Biomedical subjects

Garrett M Fitzmaurice

Publications and source records attributed to Garrett M Fitzmaurice.

At least 19 recordsLinked to original sources

PubMed2006

Pseudo-likelihood methods for longitudinal binary data with non-ignorable missing responses and covariates.

In this paper we consider longitudinal studies in which the outcome to be measured over time is binary, and the covariates of interest are categorical. In longitudinal studies it is common for the outcomes and any time-varying covariates to be missing due to missed study visits, resulting in non-monotone patterns of missingness. Moreover, the reasons for missed visits may be related to the specific values of the response and/or covariates that should have been obtained, i.e. missingness is non-ignorable. With non-monotone non-ignorable missing response and covariate data, a full likelihood approach is quite complicated, and maximum likelihood estimation can be computationally prohibitive when there are many occasions of follow-up. Furthermore, the full likelihood must be correctly specified to obtain consistent parameter estimates. We propose a pseudo-likelihood method for jointly estimating the covariate effects on the marginal probabilities of the outcomes and the parameters of the missing data mechanism. The pseudo-likelihood requires specification of the marginal distributions of the missingness indicator, outcome, and possibly missing covariates at each occasions, but avoids making assumptions about the joint distribution of the data at two or more occasions. Thus, the proposed method can be considered semi-parametric. The proposed method is an extension of the pseudo-likelihood approach in Troxel et al. to handle binary responses and possibly missing time-varying covariates. The method is illustrated using data from the Six Cities study, a longitudinal study of the health effects of air pollution.

Air Pollutants↗

PubMed2006

A marginalized pattern-mixture model for longitudinal binary data when nonresponse depends on unobserved responses.

This paper proposes a method for modeling longitudinal binary data when nonresponse depends on unobserved responses. The proposed method presumes that the target of inference is the marginal distribution of the response at each occasion and its dependence on covariates, and can accommodate both monotone and non-monotone missingness. The approach involves a marginally specified pattern-mixture model that directly parameterizes both the marginal means at each occasion and the dependence of each response on indicators of nonresponse pattern. This formulation readily incorporates a variety of nonresponse processes assumed within a sensitivity analysis. Once identifying restrictions have been made, estimation of model parameters proceeds via solution to a set of modified generalized estimating equations. The proposed method provides an alternative to standard selection and pattern-mixture modeling frameworks, while featuring certain advantages of each. The paper concludes with application of the method to data from a contraceptive clinical trial with substantial dropout.

Amenorrhea↗

PubMed2006

Estimation in regression models for longitudinal binary data with outcome-dependent follow-up.

In many observational studies, individuals are measured repeatedly over time, although not necessarily at a set of pre-specified occasions. Instead, individuals may be measured at irregular intervals, with those having a history of poorer health outcomes being measured with somewhat greater frequency and regularity. In this paper, we consider likelihood-based estimation of the regression parameters in marginal models for longitudinal binary data when the follow-up times are not fixed by design, but can depend on previous outcomes. In particular, we consider assumptions regarding the follow-up time process that result in the likelihood function separating into two components: one for the follow-up time process, the other for the outcome measurement process. The practical implication of this separation is that the follow-up time process can be ignored when making likelihood-based inferences about the marginal regression model parameters. That is, maximum likelihood (ML) estimation of the regression parameters relating the probability of success at a given time to covariates does not require that a model for the distribution of follow-up times be specified. However, to obtain consistent parameter estimates, the multinomial distribution for the vector of repeated binary outcomes must be correctly specified. In general, ML estimation requires specification of all higher-order moments and the likelihood for a marginal model can be intractable except in cases where the number of repeated measurements is relatively small. To circumvent these difficulties, we propose a pseudolikelihood for estimation of the marginal model parameters. The pseudolikelihood uses a linear approximation for the conditional distribution of the response at any occasion, given the history of previous responses. The appeal of this approximation is that the conditional distributions are functions of the first two moments of the binary responses only. When the follow-up times depend only on the previous outcome, the pseudolikelihood requires correct specification of the conditional distribution of the current outcome given the outcome at the previous occasion only. Results from a simulation study and a study of asymptotic bias are presented. Finally, we illustrate the main results using data from a longitudinal observational study that explored the cardiotoxic effects of doxorubicin chemotherapy for the treatment of acute lymphoblastic leukemia in children.

Adolescent↗

PubMed2006

A simple imputation method for longitudinal studies with non-ignorable non-responses.

Missing data are a common problem in longitudinal studies in the health sciences. Motivated by data from the Muscatine Coronary Risk Factor (MCRF) study, a longitudinal study of obesity, we propose a simple imputation method for handling non-ignorable non-responses (i.e., when non-response is related to the specific values that should have been obtained) in longitudinal studies with either discrete or continuous outcomes. In the proposed approach, two regression models are specified; one for the marginal mean of the response, the other for the conditional mean of the response given non-response patterns. Statistical inference for the model parameters is based on the generalized estimating equations (GEE) approach. An appealing feature of the proposed method is that it can be readily implemented using existing, widely-available statistical software. The method is illustrated using longitudinal data on obesity from the MCRF study.

Algorithms↗

PubMed2006

Maximum likelihood estimation of marginal pairwise associations with multiple source predictors.

Researchers interested in the association of a predictor with an outcome will often collect information about that predictor from more than one source. Standard multiple regression methods allow estimation of the effect of each predictor on the outcome while controlling for the remaining predictors. The resulting regression coefficient for each predictor has an interpretation that is conditional on all other predictors. In settings in which interest is in comparison of the marginal pairwise relationships between each predictor and the outcome separately (e.g., studies in psychiatry with multiple informants or comparison of the predictive values of diagnostic tests), standard regression methods are not appropriate. Instead, the generalized estimating equations (GEE) approach can be used to simultaneously estimate, and make comparisons among, the separate pairwise marginal associations. In this paper, we consider maximum likelihood (ML) estimation of these marginal relationships when the outcome is binary. ML enjoys benefits over GEE methods in that it is asymptotically efficient, can accommodate missing data that are ignorable, and allows likelihood-based inferences about the pairwise marginal relationships. We also explore the asymptotic relative efficiency of ML and GEE methods in this setting.

Depression↗

PubMed2006

Behavioral problems and the effects of early intervention on eight-year-old children with learning disabilities.

OBJECTIVES: 1) To investigate the comorbidity of verbal and nonverbal learning disability subtypes with several domains of behavior problems among 8-year-old children. 2) To determine whether receipt of an early intervention modified the association between childhood behavior problems and learning disabilities (LD). METHODS: This is a secondary data analysis of the Infant Health and Development Program (IHDP), a randomized clinical trial of an early intervention provided between ages 0 and 3 involving 985 children born low birthweight and premature. The findings are based on a prospective follow-up of these children at 8 years of age. RESULTS: Compared to children without verbal LD (VLD), those with VLD were twice as likely to exhibit clinical levels of total behavior problems and 89% more likely to exhibit externalizing behavior problems. Analysis of specific subscales of behavior revealed significant associations with anxious/depressed and withdrawn behaviors, as well as an increased likelihood of attention problems among children with VLD. No significant association was found between nonverbal LD (NVLD) and any type of behavior problem. Furthermore, there was a significant interaction between VLD and the intervention, in which the odds of internalizing behavior problems were greater among children with VLD. No interaction effect of the intervention occurred for any type of behavior problem among children with NVLD. CONCLUSIONS: These findings provide evidence that distinct differences exist for different learning disability subtypes with regards to behavioral outcomes and the effects of early intervention services among 8-year-old children.

Child↗

PubMed2006

Treatment outcomes for substance abuse among adolescents with learning disorders.

This paper assesses whether chemically dependent adolescents with comorbid learning disorders (LDs) derived less effective treatment results when compared to chemically dependent adolescents without LD and examines the moderating effects of prior treatments, treatment length, and treatment completion. Two hundred one adolescents were recruited between 1992 and 1993 from Massachusetts residential treatment centers and subsequently followed up 6 months after enrollment. Compared to chemically dependent teenagers without LD, those with LD were twice as likely to re-use substances at least once by follow-up. LD teenagers were more likely to attend Alcoholics/Narcotics Anonymous if they had prior admissions to treatment programs and longer treatment length. LD teenagers who completed treatment also experienced a greater decrease in current depression compared to LD teenagers not completing the treatment. This study is the first to consider outcomes of substance abuse treatment for adolescents with LD and contributes to the growing literature on comorbidity and substance abuse treatment.

Adolescent↗

PubMed2006

Trends in overweight from 1980 through 2001 among preschool-aged children enrolled in a health maintenance organization.

OBJECTIVE: To examine overweight trends over a 22-year period among preschool-aged children from primarily middle-income families enrolled in a health maintenance organization. RESEARCH METHODS AND PROCEDURES: From well-child care visits to a Massachusetts health maintenance organization, we randomly selected one visit per child per calendar year, yielding a study sample of 120,680 children seen at 366,109 visits from 1980 through 2001. Using multivariate logistic regression models accounting for repeated observations of individual children across years, we estimated trends in prevalence of overweight (weight-for-length/height > or = 95th percentile) and at-risk-for-overweight (85th to 95th percentile). RESULTS: Over the 22-year study period, the observed prevalence of overweight increased from 6.3% to 10.0% and at-risk-for-overweight increased from 11.1% to 14.4%. These increases were evident among all groups of children including infants < 6 months of age. Overall, the adjusted odds ratios were 1.21 per decade (95% confidence interval, 1.17 to 1.25) for overweight and 1.06 per decade (95% confidence interval, 1.03 to 1.08) for at-risk-for-overweight. DISCUSSION: Rates of overweight are increasing in very young children, even infants, from primarily middle-class families.

Child, Preschool↗

PubMed2006

A hybrid model for nonignorable dropout in longitudinal binary responses.

This article presents a likelihood-based method for handling nonignorable dropout in longitudinal studies with binary responses. The methodology developed is appropriate when the target of inference is the marginal distribution of the response at each occasion and its dependence on covariates. A "hybrid" model is formulated, which is designed to retain advantageous features of the selection and pattern-mixture model approaches. This formulation accommodates a variety of assumed forms of nonignorable dropout, while maintaining transparency of the constraints required for identifying the overall model. Once appropriate identifying constraints have been imposed, likelihood-based estimation is conducted via the EM algorithm. The article concludes by applying the approach to data from a randomized clinical trial comparing two doses of a contraceptive.

Algorithms↗

PubMed2006

Adolescent suicidal ideation as predictive of psychopathology, suicidal behavior, and compromised functioning at age 30.

OBJECTIVE: The purpose of this study was to examine whether suicidal ideation in a community population of adolescents represents normative adolescent angst or is predictive of psychopathology, suicidal and problem behaviors, and compromised functioning 15 years after onset. METHOD: Participants were 346 largely Caucasian individuals who were part of a single-age cohort from a working class community and whose development had been traced prospectively from ages 5 to 30. Those with suicidal ideation at age 15 were compared to those without suicidal ideation at age 15 on measures of psychopathology, suicidal ideation and behavior, problem behaviors, and adult functioning at age 30. Gender differences were assessed across all domains. RESULTS: At age 30, there were marked differences between adolescents with suicidal ideation and adolescents without suicidal ideation of both genders in most domains examined. Subjects with suicidal ideation were twice as likely to have an axis I disorder, nearly 12 times more likely to have attempted suicide by age 30, and 15 times more likely to have expressed suicidal thoughts in the past 4 years. Subjects with suicidal ideation had more problem behaviors and poorer overall functioning as assessed by multiple informants. Their self-perceptions of coping ability, self-esteem, and interpersonal relations were also lower. Although subjects with suicidal ideation among both genders had higher levels of psychopathology, suicidal ideation and behavior, and problem behaviors at age 30, male subjects with suicidal ideation had lower salaries and socioeconomic status and were less likely to have achieved residential independence. CONCLUSIONS: Findings underscore the importance of considering suicidal ideation in adolescence as a marker of severe distress and a predictor of compromised functioning, indicating the need for early identification and continued intervention.

Adaptation, Psychological↗

PubMed2006

Domains of discussion in psychotherapy: what do patients really want?

Little quantitative data exist on what the content of non-manualized psychotherapy in contemporary clinical settings actually is, and what patients and clinicians think it ought to be. This descriptive pilot study identified potential content areas to address in psychotherapy, quantified the relative importance of these domains of discussion to patients vs. clinicians, and attempted to measure the frequency with which these domains are actually addressed in current clinical practice at two academic, urban hospitals. The conteni areas assessed included health habits, avocations, work, family, friendships, community involvement, spirituality/religion, finance, sexuality, political activities, educational and cultural pursuits, and ethnicity/race. Overall, clinicians and patients agreed about the relative personal importance and importance to treatment of these content areas, and they rated work, family, friends, and sexuality as the most important domains of discussion for psychotherapy. Also, this sample of patients did not feel that clinicians undervalue the importance of religion and spirituality.

Adult↗

PubMed2005

Incidence and remission rates of overweight among children aged 5 to 13 years in a district-wide school surveillance system.

OBJECTIVES: To monitor annual changes in weight status, we determined incidence and remission rates of overweight among school-aged children with longitudinal school-based surveillance. METHODS: We estimated 1-year changes in weight status among students enrolled in public schools in Cambridge, Mass. Physical education teachers measured height and weight annually. Adjusted odds ratios (ORs) were estimated via multivariate logistic regression, accounting for repeated observations of individual children across years. RESULTS: The 1-year incidence of overweight was 4%, and the remission rate was 15%. Among overweight children, 85% remained overweight at a 1-year follow-up, while 18% of children at risk for overweight became overweight. Overweight incidence rates were higher among children aged 7 and 8 years than among those aged 11 to 13 years (boys: OR=1.68; 95% confidence interval [CI]=1.14, 2.47; girls: OR=1.86; 95% CI=1.25, 2.77). CONCLUSIONS: Both incidence and remission rates were higher among younger children. Children who were at risk for overweight were more likely to change their weight status than those who were already overweight. Our results support targeting overweight prevention efforts toward younger children and children at risk for overweight.

Adolescent↗

PubMed2005

Regression models for the analysis of longitudinal Gaussian data from multiple sources.

We present a regression model for the joint analysis of longitudinal multiple source Gaussian data. Longitudinal multiple source data arise when repeated measurements are taken from two or more sources, and each source provides a measure of the same underlying variable and on the same scale. This type of data generally produces a relatively large number of observations per subject; thus estimation of an unstructured covariance matrix often may not be possible. We consider two methods by which parsimonious models for the covariance can be obtained for longitudinal multiple source data. The methods are illustrated with an example of multiple informant data arising from a longitudinal interventional trial in psychiatry.

Adolescent↗

PubMed2005

Simple adjustments for randomized trials with nonrandomly missing or censored outcomes arising from informative covariates.

In randomized trials with missing or censored outcomes, standard maximum likelihood estimates of the effect of intervention on outcome are based on the assumption that the missing-data mechanism is ignorable. This assumption is violated if there is an unobserved baseline covariate that is informative, namely a baseline covariate associated with both outcome and the probability that the outcome is missing or censored. Incorporating informative covariates in the analysis has the desirable result of ameliorating the violation of this assumption. Although this idea of including informative covariates is recognized in the statistics literature, it is not appreciated in the literature on randomized trials. Moreover, to our knowledge, there has been no discussion on how to incorporate informative covariates into a general likelihood-based analysis with partially missing outcomes to estimate the quantities of interest. Our contribution is a simple likelihood-based approach for using informative covariates to estimate the effect of intervention on a partially missing outcome in a randomized trial. The first step is to create a propensity-to-be-missing score for each randomization group and divide the scores into a small number of strata based on quantiles. The second step is to compute stratum-specific estimates of outcome derived from a likelihood-analysis conditional on the informative covariates, so that the missing-data mechanism is ignorable. The third step is to average the stratum-specific estimates and compute the estimated effect of intervention on outcome. We discuss the computations for univariate, survival, and longitudinal outcomes, and present an application involving a randomized study of dual versus triple combinations of HIV-1 reverse transcriptase inhibitors.

Acquired Immunodeficiency Syndrome↗

PubMed2005

Relationship of physical fitness to prevalence and incidence of overweight among schoolchildren.

OBJECTIVES: We examined the relationship between comprehensive fitness tests and overweight using a school surveillance system in a racially diverse city in the United States. RESEARCH METHODS AND PROCEDURES: Trained physical education teachers measured weight, height, and fitness annually from 2001 to 2003. We compiled data for a cross-sectional analysis (11,845 measurements on 6297 students, 5 to 14 years of age) and a 1-year prospective analysis (4215 measurements on 2927 students not overweight at baseline, 5 to 13 years of age). Overweight was defined as a BMI > or =95th percentile (Centers for Disease Control and Prevention 2000 growth charts), and underfit was defined as failing at least one of five fitness tests: endurance run, abdominal strength, flexibility, upper body strength, and agility (Amateur Athletic Union and Fitnessgram). Associations between fitness and overweight were examined using multivariate logistic regression models, adjusting for sociodemographic status and repeated measurements over time. RESULTS: The mean number of fitness tests passed was lower among students with a BMI above the 80th percentile. Overweight incidence over 1 year was 7% and 2% for underfit and fit girls, respectively (odds ratio, 3.3; 95% confidence interval, 2.0 to 5.6). Not passing either the endurance run or upper body strength test was associated with overweight incidence in both boys and girls. After adjusting for baseline BMI, the endurance run remained a significant predictor of incident overweight among girls (odds ratio, 2.0; 95% confidence interval, 1.1 to 3.5). DISCUSSION: Findings support a cross-sectional inverse relationship between physical fitness and overweight among school-aged children. The direction of causation between fitness and overweight is not clearly established and merits further study.

Adolescent↗

PubMed2004

Regression analysis of multiple source and multiple informant data from complex survey samples.

In this tutorial, we describe regression-based methods for analysing multiple source data arising from complex sample survey designs. We use the term 'multiple-source' data to encompass all cases where data are simultaneously obtained from multiple informants, or raters (e.g. self-reports, family members, health care providers, administrators) or via different/parallel instruments, indicators or methods (e.g. symptom rating scales, standardized diagnostic interviews, or clinical diagnoses). We review regression models for analysing multiple source risk factors or multiple source outcomes and show that they can be considered special cases of generalized linear models, albeit with correlated outcomes. We show how these methods can be extended to handle the common survey features of stratification, clustering, and sampling weights. We describe how to fit regression models with multiple source reports derived from complex sample surveys using general purpose statistical software. Finally, the methods are illustrated using data from two studies: the Stirling County Study and the Eastern Connecticut Child Survey.

Data Collection↗

PubMed2004

Cognitive performance in childhood and early adult illness: a prospective cohort study.

STUDY OBJECTIVE: To evaluate whether cognitive performance in childhood is an early determinant of adult illness. DESIGN: Prospective cohort study covering over 30 years. SETTING: Providence, Rhode Island, USA. PARTICIPANTS: 633 people ages 30-39 followed up since birth as part of the Providence cohort of the national collaborative perinatal project. MAIN RESULTS: Higher cognitive performance at age 7 was related to a significantly reduced risk of serious illness in adulthood, OR = 0.65 (95%CI: 0.47 to 0.89) for a one standard deviation (15 point) increase in IQ score. This association was independent of both parental socioeconomic status and participant's attained level of education. CONCLUSIONS: General cognitive performance may be an important and informative early determinant of adult health. Further evaluation of this association and mechanisms linking cognitive performance and health may provide new and innovative strategies to improve disease management and reduce morbidity.

Adult↗

PubMed2003

Family disruption in childhood and risk of adult depression.

OBJECTIVE: The authors examined the risk that family disruption and low socioeconomic status in early childhood confer on the onset of major depression in adulthood. METHOD: Participants were 1,104 offspring of mothers enrolled during pregnancy in the Providence, R.I., site of the National Collaborative Perinatal Project. Measures of childhood family disruption and socioeconomic status were obtained before birth and at age 7. Structured diagnostic interviews were used to assess respondents' lifetime history of major depressive episode between the ages of 18 and 39. Survival analysis was used to identify childhood risks for depression onset. RESULTS: Parental divorce in early childhood was associated with a higher lifetime risk of depression among subjects whose mothers did not remarry as well as among subjects whose mothers remarried. These effects were more pronounced when accompanied by high levels of parental conflict. Independent of the respondents' adult socioeconomic status, low socioeconomic status in childhood predicted an elevated risk of depression. CONCLUSIONS: Family disruption and low socioeconomic status in early childhood increase the long-term risk for major depression. Reducing childhood disadvantages may be one avenue for prevention of depression. Identification of modifiable pathways linking aspects of the early childhood environment to adult mental health is needed to mitigate the long-term consequences of childhood disadvantage.

Adolescent↗