Mild cognitive impairment: how to detect it, assess it and work with the family

A 71-year-old woman arrives with her daughter. She says she is fine, that her mind wanders like everyone else's. The daughter says that this month she asked the same question four times over one lunch, that she has stopped keeping the household accounts and that last week she missed her cardiology appointment. She still cooks, still meets her friends and still lives alone without help. That gap between what the patient notices and what the family sees, with independence still intact, is the territory of mild cognitive impairment.

This guide is written for psychologists who see these people in private practice: what the condition actually is, how it is separated from normal ageing and from dementia, what can be assessed and what cannot, when to refer before going any further, which interventions have evidence behind them and how to work with the family across the years that follow-up lasts.

What mild cognitive impairment is

Mild cognitive impairment (MCI) describes cognitive performance below what would be expected for a person's age and education, without that decline stopping them from living independently. The APA Dictionary entry places it exactly there: an intermediate stage between expected ageing and dementia.

The criteria used in clinics since Petersen's work are four, and they are worth keeping close because the third one gets forgotten:

• There is a cognitive concern, raised by the person, by a close informant or by the clinician observing them.
• Impaired performance is objectified in at least one cognitive domain, measured with tests rather than impressions.
• Activities of daily living are preserved. They may take more effort, more time, or a written list; but the person does them.
• Criteria for dementia are not met.

The third point is the boundary. Once someone can no longer manage their medication, their money or getting around without supervision, we have left mild cognitive impairment behind and we are somewhere else.

Normal ageing, mild cognitive impairment and dementia

Forgetting a name, walking into a room and not knowing why, or taking longer to find a word are normal at any age and become more frequent after sixty. What changes in mild cognitive impairment is not so much the type of lapse as its intensity, its frequency and, above all, its trajectory.

The last row is the hardest to explain in session and the one that brings most relief when it is explained well. Mild cognitive impairment is not a compulsory waiting room for dementia. Some people progress, others stay stable for years and a minority return to normal performance, particularly when there was depression, a medication or a medical problem behind it that got corrected.

When the complaint does not match the results

Two situations come up often and are handled differently. Someone very worried about their memory whose performance falls within the normative range usually has an anxiety or mood problem rather than a cognitive one; the work goes in that direction, with an agreed follow-up. The reverse case — someone playing everything down while the family describes clear changes and the tests confirm them — is a different scenario: lack of insight into the deficit is itself a clinical finding and makes the informant essential.

Subtypes: not every mild cognitive impairment looks the same

The classic classification crosses two axes and is still useful for guiding prognosis and intervention.

• Amnestic MCI: episodic memory is what is affected. This is the profile most often associated with prodromal Alzheimer's disease.
• Non-amnestic MCI: memory is relatively preserved and what is impaired is executive function, processing speed, language or visuospatial ability. It points more often to vascular causes, Lewy bodies or frontotemporal degeneration.
• Each is then labelled as affecting a single domain or several. Multiple-domain amnestic MCI carries the highest probability of progression.

This distinction is not academic. A dysexecutive profile with preserved memory completely changes the intervention plan and the advice given to the family.

What brings a person with mild cognitive impairment to the consulting room

Very few arrive saying "I'm here about mild cognitive impairment". What you hear in the first session is usually something else:

• Repeating questions or stories within the same conversation.
• Losing the thread of a film or of a group conversation.
• Dropping tasks that used to be effortless: the tax return, the shopping list, keeping track of medical appointments.
• Relying on notes and reminders far more than before.
• Difficulty finding specific words, with the rest of language intact.
• Withdrawing from social activities that require following several conversations at once.

Interview both parties, and separately

The informant provides what the patient cannot, and the patient provides what the family does not see. A while alone with each of them helps. With the person you explore their own perception, mood, sleep, alcohol use and recent changes in their life. With the relative you push for specifics: not "she's worse", but what she did a year ago that she no longer does, since when, and whether the change was gradual or happened from one month to the next. The usual techniques of the clinical interview apply here, with one caveat: ask for dates and examples, because "for a while now" tells you nothing.

What to rule out before diagnosing anything

A proportion of the low scores that reach the consulting room have a treatable cause. Asking about them is not stepping into medical territory: it is gathering information that later goes into the referral. What is worth reviewing:

• Depression. It slows people down, affects attention and working memory and can produce a profile that looks like MCI. It also frequently coexists with it, so ruling out one does not rule out the other.
• Medication. Anticholinergics, benzodiazepines, hypnotics, some antihistamines. The full medication list is a first-order clinical finding.
• Sleep disorders, and obstructive sleep apnoea in particular, which produces attentional lapses and daytime sleepiness.
• Sensory deficits. Someone who cannot hear well performs worse on any verbal test and takes less part in conversation. Uncorrected hearing loss is also one of the modifiable risk factors that has gained most weight in recent years.
• Alcohol and other substances.
• Medical causes: hypothyroidism, vitamin B12 deficiency, poorly controlled vascular risk factors.

MedlinePlus keeps a plain-language page on mild cognitive impairment that works well as handout material for the family after feedback.

The assessment, step by step

The sequence that works best in private practice has five moments: interview with the patient, interview with the informant, screening, domain-by-domain testing and feedback. Between screening and testing there is usually a decision about whether to continue or refer.

A screen is not a diagnosis

The MMSE is still widely used and has a known problem with this condition: its ceiling effect. Someone with an education and genuine mild cognitive impairment can score 28 or 29 out of 30. The MoCA behaves better in this range because it loads more on executive function and delayed recall. Even so, no brief test diagnoses anything on its own: it points, and it decides whether further testing is worth it. If you plan to rely on questionnaires the patient completes at home, the cautions around online psychometric assessment apply here with extra force, because at these ages a relative sitting next to the screen is the norm and contaminates the result.

Which domains to cover

• Episodic memory, with free recall, cued recall and recognition. The difference between the three is the most informative part of the whole assessment.
• Attention and processing speed.
• Executive function: planning, flexibility, inhibition, working memory.
• Language, with naming and verbal fluency, both phonemic and semantic.
• Visuospatial ability and praxis.
• Mood and functional status, with specific scales and with the informant.

Reading the pattern matters more than any isolated score. A free-recall failure that improves with cues points to a retrieval problem, more compatible with a dysexecutive or depressive profile. A failure that improves neither with cues nor on recognition points to a consolidation problem, which is the typical amnestic pattern. Anyone doing neuropsychological assessment regularly will find the organisational side of this circuit in the guide to software for neuropsychologists.

When to refer, and to whom

Psychology does not diagnose a neurodegenerative disease on its own. The workup that rules out medical causes — blood tests, neuroimaging, medication review — belongs to medicine, and the sensible move is to refer as soon as testing objectifies impaired performance. Do it as a priority when the change has been rapid, when there are accompanying neurological symptoms, when hallucinations or marked personality changes appear, when the person is under sixty-five, or when the family describes clear worsening over a few months.

The report accompanying that referral should be short and operational: reason, tests administered, results by domain, functional status according to the informant, and an impression with the specific question you are putting to the doctor. If you want a structure to follow, the article on how to write a psychological report works as it stands.

What can be done: the interventions with evidence

This is a place to be honest about what we know and what we do not. There is no psychological intervention that halts the progression of a neurodegenerative disease. What there is are approaches with measured effects on cognitive performance, mood, independence and quality of life, which is not nothing.

Cognitive stimulation and cognitive training

Three things get used as synonyms and are not. Cognitive stimulation means general, social activities that engage several functions at once; it is the format with most support in mild to moderate dementia. Cognitive training is guided practice on specific tasks: it improves what is trained and transfer to daily life is limited, which is worth saying to the family before they buy a subscription to a brain-games app. Cognitive rehabilitation works on the patient's own functional goals — getting back to using the phone, taking over medication management again — through compensatory strategies, and it is the best fit while independence is still preserved.

In practice, what tends to work in mild cognitive impairment is a mixed plan: two or three concrete functional goals, external supports properly installed (a paper diary, phone alarms, one fixed place for keys and medication) and meaningful cognitive activity the person enjoys, which lasts far longer than a workbook.

Exercise, sleep and vascular factors

Regular aerobic exercise is the best-supported recommendation and the least followed. Controlling blood pressure, diabetes and cholesterol protects the brain through the vascular route. Sleeping well, treating sleep apnoea if present, correcting hearing and vision, keeping a social life and stopping smoking complete a list that the WHO dementia fact sheet includes among modifiable risk factors. None of these guarantees anything on its own; together and sustained over time, they are the best we have.

Mood, anxiety and apathy

A high proportion of people with mild cognitive impairment show affective symptoms, and addressing them changes daily life faster than any memory exercise. Depression in older adults often presents without verbalised sadness: irritability, somatic complaints, withdrawal from activities, slowing. The psychological approach to depression is essentially the same, with shorter sessions, very concrete goals and support from the household for between-session tasks.

Apathy deserves its own paragraph because it gets confused with depression and with laziness, and it is neither. The person is not sad: they have lost initiative. Propose something and they will take part. That changes the intervention: it is not about behavioural activation against low mood, but about structuring the day from outside and reducing the number of decisions they have to make on their own.

Working with the family without infantilising the patient

Feedback is the most delicate moment in the whole process. Three things help.

First, that the patient is present and is the main interlocutor. Talking about them in the third person in front of them is a quick way to lose the alliance. Second, giving concrete information: what was found, what it means, what it does not mean and what happens next, with dates. Third, putting the plan in writing and handing it over, because out of a forty-minute conversation in which the word "memory" has just been used, the family takes home half.

Something similar happens with supports. The household's natural instinct is to do everything for the person, and that accelerates the loss of independence. The practical rule that tends to work is replacing "I'll do it" with "let's do it together" and, once that is no longer possible, always keeping part of the task in their hands.

Caregiver strain

In mild cognitive impairment the family does not yet describe itself as caregiving, and for that very reason it does not ask for help. What does appear is tiredness, guilt about having snapped, arguments about who does what and, very often, anticipatory grief for someone who is still alive but is no longer who they were. That mixture looks a lot like what is worked on in grief therapy, even though there is no physical loss here.

What is worth opening up with the main relative: how tasks are really split between siblings, the limits of what they can sustain, agreed rather than negotiated rest each week, and the external resources available. Alzheimer's associations in most countries keep directories of local organisations with support groups and respite programmes, and referring there is part of the treatment.

Documenting a follow-up that lasts years

A case of mild cognitive impairment does not close in eight sessions. The usual shape is an initial assessment, an intervention phase and reviews every six or twelve months for years. That format has a practical consequence: when you see that person again a year from now you will remember nothing, so what you write today is all you will have.

Four notes per review are enough for the case to read in one pass: test scores with their date, what the informant says about daily life, what has changed in medication or health, and the plan until the next appointment. With that you can compare two assessments eighteen months apart and tell whether there is a trajectory or just noise. On how to build that record without turning it into a drawer of odds and ends, the article on digital clinical records goes into detail.

The practical side of seeing older adults

Some organisational details weigh more with this population than with any other. Appointments first thing in the morning work better. A sixty-minute session can be too much and is sometimes worth splitting. The phone number on file is usually the son's or daughter's, not the patient's. And spaced reviews are easily lost if nothing holds them up: nobody remembers on their own an appointment they booked eleven months ago. All of this is developed in the guide to software for geriatric psychologists.

How My Psico Agenda supports this kind of follow-up

My Psico Agenda does not assess, score tests or interpret results: that is yours. What it handles is the administrative side of a long follow-up, which in these cases is exactly where things get lost.

• Clinical records are free text, so the four notes per review get written as they are, without squeezing them into fields designed for a different kind of case. Each appointment also keeps its own note.
• WhatsApp reminders go to the phone number on the patient's file, which with this population is usually a relative's, and they prevent a good share of the no-shows on reviews booked months ahead.
• The scheduling does not impose a weekly cadence: you can leave a review six months out and block the slots you set aside for stimulation sessions.
• Per-patient statistics show days since the last session, average cadence and attendance, which is exactly what you need to know when you space appointments on purpose and want to spot who has fallen off the map.
• Each appointment can carry its therapy type — neuropsychology among them — and the practice summary later tells you how much of your work goes that way.

Individual plans start at €4.99/month + VAT — the Semilla plan, with 25 patients, 20 automatic reminders and 20 VeriFactu invoices a month — and continue at €19.99 (Júnior) and €29.99 (Sénior), already with unlimited patients. If you work in a team, practice plans start at €124.99/month + VAT with up to five therapists, each with their own account and their own patients. The first month is free, with every feature and without asking for a card, and there is no lock-in. If you work on your own, the self-employed psychologists page sums up what usually matters.