Building an Accessible Consent Process for Participants with Dementia

The form is four pages long. The participant is 78, lives with moderate Alzheimer's, and has been staring at the same paragraph for six minutes. You ask if she understands what she's signing. She nods. You're not convinced. And you're right not to be.

Building an accessible consent process for participants with dementia is one of the trickiest problems in research ethics and clinical practice, because the standard tools we use — a written information sheet, a signature, a check-the-box form — were designed for people with intact cognition. They fail quietly. Nobody catches it. The consent goes in the folder and the study proceeds.

I've sat on the other side of that table. I've watched a well-meaning research assistant push a pen toward someone who couldn't have told you what month it was, let alone what a placebo-controlled trial means. So let's talk about what actually works.

Can dementia patients provide consent?

Yes — sometimes, for some things. That's the honest answer, and it's the one most forms won't give you.

A diagnosis of dementia does not automatically strip someone of the legal capacity to consent. Capacity is assessed at the moment of decision and for that specific decision. The problem is that "the moment of decision" in dementia care is a moving target. In the morning, after breakfast and a familiar face nearby, your participant may grasp the essentials of a low-risk observational study. By 3 p.m., after a disrupted nap, the same conversation collapses.

What capacity to consent actually means

For a consent to be valid, four elements need to be present. You don't need them all at once forever — you need them at the time of the decision.

• The person has been informed in a way they can process
• They understand what they're agreeing to, in their own words
• They can weigh the choice and communicate a decision
• They are free from coercion — no family member hovering, no researcher's hopeful pause

Notice what's missing: a signature. Writing your name on a line proves nothing about understanding. It proves you can write your name.

The legal framework isn't optional

Where I work, when someone lacks capacity, a legally authorised representative can consent on their behalf — a court-appointed guardian, someone holding power of attorney for health and welfare decisions, or a next of kin depending on jurisdiction. The exact mechanism varies enormously between countries. Some places recognise advance directives that name a research proxy; others don't.

What stays constant: the representative must consent based on what the person would have wanted, not what the family finds convenient. That distinction is where most ethical failures hide.

What are the 5 requirements for informed consent?

This question gets asked constantly, and the answer shifts depending on whether you're talking about law or ethics. Here's the version I use with research teams:

1. Disclosure — you tell the person what the study involves, its risks, its benefits, and what happens to their data.
2. Comprehension — they actually understand it, not just hear it. This is where accessible formats earn their keep.
3. Voluntariness — the choice is theirs, unpressured. In dementia research, pressure often comes from the gentlest directions: a daughter saying "I think it would be good for you."
4. Capacity — the ability to make this particular decision at this particular time.
5. Ongoing consent — the right to withdraw at any point, without penalty, and the duty to re-confirm as circumstances change.

That fifth one is the one people forget. Consent signed in March doesn't cover a participant whose cognition has declined by September. It has an expiry date nobody wrote on it.

Easy-read formats and what they fix

When I first started redesigning consent materials, I made an obvious mistake: I made them shorter. Wrong move. Shorter isn't clearer. Clearer is clearer.

An easy-read consent sheet uses short sentences, one idea per paragraph, large sans-serif text, and images that support the meaning rather than decorate the page. A pictogram of a blood tube next to "we will take one small tube of blood" does more work than three sentences explaining venepuncture.

For participants who can still follow spoken language better than written, a short video walkthrough — recorded by a familiar clinician, not a professional actor — outperforms any leaflet I've tested. The catch? Videos date. A filmed walkthrough from four years ago that references a discontinued procedure is worse than useless.

Consent formats compared

Format	Best suited for	Main limitation	Effort to produce
Standard written form	Mild impairment, no reading difficulty	Fails as cognition drops	Low
Easy-read document	Mild to moderate impairment	Still requires literacy	Medium
Pictogram-supported sheet	Moderate impairment, low literacy	Images need expert review	Medium–high
Video walkthrough	Spoken comprehension stronger than reading	Becomes outdated quickly	High
Supported conversation (verbatim script + prompts)	Any stage, paired with the above	Depends entirely on the interviewer's skill	Low, but training-heavy

What are some nursing interventions for elderly patients with dementia?

Consent doesn't live in the ethics committee's office. It lives in the interaction between a nurse, a carer, or a researcher and the person in front of them. So the interventions that matter most are communication interventions.

Practical nursing approaches that help

• Anchor the conversation in the present — say who you are, why you're there, and what happens next, every single visit. Repetition isn't insulting; it's respectful.
• Use a consistent person — the same face across visits reduces anxiety, which in turn improves comprehension.
• Time it right — mid-morning after a good night's sleep is usually better than late afternoon. Sundowning is real and it will sabotage your consent conversation.
• Provide a quiet space with no competing noise
• Involve carers as informants, not deciders — they know the person's preferences and history, but they should not answer questions on their behalf.
• Watch for non-verbal signals: confusion, agitation, looking at the door, silence that lasts too long

And the worst thing you can do? Fill a silence. In dementia care, silence is information. Give the person ten seconds to process. Ten seconds feels like a year when you're the one waiting. Wait anyway.

How can I effectively communicate with patients with cognitive impairment?

Ask one question at a time. Then stop talking.

That single change — one question, then genuine silence — improved my own consent conversations more than any form redesign. People with cognitive impairment need longer to retrieve information. If you keep filling that retrieval gap with more explanation, you erase the very moment they were about to understand.

Verbal techniques

• Short sentences. Active voice. No conditionals.
• The teach-back method: ask them to tell you, in their own words, what they just agreed to. If they can't, that's your signal — not a failure on their part, but information for you.
• Repeat the key point once, differently. Not louder. Differently.
• Avoid yes/no questions when you're assessing understanding — people with dementia often answer "yes" because it's socially safe.

What about ongoing consent?

This is the piece I see missing from nearly every consent policy I've reviewed, and it's the one that matters most. Capacity fluctuates. Consent should too.

An accessible process tracks capacity across visits and re-confirms consent at each meaningful milestone — a new procedure, a dosage change, a follow-up questionnaire that's suddenly more demanding than what was originally agreed. If capacity has declined, you move to the assent-plus-proxy model: the person's willingness is sought and respected, and the legal representative formally authorises the decision.

Assent isn't a lesser version of consent. A participant who says "no" or pulls away — even without the capacity to formally consent — has told you something. Honour it.

How do cultural and ethical issues shape this?

Where I work, families often want to shield a relative from distressing information, and researchers who insist on full disclosure can fracture the trust the whole study depends on. In other settings, collective decision-making is the norm and the individual consent model itself feels alien.

You can't resolve that tension with a form. You resolve it by asking the family how decisions are made here, documenting that context, and accepting that the person's voice — however small — still needs to be sought, not assigned away.

The part nobody writes down

The best consent conversation I ever witnessed took eleven minutes and involved no paperwork at all until the very last minute. The researcher knew the participant loved gardening, so she framed the study in terms of roses. The blood draw became "a small check on how your body is doing." The participant understood. When the form finally came out, she signed it with a fountain pen her late husband had given her, and she understood that too.

Accessible consent isn't about making the form prettier. It's about the willingness to slow down, to re-explain tomorrow, and to accept that the answer might be no. If your process can't handle "no" as confidently as "yes," it isn't a consent process. It's a recruitment pipeline.