When is it time to take away my parent's car keys?

Published by Unseen Progress, an independent publisher of caregiver research. Last reviewed 2026-05-10. Part of the dementia caregiver research overview.

Short answer. A diagnosis of dementia does not, by itself, end driving — but most people with progressive dementia will need to stop within one to three years of diagnosis, and many sooner (Alzheimer's Association, 2024). The research-backed answer to "when?" is not a single moment; it is a set of observable safety markers, plus a planned conversation that ideally happens before any of those markers appear. Waiting for an obvious crash or a major incident is the most common path, and it is also the worst one — for safety, for the person's dignity, and for the relationship.

What the research says

Driving is among the most cognitively demanding everyday tasks. It requires intact attention, working memory, executive function, visuospatial processing, and reaction time, all of which are affected by progressive dementia. The Alzheimer's Association (2024) and the broader dementia-driving safety literature agree on three points:

1. Mild cognitive impairment and very early dementia do not automatically end driving. Many people drive safely for some period after diagnosis, particularly on familiar routes in light conditions. 2. Risk rises measurably with disease progression, and most people with moderate dementia are unsafe to drive even if they feel competent. 3. Self-assessment is unreliable, especially as the disease affects the very executive functions that would otherwise allow accurate self-judgement. The person with dementia is typically the last to recognise their own decline behind the wheel.

This last point is the one that surprises caregivers and matters most. The Gitlin / Kales / Lyketsos (2012) frame applies here too: the disease impairs the brain's capacity to evaluate itself. Insisting that "they say they feel fine" is the answer to the question is using a measurement instrument the disease has already broken.

What caregivers are actually noticing

The question almost always arrives in one of three forms:

1. "When do you KNOW it's time to take the keys?" — the most common phrasing in caregiver communities, often surfacing after a near miss, a wrong turn on a familiar route, or a passenger's quiet alarm. 2. "They had a small incident — is this the moment, or am I overreacting?" 3. "Their doctor hasn't said anything. Should I be the one to bring it up?"

The research answer to all three is a version of the same thing: by the time these questions are being asked, the threshold for acting has usually already been crossed.

A research-backed framework: observable safety markers

The Alzheimer's Association and the dementia driving literature converge on a set of observable markers that, taken together, indicate it is time to stop driving. No single one is dispositive; clusters of two or more are.

Cluster 1: Lapses on familiar tasks

  • Getting lost on routes driven for years.
  • Confusion at intersections that used to be automatic.
  • Difficulty changing lanes or merging.
  • Driving too slowly or hesitating at green lights.
  • Reading signs late or missing them altogether.

Cluster 2: Incidents and near-misses

  • New dents, scrapes, or unexplained damage to the car.
  • Curb-strikes when parking.
  • Other drivers honking or gesturing more frequently.
  • A near-miss the person describes calmly or does not describe at all.
  • Any at-fault crash, even minor.

Cluster 3: Cognitive symptoms with driving implications

  • Recent worsening of attention, working memory, or executive function.
  • New medication that affects alertness or reaction time.
  • Worsening visuospatial errors (knocking into doorframes, misjudging distance).
  • Episodes of disorientation in time or place.

Cluster 4: Passenger experience

Passengers — partners, adult children, friends — are often the most reliable instrument. If multiple passengers report feeling unsafe, the question is usually answered. The research consistently identifies passenger discomfort as a sensitive early marker, often appearing before clear external incidents.

The earlier conversation: planning ahead of the crisis

The single highest-leverage intervention in the literature is to have the conversation early, before any of the markers above appears. The Alzheimer's Association recommends building a written driving plan close to the time of diagnosis: who will raise the question when markers appear, what the agreed thresholds are, what alternative transport will look like, and who the trusted clinician or driving evaluator will be.

This is the same logic as the memory care thresholds conversation. Decisions made in advance, in calm, with the person's input, are different decisions from those made in crisis, in fear, with the person already cognitively unable to participate. The evidence consistently favours the planned-in-advance version.

Scripts the research supports

For the early conversation (months before any markers):

"I want to talk about driving while we still have time to plan together. The disease is going to affect this at some point. What would you want me to do when that time comes? Who do you trust to give you the honest read?"

For the conversation when markers have appeared:

"I love you. I've been noticing some things I want to talk through together — [one or two concrete examples]. I'd like us to get an evaluation together, not because I've decided anything, but because the safest thing is to get a real read from someone neither of us has to argue with."

The research-backed move is to route the decision through a clinician or formal driving evaluation, not through the family. This both produces a more accurate read and removes the relational rupture of "you took my keys."

What does not work

  • Waiting for the obvious incident. A serious crash is one of the worst possible learning instruments for this decision.
  • Trusting self-report. The disease impairs the brain's self-assessment. "I feel fine" is not the answer to the question.
  • Confiscating keys without a conversation or alternative. This produces grief, anger, and durable relational damage even when the safety judgement is correct.
  • Hoping the doctor will raise it. Many do not. The caregiver often has to start the conversation; routing the formal assessment through the clinician is the next step.
  • Allowing "just short trips" as a compromise. Most serious dementia driving incidents happen on familiar, short, daily trips, not long unfamiliar ones.

What the research suggests doing

1. Have the early conversation now, regardless of current markers. Build a written plan together. 2. Track the markers above honestly. Two or more clusters present is the threshold for a formal evaluation. 3. Route the decision through a clinician or formal driving assessment rather than making it a family argument. The Alzheimer's Association maintains guidance on driving evaluation services, and many DMVs will accept a clinician referral. 4. Plan the alternatives in parallel. Loss of driving is loss of independence; if the only conversation is about stopping, the resistance is rational. Pair the stop with a concrete, dignified alternative — a regular ride from a family member, a paid driver, a transport service.

Related questions

References

  • Alzheimer's Association. (2024). 2024 Alzheimer's Disease Facts and Figures. Alzheimer's & Dementia, 20(5).
  • Gitlin, L. N., Kales, H. C., & Lyketsos, C. G. (2012). Nonpharmacologic management of behavioral symptoms in dementia. JAMA, 308(19), 2020–2029.
  • Mittelman, M. S., Haley, W. E., Clay, O. J., & Roth, D. L. (2006). Improving caregiver well-being delays nursing home placement of patients with Alzheimer disease. Neurology, 67(9), 1592–1599.

---

Unseen Progress publishes long-form caregiver research. See the full dementia caregiver research overview for the complete framework.