Key moments, in order
We cannot do labs or imaging, so our questions are our tests. Every moment of the interview has its own positive and negative predictive value, sensitivity and specificity, and the meaning is not always in the explicit answer.
Predictive values change with prevalence, so walk in with a pretest probability, based on what you already know, of how likely certain disorders are.
Screening questions are particularly bad in bipolar disorder. Patients who say yes often do not fit once you get the narrative, patients who say no often describe clear-cut mania in the narrative, and some patients simply say yes, or no, to everything.
There is no magic screening question, and no gold question tells you the answer. Gather all the other data to adjust your probabilities, aiming for the greatest confidence you can get for the treatment decision.
Always ask follow-up questions. Even when you specify 48 hours in a row, patients may describe energy that lasted three hours.
Fully structured interviewing has no place outside research; use a semi-structured approach.
Do not stop at yes-or-no. A week of tons of energy, all day? Walk me through that day, and you may hear that they woke up exhausted and started texting everyone at 4 p.m.
It is both overdiagnosed and underdiagnosed: overdiagnosed in patients with a lot of personality pathology such as borderline personality disorder, and severely underdiagnosed in treatment-resistant depression.
Get collateral. It is the main thing that matters in bipolar disorder: a patient can come off as cluster B in the interview, and then a family member or partner describes clear-cut episodes of mania.
Collateral is not always useful. Some people completely fail to notice clear disorganized schizophrenia in a family member they live with.
Anosognosia, the inability to acknowledge the disorder, is part and parcel of bipolar disorder. Bipolar and psychotic disorders come with memory deficits and lack of insight into impairment, so you cannot simply ask and trust the answer.
People with real bipolar disorder can genuinely lack memory of past manic, hypomanic or mixed episodes once they are out of them. We lean toward a state-dependent memory explanation.
In the community, "bipolar" often means the mood swings of someone with personality issues or trauma, so do not trust the label.
Prioritize a thorough medical history, a full medication history including non-psychiatric drugs and, often most important, a detailed substance use timeline. Tell patients that leaving these out can mean a missed or wrong diagnosis and the wrong treatment.
Bipolar disorders largely do not appear in childhood or adolescence, so focus on adulthood from the twenties on. Even so, irritability, mood changes and concentration problems are often present early, before the full syndrome arrives in the 20s or 30s.
With heavy substance use, you can neither assume nor exclude bipolar disorder. Nail down periods of clear sobriety, using the DSM cutoff of at least a month.
Ask for a medication timeline, not a list. Filling in why each medication was switched and what was going on at the time can give you almost a full life history.
Do not stop at do you use any substances. Go through alcohol, then cannabis, then ask whether there was ever a time they used anything else (cocaine, painkillers, methamphetamine), and normalize it.
Explain that substances can affect the brain in ways that confuse both medication treatment and diagnosis, without coming down hard on anyone.
When assessing for bipolar disorder, ask about prescribed and unprescribed medications patients do not think of as drugs. Stimulants are the biggest.
Response to treatment mostly means an adverse reaction to an antidepressant: something weird or strange happened. Antidepressant-induced mania should clue you in to someone at least at risk for bipolar disorder.
Remember unregulated herbs and supplements with stimulant properties, and steroids such as prednisone or other less common prescriptions that can cause mania.
Do not attribute every episode to a substance or medication. People who are hypomanic or manic are more likely to use substances (cannabis and nicotine use often rises during mood episodes) even when those are not the trigger. Some people do have medication-induced episodes, but if you take the substance away and watch them, they may not have another.
Identify a theme in the patient's life early and use it as an anchor to fill in the other variables.
If they cannot give a timeline of the illness, skip to the social history: school, work, gaps, longest job, relationships. Overall functioning over time becomes your timeline and shows more about course of illness and personality than structured DSM questions can.
Ask about mania or hypomania up front only when the pre-interview probability is extremely low. Otherwise, take the history of major depressive episodes first.
The hallmark of a major depressive episode is being different from your normal self almost all day, almost every day, for two weeks, with changes especially in energy, sleep and appetite: more a wholesale change in biological functioning than a subjective state of depression.
Asking "were you depressed?" is the worst possible screener. Use something like the MADRS, whose questions are linked to the DSM criteria, or ask a high-sensitivity screening question and go to a longer rating scale if it is positive.
There is a big gap between patients saying they are depressed and having major depressive episodes. If someone says they are depressed and you go straight to a depression medication, you miss figuring out the diagnosis.
A yes is not the end; it is the start. Tell me about that. What was it like? What was happening? Why do you suppose that happened? When they stay stuck on the symptom, ask them to give you a picture of what was going on.
Build the timeline: how often has this happened, how long does it last, how much time in between, and have you ever been fully free of it for at least a month?
Ask whether they can identify particular episodes, or whether it has been there all their life without much fluctuation. That helps separate major depression from dysthymia or another condition.
If the depression never really goes away, still ask whether there are times it is consistently worse for at least two weeks in a row, and how long those last.
Patients are not answering machines. Questions stimulate memories and they start free associating into a narrative, which is a good thing; we do not always want direct answers.
After taking the depression history, ask what is different when the depression lifts. Patients sometimes start describing hypomania.
Most people with true bipolar disorder mostly experience depression, usually as discrete major depressive episodes repeated across a lifetime. A pure mania or hypomania subtype does exist: about 4 to 6 percent never have depression.
An episodic, recurrent depressive picture raises the possibility of a bipolar spectrum disorder, or at least mixed, refractory or biological-seeming depression, rather than depression from life circumstances or personality.
Kraepelin defined manic-depressive illness by recurrent mood episodes, not by polarity, so today both recurrent major depression and the bipolar disorders would fall under it. What resembles a bipolar disorder genetically and biologically is less the strict DSM definition than a recurrent, episodic, primary mood disorder.
When a patient has recurrent mood episodes and many failed medication trials, make sure there has been a mood stabilizer trial at some point.
Lithium has evidence as an augmenting agent in unipolar depression. Some clear-cut but severe unipolar depressions probably have something a little more bipolar in them, and Europe is much more willing than the US to add lithium to what looks like unipolar depression.
Start with the most extreme version: in your whole life, can you remember a time when you were totally awake, with zero sleep, for at least 48 hours?
Training builds our knowledge on the most severe inpatient cases, where you never see people when they are well between episodes. Many practitioners then think that if it is not severe, near-psychotic bipolar, it is not bipolar, and that is a big mistake.
Sample questions: has there been a time you were not your normal self for at least two or three days in a row, sleeping four or five hours or less? Or a period of at least four days with less sleep, a bit more energy or irritability, when others noticed you were different?
One of us no longer uses the four-day cutoff for screening, because it misses low-grade bipolar-type disorders that could benefit from low-dose lithium for their depression.
The DSM knows low-grade bipolar disorders exist: other specified bipolar disorder recognizes people with short-duration hypomania.
The hallmark is a change in activity level and type of activity together with a change in mood. It is not necessarily the sleep.
Long questions help, because the patient hears the one part that sticks out to them and it helps them elaborate.
Many people interpret having energy as feeling good, so drill down. Patients with agitated depression, very anxious or irritable, may say they had no energy.
The more you look for mixed episodes, the more you see them. Patients who report only depression often describe hypomanic or manic-type symptoms once you dig into their symptoms.
For patients who have used stimulants: have you ever felt like you were on that drug for two or three days when you had not used it in a month?
If a screen is positive, do not jump to the diagnostic criteria. That is confirmatory testing; still get the narrative.
Ask about energy and activity level, but not as a yes or no. Ask what their energy and activity were like during that time, and what they were doing.
People will not count hypersexuality when they are alone, since they think it only counts socially, so ask. For spending, get an idea of their baseline.
Walking through the specific DSM criteria works for a lot of conditions, but for bipolar disorder we do not find it helpful at all.
Instruments such as the Rapid Mood Screener, Mood Disorder Questionnaire, Bipolar Spectrum Diagnostic Scale and Hypomania Checklist help because they pre-write the questions better. Use the modified Rapid Mood Screener; the original's sensitivity is too low.
The Bipolar Spectrum Diagnostic Scale has the patient read a paragraph and see whether the story fits.
Keep adjusting your pre-interview probability of a disorder throughout the interview. Doing that accurately takes clinical experience and crystallized knowledge of the conditions.
Consider medication- and substance-induced bipolar-type disorders, and medically induced ones from strokes, head injuries and seizure disorders.
PTSD is the great imitator and can feel like mania or hypomania. People with a severe trauma disorder can be so hypervigilant that they do not sleep for days, so do not inappropriately treat it as bipolar disorder.
Some patients say yes to everything and some say no to everything, so you need an approach beyond yes or no. Keep the pan-positive patient, who answers yes to most things, on your list of common mimics.
Psychosis that is mostly agitation and insomnia can look like mania, which gets confusing along the schizoaffective spectrum.
In paradoxical insomnia, people believe they literally do not sleep, yet objectively they sleep, with good quality. A sleep study shows it.
Bipolar disorder is a very common reverse mimic: depressive episodes impair concentration, and mania and hypomania cause marked distractibility. Do not hear that a patient cannot concentrate and start a stimulant.
The two get mixed up mostly because people ignore episodes, take yeses at face value, and treat mood lability as a bipolar hallmark. Bipolar is overdiagnosed in patients with borderline pathology and underdiagnosed in treatment-resistant depression.
The reverse happens too. Someone can look perfectly borderline in interpersonal style and defenses while a bipolar process is going on, without pressured speech; the illness is probably aggravating their personality functioning through stress, regression, irritability or paranoia.
Having BPD does not rule out mood disorders; people with BPD are probably more likely to have them. Once a mood disorder is treated, personality tendencies may not come out with the same severity, so monitor and reassess.
Each moment is a point made in the episode, linked to the second it was said. The topic links open everything the podcast has said on that topic.