The intermediate endpoint effect in logistic and probit regression.

BACKGROUND: An intermediate endpoint is hypothesized to be in the middle of the causal sequence relating an independent variable to a dependent variable. The intermediate variable is also called a surrogate or mediating variable and the corresponding effect is called the mediated, surrogate endpoint, or intermediate endpoint effect. Clinical studies are often designed to change an intermediate or surrogate endpoint and through this intermediate change influence the ultimate endpoint. In many intermediate endpoint clinical studies the dependent variable is binary, and logistic or probit regression is used. PURPOSE: The purpose of this study is to describe a limitation of a widely used approach to assessing intermediate endpoint effects and to propose an alternative method, based on products of coefficients, that yields more accurate results. METHODS: The intermediate endpoint model for a binary outcome is described for a true binary outcome and for a dichotomization of a latent continuous outcome. Plots of true values and a simulation study are used to evaluate the different methods. RESULTS: Distorted estimates of the intermediate endpoint effect and incorrect conclusions can result from the application of widely used methods to assess the intermediate endpoint effect. The same problem occurs for the proportion of an effect explained by an intermediate endpoint, which has been suggested as a useful measure for identifying intermediate endpoints. A solution to this problem is given based on the relationship between latent variable modeling and logistic or probit regression. LIMITATIONS: More complicated intermediate variable models are not addressed in the study, although the methods described in the article can be extended to these more complicated models. CONCLUSIONS: Researchers are encouraged to use an intermediate endpoint method based on the product of regression coefficients. A common method based on difference in coefficient methods can lead to distorted conclusions regarding the intermediate effect.

Mediating mechanisms in a program to reduce intentions to use anabolic steroids and improve exercise self-efficacy and dietary behavior.

This study investigated the mediating mechanisms responsible for the effects of a program designed to reduce intentions to use anabolic steroids, improve nutrition, and increase strength training self-efficacy. Fifteen of 31 high school football teams (N = 1,506 players at baseline) in Oregon and Washington were assigned to receive the intervention. The multicomponent program addressed the social influences promoting ergogenic drug use and engaging students in healthy nutrition and strength training alternative behaviors. Although the results differed across the three dependent variables, the program appeared to work by changing team norms. Unlike prevention of other drugs, changes in knowledge and perceived severity were mediators of program effects in this study.

Longitudinal relationship between the alcohol warning label and alcohol consumption.

OBJECTIVE: The present study reports the longitudinal relationship between alcohol warning label exposure and alcohol consumption among adolescents. METHOD: In-school surveys assessed adolescents at three annual time points beginning in tenth grade. There were 649 participants measured at all three time points; 51% were female. Four effects were the focus of the covariance structure, latent growth analysis of these data: (1) an exposure effect, whereby earlier alcohol use leads to more exposure to the warning label; (2) a deterrent effect, whereby earlier alcohol warning label exposure reduces subsequent alcohol consumption; (3) a harmful effect, corresponding to a positive relationship between early exposure and subsequent consumption; and (4) both exposure and deterrent effects operating at the same time. RESULTS: There was a statistically significant exposure effect such that earlier alcohol use was associated with later exposure to the warning. The association between earlier alcohol warning label exposure and subsequent alcohol use was generally negative but not statistically significant, suggesting that the warning did not affect alcohol consumption. CONCLUSIONS: These results suggest that the alcohol warning label is having the intended effect as described in the law requiring the warning. That is, it is informing and reminding persons of the risks associated with alcohol use. The warning does not appear to significantly increase or decrease alcohol consumption.

Multilevel Modeling of Individual and Group Level Mediated Effects.

This article combines procedures for single-level mediational analysis with multilevel modeling techniques in order to appropriately test mediational effects in clustered data. A simulation study compared the performance of these multilevel mediational models with that of single-level mediational models in clustered data with individual- or group-level initial independent variables, individual- or group-level mediators, and individual level outcomes. The standard errors of mediated effects from the multilevel solution were generally accurate, while those from the single-level procedure were downwardly biased, often by 20% or more. The multilevel advantage was greatest in those situations involving group-level variables, larger group sizes, and higher intraclass correlations in mediator and outcome variables. Multilevel mediational modeling methods were also applied to data from a preventive intervention designed to reduce intentions to use steroids among players on high school football teams. This example illustrates differences between single-level and multilevel mediational modeling in real-world clustered data and shows how the multilevel technique may lead to more accurate results.

Home smoking restrictions and adolescent smoking.

The prevention of adolescent smoking has focused on peer influences to the relative neglect of parental influences. Parents socialize their children about many behaviors including smoking, and parental rules about their child's smoking have been related to lower levels of adolescent smoking. Moreover, among adults, indoor smoking restrictions have been associated with decreased smoking. Accordingly, the current study tested the relation of adolescent smoking to home smoking policy (rules regulating where adults are allowed to smoke in the home). Results showed that restrictive home smoking policies were associated with lower likelihood of trying smoking for both middle and high school students. However, for high school students this relation was restricted to homes with non-smoking parents. Home smoking policies were not associated with current regular smoking for either middle or high school students. Home smoking policies may be useful in preventing adolescent smoking experimentation, although longitudinal and experimental research is necessary to confirm this hypothesis.

The alcohol warning and adolescents: 5-year effects.

OBJECTIVES: This study, a follow-up to the authors earlier report, examined the effects of the alcohol warning label on adolescents during the first 5 years that the warning was required. METHODS: Surveys were administered to 10th-grade (n = 16,661) and 12th-grade (n = 15,856) students from the 1989-1990 school year through the 1994-1995 school year. The measures were awareness of, exposure to, and recognition memory of the alcohol warning label; beliefs about the risks listed on the warning; and open-ended statements about consequences of alcohol use, alcohol consumption, and self-reported driving after drinking. RESULTS: There were increases in warning awareness, exposure, and recognition memory. These effects leveled off approximately 3.5 years after the inclusion of the warning on alcohol beverage containers. There was no beneficial change attributable to the warning in beliefs, alcohol consumption, or driving after drinking. CONCLUSIONS: The initial positive effects of the alcohol warning label on adolescents have leveled off, consistent with theories of repeated exposure to persuasive information. The alcohol warning has not affected adolescents' beliefs about alcohol or alcohol-related behaviors.

The adolescents training and learning to avoid steroids program: preventing drug use and promoting health behaviors.

BACKGROUND: Use of alcohol and other illicit drugs by adolescent male athletes is a significant problem. Participation in sports may encourage use of drugs that enhance athletic performance, especially anabolic steroids (AS). Because, to our knowledge, no other intervention has successfully altered substance abuse by athletes, we developed and assessed the efficacy of a team-centered, sex-specific education program designed to reduce adolescent athletes' intentions to use and use of AS and alcohol and other illicit drugs. METHODS: We studied 31 high school football teams that comprised 3207 athletes in 3 successive annual cohorts (1994-1996). The intervention included interactive classroom and exercise training sessions given by peer educators and facilitated by coaches and strength trainers. Program content included discussion of sports nutrition, exercise alternatives to AS and sport supplements, and the effects of substance abuse in sports, drug refusal role-playing, and the creation of health promotion messages. Questionnaires assessing AS, the use of sport supplements and alcohol and other illicit drugs, and potential risk and protective factors were administered before and after the intervention (before and after the football season) and up to 1 year after the program. RESULTS: At season's end, intentions to use (P<.05) and actual AS use (P<.04) were significantly lower among students who participated in the study. Although AS reduction did not achieve significance at 1 year (P<.08), intentions to use AS remained lower (P = .02). Illicit drug use (marijuana, amphetamines, and narcotics) was reduced at 1 year, whether alcohol was included (P = .04) or excluded (P = .02) from the index. Other long-term effects included fewer students reporting drinking and driving (P = .004), less sport supplement use (P = .009), and improved nutrition behaviors (P<.02). CONCLUSIONS: Use of alcohol and other illicit drugs and associated harmful activities can be prevented with a sex-specific, team-centered education. School athletic teams provide an optimal environment in which to provide drug prevention and health promotion education.

Equivalence of the mediation, confounding and suppression effect.

This paper describes the statistical similarities among mediation, confounding, and suppression. Each is quantified by measuring the change in the relationship between an independent and a dependent variable after adding a third variable to the analysis. Mediation and confounding are identical statistically and can be distinguished only on conceptual grounds. Methods to determine the confidence intervals for confounding and suppression effects are proposed based on methods developed for mediated effects. Although the statistical estimation of effects and standard errors is the same, there are important conceptual differences among the three types of effects.

Multilevel mediation modeling in group-based intervention studies.

This article proposes and evaluates a method to test for mediation in multilevel data sets formed when an intervention administered to intact groups is designed to produce change in individual mediator and outcome variables. Simulated data of this form were used to compare ordinary least squares (OLS) and two multilevel estimators of the mediated effect. OLS and multilevel standard error approximations were also evaluated and recommendations given for optimal estimator choice. These methods were applied to data from an existing substance use intervention to show the impact multilevel mediation modeling can have on the conclusions drawn from real-world evaluation studies.

Designated driving among college students.

OBJECTIVE: We studied the frequency of designated driver use, predictors of designated driver use, whether designated drivers use alcohol, and barriers to using a designated driver. METHOD: Undergraduates at Arizona State University (N = 544) completed a questionnaire assessing designated driver use. The mean age of participants was 20.6 years old and 52% were male. Criterion for inclusion in the study was that participants had consumed alcohol at least one time. RESULTS: Eighty-six percent (86%) of participants had used a designated driver. Monthly alcohol use and driving to drinking destinations were positively associated with designated driver use. Ninety-four percent of participants indicated that their designated driver occasionally consumed alcohol. The most frequent reason for not using a designated driver was, "I stayed where I was drinking until I was sober" (37%). Less alcohol use among friends was associated with sounder methods of designated driver selection. CONCLUSIONS: Findings suggest those groups at risk for alcohol-related driving accidents tend to use designated drivers at high rates, but these groups tend to choose designated drivers who consume alcohol.

Effects of a community-based prevention program on decreasing drug use in high-risk adolescents.

OBJECTIVES: This study investigated the secondary prevention effects of a substance abuse primary prevention program. METHODS: Logistic regression analyses were conducted on 4 waves of follow-up data from sixth- and seventh-grade baseline users of cigarettes, alcohol, and marijuana taking part in a school-based program in Indianapolis. RESULTS: The program demonstrated significant reductions in cigarette use at the initial follow-up (6 months) and alcohol use at the first 2 follow-ups (up to 1.5 years). Models considering repeated measures also showed effects on all 3 substances. CONCLUSIONS: Primary prevention programs are able to reach and influence high-risk adolescents in a nonstigmatizing manner.

Concepts of drugs: differences in conceptual structure across groups with different levels of drug experience.

Seventy-seven college students varying in degree of drug use experience rated the perceived similarities of all possible combinations of 16 drugs classes (cigarettes, other tobacco, alcohol, marijuana, barbiturates, minor and major tranquilizers, amphetamines, amphetamine derivatives, cocaine, heroin, opiates, hallucinogens, inhalants, PCP, anti-depressants). Multi-dimensional scaling (INDSCAL) and network models (PFNET) indicated that abstainers had only one pharmacological category involving sedatives/depressants, and that they attached more importance to whether drugs were licit vs. illicit than to whether they were depressants vs. stimulants. Conceptions became more coherent, differentiated and based on pharmacological properties for more experienced drug users. In line with previous work, groups with greater experience with drugs had more sophisticated conceptions not only about the drugs they had used, but also about drugs they had not used. These findings suggest that early on in drug behavior sophisticated and interrelated concepts are developing that should be taken into account when designing interventions and information campaigns.

Injuries and use of protective equipment among college in-line skaters.

In-line skating injuries and protective gear use were explored in a sample of college students (n = 217). A minority of respondents wore protective gear. One third of skaters had experienced at least one minor injury, and a smaller percentage had experienced fractures or head injuries. Most minor injuries occurred during the first 1-2 times skating, while more serious injuries tended to occur after at least 50 times on in-line skates. Psychosocial predictors of protective gear use were explored. Four major Health Belief Model constructs (perceived barriers to wearing gear, perceived susceptibility to injury, perceived severity of injury, and perceived benefits of wearing gear) were significant predictors of protective gear use. The Health Belief Model, tested using regression and structural equation modelling, predicted gear typically worn, frequency of gear use, and injuries received while in-line skating. Implications for increasing protective gear use are described.

Effects of a multidimensional anabolic steroid prevention intervention. The Adolescents Training and Learning to Avoid Steroids (ATLAS) Program.

OBJECTIVE: To test a team-based, educational intervention designed to reduce adolescent athletes' intent to use anabolic androgenic steroids (AAS). DESIGN: Randomized prospective trial. SETTING: Thirty-one high school football teams in the Portland, Ore, area. PARTICIPANTS: Seven hundred two adolescent football players at experimental schools; 804 players at control schools. INTERVENTION: Seven weekly, 50-minute class sessions were delivered by coaches and student team leaders, addressing AAS effects, sports nutrition and strength-training alternatives to AAS use, drug refusal role play, and anti-AAS media messages. Seven weight-room sessions were taught by research staff. Parents received written information and were invited to a discussion session. MAIN OUTCOME MEASURES: Questionnaires before and after intervention and at 9- or 12-month follow-up, assessing AAS use risk factors, knowledge and attitudes concerning AAS, sports nutrition and exercise knowledge and behaviors, and intentions to use AAS. RESULTS: Compared with controls, experimental subjects at the long-term follow-up had increased understanding of AAS effects, greater belief in personal vulnerability to the adverse consequences of AAS, improved drug refusal skills, less belief in AAS-promoting media messages, increased belief in the team as an information source, improved perception of athletic abilities and strength-training self-efficacy, improved nutrition and exercise behaviors, and reduced intentions to use AAS. Many other beneficial program effects remained significant at the long-term follow-up. CONCLUSIONS: This AAS prevention program enhanced healthy behaviors, reduced factors that encourage AAS use, and lowered intent to use AAS. These changes were sustained over the period of 1 year. Team-based interventions appear to be an effective approach to improve adolescent behaviors and reduce drug use risk factors.

The Adolescents Training and Learning to Avoid Steroids (ATLAS) prevention program. Background and results of a model intervention.

OBJECTIVE: To develop and test a school-based intervention to prevent anabolic androgenic steroid use among high-risk adolescent athletes. DESIGN: Nonrandom controlled trial. SETTING: Two urban high schools. PARTICIPANTS: Fifty-six adolescent football players at the experimental school and 24 players at the control school. INTERVENTION: Eight weekly, 1-hour classroom sessions delivered by the coach and adolescent team leaders, and eight weight-room sessions delivered by research staff. The intervention addressed sports nutrition and strength training as alternatives to steroid use, drug refusal role play, and antisteroid media campaigns. OUTCOME MEASURES: A preintervention and postintervention questionnaire that assessed attitudes toward and intent to use steroids and other drugs; knowledge of drug effects; and diet, exercise, and related constructs. RESULTS: Compared with controls, experimental subjects were significantly less interested in trying steroids after the intervention, were less likely to want to use them even if their friends used them, were less likely to believe steroid use was a good idea, believed steroids were more dangerous, had better knowledge of alternatives to steroid use, had improved body image, increased their knowledge of diet supplements, and had less belief in these supplements as beneficial. CONCLUSIONS: Significant beneficial effects were found despite the sample size, suggesting that the effects of the intervention was large. This outcome trial demonstrates an effective anabolic androgenic steroid prevention program for adolescent athletes, and the potential of team-based interventions to enhance adolescents' health.

Maximizing the Usefulness of Data Obtained with Planned Missing Value Patterns: An Application of Maximum Likelihood Procedures.

Researchers often face a dilemma: Should they collect little data and emphasize quality, or much data at the expense of quality? The utility of the 3-form design coupled with maximum likelihood methods for estimation of missing values was evaluated. In 3-form design surveys, four sets of items. X, A, B, and C are administered: Each third of the subjects receives X and one combination of two other item sets - AB, BC, or AC. Variances and covariances were estimated with pairwise deletion, mean replacement, single imputation, multiple imputation, raw data maximum likelihood, multiple-group covariance structure modeling, and Expectation-Maximization (EM) algorithm estimation. The simulation demonstrated that maximum likelihood estimation and multiple imputation methods produce the most efficient and least biased estimates of variances and covariances for normally distributed and slightly skewed data when data are missing completely at random (MCAR). Pairwise deletion provided equally unbiased estimates but was less efficient than ML procedures. Further simulation results demonstrated that nun-maximum likelihood methods break down when data are not missing completely at random. Application of these methods with empirical drug use data resulted in similar covariance matrices for pairwise and EM estimation, however, ML estimation produced better and more efficient regression estimates. Maximum likelihood estimation or multiple imputation procedures. which are now becoming more readily available, are always recommended. In order to maximize the efficiency of the ML parameter estimates, it is recommended that scale items be split across forms rather than being left intact within forms.

Youth beliefs and knowledge about the risks of drinking while pregnant.

Because no published studies of young persons' knowledge and awareness of fetal alcohol syndrome are available, the awareness and beliefs about drinking while pregnant in several large samples of young persons ages 13-20 are examined. Approximately 81 percent of the entire sample that completed questionnaires in school surveys believe that drinking alcohol while pregnant can definitely harm the fetus, although males and younger persons are less likely to believe in this risk. A substantial proportion of respondents believe that occasional heavy use is not harmful and suggest a safe level of drinking that is higher than the Surgeon General's abstinence recommendations. Only 72 percent have heard of fetal alcohol syndrome, and more than one-third incorrectly report that it describes a baby born addicted to alcohol, that the syndrome can be inherited, and that it can be cured. As in prior studies of adults, beliefs about drinking while pregnant are inconsistent with the Surgeon General's recommendations. Implications for increasing the awareness of the risk of drinking while pregnant are discussed.

The risk of assaultive violence and alcohol availability in Los Angeles County.

OBJECTIVES: Although there is considerable evidence that alcohol consumption facilitates assaultive violence, the extent to which alcohol outlets in a community influence assaultive violence remains controversial. METHODS: To assess the geographic association between city-specific rates of assaultive violence and alcohol-outlet density, an ecologic analysis of the 74 larger cities in Los Angeles County was conducted for the 1990 reporting year. RESULTS: Sociodemographic factors alone accounted for 70% (R2 = .70) of the variance in the rate of assaultive violence in a multiple regression model. Adding the variable for alcohol-outlet density to the model yielded a significant positive slope. The magnitude of this relation indicates that in a typical Los Angeles County city (50,000 residents, 100 outlets, 570 offenses per year), one outlet was associated with 3.4 additional assaultive violence offenses in 1990. CONCLUSIONS: These findings indicate that higher levels of alcohol-outlet density are geographically associated with higher rates of assaultive violence. This association is independent of measured confounders, including city-level measures of unemployment, ethnic/racial makeup, income, age structure, city size, household size, and female-headed households.

Development and applications of a city-level alcohol availability and alcohol problems database.

Data on alcohol availability and problems in all cities in Los Angeles County were collected from several different sources and linked together to form a Local Alcohol Availability Database (LAAD). The two major purposes of the project are to provide a city-level alcohol availability and alcohol-related problems database needed by local community alcohol policy planners and to collect the data necessary for research on the relationship between these measures. The prevalence of drunk driving arrests is displayed on a map. We describe how the LAAD has been used to guide alcohol policy decisions. A fixed year and city effects regression model suggests that outlet density is positively related to several alcohol-related problems.

Alcohol outlet density and motor vehicle crashes in Los Angeles County cities.

The extent to which the availability of alcohol encourages alcohol consumption resulting in alcohol-related problems remains controversial. In order to address this issue we used 1990 data from 72 cities within Los Angeles County to estimate the relation between densities of four types of alcohol outlets (restaurants, bars, liquor stores, mini-markets) and rates of two types of alcohol-related motor vehicle crashes reported by police (injury, property damage). After logarithmic transformation of variables, crash rates were regressed on outlet densities and possible demographic confounders. The demographic covariates accounted for 25% of the variance in injury crashes; adding the combined outlet density to the model yielded a significant elasticity estimate (beta +/- SE) (beta = 0.55 +/- 0.13). Separate models for types of outlets yielded significant elasticities for restaurants (beta = 0.22 +/- 0.07), liquor stores (beta = 0.46 +/- 0.17) and mini-markets (beta = 0.34 +/- 0.13), but not for bars (beta = 0.08 +/- 0.07). Alcohol-related crashes resulting in property damage also showed positive associations with outlet densities, but these associations were smaller and reached statistical significance for restaurants (beta = 0.19 +/- 0.11) and bars (beta = 0.21 +/- 0.10). Direction of influence cannot be inferred from these cross-sectional findings, but they do indicate that increased alcohol availability is geographically associated with increased alcohol-related motor vehicle crashes and that this association is independent of measured confounders.