Reliability of the Washington University in St. Louis Kiddie Schedule for Affective Disorders and Schizophrenia (WASH-U-KSADS) Mania and Rapid Cycling Sections

BARBARA GELLER, M.D., BETSY ZIMERMAN, M.A., MARLENE WILLIAMS, R.N., KRISTINE BOLHOFNER, B.S., JAMES L. CRANEY, M.P.H., MELISSA P. DELBELLO, M.D., AND CESAR SOUTULLO, M.D.

Reprinted with permission from J Am Acad Child Adolesc Psychiatry, Volume 40, p. 450-455 (2001).

ABSTRACT

Objective: To investigate the reliability of the Washington University in St. Louis Kiddie Schedule for Affective Disorders and Schizophrenia (WASH-U-KSADS) mania and rapid cycling sections. Method: The 1986 version of the KSADS was modified and expanded to include onset and offset of each symptom for both current and lifetime episodes, expanded prepubertal mania and rapid cycling sections, and categories for attention-deficit/hyperactivity disorder and other DSM-IV diagnoses. To optimize diagnostic research, skip-outs were minimized. Subjects participated in the ongoing “Phenomenology and Course of Pediatric Bipolar Disorder” study. Mothers and children were interviewed separately by research nurses who were blind to diagnostic group status. In addition, ratings of off-site child psychiatrists, made from the narrative documentation given for each WASH-U-KSADS item, were compared with research nurse ratings. This work was performed between 1995 and 2000. Results: There was 100% interrater reliability, five consecutive times, as both interviewer and observer after 10 to 15 trials. The k values of comparisons between research nurse and off-site blind best-estimate ratings of mania and rapid cycling sections were excellent (0.741.00). High 6-month stability for mania diagnoses (85.7%) and for individual mania items and validity against parental and teacher reports were previously reported. Conclusions: The WASH-UKSADS mania and rapid cycling sections have acceptable reliability. J. Am. Acad. Child Adolesc. Psychiatry, 2001, 40(4):450455. Key Words: child, mania, rapid cycling, reliability, rating.

This is a brief report on the reliability of the Washington University in St. Louis Kiddie Schedule for Affective Disorders and Schizophrenia (WASH-U-KSADS) (Geller et al., 1996, 1998c) mania and rapid cycling sections.

Accepted October 31, 2000. From the Department of Psychiatry, Washington University School of Medicine, St. Louis (Dr. Geller, Ms. Zimerman, Ms. Williams, Ms. Bolhofner, Mr. Craney); and the Department of Psychiatry, University of Cincinnati College of Medicine, Cincinnati (Drs. DelBello and Soutullo). Supported by NIMH grant R01 MH-53063 (to Dr. Geller). Correspondence to Dr. Geller, Washington University School of Medicine, 660 South Euclid Avenue, Box 8134, St. Louis, MO 63110; e-mail: gellerb@medicine.wustl.edu. 0890-8567/01/4004-0450q2001 by the American Academy of Child and Adolescent Psychiatry.The background of the KSADS instruments is informative on the rationale for research interviews of children about themselves. Until the late 1970s, diagnostic assessment of children and adolescents was done largely through interviews with parents about children and observation of children in play therapy type of settings (Puig-Antich, 1980). Puig-Antich and Chambers (1978) developed the first semistructured interviews designed to directly assess children about themselves that conformed to Research Diagnostic Criteria (RDC) (Endicott and Spitzer, 1979) and later to DSM-III (American Psychiatric Association, 1980) criteria. The RDC were a precursor of the DSM-III. A semistructured interview is one in which there are no standardized specific questions; rather, there are areas to cover using the rater's judgment on how best to do so with each individual. The KSADS (Chambers et al., 1985; Puig-Antich and Chambers, 1978) instrument was modified from the adult version, the SADS (Endicott and Spitzer, 1979), by rewording items to be more relevant to the pediatric age group and by adding useful age-dependent knowledge, e.g., age-dependent normative sleep duration. The KSADS (Chambers et al., 1985; Puig-Antich and Chambers, 1978), like its adult precursor, was meant to be administered by trained clinicians who had graduate degrees in fields related to psychiatry and who had postgraduate clinical experience. Unlike the adult version, the KSADS-1978 had separate instruments for current episodes (Chambers et al., 1985; Puig- Antich and Chambers, 1978) and for lifetime episodes (Puig-Antich et al., 1980). The latter was specifically meant for use in epidemiological studies.

The KSADS became widely used in assessment, treatment, and neurobiological studies of prepubertal major depressive disorders (PMDD) (e.g., Emslie et al., 1997; Geller et al., 1992). The KSADS had nondichotomous ratings that permitted assessing levels of severity. The latter property was very important for use in measuring changes over time, which is needed for treatment studies (Emslie et al., 1997; Geller et al., 1992). In addition, the ability to use similar instrumentation and to implement similar DSM diagnostic criteria for both child-onset and adult-onset major depressive disorder (MDD) facilitated investigation of continuities across the life span.

Recently Ambrosini (2000) reviewed psychometric properties of older versions of the KSADS, including one developed by Kaufman et al. (1997). Kaufman et al. (1997) reported on a more recent version of the KSADS, called the KSADS Lifetime and Present (KSADS-PL). The latter tool revised the original KSADS by including both current and past episodes in the same instrument (similar to the WASH-U-KSADS), by using trichotomous categories rather than multiple levels of severity, and by extensive skip-outs to lessen respondent burden.

In contrast to the extensive KSADS items on PMDD, less attention was given to the category of mania, largely because prepubertal onset of mania was not believed to occur (Geller and Luby, 1997).

Beginning in the 1990s, there was controversy about the existence of a prepubertal and early adolescent bipolar disorder phenotype (PEA-BP) (Geller and Luby, 1997; Geller et al., 1995, 1998d, 2000a, 2001). Much of the controversy centered on the observation, reported by a number of investigators, that comorbid attention-deficit/ hyperactivity disorder (ADHD) is highly prevalent in PEA-BP (Biederman et al., 1995; Fristad et al., 1992; Geller et al., 1995, 1998d, 2000b). There were data to support both that comorbid ADHD is an age-dependent manifestation of child mania (Geller et al., 1995, 1998c,d, 2000b) and that a subgroup of cases had a familial diathesis for both child mania and ADHD (Faraone et al., 1997). Regardless of the etiopathogenesis of the comorbid ADHD, there was a need to assess specific criteria of mania that did not overlap with ADHD or other disruptive disorders in children (Fristad et al., 1992; Geller et al., 1998d, 2000b, 2001). Fristad et al. (1992) were the first to suggest this in a seminal paper in which they showed significant differences on mania items from the Young Mania Rating Scale (YMRS) (Young et al., 1978) in children with clinical diagnoses of mania with comorbid ADHD versus those with clinical ADHD without comorbid bipolar disorder (BP). The YMRS, however, was not designed for use in pediatric subjects.

Thus there was a need for a version of the KSADS that was developed specifically to target the assessment of prepubertal mania and hypomania and to assess the pattern of rapid cycling. Rapid cycling was important to assess because of the reported higher prevalence in child- versus adult-onset mania (Geller et al., 1995, 1998c, 2000a; Geller and Cook, 2000; Goodwin and Jamison, 1990).

A prior publication reported the validity of the WASH-U-KSADS against parental reports on the Child Behavior Checklist (CBCL) and against teacher ratings on the Teacher's Report Form (TRF) (Geller et al., 1998c). Also, high 6-month stability of both mania and hypomania diagnoses (85.7%) and of individual mania and rapid cycling items has been reported (Geller et al., 2000a). This communication supplements these aspects of the WASH-U-KSADS with interrater reliability data.

METHOD

Characteristics of the WASH-U-KSADS

Prior to the start of the "Phenomenology and Course of Pediatric Bipolar Disorder" project, the KSADS-1986 (Puig-Antich and Ryan, 1986) was modified in the following ways. Items for specific onsets and offsets were added to each symptom item so that the duration of both current and lifetime symptoms and episodes could be documented. The mania section was expanded to include more extensive areas to cover in assessing each mania symptom. For example, separate areas to cover when interviewing parents versus children about hypersexuality were added (Geller et al., 1998c).

Furthermore, on the basis of pilot data for the "Phenomenology" project (Geller et al., 1995), an extended section to document rapid cycling phenomena was added. The entire rapid cycling section, including all items in this area, is published (Geller et al., 1998d). The definition of rapid cycling was four or more episodes per year. Ultrarapid cycling was defined as between 5 and 364 and ultradian as 365 or more episodes per year. In ultradian cycling, mania needed to occur for 4 or more hours per day (Geller and Cook, 2000; Geller et al., 2001). Overlapping time periods on the WASH-U-KSADS items of mania/ hypomania and MDD were used as the definition of "mixed mania."

The levels of severity for each item on the KSADS-1986 (Puig- Antich and Chambers, 1978; Puig-Antich and Ryan, 1986) (none = 1, doubtful = 2, mild = 3, moderate = 4, severe = 5 and 6) were kept. Categories to cover ADHD and other DSM-IV diagnoses were also added. Substance use sections developed for use in substance dependency protocols for pediatric subjects were appended (Geller et al., 1998a,b). The latter considerably expanded the substance use section of the KSADS-1986.

In addition, because of the planned use in a phenomenology protocol, skip-outs were minimized. In other versions of the KSADS, there are screening questions to decrease interviewer time. If these screening questions were negative, the interviewer was instructed to skip to the next section. Because of the need to maximize phenomenology data, skip-outs were avoided.

The research nurses who performed the ratings derived a summary score from either mother or child responses, in accordance with published methods (Bird et al., 1992). In this method, separate responses from interviews of mothers and from interviews of children are amalgamated by using data given by either informant. Bird et al. (1992) reported that this method is as useful as attempts to ignore data from either informant. Mania needed ratings of ³4 and hypomania needed ratings of ³3. Establishing time frames for children's ratings was done by using birthdays, holidays (e.g., Was item present at Christmas?), start of school, and end of school as anchor points (e.g., Was item present when school term began?).

Participants

The study population for the "Phenomenology" project has been described in detail elsewhere (Geller et al., 1998c,d, 2000a,b).

Staggered entry occurred over 3.5 years between 1995 and 1999. Because the planned inpatient sites at the Washington University School of Medicine in St. Louis complex were permanently closed near the start of the study, all subjects were outpatients.

There were 93 PEA-BP (with or without comorbid ADHD), 81 ADHD (without BP or MDD), and 94 community comparison subjects (without BP, MDD, or ADHD) who participated in the ongoing "Phenomenology" study. Cases in the PEA-BP and ADHD groups were outpatients obtained by consecutive new case ascertainment to optimize generalizability, and community comparison subjects were from a survey conducted by Research Triangle Institute.

The PEA-BP phenotype required a current episode of mania or hypomania with elation and/or grandiosity as one criterion. Elated mood and/or grandiosity were included in the study design for the following reasons: (1) to diagnose DSM-IV mania or hypomania without using just criteria that overlapped with those for ADHD; (2) to obtain a study sample of PEA-BP subjects that had at least one of the two cardinal features of mania (i.e., elation and/or grandiosity); and (3) to study a child BP phenotype that was most likely to be continuous with adult BP.

To establish DSM-IV diagnoses, PEA-BP subjects needed to fit DSM-IV criteria for current mania or hypomania and to be definite cases by severity ratings on the Children's Global Assessment Scale (CGAS) (Bird et al., 1987; Shaffer et al.,1983). The CGAS was completed by raters who administered the WASH-U-KSADS. On the CGAS, a score of 0 is worst, 100 is best, and <=60 is a definite case. Final diagnoses were made in consensus conferences that included a review of all research instruments, teacher and school reports, videotapes of WASH-U-KSADS mother interviews, videotapes of WASH-UKSADS child interviews, and medical records.

Procedures for Interrater Reliability

To establish interrater reliability, we used procedures that replicated those implemented in establishing interrater reliability for studies of PMDD that used earlier versions of the KSADS (e.g., Geller et al., 1992). The standard for established interrater reliability was 100% agreement on DSM-IV diagnoses and individual criteria five consecutive times as observer and as interviewer. Raters alternated as interviewer and observer and compared and discussed ratings after each interview to resolve disagreements. Using ratings between observers and interviewers replicated methods for establishing interrater reliability on depression items in earlier versions of the KSADS (Chambers et al., 1985; Geller et al., 1992). These ratings were then discussed in consensus conferences with one of the authors (B.G.) to be sure that the narrative documentation fit the severity rating. The data collection guideline is that the narrative documentation must justify the rating with respect to onset, offset, frequency, duration, intensity, and specific examples. Thus, the narrative next to each WASH-U-KSADS item is part of using this assessment tool (e.g., part of the narrative next to a suicidal ideation item read "cut her wrists four times with a kitchen knife and wanted to die to escape her sad feelings"). Additional examples of parts of the narrative documentation were a child related feeling "high, off the charts high" before "crashing right down." The latter would be part of the documentation of a euphoric state. The elation followed by "crashing" to a state of despair was part of the narrative for cycling.

Reliability data between research nurses and best estimators were obtained from a randomly selected subsample of subjects (n = 60). For each subject, both research nurses and off-site best estimators rated using the separate parent and child interviews, as described above. To perform best estimates, off-site child psychiatrists (M.P.D., C.S.) used the narrative documentation, as described above, for each rating of WASH-U-KSADS items. These 60 cases were randomly selected from both unrecovered and recovered PEA-BP subjects at the 6-month and 1-year assessment points. Best estimators were blind to recovered versus unrecovered status and to any other information about the subjects.

Data Analyses

Comparisons between research nurse and best-estimate ratings were made using percent agreement for the diagnostic variables, which were dichotomous. For the individual WASH-U-KSADS items, ratings were assigned to three categories of severity: no pathology or doubtful (WASH-U-KSADS item rating 1 or 2); mild pathology (WASH-UKSADS item rating = 3); and moderate or marked pathology (WASHU- KSADS item rating 4 or greater). The r statistic was used to measure correlation between research nurse and best-estimate ratings for these three levels of severity. For both diagnoses and individual WASH-U-KSADS items, the k statistic was also used. The k statistic is used to estimate agreement beyond that expected by chance.

RESULTS

Raters continued to rate, alternating interviewing and observing separate interviews of mothers and children until there was 100% agreement on symptoms and diagnoses five consecutive times. The number of trials prior to 100% agreement five consecutive times was between 10 and 15 trials. Five of these 10 to 15 trials were the five in a row needed to have five consecutive interviews with full agreement. On the first 5 to 10 trials, there were usually one or two mania items that lacked agreement, and usually four or five total items without agreement on the entire WASH-U-KSADS instrument. Disagreements were usually only one number apart (e.g., one rater gave a 2, the other rater gave a 3).

High interrater reliability was similarly achieved on mania and rapid cycling section items for all subjects (93 GELLER ET AL. 452 J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 40:4, APRIL 2001 PEA-BP, 81 ADHD, and 94 community comparison). Because of the small number of positive mania items (Geller et al., 1998c) in ADHD and community comparison cases, off-site best estimation was not warranted.

Subjects in the PEA-BP group needed to have current mania or hypomania to enter the study, but they could also have had past episodes. Similar interrater reliability was obtained for both current and past episodes of mania or hypomania.

As noted above, only a random subset (n = 60) of the PEA-BP subjects (n = 93) were used for off-site blind estimation. Table 1 presents the comparisons of research nurse ratings with those of the best estimators. As noted in Table 1, k statistics for WASH-U-KSADS items were calculated using three levels. Level 1 was a rating of 1 or 2, indicating no or doubtful psychopathology. Level 2 was a rating of 3, indicating mild psychopathology. Level 3 was a rating of 4 or greater, indicating moderate or severe psychopathology. The k values ranged from 0.82 to 1.00, within the range of excellent agreement. In addition, the k statistic for grandiosity was calculated at four levels because a rating of 6 on the grandiosity item indicates grandiose delusions. The k for grandiosity using four levels was 0.74. The r for the separate WASH-U-KSADS item for grandiose delusions was 1.00 and the k was 1.00.

As reported elsewhere (Geller et al., 2000b), 87.1% of the 93 PEA-BP cases had rapid cycling and 77.4% had ultradian cycling. Therefore, 88.9% of all rapid cycling was ultradian rapid cycling as defined above and elsewhere (Geller and Cook, 2000; Geller et al., 2001). Thus, the statistics on rapid cycling in Table 1 essentially refer to ultradian cycling.

DISCUSSION

These data support that the WASH-U-KSADS (Geller et al., 1996, 1998c) version of the KSADS has acceptable reliability. Although validity against the CBCL and TRF has been reported (Geller et al., 1998c), there are limitations to CBCL and TRF comparisons. On the one hand, finding similar patterns on the CBCL to those reported by Biederman et al. (1995) is informative to the field. Also informative is the finding of similar TRF profiles to that on the CBCL (Geller et al., 1998c). Furthermore, the 6- month stability (Geller et al., 2000a) of the PEA-BP study population supports the short-term validation of child mania. Nevertheless, additional validation by longer longitudinal follow-up and by a wide array of neurobiological, molecular and family genetic, treatment, and other etiopathogenetic investigations will be necessary to characterize the phenotypic specificity of PEA-BP and to determine its continuity with adult-onset BP.

TABLE 1
Comparison of Research Nurse and Off-Site Best-Estimate Ratings of WASH-U-KSADS Mania and Rapid Cycling SectionsDSM-IV Diagnosis% AgreementkMania95.000.90Hypomania95.000.85WASH-U-KSADS ItemsrkaElated moodb0.980.92Grandiosityb0.960.82cGrandiose delusions1.001.00Flight of ideas and/or racing thoughtsb1.000.95Flight of ideas1.000.93Racing thoughts1.001.00Decreased need for sleepb1.001.00Poor judgmentb0.980.95Hypersexuality0.950.95Daredevil acts0.991.00Silliness, laughing1.001.00Uninhibited people-seeking0.970.90Irritable moodb0.960.88Accelerated speechb0.990.95Distractibilityb0.950.91Increased energyb0.980.97Hyperenergetic0.970.94Increased productivity0.980.89Sharpened thinking1.001.00Increased goal-directed activity0.980.84Increased motor agitation0.990.97Suicidal ideation1.001.00WASH-U-KSADS Item% AgreementkaRapid cycling95.000.86Note: WASH-U-KSADS = Washington University in St. Louis Kiddie Schedule for Affective Disorders and Schizophrenia. ^a The k values for WASH-U-KSADS items were calculated using three levels: level 1 was a rating of 1 or 2 that indicated no or doubtful pathology; level 2 was a rating of 3, which indicated mild pathology; level 3 was a rating of 4 or greater, indicating moderate or severe pathology. ^b DSM-IV criterion. ^c The k value for grandiosity was also calculated using four levels. Level 4 was a rating of 6, indicating grandiose delusions. The k statistic using four levels for grandiosity was 0.74. The k for the separate WASH-U-KSADS item for grandiose delusions was 1.00.

The requirement for administration by clinically experienced raters with psychiatrically relevant graduate degrees, and the time burden to respondents, make the WASH-UKSADS most useful for clinical research and least useful for epidemiological studies. The latter require rating scales that can be administered by lay raters with minimal training and require lower respondent time burden.

A comparison of the WASH-U-KSADS to the KSADS-1986 and the KSADS-PL appears elsewhere (Geller et al., 1998c). Essentially, the WASH-U-KSADS is distinguished from the other two instruments by inclusion of extensive prepubertal mania and rapid cycling items and by inclusion of specific items to document both current and lifetime onsets and offsets of each symptom. The WASH-U-KSADS retained the original KSADS structure of multiple levels of severity for each item. The epidemiological version of the KSADS, called the KSADS-E (Puig- Antich et al., 1980), is similar to the KSADS-1986 except that it uses dichotomous items instead of levels of severity and covers lifetime episodes. Thus, it differs from the WASH-U-KSADS by its lack of extensive mania and rapid cycling items, lack of multiple levels of severity, and lack of specific onsets and offsets for each symptom.

The excellent k values on the mania items (0.741.00) were higher than those reported by Chambers et al. (1985) for the KSADS depression items. This may be partly methodological, but is more likely because many years of experience with the KSADS, prior to developing the WASH-U-KSADS mania and rapid cycling sections, allowed for more precise item structure.

Prior to the advent of the KSADS, in the early 1970s, the existence of PMDD was controversial (Puig-Antich, 1980). A concept that children had "masked" depressions evidenced by nondepressive symptoms was prevalent (Cytryn et al., 1980). These "masked" symptoms included irritability, hyperactivity, poor concentration, and antisocial behaviors. Thus, having the appropriate tool to assess the cardinal features of PMDD, i.e., dysphoric mood and anhedonia, permitted the field to advance from diagnosing PMDD by nonspecific "masked" items to diagnosing PMDD with criteria that characterize MDD across the life span (e.g., sad mood, anhedonia). Thus, in children, there was a progression from viewing PMDD as "masked depression" to a consensus in the field on the existence of DSM-III and higher MDD. The latter occurred as assessment tools and validating studies became available. It remains to be seen whether a similar progression from skepticism to consensus will occur on the existence of DSM-IV PEA-BP as validating investigations are conducted.

Limitations

As described previously (Geller et al., 1998c, 2000a,b), subjects consecutively ascertained for the "Phenomenology" project had a mean socioeconomic status (SES) (Hollingshead, 1976) in the second highest of five classes. Thus these results may not generalize to lower SES groups. Also, the phenotype investigated required elated mood and/or grandiosity as one criterion. Therefore, these findings may not generalize to other definitions of child mania. These KSADS-type research methods were designed for phenomenological and validating studies. They are not feasible in clinical settings as discussed in the introduction. Usefulness for systematic research, however, is evidenced by the fact that six of the nine NIMH-funded studies on child BP (as of September 1, 2000) use the WASH-UKSADS mania and rapid cycling sections. These investigators and multiple other research groups have trained in St. Louis to replicate the WASH-U-KSADS methods in their own work.

Clinical Implications

Because the WASH-U-KSADS is a research tool, this work is not immediately relevant to clinical practice. Characterization of the PEA-BP phenotype is, however, of timely clinical importance because reliable instruments indirectly contribute to patient care. In this regard, similar to other controversies in clinical medicine, the one on child mania will be addressed best by rigorous studies that use reliable methods.

REFERENCES

Ambrosini PJ (2000), Historical development and present status of the Schedule for Affective Disorders and Schizophrenia for School-Age Children (K-SADS). J Am Acad Child Adolesc Psychiatry 39:4958

American Psychiatric Association (1980), Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (DSM-III). Washington, DC: American Psychiatric Association

Biederman J, Wozniak J, Kiely K et al. (1995), CBCL clinical scales discriminate prepubertal children with structured interviewderived diagnosis of mania from those with ADHD. J Am Acad Child Adolesc Psychiatry 34:464471

Bird HR, Canino G, Rubio-Stipec M, Ribera JC (1987), Further measures of the psychometric properties of the Children's Global Assessment Scale. Arch Gen Psychiatry 44:821824

Bird HR, Gould MS, Staghezza B (1992), Aggregating data from multiple informants in child psychiatry epidemiological research. J Am Acad Child Adolesc Psychiatry 31:7885

Chambers WJ, Puig-Antich J, Hirsch M et al. (1985), The assessment of affective disorders in children and adolescents by semi-structured interview: test-retest reliability of the Schedule for Affective Disorders and Schizophrenia for School-Age Children, Present Episode version. Arch Gen Psychiatry 42:696702

Cytryn L, McKnew DH Jr, Bunney WE Jr (1980), Diagnosis of depression in children: a reassessment. Am J Psychiatry 137:2225

Emslie GJ, Rush AJ, Weinberg WA et al. (1997), Double-blind, randomized, placebo-controlled trial of fluoxetine in children and adolescents with depression. Arch Gen Psychiatry 54:10311037

Endicott J, Spitzer RL (1979), Use of the Research Diagnostic Criteria and the Schedule for Affective Disorders and Schizophrenia to study affective disorders. Am J Psychiatry 136:5256

Faraone SV, Biederman J, Mennin D, Wozniak J, Spencer T (1997), Attentiondeficit hyperactivity disorder with bipolar disorder: a familial subtype? J Am Acad Child Adolesc Psychiatry 36:13781387

Fristad MA, Weller EB, Weller RA (1992), The Mania Rating Scale: can it be used in children? A preliminary report. J Am Acad Child Adolesc Psychiatry 31:252257

Geller B, Cook EH (2000), Ultradian rapid cycling in prepubertal and early adolescent bipolarity is not in transmission disequilibrium with Val/Met COMT alleles. Biol Psychiatry 47:605609

Geller B, Cooper TB, Graham DL, Fetner HH, Marsteller FA, Wells JM (1992), Pharmacokinetically designed double-blind placebo-controlled study of nortriptyline in 6- to 12-year-olds with major depressive disorder. J Am Acad Child Adolesc Psychiatry 31:3444

Geller B, Cooper TB, Sun K et al. (1998a), Double-blind placebo-controlled study of lithium for adolescent bipolar disorders with secondary substance dependency. J Am Acad Child Adolesc Psychiatry 37:171178

Geller B, Cooper TB, Zimerman B et al. (1998b), Lithium for prepubertal depressed children with family history predictors of future bipolarity: a double-blind, placebo-controlled study. J Affect Disord 51:165175

Geller B, Craney JL, Bolhofner K, DelBello MP, Williams M, Zimerman B (2001), One year recovery and relapse rates of children with a prepubertal and early adolescent bipolar disorder phenotype. Am J Psychiatry 158:303305

Geller B, Luby J (1997), Child and adolescent bipolar disorder: a review of the past 10 years. J Am Acad Child Adolesc Psychiatry 36:11681176

Geller B, Sun K, Zimerman B, Luby J, Frazier J, Williams M (1995), Complex and rapid cycling in bipolar children and adolescents. J Affect Disord 34:259268

Geller B, Warner K, Williams M, Zimerman B (1998c), Prepubertal and young adolescent bipolarity versus ADHD: assessment and validity using the WASH-U-KSADS, CBCL, and TRF. J Affect Disord 51:93100

Geller B, Williams M, Zimerman B, Frazier J (1996), Washington University in St. Louis Kiddie Schedule for Affective Disorders and Schizophrenia (WASH-U-KSADS). St Louis: Washington University

Geller B, Williams M, Zimerman B, Frazier J, Beringer L, Warner K (1998d), Prepubertal and early adolescent bipolarity differentiate from ADHD by manic symptoms, grandiose delusions, ultra-rapid or ultradian cycling. J Affect Disord 51:8191

Geller B, Zimerman B, Williams M et al. (2000a), Six-month stability of a prepubertal and early adolescent bipolar disorder phenotype. J Child Adolesc Psychopharmacol 10:165173

Geller B, Zimerman B, Williams M et al. (2000b), Diagnostic characteristics of 93 cases of a prepubertal and early adolescent bipolar disorder phenotype by gender, puberty and comorbid ADHD. J Child Adolesc Psychopharmacol 10:157164

Goodwin FK, Jamison KR, eds (1990), Manic-Depressive Illness. New York: Oxford University Press, pp 2255,129154, 650

Hollingshead AB (1976), Four Factor Index of Social Status. New Haven, CT: Yale University Department of Sociology

Kaufman J, Birmaher B, Brent D et al. (1997), Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): initial reliability and validity data. J Am Acad Child Adolesc Psychiatry 36:980988

Puig-Antich J (1980), Affective disorders in childhood: a review and perspective. Psychiatr Clin North Am 3:403423

Puig-Antich J, Chambers W (1978), The Schedule for Affective Disorders and Schizophrenia for School-Age Children (Kiddie-SADS). New York: New York State Psychiatric Institute

Puig-Antich J, Orvaschel H, Tabrizi MA, Chambers W (1980), The Schedule for Affective Disorders and Schizophrenia for School-Age Children-Epidemiologic Version. New York: New York State Psychiatric Institution and Yale University School of Medicine

Puig-Antich J, Ryan N (1986), The Schedule for Affective Disorders and Schizophrenia for School-Age Children (Kiddie-SADS)-1986. Western Psychiatric Institute and Clinic, 3811 O'Hara Street, Pittsburgh, PA 14213

Shaffer D, Gould MS, Brasic J et al. (1983), A children's global assessment scale (C-GAS). Arch Gen Psychiatry 40:12281231

Young RC, Biggs JT, Ziegler VE, Meyer DA (1978), A rating scale for mania: reliability, validity and sensitivity. Br J Psychiatry 133:429435