ABSTRACT
BACKGROUND AND OBJECTIVES: Opioids are involved in an increasing proportion of suicide deaths. This study examined the association between opioid analgesic prescription initiation and suicidal behavior among young people. METHODS: We analyzed Swedish population-register data on 1 895 984 individuals ages 9 to 29 years without prior recorded opioid prescriptions. We identified prescriptions dispensed from January 2007 onward and diagnosed self-injurious behavior and death by suicide through December 2013. We first compared initiators with demographically matched noninitiators. To account for confounding, we applied an active comparator design, which examined suicidal behavior among opioid initiators relative to prescription nonsteroidal antiinflammatory drug (NSAID) initiators while inverse-probability-of-treatment weighting with individual and familial covariates. RESULTS: Among the cohort, 201 433 individuals initiated opioid prescription. Relative to demographically matched noninitiators, initiators (N = 180 808) had more than doubled risk of incident suicidal behavior (hazard ratio = 2.64; 95% confidence interval [CI], 2.47-2.81). However, in the active comparator design, opioid initiators (N = 86 635) had only 19% relatively greater risk of suicidal behavior compared with NSAID initiators (N = 255 096; hazard ratio = 1.19; 95% CI,: 1.11-1.28), corresponding to a weighted 5-year cumulative incidence of 2.2% (95% CI, 2.1-2.4) for opioid and 1.9% (95% CI, 1.9-2.0) for NSAID initiators. Most sensitivity analyses produced comparable results. CONCLUSIONS: Opioid initiation may make only a small contribution to the elevated risk of suicidal behavior among young people receiving pharmacologic pain management. In weighing benefits and harms of opioid initiation, our results suggest that increased risk of suicidal behavior may not be a major concern.
Subject(s)
Analgesics, Opioid , Suicidal Ideation , Adolescent , Adult , Analgesics, Opioid/adverse effects , Anti-Inflammatory Agents, Non-Steroidal/adverse effects , Child , Humans , Pain/drug therapy , Prescriptions , Young AdultABSTRACT
The literature on latent change score models does not discuss the importance of using a precise time metric when structuring the data. This study examined the influence of time metric precision on model estimation, model interpretation, and parameter estimate accuracy in bivariate LCS (BLCS) models through simulation. Longitudinal data were generated with a panel study where assessments took place during a given time window with variation in start time and measurement lag. The data were analyzed using precise time metric, where variation in time was accounted for, and then analyzed using coarse time metric indicating only that the assessment took place during the time window. Results indicated that models estimated using the coarse time metric resulted in biased parameter estimates as well as larger standard errors and larger variances and covariances for intercept and slope. In particular, the coupling parameter estimates-which are unique to BLCS models-were biased with larger standard errors. An illustrative example of longitudinal bivariate relations between math and reading achievement in a nationally representative survey of children is then used to demonstrate how results and conclusions differ when using time metrics of varying precision. Implications and future directions are discussed.
Subject(s)
Achievement , Reading , Child , Computer Simulation , Humans , MathematicsABSTRACT
Previous research has shown functional mixed-effects models and traditional mixed-effects models perform similarly when recovering individual trajectories when data were generated following a parametric structure. We extend this previous work and compare nonlinear mixed-effects (NMEM) and functional mixed-effects models' (FMEM) ability to recover underlying trajectories when generated from an inherently nonparametric process. Nonlinear trajectories were generated using B-splines, NMEMs and FMEMs were estimated, and the accuracy of the estimated curves was examined. Sample size, number of time points per curve, and measurement design were varied across simulation conditions. Results showed the FMEMs recovered the underlying mean curve more accurately than the NMEMs, and that, the FMEMs tended to recover the underlying individual curves more accurately than the NMEMs. Progesterone cycle data were then analyzed to demonstrate the utility of both approaches, and models performed similarly when analyzing these data.
Subject(s)
Nonlinear Dynamics , Computer Simulation , Sample SizeABSTRACT
Importance: Concerns about adverse outcomes associated with opioid analgesic prescription have led to major guideline and policy changes. Substantial uncertainty remains, however, regarding the association between opioid prescription initiation and increased risk of subsequent substance-related morbidity. Objective: To examine the association of opioid initiation among adolescents and young adults with subsequent broadly defined substance-related morbidity. Design, Setting, and Participants: This cohort study analyzed population-register data from January 1, 2007, to December 31, 2013, on Swedish individuals aged 13 to 29 years by January 1, 2013, who were naive to opioid prescription. To account for confounding, the analysis compared opioid prescription recipients with recipients of nonsteroidal anti-inflammatory drugs as an active comparator, compared opioid-recipient twins and other multiple birth individuals with their nonrecipient co-multiple birth offspring (co-twin control), examined dental prescription as a specific indication, and included individual, parental, and socioeconomic covariates. Data were analyzed from March 30, 2019, to January 22, 2020. Exposures: Opioid prescription initiation, defined as first dispensed opioid analgesic prescription. Main Outcomes and Measures: Substance-related morbidity, assessed as clinically diagnosed substance use disorder or overdose identified from inpatient or outpatient specialist records, substance use disorder or overdose cause of death, dispensed pharmacotherapy for alcohol use disorder, or conviction for substance-related crime. Results: Among the included cohort (n = 1 541 862; 793 933 male [51.5%]), 193 922 individuals initiated opioid therapy by December 31, 2013 (median age at initiation, 20.9 years [interquartile range, 18.2-23.6 years]). The active comparator design included 77 143 opioid recipients without preexisting substance-related morbidity and 229 461 nonsteroidal anti-inflammatory drug recipients. The adjusted cumulative incidence of substance-related morbidity within 5 years was 6.2% (95% CI, 5.9%-6.5%) for opioid recipients and 4.9% (95% CI, 4.8%-5.1%) for nonsteroidal anti-inflammatory drug recipients (hazard ratio, 1.29; 95% CI, 1.23-1.35). The co-twin control design produced comparable results (3013 opioid recipients and 3107 nonrecipients; adjusted hazard ratio, 1.43; 95% CI, 1.02-2.01), as did restriction to analgesics prescribed for dental indications and additional sensitivity analyses. Conclusions and Relevance: Among adolescents and young adults analyzed in this study, initial opioid prescription receipt was associated with an approximately 30% to 40% relative increase in risk of subsequent substance-related morbidity in multiple designs that adjusted for confounding. These findings suggest that this increase may be smaller than previously estimated in some other studies.
Subject(s)
Analgesics, Opioid/administration & dosage , Drug Prescriptions/standards , Substance-Related Disorders/diagnosis , Adolescent , Adolescent Behavior/psychology , Analgesics, Opioid/therapeutic use , Cohort Studies , Comorbidity , Drug Prescriptions/statistics & numerical data , Female , Humans , Male , Retrospective Studies , Substance-Related Disorders/epidemiology , Sweden , Young AdultABSTRACT
College students and their friends become more similar in weight status over time. However, it is unclear which mediators explain this relationship. Using validated survey measures of diet, physical activity, alcohol intake, sleep behaviors, mental health, and food security status, we take a comprehensive look at possible factors associated with excess weight gain that may explain friends' convergence on body mass index (BMI), waist circumference, waist to hip ratio, and waist to height ratio over time. We use linear mixed models applied to a longitudinal dataset of first-year college students to examine whether these variables satisfy two criteria for potential candidate mediators of friends' influence on anthropometrics-cross-sectional similarity among friends (n = 509) and longitudinal associations with increasing anthropometrics (n = 428). While friends were similar on some survey measures (such as dining hall use, home cooked meal consumption, fruit intake, alcohol intake, hours of sleep, and stress). Only dining hall use and stress emerged as potential explanations for why friends' BMI and anthropometric change may be similar. Given that only a few variables satisfied the two criteria as potential mediators, future research may need to consider alternative measurement approaches, including real-time assessments, objective measurements, and alternative factors causing the convergence of friends' and college students' body size over time.
Subject(s)
Friends , Health Behavior , Life Style , Students/psychology , Weight Gain , Body Mass Index , Cross-Sectional Studies , Diet , Exercise , Female , Food Supply , Humans , Male , Mental Health , Sleep , Stress, Psychological/psychology , Time FactorsABSTRACT
Growth curve modeling is one of the main analytical approaches to study change over time. Growth curve models are commonly estimated in the linear and nonlinear mixed-effects modeling framework in which both the mean and person-specific curves are modeled parametrically with functions of time such as the linear, quadratic, and exponential. However, when more complex nonlinear trajectories need to be estimated and researchers do not have a priori knowledge of an appropriate functional form of growth, parametric models may be too restrictive. This paper reviews functional mixed-effects models, a nonparametric extension of mixed-effects models that permit both the mean and person-specific curves to be estimated without assuming a prespecified functional form of growth. Details of the model are presented along with results from a simulation study and an empirical example. The simulation study showed functional mixed-effects models performed reasonably well under various conditions commonly associated with longitudinal panel data, such as few time points per person, irregularly spaced time points across persons, missingness, and nonlinear trajectories. The usefulness of functional mixed-effects models is illustrated by analyzing empirical data from the Early Childhood Longitudinal Study - Kindergarten Class of 1998-1999.
Subject(s)
Algorithms , Linear Models , Models, Statistical , Child , Computer Simulation , Humans , Longitudinal StudiesABSTRACT
This didactic article aims to provide a gentle introduction to penalized splines as a way of estimating nonlinear growth curves in which many observations are collected over time on a single or multiple individuals. We begin by presenting piecewise linear models in which the time domain of the data is divided into consecutive phases and a separate linear regression line is fitted in each phase. Linear splines add the feature that the regression lines fitted in adjacent phases are always joined at the boundary so there is no discontinuity in level between phases. Splines are highly flexible raising the fundamental tradeoff between model fit and smoothness of the curve. Penalized spline models address this tradeoff by introducing a penalty term to achieve balance between fit and smoothness. The linear mixed-effects model, familiar from multilevel analysis, is introduced as a method for estimating penalized spline models. Higher order spline models using quadratic or cubic functions which further enhance a smooth fit are introduced. Technical issues in estimation, hypothesis testing, and constructing confidence intervals for higher order penalized spline models are considered. We then use data from the Early Childhood Longitudinal Study to illustrate each step in fitting a higher order penalized spline model, and to illustrate hypothesis testing, the construction of confidence intervals, and the comparison of the functions in 2 groups (boys and girls). Extensive graphical illustrations are provided throughout. Annotated computer scripts using the R package nlme are provided in online supplemental materials. (PsycINFO Database Record (c) 2019 APA, all rights reserved).