The relative benefits of meta-analysis conducted with individual participant data versus aggregated data.

The authors describe the relative benefits of conducting meta-analyses with (a) individual participant data (IPD) gathered from the constituent studies and (b) aggregated data (AD), or the group-level statistics (in particular, effect sizes) that appear in reports of a study’s results. Given that both IPD and AD are equally available, meta-analysis of IPD is superior to meta-analysis of AD. IPD meta-analysis permits synthesists to perform subgroup analyses not conducted by the original collectors of the data, to check the data and analyses in the original studies, to add new information to the data sets, and to use different statistical methods. However, the cost of IPD meta-analysis and the lack of available IPD data sets suggest that the best strategy currently available is to use both approaches in a complementary fashion such that the first step in conducting an IPD meta-analysis would be to conduct an AD meta-analysis. Regardless of whether a meta-analysis is conducted with IPD or AD, synthesists must remain vigilant in how they interpret their results. They must avoid ecological fallacies, Simpson’s paradox, and interpretation of synthesis-generated evidence as supporting causal inferences. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Obstetric complications and age at onset in schizophrenia: an international collaborative meta-analysis of individual patient data

OBJECTIVE: An excess of obstetric complications in the histories of schizophrenic patients is a well-replicated finding, but less consistent results have been found concerning the relationships between obstetric complications and family history of schizophrenia, age at onset of schizophrenia, and gender. Small sample size limited the power of previous studies that attempted to assess such relationships. The aim of this study was to use data on individual patients from all available studies to examine the links between a history of obstetric complications and family history of schizophrenia, age at onset, and gender. METHOD: Raw data from 854 schizophrenic patients concerning history of obstetric complications rated according to the Lewis and Murray scale were obtained from 11 different research groups. Weighted average estimates were calculated with the use of regression techniques. RESULTS: A significant association was found between age at onset of schizophrenia and obstetric complications: the earlier the age at onset, the more likely the history of obstetric complications. Subjects with onset of schizophrenia before age 22 were 2.7 times more likely than those with onset at a later age to have had a history of abnormal presentation at birth and 10 times more likely to have had a history of complicated Cesarean birth. No association was found between obstetric complications and family history of schizophrenia or gender. CONCLUSIONS: The association between obstetric complications and early age at onset of schizophrenia indicates that the pathophysiology of early-onset schizophrenia involves neurodevelopmental impairment.     

An analysis of group versus individual child health supervision

This study compares the effectiveness of group health supervision (three or four families counseled simultaneously) with traditional visits in conveying knowledge of child health and development, increasing perceived maternal support, and mitigating maternal depression. Subjects were recruited from a predominantly white, middle-class, suburban/rural pediatric practice. Twenty-five families were allocated to group health supervision and 25 to individual visits. A questionnaire covering knowledge of child health and development (CHDQ), the Maternal Social Support Index (MSSI), and the Center for Epidemiologic Studies Depression Scale (CESD) were administered to both groups before their 2-month and after their 10-month visits. A subset of these charts was reviewed for problem visits between 2 and 6 months. As compared with families having traditional visits, families who received the group intervention did at least as well in acquiring knowledge of child care and development and, although not statistically significant, tended to recover from postpartum depression faster and deal better with minor illnesses. The investigators found group child health supervision to be a pleasant and effective method of health care delivery.     

Correcting effect sizes computed from factor analysis of variance for use in meta-analysis.

In meta-analysis, it is possible to average results across multiple studies because the effect sizes estimated from each study are in the same metric (e.g., the standardized mean difference). However, when effect sizes are computed from a factorial analysis of variance, these estimates are influenced by the other factors in the design. A correction developed by G. V. Glass, B. McGaw, and M. L. Smith (1981) solves this problem; however, it requires information (e.g., sums of squares) that is often not available in published research. A reformulated version of the correction is presented, which requires only F values and degrees of freedom. The impact of the correction on effect size estimates is examined. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comparison of two strategies for analysis of variance in nonorthogonal designs.

Compared the M. I. Appelbaum and E. M. Cramer (see record 1974-28956-001) comparison of models strategy for analysis of data from nonorthogonal designs with the J. E. Overall and D. K. Spiegal (see record 1970-01534-001) Method 1 general linear model analysis. Data were generated by Monte Carlo methods to include known true ANOVA main and interaction effects. In the presence of a true but nonsignificant interaction, estimates of main effect parameters derived from the Method 1 general linear model analysis were significantly closer to the true values. Greater accuracy in estimation of main effects in the presence of a significant interaction was also observed. The danger of letting observed data determine the ANOVA model and the hypotheses to be tested is emphasized. (12 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Heterogeneity of variance in clinical data.

Traditional parametric (t, F) and nonparametric (Mann-Whitney-Wilcoxon U, Kruskal-Wallis H) statistics are sensitive to heterogeneity of variance (heteroscedasticity). Moreover, there are theoretical reasons to expect, and empirical results to document, the existence of heteroscedasticity in clinical data. Transformations to reduce heteroscedasticity are problematic. This article reviews the literature on robust methods that are available and that should be widely used to control rate of Type I error and maintain power. No one robust method is ideal for all situations, but such methods are superior to the traditional tests. Specific recommendations are made for application under various conditions of heteroscedasticity. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Group versus individual problem-solving.

This investigation was proposed to answer the question: "Is the performance of a group superior to the performance of an individual?" Using problems involving spatial and verbal variables, the research findings indicated that group solutions were superior to individual. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Multivariate analysis of variance.

We provide an expository presentation of multivariate analysis of variance (MANOVA) for both consumers of research and investigators by capitalizing on its relation to univariate analysis of variance models. We address several questions: (a) Why should one use MANOVA? (b) What is the structure of MANOVA? (c) How are MANOVA test statistics obtained and interpreted? (d) How are MANOVA follow-up tests obtained and interpreted? (e) How is strength of association assessed in MANOVA? (f) How should the results of MANOVA be presented? (g) Are there any alternatives to MANOVA? We use an example data set throughout the article to illustrate these points. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Osteoporosis and the individual patient

Effectiveness of applications of saprol and nonuniform constant magnetic field (NUCMF) used alone and in combination was studied on the model of experimental toxic rat hepatitis. NUCMF had no advantages over saprol in correction of hepatic tissue function. However, combination of these two modalities was more effective than each of them. Possible mechanisms of such effect are discussed.     

Group versus individual assertion training.

16 nonassertive community women (mean age 39.1 yrs) participated in a study that compared the effectiveness of group assertion therapy with the effectiveness of individual assertion therapy. A clinical treatment procedure that included behavioral rehearsal, modeling, and coaching was used in both treatment modalities. Prior research using almost identical treatment procedures, screening requirements, and assessment batteries demonstrated the greater effectiveness of this therapeutic procedure as compared with a no-treatment control condition, in which Ss did not change over time. Results of this study indicate that there are no significant differences between group versus individual assertion training. Pre-, post-, and 3-mo follow-up measures (e.g., Rathus Assertiveness Schedule, S–R Inventory of General Trait Anxiousness) demonstrated that both treatments were effective in increasing assertive behavior skills and in reducing hostility and anger. When compared with the no-treatment control condition of an earlier comparable study, the 2 treatment conditions of the present study are shown to be superior. (6 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Collective versus individual induction with single versus multiple hypotheses.

Four-person cooperative groups and 4 independent individuals solved rule induction problems by proposing 1, 2, or 4 hypotheses per trial while selecting the same amount of evidence per trial. Groups performed at the level of the best individuals and better than the 2nd-best, 3rd-best, and 4th-best individuals for both correct hypotheses and nonplausible hypotheses. Increasing the number of proposed hypotheses from 1 to 2 to 4 did not increase correct hypotheses but increased nonplausible hypotheses. Transition probabilities from hypotheses on trial t to t?+?1 indicated superior performance for the groups and best individuals for each of positive and negative hypothesis tests followed by examples and nonexamples; once the groups and best individuals proposed the correct hypothesis, they were more likely to continue to propose it than the other individuals. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Nonorthogonal two-way analysis of variance.

Examined the advantage of disadvantages of 5 different exact analyses of variance for nonorthogonal 2-way designs with respect to orthogonality of the analyses, parametric hypotheses tested, and model comparisons made by the analyses. It is proposed that experimenters, when faced with the necessity of performing a 2-way ANOVA, carefully consider these analyses with regard to the a priori information they have about the data, the questions they expect the analysis to help answer, and the questions each analysis is best equipped to answer. It is also suggested that experimenters choose the analysis that best fits their needs rather than depend on one for all situations. (16 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

An individual patient-based meta-analysis of tamoxifen versus ovarian ablation as first line endocrine therapy for premenopausal women with metastatic breast cancer

Total mercury concentration in the muscle of 417 fish of 12 common freshwater and four anadromous species from Bangladesh were low, varying from 2 to 430 ng/g fresh wt. Depending on Hg speciation, three types of accumulation mechanisms were defined. Type I covers the majority of species and describes a pattern widely accepted as 'normal', with increasing levels of organic (methyl) mercury with length (age), combined to a low and constant inorganic level. This accumulation pattern leads to a relative increase of the organic mercury fraction with age, eventually reaching 90-100% of organic mercury in full grown specimens. Type II is found in both planktivorous genera only and showed increasing levels of inorganic mercury combined to low and constant organic mercury levels, leading to a relative decrease in organic mercury fraction with age. This unexpected pattern was only reported in cases of some marine species where it seemed to be linked to demethylation mechanisms or regional influences on Hg levels. A third intermediate accumulation pattern with increasing concentrations of both the organic and the inorganic Hg fraction with age was found in one bottom dwelling species only. The implications of these observations for the accumulation mechanisms of mercury in fish are discussed.     

Prediction of residual retroperitoneal mass histology after chemotherapy for metastatic nonseminomatous germ cell tumor: multivariate analysis of individual patient data from six study groups

PURPOSE: To develop a statistical model that predicts the histology (necrosis, mature teratoma, or cancer) after chemotherapy for metastatic nonseminomatous germ cell tumor (NSGCT). PATIENTS AND METHODS: An international data set was collected comprising individual patient data from six study groups. Logistic regression analysis was used to estimate the probability of necrosis and the ratio of cancer and mature teratoma. RESULTS: Of 556 patients, 250 (45%) had necrosis at resection, 236 (42%) had mature teratoma, and 70 (13%) had cancer. Predictors of necrosis were the absence of teratoma elements in the primary tumor, prechemotherapy normal alfa-fetoprotein (AFP), normal human chorionic gonadotropin (HCG), and elevated lactate dehydrogenase (LDH) levels, a small prechemotherapy or postchemotherapy mass, and a large shrinkage of the mass during chemotherapy. Multivariate combination of predictors yielded reliable models (goodness-of-fit tests, P > .20), which discriminated necrosis well from other histologies (area under the receiver operating characteristic (ROC) curve, .84), but which discriminated cancer only reasonably from mature teratoma (area, .66). Internal and external validation confirmed these findings. CONCLUSION: The validated models estimate with high accuracy the histology at resection, especially necrosis, based on well-known and readily available predictors. The predicted probabilities may help to choose between immediate resection of a residual mass or follow-up, taking into account the expected benefits and risks of resection, feasibility of frequent follow-up, the financial costs, and the patient's individual preferences.     

Work-group versus individual differences in attitude.

To determine the relative magnitude of group and individual differences in job attitudes, responses to a 20-item Likert-type attitude scale were obtained from 190 employees, sampled from 38 work groups in 3 manufacturing plants. The hypothesis that the work groups did not differ in job attitudes was tested by an analysis of variance. The observed work-group differences in attitudes were not significant, and the lowest and highest work-group means in each of the 3 plants were not significantly far apart. In these 3 plants, at least, the more appropriate unit for administrative action or for research study on employee attitudes seems to be the individual, not the work group. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A note on variance explained in the mixed analysis of variance models.

Examines assumptions about the general linear model for interaction terms in the mixed analysis of variance. Some well-known results of S. R. Searle (1971) demonstrate that the inconsistencies between J. H. Dwyer's (see record 1975-02166-001) technique and that of G. M. Vaughn and M. C. Corballis (see record 1969-16617-001) in estimating the magnitude of effect for a mixed interaction are the direct result of specific assumptions made. If it is assumed that the interaction source of variance is a random variable, then the equations obtained by Vaughn and Corballis are correct; however, if an alternative assumption is made (i.e., that the iteraction term is fixed in one direction), then Dwyer's equations are correct. Researchers are called on to be cognizant of these two sets of assumptions and to be aware of the dramatic effects they may have on estimates of magnitude of effect for mixed interactions. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Right versus left symmetry of ulnar variance. A radiographic assessment

One hundred skeletally mature healthy volunteers underwent standardized bilateral posteroanterior radiographs in unloaded (static) and loaded (dynamic) conditions to determine the symmetry of ulnar variance. The mean age was 32 +/- 9 years (range, 19-61 years), with 58 women and 42 men. Ulnar variance was measured to the closest 0.5 mm using the method of perpendiculars. Three separate measurements were made of each radiograph in a blinded fashion by the same investigator. An intraobserver standard deviation of 0.21 was used to calculate a 95% tolerance interval of 0.7 mm (rounded up to 1 mm) as a measure of significance. The average static ulnar variance was -0.13 +/- 1.5 mm on the left and -0.29 +/- 1.6 mm on the right. The average dynamic ulnar variance was 0.93 +/- 1.5 mm on the left and 0.82 +/- 1.5 mm on the right. When compared individually, there was a greater than or equal to 1 mm side to side difference in 37% of volunteers under static and 38% under dynamic conditions. There were no significant correlations between ulnar variance measurements and patient age, gender, race, or handedness. Use of the normal wrist radiograph as a baseline for static radial length measurements is valid in only 63% of cases.     

Calculating power in analysis of variance.

Presents an overview of procedures for calculating power of the F test under 3 models of ANOVA (fixed effects, random effects, and mixed effects). A comparison of power of tests on fixed and random factors shows the latter to have substantially lower power. Consequences for designing experiments and for interpreting experimental results are discussed, and the simplicity with which power calculations are done is emphasized. (15 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)