Working up inpatient psychosis

Ask the right questions and know the common differentials to speed your diagnosis.


Not all who hear voices have true psychosis, according to Aaron Gluth, MD.

At his talk on “Psychosis for the Internist” at Southern Hospital Medicine 2018, held in Atlanta in October, Dr. Gluth, an ACP Member and assistant professor at Emory University School of Medicine in Atlanta, explained that auditory hallucinations can also occur in patients with personality disorders or recent grief. “And some people just describe an inner monologue as a ‘voice in their heads',” he said.

Photo courtesy of Dr Gluth
Photo courtesy of Dr. Gluth

Temporality can help determine whether an auditory hallucination is psychotic in nature. “If someone tells you that they've heard voices their whole life, and they've functioned fairly normally, it's less likely they have true psychosis,” Dr. Gluth said. On the other hand, “Multiple voices that keep a running commentary of what the patient is doing—that's particularly worrisome for schizophrenia and related disorders.”

If patients suggest the voices are real and are coming from someone or somewhere else, that is also more suggestive of true psychosis, he noted. Visual, tactile, and olfactory hallucinations, on the other hand, may be more suggestive of a medical or substance-induced psychosis. “They can occur in schizophrenia, but they're much less common than our run-of-the-mill auditory hallucinations,” Dr. Gluth said.

Additional symptoms to look out for include delusions, formal thought disorder, and breakdown of ego boundaries, such as ideas of reference, Dr. Gluth said. “What that means is essentially [they] take something that is most likely a coincidence and ascribe a special meaning to it,” he said. People with psychosis often have negative symptoms, which he described as the “4 A's”—anhedonia, alogia, amotivation, and autism-like features—along with social withdrawal.

Asking the right questions

Dr. Gluth suggested several questions that can help elicit symptoms, such as “Do you hear, smell, or see things that others don't? Do you hear voices when no one else is around? How many voices do you hear? What do they sound like or whose voice is it? What do they say to you?”

If a patient has delusions or paranoia, ask “Do you feel safe here? Are folks treating you OK? Do you feel like someone's been watching or keeping tabs on you? Are you worried about anything? Do you have any unusual abilities? Do you feel that others can control or know your thoughts, or vice versa?”

Hallucinations should be differentiated from illusions. “An illusion is seeing or hearing something that exists but thinking that it's something else, the classic example being a coat hanging on a door,” Dr. Gluth said. “You may look at that and think it's a person, but it isn't, and that's much less worrisome for a true psychotic illness than a hallucination.”

Remember to ask others about their impressions of patients, he advised. “These patients are often somewhat guarded, and their disorganized thought process can make them less than reliable historians, so it's really important to reach out and talk to their friends, family members,” Dr. Gluth said. “And also, in the hospital, getting additional information from staff can be helpful as well.”

Observation is a crucial part of the evaluation. Physicians should assess patients' appearance, grooming, and hygiene. For example, patients with psychosis may be wearing multiple layers of clothing that isn't appropriate for the weather. A blunted or flat affect is also common, Dr. Gluth said. Physicians should take care to assess for safety issues. When assessing safety, ask patients if they have thoughts of hurting or killing themselves or others, and if they say yes, dig deeper to assess whether they have a plan, intent, and means, he said.

Differential diagnosis

Dr. Gluth encouraged his audience to take triage of psychosis seriously, since many medical causes, such as delirium and viral encephalitis, are associated with significant morbidity and mortality. “A new diagnosis of psychosis requires extra diligence,” he said. “I would say please triage patients with psychosis and other psychiatric complaints the way we would triage physical complaints.”

The differential diagnosis should include three broad etiologies: general medical, substance-induced (including prescription medications and over-the-counter drugs), and primary psychiatric syndromes, which should always be the diagnosis of exclusion, he said.

Use mnemonics to avoid missing any broad categories, Dr. Gluth advised. While any general mnemonic for differential building can work, he recommended two that are tailored to assessment of altered mental status: AEIOU TIPS (Alcohol/drugs, acid-base; Endocrine, electrolytes; Insulin, excess or deficiency; Overdose, oxygen; Uremia; Trauma, toxin; Infection; Psychiatric; and Seizure, stroke) and I WATCH DEATH (Infection, Withdrawal, Acute metabolic, Trauma, CNS pathology, Hypoxia/hypercapnia, Deficiencies, Endocrine, Acute vascular, Toxin/drug, and Heavy metals).

Older patients are more likely to have a medical cause of psychosis than younger ones. Approximately 3% of general psychiatric inpatients with first-break psychosis will have a medical etiology, but that number jumps to 60% among the elderly, Dr. Gluth noted. “So 60% of late-life psychosis will be either delirium, dementia, medications, or some other secondary cause. If your patient has no prior psych history, especially if they're older, that can be a big tipoff,” he said.

Delirium should always be on the differential diagnosis, Dr. Gluth said, since it “can mimic any psychiatric condition, anything from anxiety to psychosis to insomnia.” He reminded his audience that a patient who is awake, alert, and oriented to person, place, and time can still have delirium. “You can be alert and oriented and that's only one domain of your cognition, but you can still meet criteria for delirium even if you know where you are,” he said.

The CAM (Confusion Assessment Method) and ICAM (Interactive Confusion Assessment Method) are quick cognitive screening tests to help identify delirium or encephalopathy, while the MoCA (Montreal Cognitive Assessment) and the Mini-Mental State Examination (MMSE) are more time-consuming but can help identify more subtle cognitive issues, such as early dementia, Dr. Gluth said. Asking the patient to recite the days of the week and months of the year backwards has a 93% sensitivity for ruling out delirium, he reported.

What to test

While there's no such thing as an obligatory workup for psychosis, a basic standard could include laboratory tests, including urinalysis and electrolytes; a pregnancy test; a urine drug screen or toxicology screen; vitamin B12 and folate levels; and always an HIV test, Dr. Gluth said. “That's really, I think, underutilized, and the CDC says we should be checking HIV in everybody anyway,” he said. Although thyroid-stimulating hormone levels and syphilis screening are fairly low-yield, they are inexpensive, and it's important that those conditions not be missed, he noted.

Other ancillary tests could be useful “if you're looking for something in particular,” Dr. Gluth said. Ceruloplasmin, for example, is fairly inexpensive and detects a potentially reversible condition, while electroencephalography (EEG) can rule out the unusual diagnosis of subclinical temporal-lobe epilepsy presenting as psychosis.

“That's pretty rare, but EEGs are fairly cheap in the grand scheme of things and also can help you identify other things, like delirium,” Dr. Gluth said. Lumbar puncture can be done if there's concern about infection or a paraneoplastic syndrome, and thiamine levels are “always worth thinking about,” since missing that diagnosis can be devastating, he noted.

Be aware that acute coronary syndrome can manifest only with neurocognitive or neuropsychiatric symptoms among elderly patients, Dr. Gluth said. In addition, N-methyl-D-aspartate (NMDA)-receptor encephalitis presents initially or solely as a psychiatric condition, such as psychosis or catatonia, in approximately 5% of cases, he noted.

When considering neuroimaging, remember that it is very low-yield in patients with first-break psychosis, especially if they have a nonfocal neurologic exam, no neurologic symptoms, and a history of head trauma, Dr. Gluth said. “Fairly large MRI studies suggest about 1% to 3% of first-break psychosis folks will have an actionable finding on MRI,” he said. “Personally, I don't necessarily recommend routine neuroimaging unless there's something specific you're looking for.”

Hospitalists might wonder when to call for a psychiatric consult, and Dr. Gluth said it's always reasonable to consider if the option is available, especially in cases that involve safety or diagnostic issues, new or decompensated psychosis, treatment failures, difficult patient or family interactions, or high-risk medications such as like clozapine or lithium. However, he pointed out that the medical acumen of psychiatrists can vary widely, from a seasoned consult liaison to a clinician who focuses on psychotherapy and takes only one week of call per year.

“For that reason, I think it's important to keep in mind that evaluating psychosis is a team effort, and I think really should involve the generalist,” he said. “And that's a role that we should embrace.”