Mounting evidence from both animal and human studies suggests that the epigenome is in constant d... more Mounting evidence from both animal and human studies suggests that the epigenome is in constant drift over the life course in response to stochastic and environmental factors. In humans, this has been highlighted by a small number of studies that have demonstrated discordant DNA methylation patterns in adolescent or adult monozygotic (MZ) twin pairs. However, to date, it remains unclear when such differences emerge, and how prevalent they are across different tissues. To address this, we examined the methylation of four differentially methylated regions associated with the IGF2/H19 locus in multiple birth tissues derived from 91 twin pairs: 56 MZ and 35 dizygotic (DZ). Tissues included cord blood-derived mononuclear cells and granulocytes, human umbilical vein endothelial cells, buccal epithelial cells and placental tissue. Considerable variation in DNA methylation was observed between tissues and between unrelated individuals. Most interestingly, methylation discordance was also present within twin pairs, with DZ pairs showing greater discordance than MZ pairs. These data highlight the variable contribution of both intrauterine environmental exposures and underlying genetic factors to the establishment of the neonatal epigenome of different tissues and confirm the intrauterine period as a sensitive time for the establishment of epigenetic variability in humans. This has implications for the effects of maternal environment on the development of the newborn epigenome and supports an epigenetic mechanism for the previously described phenomenon of 'fetal programming' of disease risk.
Background Both alcohol and cannabis use carry health risks. Both are commonly initiated in adole... more Background Both alcohol and cannabis use carry health risks. Both are commonly initiated in adolescence. To date little research has described trajectories of adolescent cannabis or alcohol use or compared their respective consequences in young adulthood. Methods The design was a 10-year eight-wave cohort study of a state-wide community sample of 1943 Victorians initially aged 14-15 years. Moderate-and high-risk alcohol use was defined according to total weekly alcohol consumption. Moderate-and high-risk cannabis use were defined as weekly and daily use, respectively. Results Around 90% of young adults used either alcohol or cannabis. Although an association existed between alcohol and cannabis use, there was a tendency for heavy users to use one substance predominantly at any one time. Weekly or more frequent cannabis use in the absence of moderate-risk alcohol use in teenagers predicted a sevenfold higher rate of daily cannabis use in young adults but only a twofold increase in high-risk alcohol use. Conversely, moderate-risk adolescent alcohol use in the absence of weekly cannabis predicted an approximately threefold increased rate of both high-risk drinking and daily cannabis use in young adulthood. Selective heavy cannabis use in both adolescence and young adulthood was associated with greater illicit substance use and poorer social outcomes in young adulthood than selective alcohol use. Conclusions Heavier teenage cannabis users tend to continue selectively with cannabis use. Considering their poor young adult outcomes, regular adolescent cannabis users appear to be on a problematic trajectory.
Australian and New Zealand Journal of Psychiatry, 2007
To examine the extent to which excessive drinking in young adults is associated with alcohol abus... more To examine the extent to which excessive drinking in young adults is associated with alcohol abuse and dependence. Cross-sectional analyses were conducted using data from the eighth wave of the Victorian Adolescent Health Cohort Study, which comprised 1943 Victorians currently aged 24-25 years drawn from 44 secondary schools across the state in 1992. The main outcome measures of interest were short-term risk drinking status (based on daily alcohol consumption) and long-term risk drinking status (based on total weekly alcohol consumption). Two out of 5 participants drank at moderate to high risk levels for short-term harm. Yet, because young people tend to drink on only 1-2 days a week, fewer (22%) were at moderate to high risk for long-term harm. Although 20% of the participants met criteria for a diagnosis of alcohol abuse or dependence, most of those in the moderate- to high-risk drinking categories were not diagnosed with either alcohol disorder. Excessive alcohol use in one or two sessions a week appears to be common in young Australian adults. While short- and long-term risky drinking is more common in those with an alcohol use disorder, the majority of moderate- and high-risk drinking is done by those who do not meet criteria for an alcohol use disorder.
Australian and New Zealand Journal of Psychiatry, 2000
Over the past two decades criteria for depressive disorder in adults have come to be widely appli... more Over the past two decades criteria for depressive disorder in adults have come to be widely applied in both adolescent and childhood depression . It is now clear that, in females, depressive symptoms become common after puberty and rates of incident disorder rise steeply from the mid-teens . Depressive disorder is estimated to be the leading cause of disease Objective: Earlier studies have suggested that symptoms of depressive disorder in adolescents may differ from those found in adults. Even so, diagnostic criteria developed in adults have come to be widely applied to younger subjects. This study examines the frequency of ICD-10 symptoms in depressive disorder and their association with severity in a large community sample of adolescents aged 15 to 18 years. Method: A six-wave prospective study of adolescent health and emotional wellbeing in 2032 Australian secondary school students provided an opportunity to conduct a two-phase study of adolescent onset depression. A self-administered computerised form of the revised Clinical Interview Schedule (CIS-R) was used as a first phase diagnostic measure. Second phase assessment using the Composite International Diagnostic Interview (CIDI) allowed the delineation of a group fulfilling criteria on both instruments. The ICD-10 symptoms and severity profiles for depression were generated with standard algorithms. Results: 1947 (95.8%) out of 2032 subjects in the designated sample completed phase 1 assessment at least once. Participation rates at phase 2 interviews were 93%. Over the 30-month study period 69 subjects (10 male, 59 female) fulfilled criteria for ICD-10 depressive episodes on both the CIS-R and CIDI. Thirty-one per cent (n = 21) had experienced a severe episode, 46% (n = 32) moderate and 23% (n = 16) mild episodes. Loss of interest and pleasure, decreased energy and fatigue, sleep disturbance, suicidal ideation and diminished concentration most clearly distinguished adolescents with depressive disorder from controls. Self-reproach and guilt, psychomotor agitation and/or retardation and appetite disturbance with weight change showed the clearest increase in frequency with increasing severity of episode. The somatic syndrome was reported by close to one in three of those with a severe depressive episode, but was uncommon in those with mild and moderate episodes. Conclusions: The ICD-10 diagnostic criteria are applicable to depressive disorder in older adolescents. With the exception of depressed mood, found in one in five noncases, all other symptoms were common in cases and uncommon in non-cases. Practitioner awareness of symptoms indicating the presence and severity of disorder should enhance early identification and choice of treatment in adolescent depression.
Reports of adolescent suicidal behaviour have generally derived from clinical settings but popula... more Reports of adolescent suicidal behaviour have generally derived from clinical settings but population-based studies are likely to provide a clearer epidemiological view. Non-fatal suicidal behaviours were studied in 1699 Australian 15- to 16-year-old secondary school students at 44 schools in the state of Victoria, Australia. Self-reported episodes of self-harm were characterized using items from the Beck Suicide Intent Scale. The 12 month weighted prevalence estimate for deliberate self-harm was 5.1%. The commonest forms were self-laceration (1.7%), self-poisoning (1.5%) and deliberate recklessness (1.8%). Self-poisoning and self-laceration were commoner in girls. The prevalence of 'true suicide attempts' was 0.2%. Most self-harmers did not perceive death as likely, plan self-harming episodes at length or inform others of the episodes. Psychiatric morbidity had the strongest association with self-harm, an association which held for all subtypes. Antisocial behaviour and substance abuse were associated with self-harm in girls but not boys. Sexual activity was independently associated with self-harm in both genders. Deliberate self-harm was common but the great majority of episodes were not 'true suicide attempts'. It is, therefore, possible that attributable mortality and morbidity may be greater in self-harmers without definite suicidal intent.
Asthma severity among mite-sensitized individuals appears to be related to the degree of mite all... more Asthma severity among mite-sensitized individuals appears to be related to the degree of mite allergen exposure. The objective of this study was to assess the long-term effectiveness of mite avoidance measures in the homes of asthmatic children in Melbourne, Australia. The concentration of house dust mite allergen (Der p 1) was measured on the child's mattress surfaces and bedroom floors in 85 homes on 10 occasions over a 16-month period. After the first three visits, all mattresses were covered with a semipermeable encasement, and carpeted bedroom floors were randomly allocated to regular applications of a placebo or an "anti-mite" shampoo. The concentration of Der p 1 recovered was initially high in the carpeted bedrooms (n = 66) (41.1, 95% confidence interval, 30.7 to 55.0 microg Der p 1 per gm) and mattresses (n = 85) (39.6, 27.2 to 57.7). During the initial observation period the concentration of mite allergen fell in the treatment, placebo, and parental control groups. During the seven treatment periods, no differences were seen between the Der p 1 concentrations in the groups using the "anti-mite" shampoo, placebo shampoo, or the parental control group (e.g., at visit 4; 12.6, 8.2 to 19.5; 14.8, 8.6 to 25.1; and 12.0, 8.1 to 17.7 microg/gm, respectively). In contrast, the concentration of Der p 1 in samples from uncarpeted floors and mattress encasements was low (4.1, 2.1 to 8.0 microg/gm and 4.2, 2.6 to 6.5 microg/gm, respectively) and insufficient dust for analysis was frequently obtained from these sites. There was no additional benefit from the use of an "anti-mite" shampoo. The absence of carpets and the use of mattress encasements was a useful long-term strategy for mite allergen avoidance.
Recent advances in statistical software have led to the rapid diffusion of new methods for modell... more Recent advances in statistical software have led to the rapid diffusion of new methods for modelling longitudinal data. Multilevel (also known as hierarchical or random effects) models for binary outcomes have generally been based on a logistic-normal specification, by analogy with earlier work for normally distributed data. The appropriate application and interpretation of these models remains somewhat unclear, especially when compared with the computationally more straightforward semiparametric or 'marginal' modelling (GEE) approaches. In this paper we pose two interrelated questions. First, what limits should be placed on the interpretation of the coefficients and inferences derived from random-effect models involving binary outcomes? Second, what diagnostic checks are appropriate for evaluating whether such random-effect models provide adequate fits to the data? We address these questions by means of an extended case study using data on adolescent smoking from a large cohort study. Bayesian estimation methods are used to fit a discrete-mixture alternative to the standard logistic-normal model, and posterior predictive checking is used to assess model fit. Surprising parallels in the parameter estimates from the logistic-normal and mixture models are described and used to question the interpretability of the socalled 'subject-specific' regression coefficients from the standard multilevel approach. Posterior predictive checks suggest a serious lack of fit of both multilevel models. The results do not provide final answers to the two questions posed, but we expect that lessons learned from the case study will provide general guidance for further investigation of these important issues.
Objectives. We sought to test the efficacy of an intervention that was designed to promote social... more Objectives. We sought to test the efficacy of an intervention that was designed to promote social inclusion and commitment to education, in reducing among students health risk behaviors and improving emotional well-being.
This study aimed to investigate risk factors for the development of intussusception in infants in... more This study aimed to investigate risk factors for the development of intussusception in infants in a developing country with a suspected high incidence and in a developed country with a low incidence. A prospective case-control study of infants <2 years of age with idiopathic intussusception confirmed by air enema or surgery was conducted at the National Hospital of Paediatrics (NHP), Vietnam (n = 533) and the Royal Children's Hospital (RCH), Australia (n = 51). Diagnosis was validated in a subset (84% NHP; 67% RCH) by an independent blinded radiologist. Risk factor assessment was performed using a standardized questionnaire. Stool specimens were assayed for bacterial, viral, and parasitic agents. The incidence of intussusception in Vietnam was 302/100,000 in infants <1 year of age (95% CI: 258-352), substantially higher than in Australia (71/100,000). A strong association with adenovirus infection was observed at both sites (cases positive at NHP: 34%, OR 8.2; cases positive at RCH: 40%, OR 44). No association was identified between intussusception and rotavirus, other enteric pathogens, oral polio vaccine, feeding practices, or living conditions. The incidence of intussusception in infants was markedly higher in Vietnam than in Australia. A strong association between adenovirus infection and intussusception was identified at both sites suggesting that adenovirus may play a role in the etiology of intussusception.
Objective To test the sensitivity and specificity of a clinical case definition of acute intussus... more Objective To test the sensitivity and specificity of a clinical case definition of acute intussusception in infants to assist health-care workers in settings where diagnostic facilities are not available. Methods Prospective studies were conducted at a major paediatric hospital in using a clinical case definition of intussusception. Diagnosis of intussusception was confirmed by air enema or surgery and validated in a subset of participants by an independent clinician who was blinded to the participant's status. Sensitivity of the definition was evaluated in 584 infants aged < 2 years with suspected intussusception (533 infants in Hanoi; 51 in Melbourne). Specificity was evaluated in 638 infants aged < 2 years presenting with clinical features consistent with intussusception but for whom another diagnosis was established (234 infants in Hanoi; 404 in Melbourne). Findings In both locations the definition used was sensitive (96% sensitivity in Hanoi; 98% in Melbourne) and specific (95% specificity in Hanoi; 87% in Melbourne) for intussusception among infants with sufficient data to allow classification (449/533 in Hanoi; 50/51 in Melbourne). Reanalysis of patients with missing data suggests that modifying minor criteria would increase the applicability of the definition while maintaining good sensitivity (96-97%) and specificity (83-89%). Conclusion The clinical case definition was sensitive and specific for the diagnosis of acute intussusception in infants in both a developing country and a developed country but minor modifications would enable it to be used more widely.
Purpose: To examine associations between social relationships and school engagement in early seco... more Purpose: To examine associations between social relationships and school engagement in early secondary school and mental health, substance use, and educational achievement 2-4 years later. Methods: School-based longitudinal study of secondary school students, surveyed at school in Year 8 (13-14-years-old) and Year 10 (16-years-old), and 1-year post-secondary school. A total of 2678 Year 8 students (74%) participated in the first wave of data collection. For the school-based surveys, attrition was Ͻ10%. Seventy-one percent of the participating Year 8 students completed the post-secondary school survey. Results: Having both good school and social connectedness in Year 8 was associated with the best outcomes in later years. In contrast, participants with low school connectedness but good social connectedness were at elevated risk of anxiety/depressive symptoms (odds ratio [OR]: 1.3; 95% confidence interval [CI]: 1.0, 1.76), regular smoking (OR: 2.0; 95% CI: 1.4, 2.9), drinking (OR: 1.7; 95% CI: 1.3, 2.2), and using marijuana (OR: 2.0; 95% CI: 1.6, 2.5) in later years. The likelihood of completing school was reduced for those with either poor social connectedness, low school connectedness, or both. Conclusions: Overall, young people's experiences of early secondary school and their relationships with others may continue to affect their moods, their substance use in later years, and their likelihood of completing secondary school. Having both good school connectedness and good social connectedness is associated with the best outcomes. The challenge is how to promote both school and social connectedness to best achieve these health and learning outcomes.
To examine the uptake and course of smoking in a representative adolescent cohort. Six-wave cohor... more To examine the uptake and course of smoking in a representative adolescent cohort. Six-wave cohort study. Secondary school students initially aged 14-15 years at 44 schools in the state of Victoria, Australia. Computerized questionnaire including 7-day retrospective recall for tobacco use. Prevalence rates for smoking in the past month rose from 25% to 31% and daily smoking 9% to 18% across the 3-year follow-up. Forty-five per cent of the sample smoked at some point but only 18% were daily smokers at the end-point. High rates of short-term cessation were observed for both experimental and daily smokers, but 70% of daily smokers relapsed within 12 months. Occasional smoking at the outset was the strongest predictor of later daily smoking and was also predictive of lower cessation and higher relapse rates. Parental divorce and parental daily smoking were associated with smoking at the outset and parental smoking was strongly predictive of the course of daily smoking. In contrast, prevalence rates of smoking in a subject&#39;s school did not significantly predict either smoking initiation or subsequent course. Female daily smokers were half as likely as males to cease smoking, a finding that accounted for gender differences in smoking prevalence in this sample. The strength of association between occasional and later daily smoking indicates the importance of primary prevention but the variability in the early course indicates that there should be much scope for promotion of adolescent efforts to quit. Both the diminished likelihood of smoking cessation in young women and parental influences on smoking course deserve further exploration.
Mounting evidence from both animal and human studies suggests that the epigenome is in constant d... more Mounting evidence from both animal and human studies suggests that the epigenome is in constant drift over the life course in response to stochastic and environmental factors. In humans, this has been highlighted by a small number of studies that have demonstrated discordant DNA methylation patterns in adolescent or adult monozygotic (MZ) twin pairs. However, to date, it remains unclear when such differences emerge, and how prevalent they are across different tissues. To address this, we examined the methylation of four differentially methylated regions associated with the IGF2/H19 locus in multiple birth tissues derived from 91 twin pairs: 56 MZ and 35 dizygotic (DZ). Tissues included cord blood-derived mononuclear cells and granulocytes, human umbilical vein endothelial cells, buccal epithelial cells and placental tissue. Considerable variation in DNA methylation was observed between tissues and between unrelated individuals. Most interestingly, methylation discordance was also present within twin pairs, with DZ pairs showing greater discordance than MZ pairs. These data highlight the variable contribution of both intrauterine environmental exposures and underlying genetic factors to the establishment of the neonatal epigenome of different tissues and confirm the intrauterine period as a sensitive time for the establishment of epigenetic variability in humans. This has implications for the effects of maternal environment on the development of the newborn epigenome and supports an epigenetic mechanism for the previously described phenomenon of &amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;#39;fetal programming&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;#39; of disease risk.
Background Both alcohol and cannabis use carry health risks. Both are commonly initiated in adole... more Background Both alcohol and cannabis use carry health risks. Both are commonly initiated in adolescence. To date little research has described trajectories of adolescent cannabis or alcohol use or compared their respective consequences in young adulthood. Methods The design was a 10-year eight-wave cohort study of a state-wide community sample of 1943 Victorians initially aged 14-15 years. Moderate-and high-risk alcohol use was defined according to total weekly alcohol consumption. Moderate-and high-risk cannabis use were defined as weekly and daily use, respectively. Results Around 90% of young adults used either alcohol or cannabis. Although an association existed between alcohol and cannabis use, there was a tendency for heavy users to use one substance predominantly at any one time. Weekly or more frequent cannabis use in the absence of moderate-risk alcohol use in teenagers predicted a sevenfold higher rate of daily cannabis use in young adults but only a twofold increase in high-risk alcohol use. Conversely, moderate-risk adolescent alcohol use in the absence of weekly cannabis predicted an approximately threefold increased rate of both high-risk drinking and daily cannabis use in young adulthood. Selective heavy cannabis use in both adolescence and young adulthood was associated with greater illicit substance use and poorer social outcomes in young adulthood than selective alcohol use. Conclusions Heavier teenage cannabis users tend to continue selectively with cannabis use. Considering their poor young adult outcomes, regular adolescent cannabis users appear to be on a problematic trajectory.
Australian and New Zealand Journal of Psychiatry, 2007
To examine the extent to which excessive drinking in young adults is associated with alcohol abus... more To examine the extent to which excessive drinking in young adults is associated with alcohol abuse and dependence. Cross-sectional analyses were conducted using data from the eighth wave of the Victorian Adolescent Health Cohort Study, which comprised 1943 Victorians currently aged 24-25 years drawn from 44 secondary schools across the state in 1992. The main outcome measures of interest were short-term risk drinking status (based on daily alcohol consumption) and long-term risk drinking status (based on total weekly alcohol consumption). Two out of 5 participants drank at moderate to high risk levels for short-term harm. Yet, because young people tend to drink on only 1-2 days a week, fewer (22%) were at moderate to high risk for long-term harm. Although 20% of the participants met criteria for a diagnosis of alcohol abuse or dependence, most of those in the moderate- to high-risk drinking categories were not diagnosed with either alcohol disorder. Excessive alcohol use in one or two sessions a week appears to be common in young Australian adults. While short- and long-term risky drinking is more common in those with an alcohol use disorder, the majority of moderate- and high-risk drinking is done by those who do not meet criteria for an alcohol use disorder.
Australian and New Zealand Journal of Psychiatry, 2000
Over the past two decades criteria for depressive disorder in adults have come to be widely appli... more Over the past two decades criteria for depressive disorder in adults have come to be widely applied in both adolescent and childhood depression . It is now clear that, in females, depressive symptoms become common after puberty and rates of incident disorder rise steeply from the mid-teens . Depressive disorder is estimated to be the leading cause of disease Objective: Earlier studies have suggested that symptoms of depressive disorder in adolescents may differ from those found in adults. Even so, diagnostic criteria developed in adults have come to be widely applied to younger subjects. This study examines the frequency of ICD-10 symptoms in depressive disorder and their association with severity in a large community sample of adolescents aged 15 to 18 years. Method: A six-wave prospective study of adolescent health and emotional wellbeing in 2032 Australian secondary school students provided an opportunity to conduct a two-phase study of adolescent onset depression. A self-administered computerised form of the revised Clinical Interview Schedule (CIS-R) was used as a first phase diagnostic measure. Second phase assessment using the Composite International Diagnostic Interview (CIDI) allowed the delineation of a group fulfilling criteria on both instruments. The ICD-10 symptoms and severity profiles for depression were generated with standard algorithms. Results: 1947 (95.8%) out of 2032 subjects in the designated sample completed phase 1 assessment at least once. Participation rates at phase 2 interviews were 93%. Over the 30-month study period 69 subjects (10 male, 59 female) fulfilled criteria for ICD-10 depressive episodes on both the CIS-R and CIDI. Thirty-one per cent (n = 21) had experienced a severe episode, 46% (n = 32) moderate and 23% (n = 16) mild episodes. Loss of interest and pleasure, decreased energy and fatigue, sleep disturbance, suicidal ideation and diminished concentration most clearly distinguished adolescents with depressive disorder from controls. Self-reproach and guilt, psychomotor agitation and/or retardation and appetite disturbance with weight change showed the clearest increase in frequency with increasing severity of episode. The somatic syndrome was reported by close to one in three of those with a severe depressive episode, but was uncommon in those with mild and moderate episodes. Conclusions: The ICD-10 diagnostic criteria are applicable to depressive disorder in older adolescents. With the exception of depressed mood, found in one in five noncases, all other symptoms were common in cases and uncommon in non-cases. Practitioner awareness of symptoms indicating the presence and severity of disorder should enhance early identification and choice of treatment in adolescent depression.
Reports of adolescent suicidal behaviour have generally derived from clinical settings but popula... more Reports of adolescent suicidal behaviour have generally derived from clinical settings but population-based studies are likely to provide a clearer epidemiological view. Non-fatal suicidal behaviours were studied in 1699 Australian 15- to 16-year-old secondary school students at 44 schools in the state of Victoria, Australia. Self-reported episodes of self-harm were characterized using items from the Beck Suicide Intent Scale. The 12 month weighted prevalence estimate for deliberate self-harm was 5.1%. The commonest forms were self-laceration (1.7%), self-poisoning (1.5%) and deliberate recklessness (1.8%). Self-poisoning and self-laceration were commoner in girls. The prevalence of &#39;true suicide attempts&#39; was 0.2%. Most self-harmers did not perceive death as likely, plan self-harming episodes at length or inform others of the episodes. Psychiatric morbidity had the strongest association with self-harm, an association which held for all subtypes. Antisocial behaviour and substance abuse were associated with self-harm in girls but not boys. Sexual activity was independently associated with self-harm in both genders. Deliberate self-harm was common but the great majority of episodes were not &#39;true suicide attempts&#39;. It is, therefore, possible that attributable mortality and morbidity may be greater in self-harmers without definite suicidal intent.
Asthma severity among mite-sensitized individuals appears to be related to the degree of mite all... more Asthma severity among mite-sensitized individuals appears to be related to the degree of mite allergen exposure. The objective of this study was to assess the long-term effectiveness of mite avoidance measures in the homes of asthmatic children in Melbourne, Australia. The concentration of house dust mite allergen (Der p 1) was measured on the child&#39;s mattress surfaces and bedroom floors in 85 homes on 10 occasions over a 16-month period. After the first three visits, all mattresses were covered with a semipermeable encasement, and carpeted bedroom floors were randomly allocated to regular applications of a placebo or an &quot;anti-mite&quot; shampoo. The concentration of Der p 1 recovered was initially high in the carpeted bedrooms (n = 66) (41.1, 95% confidence interval, 30.7 to 55.0 microg Der p 1 per gm) and mattresses (n = 85) (39.6, 27.2 to 57.7). During the initial observation period the concentration of mite allergen fell in the treatment, placebo, and parental control groups. During the seven treatment periods, no differences were seen between the Der p 1 concentrations in the groups using the &quot;anti-mite&quot; shampoo, placebo shampoo, or the parental control group (e.g., at visit 4; 12.6, 8.2 to 19.5; 14.8, 8.6 to 25.1; and 12.0, 8.1 to 17.7 microg/gm, respectively). In contrast, the concentration of Der p 1 in samples from uncarpeted floors and mattress encasements was low (4.1, 2.1 to 8.0 microg/gm and 4.2, 2.6 to 6.5 microg/gm, respectively) and insufficient dust for analysis was frequently obtained from these sites. There was no additional benefit from the use of an &quot;anti-mite&quot; shampoo. The absence of carpets and the use of mattress encasements was a useful long-term strategy for mite allergen avoidance.
Recent advances in statistical software have led to the rapid diffusion of new methods for modell... more Recent advances in statistical software have led to the rapid diffusion of new methods for modelling longitudinal data. Multilevel (also known as hierarchical or random effects) models for binary outcomes have generally been based on a logistic-normal specification, by analogy with earlier work for normally distributed data. The appropriate application and interpretation of these models remains somewhat unclear, especially when compared with the computationally more straightforward semiparametric or 'marginal' modelling (GEE) approaches. In this paper we pose two interrelated questions. First, what limits should be placed on the interpretation of the coefficients and inferences derived from random-effect models involving binary outcomes? Second, what diagnostic checks are appropriate for evaluating whether such random-effect models provide adequate fits to the data? We address these questions by means of an extended case study using data on adolescent smoking from a large cohort study. Bayesian estimation methods are used to fit a discrete-mixture alternative to the standard logistic-normal model, and posterior predictive checking is used to assess model fit. Surprising parallels in the parameter estimates from the logistic-normal and mixture models are described and used to question the interpretability of the socalled 'subject-specific' regression coefficients from the standard multilevel approach. Posterior predictive checks suggest a serious lack of fit of both multilevel models. The results do not provide final answers to the two questions posed, but we expect that lessons learned from the case study will provide general guidance for further investigation of these important issues.
Objectives. We sought to test the efficacy of an intervention that was designed to promote social... more Objectives. We sought to test the efficacy of an intervention that was designed to promote social inclusion and commitment to education, in reducing among students health risk behaviors and improving emotional well-being.
This study aimed to investigate risk factors for the development of intussusception in infants in... more This study aimed to investigate risk factors for the development of intussusception in infants in a developing country with a suspected high incidence and in a developed country with a low incidence. A prospective case-control study of infants &amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;lt;2 years of age with idiopathic intussusception confirmed by air enema or surgery was conducted at the National Hospital of Paediatrics (NHP), Vietnam (n = 533) and the Royal Children&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;#39;s Hospital (RCH), Australia (n = 51). Diagnosis was validated in a subset (84% NHP; 67% RCH) by an independent blinded radiologist. Risk factor assessment was performed using a standardized questionnaire. Stool specimens were assayed for bacterial, viral, and parasitic agents. The incidence of intussusception in Vietnam was 302/100,000 in infants &amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;lt;1 year of age (95% CI: 258-352), substantially higher than in Australia (71/100,000). A strong association with adenovirus infection was observed at both sites (cases positive at NHP: 34%, OR 8.2; cases positive at RCH: 40%, OR 44). No association was identified between intussusception and rotavirus, other enteric pathogens, oral polio vaccine, feeding practices, or living conditions. The incidence of intussusception in infants was markedly higher in Vietnam than in Australia. A strong association between adenovirus infection and intussusception was identified at both sites suggesting that adenovirus may play a role in the etiology of intussusception.
Objective To test the sensitivity and specificity of a clinical case definition of acute intussus... more Objective To test the sensitivity and specificity of a clinical case definition of acute intussusception in infants to assist health-care workers in settings where diagnostic facilities are not available. Methods Prospective studies were conducted at a major paediatric hospital in using a clinical case definition of intussusception. Diagnosis of intussusception was confirmed by air enema or surgery and validated in a subset of participants by an independent clinician who was blinded to the participant's status. Sensitivity of the definition was evaluated in 584 infants aged < 2 years with suspected intussusception (533 infants in Hanoi; 51 in Melbourne). Specificity was evaluated in 638 infants aged < 2 years presenting with clinical features consistent with intussusception but for whom another diagnosis was established (234 infants in Hanoi; 404 in Melbourne). Findings In both locations the definition used was sensitive (96% sensitivity in Hanoi; 98% in Melbourne) and specific (95% specificity in Hanoi; 87% in Melbourne) for intussusception among infants with sufficient data to allow classification (449/533 in Hanoi; 50/51 in Melbourne). Reanalysis of patients with missing data suggests that modifying minor criteria would increase the applicability of the definition while maintaining good sensitivity (96-97%) and specificity (83-89%). Conclusion The clinical case definition was sensitive and specific for the diagnosis of acute intussusception in infants in both a developing country and a developed country but minor modifications would enable it to be used more widely.
Purpose: To examine associations between social relationships and school engagement in early seco... more Purpose: To examine associations between social relationships and school engagement in early secondary school and mental health, substance use, and educational achievement 2-4 years later. Methods: School-based longitudinal study of secondary school students, surveyed at school in Year 8 (13-14-years-old) and Year 10 (16-years-old), and 1-year post-secondary school. A total of 2678 Year 8 students (74%) participated in the first wave of data collection. For the school-based surveys, attrition was Ͻ10%. Seventy-one percent of the participating Year 8 students completed the post-secondary school survey. Results: Having both good school and social connectedness in Year 8 was associated with the best outcomes in later years. In contrast, participants with low school connectedness but good social connectedness were at elevated risk of anxiety/depressive symptoms (odds ratio [OR]: 1.3; 95% confidence interval [CI]: 1.0, 1.76), regular smoking (OR: 2.0; 95% CI: 1.4, 2.9), drinking (OR: 1.7; 95% CI: 1.3, 2.2), and using marijuana (OR: 2.0; 95% CI: 1.6, 2.5) in later years. The likelihood of completing school was reduced for those with either poor social connectedness, low school connectedness, or both. Conclusions: Overall, young people's experiences of early secondary school and their relationships with others may continue to affect their moods, their substance use in later years, and their likelihood of completing secondary school. Having both good school connectedness and good social connectedness is associated with the best outcomes. The challenge is how to promote both school and social connectedness to best achieve these health and learning outcomes.
To examine the uptake and course of smoking in a representative adolescent cohort. Six-wave cohor... more To examine the uptake and course of smoking in a representative adolescent cohort. Six-wave cohort study. Secondary school students initially aged 14-15 years at 44 schools in the state of Victoria, Australia. Computerized questionnaire including 7-day retrospective recall for tobacco use. Prevalence rates for smoking in the past month rose from 25% to 31% and daily smoking 9% to 18% across the 3-year follow-up. Forty-five per cent of the sample smoked at some point but only 18% were daily smokers at the end-point. High rates of short-term cessation were observed for both experimental and daily smokers, but 70% of daily smokers relapsed within 12 months. Occasional smoking at the outset was the strongest predictor of later daily smoking and was also predictive of lower cessation and higher relapse rates. Parental divorce and parental daily smoking were associated with smoking at the outset and parental smoking was strongly predictive of the course of daily smoking. In contrast, prevalence rates of smoking in a subject&#39;s school did not significantly predict either smoking initiation or subsequent course. Female daily smokers were half as likely as males to cease smoking, a finding that accounted for gender differences in smoking prevalence in this sample. The strength of association between occasional and later daily smoking indicates the importance of primary prevention but the variability in the early course indicates that there should be much scope for promotion of adolescent efforts to quit. Both the diminished likelihood of smoking cessation in young women and parental influences on smoking course deserve further exploration.
Uploads
Papers by John B Carlin