Bipolar Test: How the MDQ Works and How Accurate Screening Is

There is no blood test, brain scan or genetic test that diagnoses bipolar disorder^[1]. What people usually mean by a bipolar test is a short self-report screening questionnaire, most often the Mood Disorder Questionnaire (MDQ), which flags who needs a full evaluation^[6].

The diagnosis itself is clinical. According to NIMH, a health care provider may complete a physical exam, order medical testing to rule out other illnesses and refer the person to a mental health professional; bipolar disorder is then diagnosed from the severity, length and frequency of symptoms over the person's lifetime and their family history^[1].

Medical tests matter because other conditions can look like bipolar disorder. NIMH notes that thyroid disease can cause similar symptoms and that some prescribed, recreational or illicit drugs can mimic or worsen mood symptoms^[1].

This article covers the screening tests: what they ask, how they are scored and how often they are right. The full diagnostic process, the difference between bipolar I and II, and common misdiagnoses are covered in how bipolar disorder is diagnosed.

The MDQ was published in 2000 as a brief self-report screen for bipolar spectrum disorders, a group that includes bipolar I, bipolar II and bipolar disorder not otherwise specified^[3]. It takes less than five minutes to complete^[6].

Its first question lists 13 yes-or-no items, each asking whether there was ever a period when the person was not their usual self and, for example, felt much more self-confident than usual, got much less sleep without missing it, had racing thoughts, or spent money in a way that caused trouble^[6]. The remaining questions ask whether several of those experiences happened during the same period, how much of a problem they caused, whether blood relatives have had bipolar disorder, and whether a health professional has ever made that diagnosis^[6].

How the MDQ is scored

The standard scoring treats a result as positive when three conditions are all met^[6]:

• Yes to 7 or more of the 13 symptom items^[6]
• Yes to several symptoms occurring during the same period of time
• A rating of "moderate problem" or "serious problem" for the trouble they caused

The MDQ asks about lifetime experience rather than the past few weeks. For that reason, ACOG notes that in pregnancy and postpartum care it needs to be given only once, unlike depression and anxiety scales that ask about the last 7 days to 1 month^[7].

Accuracy depends heavily on who takes the test. In the original 2000 validation, 198 patients at five outpatient clinics that mainly treat mood disorders completed the MDQ, and a score of 7 or more items yielded a sensitivity of 0.73 and a specificity of 0.90^[3].

The developers then tested it in the community. Of 85,358 adults in a nationwide survey sample, 711 were selected for follow-up and 695 completed a telephone diagnostic interview; the MDQ identified only 28.1% of those with bipolar I or II disorder, while clearing 97.2% of those without it^[4]. The authors concluded that the MDQ's performance in the general population differs substantially from its performance in psychiatric clinics^[4].

In practice, a negative result on a community screen does not rule bipolar disorder out^[4]. Part of the reason is recall: NIMH notes that hypomanic episodes may make a person feel very good and productive, so they may not feel like anything is wrong^[1].

Pooled results across studies

A 2022 meta-analysis in the Brazilian Journal of Psychiatry pooled 93 studies with 62,291 participants^[5]. At recommended cutoffs, pooled sensitivity was 0.71 for the MDQ, 0.75 for the Hypomania Checklist (HCL-32) and 0.71 for the Bipolar Spectrum Diagnostic Scale (BSDS), with pooled specificities of 0.77, 0.63 and 0.73^[5].

Two newer tools, the bipolarity index and the Rapid Mood Screener, showed higher overall accuracy in that analysis, and the MDQ performed best of the group for detecting bipolar II disorder^[5]. The authors still concluded that any positive screen should be confirmed by a clinical diagnostic evaluation^[5].

A positive screen means bipolar disorder is worth assessing, not that it is present. In a Rhode Island study of 480 psychiatric outpatients who completed the MDQ and full diagnostic interviews, 98 screened positive: 23.5% of them were diagnosed with bipolar disorder and 27.6% with borderline personality disorder^[2].

Among patients without bipolar disorder, borderline personality disorder was diagnosed four times as often in those who screened positive as in those who screened negative, 21.5% versus 4.1%^[2]. The authors concluded that a positive MDQ was as likely to indicate borderline personality disorder as bipolar disorder^[2].

Other conditions overlap too. In a 2005 JAMA study of 1,157 adults at an urban primary care clinic, 68.2% of those who screened positive had a current major depressive episode or an anxiety or substance use disorder^[8]. NIMH lists anxiety disorders, ADHD, substance misuse and eating disorders as conditions that commonly co-occur with bipolar disorder^[1].

That is why the screening result is only a starting point. To avoid misdiagnosis, NIMH says it is important to consider the course of the disorder over the past days and weeks rather than focusing solely on current symptoms^[1].

Online bipolar tests are self-scored checklists of mood-elevation symptoms. Their accuracy is known only when the questions match a published, validated instrument such as the MDQ, and even then the evidence comes from studies where trained interviewers checked the results^[3][4].

Three limits apply to any online version. First, a community sample is the setting where the MDQ missed the most cases, with 28.1% sensitivity^[4]. Second, the questionnaire depends on recalling past episodes, including hypomanic periods that family and friends may notice more readily than the person does^[1].

Third, the instrument itself says it is not diagnostic: the MDQ form describes it as a starting point and not a substitute for a full medical evaluation^[6]. Answers to the family history question also carry weight in a real evaluation, since NIMH describes bipolar disorder as often running in families^[1].

NIMH suggests keeping a life chart or mood journal to help recognize mood swings, a record that can give a clinician more detail than a one-time quiz^[1]. The same gap between a screening quiz and a diagnosis applies to ADHD tests.

The strongest clinical reason to screen is before treating depression. NIMH warns that if an initial depressive episode is treated with an antidepressant alone, without a mood stabilizer, a manic episode or rapid cycling may be triggered^[1].

That risk drives screening in pregnancy and postpartum care. ACOG's 2023 Clinical Practice Guideline 4 recommends screening for bipolar disorder before starting medication for anxiety or depression, if it has not been done already^[7]. ACOG's toolkit adds that 1 in 5 women who screen positive for depression may have bipolar disorder, citing a 2013 JAMA Psychiatry study, and states that consultation with or referral to psychiatry is generally indicated when bipolar disorder is suspected^[7].

Statistics on postpartum depression, one of the conditions perinatal screening targets, are covered on its own page.

Primary care is the other major setting. In the 2005 JAMA study, 9.8% of adult primary care patients screened positive for lifetime bipolar disorder; 72.3% of them had sought professional help for their symptoms, but only 8.4% reported ever receiving a bipolar diagnosis^[8].

In the same study, just 6.5% of patients who screened positive reported taking a mood-stabilizing medication in the past month^[8]. The authors called a positive bipolar screen in primary care common, clinically significant and underrecognized^[8]. For national figures on how many adults have bipolar disorder, see the bipolar disorder statistics page.

Bipolar symptoms usually start in late adolescence or early adulthood, and children occasionally have them, according to NIMH^[1]. Screening younger people raises a different question: whose answers to trust.

A 2004 study of 318 youths aged 5 to 10 and 324 aged 11 to 17 compared six screening measures against a semistructured diagnostic interview^[9]. Parent-report questionnaires performed best, with areas under the curve of 0.78 to 0.84, compared with 0.57 to 0.70 for teacher reports and 0.67 to 0.71 for adolescent self-reports^[9].

The authors concluded that parent report was more useful than teacher report or adolescent self-report and could reduce false-positive diagnoses^[9]. Combining several tests did not meaningfully improve classification^[9].