Testing structural models of DSM-IV symptoms of common forms of child and adolescent psychopathology

Confirmatory factor analyses were conducted of Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) symptoms of common mental disorders derived from structured interviews of a representative sample of 4,049 twin children and adolescents and their adult caretakers. A dimensional model based on the assignment of symptoms to syndromes in DSM-IV fit better than alternative models, but some dimensions were highly correlated. Modest sex and age differences in factor loadings and correlations were found that suggest that the dimensions of psychopathology are stable across sex and age, but slightly more differentiated at older ages and in males. The dimensions of symptoms were found to be hierarchically organized within higher-order "externalizing" and "internalizing" dimensions, which accounted for much of their variance. Major depression and generalized anxiety disorder were substantially correlated with both the "externalizing" dimension and the "internalizing" dimension, however, suggesting the need to reconceptualize the nature of these higher-order dimensions.

A randomized controlled comparison of family-based treatment and supportive psychotherapy for adolescent bulimia nervosa

CONTEXT

Evidenced-based treatment trials for adolescents with bulimia nervosa are largely absent.

OBJECTIVE

To evaluate the relative efficacy of family-based treatment (FBT) and supportive psychotherapy (SPT) for adolescents with bulimia nervosa.

DESIGN

Randomized controlled trial.

SETTING

The University of Chicago from April 1, 2001, through June 30, 2006.

PARTICIPANTS

Eighty patients, aged 12 to 19 years, with a DSM-IV diagnosis of bulimia nervosa or a strict definition of partial bulimia nervosa.

INTERVENTIONS

Twenty outpatient visits over 6 months of FBT or SPT. Participants were followed up at 6 months posttreatment.

MAIN OUTCOME MEASURES

Abstinence from binge-and-purge episodes as measured by the Eating Disorder Examination. Secondary outcome measures were Eating Disorder Examination binge-and-purge frequency and Eating Disorder Examination subscale scores.

RESULTS

Forty-one patients were assigned to FBT and 39 to SPT. Categorical outcomes at posttreatment demonstrated that significantly more patients receiving FBT (16 [39%]) were binge-and-purge abstinent compared with those receiving SPT (7 [18%]) (P = .049). Somewhat fewer patients were abstinent at the 6-month follow-up; however, the difference was statistically in favor of FBT vs SPT (12 patients [29%] vs 4 patients [10%]; P = .05). Secondary outcome assessment, based on random regression analysis, revealed main effects in favor of FBT on all measures of eating pathological features (P = .003 to P = .03 for all).

CONCLUSIONS

Family-based treatment showed a clinical and statistical advantage over SPT at posttreatment and at 6-month follow-up. Reduction in core bulimic symptoms was also more immediate for patients receiving FBT vs SPT.

Intra-individual daily and yearly variability in actigraphically recorded sleep measures: the CARDIA study

STUDY OBJECTIVES

To describe the day-to-day and year-to-year variation in sleep characteristics among early middle-aged adults.

DESIGN

Participants wore an Actiwatch (Mini Mitter, Inc) for 3 days on two occasions approximately 1 year apart.

PARTICIPANTS

N = 669 participants aged 38-50 years from the Chicago site of the Coronary Artery Risk Development in Young Adults (CARDIA) cohort study.

MEASUREMENTS AND RESULTS

Sleep measures included sleep duration, sleep latency, sleep efficiency, and time in bed. For each sleep parameter, total variance was decomposed into between-subject variance, within-subject variance from day to day, and within-subject variance from year to year. The standard deviation was calculated from the variance. Analysis yielded a within-subject daily standard deviation (SD) of 1.26 hours and a within-subject yearly SD of 0.39 hours for sleep duration. Daily SD was 30.7 minutes and yearly SD was 6.3 minutes for within-subject variability of sleep latency. Daily SD was 8.4% and yearly SD was 2.7% for within-subject variability of sleep efficiency. Finally, daily SD was 1.31 hours and yearly SD was 0.52 hours for within-subject variability of time in bed.

CONCLUSIONS

For each of the 4 sleep characteristics, nightly variability was much greater than yearly variability, meaning sleep behavior changes little in one year in this cohort of early middle-aged adults, despite large daily fluctuations. These results have important methodological implications, including that single-day measures of sleep may not accurately reflect habitual behavior.

A missing data approach to semi-competing risks problems

For event time data involving multiple mutually exclusive competing causes of failure, classic competing risks results show that marginal survival distributions are not identifiable. In a related instance, one or more failure modes may be observed provided that the failure events occur in a specific order. In such situations, sometimes referred to as semi-competing risks problems, the observations may under realistic assumptions lend information about parameters of interest that would be nonidentifiable in the strict competing risks case. Here, we present an approach that makes use of partially observable multiple modes of failures to obtain an estimate of the marginal distribution of one event type that may occur prior to the occurrence of another event type or be precluded by it. We apply the proposed method to the problem of estimating the distribution of time to tumour recurrence at specific sites among breast cancer patients participating in randomized clinical trials.

Stability of the Pittsburgh Sleep Quality Index and the Epworth Sleepiness Questionnaires over 1 year in early middle-aged adults: the CARDIA study

STUDY OBJECTIVES

To describe the stability of the Pittsburgh Sleep Quality Index (PSQI) and the Epworth Sleepiness Scale (ESS) scores over 1 year among a population-based sample of black and white early middle-aged adults.

PARTICIPANTS

More than 600 participants, aged 38 to 50 years, from the Chicago site of the Coronary Artery Risk Development in Young Adults (CARDIA) Study.

METHODS

The PSQI and ESS were completed twice, approximately 1 year apart, between 2003 and 2005. Seven PSQI 4-level component scores, a global PSQI score, and the ESS scores were calculated. A PSQI global score greater than 5 was classified as poor quality sleep, and an ESS score greater than 10 was classified as high daytime sleepiness.

RESULTS

The mean+/-SD PSQI score was 5.7+/-3.1 in Year 1 and 5.9+/-3.1 in Year 2. The mean ESS score was 7.4+/-4.3 in Year 1 and 7.2+/-4.2 in Year 2. The Pearson correlation coefficient for the PSQI score in both years in the full sample was .68 and ranged from .54 among black men to .72 among black women. The Pearson correlation coefficient for the ESS score in both years in the full sample was .76 and ranged from .70 among black men to .80 among white men. In the full sample, 76% had the same PSQI dichotomous classification, and 85% had the same ESS dichotomous classification in both years.

CONCLUSIONS

These results suggest that the PSQI and ESS are stable measures of sleep quality and sleepiness over the past year in early middle-aged adults.

Predictors of patient selection in bariatric surgery

OBJECTIVE

To identify sociodemographic and clinical predictors of patient selection in bariatric surgery.

SUMMARY BACKGROUND DATA

Population-based studies suggest that bariatric surgery patients are disproportionately privately insured, middle-aged white women. It is uncertain whether such disparities are due to surgeon decisions to operate, differences among morbidly obese individuals in access to surgery, or patients' personal preferences regarding surgical treatment.

METHODS

We conducted a national survey of 1343 U.S. bariatric surgeons. The questionnaire contained clinical vignettes generated using a balanced fractional factorial design. For each of 3 hypothetical patients unique in age, race, gender, body mass index (BMI), comorbidities, social support, functional status, and insurance, respondents were asked if they would operate. Logistic regression was used to determine the odds of selection for each characteristic while controlling for the other 7 characteristics. Subset analyses were also performed using combinations of BMI and comorbidities.

RESULTS

A total of 62.5% of eligible surgeons responded (n = 820). Patient race did not influence surgeon decisions to operate. Hypothetical patient age, BMI, and social support were most influential. In the subgroup of patients who did not meet current NIH BMI and comorbidity criteria for bariatric surgery, male sex (odds ratio [OR], 0.33; 95% confidence interval [CI], 0.14-0.76) was associated with decreased odds of selection. Overall, younger age (OR, 0.09; 95% CI, 0.07-0.11), older age (OR, 0.70; 95% CI, 0.56-0.90), limited functional status (OR, 0.66; 95% CI, 0.52-0.82), poor social support (OR, 0.37; 95% CI, 0.30-0.47), self-pay (OR, 0.72; 95% CI, 0.57-0.91), and public insurance (OR, 0.54; 95% CI, 0.43-0.67) were associated with decreased odds of selection. BMI and comorbidity criteria influenced the magnitude of these effects.

CONCLUSIONS

Patient race did not play a role in surgeon decisions to operate. Further research should examine the roles of unequal access to bariatric surgery and differing socio-cultural perceptions of morbid obesity on racial disparities. The influence of patient age, gender, insurance status, social support, and functional status on decisions to operate was mitigated by BMI and comorbidities. Policy-makers currently debating BMI and comorbidity criteria for bariatric surgery should also consider guidelines pertaining to these sociodemographic issues that influence patient selection in bariatric surgery.

Maternal depression and early positive parenting predict future conduct problems in young children with attention-deficit/hyperactivity disorder

Children with attention-deficit/hyperactivity disorder (ADHD) are at risk for adverse outcomes such as substance abuse and criminality, particularly if they develop conduct problems. Little is known about early predictors of the developmental course of conduct problems among children with ADHD, however. Parental psychopathology and parenting were assessed in 108 children who first met Diagnostic and Statistical Manual of Mental Disorders (4th ed.) criteria for ADHD at 4-7 years old. When demographic variables and baseline ADHD and conduct problems were controlled, maternal depression predicted conduct problems 2-8 years following the initial assessment, whereas positive parenting during the structured parent- child interaction task predicted fewer future conduct problems. These findings suggest that maternal depression is a risk factor, whereas early positive parenting is a protective factor, for the developmental course of conduct problems among children with ADHD.

Scale of interest versus scale of estimation: comparing alternative estimators for the incremental costs of a comorbidity

We investigate how the scale of estimation in risk-adjustment models for health-care costs affects the covariate effect, where the scale of interest for the covariate effect may be different from the scale of estimation. As an illustrative example, we use claims data to estimate the incremental costs associated with heart failure within one year subsequent to myocardial infarction. Here, the scale of interest for the effect of heart failure on costs is additive. However, traditional methods for modeling costs use predetermined scale of estimation - for example, ordinary least squares (OLS) regression assumes an additive scale while log-transformed OLS and generalized linear models with log-link assume a multiplicative scale of estimation. We compare these models with a new flexible model that lets the data determine the appropriate scale of estimation. We use a variety of goodness-of-fit measures along with a modified Copas test to assess robustness, lack of fit, and over-fitting properties of the alternative estimators. Biases up to 19% in the scale of interest are observed due to the misrepresentation of the scale of estimation. The new flexible model is found to appropriately represent the scale of estimation and less susceptible to over-fitting despite estimating additional parameters in the link and the variance functions.

Identifiability assumptions for missing covariate data in failure time regression models

Methods in the literature for missing covariate data in survival models have relied on the missing at random (MAR) assumption to render regression parameters identifiable. MAR means that missingness can depend on the observed exit time, and whether or not that exit is a failure or a censoring event. By considering ways in which missingness of covariate X could depend on the true but possibly censored failure time T and the true censoring time C, we attempt to identify missingness mechanisms which would yield MAR data. We find that, under various reasonable assumptions about how missingness might depend on T and/or C, additional strong assumptions are needed to obtain MAR. We conclude that MAR is difficult to justify in practical applications. One exception arises when missingness is independent of T, and C is independent of the value of the missing X. As alternatives to MAR, we propose two new missingness assumptions. In one, the missingness depends on T but not on C; in the other, the situation is reversed. For each, we show that the failure time model is identifiable. When missingness is independent of T, we show that the naive complete record analysis will yield a consistent estimator of the failure time distribution. When missingness is independent of C, we develop a complete record likelihood function and a corresponding estimator for parametric failure time models. We propose analyses to evaluate the plausibility of either assumption in a particular data set, and illustrate the ideas using data from the literature on this problem.

Objectively measured sleep characteristics among early-middle-aged adults: the CARDIA study

Despite mounting evidence that sleep duration is a risk factor across diverse health and functional domains, little is known about the distribution and determinants of sleep. In 2003-2004, the authors used wrist activity monitoring and sleep logs to measure time in bed, sleep latency (time required to fall asleep), sleep duration, and sleep efficiency (percentage of time in bed spent sleeping) over 3 days for 669 participants at one of the four sites of the Coronary Artery Risk Development in Young Adults (CARDIA) study (Chicago, Illinois). Participants were aged 38-50 years, 58% were women, and 44% were Black. For the entire sample, mean time in bed was 7.5 (standard deviation (SD), 1.2) hours, mean sleep latency was 21.9 (SD, 29.0) minutes, mean sleep duration was 6.1 (SD, 1.2) hours, and mean sleep efficiency was 80.9 (SD, 11.3)%. All four parameters varied by race-sex group. Average sleep duration was 6.7 hours for White women, 6.1 hours for White men, 5.9 hours for Black women, and 5.1 hours for Black men. Race-sex differences (p < 0.001) remained after adjustment for socioeconomic, employment, household, and lifestyle factors and for apnea risk. Income was independently associated with sleep latency and efficiency. Sleep duration and quality, which have consequences for health, are strongly associated with race, sex, and socioeconomic status.

Developmental transitions among affective and behavioral disorders in adolescent boys

BACKGROUND

This paper expands upon recent efforts to advance beyond the examination of concurrent comorbidity between affective and behavioral disorders by testing developmental sequences among disorders. Doing so allows for improved tests of theories, such as Capaldi and Patterson's failure model of Conduct Disorder (CD) and depression. Furthermore, Oppositional Defiant Disorder (ODD) is rarely considered distinctly from CD, minimizing the ability to identify distinct effects among behavioral disorders.

METHODS

This paper used data from the Developmental Trends Study, a clinic-referred cohort of 177 boys, along with their parents, who were assessed regularly using a structured clinical interview and a comprehensive set of other measures. Boys were recruited when they were between the ages of 7 and 12, and were reassessed annually until age 18. Predictive regression models tested the continuities among disorders, with depression, overanxious disorder, Attention Deficit Hyperactivity Disorder (ADHD), ODD and CD examined separately as outcomes.

RESULTS

Each disorder showed homotypic continuity, but a clear developmental sequence of heterotypic continuity also emerged. ADHD was predicted by no other disorders, and exclusively predicted ODD. CD was predicted only by ODD. However, ODD was also directly predictive of future anxiety and depression, and anxiety predicted future depression as well. A specific test of the failure model of CD and depression supported that model.

CONCLUSIONS

ODD appears as a pivotal developmental disorder in young males, in that ODD is notably influential in both subsequent behavioral and affective disorders. CD influences later depression only indirectly, through psychosocial impairment. Anxiety precedes depression, and ADHD is not predicted by other disorders.

Limited English proficiency and breast and cervical cancer screening in a multiethnic population

OBJECTIVES

We examined the relationship between ability to speak English and receipt of Papanicolaou tests, clinical breast examinations, and mammography in a multiethnic group of women in the United States.

METHODS

We used longitudinal data from the Study of Women Across the Nation to examine receipt of breast and cervical cancer screening among Chinese, Japanese, Hispanic, and White women who reported reading and speaking (1) only a language other than English, (2) another language more fluently than English, or (3) only English or another language and English with equal fluency. Logistic regression was used to analyze the data.

RESULTS

Reading and speaking only a language other than English and reading and speaking another language more fluently than English, were significantly and negatively associated with receipt of breast and cervical cancer screening in unadjusted models. Although these findings were attenuated in adjusted models, not speaking English well or at all remained negatively associated with receipt of cancer screening.

CONCLUSIONS

These findings suggest that language barriers contribute to health disparities by impeding adequate health communication.

Family Correlates of Oppositional and Conduct Disorders in Children With Attention Deficit/Hyperactivity Disorder

Comorbidities among children with ADHD are key determinants of treatment response, course, and outcome. This study sought to separate family factors (parental psychopathology and parenting practices) associated with comorbid Oppositional Defiant Disorder (ODD) from those associated with Conduct Disorder (CD) among children with Attention Deficit/Hyperactivity Disorder. Clinic-referred families (n = 149) were diagnosed using DSM-IV criteria. Parents completed measures of parenting practices. Comorbid ODD and CD were significantly associated with maternal negative/ineffective discipline. Comorbid CD, but not ODD, was significantly associated with lack of maternal warmth and involvement, paternal negative/ineffective discipline, and with paternal Antisocial Personality Disorder (APD). However, the risk of CD posed by parenting appeared concentrated among children without a father having APD. While consistent discipline appears important for addressing comorbid ODD and CD, paternal psychopathology and the quality of the relationship between mother and child may pose risk specifically for comorbid CD. Efforts to prevent and/or treat CD should consider not only provision of structure and prudent discipline, but also the affective qualities of the relationship between the primary caretaker and child.

Self-reported smoking, cotinine levels, and patterns of smoking in pregnancy

Non-pregnant adult smokers generally exhibit fairly stable smoking behaviour over time. In studies of this population, cotinine assays are considered a 'gold standard' measure of exposure to cigarette smoke; current smoking status can be validated with high sensitivity and specificity. In contrast, there is substantial within-person fluctuation in pregnancy smoking, as women try repeatedly to quit or cut down. As a result, cotinine measures may be of limited use for validation of amount smoked, as they are informative only about recent exposure, vary with individual smoking topography and are dependent on time lapsed since the last cigarette smoked. Thus, in reproductive epidemiology, where timing, intensity and duration of exposure are critical, self-reported history of cigarette consumption may be a more relevant fetal exposure than current smoking status. If there were substantial within-person variation over the course of pregnancy, numerous measures of cotinine would be needed to characterise patterns of fetal exposure and would not be feasible in many studies. We examined self-reported smoking patterns and compared them to patterns of urinary cotinine levels in a prospective study of 998 pregnant women, recruited 1988-92. Fluctuations in smoking were considerable and, while cotinine measures and self-reported number of cigarettes were highly correlated at any given time point across women (r=0.70), the within-person correlation between the patterns of self-reported number of cigarettes and cotinine levels was weaker (r=0.33). For researchers interested in fetal outcomes in which intensity and timing of exposure are critical, we conclude that self-reported variations in smoking during pregnancy may be a valid way to characterise detailed patterns of fetal exposure in epidemiological studies.

Estimating marginal and incremental effects on health outcomes using flexible link and variance function models

We propose an extension to the estimating equations in generalized linear models to estimate parameters in the link function and variance structure simultaneously with regression coefficients. Rather than focusing on the regression coefficients, the purpose of these models is inference about the mean of the outcome as a function of a set of covariates, and various functionals of the mean function used to measure the effects of the covariates. A commonly used functional in econometrics, referred to as the marginal effect, is the partial derivative of the mean function with respect to any covariate, averaged over the empirical distribution of covariates in the model. We define an analogous parameter for discrete covariates. The proposed estimation method not only helps to identify an appropriate link function and to suggest an underlying distribution for a specific application but also serves as a robust estimator when no specific distribution for the outcome measure can be identified. Using Monte Carlo simulations, we show that the resulting parameter estimators are consistent. The method is illustrated with an analysis of inpatient expenditure data from a study of hospitalists.

Characteristics predicting nursing home admission in the program of all-inclusive care for elderly people

PURPOSE

This study determined overall risk and predictors of long-term nursing home admission within the Program of All-Inclusive Care for the Elderly (PACE).

DESIGN AND METHODS

DataPACE records for 4,646 participants aged 55 years or older who were enrolled in 12 Medicare- and Medicaid-capitated PACE programs during the period from June 1, 1990, to June 30, 1998, were obtained. Participants were enrolled for at least 30 days and had baseline evaluations within 30 days of enrollment. Cox proportional hazard models predicting an outcome of nursing home admission of 30 days or longer were estimated.

RESULTS

The cumulative risk of admission to nursing homes for 30 days or longer was 14.9% within 3 years. Individuals enrolled from a nursing home were at very high risk for future admission, with a relative risk of 5.20 when compared with those living alone. Among individuals enrolled in PACE from the community, age, instrumental activity of daily living dependence, and bowel incontinence were predictive of subsequent nursing home admission. Asians and Blacks had a lower risk of institutionalization than Whites. However, other characteristics were not independently predictive of institutionalization, namely poor cognitive status, number of chronic conditions, activity of daily living deficits, urinary incontinence, several behavioral disturbances, and duration of program operation. Before adjusting for other variables, there was substantial site variability in risk of nursing home admission; this decreased considerably after other characteristics were adjusted for.

IMPLICATIONS

Despite the fact that 100% of the PACE participants were nursing home certifiable, the risk of being admitted to a nursing home long term following enrollment from the community is low. The presence of some reversible risk factors may have implications for early intervention to reduce risk further, although the effect of these interventions is likely to be modest. Individuals who received long-term care in a nursing home prior to enrollment in PACE remain at high risk of readmission, despite the availability of comprehensive services.

The unintended impact of welfare reform on the medicaid enrollment of eligible immigrants

BACKGROUND

During welfare reform, Congress passed legislation barring legal immigrants who entered the United States after August 1996 from Medicaid for five years after immigration. This legislation intended to bar only new immigrants (post-1996 immigrants) from Medicaid. However it may have also deterred the enrollment of legal immigrants who immigrated before 1996 (pre-1996 immigrants) and who should have remained Medicaid eligible.

OBJECTIVES

To compare the Medicaid enrollment of U.S.-born citizens to pre-1996 immigrants, before and after welfare reform, and to determine if variation in state Medicaid policies toward post-1996 immigrants modified the effects of welfare reform on pre-1996 immigrants.

DATA SOURCE/STUDY DESIGN

Secondary database analysis of cross-sectional data from 1994-2001 of the U.S. Census Bureau, Annual Demographic Survey of March Supplement of the Current Population Survey.

SUBJECTS

Low-income, U.S.-born adults (N=116,307) and low-income pre-1996 immigrants (N=24,367) before and after welfare reform.

MEASURES

Self-reported Medicaid enrollment.

RESULTS

Before welfare reform, pre-1996 immigrants were less likely to enroll in Medicaid than the U.S.-born (OR=0.55; 95 percent CI, 0.51-0.59). After welfare reform, pre-1996 immigrants were even less likely to enroll in Medicaid. The proportion of immigrants in Medicaid dropped 3 percentage points after 1996; for the U.S.-born it dropped 1.6 percentage points (p=0.012). Except for California, state variation in Medicaid policy toward post-1996 immigrants did modify the effect of welfare reform on pre-1996 immigrants.

CONCLUSIONS

Federal laws limiting the Medicaid eligibility of specific subgroups of immigrants appear to have had unintended consequences on Medicaid enrollment in the larger, still eligible immigrant community. Inclusive state policies may overcome this effect.

Bootstrap calibration of TRANSMIT for informative missingness of parental genotype data

Informative missingness of parental genotype data occurs when the genotype of a parent influences the probability of the parent's genotype data being observed. Informative missingness can occur in a number of plausible ways and can affect both the validity and power of procedures that assume the data are missing at random (MAR). We propose a bootstrap calibration of MAR procedures to account for informative missingness and apply our methodology to refine the approach implemented in the TRANSMIT program. We illustrate this approach by applying it to data on hypertensive probands and their parents who participated in the Framingham Heart Study.

Transitions in Health Status in Older Patients with Heart Failure

BACKGROUND

We aimed to determine transitions in health perception and functional status in older Medicare patients with heart failure.

METHODS

We used 1991 to 1994 data from the Medicare Current Beneficiary Survey, a database that combines Medicare claims with yearly longitudinal surveys. We identified 872 patients 65 years or older in 1991 with a diagnostic code of heart failure.

RESULTS

At baseline, 58% of the patients rated their general health perception as "fair" or "poor." Over 1 year, 18% of the patients died. Transition matrices revealed that health perception, activities of daily living, and instrumental activities of daily living were strong correlates of mortality; that dramatic changes in health status were relatively uncommon over 1 year among survivors; and that decline was common in patients with "excellent" or "very good" health perception. The prior year's health status and comorbidity were powerful predictors of the subsequent year's health status.

CONCLUSION

Many older patients with heart failure have worsening health status over time. Measures of prior health status can help predict chances of functional recovery.

Alcohol advertising in magazines and adolescent readership

CONTEXT

Adolescent drinking is a major public health concern. The federal government does not restrict alcohol advertising to adolescents, but relies on the alcohol industry for self-regulation.

OBJECTIVES

To investigate recent alcohol advertising in magazines and to determine whether advertising frequency is associated with adolescent readership.

DESIGN, SETTING, AND SUBJECTS

All alcohol advertisements were counted that appeared from 1997-2001 in 35 of 48 major US magazines, which tracked their adolescent readership (3 refused all alcohol advertisements; and advertisement counts were unavailable for 10). Variation was assessed in the advertisement placement frequency for each major category of alcohol (beer, wine and wine coolers, and distilled liquor) by a magazine's adolescent readership (age 12-19 years), young adult readership (age 20-24 years), and older adult readership (age > or =25 years); readership demographics (sex, race, and income); year; frequency of publication; and cost per advertisement.

MAIN OUTCOME MEASURE

Variation in alcohol advertising frequency by adolescent readership.

RESULTS

Adolescent readership ranged from 1.0 to 7.1 million. The alcohol industry placed 9148 advertisements at a cost of 696 million dollars. Of the 9148 advertisements, 1201 (13%) were for beer, 443 (5%) for wine, and 7504 (82%) for liquor. After adjustment for other magazine characteristics, the advertisement rate ratio was 1.6 times more for beer (95% confidence interval [CI], 1.0-2.6; P =.05) and liquor (95% CI, 1.1-2.3; P =.01) for every additional million adolescent readers. Wine industry advertising was not associated with adolescent readership.

CONCLUSIONS

Magazine advertising by the beer and liquor industries is associated with adolescent readership. Industry and federal policymakers should examine ways to regulate advertising that reaches large numbers of adolescents.

Comorbidity and the concentration of healthcare expenditures in older patients with heart failure

OBJECTIVES

To examine comorbidity and concentration of healthcare expenditures in older patients with heart failure (HF) in the Medicare program.

DESIGN

Retrospective analysis of older fee-for-service HF patients, using the 1996 Medicare Current Beneficiary Survey and linked Medicare claims.

SETTING

Variety of clinical settings.

PARTICIPANTS

One thousand two hundred sixty-six older HF patients from a nationally representative survey.

MEASUREMENTS

Medicare expenditure per person and by types of healthcare services, prevalence of comorbid conditions, and multivariate regression on the association between comorbidities and healthcare expenditure.

RESULTS

Medicare spent an average of 16,514 dollars on medical reimbursement for each HF patient in 1996. Eighty-one percent of patients had one or more comorbid diseases according to a 17-disease grouping index. The top 20% of HF patients accounted for 63% of total expenditure. Comorbidity was associated with significantly higher Medicare expenditure. HF patients with more-expensive comorbidities included those with peripheral vascular disease (24% of patients, mean total expenditure 26,954 dollars), myocardial infarction (16% of patients, mean total expenditure 29,867 dollars), renal disease (8% of patients, mean total expenditure 33,014 dollars), and hemiplegia or paraplegia (5% of patients, mean total expenditure 33,234 dollars). Diseases and disorders other than heart failure constituted a significant fraction of the causes of inpatient admissions. Comorbid conditions were more likely to be associated with expensive inpatient care, and patients with these diseases were more likely to spend more overall and more on other types of Medicare services including home health aid, skilled nursing facility, and hospice care.

CONCLUSION

Disease management should consider comorbid conditions for improving care and reducing expenditures in older patients with HF.

Informative missingness in genetic association studies: case-parent designs

We consider the effect of informative missingness on association tests that use parental genotypes as controls and that allow for missing parental data. Parental data can be informatively missing when the probability of a parent being available for study is related to that parent's genotype; when this occurs, the distribution of genotypes among observed parents is not representative of the distribution of genotypes among the missing parents. Many previously proposed procedures that allow for missing parental data assume that these distributions are the same. We propose association tests that behave well when parental data are informatively missing, under the assumption that, for a given trio of paternal, maternal, and affected offspring genotypes, the genotypes of the parents and the sex of the missing parents, but not the genotype of the affected offspring, can affect parental missingness. (This same assumption is required for validity of an analysis that ignores incomplete parent-offspring trios.) We use simulations to compare our approach with previously proposed procedures, and we show that if even small amounts of informative missingness are not taken into account, they can have large, deleterious effects on the performance of tests.

Does bone mineralization reflect economic conditions? An examination using a national US sample

BACKGROUND

Bone density is the most important determinant of osteoporotic fractures, which lead to significant morbidity, disability and mortality for older persons, particularly women. Sensitive periods for bone mineralization in infancy, childhood and adolescence overlap with those for height attainment. After peak density is reached in early adulthood, there is age-related decline. We ask whether bone density, as height, reflects economic conditions.

SUBJECTS AND METHODS

Using the third National Health and Nutrition Examination Survey, a probability sample of the US population which over-sampled Blacks and Mexican Americans, we examine the associations between adult bone mineral content (BMC) adjusted for bone and body size and three economic indicators: education, height and the poverty income ratio (PIR). We compare four groups: US-born Whites, Blacks and Mexican Americans and Mexican-born Mexican Americans.

RESULTS

Education is positively associated with BMC only for White women, although there are similar, non-significant associations for Black women and White and Black men. For women, BMC is more strongly associated with height for Mexican Americans, especially the Mexican-born, than for Whites. For men, the only significant association is a negative one with education for men born in Mexico. PIR is not significantly associated with BMC, although there is pattern of lower income being associated with lower BMC for Whites and Blacks and higher BMC for Mexican Americans.

CONCLUSION

Bone density does not reflect economic conditions as strongly or consistently as does physical stature. However, for women there is evidence that lower economic status in childhood or adolescence is associated with lower bone density.

Waxing and waning in concert: dynamic comorbidity of conduct disorder with other disruptive and emotional problems over 7 years among clinic-referred boys

Six waves of structured diagnostic assessments were conducted of 168 clinic-referred 7- to 12-year-olds, over 7 years. Wave-to-wave changes in the number of conduct disorder (CD) behaviors were paralleled by correlated changes in the numbers of symptoms of oppositional defiant disorder (ODD), attention-deficit/hyperactivity disorder (ADHD), depression, and anxiety. In addition, CD in Wave 1 predicted levels of ODD, ADHD, depression, and anxiety in later waves when initial levels of those symptoms were controlled, but only ODD in Wave 1 predicted CD in later waves when initial CD levels were controlled. These findings indicate a striking degree of dynamic comorbidity between CD and other types of psychopathology and provide an initial empirical framework for needed developmental models of comorbidity.

Performance of weighted estimating equations for longitudinal binary data with drop-outs missing at random

The generalized estimating equations (GEE) approach is commonly used to model incomplete longitudinal binary data. When drop-outs are missing at random through dependence on observed responses (MAR), GEE may give biased parameter estimates in the model for the marginal means. A weighted estimating equations approach gives consistent estimation under MAR when the drop-out mechanism is correctly specified. In this approach, observations or person-visits are weighted inversely proportional to their probability of being observed. Using a simulation study, we compare the performance of unweighted and weighted GEE in models for time-specific means of a repeated binary response with MAR drop-outs. Weighted GEE resulted in smaller finite sample bias than GEE. However, when the drop-out model was misspecified, weighted GEE sometimes performed worse than GEE. Weighted GEE with observation-level weights gave more efficient estimates than a weighted GEE procedure with cluster-level weights.

The working-class context of pregnancy smoking

The risk of smoking during pregnancy in the US is strongly associated with women's individual socioeconomic status (SES) but little is known about the influence of local area context. The aim of this study was to examine whether local-area characteristics increase the risk of smoking during pregnancy above and beyond individual SES. In a hospital-based cohort of 878 pregnant women in California, who delivered between 1980 and 1990, we compared risk of smoking during pregnancy based on individual and local-area factors. Adjusting for individual SES, neighborhood social class was related to smoking in early pregnancy. Living in a predominantly working-class area significantly increased the risk of pregnancy smoking for both working-class and non-working-class women. However, local-area economic and demographic indicators were not related to smoking early in pregnancy. Individual and family characteristics alone may be insufficient to explain smoking during pregnancy; the social class context of the places in which pregnant women live may also influence this behavior.

Adolescent outcomes of childhood conduct disorder among clinic-referred boys: predictors of improvement

Much remains to be learned about the adolescent outcomes of clinic-referred boys whose childhood conduct problems are serious enough to meet diagnostic criteria for conduct disorder (CD). Six structured diagnostic assessments were conducted over 7 years of 73 clinic-referred 7-12-year-old boys who met criteria for CD in Wave 1. There were substantial individual differences in the adolescent outcomes of CD, ranging from worsening to sustained recovery, with most boys showing persistent, but fluctuating levels of CD. Improvement in CD was not accounted for by treatment or incarceration, but more positive outcomes over Waves 2-7 were predicted prospectively with substantial accuracy, using a combination of baseline predictors: less initial severity of CD, fewer symptoms of attention-deficit hyperactivity disorder, higher child verbal intelligence, greater family socioeconomic advantage, and not having antisocial biological parents.

Youth targeting by tobacco manufacturers since the Master Settlement Agreement

The 1998 Master Settlement Agreement (MSA) between tobacco manufacturers and forty-six states bans manufacturers from targeting minors through advertising. To determine how youth targeting in magazine cigarette advertisements changed after the MSA, we analyzed magazine readership and cigarette ads in U.S. magazines from 1997 to 2000. In 2000 all three major manufacturers (Philip Morris, R.J. Reynolds, and Brown and Williamson) failed to comply with the MSA's youth-targeting ban, selectively increasing their youth targeting. Banning all magazine advertising of cigarettes may be necessary to eliminate youth targeting in magazines.

Father absence and familial antisocial characteristics

This study examined family antisocial characteristics according to whether biological fathers live at home and agree to be study participants. Antisocial symptoms were tabulated for 161 clinic-referred children and their parents. Families with fathers at home had fewer paternal, maternal, and child antisocial symptoms, and scored higher on multiple SES indicators, than did families with departed fathers. Antisocial characteristics were highest, and SES was lowest, when fathers could not be located or recruited. Results suggest that requiring father participation (as in family-trio genetic designs) screens out the more antisocial families. Of clinical interest, antisocial behavior in any family member is more likely if the father is absent and nonparticipating. The heightened antisocial behavior in children associated with absent biological fathers was not mitigated by presence of stepfathers and was not accounted for by lower SES. The ethical use of mother report on absent fathers is discussed.

Antisocial personality, substance abuse, and exposure to parental violence in males referred for domestic violence

This study investigated whether childhood disruptive behavior (hyperactivity, oppositional-defiance, conduct problems) plus adult psychopathic adjustment are associated with domestic violence. Adult males (n = 66) in diversion programs completed the Wender Utah Rating Scale (WURS), MMPI Psychopathic Deviate scale (PD), Conflict Tactics Scales representing themselves and their parents, and substance use measures. Substance use and lifespan antisocial personality (measured by high WURS and PD scores) were robust predictors of verbal and moderate physical domestic abuse. Violence in the family of origin was associated with abuse when tested alone, but failed to exhibit unique association with abuse when other predictors were taken into account. The possibility that antisocial batterers respond to contingencies by moderating physical harm, while persisting at psychological harm, is discussed.

Emergency department utilization by noninstitutionalized elders

OBJECTIVES

To the best of the authors' knowledge, no nationally representative, population-based study has characterized the proportion of elders using the emergency department (ED) and factors associated with ED use by elders. This article describes the proportion of elder Medicare beneficiaries using the ED and identifies attributes associated with elder ED users as compared with nonusers.

METHODS

The 1993 Medicare Current Beneficiary Survey was used, a national, population-based, cross-sectional survey of Medicare beneficiaries linked with Medicare claims data. The study population was limited to 9,784 noninstitutionalized individuals aged 66 years or older. The Andersen model of health service utilization was used, which explains variation in ED use through a combination of predisposing (demographic and social), enabling (access to care), and need (comorbidity and health status) characteristics.

RESULTS

Eighteen percent of the sample used the ED at least once during 1993. Univariate analysis showed ED users were older; were less educated and lived alone; had lower income and higher Charlson Comorbidity Index scores; and were less satisfied with their ability to access care than nonusers (p < 0.01, chi-square). Logistic regression identified older age, less education, living alone, higher comorbidity scores, worse reported health, and increased difficulties with activities of daily living as factors associated with ED use (p < 0.05). Need characteristics predicted ED use with the greatest accuracy.

CONCLUSIONS

The proportion of elder ED users is slightly higher than previously reported among Medicare beneficiaries. Need (comorbidity and health status) characteristics predict ED utilization with the greatest accuracy.

Body mass index in a US national sample of Asian Americans: effects of nativity, years since immigration and socioeconomic status

OBJECTIVE

To examine body mass index (BMI) and the proportion overweight and obese among adults age 18-59 in the six largest Asian American ethnic groups (Chinese, Filipino, Asian Indian, Japanese, Korean, Vietnamese), and investigate whether BMI varies by nativity (foreign-vs native-born), years in US, or socioeconomic status.

DESIGN

Cross-sectional interview data were pooled from the 1992-1995 National Health Interview Survey (NHIS).

SUBJECTS

254,153 persons aged 18-59 included in the 1992-1995 NHIS. Sample sizes range from 816 to 1940 for each of six Asian American ethnic groups.

MEASUREMENTS

Self-reported height and weight used to calculate BMI and classify individuals as overweight (BMI > or = 25 kg/m2) or obese (BMI > or = 30 kg/m2), age, sex, years in the US, household income and household size.

RESULTS

For men, the percentage overweight ranges from 17% of Vietnamese to 42% of Japanese, while the total male population is 57% overweight. For women, the percentage overweight ranges from 9% of Vietnamese and Chinese to 25% of Asian Indians, while the total female population is 38% overweight. The percentage of Asian Americans classified as obese is very low. Adjusted for age and ethnicity, the odds ratio for obese is 3.5 for women and 4.0 for men for US-vs foreign-born. Among the foreign-born, more years in the US is associated with higher risk of being overweight or obese. The association between household income for women is similar for US-born Asian Americans and Whites and Blacks, but is much weaker for foreign-born Asian Americans.

CONCLUSIONS

While these data find low proportions of Asian Americans overweight at present, they also imply the proportion will increase with more US-born Asian Americans and longer duration in the US.

Age and gender differences in oppositional behavior and conduct problems: a cross-sectional household study of middle childhood and adolescence

Behavior problems among youths cannot be understood without explaining their age and gender differences, but age and gender differences cannot be explained until they have been accurately described. In a household survey of 1,285 youths aged 9 to 17 years, there were no gender differences in oppositional behavior, but aggression, property offenses, and status offenses were more common among boys. Levels of oppositional behavior were greater at younger ages, aggression peaked near the middle of this age range, and property and status offenses were more prevalent at older ages. These findings are generally consistent with developmental models of conduct problems but are inconsistent with a recent model of gender differences and raise questions about the external validity of current taxonomies.

Low salivary cortisol and persistent aggression in boys referred for disruptive behavior

BACKGROUND

Persistent antisocial behavior in adulthood is often preceded by childhood-onset aggressive conduct disorder. Aggressive syndromes in both children and adults have previously been associated with abnormalities in peripheral responses to stress. One peripheral measure, salivary cortisol concentration, may reflect individual differences in the hypothalamic-pituitary-adrenal axis that underlie propensities for aggression, socialization, and adaptation to stress.

METHODS

The relationship between salivary cortisol levels and aggression was tested in 38 clinic-referred school-aged boys. Persistent aggression was measured by collecting disruptive behavior disorder symptoms in 4 annual clinical evaluations and peer nominations of aggression in the first 2 annual evaluations. Salivary cortisol levels were measured during years 2 and 4 of the study.

RESULTS

Low cortisol levels were associated with persistence and early onset of aggression, particularly when measures of cortisol concentrations were pooled. Boys with low cortisol concentrations at both time points exhibited triple the number of aggressive symptoms and were named as most aggressive by peers 3 times as often as boys who had higher cortisol concentrations at either sampling time.

CONCLUSIONS

This suggests that low hypothalamic-pituitary-adrenal axis activity is a correlate of severe and persistent aggression in male children and adolescents. A restricted (low) range of cortisol variability may be more indicative of persistent aggression than a low concentration of cortisol at any single point in time.

Association of parental psychopathology to the comorbid disorders of boys with attention deficit-hyperactivity disorder

This study examined whether particular forms of parental psychopathology are related to similar forms of comorbid psychopathology in offspring with attention deficit-hyperactivity disorder (ADHD). Parental disorders were assessed using maternal interviews, and child disorders were assessed using multiple-informant interviews for 111 clinic-referred boys (aged 7-12) with Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; American Psychiatric Association, 1987) ADHD. Associations between parental and child internalizing disorders and between parental and child externalizing disorders were found, but associations across categories of disorder (i.e., internalizing and externalizing) were not. Similar relationships were observed in 66 clinic-referred boys without ADHD. These findings support specific modes of familial transmission, in contrast to theories that comorbidity simply reflects more severe psychopathology in children with ADHD.

Site specificity of colorectal neoplasms in families without an inherited syndrome

BACKGROUND

Relatives of patients with bowel neoplasia have an increased risk of bowel neoplasia. If there were concordance in location of neoplasia between relatives, then location-specific screening could be used. Such concordance might also assist in the understanding of the etiology of neoplasia within individual families.

METHODS

We have investigated the concordance in anatomic location of colonic neoplasia between first-degree relatives using a new statistical technique for paired data called alternating logistic regression.

RESULTS

A total of 146 families were ascertained, none of which had clinical evidence of a hereditary predisposition to edon neoplasia. Among those with neoplasia, there was an increased risk for right-sided disease with older age (40% for less than age 60 vs. 58% for at least 70 years of age, p = 0.008). As assessed by the odds ratio, we found no significant concordance within families for location of neoplasia (odds ratio = 1.2: CI [0.7, 2.2]), although there was a suggestion that location in family members of the same generation was more strongly associated (odds ratio 1.87: CI [0.82, 4.25]).

CONCLUSIONS

The lack of concordance within families for location argues against considering family-specific bowel screening protocols and indicates that the most important causes of bowel neoplasia are not sufficiently focused on one anatomic site to facilitate etiologic research.

Evidence of environmental suppression of familial resemblance: height among US Civil War brothers

This study examines, with historical data, whether within family correlations in height varied across environments and whether variability in height was greater in worse environments. To investigate these hypotheses, brothers were identified who were mustered into the Union Army of the US Civil War, using linked records from the 1850 and 1860 censuses and military and medical records. Heights were available for 3898 men aged 18 and older, of whom 595 were further identified as belonging to 288 family sets of two, three or four brothers. Generalized estimating equations were used to concurrently model the mean height, the variance and the correlation between brothers as a function of county population. Heights decreased as county population size increased (p<0.001). The correlation between brothers' heights decreased significantly (p = 0.032) with increasing county population, and the variance increased (p = 0.026). The correlation ranged approximately from 0.63 in the least populous to 0.24 in the most populous counties. The degree of familial resemblance was lower in environments where mean height was lower, and the variability in height was greater, suggesting that the environmental contribution to the variability in height is of greater relative importance in populations reared, on average, in worse environments.

Hypothesis testing under mixture models: application to genetic linkage analysis

In this paper we propose a new class of statistics to test a simple hypothesis against a family of alternatives characterized by a mixture model. Unlike the likelihood ratio statistic, whose large sample distribution is still unknown in this situation, these new statistics have a simple asymptotic distribution to which to refer under the null hypothesis. Simulation results suggest that it has adequate power in detecting the alternatives. Its application to genetic linkage analysis in the presence of the genetic heterogeneity that motivated this work is emphasized.

Short-term consistency in self-reported physical functioning among elderly women: the Women's Health and Aging Study

The assessment of physical functioning and disability is integral to population-based and clinical research carried out among elderly people. Typically, functional status is measured through self-reported responses to questions of the form "Do you have difficulty [doing a specific task]?" Knowledge of the reliability and validity of these self-report measures is key to the interpretation of many research efforts, but data on these measurement parameters are sparse. This paper addresses this deficiency through analyses of data from the Weekly Substudy of the Women's Health and Aging Study, a cohort of Baltimore-area women aged > or =65 years with moderate to severe physical disability. Self-reported data on 20 activities, obtained weekly over a 6-month period in 1993 or 1994, were analyzed to investigate how time intervals between assessments and a subject's age and baseline level of disability influenced the consistency of self-reports of disability at both the population level and the individual level. The prevalence of self-reported difficulty increased with baseline disability and, to a lesser extent, with age group. Consistency for all items was very high over short time intervals, but it decreased substantially with increasing intervals between responses (although associations between responses remained significant at 24 weeks). Consistency did not vary with age or baseline disability. Graphic techniques and statistical methods for use with repeated binary data are also illustrated.

Discarded needles do not increase soon after the opening of a needle exchange program

This study examines the effect of a Needle Exchange Program (NEP) on the quantity and geographic distribution of discarded needles on the streets of Baltimore, Maryland, and presents methods to survey discarded needles in the community. A random sample of 32 city blocks located within high-drug-use census tracts, stratified by east and west sides of the city and by proximity to the NEP, was selected for survey. Three teams surveyed the number of needles and the number of drug vials and unbroken glass bottles ("trash") to control for practice effects. Surveillance was conducted prior to initiation of the NEP in August 1994 and 1 and 2 months thereafter. Over the three study periods, the absolute count of discarded needles was 106, 130, and 128, respectively; the number of vials and bottles was 3,048, 3,825, and 3,796, respectively. The initial nonstatistically significant increase in needles (mean change = 0.38, 95% confidence interval (CI) -0.18 to 0.93) was offset by accounting for background trash. Regression models fitted with the generalized estimating equation method, which accounted for within-block correlation over time, showed no significant increase in the number of needles after adjustment for trash during the first 2 months of the NEP's operation. These data suggest that the initiation of NEPs does not result in an increase in the number of discarded needles on the street.

Determining linkage and mode of inheritance: mode scores and other methods

The lod score method remains a popular approach for detecting linkage and estimating the recombination fraction theta between a marker locus and a trait locus. However, its implementation requires knowledge of all parameters of the genetic mechanism, including the number of loci involved and the genotype specific penetrance, which could depend on factors such as age. When some of the penetrance parameters phi are unknown, several methods are available, and have been reviewed by Hodge and Elston [(1994) Genet Epidemiol 11:329-342]. These include the "wrod score" (lod score maximized over theta under a wrong value of phi) and "mod score" (lod score maximized over both theta and phi) methods for inference on theta. It has further been proposed that the mod score also be used for estimating phi. In this paper, we review and assess the adequacy of these two methods for inferences on both phi and theta. In particular, all of the methods can be seen as variations on likelihood inference, using the information in the conditional likelihood for the marker data, given the trait data. The loss of efficiency of the mod is compared to that of the full likelihood, which utilizes all information available in the trait data. We also propose an alternative, based on the pseudo-likelihood, where phi is estimated via the trait information and plugged into the conditional likelihood. This method is compared to the mod score method, and the advantages and disadvantages of each are elucidated. In particular, it is seen that the pseudo-likelihood method can be more efficient than the mod score method if the ascertainment scheme can be modeled. As examples, both a random sample and a multiplex ascertainment scheme are considered. In addition, the pseudo-likelihood method leads to likelihood ratio tests for detecting linkage with a simple, known asymptotic reference distribution, a feature not shared by the mod score. Finally, we discuss the advantages of using the pseudo-likelihood method over the full likelihood method, both of which are valid when the ascertainment scheme is known.