Garrett M. Fitzmaurice

Comprehensive Treatment of Extensively Drug-Resistant Tuberculosis

BACKGROUND Extensively drug-resistant tuberculosis has been reported in 45 countries, including countries with limited resources and a high burden of tuberculosis. We describe the management of extensively drug-resistant tuberculosis and treatment outcomes among patients who were referred for individualized outpatient therapy in Peru. METHODS A total of 810 patients were referred for free individualized therapy, including drug treatment, resective surgery, adverse-event management, and nutritional and psychosocial support. We tested isolates from 651 patients for extensively drug-resistant tuberculosis and developed regimens that included five or more drugs to which the infecting isolate was not resistant. RESULTS Of the 651 patients tested, 48 (7.4%) had extensively drug-resistant tuberculosis; the remaining 603 patients had multidrug-resistant tuberculosis. The patients with extensively drug-resistant tuberculosis had undergone more treatment than the other patients (mean [+/-SD] number of regimens, 4.2+/-1.9 vs. 3.2+/-1.6; P<0.001) and had isolates that were resistant to more drugs (number of drugs, 8.4+/-1.1 vs. 5.3+/-1.5; P<0.001). None of the patients with extensively drug-resistant tuberculosis were coinfected with the human immunodeficiency virus (HIV). Patients with extensively drug-resistant tuberculosis received daily, supervised therapy with an average of 5.3+/-1.3 drugs, including cycloserine, an injectable drug, and a fluoroquinolone. Twenty-nine of these patients (60.4%) completed treatment or were cured, as compared with 400 patients (66.3%) with multidrug-resistant tuberculosis (P=0.36). CONCLUSIONS Extensively drug-resistant tuberculosis can be cured in HIV-negative patients through outpatient treatment, even in those who have received multiple prior courses of therapy for tuberculosis.

Laboratory-based versus non-laboratory-based method for assessment of cardiovascular disease risk: the NHANES I Follow-up Study cohort

BACKGROUND Around 80% of all cardiovascular deaths occur in developing countries. Assessment of those patients at high risk is an important strategy for prevention. Since developing countries have limited resources for prevention strategies that require laboratory testing, we assessed if a risk prediction method that did not require any laboratory tests could be as accurate as one requiring laboratory information. METHODS The National Health and Nutrition Examination Survey (NHANES) was a prospective cohort study of 14 407 US participants aged between 25-74 years at the time they were first examined (between 1971 and 1975). Our follow-up study population included participants with complete information on these surveys who did not report a history of cardiovascular disease (myocardial infarction, heart failure, stroke, angina) or cancer, yielding an analysis dataset N=6186. We compared how well either method could predict first-time fatal and non-fatal cardiovascular disease events in this cohort. For the laboratory-based model, which required blood testing, we used standard risk factors to assess risk of cardiovascular disease: age, systolic blood pressure, smoking status, total cholesterol, reported diabetes status, and current treatment for hypertension. For the non-laboratory-based model, we substituted body-mass index for cholesterol. FINDINGS In the cohort of 6186, there were 1529 first-time cardiovascular events and 578 (38%) deaths due to cardiovascular disease over 21 years. In women, the laboratory-based model was useful for predicting events, with a c statistic of 0.829. The c statistic of the non-laboratory-based model was 0.831. In men, the results were similar (0.784 for the laboratory-based model and 0.783 for the non-laboratory-based model). Results were similar between the laboratory-based and non-laboratory-based models in both men and women when restricted to fatal events only. INTERPRETATION A method that uses non-laboratory-based risk factors predicted cardiovascular events as accurately as one that relied on laboratory-based values. This approach could simplify risk assessment in situations where laboratory testing is inconvenient or unavailable.

Socio-economic status, family disruption and residential stability in childhood: relation to onset, recurrence and remission of major depression.

BACKGROUND Childhood adversity significantly increases the risk of depression, but it is unclear whether this risk is most pronounced for depression occurring early in life. In the present study, we examine whether three aspects of childhood adversity--low socio-economic status (SES), family disruption, and residential instability--are related to increased risk of depression during specific stages of the life course. We also examine whether these aspects of childhood adversity are related to the severity of depression. METHOD A sample of 1089 of the 4140 births enrolled in the Providence, Rhode Island cohort of the National Collaborative Perinatal Project was interviewed between the ages of 18 and 39. Measures of parental SES, childhood family disruption and residential instability were obtained upon mothers enrolment and at age 7. Age at onset of major depressive episode, lifetime number of depressive episodes, and age at last episode were ascertained via structured diagnostic interviews. Survival analysis was used to identify risk factors for depression onset and remission and Poisson regression was used to model the recurrence rate of depressive episodes. RESULTS Low parental SES, family disruption and a high level of residential instability, defined as three or more family moves, were related to elevated lifetime risks of depression; the effects of family disruption and residential instability were most pronounced on depression onset by age 14. Childhood adversity was also related to increased risk of recurrence and reduced likelihood of remission. CONCLUSIONS Childhood social disadvantage significantly influences risk of depression onset both in childhood and in adulthood. Early childhood adversity is also related to poor prognosis.

Statistical methods for assessing agreement

In a previous column 1 we discussed the reliability of a measurement instrument and some of the potential consequences of unreliable measurement. Reliability was broadly defined as the extent to which repeated measurements, taken under the same conditions, are similar to one another. Our previous discussion of reliability focused on the reliability of quantitative data. In many applications, however, the variable of interest is a dichotomous rating (e.g., a diagnosis of the presence or absence of a disorder, a positive or negative test result) where the true classification is either unknown or difficult to ascertain. In that case it is of interest to determine the degree to which there is agreement when more than a single rater classifies subjects. Because the degree of agreement can be considered an upper limit on the reliability of the ratings, poor agreement among raters should prompt concern about the accuracy of the ratings and the quality of the resulting classification. In this column we focus on the seemingly straightforward task of quantifying the degree of agreement between two raters when their ratings produce dichotomous classifications (e.g., presence or absence of a disorder). The generic term raters is used here to refer to clinicians, observers, informants, judges, experts, diagnostic tests, and so on; the term ratings denotes the resulting classifications (e.g., positive or negative, present or absent). For example, consider the case where a sample of N subjects is classified for the presence or absence of an eating disorder by two clinicians. The two raters assess each of the N subjects and make diagnoses independently of each other. That is, the diagnosis of any particular subject made by the first clinician is completely unknown to the second clinician, and vice versa. The overall results of the clinician diagnoses can then be summarized in terms of a two-by-two contingency table. This two-by-two table has rows corresponding to the two possible diagnostic categories (presence or absence of an eating disorder) when the diagnosis is made by the first clinician and columns corresponding to the same two diagnostic categories when the diagnosis is made by the second clinician. The four internal cells of the contingency table contain frequency counts of the number of subjects occurring in the diagnostic categories. Figure 1 shows such a table. Note that the upper left-hand corner cell of the table contains the number of subjects who are diagnosed with an eating disorder by both clinicians (i.e., a number of subjects for whom the two clinicians concur regarding the presence of an eating disorder). Given that the rows and columns of the two-by-two contingency table correspond to diagnoses made by two clinicians, how can we quantify the degree of agreement among the clinicians? In this article we consider some of the more commonly used measures of agreement.

Performance of generalized estimating equations in practical situations.

Moment methods for analyzing repeated binary responses have been proposed by Liang and Zeger (1986, Biometrika 73, 13-22), and extended by Prentice (1988, Biometrics 44, 1033-1048). In their generalized estimating equations (GEE), both Liang and Zeger (1986) and Prentice (1988) estimate the parameters associated with the expected value of an individuals vector of binary responses as well as the correlations between pairs of binary responses. In this paper, we discuss one-step estimators, i.e., estimators obtained from one step of the generalized estimating equations, and compare their performance to that of the fully iterated estimators in small samples. In simulations, we find the performance of the one-step estimator to be qualitatively similar to that of the fully iterated estimator. When the sample size is small and the association between binary responses is high, we recommend using the one-step estimator to circumvent convergence problems associated with the fully iterated GEE algorithm. Furthermore, we find the GEE methods to be more efficient than ordinary logistic regression with variance correction for estimating the effect of a time-varying covariate.

A CAVEAT CONCERNING INDEPENDENCE ESTIMATING EQUATIONS WITH MULTIVARIATE BINARY DATA

Clustered binary data occur commonly in both the biomedical and health sciences. In this paper, we consider logistic regression models for multivariate binary responses, where the association between the responses is largely regarded as a nuisance characteristic of the data. In particular, we consider the estimator based on independence estimating equations (IEE), which assumes that the responses are independent. This estimator has been shown to be nearly efficient when compared with maximum likelihood (ML) and generalized estimating equations (GEE) in a variety of settings. The purpose of this paper is to highlight a circumstance where assuming independence can lead to quite substantial losses of efficiency. In particular, when the covariate design includes within-cluster covariates, assuming independence can lead to a considerable loss of efficiency in estimating the regression parameters associated with those covariates.

Sierra Leone's former child soldiers: a longitudinal study of risk, protective factors, and mental health

OBJECTIVE To investigate the longitudinal course of internalizing and externalizing problems and adaptive/prosocial behaviors among Sierra Leonean former child soldiers and whether postconflict factors contribute to adverse or resilient mental health outcomes. METHOD Male and female former child soldiers (N = 260, aged 10 to 17 years at baseline) were recruited from the roster of an non-governmental organization (NGO)-run Interim Care Center in Kono District and interviewed in 2002, 2004, and 2008. The retention rate was 69%. Linear growth models were used to investigate trends related to war and postconflict experiences. RESULTS The long-term mental health of former child soldiers was associated with war experiences and postconflict risk factors, which were partly mitigated by postconflict protective factors. Increases in externalizing behavior were associated with killing/injuring others during the war and postconflict stigma, whereas increased community acceptance was associated with decreases in externalizing problems (b = -1.09). High baseline levels of internalizing problems were associated with being raped, whereas increases were associated with younger involvement in armed groups and social and economic hardships. Improvements in internalizing problems were associated with higher levels of community acceptance and increases in community acceptance (b = -0.86). Decreases in adaptive/prosocial behaviors were associated with killing/injuring others during the war and postconflict stigma, but partially mitigated by social support, being in school and increased community acceptance (b = 1.93). CONCLUSIONS Psychosocial interventions for former child soldiers may be more effective if they account for postconflict factors in addition to war exposures. Youth with accumulated risk factors, lack of protective factors, and persistent distress should be identified. Sustainable services to promote community acceptance, reduce stigma, and expand social supports and educational access are recommended.

Longitudinal Data Analysis

Preface. Acknowledgments. Acronyms. 1. Introduction. 1.1 Advantages of Longitudinal Studies. 1.2 Challenges of Longitudinal Data Analysis. 1.3 Some General Notation. 1.4 Data Layout. 1.5 Analysis Considerations. 1.6 General Approaches. 1.7 The Simplest Longitudinal Analysis. 1.8 Summary. 2. ANOVA Approaches to Longitudinal Data. 2.1Single-Sample Repeated Measures ANOVA. 2.2 Multiple-Sample Repeated Measures ANOVA. 2.3 Illustration. 2.4 Summary. 3. MANOVA Approaches to Longitudinal Data. 3.1 Data Layout for ANOVA versus MANOVA. 3.2 MANOVA for Repeated Measurements. 3.3 MANOVA of Repeated Measures-s Sample Case. 3.4 Illustration. 3.5 Summary. 4. Mixed-Effects Regression Models for Continuous Outcomes. 4.1 Introduction. 4.2 A Simple Linear Regression Model. 4.3 Random Intercept MRM. 4.4 Random Intercept and Trend MRM. 4.5 Matrix Formulation. 4.6 Estimation . 4.7 Summary. 5. Mixed-Effects Polynomial Regression Models. 5.1 Introduction. 5.2 Curvilinear Trend Model. 5.3 Orthogonal Polynomials. 5.4 Summary. 6. Covariance Pattern Models. 6.1 Introduction. 6.2 Covariance Pattern Models. 6.3 Model Selection. 6.4 Example. 6.5 Summary. 7. Mixed Regression Models with Autocorrelated Errors. 7.1 Introduction. 7.2 MRMs with AC Errors. 7.3 Model Selection. 7.4 Example. 7.5 Summary. 8. Generalized Estimating Equations (GEE) Models. 8.1 Introduction. 8.2 Generalized Linear Models (GLMs). 8.3 Generalized Estimating Equations (GEE) Models. 8.4 GEE Estimation. 8.5 Example. 8.6 Summary. 9. Mixed-Effects Regression Models for Binary Outcomes. 9.1 Introduction. 9.2 Logistic Regression Model. 9.3 Probit Regression Models. 9.4 Threshold Concept. 9.5 Mixed-Effects Logistic Regression Model. 9.6 Estimation. 9.7 Illustration. 9.8 Summary. 10. Mixed-Effects Regression Models for Ordinal Outcomes. 10.1 Introduction. 10.2 Mixed-Effects Proportional Odds Model. 10.3 Psychiatric Example. 10.4 Health Services Research Example. 10.5 Summary. 11. Mixed-Effects Regression Models for Nominal Data. 11.1 Mixed-Effects Multinomial Regression Model. 11.2 Health Services Research Example. 1 1.3 Competing Risk Survival Models. 11.4 Summary. 12. Mixed-effects Regression Models for Counts. 12.1 Poisson Regression Model. 12.2 Modified Poisson Models. 12.3 The ZIP Model. 12.4 Mixed-Effects Models for Counts. 12.5 Illustration. 12.6 Summary. 13. Mixed-Effects Regression Models for Three-Level Data. 13.1 Three-Level Mixed-Effects Linear Regression Model. 13.1.1 Illustration. 13.2 Three-Level Mixed-Effects Nonlinear Regression Models. 13.3 Summary. 14. Missing Data in Longitudinal Studies. 14.1 Introduction. 14.2 Missing Data Mechanisms. 14.3 Models and Missing Data Mechanisms. 14.4 Testing MCAR. 14.5 Models for Nonignorable Missingness. 14.6 Summary. Bibliography. Topic Index.

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.

Effects of marital transitions on changes in dietary and other health behaviours in US male health professionals

Study objective: To examine the effect of change in marital status on health behaviours among men. Design: Longitudinal study of repeated measures of marital status and health behaviours collected at four year intervals (1986–90; 1990–94). Setting: US male health professionals. Participants: 38 865 men aged 40–75 in 1986. Main results: Relative to men who stayed married over four years, men who became widowed increased their alcohol consumption. Men who become divorced or widowed experienced decreases in body mass index. Compared with men who remained unmarried, men who remarried exhibited increases in body mass index along with decreased physical activity. Becoming divorced or widowed was associated with decreased vegetable intake while remarriage was linked to greater consumption. Conclusions: Marital termination may adversely affect health and dietary behaviours among men.