 60 in another. Your readers deserve to know who the finding is really about.
Step 8: Use a simple layout readers can trust
Build every health article in the same order:
- What the study found, in one sentence
- Who ran it and how many people were in it
- What this means for your plate
- The limits of the study
- A short list of practical takeaways
Keep the study facts separate from your own opinion, so readers can see where the reporting ends and your view begins.
Put your links right where the claim is, the first time it appears, not buried at the bottom. It shows readers you read the study and not just a headline about it. A short “Sources” list at the end is a nice extra for people who save or print the page.
Step 9: Know when to add a disclaimer or hire an expert
A disclaimer is a short note telling readers the article is information, not medical advice. A medical reviewer is a trained professional, like a doctor or registered dietitian, who checks your draft before it goes live.
| If your article covers | Disclaimer? | Medical reviewer? |
|---|---|---|
| Recipes or what’s in season | Usually not needed | No |
| A new nutrition study | Yes | A good idea |
| Supplements or doses | Yes | Yes |
| Pregnancy, children, drug interactions or a medical condition | Yes | Yes |
Write one disclaimer, put it across your whole site, and add an extra line to any article that makes a stronger claim than usual.
If you promote supplements as an affiliate, be extra careful. The Federal Trade Commission (FTC) expects health claims in ads to be backed by solid scientific evidence, and that includes affiliate content. If you can’t afford a reviewer, stick to summarizing single studies with their limits, and never turn animal or lab-dish results into advice for people. (This is general guidance, not legal advice.)
Step 10: Fact-check once, use it many times
The fact check is the slow part, so do it once and save it. Keep a simple note for every story:
- The source you checked
- The claims you confirmed
- The warnings and limits you added
- The date you checked it
Every article you write from that story uses the same checked facts. A second site, an email or a social post then takes minutes instead of another hour of research.
This is also where AI fits, and where it stops. Writing tools can turn a trending story into a draft and social posts on a schedule. My own app, NewsMasher, reads the original reporting before it writes in your voice. But no tool replaces the checklist on a health topic. Turn off automatic publishing for health stories, check each draft yourself, and only publish once it passes.
Quick answers
Can I write health articles without a medical degree?
Yes. Google’s raters look at the knowledge the page shows. Use original sources, match your words to the proof, and keep your headlines honest. For higher-risk topics, add a credentialed reviewer.
Can I republish health news from other websites?
Not word for word. Copying someone else’s article is copyright infringement. Summarizing a study in your own words, with your own sources and your own take, is fine. Link to the original reporting.
How new does a study need to be?
Older studies can still be useful if newer research hasn’t replaced them. But for any safety or dose claim, check that current advice from groups like the FDA, NHS or WHO still agrees.
Does Google punish AI-written health articles?
Google judges the quality of the page, not the tool that wrote the first draft. The danger is that AI drafts often overstate findings and skip sources. Check every health article yourself before it goes live.
Trust is the whole game
Health readers are trusting you with something that matters. When you get it right, they come back, they share your work, and they believe you the next time.
This checklist turns that trust into a habit. Pick one claim. Find the real source. Use honest words. Check it once, and build on it. Your site becomes the place people go when they want the truth about what’s on their plate.
Want more plain-English guides like this one? Join the Pit Crew, my weekly email. One email a week, no pitch parade.
