Reflections on the concept of optimality of single decision point treatment regimes.

In many areas, applied researchers as well as practitioners have to choose between different solutions for a problem at hand; this calls for optimal decision rules to settle the choices involved. As a key example, one may think of the search for optimal treatment regimes (OTRs) in clinical research, that specify which treatment alternative should be administered to each patient under study. Motivated by the fact that the concept of optimality of decision rules in general and treatment regimes in particular has received so far relatively little attention and discussion, we will present a number of reflections on it, starting from the basics of any optimization problem. Specifically, we will analyze the search space and the to be optimized criterion function underlying the search of single decision point OTRs, along with the many choice aspects that show up in their specification. Special attention is paid to formal characteristics and properties as well as to substantive concerns and hypotheses that may guide these choices. We illustrate with a few empirical examples taken from the literature. Finally, we discuss how the presented reflections may help sharpen statistical thinking about optimality of decision rules for treatment assignment and to facilitate the dialogue between the statistical consultant and the applied researcher in search of an OTR.

Estimating the quality of optimal treatment regimes.

When multiple treatment alternatives are available for a disease, an obvious question is which alternative is most effective for which patient. One may address this question by searching for optimal treatment regimes that specify for each individual the preferable treatment alternative based on that individual's baseline characteristics. When such a regime has been estimated, its quality (in terms of the expected outcome if it was used for treatment assignment of all patients in the population under study) is of obvious interest. Obtaining a good and reliable estimate of this quantity is a key challenge for which so far no satisfactory solution is available. In this paper, we consider for this purpose several estimators of the expected outcome in conjunction with several resampling methods. The latter have been evaluated before within the context of statistical learning to estimate the prediction error of estimated prediction rules. Yet, the results of these evaluations were equivocal, with different best performing methods in different studies, and with near-zero and even negative correlations between true and estimated prediction errors. Moreover, for different reasons, it is not straightforward to extrapolate the findings of these studies to the context of optimal treatment regimes. To address these issues, we set up a new and comprehensive simulation study. In this study, combinations of different estimators with .632+ and out-of-bag bootstrap resampling methods performed best. In addition, the study shed a surprising new light on the previously reported problematic correlations between true and estimated prediction errors in the area of statistical learning.

Studying treatment-effect heterogeneity in precision medicine through induced subgroups.

Precision medicine, in the sense of tailoring the choice of medical treatment to patients' pretreatment characteristics, is nowadays gaining a lot of attention. Preferably, this tailoring should be realized in an evidence-based way, with key evidence in this regard pertaining to subgroups of patients that respond differentially to treatment (i.e., to subgroups involved in treatment-subgroup interactions). Often a-priori hypotheses on subgroups involved in treatment-subgroup interactions are lacking or are incomplete at best. Therefore, methods are needed that can induce such subgroups from empirical data on treatment effectiveness in a post hoc manner. Recently, quite a few such methods have been developed. So far, however, there is little empirical experience in their usage. This may be problematic for medical statisticians and statistically minded medical researchers, as many (nontrivial) choices have to be made during the data-analytic process. The main purpose of this paper is to discuss the major concepts and considerations when using these methods. This discussion will be based on a systematic, conceptual, and technical analysis of the type of research questions at play, and of the type of data that the methods can handle along with the available software, and a review of available empirical evidence. We will illustrate all this with the analysis of a dataset comparing several anti-depressant treatments.

The impact of self-distancing on emotion explosiveness and accumulation: An fMRI study.

Emotions unfold over time with episodes differing in explosiveness (i.e., profiles having a steep vs. a gentle start) and accumulation (i.e., profiles increasing over time vs. going back to baseline). In the present fMRI study, we wanted to replicate and extend previous findings on the psychological and neural mechanisms underlying emotion explosiveness and accumulation. Specifically, we aimed to: (a) replicate the finding that different neural mechanisms are associated with emotion explosiveness and accumulation, (b) replicate the finding that adopting a self-distanced (vs. self-immersed) perspective decreases emotion explosiveness and accumulation at the level of self-report, and (c) examine whether adopting a self-distanced (vs. self-immersed) perspective similarly modulates activity in the brain regions associated with emotion explosiveness and accumulation. Participants in an fMRI scanner were asked to adopt a self-immersed or self-distanced perspective while reading and thinking about negative social feedback, and to report on felt changes in negative affect during that period using an emotion intensity profile tracking approach. We replicated previous findings showing that emotion explosiveness and accumulation were related to activity in regions involved in self-referential processing (such as the medial prefrontal cortex) and sustained visceral arousal (such as the posterior insula), respectively. The finding that adopting a self-distanced (vs. self-immersed) perspective lowers emotion explosiveness and accumulation was also replicated at a self-report level. However, perspective taking did not impact activity in the neural correlates of emotion explosiveness and accumulation.

The relation between rumination and temporal features of emotion intensity.

Intensity profiles of emotional experience over time have been found to differ primarily in explosiveness (i.e. whether the profile has a steep vs. a gentle start) and accumulation (i.e. whether intensity increases over time vs. goes back to baseline). However, the determinants of these temporal features remain poorly understood. In two studies, we examined whether emotion regulation strategies are predictive of the degree of explosiveness and accumulation of negative emotional episodes. Participants were asked to draw profiles reflecting changes in the intensity of emotions elicited either by negative social feedback in the lab (Study 1) or by negative events in daily life (Study 2). In addition, trait (Study 1 & 2), and state (Study 2) usage of a set of emotion regulation strategies was assessed. Multilevel analyses revealed that trait rumination (especially the brooding component) was positively associated with emotion accumulation (Study 1 & 2). State rumination was also positively associated with emotion accumulation and, to a lesser extent, with emotion explosiveness (Study 2). These results provide support for emotion regulation theories, which hypothesise that rumination is a central mechanism underlying the maintenance of negative emotions.

Comparing Four Methods for Estimating Tree-Based Treatment Regimes.

When multiple treatment alternatives are available for a certain psychological or medical problem, an important challenge is to find an optimal treatment regime, which specifies for each patient the most effective treatment alternative given his or her pattern of pretreatment characteristics. The focus of this paper is on tree-based treatment regimes, which link an optimal treatment alternative to each leaf of a tree; as such they provide an insightful representation of the decision structure underlying the regime. This paper compares the absolute and relative performance of four methods for estimating regimes of that sort (viz., Interaction Trees, Model-based Recursive Partitioning, an approach developed by Zhang et al. and Qualitative Interaction Trees) in an extensive simulation study. The evaluation criteria were, on the one hand, the expected outcome if the entire population would be subjected to the treatment regime resulting from each method under study and the proportion of clients assigned to the truly best treatment alternative, and, on the other hand, the Type I and Type II error probabilities of each method. The method of Zhang et al. was superior regarding the first two outcome measures and the Type II error probabilities, but performed worst in some conditions of the simulation study regarding Type I error probabilities.

The neural basis of emotions varies over time: different regions go with onset- and offset-bound processes underlying emotion intensity.

According to theories of emotion dynamics, emotions unfold across two phases in which different types of processes come to the fore: emotion onset and emotion offset. Differences in onset-bound processes are reflected by the degree of explosiveness or steepness of the response at onset, and differences in offset-bound processes by the degree of accumulation or intensification of the subsequent response. Whether onset- and offset-bound processes have distinctive neural correlates and, hence, whether the neural basis of emotions varies over time, still remains unknown. In the present fMRI study, we address this question using a recently developed paradigm that allows to disentangle explosiveness and accumulation. Thirty-one participants were exposed to neutral and negative social feedback, and asked to reflect on its contents. Emotional intensity while reading and thinking about the feedback was measured with an intensity profile tracking approach. Using non-negative matrix factorization, the resulting profile data were decomposed in explosiveness and accumulation components, which were subsequently entered as continuous regressors of the BOLD response. It was found that the neural basis of emotion intensity shifts as emotions unfold over time with emotion explosiveness and accumulation having distinctive neural correlates.

Principal Covariates Clusterwise Regression (PCCR): Accounting for Multicollinearity and Population Heterogeneity in Hierarchically Organized Data.

In the behavioral sciences, many research questions pertain to a regression problem in that one wants to predict a criterion on the basis of a number of predictors. Although in many cases, ordinary least squares regression will suffice, sometimes the prediction problem is more challenging, for three reasons: first, multiple highly collinear predictors can be available, making it difficult to grasp their mutual relations as well as their relations to the criterion. In that case, it may be very useful to reduce the predictors to a few summary variables, on which one regresses the criterion and which at the same time yields insight into the predictor structure. Second, the population under study may consist of a few unknown subgroups that are characterized by different regression models. Third, the obtained data are often hierarchically structured, with for instance, observations being nested into persons or participants within groups or countries. Although some methods have been developed that partially meet these challenges (i.e., principal covariates regression (PCovR), clusterwise regression (CR), and structural equation models), none of these methods adequately deals with all of them simultaneously. To fill this gap, we propose the principal covariates clusterwise regression (PCCR) method, which combines the key idea's behind PCovR (de Jong & Kiers in Chemom Intell Lab Syst 14(1-3):155-164, 1992) and CR (Späth in Computing 22(4):367-373, 1979). The PCCR method is validated by means of a simulation study and by applying it to cross-cultural data regarding satisfaction with life.

Physical activity and relaxation in the work setting to reduce the need for recovery: what works for whom?

BACKGROUND: To recover from work stress, a worksite health program aimed at improving physical activity and relaxation may be valuable. However, not every program is effective for all participants, as would be expected within a "one size fits all" approach. The effectiveness of how the program is delivered may differ across individuals. The aim of this study was to identify subgroups for whom one intervention may be better suited than another by using a new method called QUalitative INteraction Trees (QUINT). METHODS: Data were used from the "Be Active & Relax" study, in which 329 office workers participated. Two delivery modes of a worksite health program were given, a social environmental intervention (group motivational interviewing delivered by team leaders) and a physical environmental intervention (environmental modifications). The main outcome was change in Need for Recovery (NFR) from baseline to 12 month follow-up. The QUINT method was used to identify subgroups that benefitted more from either type of delivery mode, by incorporating moderator variables concerning sociodemographic, health, home, and work-related characteristics of the participants. RESULTS: The mean improvement in NFR of younger office workers in the social environmental intervention group was significantly higher than younger office workers who did not receive the social environmental intervention (10.52; 95 % CI: 4.12, 16.92). Furthermore, the mean improvement in NFR of older office workers in the social environmental intervention group was significantly lower than older office workers who did not receive the social environmental intervention ( -10.65; 95 % CI: -19.35, -1.96). The results for the physical environmental intervention indicated that the mean improvement in NFR of office workers (regardless of age) who worked fewer hours overtime was significantly higher when they had received the physical environmental intervention than when they had not received this type of intervention (7.40; 95 % CI: 0.99, 13.81). Finally, for office workers who worked more hours overtime there was no effect of the physical environmental intervention. CONCLUSIONS: The results suggest that a social environmental intervention might be more beneficial for younger workers, and a physical environmental intervention might be more beneficial for employees with a few hours overtime to reduce the NFR. TRIAL REGISTRATION: NTR2553.

Qualitative Treatment-Subgroup Interactions in a Randomized Clinical Trial of Treatments for Adolescents with ADHD: Exploring What Cognitive-Behavioral Treatment Works for Whom.

OBJECTIVE: This study explored qualitative treatment-subgroup interactions within data of a RCT with two cognitive behavioral treatments (CBT) for adolescents with ADHD: a planning-focused (PML) and a solution-focused CBT (SFT). Qualitative interactions imply that which treatment is best differs across subgroups of patients, and are therefore most relevant for personalized medicine. METHODS: Adolescents with ADHD (N = 159) received either PML or SFT. Pre-, post- and three-month follow-up data were gathered on parent-rated ADHD symptoms and planning problems. Pretreatment characteristics were explored as potential qualitative moderators of pretest to follow-up treatment effects, using an innovative analyses technique (QUINT; Dusseldorp & Van Mechelen, 2014). In addition, qualitative treatment-subgroup interactions for the therapeutic changes from pre- to posttest and from post- to follow-up test were investigated. RESULTS: For the entire time span from pretest to follow-up only a quantitative interaction was found, while from posttest to follow-up qualitative interactions were found: Adolescents with less depressive symptoms but more anxiety symptoms showed more improvement when receiving PML than SFT, while for other adolescents the effects of PML and SFT were comparable. DISCUSSION: Whereas subgroups in both treatments followed different trajectories, no subgroup was found for which SFT outperformed PML in terms of the global change in symptoms from pretest to three months after treatment. This implies that, based on this exploratory study, there is no need for personalized treatment allocation with regard to the CBTs under study for adolescents with ADHD. However, for a subgroup with comorbid anxiety symptoms but low depression PML clearly appears the treatment of preference. TRIAL REGISTRATION: Nederlands Trial Register NTR2142.

QUINT: A tool to detect qualitative treatment-subgroup interactions in randomized controlled trials.

OBJECTIVE: The detection of subgroups involved in qualitative treatment-subgroup interactions (i.e., for one subgroup of clients treatment A outperforms treatment B, whereas for another the reverse holds true) is crucial for personalized health. In typical Randomized Controlled Trials (RCTs), the combination of a lack of a priori hypotheses and a large number of possible moderators leaves current methods insufficient to detect subgroups involved in such interactions. A recently developed method, QUalitative INteraction Trees (QUINT), offers a solution. However, the paper in which QUINT has been introduced is not easily accessible for non-methodologists. In this paper, we want to review the conceptual basis of QUINT in a nontechnical way, and illustrate its relevance for psychological applications. METHOD: We present a concise introduction into QUINT along with a summary of available evidence on its performance. Subsequently, we subject RCT data on the effect of motivational interviewing in a treatment for substance abuse disorders to a reanalysis with QUINT. As outcome variables, we focus on measures of retention and substance use. RESULTS: A qualitative treatment-subgroup interaction is found for retention. By contrast, no qualitative interaction is detected for substance use. CONCLUSIONS: QUINT may lead to insightful and well-interpretable results with straightforward implications for personalized treatment assignment.

KSC-N: Clustering of Hierarchical Time Profile Data.

Quite a few studies in the behavioral sciences result in hierarchical time profile data, with a number of time profiles being measured for each person under study. Associated research questions often focus on individual differences in profile repertoire, that is, differences between persons in the number and the nature of profile shapes that show up for each person. In this paper, we introduce a new method, called KSC-N, that parsimoniously captures such differences while neatly disentangling variability in shape and amplitude. KSC-N induces a few person clusters from the data and derives for each person cluster the types of profile shape that occur most for the persons in that cluster. An algorithm for fitting KSC-N is proposed and evaluated in a simulation study. Finally, the new method is applied to emotional intensity profile data.

Not Just a Sum? Identifying Different Types of Interplay between Constituents in Combined Interventions.

MOTIVATION: Experiments in which the effect of combined manipulations is compared with the effects of their pure constituents have received a great deal of attention. Examples include the study of combination therapies and the comparison of double and single knockout model organisms. Often the effect of the combined manipulation is not a mere addition of the effects of its constituents, with quite different forms of interplay between the constituents being possible. Yet, a well-formalized taxonomy of possible forms of interplay is lacking, let alone a statistical methodology to test for their presence in empirical data. RESULTS: Starting from a taxonomy of a broad range of forms of interplay between constituents of a combined manipulation, we propose a sound statistical hypothesis testing framework to test for the presence of each particular form of interplay. We illustrate the framework with analyses of public gene expression data on the combined treatment of dendritic cells with curdlan and GM-CSF and show that these lead to valuable insights into the mode of action of the constituent treatments and their combination. AVAILABILITY AND IMPLEMENTATION: R code implementing the statistical testing procedure for microarray gene expression data is available as supplementary material. The data are available from the Gene Expression Omnibus with accession number GSE32986.

Variability in anger intensity profiles: structure and predictive basis.

The aim of this study is to describe variability in the shape and amplitude of intensity profiles of anger episodes and how it relates to duration, and to investigate whether this variability can be predicted on the basis of appraisals and emotion regulation strategies used. Participants were asked to report on a wide range of recollected anger episodes. By means of K-spectral centroid clustering, two prototypical shapes of anger intensity profiles were identified: early- and late-blooming episodes. Early-blooming episodes are relatively short and reach their peak immediately. These profiles are associated with low-importance events and adaptive regulation. Late-blooming episodes last longer and reach their peak (relatively) late in the episode. These profiles are related to high-importance events and maladaptive regulation. For both early- and late-blooming profiles, overall amplitude is positively associated with event importance and the use of maladaptive regulation strategies and negatively with the use of adaptive ones.

To share, or not to share? Examining the emotional consequences of social sharing in the case of anger and sadness.

Previous research has shown the relation between social sharing and emotional processing to be notoriously complex. In the present study, we unraveled this complexity by, for the first time, taking 3 key aspects of this relation into account simultaneously: the nature of the emotion, the timing of possible sharing effects, and the multicomponential character of emotions. Using the day reconstruction method, we first identified an intense anger or sadness target episode for each participant. In a second phase, participants repeatedly reported their sharing behavior and intensity of different emotion components over 5 days. Growth curve analyses revealed that sharing anger leads to several immediate and delayed beneficial effects, whereas sharing sadness leads to limited positive effects that emerge later on. This implies that all 3 aspects under study, as well as their interplay, are of critical importance in the relation between sharing and emotional processing.

Qualitative interaction trees: a tool to identify qualitative treatment-subgroup interactions.

When two alternative treatments (A and B) are available, some subgroup of patients may display a better outcome with treatment A than with B, whereas for another subgroup, the reverse may be true. If this is the case, a qualitative (i.e., disordinal) treatment-subgroup interaction is present. Such interactions imply that some subgroups of patients should be treated differently and are therefore most relevant for personalized medicine. In case of data from randomized clinical trials with many patient characteristics that could interact with treatment in a complex way, a suitable statistical approach to detect qualitative treatment-subgroup interactions is not yet available. As a way out, in the present paper, we propose a new method for this purpose, called QUalitative INteraction Trees (QUINT). QUINT results in a binary tree that subdivides the patients into terminal nodes on the basis of patient characteristics; these nodes are further assigned to one of three classes: a first for which A is better than B, a second for which B is better than A, and an optional third for which type of treatment makes no difference. Results of QUINT on simulated data showed satisfactory performance, with regard to optimization and recovery. Results of an application to real data suggested that, compared with other approaches, QUINT provided a more pronounced picture of the qualitative interactions that are present in the data.

Performing DISCO-SCA to search for distinctive and common information in linked data.

Behavioral researchers often obtain information about the same set of entities from different sources. A main challenge in the analysis of such data is to reveal, on the one hand, the mechanisms underlying all of the data blocks under study and, on the other hand, the mechanisms underlying a single data block or a few such blocks only (i.e., common and distinctive mechanisms, respectively). A method called DISCO-SCA has been proposed by which such mechanisms can be found. The goal of this article is to make the DISCO-SCA method more accessible, in particular for applied researchers. To this end, first we will illustrate the different steps in a DISCO-SCA analysis, with data stemming from the domain of psychiatric diagnosis. Second, we will present in this article the DISCO-SCA graphical user interface (GUI). The main benefits of the DISCO-SCA GUI are that it is easy to use, strongly facilitates the choice of model selection parameters (such as the number of mechanisms and their status as being common or distinctive), and is freely available.

Inferring protein-protein interaction complexes from immunoprecipitation data.

BACKGROUND: Protein-protein interactions in cells are widely explored using small-scale experiments. However, the search for protein complexes and their interactions in data from high throughput experiments such as immunoprecipitation is still a challenge. We present "4N", a novel method for detecting protein complexes in such data. Our method is a heuristic algorithm based on Near Neighbor Network (3N) clustering. It is written in R, it is faster than model-based methods, and has only a small number of tuning parameters. We explain the application of our new method to real immunoprecipitation results and two artificial datasets. We show that the method can infer protein complexes from protein immunoprecipitation datasets of different densities and sizes. FINDINGS: 4N was applied on the immunoprecipitation dataset that was presented by the authors of the original 3N in Cell 145:787-799, 2011. The test with our method shows that it can reproduce the original clustering results with fewer manually adapted parameters and, in addition, gives direct insight into the complex-complex interactions. We also tested 4N on the human "Tip49a/b" dataset. We conclude that 4N can handle the contaminants and can correctly infer complexes from this very dense dataset. Further tests were performed on two artificial datasets of different sizes. We proved that the method predicts the reference complexes in the two artificial datasets with high accuracy, even when the number of samples is reduced. CONCLUSIONS: 4N has been implemented in R. We provide the sourcecode of 4N and a user-friendly toolbox including two example calculations. Biologists can use this 4N-toolbox even if they have a limited knowledge of R. There are only a few tuning parameters to set, and each of these parameters has a biological interpretation. The run times for medium scale datasets are in the order of minutes on a standard desktop PC. Large datasets can typically be analyzed within a few hours.

SCA with rotation to distinguish common and distinctive information in linked data.

Often data are collected that consist of different blocks that all contain information about the same entities (e.g., items, persons, or situations). In order to unveil both information that is common to all data blocks and information that is distinctive for one or a few of them, an integrated analysis of the whole of all data blocks may be most useful. Interesting classes of methods for such an approach are simultaneous-component and multigroup factor analysis methods. These methods yield dimensions underlying the data at hand. Unfortunately, however, in the results from such analyses, common and distinctive types of information are mixed up. This article proposes a novel method to disentangle the two kinds of information, by making use of the rotational freedom of component and factor models. We illustrate this method with data from a cross-cultural study of emotions.