Methodological and Statistical Practices of Using Symptom Networks to Evaluate Mental Health Interventions: A Review and Reflections

The network approach to psychopathology, which assesses associations between individual symptoms, has recently been applied to evaluate treatments for mental disorders. While various options for conducting network analyses in intervention research exist, an overview and an evaluation of the various approaches are currently missing. Therefore, we conducted a review on network analyses in intervention research. Studies were included if they constructed a symptom network, analyzed data that were collected before, during or after treatment of a mental disorder, and yielded information about the treatment effect. The 56 included studies were reviewed regarding their methodological and analytic strategies. About half of the studies based on data from randomized trials conducted a network intervention analysis, while the other half compared networks between treatment groups. The majority of studies estimated cross-sectional networks, even when repeated measures were available. All but five studies investigated networks on the group level. This review highlights that current methodological practices limit the information that can be gained through network analyses in intervention research. We discuss the strength and limitations of certain methodological and analytic strategies and propose that further work is needed to use the full potential of the network approach in intervention research.

Recognize the Value of the Sum Score, Psychometrics' Greatest Accomplishment

The sum score on a psychological test is, and should continue to be, a tool central in psychometric practice. This position runs counter to several psychometricians' belief that the sum score represents a pre-scientific conception that must be abandoned from psychometrics in favor of latent variables. First, we reiterate that the sum score stochastically orders the latent variable in a wide variety of much-used item response models. In fact, item response theory provides a mathematically based justification for the ordinal use of the sum score. Second, because discussions about the sum score often involve its reliability and estimation methods as well, we show that, based on very general assumptions, classical test theory provides a family of lower bounds several of which are close to the true reliability under reasonable conditions. Finally, we argue that eventually sum scores derive their value from the degree to which they enable predicting practically relevant events and behaviors. None of our discussion is meant to discredit modern measurement models; they have their own merits unattainable for classical test theory, but the latter model provides impressive contributions to psychometrics based on very few assumptions that seem to have become obscured in the past few decades. Their generality and practical usefulness add to the accomplishments of more recent approaches.

Evaluating potential effects of distress symptoms' interventions on suicidality: Analyses of in silico scenarios

BACKGROUND

Complexity science perspectives like the network approach to psychopathology have emerged as a prominent methodological toolkit to generate novel hypotheses on complex etiologies surrounding various mental health problems and inform intervention targets. Such approach may be pivotal in advancing early intervention of suicidality among the younger generation (10-35 year-olds), the increasing burden of which needs to be reversed within a limited window of opportunity to avoid massive long-term repercussions. However, the network approach currently lends limited insight into the potential extent of proposed intervention targets' effectiveness, particularly for target outcomes in comorbid conditions.

METHODS

This paper proposes an in silico (i.e., computer-simulated) intervention approach that maps symptoms' complex interactions onto dynamic processes and analyzes their evolution. The proposed methodology is applied to investigate potential effects of changes in 1968 community-dwelling individuals' distress symptoms on their suicidal ideation. Analyses on specific subgroups were conducted. Results were also compared with centrality indices employed in typical network analyses.

RESULTS

Findings concur with symptom networks' centrality indices in suggesting that timely deactivating hopelessness among distressed individuals may be instrumental in preventing distress to develop into suicidal ideation. Additionally, however, they depict nuances beyond those provided by centrality indices, e.g., among young adults, reducing nervousness and tension may have similar effectiveness as deactivating hopeless in reducing suicidal ideation.

LIMITATIONS

Caution is warranted when generalizing findings here to the general population.

CONCLUSION

The proposed methodology may help facilitate timely agenda-setting in population mental health measures, and may also be augmented for future co-creation projects.

Sex differences in the symptom network structure of depression: Findings from a nationwide sample of the Spanish adult population

BACKGROUND

Sex differences in the prevalence and clinical features of depression have been widely described. However, some authors argue that categorical diagnostic systems do not adequately capture the complexity of depression. The aim of this study was to examine sex differences in the symptom network structure of depressive symptoms among individuals with a major depressive episode.

METHODS

The study sample consisted of 510 participants (age 62.17 ± 14.43, 71.96 % women) from a nationwide study of the Spanish non-institutionalised adult population (Edad con Salud). To estimate the presence of a 12-month major depressive episode according to DSM-IV criteria, participants were administered an adapted version of the Composite International Diagnostic Interview (CIDI 3.0). A network analysis was carried out to determine possible interrelationships between different depressive symptoms by sex.

RESULTS

Men and women showed a similar overall structure and network strength. However, sex-specific variations emerged in relation to individual symptom associations and symptom centrality. Specifically, for individual symptom associations "loss of confidence" and "suicide attempts" were more strongly related in women, and "suicidal ideation" and "impaired thinking" in men. For symptom centrality, "anxiety" played a central role in men's symptomatology, whereas "hopelessness", "loss of confidence", "distress" and "slowness of movement" were the most central symptoms in the women's group.

LIMITATIONS

Reliance on cross-sectional data precludes us from determining the direction and temporality of the association between different symptoms.

CONCLUSIONS

This study suggests that specific symptoms should be prioritised in the prevention, diagnosis assessment and treatment of depressed patients based on sex.

Network analysis of neuropsychiatric symptoms in Alzheimer's disease

BACKGROUND

Neuropsychiatric symptoms due to Alzheimer's disease (AD) and mild cognitive impairment (MCI) can decrease quality of life for patients and increase caregiver burden. Better characterization of neuropsychiatric symptoms and methods of analysis are needed to identify effective treatment targets. The current investigation leveraged the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS) to examine the network structure of neuropsychiatric symptoms among symptomatic older adults with cognitive impairment.

METHODS

The network relationships of behavioral symptoms were estimated from Neuropsychiatric Inventory Questionnaire (NPI-Q) data acquired from 12,494 older adults with MCI and AD during their initial visit. Network analysis provides insight into the relationships among sets of symptoms and allows calculation of the strengths of the relationships. Nodes represented individual NPI-Q symptoms and edges represented the pairwise dependency between symptoms. Node centrality was calculated to determine the relative importance of each symptom in the network.

RESULTS

The analysis showed patterns of connectivity among the symptoms of the NPI-Q. The network (M = .28) consisted of mostly positive edges. The strongest edges connected nodes within symptom domain. Disinhibition and agitation/aggression were the most central symptoms in the network. Depression/dysphoria was the most frequently endorsed symptom, but it was not central in the network.

CONCLUSIONS

Neuropsychiatric symptoms in MCI and AD are highly comorbid and mutually reinforcing. The presence of disinhibition and agitation/aggression yielded a higher probability of additional neuropsychiatric symptoms. Interventions targeting these symptoms may lead to greater neuropsychiatric symptom improvement overall. Future work will compare neuropsychiatric symptom networks across dementia etiologies, informant relationships, and ethnic/racial groups, and will explore the utility of network analysis as a means of interrogating treatment effects.

Network Analysis of Neuropsychiatric Symptoms in Alzheimer's Disease

Background: Neuropsychiatric symptoms due to Alzheimer's disease (AD) and mild cognitive impairment (MCI) can decrease quality of life for patients and increase caregiver burden. Better characterization of neuropsychiatric symptoms and methods of analysis are needed to identify effective treatment targets. The current investigation leveraged the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS) to examine the network structure of neuropsychiatric symptoms among symptomatic older adults with cognitive impairment. Methods: The network relationships of behavioral symptoms was estimated from Neuropsychiatric Inventory Questionnaire (NPI-Q) data acquired from 12,494 older adults with MCI and AD during their initial visit. Network analysis provides insight into the relationships among sets of symptoms and allows calculation of the strengths of the relationships. Nodes represented individual NPI-Q symptoms and edges represented the pairwise dependency between symptoms. Node centrality was calculated to determine the relative importance of each symptom in the network. Results: The analysis showed patterns of connectivity among the symptoms of the NPI-Q. The network ( M =.28) consisted of mostly positive edges. The strongest edges connected nodes within symptom domain. Disinhibition and agitation/aggression were the most central symptoms in the network. Depression/dysphoria was the most frequently endorsed symptom, but it was not central in the network. Conclusions: Neuropsychiatric symptoms in MCI and AD are highly comorbid and mutually reinforcing. The presence of disinhibition and agitation/aggression yielded a higher probability of additional neuropsychiatric symptoms. Interventions targeting these symptoms may lead to greater neuropsychiatric symptom improvement overall. Future work will compare neuropsychiatric symptom networks across dementia etiologies, informant relationships, and ethnic/racial groups, and will explore the utility of network analysis as a means of interrogating treatment effects.

Using theory to guide exploratory network analyses

The use of exploratory network analysis has increased in psychopathology research over the past decade. A benefit of exploratory network analysis is the wealth of information it can provide; however, a single analysis may generate more inferences than what can be discussed in one manuscript (e.g., centrality indices of each node). This necessitates that authors choose which results to discuss in further detail and which to omit. Without a guide for this process, the likelihood of a biased interpretation is high. We propose that the integration of theory throughout the research process makes the interpretation of exploratory networks more manageable for the researcher and more likely to result in an interpretation that advances science. The goals of this paper are to differentiate between exploratory and confirmatory network analyses, discuss the utility of exploratory work, and provide a practical framework that uses theory as a guide to interpret exploratory network analyses.

Critical transitions in degree mixed networks: A discovery of forbidden tipping regions in networked spin systems

Critical transitions can be conceptualized as abrupt shifts in the state of a system typically induced by changes in the system's critical parameter. They have been observed in a variety of systems across many scientific disciplines including physics, ecology, and social science. Because critical transitions are important to such a diverse set of systems it is crucial to understand what parts of a system drive and shape the transition. The underlying network structure plays an important role in this regard. In this paper, we investigate how changes in a network's degree sequence impact the resilience of a networked system. We find that critical transitions in degree mixed networks occur in general sooner than in their degree homogeneous counterparts of equal average degree. This relationship can be expressed with parabolic curves that describe how the tipping point changes when the nodes of an initially homogeneous degree network composed only of nodes with degree k1 are replaced by nodes of a different degree k2. These curves mark clear tipping boundaries for a given degree mixed network and thus allow the identification of possible tipping intersections and forbidden tipping regions when comparing networks with different degree sequences.

Editorial of the Psych Special Issue “Computational Aspects, Statistical Algorithms and Software in Psychometrics”

Statistical software in psychometrics has made tremendous progress in providing open source solutions (e [...]