Assessment of Psychosis Part 1: Base Rates and When To Worry

Clinicians often intuitively believe that psychosis and psychotic-like experiences are "rare" in children and adolescents. Let's take a look at the prevalence rates and see if that's true. The lifetime prevalence of schizophrenia is about 1% (range: 0.46-1.2%). For all disorders with symptoms of psychosis - including other thought disorders and mood symptoms with psychotic features - the lifetime prevalence is about 3%. Schizophrenia has a bimodal mean age of onset. The first high peak, in both genders, is between the ages of 15-24. There is a second peak, which is moderate in men and high in women, between the ages of 55-64. Here's the same information shown in a graph from one classic study on prevalence rates (note they did not find the second peak in the men in their sample):


Because the average age of onset peaks in the late teens and early twenties, this means the prevalence rate in adolescents approaches the adult prevalence. That is, the prevalence rate of thought disorders in adolescents is just under 1%. The prevalence rate of all disorders with psychotic symptoms is just under 3%.

To put that in perspective, the most recent estimates of the prevalence rate of autism in children and adolescents - even if you go by the widest criteria - is now 1 out of every 59, or 1.7%. So in adolescents, teenagers experiencing psychotic symptoms outnumbers teenagers with autism by almost 2 to 1. So, psychosis in adolescents is not necessarily what most psychologists think of as rare. In the United States, approximately 100,000 adolescents and young adults (that is, 19-21 year olds) experience First Episode Psychosis each year (McGrath, Saha, Chant, et al., 2008).


However, psychosis in children is extraordinarily rare. When looking specifically at individuals under 19 years of age who develop psychosis, researchers usually define two distinct groups:

  • Adolescent Onset - Onset of symptoms before age 19

  • Childhood Onset - Onset of symptoms before age 13

These two groups have very different prevalence rates, which is extremely helpful to be aware of when estimating how concerned you should be about a child or adolescent you are evaluating. Let's look at adolescent onset first. Adolescent Onset:

  • Prevalence: Approximately 3 in 100

  • Mean Age of Onset: 14.5 years

  • Mean Age of First Diagnosis: 15.6 years

  • Outcome: Similar to adult outcomes

  • Additional Info: Typically a longer Duration of Untreated Psychosis (DUP) than in adults. Average DUP is 17 months, but up to 5 years DUP is not uncommon.


Childhood Onset:

  • Prevalence: Less than 1 in 10,000

  • Mean Age of Onset: No clear peak - increased risk as child approaches adolescence

  • Outcome: Associated with much worse outcomes

  • Additional Info: Highly comorbid with autism spectrum disorders (up to 50%)


So, to reiterate, psychosis in children is incredibly rare. Just going by prevalence rates alone, you should be about 300 times more worried about an adolescent than you should be about a child when the referral question is possible psychosis.


In addition to age, another important risk factor is family history. About 80% of the risk of psychotic disorders is inherited. Here is a helpful table from Chan 2017 showing relative risk:

Keep in mind that schizophrenia and bipolar disorder appear to share the same genetic risk, so family history of bipolar disorder is also extremely important to assess.


To summarize and add to the information above, in studies looking at trying to predict which individuals will go on to develop psychosis (e.g., Carrion et al 2017), the best predictors are:

  1. Baseline age

  2. Family history

  3. Decline in social functioning (covered in part 2 of this series)

  4. Unusual thought content and suspiciousness (covered in Part 3 of this series and below)

  5. Specific neurocognitive markers (covered in part 4 of this series)


While this list seems pretty straightforward, when looking at #4, we are faced with a whole new question about prevalence. Which is: How common are psychotic-like experiences in the general population? Should we be concerned any time a child or adolescent says they are hearing voices, or thinks someone can read their thoughts?


First we should probably start with a quick definition of psychotic-like experiences (PLEs). Psychotic-like experiences are experiences that are unlikely to be real or to reflect reality, in a way that can usually be easily shown to be untrue or inaccurate. In both adults and children, PLEs usually fall into one of four primary categories:

  • Bizarre experiences (e.g., auditory hallucinations)

  • Perceptual abnormalities (e.g., feeling something inside your skin)

  • Persecutory ideas (e.g., feeling you are being watched or followed)

  • Magical thinking (e.g., feeling that if you do a particular action, you can prevent a particular outcome)


In the same way clinicians often think psychosis is extremely rare, clinicians also often think PLEs are quite rare in the general population. They may also believe that PLEs are clear predictors of likely psychosis. However, neither of these commonly-held beliefs turn out to be true.


Psychotic-like experiences are actually quite common, both in the general population and in children and adolescents. Of course, ranges of prevalence rates depend heavily on how the individuals are asked about psychotic-like experiences. For example, if you ask "do you experience symptoms of psychosis?" you get very low prevalence rates, while a question such as "does your mind ever play tricks on you?" will result in very high prevalence rates. In studies I have seen, prevalence rates for psychotic-like experiences vary from 0.6% to 84% depending on how the questions are asked. In most studies, prevalence rates of PLEs for adults & adolescents are about 28% when all types of psychotic-like experiences are included, with only 2% of those individuals having psychosis. Here's a useful table showing prevalence rates of various types of psychotic-like experiences in adults:


In most studies, prevalence rates of PLEs in children range from 6-17%, with many studies showing even higher rates up to 25% (recall that the prevalence rate of psychosis in children is 0.01%). The most fascinating piece of data for me is that only about 15% of children and teens who experience PLEs are bothered by these experiences.


The best source of information about auditory hallucinations in children, in my opinion, is Agna Bartels-Velthuis' and colleagues' longitudinal series of articles. They first assessed prevalence rates of auditory hallucinations in a large sample of children ages 7-8 (in their sample, the original prevalence rate was 9%, with again only about 15% of those children reporting distress associated with their PLEs). They then followed up with these children 5 years later, providing information about the stability and course of auditory hallucinations. In 2016, they followed up with those children again, now 11 years after the original article.


Here are the links to those articles:

A very recent comprehensive review, which covers these studies and many more, is Maijer et al (2019), Hallucinations in Children and Adolescents: An Updated Review and Practical Recommendations for Clinicians.


In most studies, researchers find auditory hallucinations persist longer than 3 months in only about half of children who have them. Only about a third of auditory hallucinations persist longer than 1 year. For example, Askenazy and her colleagues (2007) followed a community sample of 5-12 year olds, 17% of whom experienced auditory hallucinations. About 50% of those children no longer reported these at 3 months, while 70% of the children no longer had symptoms at 12 months. None of the children in the study met criteria for a diagnosis of schizophrenia at any time.


In general, PLEs are not good predictors of psychosis. For example, the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort is a population-based cohort of almost 7000 kids in the UK. When these children were 12, about 6% experienced definite PLEs and 8% experienced probable PLE. However, almost 80% were no longer experiencing those symptoms by age 18 (Hameed et al 2018; Zammit et al 2013). At age 18, while 4% of the cohort had definite PLEs and 4% had probable PLEs, only 1.7% met criteria for a psychotic disorder. So, a lot of kids who experience PLEs do not go on to experience a clear psychotic disorder as adolescents/adults.


In Zammit et al, the odds ratios for having PLEs at age 18 was definitely increased by having probable PLEs (Odds Ratio :5.6) or definite PLEs (Odds Ratio: 12.7) at age 12. However, they concluded that:

"Despite evidence for a continuum of psychotic experiences from as early as age 12, positive predictive values for predicting psychotic disorders were too low to offer real potential for targeted interventions."

Since PLEs are quite common, most clinicians have experience with a concerned parent, teacher, or other adult reporting that the child or teen is experiencing auditory hallucinations, and asking worriedly, "Is this psychosis?" As can be seen from the data above, in almost all cases, the answer is "no" if the child is 12 or younger. In older adolescents, the answer is more likely to be "maybe." However, many more adolescents experience PLEs than experience psychosis.


Of course, some individuals who experience PLEs are experiencing psychosis or go on to experience psychosis. So, a common question I get from clinicians is "When should I be concerned about auditory hallucinations?"


Based on my reading of the literature, this is my list of when you should raise your index of suspicion that the individual is at high risk of psychosis. Specifically, you should be more concerned about auditory hallucinations when they are:

  • Complex (e.g., fully-formed words or sentences, as opposed to a simple hallucination like ringing in the ears or footsteps, or a perceptual distortion like hearing one's name rather than what was actually said)

  • Involve multiple voices (especially if conversing with each other, arguing, or commenting on what the child/teen is doing)

  • Include specific, commanding content (e.g., telling the child/teen what to do, especially if the behavior is out of the child/teen's usual repertoire)

  • Frequent

  • Have distressing content (as opposed to negative or neutral content)

  • Cause impairment in functioning

  • The individuals reports lack of control over the voices, or feels the voices are very powerful

  • The individual reports a clear external attribution (e.g., the child/teen is sure that the voices are not imaginary or their own thoughts, or if they are their own thoughts, it sounds like their thoughts are being said by someone else)


Here's a nice, quick article (written for pediatricians, but useful) Fifteen minute consultation on children 'hearing voices': when to worry and when to refer (Garralda 2015).


Another question I get quite frequently is, "Well, if PLEs are not good predictors of psychosis, what is going on?" Most studies have found that hearing voices before adolescence is associated, when the child is not distressed by the symptoms, with:

  • No disorder at all (e.g., about 65% of a Japanese cohort of 11-12 year-olds who experienced auditory hallucinations (prevalence of 21%) had no DSM diagnosis)

  • Anxiety/depression in some children and

  • ADHD/behavioral disorders in other children


What about cases where the child is distressed and/or is experiencing hallucinations serious enough to merit an ER visit? Edelsohn and her colleagues (2003) found in a series of 62 consecutive children admitted to a psychiatric ER for PLEs, that:

  • 34% had depression/anxiety

  • 22% had ADHD

  • 21% had disruptive behavior disorders

  • 23% had other DSM disorders

  • Notably, 0% were found to have psychosis


Other studies have shown increased rates of auditory hallucinations in PTSD and in kids with Tourette's plus ADHD. Some research has also shown that somatic complaints are often elevated in children who report PLEs. Bartels-Velthius' studies suggest girls are more likely to experience anxiety and distress related in some way to hearing voices. Most children who report hearing voices also endorse an immediate environmental stressor that preceded the onset of the auditory hallucinations (see Bartels-Velthius, linked earlier, or Escher et al 2002).


These patterns of associated psychopathology are slightly different in adolescents and adults. In teens and adults PLEs are most likely to be correlated with mood disorders (e.g., Armando et al 2010).


Given the information above, how confident are you that Roger, our case example, is experience psychotic like experiences? How concerned are you that he may be experiencing psychosis? Let's turn now to part 2, where we look at the usual course of psychosis.






©2018 by Stephanie Nelson, Ph.D.