ADHD and Hyperkinetic Disorder Symptoms in Australian Adults

Abstract

Objective: ADHD and Hyperkinetic Disorder (HKD) have the same 18 symptoms, covering inattention (IA), hyperactivity (HYP), and impulsivity (IMP). This study was aimed at providing descriptive scores for the different symptom groups in these disorders and how these scores varied by age and gender, the percentages of individuals meeting the symptom thresholds for the different ADHD types and HKD, the factor structure, and gender invariance of these symptoms in adults. Method: To accomplish this, 852 adults provided self-ratings for a scale comprising the 18 ADHD/HKD symptoms. Results: The findings showed that age and gender had minimal effects on the ADHD symptoms. Also, in terms of symptom counts, 2% had HKD and 6.3% had ADHD (inattentive = 1.6%, hyperactive-impulsive = 2.7%, and combined type = 2.0%). Confirmatory factor analysis (CFA) provided most support for the three-factor HKD model, involving separate factors for the IA, HYP, and IMP symptoms. This model showed full measurement invariance across gender. Conclusion: The theoretical and clinical implications of the findings are discussed.

Keywords

ADHD hyperkinetic disorder adults Current Symptoms Scale descriptive scores incidence rates factor models gender invariance Australia

Studies have shown that ADHD (American Psychiatric Association [APA], 2000) and its counterpart, “Hyperkinetic Disorder” (HKD; International Classification of Diseases–10 [ICD-10], World Health Organization [WHO], 1993) are fairly stable from childhood to adulthood (Barkley, Fischer, Smallish, & Fletcher, 2002) and are valid adult disorders (Biederman, Faraone, Monuteaux, Bober, & Cadogen, 2004; Clarke, Heussler, & Kohn, 2005). To date, there has been little Australian research on ADHD in adults. Based on self-ratings of the ADHD symptoms provided by a group of Australian adults, this study computed descriptive scores for the different symptom groups for ADHD and HKD, the percentage of adults who met symptom thresholds for the different ADHD types and HKD, the factor structure of the ADHD/HKD symptoms, and measurement invariance for the ADHD ratings across gender. It is worth noting that none of these issues have been explored for an Australian adult sample.

The first official recognition of ADHD was in Diagnostic and statistical manual of mental disorders (2nd ed.; DSM-II; APA, 1968). The disorder was labeled “Hyperkinetic Reaction Disorder.” When revised, DSM-III (3rd ed.; APA, 1980) listed three separate dimensions for the diagnosis of this disorder: inattention (IA), impulsivity (IMP), and hyperactivity (HYP); and called it “Attention Deficit Disorder With Hyperactivity” (ADDH). DSM-III-R (3rd ed., revised; APA, 1987) however provided a single list of 14 symptoms, covering IA, IMP, and HYP. The disorder was called “Attention-Deficit/Hyperactivity Disorder” (ADHD). The more recent editions of DSM-IV (4th ed.; APA, 1994) and DSM-IV-TR (4th ed., text revision; APA, 2000) have retained the same diagnostic label “Attention-Deficit/Hyperactivity Disorder.” According to DSM-IV and DSM-IV-TR, the core symptoms of ADHD/HKD are age inappropriate IA, HYP, and IMP. For the diagnosis of ADHD, DSM-IV has 9 IA symptoms, 6 HYP symptoms, and 3 IMP symptoms. DSM-IV suggests that there are three types of ADHD. These are the combined type (at least 6 IA, and 6 HYP and IMP symptoms), predominantly inattentive type (at least 6 IA symptoms), and predominantly hyperactive-impulsive type (at least 6 symptoms from the HYP and IMP symptom groups). Thus, DSM-IV makes no distinction between the HYP and IMP symptoms for diagnosis, thereby implying a two-factor ADHD model: IA and HYP + IMP.

In the current edition of ICD-10 (WHO, 1993), ADHD is called “Hyperkinetic Disorder” (HKD). DSM-IV and ICD-10 have the same 18 symptoms for diagnosis, with only slight differences in the way the symptoms are worded. Both have the same 9 IA symptoms for the diagnosis of ADHD/HKD. Although HKD has the same 9 HYP and IMP symptoms, it lists 5 symptoms under the HYP group, and 4 symptoms under the IMP group. Unlike ADHD, for HKD there are no subtypes, but a single disorder. The diagnosis of HKD requires at least 6 IA symptoms, 3 HYP symptoms, and 1 IMP symptom to be present. Thus, from a diagnostic viewpoint, HKD reflects a three-factor model. This model is comparable with how the 3 core symptom groups are considered separately in DSM-III for the diagnosis of ADDH (APA, 1980).

Although the ADHD/HKD symptoms provided in DSM-IV/ICD-10 are worded for use with children, they are also used for the diagnosis of ADHD/HKD in adults. For this, the symptoms are changed to reflect adult experiences, such as making reference to work instead of school. To date, a number of ADHD self-rating scales for use with adults have been developed for screening ADHD. In many instances, these scales have limited their ADHD item pool to correspond to the 18 DSM-IV ADHD symptoms. Examples of such scales include the Current Symptoms Scale (CSS; Barkley & Murphy, 1998), the Adult Self-Report Scale (ASRS; Kessler et al., 2005), and the ADHD Rating Scale–IV (ADHD-RS-IV; DuPaul, Power, Anastopoulos, & Reid, 1998). As DSM-IV ADHD and ICD-10 HKD symptoms are similar, these scales can also be used for screening HKD. Generally, total scores of at least 1.5 SD above the normative mean scores for the relevant factors (IA and HYP + IMP for ADHD; IA, HYP, and IMP for HKD) are used to identify individuals with evaluated levels of the symptoms groups (Barkley & Murphy, 1998).

ADHD rating scales have also been used extensively in research. One area of research that has used ADHD rating scales is for establishing the number of individuals who meet symptom thresholds for the different ADHD types. For this, the dichotomous rescoring method proposed by Barkley and Murphy (1998) has been used. As an example, in the CSS (Barkley & Murphy, 1998), which has four response categories, coded as 0 = never or rarely, 1 = sometimes, 2 = often, and 3 = very often, the first two response options (i.e., options 0 and 1) are recoded as the symptom being absent (rescored 0), while the next two response options (i.e., options 2 and 3) are recoded as the symptom being present (rescored 1). The recoded scores are used directly to ascertain the presence of each symptom group. These scores can then be used to find the number of individuals who meet symptom thresholds for the different ADHD types. In terms of ADHD, a total score of 6 or more for the recoded scores for only the IA symptom group is taken as having met the threshold for the inattentive type. Similarly, a total score of 6 or more for the recoded scores for only HYP + IMP symptom group is taken as having met the threshold for the hyperactive-impulsive type. A total score of 6 or more for the recoded scores for both IA and HYP + IMP symptom groups can be taken as having met the threshold for the combined type. In terms of HKD, the presence of at least six, three, and one symptoms for the IA, HYP, and IMP symptom groups, respectively, can be taken as having met the threshold for HKD. Using the dichotomous scoring approach, Murphy and Barkley (1996) reported that for a convenient sample of adults, the number meeting symptom thresholds for all types of ADHD was 4.7% (1.3% for only the inattentive type, 2.5% for the hyperactive-impulsive type, and 0.9% for the combined type). For a sample of university students, Heiligenstein, Conyers, Berns, and Smith (1998) reported an overall rate of 4% (2.2% for only the inattentive type, 0.9% for only the hyperactive-impulsive type, and 0.9% for the combined type).

Another area in which ADHD rating scales have been used is for understanding the factor structure of the ADHD symptoms in adults (DuPaul et al., 2001; Glutting, Youngstrom, & Watkins, 2005; Kooij et al., 2008; Smith & Johnson, 1998; Span, Earleywine, & Strybel, 2002). With one exception (Span et al., 2002), the other studies used rating scales that included the 18 DSM-IV ADHD symptoms as well as a number of additional ADHD-related items. The findings in these studies have been mixed. Results indicate support for the two-factor model (DuPaul et al., 2001; Smith & Johnson, 1998), and also a three-factor model, involving separate factors for the DSM-IV IA, HYP, and IMP symptoms (Glutting et al., 2005; Span et al., 2002). It is to be noted that the DSM-IV three-factor model is different from the three-factor model proposed in ICD-10 for HKD in terms of the number of symptoms for the HYP and IMP factors. As already noted, DSM-IV has 6 symptoms in the HYP factor and 3 symptoms in the IMP factor. In contrast, ICD-10 has 5 symptoms in the HYP factor and 4 symptoms in the IMP factor because the symptom for “talks excessively” that is listed as an HYP symptom in DSM-IV is listed as an IMP symptom in ICD-10. For convenience, and to avoid confusion, the three-factor model that is based on DSM-IV will be referred to here as the three-factor DSM-IV model, while the three-factor model that is based on ICD-10 will be referred to here as the three-factor ICD-10 model. Barkley, Murphy, and Fischer (2008) have proposed a different three-factor model. Based on exploratory factor analysis involving a number of samples, their model suggests factors for IA, HYP-IMP, and verbal IMP. The IA factors comprised all the nine IA symptoms. The HYP-IMP factor comprised the six DSM-IV HYP symptoms plus the IMP symptom on “difficulty await turn.” The VI factor comprised the 2 remaining IMP symptoms (“blurts out answers before question” and “interrupts/intrudes on others”).

Span et al. (2002) used confirmatory factor analysis (CFA) to compare the two- and three-factor DSM-IV models, and also a one-factor model. The one-factor model involved a single factor encompassing all symptoms, corresponding to the way the symptoms are to be used for the diagnosis of ADHD with DSM-III-R (APA, 1987). They found that the three-factor DSM-IV model showed better fit than the two- and one-factor models. Kooij et al. (2008) compared the one-, two-, and three-factor DSM-IV ADHD models as well as two- and three-factor models with cross-loadings. The CFA was based on self-ratings of the 18 DSM-IV symptoms and five more ADHD-related items. Kooij et al. reported that the three-factor model with cross-loadings showed adequate fit and was the best fitting model. This model had eight items that cross-loaded on all three factors and another six items that cross-loaded on two factors. For the items that cross-loaded, only the HYP symptom relating to “talks excessively” had a cross-loading of more than .35, and it cross-loaded equally on the HYP and IMP factors. Despite finding support for the three-factor model, Glutting et al. (2005) found that this symptom loaded on the IMP rather than the HYP factor. As “talks excessively” is an IMP symptom in ICD-10, this finding can be interpreted as possible support for the three-factor ICD-10. Overall, although existing data from the CFA studies appear to be more supportive of the three-factor DSM-IV model over the two- and one-factor models, there are some data suggestive of support for the three-factor ICD-10 model. To date, there has been no direct test of the three-factor ICD-10 model.

In research involving children and adolescents, ADHD rating scales have been used extensively to evaluate measurement invariance for the ADHD symptoms across gender (Burns, Walsh, Gomez, & Hafetz, 2006; Gomez, 2007; Reid et al., 2000). Invariance deals with whether the expected scores on a measure are the same or different across different groups when the groups have the same level of the underlying latent trait scores (Reise, Widaman, & Pugh, 1993). Invariance is inferred when the expected scores are the same, while noninvariance is inferred if this is not the case. If there is weak or no support for measurement invariance, then it follows that the groups in question cannot be accurately compared on the given measure for observed or latent scores. For adult ratings of the ADHD symptoms, Smith and Johnson (1998) reported gender invariance for the two-factor ADHD model. However, a close examination of the results reported by these authors reveals a serious computational error. This study used the Satorra–Bentler chi-square statistic (S-Bχ²), which corrects for the lack of normality in the data set, to ascertain statistical fit and differences between nested models. As the difference in two S-Bχ² values is not distributed as a chi-square variate, it is necessary to use the correction formula proposed by Satorra and Bentler (1999) for computing difference in two S-Bχ² values. However, Smith and Johnson did not apply the correction formula in their analyses, thereby invalidating their results. Kooij et al. (2008) reported gender invariance for their three-factor model with cross-loadings. However, as this model involved five more items than the 18 ADHD symptoms and had eight items that cross-loaded on all three factors and another six items that cross-loaded on two factors, it cannot be taken with certainty that there will be gender invariance for the 18 DSM-IV ADHD symptoms. As far as it can be ascertained, no other study has provided acceptable gender invariance data for the DSM-IV ADHD symptoms in adults. This is worthy of exploration given it will have implications for the use of the same measures for diagnosis and screening of ADHD in men and women, as is currently the practice.

There were four major aims in this study. Based on self-ratings of the ADHD symptoms in the CSS (Barkley & Murphy, 1998), provided by a group of Australian adults, the first aim of the study was to compute descriptive scores for the original ratings (summary scores) for the symptom groups of ADHD (IA and HYP + IMP) and the three symptom groups of HKD (IA, HYP, and IMP). As mentioned previously, such scores are useful for screening HKD and the different ADHD types. Associated with this, the effects of age and gender on these scores were examined. The second aim was to use the dichotomous scoring method to ascertain the percentages of adults who met symptom threshold numbers for the different ADHD types and HKD. The third aim was to examine the factor structure of the ADHD symptoms. CFA was used to test the two-factor, three-factor DSM-IV, three-factor Barkley et al. (2008), and three-factor ICD-10 models. The fourth aim was to use the multiple-group CFA procedure to examine measurement invariance for the ADHD ratings across gender. It is worthy of note that so far none of these issues have been explored for an Australian adult sample.

Method

Participants

A total of 852 adults completed ratings of the ADHD symptoms. In all, there were 502 women and 350 men. The mean age of all participants together was 28.32 years (SD = 9.73, range = 18-50 years). The mean age of women (M = 28.16, SD = 9.50) and men (M = 28.43, SD = 9.90) did not differ significantly, t(850) = 0.39, ns. Participants were recruited in Australia through several sources from the State of Victoria. The participants from the general community were recruited from educational institutions, workplaces, sporting venues, social clubs, and shopping centers. In terms of ethnicity, 94% were from predominantly European background, 4% were from an Asian background, and the remaining 3% were from other backgrounds, including Indigenous Australian and Maori. These figures compare with around 90% European, 7% Asian, 3% of Others (including Indigenous Australian) in the general Australian population (Australian Bureau of Statistics, 2007). There was close match in ethnicity between the Australian general population and the group involved in the study, χ²(df = 2) = 0.90, p = ns. In terms of educational background, approximately 93% had completed secondary education.

Measure

ADHD ratings were obtained using Barkley’s CSS (Barkley & Murphy, 1998). This measure contains the 18 symptoms of ADHD and the 8 symptoms of Oppositional Defiant Disorder (ODD), all of which correspond closely with the list of ADHD and ODD symptoms specified in the DSM-IV for these disorders. For this study, the ratings for the ODD symptoms were not used. Participants indicate how often they have experienced each symptom over the past 6 months by circling a number from 0 to 3 (0 = never or rarely, 1 = sometimes, 2 = often, 3 = very often). In the current study, the Cronbach’s alpha for all 18 ADHD symptoms was .89, while for the 9 IA symptoms it was .86, and for the 9 HYP + IMP symptoms it was .81. For the group of 6 HYP and the 3 IMP symptoms as specified in DSM-IV, the values were .72 and .78, respectively. For the group of 5 HYP and 4 IMP symptoms as specified in ICD-10, the values were .70 and .79, respectively.

Procedure

Prior to data collection, ethics approval to conduct this study was obtained from the University of Ballarat Human Research Ethics Committee. Participation was voluntary and anonymous, and consent was implied by the return of the questionnaire. Potential participants were approached randomly in their respective workplaces, at various sporting venues, university campuses, social clubs, and other locations and invited to participate in the study. This involved informing them about the nature of the study and what it entailed. If they were interested in participating, they were given the questionnaire package. Apart from the CSS, this package included other questionnaires (not examined in this article) and a plain language statement about the study. The general instruction for participants was to complete the questionnaires as quickly as possible, without spending too much time on any one question. Participants were also asked to complete the questionnaires by themselves, and to refrain from seeking assistance from parents, partners, friends, colleagues, or others. Participants were informed that they could return the questionnaire package via the reply paid envelopes given to them or directly to the researchers, or at appropriate (secure) locations at their workplaces or campuses. A debriefing statement was also included thanking participants for their time.

CFA Statistical Procedures

All the CFA analyses in the study were conducted using Mplus (Version 4.1) software (Muthen & Muthen, 2006), and they used the mean and variance-adjusted weighted least squares or WLSMV. This is a robust estimator for ordered-categorical scores. Brown (2006) has indicated that this estimator performs well for variables with floor and ceiling effects. Thus, it is well suited for this study because the ADHD items involved ordered category scores, and these scores can be expected to have floor effect because the ratings were from a community sample.

For this study, statistical fit was ascertained using the robust WLSMV fit function χ² and two approximate fit indices. The WLSMVχ² shows the closeness of fit between the unrestricted sample covariance matrix and the restricted (model) covariance matrix, taking into consideration nonnormality and the categorical nature of the scores. The approximate fit indices used in the current study were the root mean square error of approximation (RMSEA) and the comparative fit index (CFI). The guidelines suggested by Hu and Bentler (1998) are that RMSEA values close to .06 or below be taken as good fit, .07 to .08 as moderate fit, .08 to .10 as marginal fit, and >.10 as poor fit. For the CFI, values close to .95 or above are taken as indicating good fit, and values close to .90 and .95 are taken as acceptable fit (Hu & Bentler, 1998). These fit values have also being found acceptable for order-categorical data (Muthen & Muthen, 2006).

Measurement invariance across the groups was tested using the procedure proposed by Millsap and Yun-Tein (2004) for the WLSMV estimator with theta parameterization. This procedure involves comparing progressively more constrained models that test configural invariance, metric invariance, threshold invariance, and error variance invariance. For multiple-group analyses, the differences in the various nested models were examined using the WLSMVχ² difference test that can be obtained directly in Mplus. It is worth noting that the difference in WLSMVχ² of nested models is not distributed as the usual chi-square. An α value of .01 was used to allow for more stringent Type II error control in the models compared (Green & Babyak, 1997).

Results

Descriptive Scores for the ADHD and HKD Symptom Groups for the CSS

Table 1 provides the mean and standard deviation scores for the three different age groups (18-29 years, 30-49 years, and above 50 years) for the summary scores and for the symptom groups relevant to DSM-IV ADHD and ICD-10 HKD. It also includes these scores for all 18 symptoms together.

Table 1.

Mean and Standard Deviation Scores for the DSM-IV ADHD and ICD-10 HKD Symptom Groups for Males and Females and Both Groups Together.

	All	Males	Females
IA (DSM-IV and ICD-10)
18-29 years (M = 223, F = 315)	12.37 (7.49)	7.76 (4.73)	6.18 (4.07)
30-49 years (M = 96, F = 136)	11.10 (8.31)	7.21 (5.12)	5.70 (4.99)
50+ years (M = 31, F = 51)	9.59 (6.90)	5.35 (3.50)	4.73 (3.59)
All (M = 350, F = 502)	11.74 (7.70)	7.20 (4.78)	5.90 (4.31)
HYP + IMP (DSM-IV)
18-29 years (M = 223, F = 315)	6.63 (4.55)	7.27 (4.83)	6.19 (4.30)
30-49 years (M = 96, F = 136)	5.87 (4.22)	6.53 (4.39)	5.40 (4.05)
50+ years (M = 31, F = 51)	4.82 (3.61)	4.74 (3.11)	4.86 (3.91)
All (M = 350, F = 502)	6.25 (4.42)	6.85 (4.63)	5.84 (4.22)
HYP (ICD-10)
18-29 years (M = 223, F = 315)	3.57 (2.57)	3.94 (2.64)	3.31 (2.43)
30-49 years (M = 96, F = 136)	3.19 (2.75)	3.70 (2.87)	2.82 (2.62)
50+ years (M = 31, F = 51)	2.85 (2.47)	2.71 (2.18)	2.94 (2.64)
All (M = 350, F = 502)	3.40 (2.62)	3.77 (2.68)	3.14 (2.54)
IMP (ICD-10)
18-29 years (M = 223, F = 315)	3.06 (2.64)	3.33 (2.87)	2.87 (2.45)
30-49 years (M = 96, F = 136)	2.69 (2.15)	2.83 (2.21)	2.58 (2.10)
50+ years (M = 31, F = 51)	1.96 (1.57)	2.03 (1.47)	1.92 (1.64)
All (M = 350, F = 502)	2.85 (2.45)	3.08 (2.63)	2.70 (2.30)
Total (DSM-IV and ICD-10)
18-29 years (M = 223, F = 315)	13.34 (7.99)	14.73 (8.49)	12.37 (7.49)
30-49 years (M = 96, F = 136)	12.15 (8.56)	13.74 (8.73)	11.10 (8.31)
50+ years (M = 31, F = 51)	9.78 (6.50)	10.10 (5.90)	9.59 (6.90)
All (M = 350, F = 502)	12.69 (8.09)	14.05 (8.44)	11.74 (7.70)

Note. DSM-IV = Diagnostic and Statistic Manual of Mental Disorders (4th ed.); ICD-10 = International Classification of Diseases–10; HKD = Hyperkinetic Disorder; IA = inattention; HYP = hyperactivity; IMP = impulsivity.

The effects of age and gender for the scores provided in Table 1 were examined using 2 (gender) × 3 (age: 18-29 years, 30-49 years, and above 50 years) ANOVAs.

The results are summarized in Table 2. As shown, for all scores examined, there was no Gender × Age interaction effect. There were main effects for gender for only the IA and total scores. For both scores (relevant to both DSM-IV and ICD-10), males scored higher. Based on Cohen’s (1992) guidelines (small = .20, medium = .50, and large = .80), for both instances, the effects sizes were small. There were main effects for age for all the scores, except the ICD-10 HYP symptom group. Post hoc analyses (based on Bonferroni correction for an α of .05) for the other summary scores indicated that in all instances, no significant differences were found between the 18- to 29-year-old and 30- to 49-year-old age groups, with both these groups scoring higher than the above-50-years-old group. However, the effects sizes for all differences were small based on Cohen’s guidelines.

Table 2.

Results of the ANOVAs for Gender and Age Group Differences for the DSM-IV ADHD and ICD-10 Hyperkinetic Disorder Symptom Groups and Total Scores.

		Post hoc comparison
Symptom groups	Results of ANOVA	Gender (d)	Age (d)
IA (DSM-IV and ICD-10)	Gender (G): F(1, 846) = 7.49**	M > F (0.29)	1 = 2
	Age (A): F(1, 846) = 5.35**		1 > 3 (0.38)
	G × A: F(2, 846) = 0.28		2 > 3 (0.20)
HYP + IMP (DSM-IV)	Gender (G): F(1, 846) = 2.98	M = F	1 = 2
	Age (A): F(1, 846) = 7.70**		1 > 3 (0.41)
	G × A: F(2, 846) = 0.64		2 > 3 (0.27)
HYP (ICD-10)	Gender (G): F(1, 846) = 3.11	M = F	1= 2
	Age (A): F(1, 846) = 4.07**		1 = 3
	G × A: F(2, 846) = 1.31		2 = 3
IMP (ICD-10)	Gender (G): F(1, 846) = 1.45	M = F	1 = 2
	Age (A): F(1, 846) = 8.11***		1 > 3 (0.43)
	G × A: F(2, 846) = 0.26		2 > 3 (0.39)
Total (DSM-IV and ICD-10)	Gender (G): F(1, 846) = 6.20**	M > F (0.29)	1 = 2
	Age (A): F(1, 846) = 7.84**		1 > 3 (0.46)
	G × A: F(2, 846) = 0.54		2 > 3 (0.31)

Note. DSM-IV = Diagnostic and Statistic Manual of Mental Disorders (4th ed.); ICD-10 = International Classification of Diseases–10; IA = inattention; HYP = hyperactivity; IMP = impulsivity.

p < .01. ***p < .001.

Percentage of Individuals Meeting ADHD and HKD Symptom Threshold Numbers

The rates of individuals with ADHD and HKD symptom threshold numbers were computed using recoded dichotomous scores. For all ADHD types, the rate was 6.3%. The rates for the inattentive, hyperactive-impulsive, and combined types were 1.6%, 2.7%, and 2.0%, respectively. For HKD, the overall rate was 2.0%. Of the 17 adults who were at or above the threshold for HKD, 15 were also above the threshold for the combined type.

CFA of the ADHD Symptoms

The fit values for the two-factor model were χ²(df = 75) = 604.37, p < .001, RMSEA = .09, CFI =.87. Thus, there was poor fit for this model. The fit values for the three-factor DSM-IV model were χ²(df = 79) = 481.67, p < .001, RMSEA = .08, CFI =.90. This model showed only adequate fit for both approximate fit indices. For the Barkley et al. (2008) three-factor model, the fit values were χ²(df = 77) = 568.15, p < .001, RMSEA = .09, CFI = .88. The fit values indicated poor fit. The fit values for the three-factor ICD-10 model were χ²(df = 79) = 373.66, p < .001, RMSEA = .06, CFI =.93. The RMSEA indicated good fit. The CFA value of this model was acceptable and close to good fit. Overall, there was support for good or at least close to good fit for this model, and this model was the best fitting model.

Measurement Invariance Across Males and Females

As the three-factor ICD-10 model showed good or at least close to good fit and better fit than the one-, two-, and three-factor DSM-IV ADHD models, this model was used for testing gender invariance. For males and females separately, the fit values for the three-factor ICD-10 model were χ²(df = 66) = 210.50, p < .001, RMSEA = .08, CFI =.92, and χ²(df = 79) = 202.24, p < .001, RMSEA = .06, CFI =.95, respectively. Thus, the fit for males was good, while the fit for females was at least adequate.

As indicated previously, gender invariance was examined using the guidelines recommended by Millsap and Yun-Tein (2004). Table 3 shows the results of these analyses. The configural invariance (M1) was tested first. As shown in Table 3, for this model (M1), there was at least adequate fit, thereby supporting configural invariance. The next model tested was the metric invariance model (M2). Table 3 shows that there was no difference between M2 and M1, thereby supporting metric invariance. The invariance for thresholds was tested next (M3). As shown in the table, there was no difference between M3 and M2, thereby supporting threshold invariance. The next level of invariance that was tested was the error variances invariance model (M4). Table 3 shows that there was no difference between M4 and M5, thereby supporting error variances invariance. Overall therefore, there was support for invariance for the full measurement model.

Table 3.

Results of Test for Invariance Across Females and Males.

Model (M)	χ²	df	CFI	RMSEA	ΔM	Δdf	Δχ²
M1: Baseline invariance	431.74***	144	.928	.068		—	—
M2: Full factor loading invariance (M1 with equality constrained on all factor loadings)	439.98***	151	.928	.067	M2 – M1	14	27.03
M3: Threshold invariance (M2 with equality constrained on all thresholds)	391.15***	150	.940	.061	M3 – M2	18	11.74
M4: Full uniqueness invariance (M3 with equality constrained on error variances)	349.40***	142	.948	.059	M4 – M3	16	22.76

Note. χ² = weighted least square with mean and variance-adjusted chi-square (WLSMVχ²); CFI = comparative fit index; RMSEA= root mean square error of approximation.

***

p < .001.

Discussion

The first aim of this study was to provide descriptive scores that could be used to facilitate screening and diagnosis of ADHD and HKD. The mean and standard deviation for the different ADHD and HKD symptom groups and all symptoms together are provided in this article (Table 1). These scores have been presented for three different age groups (18-29 years, 30-49 years, and above 50 years) and across all ages for males and female separately, and together. With the exception of the mean score for ICD-10 HYP, age had an effect on all the other symptom group scores. For all scores, the 18 to 29 years group did not differ from the 30 to 49 years age group, and both these groups scored higher than the above 50 years group. The decline of the ADHD and HKD symptoms with age found in the current study has also been reported by others (Biederman, Mick, & Faraone, 2000; Murphy & Barkley, 1996). However, the effect sizes for these differences in the current study were small, thereby suggesting that the general decline of the ADHD and HKD symptoms with age is of little practical importance.

The findings here showed higher scores for males for IA and total scores. All other scores did not vary by gender. Again the effect sizes for these differences were small. Thus, there was minimal gender effect. Previous studies have also found minimal gender effects for the ADHD symptom groups in both community (Kooij et al., 2008; Murphy & Barkley, 1996) and diagnosed ADHD (Kooij et al., 2008; Rasmussen & Levander, 2009) groups. On the basis of the results of age and gender effects, it could appear that when using the descriptive scores provided here, there is no need to use gender- or age-specific scores. The overall scores for the different symptom groups can be used for all adults.

The second aim of the current study was to examine the percentages of individuals with symptom counts at or above the threshold numbers for the different ADHD types and for HKD. The finding here indicated that the rate for all ADHD types was 6.3%. For HKD, it was 2.0%. The rates for the inattentive, hyperactive-impulsive, and combined types were 1.6%, 2.7%, and 2.0%, respectively. As far as it can be ascertained, this is the first study to provide incidence rates based on symptom threshold numbers for the different ADHD types and HKD for Australian adults. In this respect, the overall rate for ADHD found in this study is comparatively higher than that reported in most previous studies that have used a similar approach as in the current study. For instance, Murphy and Barkley (1996) reported an overall rate of 4.7% for a convenience sample of adults. For a sample of university students, Heiligenstein et al. (1998) reported a rate of 4%. It needs to be stressed however that the rates reported here are not prevalence rates as prevalence estimates are based on clinical interviews, involving larger national representative samples, and also the presence of impairment in occupational, social, or academic settings.

The finding here of higher incidence rate for the hyperactive-impulsive type, compared with the inattentive and combined types, is consistent with existing data (Kooij et al., 2008; Murphy & Barkley, 1996). This is a surprising finding as this ADHD type is extremely rare in children (Gomez, Harvey, Quick, Scharer, & Harris, 1999). In view of this, Kooij et al. (2008) have raised the possibility that the HYP and IMP symptoms that were designed for children are inappropriate for adult use or there may be qualitative differences between adult self-rating of these ADHD symptoms and parents and teachers ratings of children on these ADHD symptoms. These possibilities may be worthy of further exploration in future studies.

An interesting finding in this study was that the symptom threshold rate for the ADHD combined type was the same as HKD. More importantly, of the 17 individuals who fell at or above the threshold for HKD, 15 also fell at or above the threshold for ADHD combined type. At the general level, this finding suggests that in most instances, individuals diagnosed as HKD are also likely to qualify for a diagnosis of ADHD combined type. If so, this could have implications for what may be appropriate symptom threshold number for the hyperactive and impulsive symptoms. For both DSM-IV ADHD and ICD-10 HKD, there is a need for at least six IA symptoms to be present. Across the HYP and IMP symptoms, DSM-IV requires at least six symptoms. In contrast, ICD-10 diagnosis of HKD requires at least three HYP symptoms and one IMP symptom to be present. Thus, only four symptoms across the HYP and IMP groups (and not six symptoms) are required for HKD, with the proviso that one has to be an IMP symptom. Given that virtually all the individuals in the HKD group were also in the ADHD combined group, it could be argued that a threshold of four symptoms may be all that is required to qualify for the ADHD combined group. Indeed, a lower cutoff threshold has also been suggested by others (Biederman et al., 2000; Kooij et al., 2008; Murphy & Barkley, 1996), and follow-up studies by Biederman et al. (2000) supported a lower symptom threshold for adults. Consistent with the proposal here, Kooij et al. (2008) have shown that a cutoff of four or more symptoms of IA or of HYP + IMP was associated with a significant increase in overall psychosocial impairment.

The third aim of the current study was to examine the factor structure of adult self-ratings of the ADHD symptoms. Overall, the findings here were not supportive of the two-factor ADHD model or the Barkley et al. (2008) three-factor model. There was adequate support for the three-factor DSM-IV model and also the three-factor ICD-10 model. The findings here supporting the three-factor DSM-IV model and not the two-factor model are comparable with the results of existing studies in this area (Kooij et al., 2008; Span et al., 2002). However, the findings in the current study extend that of previous studies. In terms of the 2 three-factor models examined, the results indicated better fit for the three-factor ICD-10 model than the three-factor DSM-IV model. This is a new finding.

The findings for the factorial structure of the ADHD symptoms have important diagnostic and clinical implications. The lack of support for the two-factor ADHD model implies that the current bidimensional model proposed in DSM-IV for ADHD may not be appropriate for ADHD in adults. Although the two-factor ADHD model has been supported in ratings of children (Gomez, Burns, Walsh, & Alves de Moura, 2003; Gomez et al., 1999), there has also been support for the three-factor model, with the three-factor model showing a better fit than the two-factor model (Gomez et al., 1999; Scholte, van Berckelaer-Onnes, & van der Ploeg, 2001). Thus, it would be argued that the three-factor model is a more parsimonious and robust ADHD model. This argument also means that it may be more prudent for the next edition of DSM to consider a three-factor ADHD model, as is currently the case in ICD-10 for the HKD.

Kooij et al. (2008) have proposed two possible explanations for support for the three-factor and not the two-factor ADHD model in adults. One is that there may be changes to this disorder across the life span, with separation of the IA symptoms from the HYP and IMP symptoms during childhood and adolescence, and then further separation of the HYP and IMP symptoms in later years. In this respect, Span et al. (2002) have suggested that HYP and IMP may function as a combined construct in childhood because self-monitoring skills have not developed fully. The lack of support for the two-factor model in adults raises questions about the existence of the hyperactive-impulsive ADHD type in adults (Kooij et al., 2008). Indeed, this study found the HYP factor to be more closely correlated with the IA factor than the IMP factor, thereby implying that the better two-factor model for the ADHD symptoms in adults is IA + IMP and HYP factors rather than the DSM-IV proposed IA and HYP + IMP factors. Indeed, the existence of the hyperactive-impulsive ADHD type in children has also been questioned (Gomez et al., 1999).

The fourth aim of this study was to examine measurement invariance for ratings of the symptoms in the CSS across gender. Given that the three-factor HKD model showed the best fit, gender invariance was tested for this model using the multiple-group CFA procedure approach for ordered-categorical data. The results showed support for configural invariance, metric invariance, threshold invariance, and error variances invariance, thereby indicating support for the full measurement model. Kooij et al. (2008) reported gender invariance for their three-factor ADHD model with cross-loadings. However, their three-factor model was not the typical simple structure ADHD model as it involved 5 more than the 18 ADHD symptoms and had eight items that cross-loaded on all three factors and another six items that cross-loaded on two factors. Although Smith and Johnson (1998) reported gender invariance among adults for the two-factor ADHD model, as already noted, there is computational error in their analyses, thereby invalidating their findings. Thus, as far as it can be ascertained, this is the first study to provide valid gender invariance information for the ADHD/HKD symptoms in terms of their simple structure.

Strictly speaking, the support for gender invariance found in the current study needs to be seen as being specific to CSS. However, as the CSS is similar to the other ADHD rating scales, it is conceivable that measurement invariance may also be present for the other adult ADHD rating scales and for the ADHD symptoms in general. However, these would need to be empirically validated in future studies. The findings here in support of measurement invariance across males and females for ratings of the CSS mean that this scale is measuring conceptually similar constructs and in the same manner across these groups. Expressed differently, the CSS does not function differently across males and females, and the observed scores obtained in the CSS can be used with confidence for comparing males and females. This also means that the gender differences findings reported above were not confounded by differences in measurement properties in CSS arising from differential responses made by males and females.

In conclusion, gender and age differences for the mean scores for different ADHD and HKD symptom groups were minimal. The overall percentages of individuals who had symptom score at or above the threshold number for HKD was 2%. The rates for the inattentive, hyperactive-impulsive, and combined types were 1.6%, 2.7%, and 2.0%, respectively. Findings in the study supported the three-factor model, especially as reflected in ICD-10. For this model, there was full measurement invariance across the ratings provided by males and females. The findings in this study, however, need to be viewed with several important limitations in mind. First, as ethics approval for this study did not permit collection of information about individuals prior to inviting them to participate, there is no information about those who did not respond to the invitation to participate in this study, and therefore how this affected the results. As the sample examined may not be representative of the general population, it would be prudent to suggest that the findings provided in this study have to be viewed with some caution. There is certainly a need to replicate these findings with more representative and larger samples. Second, as this study examined a community sample, it is uncertain if the findings for factor structure and gender invariance are applicable to clinically diagnosed ADHD adults. Third, in terms of the support for the three-factor models over the two-factor models, it is possible that the separation of the HYP and IMP symptoms into different factors found in this study may be related to the way that these symptoms are worded in the CSS. A close examination of these items in the child and adult versions suggest that they refer to the same overt behaviors. Several researchers have argued that in adults, the HYP and IMP symptoms are less overt and are expressed more in terms of mental restlessness (Downey, Stelson, Pomerleau, & Giordani, 1997; Weyandt et al., 2003). Indeed, DSM-IV (APA, 1994) specifies that in adults, these symptoms “maybe limited to subjective feelings of restlessness” (p. 84). Given this, it cannot be ruled out that if the HYP and IMP symptoms were worded to reflect mental restlessness, a combined factor for these symptoms would have emerged. Future studies may wish to examine this. Notwithstanding these limitations to the generalization of the findings of this study, it could however be argued that the results of the current do provide impetus for more studies in this area.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biography

Rapson Gomez is Professor of Clinical Psychology at the University of Ballarat (Australia). He major research focus is on childhood psychological disorders.

References

American Psychiatric Association. (1968). Diagnostic and statistic manual of mental disorders (2nd ed.). Washington, DC: Author

American Psychiatric Association. (1980). Diagnostic and statistic manual of mental disorders (3rd ed.). Washington, DC: Author.

American Psychiatric Association. (1987). Diagnostic and statistic manual of mental disorders (3rd ed., rev.). Washington, DC: Author.

American Psychiatric Association. (1994). Diagnostic and statistic manual of mental disorders (4th ed.). Washington, DC: Author.

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

Australian Bureau of Statistics. (2007). Ancestry by country of birth of parents—Time series statistics (2001, 2006 Census Years): Australia. Canberra, Australia: Author.

Barkley

R. A.

Fischer

Smallish

Fletcher

(2002). The persistence of attention-deficit/hyperactivity disorder into young adulthood as a function of reporting source and definition of disorder. Journal of Abnormal Psychology, 111, 279-289.

Barkley

R. A.

Murphy

K. R.

(1998). Attention deficit hyperactivity disorder: A clinical workbook (2nd ed.). New York, NY: Guilford.

Barkley

R. A.

Murphy

K. R.

Fischer

(2008). ADHD in adults: What the science says. New York, NY: Guilford.

10.

Biederman

Faraone

S. V.

Monuteaux

M. C.

Bober

Cadogen

(2004). Gender effects on attention-deficit/hyperactivity disorder in adults, revisited. Biological Psychiatry, 55, 692-700.

11.

Biederman

Mick

Faraone

S. V.

(2000). Age-dependent decline of symptoms of attention deficit hyperactivity disorder: Impact of remission definition and symptom type. American Journal of Psychiatry, 157, 816-818.

12.

Brown

T. A.

(2006). Confirmatory factor analysis for applied research. New York, NY: Guilford.

13.

Burns

G. L.

Walsh

J. A.

Gomez

Hafetz

(2006). Measurement and structural invariance of parent ratings of ADHD and ODD symptoms across gender for American and Malaysian children. Psychological Assessment, 18, 452-457.

14.

Clarke

Heussler

Kohn

M. R.

(2005). Attention deficit disorder: Not just for children. Internal Medicine Journal, 35, 721-725.

15.

Cohen

(1992). A power primer. Psychological Bulletin, 112, 155-159.

16.

Downey

K. K.

Stelson

F. W.

Pomerleau

O. F.

Giordani

(1997). Adult attention deficit hyperactivity disorder: Psychological test profiles in a clinical population. Journal of Nervous and Mental Disease, 185, 32-38.

17.

DuPaul

G. J.

Power

T. J.

Anastopoulos

A. D.

Reid

(1998). ADHD Rating Scale–IV: Checklists, norms, and clinical interpretation. New York, NY: Guilford.

18.

DuPaul

G. J.

Schaughency

E. A.

Weyandt

L. L.

Tripp

Kiesner

Ota

Stanish

(2001). Self-report of AD/HD symptoms in university students: Cross-gender and cross-national prevalence. Journal of Learning Disabilities, 34, 370-379.

19.

Glutting

J. J.

Youngstrom

E. A.

Watkins

M. W.

(2005). ADHD and college students: Exploratory and confirmatory factor structures with student and parent data. Psychological Assessment, 17, 44-55.

20.

Gomez

(2007). Testing gender differential item functioning for ordinal and binary scored parent-rated ADHD symptoms. Personality and Individual Differences, 42, 733-742.

21.

Gomez

Burns

G. L.

Walsh

J. A.

Alves de Moura

M. A.

(2003). A multitrait-multisource confirmatory factor analytic approach to the construct validity of ADHD rating scale. Psychological Assessment, 15, 3-16.

22.

Gomez

Harvey

Quick

Scharer

Harris

(1999). DSM-IV AD/HD: Confirmatory factor models, prevalence, and gender and age differences based on parent and teacher ratings of Australian primary school children. Journal of Child Psychology and Psychiatry, 40, 265-274.

23.

Green

S. B.

Babyak

M. A.

(1997). Control of Type I errors with multiple tests of constraints in structural equation modeling. Multivariate Behavioral Research, 32, 39-51.

24.

Heiligenstein

Conyers

L. M.

Berns

A. R.

Smith

M. A.

(1998). Preliminary normative data on DSM-IV attention deficit hyperactivity disorder in college students. Journal of American College Health, 46, 185-188.

25.

Bentler

P. M.

(1998). Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological Methods, 3, 424-453.

26.

Kessler

R. C.

Adler

Ames

Demler

Faraone

Hiripi

Walters

E. E.

(2005). The World Health Organization Adult ADHD Self-Report Scale (ASRS): A short screening scale for use in the general population. Psychological Medicine, 35, 245-256.

27.

Kooij

J. J. S.

Boonstra

A. M.

Swinkels

S. H. N.

Bekker

E. M.

de Noord

Buitelaar

J. K.

(2008). Reliability, validity, and utility of instruments for self-report and informant report concerning symptoms of ADHD in adult patients. Journal of Attention Disorders, 11, 445-458.

28.

Millsap

R. E.

Yun-Tein

(2004). Assessing factorial invariance in ordered-categorical measures. Multivariate Behavioral Research, 39, 479-515.

29.

Murphy

Barkley

R. A.

(1996). Prevalence of DSM-IV symptoms of ADHD in adult licensed drivers: Implications for clinical diagnosis. Journal of Attention Disorders, 1, 147-161.

30.

Muthen

L. K.

Muthen

B. O.

(2006). Mplus user’s guide (4th ed.). Los Angeles, CA: Author.

31.

Rasmussen

Levander

(2009). Untreated ADHD in adults: Are there sex differences in symptoms, comorbidity, and impairment? Journal of Attention Disorders, 12, 353-360.

32.

Reid

Riccio

C. A.

Kessler

R. H.

DuPaul

G. J.

Power

T. J.

Anastopoulos

A. D.

. . .Noll

M. B.

(2000). Gender and ethnic differences in ADHD as assessed by behaviour ratings. Journal of Emotional and Behavioral Disorders, 8, 38-48.

33.

Reise

S. P.

Widaman

K. F.

Pugh

R. H.

(1993). Confirmatory factor analysis and item response theory: Two approaches for exploring measurement invariance. Psychological Bulletin, 114, 552-566.

34.

Satorra

Bentler

P. M.

(1999). A scaled difference chi-square test statistic for moment structure analysis (UCLA Statistics Series No. 260). Los Angeles: Department of Psychology, University of California.

35.

Scholte

E. M.

van Berckelaer-Onnes

I. A.

van der Ploeg

J. D.

(2001). DSM-IV related ADHD symptom ratings by professional caretakers in residential treatment centres. Journal of Child Psychology and Psychiatry, 42, 341-346.

36.

Smith

E. V.

Johnson

B. D.

(1998). Factor structure of the DSM-IV criteria for college students using the Adult Behavior Checklist. Measurement and Evaluation in Counseling and Development, 31, 164-183.

37.

Span

S. A.

Earleywine

Strybel

T. Z.

(2002). Confirming the factor structure of attention deficit hyperactivity disorder symptoms in adults, nonclinical sample. Journal of Psychopathology and Behavioural Assessment, 24, 129-136.

38.

Weyandt

L. L.

Iwaszuk

Fulton

Ollerton

Beatty

Fouts

Greenlaw

(2003). The Internal Restlessness scale: Performance of college students with and without ADHD. Journal of Learning Disabilities, 36, 382-389.

39.

World Health Organization. (1993). The ICD-10 classification of mental and behavioural disorders: Clinical descriptions and diagnostic guidelines 1992. Diagnostic criteria for research 1993. Geneva, Switzerland: Author.