Bibliographic reference

Study type

Study quality

Number of patients

Patient characteristics

Aim study/instrument

Prognostic factor(s)

Length of follow-up

Outcome measures and effect size

Psychometric qualities

Additional comments

Source of funding

Barkley RA, Fischer M, Smallish L, Fletcher K. The persistence of attention-deficit/hyperactivity disorder into young adulthood as a function of reporting source and definition of disorder. J Abnorm Psychol. 2002 May;111(2):279-89.

This study is also referred to as the Milwaukee study.

Study design:

Prospective cohort study.

Study quality:

Strengths:

• Long follow-up period.

Weaknesses:

• No clinical diagnosis were made.

• Low sample size

• Relative young ages at follow up.

• Changing DSM-criteria.

• Subthreshold ADHD was compared to control group but not to DSM-III-R group.

n=239

(n=158 diagnosed as hyperactive, n=81 matched community control)

Age at inclusion 4-12 years.

Inclusion criteria:

To be considered hyperactive, these children were all clinically referred and had to:

•have scores on both the CPRS-R and the WWPARS that met or exceeded 2 SDs above the mean for severity for same-age, same-sex normal children;

•have scores on the HSQ that indicated significant behavioral problems in at least 6 of the 14 problem situations on this scale (a score exceeding +1 SD);

•have poor sustained attention, poor impulse control, and excessive activity level;

•have developed their behavior problems prior to 6 years of age.

Setting: US

Loss to follow up:

19 pts were lost to follow up (n=147 Hyperactive and n=73 Control group, total n=220)

Patient characteristics:

• Age at follow up 19-25 y (mean Hyperactive 21.1 (SD 1.3); mean control 20.5 (SD 0.6).

• Gender: Males in Hyperactive group 92% and in control 87%).

Intervention n=147:

Hyperactive group.

Comparison n=73:

A matched community control group.

Aim study:

Comparing the DSM-III-R threshold (8 of 14 ADHD symptoms) to a developmentally referenced criterion (DRC) for recent ADHD symptoms.

The clinical threshold of this DRC is + 2 SD above the mean in an same age population. DRC threshold was set at 5 of 14 ADHD symptoms (parent reports).

This means that instead of 99% (when DSM-III-R norms are used), 95.4% falls within the normal range, and 4.6% receives the ADHD-diagnosis (calculated MH).

DRC criteria based on parent-reports identified more pts with hyperactivity with current ADHD then did DSM-III-R (66% vs 46%).

Are these pts that met DRC-criteria but not DSM-II-R criteria impaired in a meaningful way? If so, then you could argue that they actually have the disorder and not merely a high level of symptoms.

ADHD was defined in 3 different ways in this study:

• DSM-III-R: 8 or more of 14 ADHD symptoms.

• DRC criteria: 5 or more of 14 ADHD-symptoms (parent report).

• Meeting DRC but not DSM-III-R criteria: 5, 6 or 7 ADHD symptoms.

Length of follow-up:

13 years

Outcome measures:

• Structured interview of disruptive behaviour (14 ADHD items, DSM-III-R).

• Parent interview of ADHD symptoms (18 items for ADHD, DSM-IV).

• Conners’ Parent Rating Scale – Revised (CPRS-R).

• Homes Situation Questionnaire.

•Werry-Weiss- Peters Activity Rating Scale (WWPARS).

• High school transcripts.

• Employer ratings of job performance

• Criminal records

• Young Adult Self-Report for Child Behavior Checklist (YASR)

Results:

The DRC-approach identified more hyperactive pts as having current ADHD then did the DSM-III-R (66% vs 46%), both based on parent reports.

Comparison of hyperactivity pts that met DRC-criteria but not DSM-III-R criteria (n=27) were compared with the control group (n=73) on 8 outcome measures on impairment (based on self-report, official records, and employer evaluations, but not on parent reports).

This subgroup of DRC-group had 5, 6 or 7 ADHD-symptoms.

The subgroup was significantly more impaired on 6 of the 8 outcomes (p<0.01) than the control group. The subgroup had:

• Fewer years of education,

• a lower GPA,

• a lower class ranking in their last year of high school,

•more symptoms of ADHD at work,

• a lower job performance rating from their employer, and

•more arrests.

No differences were found in nr of current friends and nr of jobs fired, the only self-report measures. See table 3 on p. 284 in article for mean, SD, F, and p values.  

Additional comments:

This article also investigated differences in prevalence of ADHD comparing self-reports and parent reports.

This 13-year longitudinal

study of hyperactive children into young adulthood found a

low frequency of self-reported ADHD at follow-up, but 5-9 times higher frequency when parental reports were utilized. Depending on the diagnostic criteria used (DSM–III–R vs. DRC), between 46% and 66% of the hyperactive group could be considered to have ADHD in young adulthood.

Source of funding:

None mentioned.

Conclusion:

Pts with DRC or subthreshold ADHD (5-7 of 14 DSM-III-R symptoms) are more impaired (not based on self report) than pts in the control group. They should qualify for the diagnosis of ADHD, they seem to have outgrown the criteria but not the disorder. Age based comparisons may be more sensitive to valid cases of the disorder in adulthood then a fixed symptom threshold for diagnosis that is extended from childhood to adulthood, as in DSM-III-R and DSM-IV.

Biederman J, Mick E, Faraone SV. Age-dependent decline of symptoms of attention deficit hyperactivity disorder: impact of remission definition and symptom type. Am J Psychiatry. 2000 May; 157(5):816-8.

Design:

Cohort study

Study quality:

• Strenghts are the longitudinal design, the number of ADHD-patients, and loss to follow-up is unrelated to key characteristics.

• Weaknesses are the generalisability to other patient groups (girls, ethnic minorities, etc.), no information on inclusion criteria, potential confounding by treatment was not accounted for.

n=128

Patiënt characteristics:

• Diagnosis: ADHD

• Age: children, maximum age 16 years

• Sex: 100% boys

• Etnicity: 100% Caucasion

• Setting: clinically referred patients.

Loss to follow-up:

Originally there were 140 pts, but 12 were lost to follow up. However, no significant difference between groups.

Aim study:

To examine remission with regard to ADHD-symptoms.

3 types of remission;

• Syndromatic: loss of full diagnosis (< 57% of symptoms).

• Symptomatic lower than thresshold (< 36% of symptoms).

• Functional: < than 36% of symptoms and GAF score > 60.

3 symptom clusters, according to the 14 DSM-III-R symptoms for ADHD:

Inattentive (6), hyperactive (4) and impulsive (4).

Follow up was 4 years.

There were 3 measurements that resulted in data for 5 time points:

• Measurements took place at baseline, 1 year follow up and 4 year follow up.

• Time points at onset (retrospectively), baseline, 1 year follow up, 3 years follow up (retrospectively) and 4 years follow-up.

Outcome measures:

Schedule for Affective Disorders and Schizofrenia for School-Age Children – Epidemiologic Version (interview with mother and/or pts).

Results:

Age was significantly related to all types of remission (Wald X2>5.1, df=1, p<0.02): there was an increase in remission with age.

Remission type:

The increase was the highest for syndromatic, then symptomatic, then functional remission.

Symptom type:

The increase of remission in inattentiveness was lower than that of hyperactivity or impulsivity.

Source of funding:

NIMH grant (government).

Conclusion:

These results indicate that differences in reported remission rates reflect the definition used rather than the disorder’s course. They provide systematic support for the clinical observation that hyperactivity and impulsivity symptoms tend to decline at a higher rate than inattention symptoms.

Additional comments:

This study did not investigate the symptom threshold to make an ADHD diagnosis, but the symptom profile of ADHD over time.

Heiligenstein E, Conyers LM, Berns AR & Smith MA (1998). Preliminary normative data on DSM-IV attention deficit hyperactivity disorder in college students. Journal of American College Health, 46, 185-188.

Design:

Cross-sectional

Study quality:

• Strength is a large sample.

• Weaknesses are the non-representative community sample (generalisability), only self-report rating current symptoms only, no impairment measures were used, no clinical diagnosis, and results may not be generizable to a clinic ADHD population.

n=468

Pt characteristics:

• US-college students

• Sex: M 56%, F 44%

• Age: range 17-46, M 20.6 (SD 4).

• Ethnicity: 91% Caucasian.

Loss to follow-up:

No follow-up

Aim study:

To present data that determine thresholds specific to college students and to evaluate the implications of selecting different thresholds for a disorder.

Outcome measures:

• Current DSM-IV symptom count with the ADHD Rating Scale (ARS, 18 items). The ARS was modified.

• The optimal cutoff score for current ADHD symptoms, based the 93 percentile (mean + 1.5 SD). This percentile is the customary threshold for determining differences from the norm in research on ADHD.

• Number of pts meeting the DSM-IV and modified thresholds.

Psychometric quality:

An adjusted (non validated) rating scale was used,

Length of follow up:

No follow up.

Results (n=448):

Mean (SD) symptom count score and cutoff M + 1.5 SD:

• Inattention: range 0-8, M 0.82 (SD 1.6), cutoff 3.2

• Hyperactivity/Impulsiveness: range 0-8, M 1.2 (SD 1.5), cutoff 3.2

Calculated cutoff (M + 1.5 SD):

• Inattention: cutoff 3.2

• Hyperactivity/Impulsiveness: cutoff 3.2

Authors conclude that data indicate that a cutoff score of 4 out of 9 current symptoms of Inattention and Hyperactivity/Impulsiveness would be sufficient to identify a college student as distinct from his or her peers (DSM-IV threshold is 6).

% pts meeting criteria for ADHD:

• DSM-IV (6 of 9): 4%

• Modified cutoff (4 of 9): 11%

Source of funding:

No report on source of funding or conflicting interests.

Conclusion:

This cross-sectional study in college students on current ADHD symptoms suggest that cutoff scores of 4 for Inattentive and Hyperactive/Impulsive symptoms could be sufficient to identify a college student as distinct from the norm.

Additional comments:

Symptoms counts above the threshold do not necessarily denote an ADHD diagnosis.

Kooij JJ, Buitelaar JK, van den Oord EJ, Furer JW, Rijnders CA, Hodiamont PP. Internal and external validity of attention-deficit hyperactivity disorder in a population-based sample of adults. Psychol Med. 2005 Jun;35(6):817-27.

Design:

Cross-sectional.

Study quality:

• Strength is the large community sample.

• Weaknesses are that it is a cross-sectional and not a longitudinal cohort study; low N of patients with full or subthreshold ADHD; only self-report measures; no diagnostic interview or diagnostic assessment by clinician; psychometric quality of scales.

n=1.813

(40% of a sample of 4.517 people who were asked to participate)

Patient characteristics:

• Population based sample.

• Age: range 18-75 years, 37.5% was 30-44 years.

• Sex: 44.7% men

Loss to follow-up:

No follow up.

Aim study:

To present data on prevalence and impairment of functioning of adult ADHD obtained directly from a community study. To estimate the diagnostic symptom threshold in adults, the number of current symptoms was plotted against the amount of impairment.

No follow-up.

Outcome measures:

All self-report measures.

• An adjusted ADHD DSM-IV rating scale measuring current (23 items) and childhood symptoms (3 items),

• a 4-item scale to measure psycho-social impairment,

• and the GHQ-28 to measure general liability to psychopathology.

Psychometric quality:

• Symptom rating scale was adjusted but not validated.

• Impairment scale was not tailored to ADHD.

Results:

Current symptoms cutoff point:

• Pts with >4 inattentive symptoms were significantly more impaired than pts with 0, 1 or 2 symptoms [overall model F(7, 1709)=144.631, p<0.000, ‘number of symptoms’ F(6, 1709)=7.371, p<0.000].

• Pts with >4 Hyperactive/Impulsive symptoms were significantly more impaired than pts with 0, 1, 2 or 3 symptoms [overall model F(7, 1709)=153.963, p<0.000, ‘number of symptoms’ F(6, 1709)=10.973, p<0.000].

• The cutoff of 4 symptoms was found for men and women and young and old subjects separately.

Prevalence ADHD (based on current and childhood symptoms):

• Prevalence increased from 1.0% (95% CI 0.6-1.6) to 2.5% (95% CI 1.9-3.4) when the cutoff was changed from 6 to 4 current items.

• The Hyperactive/Impulsive subtype was the most common in both cases, compared to the inattentive and the combined subtype.

Source of funding:

Funding by foundations, government, non-profit.

Conclusion:

The threshold of 6 out of 9 symptoms may be too restrictive to be applied to adults.

On average, the presence of 4 or more ADHD symptoms of either Inattentive and/or Hyperactive/Impulsive symptoms is associated with significantly increased self –perceived psychosocial impairment. This indicates that adults may have on average less symptoms of ADHD than children and adolescents, and that lower symptom levels in adults may not imply better functioning.

Additional comments:

Further epidemiological research including a clinical diagnosis is warranted.

Millstein RB, Wilens TE, Biederman J, Spencer TJ (1997). Presenting ADHD symptoms and subtypes in clinically referred adults with ADHD. Journal of Attention Disorders, 2, 159-166.

Design:

Cross sectional study.

Study quality:

Strength:

• Diagnostic assessment (not symptom rating scales).

• Sample size.

Weaknesses:

• Cross-sectional and not a longitudinal cohort study.

• Generalization: No pts that did not meet full ADHD criteria in adulthood.

• No quantative results on number of symptoms were presented for childhood and adulthood and for type of symptoms.

• No information on significance for reduction of symptoms from childhood to adulthood.

n=149

Patient characteristics:

• Setting: outpatient ADHD adults referred to a clinic.

• Age: 19-60 years, M 37 (SD 11).

• Sex: M 59%.

• 97% had one or more psychiatric co-morbidities.

Loss to follow-up:

No follow up.

Aim study:

An evaluation of the influence of age on the presentation of ADHD in adults.

Pts were assessed and DSM III-R diagnosis were made with an diagnostic interview (SCID) and psychiatric examination. Current and childhood ADHD-symptoms and impairment was assessed.

Outcome measures:

Amount of ADHD-symptoms in childhood (kiddie SADDS-E) and current in adulthood (SCID and DSM III-R criteria).

Results:

• 83% was moderate to severe impaired within the last month due to ADHD symptoms.

• A 11% reduction in nr of ADHD symptoms was found between childhood and current reports.

• No association of age in adulthood and number of ADHD symptoms was found (p>0.05).

• Hyperactive/Impulsive symptoms decreased with 15%, inattentive symptoms decreased with 5%.

• No gender differences were found in current ADHD-symptoms.

Source of funding:

None mentioned.

Conclusion:

The number of ADHD-symptoms between childhood and adulthood reduced with 11%.

Additional comments:

-

Murphy K & Barkley RA (1996). Prevalence of DSM -IV symptoms of ADHD in adult licensed drivers: Implications for clinical diagnosis. Journal of Attention Disorders 3, 147-161.

Design:

Cross-sectional study

Study quality:

Weaknesses:

• Non-representative community sample.

• Cross-sectional and not a longitudinal cohort study.

• No diagnostic interview or diagnostic assessment by clinician.

• Only self-report scales, psychometric quality unknown.

• No ADHD related impairment measures.

n=720

Patient characteristics:

• Community sample of people with a drivers license in the US.

• Age: 17-84 (M35;SD13.2)

• Sex: M 60%, F 40%.

• Ethnicity: 86% white.

Aim study:

Investigating the relationship between DSM-IV ADHD symptoms and impairment scores across age groups.

Follow up: no follow-up

Measures:

• Two self-report rating scales with 18 items. Each item was rated on a 4-point scale (item range 0-3).

o One scale was completed based on the adults current behavior within the past 6 months.

o The second was based on their retrospective recall of their own behavior when they were between 5-12 years of age.

• Impairment was measured with:

o Level of education and

o Social economic status (Hollingshead Index of Social class).

6 outcome measures:

• 3 summation scores (itemscores on Inattention items, Hyperactive/Impulsive items and total ADHD items list).

• 3 symptom counts (positive symptoms, score 2 or 3, on Inattention items, Hyperactive/Impulsive items and total ADHD items list).

Psychometric quality:

Both ADHD-scales were self-constructed measures, psychometric quality unknown.

Results:

• Mean (SD) symptom count score per age group (total ADHD; Inattention; Hyperactivity/Impulsiveness respectively):

o 17-29 y: 3.3 (3.5); 1.3 (1.8); 2.1 (2.0).

o 30-49 y: 2.3 (2.0); 0.9 (1.6); 1.5 (1.8).

o 50+: 1.2 (2.0); 0.4 (1.0); 0.8 (1.3).

• Age was significantly related to all 3 symptom count scores: The higher the age, the lower the amount of symptoms.

o Inattention F=9.14, df=2/674, p<.001;

o Hyperactive/Impulsive, F=19.98, df= 2/671, p<.001; and

o Total ADHD, F=18.16, df=2/645, p<.001.

No gender effects.

• Prevalence:

Based on current and childhood ADHD-symtoms: 4.7% (1.3% inattentive type; 2.5% Hyperactive/Impulsive type; 0.9% combined type).

• Impairment was significant correlated to all 3 summery scores: The higher the level of current ADHD symptoms, the less education obtained and the lower the socio-economic status of current occupation (correlations -.15 to -.21, p<0.001).

Source of funding:

Funding from university, government.

Conclusion:

This study shows that higher degrees of ADHD symptoms were associated with lower educational attainment and lower occupational status.

The current diagnostic threshold for ADHD in adults may be too high. The cutoff of 6 of 9 Inattention symptoms and/or 6 of 9 Hyperactive/Impulsive symptoms recommended

in the DSM-IV sets a threshold of deviance that is quite statistically extreme for an adult population, falling at approximately 2.5 to 3 standard deviations above the mean (>99th percentile) for most ratings of current behavior.

A threshold should be chosen on the extend of impairment associated with them as to be clinically and socially meaningful. The authors argue to set the threshold at a symptom count that is endorsed by for example 7% of the age-appropriate population (mean + 1.5 SD, or the 93rd percentile).

Additional comments:

There is a need for more research in which developmentally referenced criteria are taken into account.