|
QUEST -
Childhood
AUTHOR
David Novosad
Chairman’s Rounds, ?
Citation:
Poulton R,
Caspi A, Moffitt TE, et al. Children’s self reported psychotic
symptoms and adult schizophreniform disorder: A 15-year longitudinal
study. Arch Gen Psychiatry 2000; 57: 1053-1058.
Clinical
question:
Are there childhood risk factors for the development of adult
schizophrenia?
Study Design
Type:
Prospective
cohort
Methods:
Patient population:
A birth cohort
born between April 1, 1972 and March 31, 1973 in Dunedin, New
Zealand. 91% (1037) of births participated at age 3. Data used
from age 11 and age 26 assessments. 789 births had psychiatric
interviews at age 11. 972 had psychiatric interviews at age 26.
761 were psychiatrically assessed at both ages. Those included vs.
omitted (assessed at age 26 and not age 11) did not differ
significantly in psychiatric diagnosis. Study members assessed via
DISC-C for DSM-III at age 11. 5 questions from schizophrenia
section scored from 0-2. Children placed into 3 groups (no
symptoms, weak symptoms, and strong symptoms). Study members
assessed via DIS at age 26.
Outcomes: Schizophreniform disorder. Eliminated symptoms
with plausible explanations and related to alcohol or drug use or a
major depressive episode. For diagnosis, study members must have 1
hallucination symptoms and 2 other symptoms reported as “yes,
definitely” from criterion A of DSM. For diagnosis, study members
must also have evidence of impairment in at least one area of social
or occupational functioning (long term unemployment, poor money
management skills, not in a relationship, paranoia, social
isolation, poor personal grooming.) These symptoms must occur for
at least one month. Several of these symptoms were determined by
informants and observers as well as self-report.
Analysis: Evaluated relationship with Mantel-Haenszel x2
test and logistic regression equations. Equations contained dummy
variables representing both strong and weak symptom groups. Control
group was children at age 11 with no symptoms. Results reported
with ORs and 95% CIs.
Validity
Criteria:
Patient sample
was clearly defined. The 3 age 11 risk groups did not differ by
parental SES or on WIS for Children-Revised Full Scale IQ test.
Prevalence of mental disorders in Dunedin sample matched prevalence
in US National Comorbidity Survey. All patients were accounted for
and 95% of patients alive at age 26 were assessed. Outcome criteria
were objective and unbiased and explained in a detailed and
transparent manner. Sex, social class origins, and age-11 IQ scores
were controlled for in the analysis of the cohort.
Main Results:
Weak
symptom group was more likely to meet criteria for adult
schizophreniform disorder (OR, 5.1; 95% CI, 1.7-18.3). Strong
symptom group was even more likely to meet criteria for diagnosis
(OR, 16.4; 95% CI, 3.9-67.8). Odds ratios were unchanged after
controlling for sex, social class origins, and age-11 IQ scores.
See Table 1. Strong symptom group members were not more likely to
have a diagnosis of mania or depressive disorder. Both weak and
strong symptom groups were more likely than no symptom children to
develop an anxiety disorder, although less likely than for
schizophreniform disorder. See Table 2. No children in the cohort
had a diagnosis of childhood schizophrenia. One half of strong
symptom group had no psychiatric diagnosis at age 11 and those
without psychiatric diagnosis were just as likely to report
delusions or hallucinations at age 26.
Conclusions:
This
prospective cohort study demonstrates an association between
childhood psychotic symptoms and adult schizophreniform disorder.
The strengths of the study included prospective cohort design with
all patients accounted for and transparent reporting of methods and
results. Weaknesses included those inherent to the study design.
Other weaknesses are the sample has not yet passed through the
entire risk period for psychosis, not all interviews at age 26 were
conducted by a psychiatrist, psychotic symptoms were also associated
with anxiety disorders, and findings of this study are based on
small groups.
Return
to top of the page
AUTHOR
Doug Burgess
SSRIs and routine specialist care with and without cognitive
behavioral therapy in adolescents
Chairman’s Rounds, August 13, 2007
Goodyer I, et al. Selective serotonin reuptake inhibitors (SSRIs
and routine specialist care with and without cognitive behavioral
therapy in adolescents with major depression: randomized controlled
trial. BMJ, 2007;335:142
Is the combination of SSRI and CBT superior to SSRI alone in
improving the general functioning of adolescents with moderate to
severe major depression?
Depression in adolescence is a serious disorder that is both
prevalent and carries relatively high risk for suicide and the
development of other psychiatric disorders later in life. Results
from the treatment of adolescent depression study (TADS) indicated
that fluoxetine in conjunction with CBT was superior to fluoxetine
alone in treating adolescent depression and possibly preventing
suicide. Subsequent studies have failed to reproduce these results
raising questions about the study’s validity.
Therapy Validity
Criteria:
Follow up was
complete and all patients who entered the trial were accounted for
at its completion. Patients were randomized and the randomization
was concealed from the evaluators (although not the physicians).
The groups were similar and all participants were analyzed within
the groups they were initially assigned. The study was not blinded
to patients or physicians although assessors were not aware of group
allocation. With the exception of CBT participation, groups were
treated equally.
Randomized
Controlled Trial criteria for appropriate design:
The trial was funded by NS Health Technology Assessment,
University hospitals and a non-profit mental health trust. Although
physicians and participants of the trial were aware of the groups
they had been assigned to, evaluators remained blinded. Patients
were followed up at 12 weeks and 28 weeks. They were chosen from
six specialist outpatient psychiatry clinics and referred to 2 major
University clinics in England. In all, 510 patients between the
ages of 11 and 17 who scored 7 or higher on the Health of the Nation
outcome scales were recruited. (Validated scale that indicates
moderate to severe depression). Exclusion criteria included
schizophrenia, bipolar disorder, pregnancy, need for immediate
admission and previous SSRI+CBT with no effect. Suicidal ideation
was not an exclusion criterion. All told, 103 patients were
randomized to SSRI alone and 105 were randomized to SSRI +CBT. The
Health of the Nation outcome scale was the primary outcome measure.
Secondary outcomes were the participant rated mood and feelings
questionnaire, the observer rated revised children’s depression
rating scale, the children’s global assessment scale and the
clinical global impression improvement scale. 191/211 participants
who entered the study remained for the duration.
Main results:
Mean
values for the two treatments were similar at corresponding
assessments. For the primary outcome measure (Health of the Nation)
the treatment effect measured across follow up points was 0.001
(-1.52 to 1.52, P=0.99). There was also no evidence of an
interaction between baseline severity and treatment for any outcome
measure. At 28 weeks, 57/94 (61%) in the SSRI group and 52/98 (53%)
of the SSRI+CBT were much or very much improved. 62% of patients in
the SSRI+CBT group reported adverse events while only 59% of the
SSRI group reported side effects.
Conclusions:
This study indicates that in adolescents with moderate to severe
major depression, treatment with an SSRI combined with CBT does not
offer significant benefit over treatment with SSRI alone. This
conclusion stands in contrast to the conclusion of TADS, which led
to the recommendation that children be started on an SSRI only if
they also initiate CBT. While the study attempted to replicate
“real world” clinical settings, the degree of follow up SSRI only
participants received was more intense than that offered in most
clinical settings. The similar response rates between the two
groups could underscore the importance of close follow up in the
treatment of depression. This study can not be used as evidence to
support the efficacy of SSRI efficacy as no placebo was utilized for
control.
Return
to top of the page
AUTHOR
Leigh Fylstra
Adult Outcome for
Children With Autism
Chairman’s
Rounds, March 19, 2007
Citation:
Howlin P, Goode S, Hutton J, and Rutter M. 2004. Adult outcome
for children with autism. Journal of Child Psychology and
Psychiatry 45: 212-229.
Clinical
Question:
What
factors influence long-term outcomes in individuals with autism with
an IQ above 50?
Background:
studies in pt’s with autism have found inconclusive evidence
concerning prognostic factors in long term outcomes, which is a
major concern for parents. Studies show poorest outcomes with IQs
below 50, but anyone with IQ below 50 rarely achieves independence.
This study investigates the long-term outcomes in pt’s with autism
with IQs greater than 50.
Validity Criteria for Prognostic study
The sample size was sufficiently powered. Pts were
generally representative of clinical practice, except this
population was potentially more severely handicapped, as the
diagnostic criteria were more strict in 1950s-60s (n =47). Patients
were relatively homogenous with respect to prognostic risk, although
they likely differed in parental support and availability of
educational programs (as intensive educational programs not
available in 50-60s). The paper itself suggests that studying pt’s
with autism is difficult d/t heterogeneity. Patients were not
necessarily at a similar point in disease progression or severity,
as an example, ADI scores differed with respect to autistic-like
behaviors: 12% no problems, 42& mild, 35% moderate, 11% severe.
Follow-up was complete and all pt’s accounted for: 79 initially
evaluated, only 68 available for f/u- 1 death, 2 refused, 6
untraceable, 2 admin error. It is unlikely that pt’s were treated
equally from initial evaluation to f/u, as factors were not
controlled. The major outcome criteria was a score on “overall
social outcome” scale created by the investigators; however, this is
problematic, as the scale is neither validated, nor standardized.
Study Design of
Prospective
Cohort – not really a prospective cohort b/c there is no control
group, it is more an observational study
Setting:
London academic hospital. Population/Patients:
N = 68, pt’s with autism dx’d since 1950 meeting 4 core criteria,
outpts in London, all white, mostly male 61:7, without medical
problems related to autism, without major physical or sensory
impairments, and without prolonged institutional care. All dx’d
before age 15 with IQ above 50. Prognostic factors
for the study early cognitive ability/IQ level.
Outcomes
were not specifically determined a priori, but the study focused on
long-term outcomes for pt’s in certain IQ bands in specific areas of
functioning- social (work, friendships, independence), cognitive,
linguistic and behavioral functioning. The study also investigated
the stability of IQ over time. Follow-up Period:
average time gap between initial assessment and f/u was 22.1 years,
ranging from 6.8 to 41.3 yrs
Main results:
·
Overall social
functioning:
Composite score of work, friendships, and independence created by
investigators.
0 = Very good
(0-2) 1 = Good (3-4) 2 = Fair
(5-7) 3 = Poor (8-10) 4 = Very
Poor (total 11)
o
Work:
0-3, most were engaged in some type of work, but the majority of it was
supported/volunteer (21%), arranged by residential facility (22%), or
was considered a “general work/leisure program” (40%), only a few had
independent jobs (13%), and these were low-paying and found by parental
contacts
o
Friendship:
0- 3, most had no friends (56%), a few had at least 1 friend and shared
mutual participation in activities (26%), and a small amount had friends
arranged only in social groups (15%)
o
Independence:
0-5, most lived with parents or in residential facilities, only 4% lived
independently, 6% in supported housing
·
Tables 4a,b:
PIQ > 70
showed better fx in all areas that were measured separately, with p
values < 0.05
·
“Overall Social Outcomes,” score was compiled:
o
Fig 4: “Good” outcome predictors:
nothing
alone seemed to be a good predictor
·
PIQ >
100 & VIQ > 70 showed only “Good” and “Fair” outcomes, no
“Poor,” but only 4% of pts were in this category, so hard to draw
conclusions
·
PIQ >
70 & VIQ > 70 = 5 “Very/Good,” 4 “Fair,” 2 “Poor”
·
PIQ >
70 & VIQ > 50 = 8 “Very/Good,” 7 “Fair,” 5 “poor”
·
Conclusion:
PIQ or VIQ alone were not good predictors of outcome
o
Fig 5: “Poor” outcome predictors:
·
Pt’s
with PIQ < 70 & VIQ < 30 ALL were “V/Poor,” and even PIQ
<70 & VIQ <50 were mostly “V/Poor.”
·
Conclusion:
if PIQ < 70, very likely to have poor outcome
Conclusions:
The results indicate that pt’s with PIQ above 70 have variable overall
outcomes, as this group could still have poor outcomes due to autistic
behaviors (fixed routines, preoccupations, need for predictability, and
anxiety about change), which could essentially overcome the benefits of
a relatively high IQ. More predictive is a poor outcome with PIQ below
70, as very few pt’s in this group did well. So, a better cutoff of
good vs poor outcomes may be 70 instead of 50, as previously believed.
However, other factors may influence outcomes when PIQ is above 70, such
as, family support, which played a major role in job acquisition in this
study.
The study had multiple limitations. One concern is the main outcome
measure, a scale the investigators created, which is neither
standardized or validity. Weaknesses also include recall bias of
parents, who did most of the reporting: i.e. recalling if pt had useful
language at age 5, or how many friends pt had as a child. Another
limitation of the study was the IQ testing, as different scales were
used in a hierarchical manner to measure Performance IQ: WISC-R or
WPPSI (44%), Merrill Palmer (50%), Leiter (3%), Stanford Binet (1%), and
Vineland Adaptive Behavior Scales (1%). IQ was stable from childhood
to adulthood, so contrary to former studies, early testing may be more
reliable than previously expected, if one chooses the right tests.
This study may be
clinically useful in answering parental questions concerning prognosis
in children with autism.
Synopsis:
In
this prospective, long-term follow-up study of 68 patients with autism
in London, which investigated factors affecting long term outcomes in
overall functioning, the only clear prognostic indicator was a poor
prognosis associated with a Performance IQ of below 70. Pt’s who did
well more often had PIQs greater than 70; however, outcomes for all pt’s
with PIQs above 70 were variable, as other factors likely affected
outcomes, such as functional impairment due to autistic-type behaviors.
Leigh Fylstra, MD; CAT 3/19/07
Return
to top of the page
AUTHOR
Maria Almond
Firearm Violence
and Serious Violent Behavior
Chairman’s Rounds, May 7, 2007
Citation:,
Bingenheimer, J et al. “Firearm Violence Exposure and Serious Violent
Behavior.” Science. 2005. 308:1323.
Clinical Question:
Does
exposure to firearm violence increase propensity towards serious violent
behavior in the future?
Study Design
Type:
Cohort
Relevant Criteria
for Appropriate Design:
Setting: 78 neighborhoods in Chicago, IL
Population/Patients:
1517 adolescents ages 12yrs or 15yrs and their caregivers from PHDCN
Longitudinal
Cohort
Study—stratified probability samples from 21 socio-economic and racial
strata. Used propensity
scores
to group subjects with same identified risk factors for violence
exposure, then compared
outcomes
within these groups.
Prognostic factors,
Risk factors, and/or Exposures: Exposure to firearm violence (shot or
shot at, or seen someone
shot or shot at) in the past 12 months
Outcomes:
Perpetration of violence (carried hidden weapon, attacked someone with a
weapon, shot someone, shot at
someone, been in a gang fight in which someone was hurt or threatened
with harm) over previous 12
months
Follow-up Period: 3
interview assessments over 5 years.
Question Type:
Correlation
Validity Criteria
for Appropriate Question Type:
Population: 1517 adolescents from two age groups—12yrs and
15yrs—enabling 5 year follow-up while population still mostly within
adolescent age-range. Based from 78 neighborhoods in urban Chicago.
Total 153 pre-exposure covariatesàfrom
baseline survey 139 covariates in 10 domains: demographic background,
family hx, home environment, temperament, health and physical
development, social support, peer influences, vocab/reading proficiency,
school-related factors, behavioral patterns, previous exposure to
violence. Additionally 14 neighborhood social and economic
characteristics via census data and independent survey of prob sample of
adult residents. Because of vast amounts of information available from
baseline assessment, able to create groups with similar likelihood of
exposure for comparisonàpropensity
scores.
Exposure/Outcome
Criteria: Given both exposure and outcome data assessed by self-report,
may be retrospective reporting bias. Unable to determine direction or
magnitude of such bias.
Follow-up: over 5
yrs.
1)
Assessment 1: n=1517
àbaseline
data
2)
Assessment 2: n=1239 (81.7%)
à
exposure data
3)
Assessment 3: n=984 (80.3%)
à
outcome data
Concern
for attrition bias, esp if differential and associated with one
particular group. In subsequent analyses, did note find that attrition
strongly related to baseline covariates measured. However, may have
been unmeasured pre-exposure covariate that may have influenced both
exposure and outcome uncorrelated with the other 153 outcome measures.
Main Results:
Of the 984 subjects
who completed all three assessments, 87% (n=856) were classified as
non-perpetrators and 12.4% (n=122) had perpetrated violence with 0.6%
(n=6) unclassified. The investigators then used a series of
maximum-likelihood logistic regression models to obtain estimates of
association between exposure to firearm violence and perpetration of
serious violence. As well, adjustment calculations by regression to
exclude confounders such as race/ethnicity, age, sex, family
socioeconomic index, neighborhood of residence, previous violence
exposure, self-reported violent crime, self- and caregiver-reported
delinquency were done. Results showed adjusted OR=2.47 (t966=3.64,
P=0.0003).
Conclusions:
Study examines the effect of an individual’s contact with
violence and how even one exposure can slightly more than double the
odds (OR=2.47) of perpetuating violence oneself. Extensive data
gathering over a large population and with a broad range collection of
confounding factor data allowed researchers to begin to isolate possible
causal mechanisms. Because exposure to violence cannot be done ethically
within the setting of an RCT, this observational study, following a
cohort over time, grouped by likelihood of exposure based on
risk-factors, and waiting for both exposure and outcome data allows one
to come close to approximating the independent contribution of gun
violence itself. Major limitations, of course, include retrospective,
self-reporting bias and the possibility that despite the 153 covariates
collected, another factor uncorrelated with the other covariates could
be influencing the results. Also possible that there may be a
difference between estimated propensity and true propensity. As well,
because the study eliminated subjects from the lowest and highest
propensity scoring groups, unable to determine whether those who with
very low risk towards violence would become violent, simply pointing to
the fact that those with some risk for violence if exposed, are more
likely to perpetuate violence. However, generally this cohort study
demonstrates the strong impact that exposure to violence has upon
adolescents, specifically in determining one’s risk for perpetuating
violence.
Summary:
An
adolescent in urban Chicago between the ages of 12-17 exposed to firearm
violence will have 2.5 times the odds of perpetuating violence within
the next 3 years.
Return
to top of the page
Author
Brita Klein
A double-blind, randomized, placebo-controlled trial of fluoxetine in
children and adolescents with depression
Chairman’s Rounds, April 2, 2007
Citation:
Emslie G, Rush J, Weinbert W, Kowatch R,
Hughes C, et al. A double-blind, randomized, placebo-controlled trial of
fluoxetine in children and adolescents with depression. Archives of
General Psychiatry 1997;54:1031-37.
Clinical Question:
Is fluoxetine effective in treating depression in children and
adolescents?
Background:
SSRIs are an effective treatment for depression in adults, but the same
results have not been seen in children. Fluoxetine is the currently the
only SSRI that is approved for treatment of depression in children. I
decided to review the evidence for this.
Question Type:
Therapy
Validity Criteria
for Therapy Study: Was follow-up complete? Yes. Were all patients who
entered the trial properly accounted for at its conclusion? Yes, all
patients who dropped out were accounted for by the method of last
observation carried forward. Were patients randomized? Yes-stratified by
age and sex. Was randomization concealed? Yes. Were patients analyzed in
the groups to which they were randomized? Yes. Were all randomized
patient data analyzed? Yes. Were pts in tx and control groups similar?
Yes, except for significantly greater incidence comorbid anxiety in the
tx group (p=0.4). Were patients/clinicians/assessors aware of group
allocation? No. Aside from intervention, were groups treated equally?
Yes.
Study Design Type:
Randomized Controlled Trial
Relevant
Criteria for Appropriate Design:
Patients/Setting:
Children and adolescents aged 7-17 years who were self-referred or
referred by other practitioners to the University of Texas Southwestern
Medical Center Mood Disorders Program in Dallas, Texas. Patients were
evaluated and had to meet DSM-III-R criteria for nonpsychotic MDD x 3
weeks, have a CDRS-R score >40 at week 3 of evaluation, be in good
general health and of normal intelligence. Patients were excluded if
they had bipolar I/II disorder, hx of psychotic depression, independent
sleep-wake disorder, alcohol or substance abuse within the past year, hx
of an eating disorder, hx of previous adequate tx with fluoxetine, and
if there was a family hx of bipolar disorder. Patients who met
inclusion/exclusion criteria were enrolled in a placebo run-in week.
Patients who responded to placebo during this run-in period were
excluded. 96 patients were randomized after the placebo run-in period.
Intervention/Exposure: fluoxetine 20mg po q day vs placebo.
Duration: 8 weeks.
Outcomes:
The Clinical Global Impressions (CGI) scale improvement rating and the
Childhood Depression Rating Scale-Revised (CDRS-R) were selected, a
priori, as the major outcome measures. They were measured weekly for the
duration of the study. Response was defined as a CGI rating of 1 or 2
(very much improved or much improved), and an improvement in the group
mean score of the CDRS-R. Time to response/remission and differences in
response based on age/sex were also examined. In addition, the evaluator
completed the Children’s Global Assessment Scale and Brief Psychiatric
Rating Scale-Children weekly, and the patient/parents completed the Beck
Depression Inventory or Children’s Depression Inventory (depending on
age) and the Weinberg Screening Affective Scale at the beginning and end
of treatment.
Main Results:
-Drop outs in
placebo group: 19 for nonresponse, 1 for adverse effects, and 2 for
protocol violation. Drop outs in fluoxetine group: 7 for nonresponse, 4
for ae (mania and rash), and 3 for protocol violation.
-At week 8, 56%
(27/48) of patients randomized to the fluoxetine group had a CGI
“response”, while only 33% (16/48) of patients in the placebo group had
a CGI “response” (using LOCF). NNT = 5.
-Weekly mean CDRS-R
scores were significantly different between groups starting at week 5.
Final mean CDRS scores in tx group = 38.4, and in placebo group = 47.1
(p=0.008). Effect size= ?
- No significant
differences in CGI/CDRS-R between the sexes or children vs. adolescents.
Conclusions/Limitations:
Depressed children/adolescents who took fluoxetine 20mg po q day
appeared to have a significant reduction in symptoms as compared to
placebo. However, few reached the cut-off score on the CDRS-R for
remission of depression (score of 28) indicating that symptoms were
reduced, but patients were still depressed. This could have been b/c
the trial only last 8 weeks. Significantly more subjects in the
fluoxetine group also had comorbid anxiety and improvement in sxs could
have been be/c of improvement in anxiety. Additionally, many more
patients dropped out of the placebo group b/c of lack of effect, and the
LOCF in these patients would have underrepresented any improvement
simply from the natural hx of the disorder. Indeed, when using only
completers from this 8 week study, the difference between CDRS-R scores
was no longer significant (but power reduced). I would be interested in
seeing results from larger, longer studies. In the absence of that,
however, I do feel that this study shows that fluoxetine improves
depressive sxs in children/adolescents. I would start fluoxetine in
appropriately selected pediatric patients.
Synopsis:
This study was a valid 8 week RCT of 96 patients that showed
significantly greater improvement in depressive symptoms, as measured by
CGI and CDRS-R scores, in depressed children/adolescents who were
randomized to fluoxetine 20mg po q day vs. placebo.
Return
to top of the page
Author:
Sarah Rivelli, MD
Reference:
Wolraich ML, et al .
The Effect of Sugar on Behavior
or Cognition in Children:
a meta-analysis.
JAMA 274(20) Nov 1995, 1617-1621.
BACKGROUND:
Sugar may negatively affect behavior (the "twinkle defense").
Mechanism: allergy
to sugar, or
functional reactive hypoglycemia. In the 1970s, cross-sectional studies
have shown
association between sugar and hyperactivity.
METHODS:
MEDLINE and PsychINFO literature searches using:
sugar, sucrose and ADHD.
Design:
Meta-analysis of randomized trials 4 16 articles,
23 studies (conducted 1982-1994).
Inclusion
criteria:
subjects consumed known qty sugar,
placebo used;
subjects, parents and staff
blinded; means and SD reported.
Intervention:
subjects randomized to sugar vs pbo (aspartame or saccharin).
Outcomes:
Subjective ratings of observers
(parents, teachers,
experimenters); counting
defined behaviors; measures of skills and activities
using standardized measures
(neuropsych
tests, actometer, continuous performance
test, paired
associated learning,
Figure matching, drawing, motor
skills, academic tests);
aggression; self-reported mood.
Analysis: Calculation of effect sizes for dependent measures:
Effect size estimator d=[sugar - pbo]/pooled SD
- Calculated a single effect size for
each theoretically meaningful measurement
construct.
- Multiple effect
sizes from a single study within a measurement
construct were averaged. (from 532 dependent variables-*14
constructs).
- A weighted mean d for each of the 14 constructs (weighted by inverse variance to reflect sampling error)
was calculate.
Homogeneity
statistic Q (tests that the sample of effect
sizes estimates a single underlying population parameter)
calculated.
VALIDITY:
- Systematic review of randomized trials? Yes.
- Description
of how validity of the individual studies was assessed? Yes.
- Results consistent study to study? Yes.
-
Treatment groups similar at baseline? Yes.
- Pts analyzed
in groups that they were randomized? Yes.
- Pt and clinician blinded? Yes.
RESULTS:
- 23 within-subject design studies, sucrose challenge (1.25-5.6g/kg body weight) - except for one that used high-sucrose diet.
Placebo = aspartame (13), saccharin (2), both (6). Subjects were either
normal (8), had ADHD (5), psych inpts
(1), "sugar
reactors" (6),
Prader-Willi syndrome (1).
- 12/14 of the constructs' distributions were homogenous (Q statistic NS); motor skills and academic test
were not due to one outlier study, when this study was
excluded, the Q statistic became NS.
-Exploration
of effect size distribution by subject type and age
group failed to identify any important modifiers of the effect
of sugar on the measurements.
- For all
constructs, the CI of the effect size included zero.
COMMENTS:
Strengths:
Meta-analysis of randomized trials.
Weakness:
Trials could have been missed by search strategy or inclusion criteria,
a small effect could have been missed.
BOTTOM LINE:
Sugar does not affect the behavior or cognitive performance of children. (Why do we feel
so sure it does? It may be due to expectancy: Parents believe children
adversely affected by sugar --> Affects how parents interact and
perceive children.)
Return
to top of the page
Author:
Ana
Carla P. Smith
Influence
of supplementary vitamins, minerals and essential fatty acids on the
antisocial behavior of young adult prisoners.
Bottom Line:
Dietary interventions should be considered when addressing antisocial
behaviors. Reference:
Gesch CB, Hammond SM, Hampson SE, Eves A, Crowder MJ. Influence of
supplementary vitamins, minerals and essential fatty acids on the
antisocial behavior of young adult prisoners: randomized,
placebo-controlled trial. British Journal of Psychiatry.
2002: 181: 22-28.
Methods:
Design
- double-blind,
randomized controlled clinical trial
Setting
- unnamed
prison in
England
: study was managed from
University
of
Surrey
.
Patient
population /Inclusion criteria - 231
volunteer prisoners 18 yrs of age or older.
Exclusion
criteria - 1.
age < 18 yrs.
Interventions
- 1.
'Forceval':
vitamin/mineral supplement vs a vegetable oil-based placebo w/ an
identical opaque bi-colored gelatin
shell. 2. 'Efamol Marine': essential fatty acid supplement (providing
omega-3 and -6 fatty acids w/o an obvious aftertaste that
would have compromised the blind) vs vegetable oil-based placebo of
identical color w/ identical clear gelatin shell.
Main
outcome measure: 1. Governor
reports - prison
reports of serious incidents (e.g. violence),
and Minor reports
- prison reports
of minor infractions (e.g. failure
to comply w/ requirements) over a specific time period formed the measure
of antisocial behavior. 2. Dietary
intake: 7-day food diaries
(prisoners recorded which of the available food choices they ate & how
much. Also recorded 4 of extra items consumed, including spreads, sugar,
etc.) Quantity & type of all food consumed was entered into a
computer program. Participation varied between 2 weeks & 9 months;
average time spent on supplementation = 142 days for both
groups.
Analysis
- negative binomial (mixed
Poisson) regression analysis, to account for individual differences in
rates of disciplinary action.
differences in baseline rates of offending, and how long a person had been
in the trial.
Follow-up
/Assessments- compliance
monitored daily; food diaries monitored weekly; final analysis of reports
done at 9 months.
Validity:
Randomization
- yes.
Participants entered trial en bloc in September 1996 & underwent
psychometric assessment. Their baseline
disciplinary records were obtained. A stratified randomization was
conducted in each of the 4 main prison wings, employing
a random number generator to allocate groups (each wing had active &
placebo group, so they were matched in terms of
disciplinary incidents & daily events). Participants recruited
subsequently over the next 8 months were first grouped by wing, &
then randomly allocated using r.n.g.
Blinding
- yes.
Blister
packs contained either all active or all placebo supplements (5 pills
total), & were stamped w/ an I 1-digit code
during manufacture. Research staff were only provided w/ details of the
respective code allocated for each participant. Each
day, coded packs were labeled w/ participant's name, cell & prison #.
& packs were given each day when prisoners were locked
in cells for lunch. Compliance was monitored & logged through officers
returning the used packs each day & routine cell searches.
Intention
to treat -
yes
& no - FIT
analysis was performed, but not reported as main measure. Authors mention
that w/ IIT, outcomes did not
differ significantly from those reported in the paper.
Analysis
in group assigned -yes.
hut.... (see
fig. I on pg 26). Of 115
allocated to placebo, 90 were analyzed; of 116 allocated to active.
83 were analyzed.
Groups
similar at baseline - yes:
there were no statistically significant differences btw active &
placebo groups on any of the baseline
psychometric scores (see Table 3 on pg 25) testing intelligence, verbal
ability, anger, anxiety, malaise & depression. Groups
were also equivalent in average rates of disciplinary incidents prior to
supplementation. No comment on age. sex. SES. education,
weight, BMI. nutritional status, vitamin deficiencies. etc.
Groups
treated similarly outside of intervention - yes.
Results:
Outcomes
- prisoners
who received active capsules committed on average 11.8 infringements per
1000 person-days. a reduction of
26.3% (95%Cl 8.3 -44.3%) compared to placebo (statistically significant at
p<0.03) (1ff analysis). There were no statistically
significant differences between groups on dietary intake or accuracy of
judgment about group allocation (see Table 4 on
pg 25). With non-completers removed from analysis. active group showed a
reduction in disciplinary incidents from 16 to 10.4.
a35.1% reduction (p< 0.001, 95%CI 16.3 - 53.9%).
whereas placebo group reduced their rate off offending by 6.7% (p> 0.1.
95%CI -15.1 -28.7%: not significant; negative figures indicate an
increased rate of offending). The greatest reduction occurred
for the most serious incidents (Governor reports): 37% (p< 0.005, 95%CI
11.6 - 62.4%). Placebo group reduced their Governor
reports by a non-significant 10.1% (p >0.1. 95%CI -16.9 -
37.1%). Active group showed a significant reduction in minor
reports as well: 33% (p< 0.025, 95%CI 0.9 - 65.7%),
whereas the placebo group showed little reduction: 6.5% (p> 0.1, 95%CI
-28.5 -41.5%.
Adverse
Effects - none
Compliance
w/ therapies - mean
compliance rate for placebo = 89.83% (95%CI 87.43-92.23%). Mean compliance
rate for active grp = 90.67 (95%Cl
88.47-92.87%). Difference in compliance btw groups not statistically
significant.
Comments
/ Application:
Are
the likely benefits worth the potential harms & costs? yes.
Strengths:
1. double-blind,
randomized controlled trial. 2. large sample size. 3.
Weaknesses:
1. Behavior
in institutions may not be generalizable to the general or outpatient
population. 2. No measure of pt's nutritional status at baseline vs after
intervention: does supplementation only work for those w/ antisocial
behavior already consuming poor
diets? 3. Biochemical measures not performed.
Next
steps for further study of this
problem. Repeat
study indifferent settings, w/ different age-groups, sexes, etc. Assess nutritional
status from blood before & during supplementation. Longer f/u.
Return
to top of the page
Clinical
Question:
Does nutrition supplements reduce behavioral misdeeds in
schoolchildren?
Reference
Schoenthaler SJ and Bier ID The effect of vitamin-mineral
supplementation on Juvenile delinquency among American schoolchildren: a
randomized, double-blind placebo-controlled trial. The Journal of
alternative and complementary medicine V6 pp7-17, 2000
Methods
Design
- A stratified randomized, double-blind, placebo-controlled trial with
pretest and posttest measures of antisocial behavior
on school property.
Setting - Two
"working class" , primarily Hispanic elementary schools in the
Cartwright District of phoenix,
Arizona
.
Patient Population - all school
students aged 6-12 in classes of 29 teachers who volunteered to distribute
the tablets daily.
Inclusion criteria: all
students whose parent signed consent.
Exclusion criteria: none
Screening/enrollment
methods: All tested for nonverbal intelligence, academic
performance, hyperkinesis, and delinquency prior to intervention. A
stratified randomized design was based on pre intervention nonverbal
intelligence scores.
Intervention / Control: Active
tablets grp treated with daily vitamin-mineral supplementation at 50% of
the
U.S.
recommended daily allowance (RDA) for 4 months versus placebo.
Assessments:
Violent and nonviolent delinquency as measured by official
school disciplinary records.
Analysis - Non-ITT. This
classical randomized two-group design, with one pretest and one post-test,
required an analysis of covariance with no interaction. Post intervention
disciplinary actions as the covariate, and group assignment as the factor.
The treatment effect of supplementation, which was adjusted for baseline
differences, was the primary variable of interest. Because of the skewing
of the rule violation rates per person, the data were normalized using log10
transformations prior to inferential testing. Analysis of covariance
allowed for numerous variables to serve as potential covariates: school,
teacher's class, grade, age, sex, and student's 10.
VALIDITY
It
is a randomized, double-blind, placebo-controlled trial
Treatment groups appears to be similar at baseline but not explicitly
expressed. Students starting trial are not always accounted for at
conclusion
Students were not always analyzed in groups to which they were randomized.
Students and teachers appears to be blinded to treatment. Did not mention
in detail about other parties. Groups appears to be treated similarly
outside of intervention
Do the study population characteristics describe your patient? Not really.
Results
Of
the 468 students randomly assigned to active or placebo tablets, the 80
who were disciplined at least once between September
1s` and May 1s` served as the research sample. During intervention, the 40
children who received active tablets were disciplined, on average,
1 time each, a 47% lower mean rate of antisocial behavior than the 1.875
times each for the 40 children who received placebos (95% confidence
interval, 29% to 65%, < 5.020). The children who took active tablets
produced lower rates of antisocial behavior in 8 types of recorded
infractions: threats/fighting, vandalism, being disrespectful, disorderly
conduct, defiance, obscenities, refusal to work or serve, endangering
others, and non specified offenses.
CONCLUSIONS:
Poor
nutritional habits in children that lead to low concentrations of
water-soluble vitamins in blood, impair brain function and subsequently
cause violence and other serious antisocial behavior. Correction of
nutrient intake, either through a well-balanced diet or low-dose
vitamin-mineral supplementation, corrects the low concentrations of
vitamins in blood, improves brain function and subsequently lowers
institutional violence and antisocial behavior by almost half. This paper
adds to the literature by enabling previous research to be generalized
from older incarcerated subjects with a history of antisocial behavior to
a normal population of younger children in an educational setting.
Comments
Strengths
It is a randomized, double-blind, placebo-controlled trial in a
difficult study (dynamic outcome).
Weakness:
internal validity appears fair. Poor external validity (generalized).
Incomplete presented data, unable to determined the true conclusion. Non
ITT, stretching conclusion.
Study in context of other available literature and/or current standard-of
care: numerous studies in juvenile correctional institutions, but extended
to elementary school population.
How will this study affect your management of the putative patient?
Although the conclusion is a stretch, it likely to affect my management of
using vitamins due to low risk and other potential benefits.
Next steps for further study of this problem. Contracted therapy - none.
In
summary, given these limitations, this study provides a poor quality
evidence of a moderate benefit for general school population but a fair
quality evidence of a moderate (cohn's d = 0.48) benefit for the
"trouble students", and therefore does support clinical use of
vitamin-mineral supplementation in the habitual violators in schools.
Return
to top of the page
Chairman's Rounds
8/23/04
Sarah Parker, MD, MPH
Bottom Line: In this
randomized trial evaluating the effects of problem-solving skills
training (PSST) and parent management training
(PMT) (PSST, PMT vs. PSST +PMT) in children (age 713) with antisocial
behavior,_ the combined treatment lead to a greater proportion with
range of "nonclinical" functioning than either alone. NNT was
an overall 3 which appears clinically significant; however major limitations include failure to
include drop-outs in analyzation, as well
as other typical limitations
in behavior studies: self-report phenomenon, selection bias, regression to
mean, lack of placebo, lack of long-term outcome. Despite these
shortcomings the results appear
clinically significant.
Background: "Antisocial
behavior includes aggressive acts, theft, vandalism, fire setting,
lying, truancy, and other acts violating social rules and
expectations." Parents and family are related to antisocial behavior (stress
psychopathology, social isolation, poor parental relations, affect
onset, maintenance, escalation). Parent dysfunction and family adversity
predict dropping out, degree of therapeutic change, and
maintenance of treatment. Of treatments employed PSST and PMT are two of many with promising results for documented change. PSST
focuses on individual
child and cognitive-behavioral
scenarios patients bring to interpersonal situations. PMT
focuses on child-rearing
practices, parent-child interaction, and contingencies focusing on
prosocial behavior at home/school. Both have shown improvement
alone.
Reference:
Kazdin, Alan, Siegel Todd, and Debra Bass. Cognitive Problem-Solving Skills Training and Parent Management
Training in the Treatment of
Antisocial Behavior in Children. Journal
of Consulting and Clinical Psychology 1992, 60 (5), 733-747.
Methods: 97 children (21 girls and 76 boys)
referred to treatment at Child Conduct Clinic (see back page for
diagram), randomized control trial, in clinic
Measures
were self-report:
To
examine child antisocial behavior:
Parents-CBCL
(118 item 0-2 scale), TAB (30 item, 5 pnt severity, 3 nt duration), PDR,
TeacherCBCL (teacher report form) , 4. IRT, (,SA ; 4
tm I Sjo -
s ccvk J Children- CATS (30 item-forced to one of three), SRD (37
items)-children asked directly
To
examine parent dysfunction: Parent: PSI (120 item 5 point
scale)-scales of own perception of self and child,
life, BDI (21 item 3 point scale), SCL-90 (Hopkins Symptom Checklist (90
items on 5-pnt), Family Environment Scale
Then: PEI, CEI, TEI- to evaluate progress and acceptability of treatment Each condition 6-8 months
PSST: 25
individual administered about 50 minutes once a week, parents actively
involved as well to watch,
assist, foster problem-solving
PMT:
16 treatment sessions over 6-8 months
for 1.5 -2 hour, q wk then qow for last 3**
Validity: randomized, not blinded, 7 clinicians (MSW or clinical
psychology), followed treatment manuals and trained
prior to delivery, ongoing supervision, weekly review, treatment groups were similar at baseline, PATIENTS starting trial were
NOT accounted for at conclusion, groups ANOVA analysis
not different,
study population characteristics do
describe those patients who stay and comply with treatment
Results:
Diff. in drop-out were lower on WISC IQ, did not do family
differences
For NNT and
TABLE 5 (see back)-best clinical, separate sheet with all tables (lots
of data)
Comment: In
this randomized study, it appears combo treatment is more effective
(clinically significant) than either
PSST or PMT alone despite shortcomings of validity. What
do you think?!
Return
to top of the page
Clinically Appraised Topic: Chairman's Rounds
Resident name: Damon
Tweedy
Date:
8/9/04
Context: Violent behavior remains
a leading cause of death and disability in the
United States
,
particularly among adolescents and
young adults. Longitudinal studies have shown continuity in aggressive
behavior over time, in that children who have aggressive behavior in the
elementary school period are more likely to display antisocial and
violent behavior as adolescents and young adults.
Clinical Question:
Does implementation of a violence prevention curriculum lead to
a reduction in aggressive behavior and an increase in pro social behavior among
elementary school students?
Article:
Grossman
DC
.
Neckerman HJ. Koepsell TD. Liu PY. Asher KN. Beland K. Frey K. Rivara
FP. Effectiveness of a violence
prevention curriculum among children in elementary school. A randomized
controlled trial. JAMA. 277(20):1605-11, 1997 May 28.
Methods:
Design:
Randomized
controlled trial, with 12 elementary schools representing the units of
randomization Setting:
12 elementary schools in
King
County
,
Washington
Patient Population:
A.
Population: 790 students from a total of 49
classrooms at the 12
King
County
elementary schools
B.
Inclusion Criteria: active parental
consent amongst the students in the designated classrooms
C.
Exclusion Criteria: no parental consent obtained
Intervention:
Thirty lessons of Second Step: A Violence Prevention Curriculum, each lasting about 35 minutes, were taught once or twice a week
between December 1993 and May 1994
Outcome Measures: Aggressive and
pro social behavior changes were measured 2 weeks and 6 months after participation
in the curriculum by three
different methods: 1) parent reports, (2) teacher reports, and (3)
professional observation by
trained, blinded observers of a random sub-sample of 588 students (12
per class) in the classroom setting and in the
playground/cafeteria settings.
Results:
·
Table I shows baseline characteristics. Authors note
that a "somewhat" larger proportion of subjects in control schools
were black and were identified as receiving special education services,
while a higher proportion of
subjects in the intervention group were Asians (unclear if these
differences are significant).
·
Parent-Reported data did not reveal any significant
changes between intervention and controls
|