The Limited Prosocial Emotions Specifier for Conduct Disorder Among Detained Girls

Abstract

This study examines the prevalence and clinical usefulness of the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5) specifier “with Limited Prosocial Emotions” (LPE) in detained girls. Detained girls (N = 85; M_age = 16.24) and their parents were interviewed with a structured diagnostic interview to identify girls with conduct disorder (CD), and both informants completed the Antisocial Process Screening Device to assess the LPE specifier. Psychiatric disorders other than CD, aggression, and offending were assessed through standardized self-report tools. Different approaches were used to deal with diagnostic information from multiple informants. The prevalence of CD + LPE girls was lower when using self-report (12.9%) compared with parent-report (38.8%), suggesting that parents indeed are important to identify CD + LPE girls. However, including parental information did not result in a better differentiation between CD + LPE and CD-only girls. Specifically, the LPE specifier only enabled to identify a group of seriously antisocial girls with higher levels of proactive aggression, though solely when using self-report.

Keywords

Antisocial Process Screening Device callous-unemotional traits conduct disorder forensic psychiatry

Callous-unemotional (CU) traits have become increasingly emphasized in theoretical models and empirical studies on the etiology of conduct problems (Frick, Ray, Thornton, & Kahn, 2014a). In the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association [APA], 2013), CU traits have been added as a specifier—“with Limited Prosocial Emotions”—for the diagnosis of conduct disorder (CD). To meet criteria for the “with Limited Prosocial Emotions” (LPE) specifier, at least two of the following characteristics must be present over at least 12 months and in multiple relationships and settings: (a) lack of remorse or guilt, (b) callous–lack of empathy, (c) shallow or deficient affect, and (d) unconcerned about performance (APA, 2013). Overall, it is expected that this LPE specifier designates a group of severe antisocial and aggressive youths, provides greater information about current and future impairment, and supports treatment planning for youths with CD (Frick & Nigg, 2012; Kimonis et al., 2014).

Few studies, however, tested the clinical usefulness of this LPE specifier as categorically defined by DSM-5. In community and clinic-referred samples, 5- to 17-year-olds with CD who met criteria for the LPE specifier (CD + LPE) showed higher rates of aggression, cruelty, and symptoms of attention-deficit/hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD) than youths who only met criteria for CD (CD-only; Kahn, Frick, Youngstrom, Findling, & Youngstrom, 2012). In community girls (ages 6-8), CD + LPE girls displayed more externalizing disorder symptoms, relational aggression, bullying, global impairment, and less anxiety than CD-only girls (Pardini, Stepp, Hipwell, Stouthamer-Loeber, & Loeber, 2012), whereas clinic-referred CD + LPE children (ages 6-11) had greater impairment at pretreatment than CD-only children (Kolko & Pardini, 2010). Available evidence suggests that the LPE specifier in detained adolescents is of restricted usefulness (Colins & Andershed, 2014; Colins & Vermeiren, 2013), a finding that clearly runs counter to the aforementioned evidence stemming from community and clinic-referred samples. Specifically, detained CD + LPE boys and girls were not significantly different from CD-only boys and girls regarding ADHD, ODD, substance use disorder (SUD), major depression, and anxiety disorders. However, CD + LPE girls were more aggressive, rule-breaking, and delinquent, than CD-only girls, a finding that was not replicated among detained boys (Colins & Andershed, 2014; Colins & Vermeiren, 2013). Clearly, empirical evidence in support of or against the LPE specifier is thin (Lahey, 2014).

Parents of detained adolescents are difficult to locate, and/or often unwilling or unable to provide information (Colins, Vermeiren, Schuyten, Broekaert, & Soyez, 2008). As a consequence, all studies on the DSM-5 LPE specifier solely relied on youth self-report (Colins & Andershed, 2014; Colins & Vermeiren, 2013). This is unfortunate, not only because DSM-5 explicitly states that self-report of LPE must be extended with information from others but also because the sole reliance on uncorroborated self-ratings hampers firm conclusions about the usefulness of the LPE specifier in detained adolescents (Colins & Andershed, 2014). Notwithstanding that parents of detained adolescents may provide relevant information (Colins et al., 2012), approaching these informants still is a time-consuming investment for which detention facilities and researchers often lack budget and personnel. Therefore, it is important to test whether gathering parental information about detained youths is worth the effort, for example, because their information results in stronger and more differences between CD + LPE and CD-only girls. Yet, including multiple informants inevitably confronts clinicians and researchers with the question how to deal with these multiple sources (Colins et al., 2008; Fink, Tant, Tremba, & Kiehl, 2012). There are various ways to deal with multiple informants. The usefulness of parent and adolescent information can be assessed independently from each other (i.e., optimal informant approach; Loeber, Green, Lahey, & Stouthamer-Loeber, 1989), but parent- and youth-reports can also be combined in several ways. The first and most commonly used strategy is to consider a disorder present if the girl met criteria for this disorder according to at least one informant (i.e., “OR” rule; Ko, Wasserman, McReynolds, & Katz, 2004). In addition, one can consider a disorder present if reported by both informants (i.e., “AND” rule), by the girl only (i.e., “Unique Girl” rule), or by the parent only (i.e., “Unique Parent” rule; Colins et al., 2012).

The overall aim of this study was to examine the prevalence of detained girls who met the DSM-5 LPE specifier, and to test whether CD + LPE girls differed from CD-only girls on clinically important features. First, prevalence rates of CD and LPE were explored using information from multiple informants alone or in conjunction. We hypothesized that detained girls would more frequently identify CD than their parents, while the reverse would be true for LPE (Colins et al., 2008; Fink et al., 2012; Kahn et al., 2012; Ko et al., 2004). We also apply the “OR” and “AND” rules, and we hypothesized that the former will yield higher prevalence rates of CD and LPE than the latter rule. Second, CD + LPE and CD-only girls were compared regarding psychiatric morbidity, aggression, and offending, while using various informant approaches. Based on prior work among detained girls (Colins & Andershed, 2015), we expected that CD + LPE girls compared with CD-only girls would display higher levels of aggression and offending but would be similar regarding psychiatric morbidity. Finally, it has recently been argued that being too DSM centric may limit our understanding of the role of the LPE specifier in designating a distinct subgroup of juveniles with serious conduct problems that may not meet criteria for CD (Frick, Ray, Thornton, & Kahn, 2014b). Therefore, we also compared LPE and non-LPE girls on the above-mentioned variables of interest when using the specifier in a non-DSM CD centric manner (i.e., without requiring that girls also meet criteria for CD).

Method

Participants

Participants included girls who were placed in an all-girl Youth Detention Center (YDC) in Flanders, Belgium, and one of their parents. Placement in a YDC is only possible following referral by a juvenile judge because of a criminal offense (e.g., shoplifting, burglary, fighting, or threatening) or an urgent problematic educational situation (e.g., persistent truancy, running away, aggression, or prostitution), and is considered the harshest measure a juvenile judge can impose. Girls were eligible to participate if the following criteria were met: (a) adjudicated to be placed in a YDC for at least 1 month, (b) sufficient knowledge of Dutch, and (c) sufficient cognitive abilities. The latter criteria were based upon both staff’s and interviewer’s assessment of the girl’s ability to participate in Dutch conversations and to read and comprehend the informed assent form. Between February 2012 and June 2014,¹ 169 girls were eligible to participate. Two girls could not be approached due to acute psychiatric crisis, 14 girls refused participation, and six parents refused their daughter’s participation, resulting in a participation rate of 87% (n = 147).

We aimed to include one parent for each girl. A parent could participate if the following criteria were met: (a) sufficient contact with his or her daughter during the past year, varying from daily until at least monthly; and (b) sufficient knowledge of Dutch. The latter criterion was based on the girl’s, staff’s, and interviewer’s assessment of the parent’s ability to participate in Dutch conversations and to read and comprehend the informed consent form. For the total sample of 147 girls, 115 girls had at least one parent meeting inclusion criteria. Fourteen girls did not provide informed assent to contact their parents, and for 16 girls, the parents did not provide informed consent themselves, resulting in a final sample of 85 pairs of girls and one of their parents (participation rate = 74%).

Procedure

This study was approved by the Institutional Review Board of the Faculty of Psychology and Educational Sciences at Ghent University (2011/59) and by the board of the YDC. Participants were approached and assessed following a standardized protocol. The girls were approached individually, receiving oral and written information about the aims, content, and duration of the study. They were assured that their information would be treated confidentially and that refusal to participate would not affect their judicial status or stay in the YDC. The girls had to give written informed assent before starting the assessment. At the moment the girls entered the YDC, their parents/caretakers received an informed consent letter including information about the aims and practical aspects of the study and could refuse the girls’ participation. The assessment took place in a private area in the YDC, within the first 3 weeks of placement. The Diagnostic Interview Schedule for Children, Version IV (DISC-IV) and a set of self-report questionnaires were administered to the girls by the first author or final year university students, who were all trained in using the DISC-IV. None of the assessors were YDC staff. Afterward, the girls received oral and written information about the aim of contacting their parents/caretakers. After receiving the girls’ written informed assent to contact their parent/caretaker, an informed consent letter concerning their own participation was sent to these adults. The first author then tried to contact one parent/caretaker for each girl at least 10 times over a 1-month period at varying times during the day, to check their willingness to cooperate and to make a telephone appointment at a time that suited the parent/caretaker the best. In most cases, the telephone assessment was conducted by the first author within 3 weeks after the girl had been assessed, including only some modules from the DISC-IV and the LPE measure of interest (see “Measures” section). Neither girls nor their parents received financial compensation.

Measures

Psychiatric Disorders

The DISC-IV (Shaffer, Fisher, Lucas, Dulcan, & Schwab-Stone, 2000) is a highly structured diagnostic interview, designed for interviewing children and adolescents 9 to 17 years of age. The Dutch DISC-IV (Ferdinand & Van der Ende, 2002) was used to assess the past-year prevalence of CD, ADHD, ODD, any SUD (i.e., alcohol, marijuana, and/or other drug use disorder), any mood disorder (i.e., major depressive or dysthymic disorder), and any anxiety disorder (i.e., posttraumatic stress disorder and/or separation anxiety disorder). For practical reasons (e.g., only one interviewer to approach and interview parents), only criteria for CD were assessed by both parents and the girls themselves. All the other aforementioned disorders were assessed by means of self-report only. The DISC-IV is a reliable and valid questionnaire in clinical and community samples (Crowley, Mikulich, Ehlers, Whitmore, & MacDonald, 2001; Shaffer et al., 2000). It is important to note that the DISC-IV assesses DSM-IV (Diagnostic and Statistical Manual of Mental Disorders [4th ed.]; APA, 1994) psychiatric disorders. However, as the main diagnostic criteria of CD remained unchanged in the DSM-5, the DISC-IV is equally valid.

Limited Prosocial Emotions (LPE)

The LPE specifier was assessed using the Antisocial Process Screening Device (APSD; Frick & Hare, 2001). The Dutch self-report version of the APSD (APSD-SR; Bijttebier & Decoene, 2009) consists of 20 items that tap psychopathic-like traits and are answered on a 3-point rating scale: not at all true (0), sometimes true (1), or definitely true (2). In line with all previous studies that used the APSD (Colins & Andershed, 2014; Kahn et al., 2012; McMahon, Witkiewitz, & Kotler, 2010; Pardini et al., 2012), the girls were identified as meeting the LPE specifier threshold if they had a (reversed) score of 2 (definitely true) on at least two of the four items from the APSD CU factor that corresponded to the four DSM-5 LPE specifier criteria. Despite concerns regarding the factor structure and reliability of the APSD-SR in juvenile justice–involved youths (Colins, Bijttebier, Broekaert, & Andershed, 2014; Poythress et al., 2006), the APSD is the most widely used measure to study the clinical usefulness of the LPE specifier (Colins & Andershed, 2014). A strength of the APSD is that this tool also has a parent version, which enables to assess the specifier by multiple informants.

Aggression

Aggression was measured by means of the Reactive–Proactive Aggression Questionnaire (RPQ; Raine et al., 2006). The RPQ (Dutch version; Cima, Raine, Meesters, & Popma, 2013) is a self-report measure that includes 11 items that focus on reactive aggression (e.g., gotten angry when frustrated) and 12 items that focus on proactive aggression (e.g., had fights with others to show who was on top). All items must be answered on a 3-point Likert-type scale: never (0), sometimes (1), or often (2). The internal consistency as indexed by Cronbach’s alpha was .81 for both RPQ scales.

Self-Reported Delinquency

Delinquency was measured using the youth-report questionnaire developed by the Research and Documentation Center of the Ministry of Safety and Justice in the Netherlands (van der Laan & Blom, 2005). All items begin with the standardized question “Have you ever . . .” The lifetime violent offending score reflects the total number of reported violent offenses (seven items; for example, fighting and threats). Lifetime nonviolent offending refers to 15 items capturing property offenses (e.g., shoplifting and vandalism), two items capturing insults, and three items capturing dealing drugs. Cronbach’s alphas for nonviolent and violent offending were .87 and .75, respectively.

Sociodemographics

Standardized information regarding age, origin, socioeconomic status (SES), family situation, school attendance, and detention history was gathered by means of a self-report questionnaire (see also Colins, Vermeiren, Schuyten, & Broekaert, 2009). Girls were placed in the low (vs. moderate to high) SES category, when both parents were unemployed or worked as (un)skilled laborers.

Statistical Analyses

First, prevalence rates of CD and the LPE specifier were presented when using different rules to combine youth- and parent-reports. Parent–youth (dis)agreement on the diagnosis of CD and the LPE specifier was explored in multiple ways. Chi-square tests were used to investigate differences between parent- and youth-reported prevalence rates. Cohen’s kappa (κ) statistics were used to examine the overall level of parent–youth agreement: κ < .40 is considered poor, .39 < κ< .60 moderate, and κ > .59 good (Landis & Koch, 1977). To gain a more detailed insight in the nature of agreement between both raters, we also presented indices of positive agreement (PA) and negative agreement (NA). These figures indicate the agreement between youth and parents on the presence (i.e., PA) or absence (i.e., NA) of CD and LPE, respectively. The McNemar test was used to test whether parents or youths significantly reported more unique diagnostic information. Second, the clinical utility of the LPE specifier for CD was scrutinized, again adopting different rules to combine youth- and parent-reports. Even though our sample of 85 youth–parent dyads is very large compared with prior work with detained youths (e.g., 35 out of 160 parents; Fink et al., 2012), we needed to be selective in the number of group comparisons. Given the goals of this article, we only focused on CD + LPE and CD-only girls comparisons. Differences between both groups were examined using Fisher’s exact statistics for categorical variables, given the rather small sample size. For continuous variables and because assumptions of normality and homoscedasticity were often violated, nonparametric tests (i.e., Mann–Whitney U tests) were used (details available upon request from the first author). DAG_Stat was used to compute Cohen’s kappa and the indices of PA and NA (Mackinnon, 2000). SPSS 21.0 was used for all the other analyses. The performed tests were two-tailed, with p < .05 as the standard for statistical significance.

Results

Descriptives

Overall, girls whose parent participated were not significantly different from girls whose parent did not participate regarding sociodemographic variables, CD, LPE, psychiatric morbidity, aggression, and offending (details available upon request from the first author), with one exception: Girls whose parent participated, M = 12.07, SD = 4.19, reported significantly higher rates of reactive aggression than girls whose parent did not participate, M = 9.27, SD = 5.19, t = 2.95(112), p = .004. The age of the participants (n = 85) ranged from 13.52 to 17.92 years, M = 16.24, SD = 1.16. The majority of the sample was from Belgian origin (n = 65, 76.5%), five (5.9%) girls were from Moroccan origin, five (5.9%) girls from Turkish origin, and 10 (11.8%) girls from other origins (e.g., Spanish). The SES was moderate to high for 35 (41.2%) girls, and 23 (27.1%) girls did not live with their biological parents prior to detention. In addition, 47 (55.3%) girls had been attending school during the past month before placement, and 15 (17.6%) had been detained in the past. Prevalence rates for any SUD (n = 58, 68.2%) were the highest, followed by any mood disorder (n = 34, 40.0%), ODD (n = 30, 35.3%), any anxiety disorder (28, 32.9%), and ADHD (n = 22, 25.9%). The mean score for reactive and proactive aggression was 12.07, SD = 4.19, and 4.70, SD = 4.07, respectively. For nonviolent and violent offending, the mean score was 4.40, SD = 4.26, and 1.25, SD = 1.49, respectively. Most participating parents/caretakers were biological parents (n = 72, 84.7%), female (n = 59, 69.4%) and of Belgian origin (n = 73, 85.9%).

Prevalence and Parent–Youth Agreement: Total Sample

CD

There was no significant difference, χ² = .21(1), p = .645, in the prevalence of youth-reported CD (n = 40, 47.1%) and parent-reported CD (n = 43, 50.6%). Applying the “OR”- and “AND” rules resulted in a prevalence of 74.1% (n = 63) and 9.4% (n = 8), respectively. The low kappa value, κ = −.30, corresponded with the low prevalence of CD while applying “AND” rule, and is indicative of overall poor parent–youth agreement. The results revealed a low level of PA (.23) compared with a higher level of NA (.44). The McNemar test, finally, indicated that girls (n = 21, 24.7%) did not significantly report unique CD more frequently than their parents (n = 34, 40.0%), McNemar test = 2.62(1), p = .106 (Table 1).

Table 1:

Number and Percentage of Girls That Met Criteria for CD, for LPE symptoms, and the LPE Specifier (n = 85)

	Prevalence Rates Optimal Informant Approach						Prevalence Rates Using Informant Reports in Conjunction
	Girl	Parent					OR	AND	Unique Girl	Unique Parent	McNemar (1); p
	n (%)	n (%)	χ²(1); p	κ [95% CI]	PA [95% CI]	NA [95% CI]	n (%)	n (%)	n (%)	n (%)	McNemar (1); p
Criterion (a): Lack of remorse	12 (14.1)	30 (35.3)	10.25; .001**	−.01 [−.19, .16]	.19 [.03, .35]	.73 [.65, .82]	38 (44.7)	4 (4.7)	8 (9.4)	26 (30.6)	8.50; .004**
Criterion (b): Lack of empathy	2 (2.4)	28 (32.9)	27.36; <.001**	−.05 [−.11, .02]	.00 [.00, .00]	.79 [.71, .86]	30 (35.3)	0 (.0)	2 (2.4)	28 (32.9)	20.83; <.001**
Criterion (c): Shallow affect	38 (44.7)	55 (64.7)	6.86; .009**	−.12 [−.31, .08]	.47 [.35, .60]	.36 [.23, .50]	71 (83.5)	22 (25.9)	16 (18.8)	33 (38.8)	5.22; .022*
Criterion (d): Unconcerned performance	18 (21.2)	43 (50.6)	15.95; <.001**	−.05 [−.22, .12]	.26 [.12, .41]	.59 [.48, .70]	53 (62.4)	8 (9.4)	10 (11.8)	35 (41.2)	12.80; <.001**
Met one of the LPE specifier criteria	35 (41.2)	22 (25.9)	4.46; .035*	.46 [.23, .68]	.58 [.39, .77]	.88 [.82, .94]	27 (31.8)	11 (12.9)	7 (8.2)	9 (10.6)	.06; .804
Met two of the LPE specifier criteria	11 (12.9)	21 (24.7)	3.85; .050	.34 [.10, .58]	.43 [.20, .66]	.89 [.83, .94]	22 (25.9)	6 (7.1)	2 (2.4)	14 (16.5)	7.56; .004**
Met three of the LPE specifier criteria	3 (3.5)	16 (18.8)	10.01; .002**	−.06 [−.13, −.00]	.00 [.00, .00]	.87 [.82, .93]	19 (22.4)	0 (.0)	3 (3.5)	16 (8.8)	7.58; .004**
Met four of the LPE specifier criteria	1 (1.2)	11 (12.9)	8.97; .003**	−.02 [−.06, .02]	.00 [.00, .00]	.92 [.88, .97]	12 (14.1)	0 (.0)	1 (1.2)	11 (12.9)	6.75; .006**
Met the LPE specifier threshold	15 (17.6)	48 (56.5)	27.46; <.001**	−.03 [−.17, .11]	.20 [.07, .34]	.58 [.47, .69]	53 (62.4)	6 (7.1)	6 (7.1)	41 (48.2)	24.60; <.001**
CD	40 (47.1)	43 (50.6)	.21; .645	−.30 [−.49, −.11]	.23 [.10, .35]	.44 [.32, .57]	63 (74.1)	8 (9.4)	21 (24.7)	34 (40.0)	2.62; .106
CD-only	29 (34.1)	10 (11.8)	12.01; .001**	.18 [−.06, .42]	.30 [.07, .52]	.87 [.81, .93]	23 (27.1)	4 (4.7)	14 (16.5)	5 (5.9)	3.37; .064
CD + LPE	11 (12.9)	33 (38.8)	14.84; <.001**	−02 [−.18, .15]	.18 [.03, .34]	.71 [.62, .80]	40 (47.1)	4 (4.7)	7 (8.2)	29 (34.1)	12.25; <.001**

Note. CD = conduct disorder; LPE = with limited prosocial emotions; κ = Cohen’s kappa; CI = confidence interval; PA = positive agreement; NA = negative agreement.

p < .05. **p < .01.

LPE Specifier

The prevalence of the LPE specifier was significantly lower when based upon youth-report (n = 15, 17.6%) than when based upon parent-report (n = 48, 56.5%), χ² = 27.46(1), p < .001. When applying the “OR”- and “AND” rules, the prevalence was 62.4% (n = 53) and 7.1% (n = 6), respectively. The low “AND” rule prevalence was corroborated with a poor kappa value, κ = −.03, indicating that, overall, girls and parents rarely agreed on the presence of LPE. Parents and girls agreed more that the girls were without (NA = .58) than with (PA = .20) the LPE specifier. Finally, significantly more parents (n = 41, 48.2%) than girls (n = 6, 7.1%) uniquely identified the LPE specifier, McNemar test = 24.60(1), p < .001. Table 1 also shows the prevalence of girls that met a specific LPE specifier criterion and a particular number of LPE specifier criteria.

Prevalence and Parent–Youth Agreement: Girls with CD

CD + LPE

The prevalence of youth-reported CD + LPE (n = 11, 12.9%) was significantly lower than parent-reported CD + LPE (n = 33, 38.8%), χ² = 14.84(1), p < .001. The “OR”- and “AND” rules demonstrated a prevalence of 47.1% (n = 40) and 4.7% (n = 4), respectively. The low kappa value, κ = −.02, accorded with the low “AND” rule prevalence, indicating overall poor parent–youth agreement. The level of PA was low (.18), whereas the level of NA was much higher (.71). Last, significantly more parents (n = 29, 34.1%) than girls (n = 7, 8.2%) uniquely reported CD + LPE, McNemar test = 12.25(1), p < .001 (Table 1).

CD-Only

The prevalence of youth-reported CD-only (n = 29, 34.1%) was significantly higher than parent-reported CD-only (n = 10, 11.8%), χ² = 12.01(1), p = .001. According to the “OR”- and “AND” rules, the prevalence of CD-only was 27.1% (n = 23) and 4.7% (n = 4), respectively. The low “AND” rule prevalence was supported by an overall poor level of parent–youth agreement, κ = .18. The PA and NA were .30 and .87, respectively. Parents (n = 5, 5.9%) and girls (n = 14, 16.5%) did not significantly differ in uniquely reported CD-only, McNemar test = 3.37 (1), p = .064 (Table 1).

Between-Group Comparisons: CD + LPE versus CD-Only

Using girls as optimal informant (Table 2), higher levels of proactive aggression were found in CD + LPE girls, M (SD) = 9.00 (2.75) than in CD-only girls, M (SD) = 6.00 (3.71), U = 211.00, p = .018. This finding was not replicated when using parents as optimal informant. Using the “OR” rule (Table 2), no significant differences were revealed between CD + LPE and CD-only girls. Due to the small numbers of CD + LPE girls identified by the “AND”-, the “Unique Girl”-, and “Unique Parent” rules (see Table 1), between-group comparisons that were based on these approaches were not performed.

Table 2:

Between-Group Differences (CD-Only vs. CD + LPE) Using the Optimal Informant Approach and the “OR” Rule (n = 85)

	Girl			Parent			OR
	CD-only (1)(n = 29)	CD + LPE (2) (n = 11)	(1) vs. (2)	CD-only (1)(n = 10)	CD + LPE (2) (n = 33)	(1) vs. (2)	CD-only (1)(n = 23)	CD + LPE (2)(n = 40)	(1) vs. (2)
	n (%)	n (%)	p	n (%)	n (%)	p	n (%)	n (%)	p
Attention-deficit/hyperactivity disorder	7 (24.1)	5 (45.5)	.254	2 (20.0)	10 (30.3)	.698	7 (30.4)	12 (30.0)	1.000
Oppositional defiant disorder	12 (41.4)	7 (63.6)	.293	5 (50.0)	13 (39.4)	.717	10 (43.5)	16 (40.0)	.797
Any substance use disorder	26 (89.7)	11 (100.0)	.548	6 (60.0)	22 (66.7)	.719	18 (78.3)	29 (72.5)	.766
Any mood disorder	17 (58.6)	5 (45.5)	.498	4 (40.0)	14 (42.4)	1.000	12 (52.2)	17 (42.5)	.600
Any anxiety disorder	12 (41.4)	2 (18.2)	.266	5 (50.0)	8 (24.2)	.140	10 (43.5)	9 (22.5)	.085
	M (SD)	M (SD)	p	M (SD)	M (SD)	p	M (SD)	M (SD)	p
Reactive aggression	13.57 (3.74)	13.91 (3.53)	1.000	11.11 (3.06)	12.64 (4.37)	.333	12.77 (3.62)	12.63 (4.12)	.836
Proactive aggression	6.00 (3.71)	9.00 (2.75)	.018*	4.38 (3.20)	4.91 (3.96)	.859	5.23 (3.94)	5.38 (3.92)	.862
Nonviolent offending	6.21 (4.53)	8.00 (4.27)	.315	4.00 (3.89)	4.39 (3.89)	.745	5.05 (4.62)	5.28 (4.41)	.768
Violent offending	1.66 (1.61)	2.55 (1.51)	.124	.80 (1.48)	1.33 (1.59)	.327	1.09 (1.41)	1.60 (1.66)	.241

Note. There were no significant group differences regarding girls’ age and origin. CD = conduct disorder; LPE = with limited prosocial emotions.

p < .05. **p < .01.

Non-DSM CD Centric Between-Group Comparisons: Lpe Versus Non-LPE

Using girls as optimal informants, LPE (vs. non-LPE) girls had significantly higher levels of proactive aggression (M = 8.50, SD = 2.50 vs. M = 3.91, SD = 3.89; U = 803.50, p < .001), nonviolent offenses (M = 6.93, SD = 4.56 vs. M = 3.86, SD = 4.02; U = 735.50, p = .010), and violent offenses (M = 2.33, SD = 1.50 vs. M = 1.01, SD = 1.39; U = 784.50, p = .002). Using parents as optimal informants did not reveal significant group differences (Table 3). Using the “OR” rule, LPE (vs. non-LPE) girls had higher levels of nonviolent offenses (M = 5.28, SD = 4.41 vs. M = 5.05, SD = 4.62; U = 1,055.00, p = .029) and violent offenses (M = 1.60, SD = 1.66 vs. M = 1.09, SD = 1.41; U = 1,072.50, p = .031; Table 3). Unfortunately, the number of LPE girls identified by the “AND”-, “Unique Girl”-, and “Unique Parent” rules was too low to perform between-group comparisons.

Table 3:

Non-DSM CD Centric Between-Group Differences (Non-LPE vs. LPE) Using the Optimal Informant Approach and the “OR” Rule (n = 85)

	Girl			Parent			OR
	Non-LPE (1)(n = 70)	LPE (2)(n = 15)	(1) vs. (2)	Non-LPE (1)(n = 37)	LPE (2)(n = 48)	(1) vs. (2)	Non-LPE (1)(n = 32)	LPE (2)(n = 53)	(1) vs. (2)
	n (%)	n (%)	p	n (%)	n (%)	p	n (%)	n (%)	p
Attention-deficit/hyperactivity disorder	17 (24.3)	5 (33.3)	.521	10 (27.0)	12.0 (25.0)	1.000	9 (28.1)	13 (24.5)	.800
Oppositional defiant disorder	22 (31.4)	8 (53.3)	.139	14 (37.8)	16 (33.3)	.819	12 (37.5)	18 (34.0)	.816
Any substance use disorder	45 (64.3)	13 (86.7)	.131	25 (67.5)	33 (68.8)	1.000	20 (62.5)	38 (71.7)	.625
Any mood disorder	29 (41.4)	5 (33.3)	.772	16 (43.2)	18 (37.5)	.658	14 (43.8)	20 (37.7)	.651
Any anxiety disorder	25 (35.7)	3 (20.0)	.365	14 (37.8)	14 (29.2)	.362	12 (37.5)	16 (30.2)	.477
	M (SD)	M (SD)	p	M (SD)	M (SD)	p	M (SD)	M (SD)	p
Reactive aggression	11.84 (4.29)	13.13 (3.58)	.446	11.44 (4.07)	12.54 (4.25)	.325	12.77 (3.62)	12.63 (4.11)	.448
Proactive aggression	3.91 (3.89)	8.50 (2.50)	.001**	4.80 (4.37)	4.62 (3.88)	1.000	5.23 (3.94)	5.38 (3.92)	.441
Nonviolent offending	3.86 (4.02)	6.93 (4.56)	.010*	4.33 (4.67)	4.46 (3.97)	.458	5.05 (4.62)	5.28 (4.41)	.029*
Violent offending	1.01 (1.39)	2.33 (1.50)	.002**	1.14 (1.51)	1.33 (1.48)	.419	1.09 (1.41)	1.60 (1.66)	.031*

Note. There were no significant group differences regarding girls’ age and origin. DSM = Diagnostic and Statistical Manual of Mental Disorders; CD = conduct disorder; LPE = with limited prosocial emotions.

p < .05. **p < .01.

Discussion

The overall aim of this study was to examine the prevalence of detained girls who met the DSM-5 LPE specifier and to test whether CD + LPE girls differed from CD-only girls on clinically important features. In line with prior prevalence studies among detained girls (Teplin, Abram, McClelland, Dulcan, & Mericle, 2002; Van Damme, Colins, & Vanderplasschen, 2014), a substantial proportion of girls met criteria for CD (47.1%). In addition, 27.5% of CD girls met the LPE specifier threshold, a finding that also converges well with prior APSD-SR work in detained girls (26%; Colins & Andershed, 2015) and clinic-referred youths (21%; Kahn et al., 2012). Using parents as informants, almost 77% of the CD girls were also identified as being with LPE. This is remarkably higher than the 10%-11% and 31% APSD parent version–based prevalence among community and clinic-referred youths with CD, respectively (Kahn et al., 2012; McMahon et al., 2010). On the one hand, these findings suggest that parents of detained youths indeed are important to identify CD + LPE girls. On the other hand, unique parent information is not necessarily synonymous with accurate information. Indeed, the reliability of CD- and LPE-related information provided by parents of detained girls might be limited for various reasons, such as having too limited contact to accurately estimate the frequency of certain behaviors or traits and overestimating symptoms due to parental stress caused by their child’s behavior or due to features from parents themselves (e.g., depression; Colins et al., 2008; De Los Reyes & Kazdin, 2005).

In line with all prior work in detained youths who used the APSD or alternative tools to assess the DSM-5-defined LPE specifier (Colins & Andershed, 2015; Colins & Vermeiren, 2013), CD + LPE and CD-only girls did not differ in the prevalence of ADHD, ODD, SUD, and affective and anxiety disorders, regardless of the informant being used. Possibly, the often reported comorbidity between CD and other psychiatric disorders in detained girls (Teplin et al., 2002; Van Damme et al., 2014) leaves little room for the LPE specifier to identify CD girls with different levels of mental health problems (Colins & Andershed, 2015). Also, CD + LPE girls and CD-only girls were not different in mean levels of violent and nonviolent offenses, a finding that runs counter to the view that the LPE specifier will identify a severe antisocial subgroup of girls (Frick & Dickens, 2006). Our sample may exhibit a ceiling effect in terms of behavioral problems, thereby restricting the likelihood to detect differences between CD + LPE and CD-only girls. Importantly, CD + LPE girls did show the highest levels of proactive aggression while being similar to CD-only girls in their levels of reactive aggression. This finding converges with prior work in detained girls showing that CU traits were related to proactive aggression but not to reactive aggression (Marsee & Frick, 2007). Taken together, our study contributes to the literature by showing that detained CD + LPE girls are the most severe antisocial girls if one focuses on a specific, and relatively rare, but severe form of aggression.

Recently, it has been argued that being too DSM centric may limit our understanding of the potential role of the DSM-5 specifier in designating a subgroup of youths with serious conduct problems that may not meet criteria for CD (Frick et al., 2014b). Our study showed that using the LPE specifier in a non-DSM CD centric manner (i.e., without requiring that girls also meet criteria for CD) did not identify differences in psychiatric morbidity between LPE and non-LPE girls. Interestingly, LPE girls reported more violent and nonviolent offenses and higher levels of proactive aggression than non-LPE girls, suggesting that using the LPE specifier in a non-DSM CD centric manner increases the ability to identify a more severe subset of antisocial girls in detention. Future studies are warranted to see whether these non-DSM CD centric findings can be replicated in other samples of detained youths and by means of alternative tools to assess the LPE criteria. In this respect, a recent study showed that the likelihood of such alternative tools to assess the LPE specifier may depend on the number of items used to assess the LPE specifier as well as to the coding method (i.e., what item score is needed to endorse an LPE specifier criterion; Kimonis et al., 2014).

Finally, our results must be considered against the DSM-5’s emphasis to extend self-report with report from parents (APA, 2013). The current study results point to the importance of obtaining measures of predictor and outcome variables from different informants, as a way to control for potential method bias (Podsakoff, MacKenzie, & Podsakoff, 2012). Support for the clinical value of the LPE specifier in identifying a subgroup of seriously antisocial girls, whether in a DSM centric manner or not, was only revealed when using girls as optimal informant or when applying the “OR” rule, but not when parent-reports were considered as an optimal source of information. Because measures of aggression and offending were solely based on youth self-report, shared method variance is likely to explain our findings in support of the LPE specifier (i.e., the higher levels of proactive aggression in CD + LPE girls, Table 2; and the higher levels of proactive aggression and offenses in LPE girls, Table 3).

Detention facilities often have limited resources to expend on locating and interviewing parents (Ko et al., 2004), indicating that it is relevant to know whether or not the energy and time to recruit parents is worth the effort. Altogether, our findings suggest that gathering parental information might be useful to gain unique information on the prevalence of CD + LPE, but that gathering parental information on CD and LPE is not worth the effort if identifying a subgroup of seriously antisocial girls is the ultimate purpose. Interestingly, the DSM-5 also states that reports from informants other than parents, such as teachers and peers, should be considered in the assessment of LPE. But this may even be more challenging, given the often disrupted school career and high dropout of detained adolescents (Kroll et al., 2002) and given the unlikelihood that peers will provide information that (allegedly) may be used against the detained girl or boy. In that case, alternative sources, such as clinical ratings or observational information of detention personnel, are urgently needed. Yet, training detention staff to observe and report about their observations in a standardized manner and empirically testing the usefulness of this source of information will cost considerable time and (financial) efforts.

This study has several strengths, including the use of an understudied but highly relevant population to test the usefulness of the DSM-5 specifier, and the use of multiple informants and well-validated questionnaires. As always, the results should be interpreted in the context of some limitations. First, notwithstanding that the DSM-5 states that LPE criteria must be present over at least 12 months, the APSD does not refer to any time frame in particular. Consequently, parents may have recalled and utilized more historic factors in rating CD or LPE, compared with their daughters. To adequately test the LPE specifier as operationalized within the DSM-5, future studies are warranted that also assess the specified time frame. Second, this study only used one measure to assess LPE criteria. Therefore, it is possible that the DSM-5 LPE specifier construct has utility, but that the measurement of it (i.e., via the APSD) is inadequate. Prior work in detained adolescents on CD + LPE (Colins & Andershed, 2015) or LPE (Kimonis et al., 2014) indeed suggests that using alternative measures (with more items to assess the LPE criteria) may increase support in favor of the LPE specifier. Third, although our sample is unique and difficult to recruit, we cannot exclude the possibility that the relative small number of girls in some of the groups has restricted the power to reveal significant between-group differences. Power issues also hampered to test the usefulness of the “AND”-, the “Unique Youth”-, and “Unique Parent” rules, and of combining the LPE specifier with the age-of-onset subtyping (e.g., APA, 2013; Colins & Vermeiren, 2013; Frick & Dickens, 2006). Fourth, and in line with prior work on CD among detained youths (Colins et al., 2008; Ko et al., 2004), parental information was gathered by telephone, while a face-to-face interview might be more appropriate. Fifth, due to the cross-sectional study design, it remains to be seen how stable the LPE specifier assignment is and whether this specifier has prognostic usefulness. Therefore, longitudinal studies among detained girls are urgently warranted, especially because it has been shown that only 14.5% of girls initially classified as CD + LPE in childhood (ages 6-8) were identified as such 6 years later (Pardini et al., 2012) and that CD + LPE (vs. CD-only) children are not at increased risk for future antisocial behavior (Kolko & Pardini, 2010; Pardini et al., 2012) and recidivism (Colins & Vermeiren, 2013). Finally, because of our focus on the DSM-5 LPE specifier, we did not consider other categorical approaches (e.g., Rowe et al., 2010) or dimensional approaches (e.g., Pardini et al., 2012) to incorporate CU traits into the diagnosis of CD. Thus, studies on these topics are needed.

In conclusion, this study showed that the prevalence of the DSM-5 LPE specifier was the highest when using parent-reports, a finding that seems to underscore the relevance of using parent-ratings of limited prosocial emotions. However, including parental information did not result in a better differentiation between CD + LPE and CD-only girls, or between LPE versus non-LPE girls. This suggests that the lack of support for the clinical usefulness of the DSM-5 specifier in prior studies among detained adolescents cannot solely be explained by their sole reliance on self-report. Altogether, our findings suggest that self-report remains an important and cost-effective source of information that must be used in future studies on the DSM-5 specifier in detained adolescents.

Footnotes

This study is funded by the Special Research Fund from Ghent University.

The authors declare no conflicts of interest.

Notes

Lore Van Damme works as a PhD student at the Department of Special Needs Education, Ghent University (Belgium). Her research consists of a prospective cohort study focusing on detained girls’ psychopathology, quality of life, and social adaptation prior to, during, and after detention.

Olivier F. Colins, PhD, is a senior researcher at the Department of Child and Adolescent Psychiatry, Curium/Leiden University Medical Center (LUMC; The Netherlands). His main research areas of interest are forensic mental health screening and assessment; the manifestation, stability, and outcomes of psychopathic personality; and empathy-related deficits in adolescents with aggressive conduct disorder.

Wouter Vanderplasschen, PhD, is associate professor at the Department of Special Needs Education of Ghent University (Belgium). He is a senior researcher in the field of case management, and substance abuse treatment and recovery.

References

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

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.

Bijttebier

Decoene

(2009). Assessment of psychopathic traits in children and adolescents further validation of the Antisocial Process Screening Device and the Childhood Psychopathy Scale. European Journal of Psychological Assessment, 25, 157-163. doi:10.1027/1015-5759.25.3.157

Cima

Raine

Meesters

Popma

(2013). Validation of the Dutch Reactive Proactive Questionnaire (RPQ): Differential correlates of reactive and proactive aggression from childhood to adulthood. Aggressive Behavior, 39, 99-113. doi:10.1002/Ab.21458

Colins

O. F.

Andershed

(2015). The DSM-5 with limited prosocial emotions specifier for conduct disorder among detained girls. Law and Human Behavior, 39(2), 198-207. doi:10.1037/lhb0000108

Colins

O. F.

Bijttebier

Broekaert

Andershed

(2014). Psychopathic-like traits among detained female adolescents: Reliability and validity of the Antisocial Process Screening Device and the Youth Psychopathic Traits Inventory. Assessment, 21, 195-209. doi:10.1177/1073191113481997

Colins

O. F.

Vermeiren

R. R. J.

(2013). The usefulness of DSM-IV and DSM-5 conduct disorder subtyping in detained adolescents. Journal of Nervous and Mental Disease, 201, 736-743. doi:10.1097/Nmd.0b013e3182a20e94

Colins

O. F.

Vermeiren

Schuyten

Broekaert

(2009). Psychiatric disorders in property, violent, and versatile offending detained male adolescents. American Journal of Orthopsychiatry, 79, 31-38. doi:10.1037/A0015337

Colins

O. F.

Vermeiren

Schuyten

Broekaert

Soyez

(2008). Informant agreement in the assessment of disruptive behavior disorders in detained minors in Belgium: A diagnosis-level and symptom-level examination. Journal of Clinical Psychiatry, 69, 141-148.

10.

Colins

O. F.

Vermeiren

Vahl

Markus

Broekaert

Doreleijers

(2012). Parent-reported attention-deficit hyperactivity disorder and subtypes of conduct disorder as risk factor of recidivism in detained male adolescents. European Psychiatry, 27, 329-334. doi:10.1016/j.eurpsy.2011.01.001

11.

Crowley

T. J.

Mikulich

S. K.

Ehlers

K. M.

Whitmore

E. A.

MacDonald

M. J.

(2001). Validity of structured clinical evaluations in adolescents with conduct and substance problems. Journal of the American Academy of Child & Adolescent Psychiatry, 40, 265-273. doi:10.1097/00004583-200103000-00005

12.

De Los Reyes

Kazdin

A. E.

(2005). Informant discrepancies in the assessment of childhood psychopathology: A critical review, theoretical framework, and recommendations for further study. Psychological Bulletin, 131, 483-509. doi:10.1037/0033-2909.131.4.483

13.

Ferdinand

Van der Ende

(2002). Diagnostic Interview Schedule for Children. Authorized Dutch translation of the NIMH DISC-IV ( Shaffer

Fisher

Lucas

& the NIMH editorial board, Eds.). Rotterdam: Department of Child and Adolescent Psychiatry, Sophia Children’s Hospital, Erasmus University.

14.

Fink

B. C.

Tant

A. S.

Tremba

Kiehl

K. A.

(2012). Assessment of psychopathic traits in an incarcerated adolescent sample: A methodological comparison. Journal of Abnormal Child Psychology, 40, 971-986. doi:10.1007/s10802-012-9614-y

15.

Frick

P. J.

Dickens

(2006). Current perspectives on conduct disorder. Current Psychiatry Reports, 8, 59-72. doi:10.1007/s11920-006-0082-3

16.

Frick

P. J.

Hare

R. D.

(2001). Antisocial Process Screening Device technical manual. Toronto, Ontario, Canada: Multi-Health Systems.

17.

Frick

P. J.

Nigg

J. T.

(2012). Current issues in the diagnosis of attention deficit hyperactivity disorder, oppositional defiant disorder, and conduct disorder. Annual Review of Clinical Psychology, 8(8), 77-107. doi:10.1146/annurev-clinpsy-032511-143150

18.

Frick

P. J.

Ray

J. V.

Thornton

L. C.

Kahn

R. E.

(2014a). Can callous-unemotional traits enhance the understanding, diagnosis, and treatment of serious conduct problems in children and adolescents? A comprehensive review. Psychological Bulletin, 140, 1-57. doi:10.1037/A0033076

19.

Frick

P. J.

Ray

J. V.

Thornton

L. C.

Kahn

R. E.

(2014b). The road forward for research on callous-unemotional traits: Reply to Lahey (2014). Psychological Bulletin, 140, 64-68. doi:10.1037/A0033710

20.

Kahn

R. E.

Frick

P. J.

Youngstrom

Findling

R. L.

Youngstrom

J. K.

(2012). The effects of including a callous-unemotional specifier for the diagnosis of conduct disorder. The Journal of Child Psychology and Psychiatry, 53, 271-282. doi:10.1111/j.1469-7610.2011.02463.x

21.

Kimonis

E. R.

Fanti

Goldweber

Marsee

M. A.

Frick

P. J.

Cauffman

(2014). Callous-unemotional traits in incarcerated adolescents. Psychological Assessment, 26, 227-237. doi:10.1037/A0034585

22.

S. J.

Wasserman

G. A.

McReynolds

Katz

L. M.

(2004). Contribution of parent report to voice DISC-IV diagnosis among incarcerated youths. Journal of the American Academy of Child & Adolescent Psychiatry, 43, 868-877. doi:10.1097/01.chi.0000128788.03192.fa

23.

Kolko

D. J.

Pardini

D. A.

(2010). ODD dimensions, ADHD, and callous-unemotional traits as predictors of treatment response in children with disruptive behavior disorders. Journal of Abnormal Psychology, 119, 713-725. doi:10.1037/A0020910

24.

Kroll

Rothwell

Bradley

Shah

Bailey

Harrington

R. C.

(2002). Mental health needs of boys in secure care for serious or persistent offending: A prospective, longitudinal study. The Lancet, 359, 1975-1979. doi:10.1016/S0140-6736(02)08829-3

25.

Lahey

B. B.

(2014). What we need to know about callous-unemotional traits: Comment on Frick, Ray, Thornton, and Kahn (2014). Psychological Bulletin, 140, 58-63. doi:10.1037/A0033387

26.

Landis

J. R.

Koch

G. G.

(1977). Measurement of observer agreement for categorical data. Biometrics, 33, 159-174. doi:10.2307/2529310

27.

Loeber

Green

S. M.

Lahey

B. B.

Stouthamer-Loeber

(1989). Optimal informants on childhood disruptive behaviors. Development and Psychopathology, 1, 317-337.

28.

Mackinnon

(2000). A spreadsheet for the calculation of comprehensive statistics for the assessment of diagnostic tests and inter-rater agreement. Computers in Biology and Medicine, 30, 127-134.

29.

Marsee

M. A.

Frick

P. J.

(2007). Exploring the cognitive and emotional correlates to proactive and reactive aggression in a sample of detained girls. Journal of Abnormal Child Psychology, 35, 969-981. doi:10.1007/s10802-007-9147-y

30.

McMahon

R. J.

Witkiewitz

Kotler

J. S.

(2010). Predictive validity of callous-unemotional traits measured in early adolescence with respect to multiple antisocial outcomes. Journal of Abnormal Psychology, 119, 752-763. doi:10.1037/a0020796

31.

Pardini

Stepp

Hipwell

Stouthamer-Loeber

Loeber

(2012). The clinical utility of the proposed DSM-5 callous-unemotional subtype of conduct disorder in young girls. Journal of the American Academy of Child & Adolescent Psychiatry, 51, 62-73. doi:10.1016/j.jaac.2011.10.005

32.

Podsakoff

P. M.

MacKenzie

S. B.

Podsakoff

N. P.

(2012). Sources of method bias in social science research and recommendations how to control it. Annual Review of Psychology, 63, 539-569. doi:10.1146/annurev-psych-120710-100452

33.

Poythress

N. G.

Douglas

K. S.

Falkenbach

Cruise

Lee

Murrie

D. C.

Vitacco

(2006). Internal consistency reliability of the self-report antisocial process screening device. Assessment, 13, 107-113. doi:10.1177/1073191105284279

34.

Raine

Dodge

Loeber

Gatzke-Kopp

Lynam

Reynolds

Liu

(2006). The Reactive-Proactive Aggression Questionnaire: Differential correlates of reactive and proactive aggression in adolescent boys. Aggressive Behavior, 32, 159-171. doi:10.1002/ab.20115

35.

Rowe

Maughan

Moran

Ford

Briskman

Goodman

(2010). The role of callous and unemotional traits in the diagnosis of conduct disorder. The Journal of Child Psychology and Psychiatry, 51, 688-695. doi:10.1111/j.1469-7610.2009.02199.x

36.

Shaffer

Fisher

Lucas

C. P.

Dulcan

M. K.

Schwab-Stone

M. E.

(2000). NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses. Journal of the American Academy of Child & Adolescent Psychiatry, 39, 28-38. doi:10.1097/00004583-200001000-00014

37.

Teplin

L. A.

Abram

K. M.

McClelland

G. M.

Dulcan

M. K.

Mericle

A. A.

(2002). Psychiatric disorders in youth in juvenile detention. Archives of General Psychiatry, 59, 1133-1143. doi:10.1001/archpsyc.59.12.1133

38.

Van Damme

Colins

Vanderplasschen

(2014). Gender differences in psychiatric disorders and clusters of self-esteem among detained adolescents. Psychiatry Research, 220, 991-997. doi:10.1016/j.psychres.2014.10.012

39.

van der Laan

A. M.

Blom

(2005). Jeugddelinquentie: risico’s en bescherming: Bevindingen uit de WODC monitor zelfgerapporteerde jeugdcriminaliteit 2005 [Juvenile delinquency: risks and protective factors; findings of the WODC Youth Delinquency Survey 2005]. Den Haag, The Netherlands: Boom Juridische Uitgevers.