When I wrote the national drink spiking advice for the Metropolitan Police Service, victim support organisations praised it for hitting exactly the right tone. But what they didn’t see was how much it cost me personally.

I spent weeks immersed in harrowing accounts from survivors, researching predatory behaviours and documenting the appalling aftermath of these crimes. I couldn’t sleep. I couldn’t stop thinking about it. And I certainly couldn’t talk about it at dinner parties.

This is the reality of creating trauma-informed content that actually works. It’s not about theoretical frameworks or academic principles. It’s about understanding how trauma fundamentally changes how people process information and then having the courage to fight for their needs against organisational pressure to water things down.

The tricky territory of experience

There’s something I’ve come to realise that complicates this work even further: the subtle but significant difference between ‘lived experience’ and ‘living experience’.

I see this a lot in my trauma-informed work. Even something as simple as helping my mum navigate digital services has shown me the gap between my “lived” tech-native experience and her “living” experience of trying to learn new systems in real time. It’s a constant reminder to check my assumptions and meet people where they are, not where I think they should be based on my own past.

My past work with vulnerable populations gives me valuable insight, certainly. But there’s a danger in assuming my past experiences – even working directly on trauma-informed projects – give me complete understanding of what people need today. The contexts shift. Social norms evolve. The kaleidoscope keeps turning.

I’ve seen this play out in government projects where well-meaning designers proudly declare “I’m good at empathy” while simultaneously dismissing user behaviours they don’t personally understand. “Why do domestic abuse victims keep returning to their abusers?” they’d wonder, revealing how shallow their supposed empathy actually ran.

This isn’t just academic navel-gazing. It affects real people. At Cancer Research UK, we had designers who’d navigated the healthcare system themselves and felt this gave them supreme authority over what “all patients” needed. They’d sometimes reject user research findings that contradicted their personal experience, pulling up the ladder behind them with a “I managed it, why can’t they?” attitude.

When I notice myself doing this – and I catch myself at it regularly – I’ve learned to challenge my assumptions. I wiggle them like loose Jenga blocks, testing whether what seemed solid in my understanding might actually be ready to topple over when examined honestly.

What trauma does to information processing

Let’s get something straight: when people are traumatised, terrified or vulnerable, they simply cannot process information the same way. Full stop.

I saw this firsthand at Cancer Research UK – and later when my dad developed cancer. Content that tested beautifully with general audiences completely overwhelmed people who’d just received a diagnosis. Their cognitive capacity was decimated by fear. They couldn’t retain information. They couldn’t weigh options. They needed something entirely different.

Now, imagine how much that is amplified for people who think that they have been spiked with drugs or alcohol – or both – late at night in a bar or club. How do we create content designed for people “living” the experience when we can realistically only research and test it by talking to those with “lived” experience? I did – and victim support organisations described the result as “excellent”: “We wouldn’t change a word”. That’s the real litmus test for trauma-informed content design.

This isn’t about making things “emotionally sensitive”, but about making them functional. If your users can’t absorb what you’re telling them because they’re in fight-or-flight mode, your content has failed no matter how grammatically perfect it might be.

What actually works in practice

Forget the academic models for a minute. Here’s what I’ve found genuinely works when creating content for vulnerable people:

• Give control back immediately.
  Trauma strips away control. Your content must return it from the first moment. At the Met Police, every page of the domestic abuse advice had a prominent exit button. Not because it looked nice, but because someone being watched by their abuser might need to leave instantly. Victim support organisations praised the spiking service because “all the way it gave the victim the choice of what to do next”. It gave them control back.
• Tell people why you’re asking.
  Every question feels invasive when you’re vulnerable. “We need your income details to check what benefits you qualify for” is infinitely better than “Section 3: Financial Information (Required)”. When I wrote Universal Credit, I typically spent four days a week locked in a room with forty experts and me as the content specialist. We would discuss on average one question a day. Much of the debate was because I questioned why the information was needed in the first place. For housing benefits, we were able to scrap seventy percent of the questions as a result. That matters, because people naturally mistrust why you’re asking – especially on government websites.
• Use direct language, not euphemisms.
  At Universal Credit, we discovered people needed absolute clarity. They didn’t want “your claim may be affected if your circumstances change” – they needed “you must tell us within 14 days if you start work or your benefits will stop”. Our research also taught us to put key information on payments or tasks required at the start of our letters and emails. Traditionally, letters from the Department for Work and Pensions concluded with such details, but we found that recipients fastforwarded to the end and skipped the rest.
• Create emotional safety before information transfer.
  No one absorbs complex information when they feel under threat. Start by acknowledging their situation: “Reporting sexual assault is difficult. You can take breaks at any time during this process.” For our spiking advice, we always told people to seek help and medical support first and report later.

The danger of proximity without participation

Being adjacent to a space doesn’t make you a substitute for the people in it. This is a lesson I’ve learned repeatedly, sometimes painfully.

Working on the Police.uk  domestic abuse content gave me proximity to the experiences of domestic abuse survivors. But it didn’t make me one of them.

When a senior stakeholder suggested removing specific descriptions of pet abuse because they made him “uncomfortable”, I didn’t argue from my own authority. Instead, I brought in voices of actual survivors who explained why naming these behaviours explicitly helped them recognise their own situations.

The UK’s policing websites still continue to refer animal abuse to animal welfare charities rather than acknowledge it as a possible indicator of domestic violence – and that makes me uncomfortable.  

Remember that proximity without participation is just observation. And observation without continuous validation becomes outdated remarkably quickly. What was true about how traumatised people processed information three years ago might not hold today. Social contexts change, support systems evolve, even vocabulary shifts.

Why most organisations get this catastrophically wrong

I’ve sat in countless meetings where organisations prioritise their own needs over their users’ emotional safety. Legal teams demand language that protects the organisation rather than supports the user. Tech teams build journeys based on database structure rather than human need. Senior stakeholders water down direct language because it makes them uncomfortable.

Your job is to fight back. Not with vague appeals to best practice, but with specific evidence about how trauma affects decision-making.

When a stakeholder wanted to remove direct language about drink spiking symptoms from Police.uk, I didn’t debate writing style. I showed them research on how spiking victims often don’t recognise their own situation without explicit description. “How spiking can make you feel” became a section of the new service most welcomed by victim support organisations.

How to actually do this tomorrow

If you’re responsible for content that vulnerable people might use, here’s what to do:

• Test with people with current, relevant experience.
  Not just those who’ve “been there” in the past, but those living the reality today. At Cancer Research UK, content that sailed through testing with former patients repeatedly failed with those in active treatment.
• Challenge your own assumptions constantly.
  If you think you understand a user group because of your past work, deliberately look for evidence that might contradict your understanding. The world changes, contexts shift and our knowledge must evolve too.
• Create genuine escape routes at every stage.
  Never trap users in a process. Remember: someone filling in your form might have an abuser looking over their shoulder.
• Read your content aloud.
  Ask: “Would I actually say this to someone sitting across from me who’s just experienced trauma?” If it sounds like corporate speak or legal jargon when read aloud, rewrite it.
• Listen more than you speak.
  The “I’ve managed, why can’t they?” mentality creeps in when we stop genuinely listening to people’s current realities. Success in past trauma-informed work doesn’t exempt you from the need to listen deeply to current experiences.

How you know if it’s working

Standard metrics won’t tell you if your trauma-informed approach is effective. Watch for these signals instead:

• Drop-off patterns.
  When users abandon your journey in clusters around sensitive questions, that’s a trauma-informed red flag.
• Support service contact.
  Are people calling your helpline because they’re confused or distressed by your content?
• Time on page.
  Unusually long time spent on sensitive content often indicates people struggling emotionally with what you’re asking.

But the most telling metric is when victim support organisations spontaneously contact you to say your content felt supportive rather than institutional. That happened with the spiking advice and no usability score could ever mean more.

The uncomfortable truth about AI and trauma-informed content

I’ve tested dozens of AI content tools and they all share the same fundamental flaw: they cannot truly understand trauma because they haven’t lived it. They default to clinical detachment or saccharine empathy, neither of which serves vulnerable users.

This doesn’t mean AI has no place. I use it to generate structural outlines and identify potential gaps. But I would never let AI write final content for trauma survivors without substantial human editing.

The future belongs to those who can combine AI efficiency with genuine human empathy, using technology to handle routine elements while preserving human oversight for emotionally complex content. AI can also analyse vast numbers of disturbing crime dossiers, for example, without suffering burnout or secondary trauma.

The question that matters most

After 30 years in this field, the most powerful question I ask when reviewing content is simple: “If someone was having the worst day of their life when they encountered this content, would it help them or add to their burden?”

Let that guide your work. But keep checking whether your understanding needs updating. What helped someone yesterday might not help them today. The kaleidoscope keeps turning, showing familiar elements in new arrangements.

Because trauma-informed content isn’t just about being kind or empathetic – though it is both those things. It’s about creating content that actually works for people when they need it most.

And that’s not academic theory. It’s the most practical content design principle I know. Even when I think I’ve mastered it, I keep finding new aspects to learn, new assumptions to challenge and new ways my past experience might not fully capture today’s realities.​​​​​​​​​​​​​​​​