Parental attitudes towards early intervention in children at high risk for affective disorders

Journal of Affective Disorders 70 (2002) 117–124 www.elsevier.com / locate / jad

Special report

Parental attitudes towards early intervention in children at high risk for affective disorders Robert M. Post*, Gabriele S. Leverich, Emily Fergus, Rachel Miller, David Luckenbaugh Biological Psychiatry Branch, National Institute of Mental Health, NIH, Bldg. 10, Rm. 3 S239, 10 Center Drive, MSC-1272, Bethesda, MD 20892 -1272, USA Received 19 May 2000; accepted 21 December 2000

Abstract Background: Parents volunteered to complete surveys on attitudes toward treatment intervention in children at a theoretically high (20–30%) or very high (70%) risk for affective disorders because of an assumed uni-lineal or bi-lineal family history of bipolar illness. Methods: Questions focused on examining at what ages and stage of symptom and syndrome evolution parents would wish their child to begin treatment with different types of therapeutic approaches and clinical trial designs. Sixty percent of the respondents had a personal history of unipolar or bipolar affective disorders. Results: In 156 completed surveys, 83% of parents favored acute medication intervention and 67% favored long-term medication treatment for those children at very high risk at or before the development of severe symptoms (i.e. even prior to meeting full diagnostic thresholds). On the average, parents indicated that they would enter their child in a trial of: two types of psychotherapy at a point in illness evolution between moderate and severe symptoms, two types of open medications between severe symptoms and a definite diagnosis, two blind medications at a definite diagnosis, and a blind trial of placebo and medication after a definite diagnosis but before multiple recurrences of the illness. Parents, primarily on the basis of perceived safety, would allow their children to use medications that have been found to be effective in adults. Limitations: In addition to a number of methodological limitations, responders to the survey were self-selected. Conclusions: The results indicate a willingness on the part of most parents to treat a child at high risk for affective illness early in the course of symptom evolution, even prior to a full syndromic illness or diagnosis. This and other parental views of the risk–benefit and ethical dimensions of early intervention may be helpful in the initiation and design of studies aimed at assessing the efficacy of early interventions in childhood-onset bipolar illness and its prodromes.  2002 Elsevier Science B.V. All rights reserved. Keywords: Bipolar disorder; Mania; Childhood; Early intervention

1. Introduction *Corresponding author. Tel.: 11-301-496-4805; fax: 11-301402-0052.

In contrast to several decades ago, there is now increased recognition of the prevalence and severity

0165-0327 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved. PII: S0165-0327( 01 )00299-3

118

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

of unipolar and bipolar affective disorders in children and adolescents. The cohort effect, as described by Gershon et al. (1987) and Weissman et al. (1992), reflects that each birth cohort since World War II appears to have an increased incidence as well as an earlier age of onset of both unipolar and bipolar affective illness. The psychosocial and neurobiological mechanisms for this phenomenon are not known, but are widely discussed. They range from the breakdown of the family unit and increases in substance use to potential genetic mechanisms, such as those of genetic anticipation. A number of studies have suggested that the onset of affective disorders in offspring of parents with affective illnesses may precede that of their parents by as much as 10 years (McInnis et al., 1993; McMahon et al., 1994). Additionally, there is increasing concern about the potential morbidity and mortality of the childhoodand adolescent-onset of affective illness. In adults, there is a 10–20% lifetime mortality by suicide (Goodwin and Jamison, 1990). In the follow-up studies of Weissman et al. (1999), 27% of patients with childhood-onset unipolar depression attempted suicide (n 5 23). The prognosis regarding functional outcome of adolescent-onset bipolar disorder is equally troubling, with the small study by Gillberg et al. (1993) indicating that 85.7% of patients hospitalized with adolescent mania emerged as highly disabled upon follow-up evaluation one to three decades later. These empirical data demonstrate the potential importance of intervening earlier and more effectively in children and adolescents with unipolar and bipolar affective disorders in order to alter this potentially adverse course. This is particularly critical given that (a) several studies of adult-onset bipolar illness reveal an average delay of 5–10 years between the onset of first symptoms sufficient to reach diagnostic thresholds and first treatment (Suppes et al., 2001; Lish et al., 1994; Egeland et al., 1987) and (b) the issue of early recognition and treatment in children appears to be even more complex than in adults. Current problems in the field include concerns about the efficacy and safety of somatic treatment in children (Werry, 1993), as well as ongoing controversies about diagnostic criteria, thresholds for childhood-onset bipolar disorder, and their congruence with adult presentations of the

illness (Anthony and Scott, 1960; Weinberg and Brumback, 1976; Strober, 1992; Werry, 1992; Weller et al., 1995; Geller and Luby, 1997; Biederman et al., 1998; Sanchez et al., 1999; Carlson et al., 2000). In an attempt to formulate the most appropriate designs for early intervention studies that balance efficiency, practicality, safety, and ethics, we decided not only to consult our colleagues about what they thought were the most ethical approaches to studies of early intervention in childhood bipolar disorder, but also to directly explore parental attitudes on this topic. As such, we designed a survey that would approach the issues of treatment intervention in children: at given thresholds of illness severity, at different age cutoffs, and at different degrees of established efficacy, safety, and side-effect potentials for various treatments. We were particularly interested in assessing attitudes toward intervention in children at high risk (one parent with bipolar disorder) and at very high risk (a bi-lineal family history of affective disorder with bipolar disorder on at least one side) (Gershon et al., 1982; Lapalme et al., 1997), because the clinical, ethical, and safety risk– benefit ratios would likely be shifted toward earlier intervention compared with those at low risk.

2. Methods The survey was advertised in the Bipolar Network News (Leverich et al., 2001) and several mental health advocacy group newsletters, and was also sent on request to parents. The advertisement indicated that surveys would be coded with numbers insuring anonymity. Out of 190 biological parents (mothers and fathers) who reported their own diagnoses on the first 156 forms received, 70 (36.8%) had a bipolar diagnosis, 45 (23.7%) had a unipolar diagnosis, and 59 (31.1%) reported having no diagnosis. The survey focused on different time frames and types of potential intervention in children at high risk (Section A). These included: (1) prior to symptom onset, (2) at onset of minor symptoms, (3) at onset of moderate symptoms, (4) at onset of severe symptoms, (5) at time of definite diagnosis, (6) after diagnosis and multiple episodes, and (7) never. Interventions included: (a) watchful observation, (b) acute counseling, (c) long-term counseling, (d) acute

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

medication (targeted), (e) long-term medication, (f) electroconvulsant therapy (ECT), (g) repeated transcranial magnetic stimulation (rTMS) of the brain, and (h) rTMS if and when it was clearly effective in adults and did not appear to have serious sideeffects. Parents were asked at what stage of illness evolution they would consider each of these types of interventions, if the chance of their children developing an affective illness was either 20–30% or 70%. These percentages were specified based on preliminary data suggesting that these may be the incidences of affective illness in offspring of families where there was a uni-lineal and bi-lineal family history of bipolar illness respectively (Gershon et al., 1982). In addition, this preliminary data was not inconsistent with the meta-analysis of Lapalme et al. (1997) for one affected parent. In relation to potential research participation, parents rated at what point in their child’s course of illness they were willing to enter their child into a series of different trial designs (Section B). Trial designs included: (a) an open trial of two types of psychotherapy, (b) an open trial of psychotherapy versus a medication, (c) an open trial of two different medications, (d) a blind trial of two different medications, (e) a blind trial of placebo versus an active medication (noted as ‘‘scientifically the best type of trial to prove efficacy’’), and (f) other (‘‘please specify’’). In Section C, parents rated when they would consider starting their child on particular medication treatments. Medications listed on the survey included: antidepressants (tricyclics, monoamine oxidase inhibitors (MAOIs), serotonin-selectives (SSRIs), and bupropion), mood stabilizers (lithium, carbamazepine, valproate, lamotrigine, gabapentin), calcium-channel blockers, and major tranquilizers (haloperidol, clozapine, risperidone, and olanzapine). Each drug or drug class was given a general rating of: efficacy in adults (1 to 1 1 1 1), efficacy in children (usually questionable to unknown), and the degree of side-effects rated as ‘rare’ to ‘frequent.’ In addition, the need for dietary restriction was noted for MAOIs and weekly blood monitoring for clozapine. Section D included a series of questions for parents related to the age at which they would

119

consider medication intervention for their child, as a function of stage of illness evolution, with agents known to be effective in adult bipolar illness, but with unknown efficacy in children. Chi square analyses ( x 2 ) were completed to determine if significant differences were present based on type of potential treatment selected and percentage of risk of illness. Post hoc contrasts were used to test for differences among the graded levels of the potential treatment variable. Analyses of variance (ANOVAs) were performed to determine the mean points in their child’s illness at which parents would institute the various types of interventions. We used the Greenhouse–Geisser epsilon to adjust the degrees of freedom in a repeated measures ANOVA when the assumption of sphericity (or compound symmetry or homogeneity of variance–covariance matrices) was violated.

3. Results More than 60% of parents thought that acute medication interventions were warranted at the onset of moderate symptoms in children at a very high (70%) hypothetical risk of developing an affective illness. This increased to 80% of parents endorsing medication at the onset of severe symptoms in these children. Ninety-nine percent of parents indicated that they would agree to treat their children with acute medication by the time of a definite diagnosis. The figures for long-term medication were somewhat lower, with approximately 45% of parents indicating they would begin this type of treatment at moderate symptoms, 65% with severe symptoms, and 93% with a definite diagnosis. Only 7.1% of parents indicated that they would wait for the occurrence of multiple episodes and a diagnosis, in order to institute long-term treatment in children at very high risk (Table 1). Although the difference was not significant, the percentage of parents endorsing treatment intervention at a given point in illness were generally lower for children at a 20–30% hypothetical risk of affective illness compared with a 70% hypothetical risk. The point of symptom and illness progression at which a parent would intervene also interacted with a

120

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

Table 1 Percentage of parents who would institute different types of treatment at various points of symptom evolution in children who were at (a) high (20–30%) or (b) very high (70%) risk of developing an affective illness Time of onset of treatment

Type of treatment Careful observation

Counselling

Medication

Acute

LT

a

(a) Percent of parents endorsing time and type of treatment for a child at high risk Prior to symptom onset 46.2 3.8 2.4 Mild symptoms 35.6 30.0 15.7 Moderate symptoms 14.4 42.3 30.7 Severe symptoms 2.3 13.1 25.2 At diagnosis 1.5 6.9 13.4 After dx and multiple episodes – 2.3 8.7 Never – 1.5 3.9 ( b) Percent of parents endorsing treatment for a child at very high risk b Prior to symptom onset 57.7 6.5 Mild symptoms 30.3 43.2 Moderate symptoms 10.6 35.3 Severe symptoms – 9.4 At diagnosis 1.4 3.6 After dx and multiple episodes – 0.7 Never – 1.4

5.1 22.6 35.0 21.2 8.8 4.4 2.9

Acute

LT a

ECT

– 11.5 32.8 21.4 26.7

1.5 6.9 21.5 19.2 32.3

0.8 1.6 – 3.1 3.1

6.9 0.8

17.7 0.8

28.3 63.0

0.7 15.6 44.7 20.6 17.7

2.1 7.8 36.2 21.3 24.8

0.7 2.2 0.7 5.2 3.7

0.7

7.1 0.7

32.6 54.8

–

a

LT, long term or prophylactic treatment. Percentages were not significantly different for percent of parents endorsing treatment as a function of the hypothetical high risk (20–30%) or very high risk (70%) of a child. dx5diagnosis; ECT5electroconvulsive therapy. b

child’s age. The average age at which parents recommended intervention for a high-risk child with severe symptoms was 7.9 years, with moderate symptoms was 9.5 years, and for primary preventive treatment (i.e. before the onset of symptoms) was 13 years. On the average, parents preferred acute counseling at the onset of moderate symptoms, long-term counseling and acute medication in between moderate and severe symptoms, and long-term medication at the onset of severe symptoms. On the average, parents would consider entering their child in different clinical trial designs at: greater than moderate symptoms for an open trial of two psychotherapies; at severe symptoms for an open randomized trial of medications versus therapy; between severe symptoms and a definite diagnosis for a trial of two different medications; at a definite diagnosis for a blind trial of two different medica-

tions; and between a definite diagnosis and diagnosis with multiple episodes for a blind trial comparing a placebo versus a medication (Fig. 1). The average point in the course of illness at which parents would consider the use of SSRIs, and bupropion, as well as lithium, carbamazepine, and valproate for their high-risk child is severe symptoms. Tricyclic antidepressants would be used between severe symptoms and definite diagnosis, while the putative anticonvulsants lamotrigine and gabapentin, or calcium-channel blockers would be considered for use ‘at diagnosis.’ On the average, all of the typical and atypical major tranquilizers and the MAOIs were considered for use at a higher threshold (i.e. later in the course of illness). No differences in parents’ preferences were found when comparing those parents with and without a diagnosis of bipolar illness.

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

121

Fig. 1. Average point in illness progression (mean6standard error of the mean) when parents would have children at high risk participate in different types of treatment trials.

4. Discussion

4.1. Limitations of this study This group of subjects was highly self-selected, with 60% of the parents who requested and responded to the survey reporting personal experience with either unipolar or bipolar affective illness. This sample is clearly not intended to represent an unbiased sample nor to be representative of the general population. It remains unclear how the nonresponders to the survey and parents more naive to psychiatric illness in themselves or others would have responded to the questionnaire. Another limitation of the study was the stated designation of the 20–30% and 70% hypothetical risks of a child developing an affective illness (Gershon et al., 1982; Lapalme et al., 1997). It was not clear whether survey participants clearly under-

stood that the risk referred to a child developing any affective illness as opposed to bipolar illness in particular. However, the relatively small and statistically nonsignificant differences in time to intervention as a function of a 20–30% versus 70% hypothetical risk (in Table 1) suggest that perceived differences in relative high risk are not as major a factor as one might have surmised in the parents’ decision of when to intervene. It would have been useful to compare these choices to those made for children at low risk in the general population. In the survey, symptom presentation prior to diagnosis was only referred to globally as that associated with mild, moderate, or severe dysfunction in the child’s usual social or educational environment. The symptoms themselves were not specified in the survey nor was the exact diagnosis (e.g., a conduct disorder, major depression, or classic mania). The details of each of these may have

122

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

changed the percentage of parents endorsing a given question. Finally, only a brief overview of the relative efficacy and side-effect risks for the different treatments in adults was indicated. How the assessments might have changed with much more specific and detailed information about each treatment is also uncertain.

4.2. Implications Given these caveats, perhaps the most striking findings of the survey were those indicating that 84% of parents endorsed acute medication treatment and 70% endorsed long-term medication treatment at or before the time of onset of severe symptoms (i.e. even prior to a definite diagnosis for children at a 20–30% hypothetical risk of experiencing an affective disorder). Treatment preferences were not significantly different for those children at a much higher (70%) risk. Parents perceived differential risk among various medications, and were much more likely to consider SSRIs and bupropion rather than an MAOI. In addition, lithium, carbamazepine, and valproate were favored over the lesser studied agents such as lamotrigine, gabapentin, and calcium-channel blockers even though the latter two compounds were rated safe with few serious side-effects in adults. On the average, the typical and atypical antipsychotics (major tranquilizers) were considered for use only after a full diagnosis had been achieved. When asked at what point they would consider allowing a child at high risk to participate in treatment efficacy research protocols, parents responded in an ever more conservative fashion as the implied risk of the trial increased. Parents would first consider an open clinical trial of two psychotherapies or an open trial of two medications, followed by a blind clinical trial of two active medications, and lastly only after a definite diagnosis had been made, a blind trial comparing a medication to placebo. These data suggest that the majority of parents are in favor of instituting early treatment for children at high risk around the time of development of severe symptoms, even prior to a definite diagnosis. On the average, parents would be willing to enter their children into open clinical trials at the point of severe

symptoms. However, entering their children into blind clinical trials would be contingent upon a definite diagnosis. Despite the considerable methodological limitations enumerated above, the data from the current survey support the general view that parents are interested in having children at high risk for affective disorder treated early in the course of their symptom evolution. This view is held despite the statements and knowledge that formal data from controlled clinical trials on efficacy of most antidepressants and mood stabilizers in children and adolescents are limited. Although clinical trials for early intervention in patients with premonitory symptoms of schizophrenia are underway in several clinical research settings in the U.S. and Australia (Edwards et al., 1999; McGorry, 1998), relatively few parallel attempts are occurring in childhood- and adolescent-onset bipolar illness in those at high risk. However, the recognition that extreme affective dysfunction and impairment can occur in very young, even pre-latency individuals (Geller and Luby, 1997; Emslie et al., 1995; Strober et al., 1998; Papolos and Papolos, 2000), suggests the potential importance of intervening in this younger age group. The not-always-benign longterm outcome of bipolar illness, even when compared with schizophrenia (Gillberg et al., 1993), would also appear to impel greater attention and efforts at early intervention. The prominent controversies around the diagnosis of childhood-onset bipolar illness (Anthony and Scott, 1960; Weinberg and Brumback, 1976; Strober, 1992; Werry, 1992, 1993; Weller et al., 1995; Biederman et al., 1998; Sanchez et al., 1999, Geller and Luby, 1997; Carlson et al., 2000), as well as the potential earlier age of onset of bipolar disorders in some children compared with that in schizophrenia, have likely contributed to the slower initiation of early intervention trials in bipolar illness. Nonetheless, the results of this survey suggest that the majority of parents of children at high risk would be supportive of these efforts. Parents appear willing to consider and endorse a continuum of thresholds for intervention in children at high risk for an affective disorder. Such a viewpoint may also be a useful one for the academic community, as it may help circumvent some of the

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

controversies about the prodromal and atypical presentations of childhood bipolar illness and their similarities and discrepancies when compared to adult diagnostic criteria. In the unipolar affective disorders, a recognized continuum of increasing severity proceeds from minor depression and dysthymia to recurrent brief and major depression to melancholic and psychotic subtypes. Perhaps a similar concept could be developed for the childhoodonset bipolar-like disorders, such that successive thresholds of symptom-driven dysfunction can be reliably defined, recognized, and utilized. This would mitigate some of the controversies around the prevailing concept of a single cut-off for diagnosis and treatment. Some advocacy groups have recommended that patients participate in formulating clinical trial designs. We would hope that this survey on parental attitudes toward treatment provides a first step towards facilitating the development of early intervention studies in a unique sub-sample of children at high risk for an affective disorder by virtue of a uni-lineal or bi-lineal family history of bipolar illness. Knowledge of these positive parental attitudes toward early intervention may help stimulate further initiatives in this area. Crucial questions in need of systematic exploration in the future include: whether treatment effective in adults will ultimately prove to be effective in children; whether a good balance of efficacy and safety of an agent in children can be achieved; and whether early intervention will be able to ward off the development of more fullblown bipolar illness altogether.

References Anthony, J., Scott, P., 1960. Manic-depressive psychosis in childhood. Child Psychol. Psychiatry 4, 53–72. Biederman, J., Klein, R.D., Pine, D.F., 1998. Resolved: mania is mistaken for ADHD in prepubertal children. J. Am. Acad. Child Adolesc. Psychiatry 37, 1091–1099. Carlson, G.A., Bromet, E.J., Sievers, S., 2000. Phenomenology and outcome of subjects with early- and adult-onset psychotic mania. Am. J. Psychiatry 157, 213–219. Edwards, J., McGorry, P.D., Waddell, F.M., Harrigan, S.M., 1999. Enduring negative symptoms in first episode psychosis: comparison of six methods using follow up data. Schizophr. Res. 40, 147–158. Egeland, J.A., Blumenthal, R.L., Nee, J., Sharpe, L., Endicott, J.,

123

1987. Reliability and relationship of various ages of onset criteria for major affective disorder. J. Affect. Disord. 12, 159–165. Emslie, G.J., Kennard, B.D., Kowatch, R.A., 1995. Affective disorders in children: diagnosis and management. J. Child Neurol. 10 (Suppl. 1), S42–S49. Geller, B., Luby, J., 1997. Child and adolescent bipolar disorder: a review of the past 10 years. J. Am. Acad. Child Adolesc. Psychiatry 36, 1168–1176. Gershon, E.S., Hamovit, J., Guroff, J.J., Dibble, E., Leckman, J.F., Sceery, W., Targum, S.D., Nurnberger, J.I., Goldin, L.R., Bunney, Jr. W.E., 1982. A family study of schizoaffective, bipolar I, bipolar II, unipolar, and normal control probands. Arch. Gen. Psychiatry 39, 1157–1167. Gershon, E.S., Hamovit, J., Guroff, J.J., Nurnberger, J.I., 1987. Birth-cohort changes in manic and depressive disorder in relatives of bipolar and schizoaffective patients. Arch. Gen. Psychiatry 44, 314–319. Gillberg, I.C., Hellgren, L., Gillberg, C., 1993. Psychotic disorders diagnosed in adolescence. Outcome at age 30 years. J. Child. Psychol. Psychiatry 34, 1173–1185. Goodwin, F.K., Jamison, K.R., 1990. Manic-Depressive Illness. Oxford University Press, New York. Lapalme, M., Hodgins, S., LaRoche, C., 1997. Children of parents with bipolar disorder: a metaanalysis of risk for mental disorders. Can. J. Psychiatry 42, 623–631. Leverich, G.S., Nolen, W.A., Rush, A.J., McElroy, S.L., Keck, P.E., Jr., Denicoff, K.D., Suppes, T., Altshuler, L.L., Kupka, R., Kramlinger, K.G., Post, R.M., 2001. The Stanley Foundation Bipolar Treatment Outcome Network: I. Longitudinal methodology. J. Affect. Disord., in press. Lish, J.D., Dime-Meenan, S., Whybrow, P.C., Price, R.A., Hirschfeld, R.M.A., 1994. The National Depressive and ManicDepressive Association (NDMDA) survey of bipolar members. J. Affect. Disord. 31, 281–294. McGorry, P.D., 1998. ‘A stitch in time’ . . . the scope for preventive strategies in early psychosis. Eur. Arch. Psychiatry Clin. Neurosci. 248, 22–31. McInnis, M.G., McMahon, F.J., Chase, G.A., Simpson, S.G., Ross, C.A., De Paulo, J.R., 1993. Anticipation in bipolar affective disorder. Am. J. Hum. Genet. 53, 385–390. McMahon, F.J., Stine, O.C., Chase, G.A., Meyers, D.A., Simpson, S.G., De Paulo, J.R., 1994. Influence of clinical subtype, sex, and lineality on age of onset of major affective disorder in a family sample. Am. J. Psychiatry 151, 210–215. Papolos, D.F., Papolos, J., 2000. The Bipolar Child: The Definitive and Reassuring Guide To Childhood’s Most Misunderstood Disorder. Broadway Books, New York. Sanchez, L., Hagino, O., Weller, E., Weller, R., 1999. Bipolarity in children. Psychiatr. Clin. North Am. 22, 629–644. Strober, M., 1992. Relevance of early age-of-onset in genetic studies of bipolar affective disorder. J. Am. Acad. Child Adolesc. Psychiatry 31, 606–610. Strober, M., DeAntonio, M., Schmidt-Lackner, S., Freeman, R., Lampert, C., Diamond, J., 1998. Early childhood attention deficit hyperactivity disorder predicts poor response to acute lithium therapy in adolescent mania. J. Affect. Disord. 51, 145–151.

124

R.M. Post et al. / Journal of Affective Disorders 70 (2002) 117 – 124

Suppes, T., Leverich, G.S., Keck, P.E., Nolen, W., Denicoff, K.D., Altshuler, L.L., McElroy, S.L., Rush, A.J., Kupka, R., Frye, M.A., Bickel, M., Post, R.M., 2001. The Stanley Foundation Bipolar Treatment Outcome Network: II. Demographics and illness characteristics of the first 261 patients. J. Affect. Disord., in press. Weinberg, W.A., Brumback, R.P., 1976. Manic in childhood: case studies and literature review. Am. J. Dis. Child. 130, 380–385. Weissman, M.M., Fendrich, M.M., Warner, V., Wickramaratne, P., 1992. Incidence of psychiatric disorder in offspring at high and low risk for depression. J. Am. Acad. Child Adolesc. Psychiatry 31, 640–648. Weissman, M.M., Wolk, S., Wickramaratne, P., Goldstein, R.B.,

Adams, P., Greenwald, S., Ryan, N.D., Dahl, R.E., Steinberg, D., 1999. Children with prepubertal-onset major depressive disorder and anxiety grown up. Arch. Gen. Psychiatry 56, 794–801. Weller, E.B., Weller, R.A., Fristad, M.A., 1995. Bipolar disorder in children: misdiagnosis, underdiagnosis, and future directions. J. Am. Acad. Child Adolesc. Psychiatry 34, 709–714. Werry, J.S., 1992. Child psychiatric disorders: are they classifiable? Br. J. Psychiatry 162, 472–480. Werry, J.S., 1993. Long-term drug use in psychiatric disorders in children: facts, controversies and the future. Acta Paedopsychiatr. 56, 113–118.