National Survey Shows High Prevalence of Impulse Control Disorders

November 1, 2005

Published a decade ago, the original National Comorbidity Survey focused largely on anxiety and depression. In an exclusive interview, the survey's designer, Ronald C. Kessler, Ph.D., talks with Psychiatric Times about the just-published replication study, which found that the combined lifetime prevalence of impulse control disorders is higher than that for either mood or substance use disorders.

Psychiatric Times

November 2005


Issue 13

Although most previous epidemiological surveys of mental disorders in adultsneglected to track such impulse control disorders as attention-deficit/hyperactivity disorder, intermittent explosive disorder (IED), conduct disorder(CD) and oppositional defiant disorder (ODD), the recently completed National Comorbidity Survey Replication (NCS-R) found that theircombined lifetime prevalence is higher than that for either mood disorders orsubstance use disorders.

"Surprisingly, impulse control disorders ... were found in 8.9% (12-monthprevalence) and 24.8% (lifetime prevalence) of the population with a greaterproportion at the serious level than either anxiety or substance disorders,"noted a commentary accompanying publication of four papers on the NCS-R in the Archives of General Psychiatry (Insel and Fenton, 2005).

For comparison, the lifetime prevalence of mood disorders was 20.8% and ofsubstance use disorders was 14.6%. Only the lifetime prevalence of anxiety disordersat 28.8% exceeded that of impulse control disorders (Kessler et al., 2005a).

The $20 million NCS-R was a collaborative project between Harvard UniversityMedical School,the University of Michigan Institute forSocial Research and the National Institute of Mental Health Intramural ResearchProgram. It is a nationally representative survey of 9,282 English-speakinghousehold residents age 18 and older in the coterminous United States.

Ronald C. Kessler, Ph.D., professor of health care policy at Harvard MedicalSchool and designer of the NCS-R, pointed out that all major psychiatricepidemiological studies in the past, particularly the Epidemiologic Catchment Area (ECA) study in the 1980s (Robins and Regier, 1991) and the National ComorbiditySurvey (NCS) in the 1990s (Kessler et al., 1994), have essentially focused onanxiety and depression.

"Over the last decade, we have come to recognize that violence is not onanybody's radar screen, so I thought that when we did this survey, we reallyneeded to talk about hostility, not just anxiety and sadness," Kessler told Psychiatric Times in an exclusiveinterview. "People who study moods think of it as a triumvirate--anxiety,hostility and depression, but hostility seems to have fallen off the radarscreen in psychiatric epidemiology, so that was the simple-minded notion when Ifirst started approaching this issue."

The NCS-R survey conducted from February 2001 to April 2003 implementedseveral methodological innovations, according to Kessler. It used the internationalWorld Health Organization Composite International Diagnostic Interview(WHO-CIDI), a fully structured, lay-administered interview, to generate DSM-IV diagnoses, but then addedin-depth clinical validation of field research diagnoses based on the clinician-administeredStructured Clinical Interview for DSM-IV(SCID) reinterviews, dimensional self-ratings onclinical anchored scales, inclusion of subthresholddiagnostic syndromes, assessment of severity, assessments of service use andevaluations of treatment adequacy along with an expanded set of diagnoses.

Kessler said he grouped ADHD, IED, CD and ODD as impulse control disorders,because he thought that the most defining characteristic that they shared wasthe impulsivity. He noted that these disorders have not been well studied inadults, particularly IED.

"We have an enormous number of articles about panic attacks ... but you cancount on one hand all the things that have been done about anger attacks," hesaid.

The lack of research on IED is mystifying to Kessler, because "when you lookat the data, you find it is roughly as prevalent as panic disorder [12-monthprevalence is 2.6% for IED and 2.7% for panic disorder], it certainly has animpact on people's lives as much as panic disorder, and it certainly has moreimpact on the lives of loved ones than panic disorder."

He acknowledged that many are confused as to whether IED should beclassified as an illness and whose job it is to take care of the people whosuffer from it.

"It is clearly not something that is seen as being in the core ofpsychiatry," he said. "But anger attacks are, in fact, very common in thepopulation. And we find that both people with anger attacks and those withpanic attacks have a strong family history of mood disorder. The anger attackshave an earlier age of onset and are more fundamental in being involved in highcomorbidity," Kessler said.

Also, there is an interesting mating correlation, he added. Women with panicattacks are often married to men who have anger attacks.

With regard to the inclusion of ADHD under the impulse control category,Kessler said, "There is a lot of evidence suggesting ADHD does continue intoadulthood. It is not the same thing as being aggressive, but we were looking atthe fingers that radiate out from that interest in aggressiveness."

Even though conduct and oppositional defiant disorders usually begin inchildhood (median age of onset=11), Kessler said that if you read DSM-IV-TR carefully, it does not sayanything about their persistence into adulthood. So the researchers asked aboutODD and CD characteristics in adults.

At the request of NIMH, NCS-R researchers also conducted a subsampling, asking questions about pathological gambling,but those data are still unpublished.

According to the commentary by Insel and Fenton(2005), nearly half of all lifetime cases of impulse control disorder havenever been treated. The low proportion of cases that ever seek treatment forimpulse control disorder could reflect the perceptions, both on the part of thepeople with the disorders and of society at large, that their problems are lessrelevant to the mental health care system than to other systems (e.g., socialservices, education, criminal justice), wrote Wang and colleagues (2005) in theNCS-R article on treatment.

Kessler said it might be part of a broader problem. "There are a lot ofpeople with complicated comorbidities in treatmentwho say that they are not getting treatment for particular things that aredifficulties. Whether that is just their perception or reality, I don't know,"he said.

For instance, he noted quite a few people in the population have adult ADHD."They have depression, they have anxiety, they have drinking problems, they getdivorced and they get into car accidents. As a result, a lot of them are intreatment for mental health problems," Kessler said. "We ask, 'What are yougetting treated for?' They say, 'Well, my depression or my alcoholism.' We ask,'Are you getting treated for your attention problems?' and they say no. Theydon't know about that."

A similar scenario occurs with IED, Kessler said, where individuals with IEDenter the treatment system because they are getting divorced, keep losing theirjobs, and often are having problems with alcohol and/or depression. "When we askif they are getting treatment for their anger problems, they say no. They don'tknow about that."

Having impulse control disorders as comorbiditieshas clinical significance, according to Kessler. "It turns out that havingthose comorbidities [ADHD or IED] is pretty stronglycorrelated to severity and course of other illnesses, so people who have ADHDor IED tend to be more severely depressed and more persistently depressed,"Kessler said.

Although stating that he is a researcher and not a clinician, Kesslerbelieves that when clinicians are unaware their patients may have impulsecontrol disorders, it is a problem. "Digging into some of theselesser-considered but important issues might have some implications fortreatment planning," he added.

Further information about impulsivity, aggressiveness and violence isforthcoming, Kessler said, adding that a lot of information about hostility wasbuilt into the survey. For example, the researchers gathered as yet unpublisheddata on gun ownership, marital violence, parent-child violence and frequency ofphysical fights.

Other NCS-R Findings

According to the NCS-R reports, about half (46.4%) of all people in theUnited States will meet the criteria for one or more DSM-IV disorders in theirlife, and during any year, one of every four people (26.2%) in this country is "mentallyill" (Kessler et al., 2005a, 2005b).

"There has been a lot of concern among clinicians and policy-makers aboutthe extremely high prevalence estimates that we found, but the study showedthat only about one out of five of people with a mental disorder has a seriousdisorder," said Kessler. (Those numberstranslate to about 5.7% of the adult U.S. population, 22.3% of the 26.2%overall 12-month prevalence--Ed.)

"Those proportions are manageable enough. They really are of a similarmagnitude to people who have serious heart disease or diabetes," he said.

When 12-month prevalence is categorized by disorder, the most prevalentdisorders were specific phobia (8.7%), social phobia (6.8%) and majordepressive disorder (6.7%). When it is categorized by class, the most prevalentclass was anxiety disorders (18.1%), followed by mood disorders (9.5%), impulsecontrol disorders (8.9%) and substance disorders (3.8%). More than 40% of the12-month cases were comorbid, often affecting illnessseverity (Insel and Fenton, 2005; Kessler et al.,2005b).

Of the 12-month cases, 22.3% were classified as serious; 37.3%, moderate;and 40.5%, mild (Kessler et al., 2005b). Although there are higher rates oftreatment among those with serious mental disorders than those with moderate ormild ones, there are still many who are not getting treated, said Kessler.

Among those receiving treatment, he added, there are many cases "where thequality of treatment is clearly not adequate in relation to publishedguidelines." Kessler said, "It seems to me that we have to shine a spotlight onthat important segment of the population and do more for them than what we havein the past."

The proportion of people who reported 12-month mental health service use ishigher now (17%) than compared to the 13% reported a decade ago in the baselineNCS survey (NIMH, 2005). The expansion was mainly in the general medical sector,with more primary care physicians providing psychiatric services.

In the NCS-R, people with mental or substance abuse disorders were morelikely to get treatment from a primary care physician/nurse or other generalmedical doctor (22.8%), or from a nonpsychiatristmental health specialist (16%), such as a psychologist, social worker orcounselor, than from a psychiatrist (12%), though the survey did show that theadequacy of treatment (measured by number of visits) is best when provided bymental health practitioners. About 9.7% sought help from a counselor orspiritual advisor outside of a mental health setting; and 6.9% used acomplementary/alternative source, such as a chiropractor or self-help group.This held true even for those with severe mood disorders.

Traditionally underserved groups, such as the elderly, racial/ethnicminorities, and those with low income or without insurance, had the greatestunmet need for treatment (NIMH, 2005).

"It is clear that given the current state of affairs, it is very difficultfor a primary care physician to deal with the massive influx of people who havehad mental health problems and are coming to them over this past decade. At thesame time, it is also clear that the number of psychiatrists in Americais just too small to see all the people who want to get treatment; this is acomplicated issue," Kessler said.

One possible solution, he said, could involve having one psychiatrist run anorganization that has, for example, four clinical psychologists and 10 socialworkers.

"There are people with different levels of severity ... who are gettingtreatment that makes sense, given the severity of their illness and where theyare in their life course. The psychiatrist is someone who can intervene at allthose levels, but is able to triage in a way that gives the appropriate amountof intervention," said Kessler, adding that such a service delivery model "islikely to evolve over the next decade."

Other Studies

Currently, the NCS-R provides a guide to what is happening in the United States;however, the methodology used in the study is being employed in epidemiologicalstudies elsewhere.

"The NCS-R is one of 30 national surveys being done around the world inconjunction with the World Health Organization's World Mental Health Surveyinitiative. The NCS-R was the first of them, but there are also surveys in mostof the major Western European countries, several countries in Eastern Europe,Africa, Asia, every region of the world,"Kessler said. "More than a quarter of a million people are being interviewedwith this instrument. And we have a consortium of researchers in these variouscountries that are coming together and analyzing these data collaboratively."

In the United States,the NCS-R is just one part of a coordinated program of new psychiatricepidemiological studies that will be completed over the next several years,according to Kessler. These include the NCS-A study of adolescent mental healthin the United States;the NCS-2, a 10-year follow-up of the original NCS; the National Study ofAfrican American Life, and the National Study of Latino and Asian Americans.

The NCS was a nationally representative household survey of DSM-III-R disorders in the three-yeartime span from 1990 to 1992.

The NCS-2 is a follow-up survey of 4,375 NCS respondents (76.6% conditionalresponse rate) reinterviewed in 2000 through 2002.The NCS-2 outcomes include hospitalization for mental health or substancedisorders, work disability due to these disorders, suicide attempts, andserious mental illness.

"In the NCS-2, we are asking what happened to those 18-year-olds now thatthey are 28. How many of them are still depressed, have recovered or havegotten worse? So we are doing the life history kind of story," Kessler said.

Reports are now emerging for the NCS-2. One report looking at level ofseverity was published in the Archives ofGeneral Psychiatry (Kessler et al., 2003).

"We looked at people who were mild cases 10 years ago to see what happenedto them, and we found a lot of them have substantial problems today," Kesslersaid. "The question is, is there something we could do to nip things in thebud, to treat [such individuals] while they still have a mild case?"

Asked about the adolescent study, Kessler said it is difficult to seedisorders declaring themselves in young children, whereas in the adolescentyears, "you can start seeing depression, substance problems and so forth. So weare quite interested in looking descriptively at what the prevalence, severityand correlates of those disorders are, and also in using [these data] as abaseline for following this cohort over time as they enter adulthood," he said.

He noted that in contrast to physical illness, mental illness exacts a heavytoll beginning in adolescence. "The typical person with arthritis has an age ofonset of 55, whereas the typical person with a mental disorder has an age ofonset of 12. What it means is that mental disorders have the ability to make afundamental impact on the trajectory of your life by adversely affecting thebasic building blocks, such as educational attainment, moving into adulthood,marriage, entering an occupation and maintaining a stable career," he said.

For the study, researchers have interviewed 10,000 adolescents (ages 13 to17) and 10,000 of their parents.

"We have a household subsample and a school subsample, and we have to do some complicated blending ofthe samples at the same time as blending the reports of the parents andchildren," Kessler said. "I am hoping that in the next six months, we will haveall the data in perfect shape, so we will be able to publish core papers foradolescents as we just did for the adults, and then move forward from there."


Insel TM, Fenton WS (2005), Psychiatricepidemiology: it's not just about counting anymore. Arch Gen Psychiatry62(6):590-592 [comment].

Kessler RC, Berglund P, Demler O et al. (2005a),Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in theNational Comorbidity Survey Replication. [Published erratum Arch Gen Psychiatry 62(7):768.] Arch GenPsychiatry 62(6):593-602 [see comment].

Kessler RC, Chiu WT, Demler O etal. (2005b), Prevalence, severity, and comorbidity of12-month DSM-IV disorders in the National ComorbiditySurvey Replication. [Published erratum Arch Gen Psychiatry62(7):709.] Arch Gen Psychiatry 62(6):617-627 [see comment].

Kessler RC, McGonagle KA, Zhao S et al. (1994),Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in theUnited States. Results from the National ComorbiditySurvey. Arch Gen Psychiatry 51(1):8-19.

Kessler RC, Merikangas KR, Berglund P et al.(2003), Mild disorders should not be eliminated from the DSM-V. Arch GenPsychiatry 60(11):1117-1122.

NIMH (2005), Mental illness exacts heavy toll, beginning in youth. Availableat: Accessed Aug. 8.

Robins LN,Regier DA, eds. (1991), Psychiatric Disorders in America: TheEpidemiologic Catchment Area Study. New York: The Free Press.

Wang PS, Berglund P, Olfson M et al. (2005),Failure and delay in initial treatment contact after first onset of mentaldisorders in the National Comorbidity SurveyReplication. Arch Gen Psychiatry 62(6):603-613 [see comment].