The issue of cognitive testing of older adults has received considerable attention since President Trump disclosed his own results (“Person, man, woman, camera, TV”). The test he was given, the Montreal Cognitive Assessment (MoCA), is a simple and rapid screening tool and the one most widely used. But how useful are such tests? Geriatrician Mark Williams, MD, talks about their role in determining mental function.
Medscape: Can you briefly review the pros and cons of using a standardized tool — any tool — to assess subtle changes in cognition?
Williams: Evaluating mental function is a critically important clinical skill. Dementia is probably the most feared chronic condition that older people worry about. It creates a huge burden of suffering and is a major public health issue, so doing this evaluation accurately is very important.
There are a number of quick screening tests, and many of them rely on asking patients a series of what I call “magic questions.” The tasks might include listing objects, remembering three items, things of that nature. These tests are quick and easy, and the results — when completely normal — can be reassuring for healthcare professionals. But I worry that they may be measuring what’s easy to measure and not measuring what’s important to measure.
Let me give you a simple example. How many health professionals would feel comfortable with me giving them a 10- or 30-item questionnaire, based on standardized student tests, and using those results to judge that individual’s clinical competence?
I’ve sometimes begun a classroom lecture for physicians with just such a challenge. I ask the audience to name the three hypothalamic hormones that rise during an episode of hypoglycemia. And you have to get all three correct to get credit. There’s usually nervous laughter in the room and an awareness of the underlying premise of judging something as complicated as a person’s memory (or professional competence) by their answers to an arcane set of questions.
My other big concern with these tools is that all items are equally weighted in determining a final score. So if a person scores 25 out of a possible 30 on a screening test — I’m not sure what to do with that number.
Here’s my analogy to this particular dilemma. What if the laboratory in any major hospital decided to aggregate results of a chemistry panel? Instead of giving you each value, the lab reported that person A has 25 out of 30 normal chemistries. Is that person sick or not? If on repeat a year later that patient now has 23 out of 30 normal results, does that mean that they are worse? And the answer is, I don’t know.
I’d be the first to say that if a person has 30 out of 30 normal chemistries, they are probably healthy, but I’m not going to bet the ranch on that. And if they have 10 out of 30 normal chemistries, they are probably pretty sick. But I likely don’t need the lab to tell me that. I can look at the person and see that they’re really sick or not.
Is this concern across the board? Do you think one tool has any advantage over another?
I question the basic premise that any standardized set of questions can allow us to answer the core clinical question that must be assessed when it comes to mental function: Can this person perceive and act on risks in the environment?
To me, that’s the clinically relevant question. We use our mental capacity to appreciate and act on risks. Please note that I did not say appreciate and reduce risks; we regularly choose to take risks. If I can be convinced that a person is aware of a risk and its potential consequences and chooses to take it, then that’s fine.
None of the questionnaire-based mental status tests help me answer that critical, clinically relevant question. I cannot make the conceptual leap of faith that because a person cannot draw intersecting pentagons or subtract 7 from 100, that they’re going to leave a stove on in the kitchen or wander at night or get lost in their familiar environment.
The chemistry panels you referred to earlier are often used as broad screening tools. Everybody gets one when they go in for an annual physical. Of course, they can also be diagnostic. Mental status questionnaires are never diagnostic. But even with the limitations you have noted, do they have a role in screening? Do you think clinicians and patients sometimes mistake them as diagnostic tools?
The tools have an important role in population-based research as a quick way to determine broad population-based norms and the prevalence of a condition such as dementia. To me, the problem is that it’s notoriously treacherous to apply that screening tool to an individual sitting in front of you. If the goal is screening and there is no intent to make a diagnostic assessment, then I think combining the results of a “normal” screen with a more global assessment made during the visit does provide some reassurance. Where I have real problems with the questionnaires is when they are the sole tools used to make a determination that an individual has dementia.
One of the challenges is that our taxonomy is autopsy-based. So essentially when we try to put a label on a condition — does this patient have Alzheimer’s disease or not? — those diagnostic criteria are based on the pathology that a neuropathologist would find during a postmortem exam. And that doesn’t help me or the individual person very much.
Can you describe your usual assessment methodology?
Let me step back before answering that question and give you an example to illustrate why I have difficulty explaining this process.
When I was growing up in rural North Carolina, one of my favorite magazines to read was called Highlights for Children, which had a page that asked the child reader to find the hidden pictures. Maybe a baseball embedded in the image somewhere. Now imagine that the person you’re with can’t find the baseball and asks you to tell them how you were able to find it. That’s the dilemma. How would you teach someone else to become a very good hidden picture finder? The analogy is that I’m looking at the whole picture — that older patient sitting in front of me — and looking for clues that tell me how this individual is functioning cognitively.
I’m looking for clues that help me answer that key question I’ve defined previously: Can this person perceive and act on risks?
And when I “size somebody up,” I generally look at their appearance, dress, language, and behaviors — and appreciate whether that all gels.
Generally speaking, like in the hidden picture finding in Highlights for Children, when a person has dementia, something won’t quite fit. There is an incongruity in how they are presenting. For example, I remember seeing an older realtor in the community, very elegantly dressed. His appearance and dress implied someone well-educated and affluent. But when I asked him specifically where I should buy a house in his community, he responded that I should find a good home in a nice neighborhood. His lack of linguistic specificity didn’t match with what I anticipated based on his appearance and dress.
Another example. I asked a well-groomed, neatly dressed patient where he lived. His response was to turn to his wife and ask her to answer me. Again, a discrepancy between the language and the linguistic content and the person’s appearance and dress.
Is that methodology equally applicable to somebody who is from a different ethnic background, perhaps less educated, maybe does not speak English as their primary language?
Yes, I believe it is. Remember, the goal of my assessment is not to make a diagnosis of dementia. My goal is to appreciate the truth in the light of the moment, the reality behind the appearance. Say I determine that a patient has a problem with memory. Is it a delirium? Encephalopathy? If so, how far can I go in interpreting the details of this picture that the person is giving me?
My basic premise is that each of us holds a picture of who we are (or think we are), and this self-expression is shared with others through the choices we make. There are no random events. The man who comes into my office wearing his Phi Beta Kappa tie pin is telling me something. The person with their sign of the zodiac necklace is telling me something. I’m not making value judgments. Rather, I’m just taking advantage of the appearance, the dress, the language, and the behavior in order to gain some insight into the person.
So let me just give you some basic things that tend to go through my mind in the first millisecond or two when I see somebody. This is not meant to be comprehensive, but it gives you the flavor. The first thing that I ask myself, rhetorically, is, am I able to capture this person’s attention? If I can, how much energy do I have to use to hold their attention? If I have difficulty in capturing and maintaining a person’s attention, then they may have a delirium, and that’s a whole different set of concerns. If I can capture their attention but they have another illness — they’ve just been hospitalized for pneumonia, they have a history of active lupus, or perhaps abnormal vital signs — maybe I’m dealing with encephalopathy or an effect of medications.
Assuming that I can capture their attention and there are no other obvious active illnesses, then the second question is, are there barriers to communication? Does the person speak the same language that I speak? Can they understand me? Can they hear okay? Are there distractions in the room? If any of these are the case, then I have to be very careful about making a judgment about their mental function and probably should just say there’s a communication barrier and try to remove or eliminate that barrier before we go further.
Then I look at overall appearance. Is there facial asymmetry? Does the person look parkinsonian? Do they have a masked face? Is there a resting tremor?
And on and on. There are clues right in front of the attentive clinician that can have important diagnostic specificity.
But each of those clues is just one piece of the puzzle; it’s not the whole puzzle. The goal is not to make a final determination about the person’s cognitive function or even apply a particular diagnostic label as soon as possible. It’s to be accurate, not fast.
Is there a role for looking beyond that patient to the patient’s family and how the patient may interact — or not — with their family member? Or should this process be solely between clinician and patient?
I think it is important at some point to have the discussion with the individual alone. But having said that, if you walk into the room and there are four additional family members present, then that tells me something. This is a gross oversimplification, but if the person comes in by themselves and says, “Doc, I live alone and I think I have a problem with my memory,” they probably don’t. If a person comes in with two family members and the person’s spouse says, “I’m concerned about my loved one,” then they probably do have a problem.
I do pay attention to the physical distance between patient and family, which I think can parallel the emotional distance. So if the person is sitting on one side of the room and everybody else is sitting as far away from them as they can, that tells me something. Family members who jump in and try to interrupt and help to answer questions for the patient tell me a lot.
At some point I usually say to any family present that I’d like to speak to your family member right now. After I talk with them, then I’ll ask you some questions, too. But let them answer for now.
I will then say, “Now, I’d like to do a few things, so please leave the room and let me have a few minutes with your parent or spouse alone.”
Over the course of 40 years, I have seen just about every response you can imagine. I’ve seen older adults ask, “Are they gone? They’re trying to put me in a nursing home. You’ve got to help me out. I don’t know what to do.”
I’ve had others who appear more befuddled because the family is covering for them.
Every permutation and combination you can imagine is possible. It’s worth having a little bit of time just to get a feel for the dynamics and how the person feels about whether the family is trying to help, trying to impede, or trying to manipulate them for secondary gain.
Given that most PCPs are not geriatricians and that visits typically last only 10 to 15 minutes, do you think this methodology is reasonably reproducible for most clinicians?
I would say, first off, that determining somebody’s mental function is one of the most important things that a physician is ever asked to do. Remember, if two physicians agree that somebody has a problem, that individual’s legal rights could be completely taken away and other people can be substituted in for their judgment. So I think it’s really critical.
I would also repeat my emphasis on the fact that are we not measuring something that is easy to measure. The MoCA is easy to do. The Short Portable Mental Status Questionnaire is easy to do. You get a nice comfortable number, but it doesn’t really tell you what you want to know.
I would argue that the assessment I’ve outlined does not take any extra amount of time. It’s focusing your time and energy on what’s important to focus it on. Like anything else in medicine, it might take someone inexperienced a little longer to do it. It is just like learning to ride a bicycle. It takes a little longer to do it until you get comfortable just riding it.
I really believe that the person and the family can tell that you’re paying attention, that you’re giving this encounter the most precious thing you have, which is your total undivided attention. I will say to the person and their family, “We’re not going to have a marathon visit here. I just want to get to know what’s important to you and your fears, concerns, and strengths.”
I will emphasize that I’m going to take my time and we’re going to try to get to the bottom of things. This may mean I’ll need to ask them to return for another visit or even two or three visits. I’m not in a hurry to put a label on this situation. Eventually we’ll get to the bottom of whatever it is that is going on, and then together we’ll go from there.
When you couple taking your time, focusing, and add in a careful, conscientious physical examination, you have real power to help resolve the predicament.
I’m not trying to guess what the pathologist would find. Don’t try to have me pin a label on someone. Sometimes early on I might say to a family, “This doesn’t look like Alzheimer’s.” Everybody will say, “Phew, I’m glad it’s not Alzheimer’s.” And I might say, “Well, this still might be an incurable problem,” and they respond, “Yeah, but it’s not Alzheimer’s.”
People are terrified of that word, aren’t they?
They really are, and it’s a labeling phenomenon. Again, I think that there’s a lot of fear combined with a lot of, unfortunately, hype around drugs that are oversold. It ends up creating more harm than good in the long run.
There has been so much recent attention on standardized assessment tools that the developer of the MoCA, neurologist Ziad Nasreddine, and his peers are growing increasingly concerned that the test might not be as accurate anymore, because too many elements have been shared online. You have already expressed your skepticism about this tool. Do you think this widespread press will ultimately render it useless?
I think that if a screen like this is recognized as just that ― a screen ― and the results completely normal, then that’s helpful, but not sufficient. So I don’t know that I’d want for the tools to be considered completely useless. But the problem is for me that life is short and there’s no reason to spend a lot of time doing something that’s not going to help you.
A simple screening tool doesn’t really tell you anything that you don’t already know if you’re paying attention to the patient in front of you.
Mark E. Williams, MD, is an internist and geriatrician with an active clinical practice in Wilmington, North Carolina. He is the author of five books and more than 100 articles on various aspects of geriatric medicine. He is the author of the series The Art and Science of Aging Well.