The importance of cluster analysis for enhancing clinical practice: an example from irritable bowel syndrome

Measuring health literacy to inform actions to address health inequities: a cluster analysis approach based on the Australian national health literacy survey

BACKGROUND

Measuring health literacy can inform interventions to address health inequities. This study used cluster analysis to examine health literacy data to determine if it can provide more insightful information than standard descriptive analysis to better inform intervention development.

METHODS

Using data from the Australian National Health Survey (2018), this study compared descriptive analysis and cluster analysis results of two states-New South Wales (NSW) and Victoria-generated from the Health Literacy Questionnaire (HLQ). Based on the nine scale scores of the HLQ, a hierarchical cluster analysis using Ward's method for linkage was undertaken.

RESULTS

The number of NSW and Victoria respondents was 1018 and 923, respectively. The nine HLQ scale full sample mean scores from both states were similar. However, the cluster analyses identified 11 clusters for NSW and 12 clusters for Victoria. While six clusters from each state presented similar health literacy patterns, five and six clusters from NSW and Victoria, respectively, displayed unique health literacy patterns.

CONCLUSIONS

The results demonstrate that descriptive analysis only provides an overview and may lead to one-size-fits-all interventions. The varying health literacy patterns among subgroups resulting from the cluster analysis pave the way to inform tailored actions to improve health equity.

How are poor sleepers with other clinical conditions affected by maladaptive personality traits? A neural network-based analysis

Background

Psychopathology research mainly focused on the cross-sectional and longitudinal associations between personality and psychiatric disorders without considering the moment-to-moment dynamics of personality in response to environmental situations. The present study aimed to both cluster a young sample according to three mixed clinical conditions (poor sleep quality, depression, and somatization) and to predict the derived clusters by maladaptive personality traits and sex differences using a deep machine learning approach.

Methods

A sample of 839 adults aged 18-40 years (64% female) from the west of Iran were clustered according to the mixed clinical conditions using the cluster analysis techniques. An Artificial Neural Network (ANN) modeling is used to predict the derived clusters by maladaptive personality traits and biological gender. A receiver operating characteristic (ROC) curve was used to identify independent variables with high sensitivity specific to the derived clusters.

Results

The cluster analysis techniques suggested a fully stable and acceptable four-cluster solution for Depressed Poor Sleepers, Nonclinical Good Sleepers, Subclinical Poor Sleepers, and Clinical Poor Sleepers. The ANN model led to the identification of one hidden layer with two hidden units. The results of Area under the ROC Curve were relatively to completely acceptable, ranging from.726 to.855. Anhedonia, perceptual dysregulation, depressivity, anxiousness, and unusual beliefs are the most valuable traits with importance higher than 70%.

Conclusion

The machine learning approach can be well used to predict mixed clinical conditions by maladaptive personality traits. Future research can test the complexity of normal personality traits connected to mixed clinical conditions.

Burden and positive aspects of caregiving: cluster profiles of dementia caregiving experiences

OBJECTIVES

Although caregiver burden is common in the context of dementia caregiving, the caregiving role is linked to beneficial outcomes too. Individuals reporting higher positive aspects of caregiving tend to exhibit lower burden relative to those reporting few. The goal of this retrospective review of outpatient memory clinic medical records was to demonstrate whether and how constructs of burden and positive aspects of caregiving coexist within individual caregivers, and to explore potential contributors to caregiver profiles created based upon these constructs.

METHOD

Cluster analyses were conducted on 1160 caregivers from an initial intake interview meeting criteria on primary measures of Positive Aspects of Caregiving and the Zarit Burden Interview and repeated with 225 caregivers meeting inclusion criteria on all measures. Samples were compared for similarity, and the smaller sample (n = 225) was deemed appropriately representative. Multinomial logistic regressions examined cluster predictors in sample with 225 caregivers.

RESULTS

Results suggested a three-cluster solution: a High Burden group, a High Positive Experiences group, and a Low-Moderate Experiences group showing low burden and moderate positive experiences. Greater behavioral problems predicted belonging to the High Burden cluster. Greater care recipient dependence predicted belonging to the High Positive Experiences cluster while greater independence predicted the Low-Moderate Experiences cluster.

CONCLUSION

Findings suggest that burden and positive aspects of caregiving do not simultaneously present in caregivers at high levels. Supportive caregiver interventions might be tailored to profiles demonstrated here. Future research should investigate other potential contributors to experiences of burden and positive aspects of caregiving.

The impact of symptom clusters on endocrine therapy adherence in patients with breast cancer

BACKGROUND

When taken as prescribed, endocrine therapy is effective in reducing risk of recurrence and mortality in the treatment of patients with breast cancer. However, treatment side effects can act as a barrier to medication adherence. Existing research has not identified any specific side effects as consistent predictors of nonadherence. Our aim was to explore the influence of symptom clusters on self-reported adherence in patients with breast cancer.

METHODS

A cross-sectional online survey was conducted, including patients with breast cancer currently or previously prescribed endocrine therapy (N = 1051). This included measures of self-reported endocrine therapy adherence and common symptoms among this population (insomnia, depression, anxiety, fatigue, musculoskeletal, and vasomotor symptoms).

RESULTS

Unintentional nonadherence was higher than intentional nonadherence (50.8 % vs 31.01 %). The most troublesome symptom was insomnia (73.83 % displayed probable insomnia disorder). K-means cluster analysis identified 2 symptom clusters: overall High symptoms, and overall Low symptoms. Participants in the Low symptoms cluster were significantly more likely to be classed as adherent based on unintentional and intentional items.

CONCLUSIONS

Nonadherence was high in the current sample, and significantly more likely in participants reporting overall severe symptoms. Clinicians should be aware of the scale of common side effects and facilitate open conversation about potential barriers to adherence. Follow-up care should include assessment of common symptoms and signpost patients to appropriate support or treatment when required. Future research should explore potential for a central symptom to act as a target for intervention, to relieve overall side effect burden and facilitate better medication adherence.

Patterns of cognitive-emotional change after cognitive-behavioural treatment in emotional disorders: A 12-month longitudinal cluster analysis

INTRODUCTION

The aim of this study was to use cluster analysis based on the trajectory of five cognitive-emotional processes (worry, rumination, metacognition, cognitive reappraisal and expressive suppression) over time to explore differences in clinical and performance variables in primary care patients with emotional symptoms.

METHODS

We compared the effect of adding transdiagnostic cognitive-behavioural therapy (TD-CBT) to treatment as usual (TAU) according to cluster membership and sought to determine the variables that predicted cluster membership. 732 participants completed scales about cognitive-emotional processes, anxiety and depressive symptoms, functioning, and quality of life (QoL) at baseline, posttreatment, and at 12 months. Longitudinal cluster analysis and logistic regression analyses were carried out.

RESULTS

A two-cluster solution was chosen as the best fit, named as "less" or "more" improvement in cognitive-emotional processes. Individuals who achieved more improvement in cognitive-emotional processes showed lower emotional symptoms and better QoL and functioning at all three time points. TAU+TD-CBT, income level, QoL and anxiety symptoms were significant predictors of cluster membership.

CONCLUSIONS

These results underscore the value of adding TD-CBT to reduce maladaptive cognitive-emotional regulation strategies. These findings highlight the importance of the processes of change in therapy and demonstrate the relevance of the patient's cognitive-emotional profile in improving treatment outcomes.

Subtyping irritable bowel syndrome using cluster analysis: a systematic review

BACKGROUND

Irritable bowel syndrome (IBS) is a common chronic functional gastrointestinal disorder associated with a wide range of clinical symptoms. Some researchers have used cluster analysis (CA), a group of non-supervised learning methods that identifies homogenous clusters within different entities based on their similarity.

OBJECTIVE AND METHODS

This literature review aims to identify published articles that apply CA to IBS patients. We searched relevant keywords in PubMed, Embase, Web of Science, and Scopus. We reviewed studies in terms of the selected variables, participants' characteristics, data collection, methodology, number of clusters, clusters' profiles, and results.

RESULTS

Among the 14 articles focused on the heterogeneity of IBS, eight of them utilized K-means Cluster Analysis (K-means CA), four employed Hierarchical Cluster Analysis, and only two studies utilized Latent Class Analysis. Seven studies focused on clinical symptoms, while four articles examined anocolorectal functions. Two studies were centered around immunological findings, and only one study explored microbial composition. The number of clusters obtained ranged from two to seven, showing variation across the studies. Males exhibited lower symptom severity and fewer psychological findings. The association between symptom severity and rectal perception suggests that altered rectal perception serves as a biological indicator of IBS. Ultra-slow waves observed in IBS patients are linked to increased activity of the anal sphincter, higher anal pressure, dystonia, and dyschezia.

CONCLUSION

IBS has different subgroups based on different factors. Most IBS patients have low clinical severity, good QoL, high rectal sensitivity, delayed left colon transit time, increased systemic cytokines, and changes in microbial composition, including increased Firmicutes-associated taxa and depleted Bacteroidetes-related taxa. However, the number of clusters is inconsistent across studies due to the methodological heterogeneity. CA, a valuable non-supervised learning method, is sensitive to hyperparameters like the number of clusters and random initialization of cluster centers. The random nature of these parameters leads to diverse outcomes even with the same algorithm. This has implications for future research and practical applications, necessitating further studies to improve our understanding of IBS and develop personalized treatments.

Positive psychosocial factors and the development of symptoms of depression and posttraumatic stress symptoms following acute myocardial infarction

Introduction

Acute myocardial infarction (MI) is a potentially fatal condition, leading to high psychological distress and possibly resulting in the development of depressive symptoms and posttraumatic stress symptoms (PTSS). The aim of this study was to investigate the association of clusters of positive psychosocial factors (resilience, task-oriented coping, positive affect and social support) with both MI-induced depressive symptoms and PTSS, independent of demographic factors.

Methods

We investigated 154 consecutive patients with MI, 3 and 12 months after hospital discharge. All patients completed the short version of the German Resilience Scale, the Coping Inventory for Stressful Situations (CISS), the Enriched Social Support Inventory (ESSI) and the Global Mood Scale (GMS). The level of interviewer-rated MI-induced posttraumatic stress disorder (PTSD) symptoms at 3- and 12-months follow-up was evaluated through the Clinician-Administered PTSD Scale (CAPS). Depressive symptoms were assessed at 3- and 12-month follow-up with the Beck Depression Inventory (BDI-II).

Results

Three different clusters were revealed: (1) lonely cluster: lowest social support, resilience and average task-oriented coping and positive affect; (2) low risk cluster: highest resilience, task-oriented coping, positive affect and social support; (3) avoidant cluster: lowest task-oriented coping, positive affect, average resilience and social support. The clusters differed in depressive symptoms at 3 months (F = 5.10; p < 0.01) and 12 months follow-up (F = 7.56; p < 0.01). Cluster differences in PTSS were significant at 3 months (F = 4.78, p < 0.05) and 12 months (F = 5.57, p < 0.01) follow-up. Differences in PTSS subscales were found for avoidance (F = 4.8, p < 0.05) and hyperarousal (F = 5.63, p < 0.05), but not re-experiencing, at 3 months follow-up. At 12 months follow-up, cluster differences were significant for re-experiencing (F = 6.44, p < 0.01) and avoidance (F = 4.02, p < 0.05) but not hyperarousal.

Discussion

The present study contributes to a better understanding of the relationships among different positive psychosocial factors, depressive symptoms and PTSS following acute MI. Future interventions may benefit from taking into account positive psychosocial factors to potentially reduce patients' depressive symptoms and PTSS after MI.

Association of Neighborhood Deprivation With Thrombolysis and Thrombectomy for Acute Stroke in a Health System With Universal Access

BACKGROUND AND OBJECTIVES

The association between socioeconomic status and acute ischemic stroke treatments remain uncertain, particularly in countries with universal health care systems. This study aimed to investigate the association between neighborhood-level material deprivation and the odds of receiving IV thrombolysis or thrombectomy for acute ischemic stroke within a single-payer, government-funded health care system.

METHODS

We conducted a population-based cohort study using linked administrative data from Ontario, Canada. This study involved all community-dwelling adult Ontario residents hospitalized with acute ischemic stroke between 2017 and 2022. Neighborhood-level material deprivation, measured in quintiles from least to most deprived, was our main exposure. We considered the receipt of thrombolysis or thrombectomy as the primary outcome. We used multivariable logistic regression models adjusted for baseline differences to estimate the association between material deprivation and outcomes. We performed a sensitivity analysis by additionally adjusting for hospital type at initial assessment. Furthermore, we tested whether hospital type modified the associations between deprivation and outcomes.

RESULTS

Among 57,704 patients, those in the most materially deprived group (quintile 5) were less likely to be treated with thrombolysis or thrombectomy compared with those in the least deprived group (quintile 1) (16.6% vs 19.6%, adjusted odds ratio [aOR] 0.76, 95% CI 0.63-0.93). The association was consistent when evaluating thrombolysis (13.0% vs 15.3%, aOR 0.78, 95% CI 0.64-0.96) and thrombectomy (6.4 vs 7.8%, aOR 0.73, 95% CI 0.59-0.90) separately. There were no statistically significant differences between the middle 3 quintiles and the least deprived group. These associations persisted after additional adjustment for hospital type, and there was no interaction between material deprivation and hospital type (p interaction >0.1).

DISCUSSION

We observed disparities in the use of thrombolysis or thrombectomy for acute ischemic stroke by socioeconomic status despite access to universal health care. Targeted health care policies, public health messaging, and resource allocation are needed to ensure equitable access to acute stroke treatments for all patients.

The interplay between social media use and problematic internet usage: Four behavioral patterns

Objective

The study aims to identify typical interplay between the use of social media apps on smartphones and Problematic Internet Usage (PIU).

Method

Our study utilizes data from a smartphone app that objectively monitors user usage, including the apps used and the start and finish times of each app session. This study included 334 participants who declared a need to be aware of their smartphone usage and control it. Problematic Internet Usage (PIU) was measured using the Problematic Internet Use Questionnaire-Short Form-6 (PIUQ-SF6). The total PIU score can range from 6 to 30, with a score above 15 indicating that a person is at risk of PIU. Time spent on Social Media (SM) apps of Facebook, WhatsApp, and Instagram, and whether people used each of these apps were studied along with the total PIU score. K-Prototype clustering was utilized for the analysis.

Results

Four distinct clusters, typifying the relationship between social media use and PIU, were identified. All the individuals in Cluster 1 (Light SM Use Cluster; Cluster size = 270, 80.84% of total dataset) spent between 0 and 109.01 min on Instagram, between 0 and 69.84 min on Facebook, and between 0 and 86.42 min on WhatsApp and its median PIU score was 17. Those who were in cluster 2 (Highly Visual SM Cluster; Cluster size = 23, 6.89% of total dataset) all used Instagram, and each member spent between 110 and 307.63 min on Instagram daily. The cluster median PIU score and average daily usage of Instagram were respectively 20 and 159.66 min. Those who were in Cluster 3 (Conversational SM Cluster; Cluster size = 19, 5.69% of total dataset) all used WhatsApp, and spent between 76.68 and 225.22 min on WhatsApp daily. The cluster median PIU score and average time spent per day on WhatsApp were 20 and 132.65 min, respectively. Those who were in Cluster 4 (Social Networking Cluster; (Cluster size = 22, 6.59% of total dataset) all used Facebook, and each spent between 73.09 and 272.85 min daily on Facebook. The cluster median PIU score and average time spent per day on Facebook were 18 and 133.61 min respectively.

Conclusion

The clusters indicate that those who use a particular social media app spend significantly less time on other social media apps. This indicates that problematic attachment to social media occurs primarily for one of three reasons: visual content and reels, conversations with peers, or surfing network content and news. This finding will help tailor interventions to fit each cluster, for example by strengthening interpersonal skills and resistance to peer pressure in the case of Cluster 3 and increasing impulse control in the case of Cluster 2.

Cluster-Analytic Identification of Clinically Meaningful Subtypes in MCAS: The Relevance of Heat and Cold

BACKGROUND

Mast cell activation syndrome (MCAS) is a clinically heterogeneous disease with allergy-like symptoms and abdominal complaints. Its etiology is only partially understood and it is often overlooked.

AIMS

The aim of this study was to identify subgroups of MCAS patients to facilitate diagnosis and allow a personalized therapy.

METHODS

Based on data from 250 MCAS patients, hierarchical and two-step cluster analyses as well as association analyses were performed. The data used included data from a MCAS checklist asking about symptoms and triggers and a set of diagnostically relevant laboratory parameters.

RESULTS

Using a two-step cluster analysis, MCAS patients could be divided into three clusters. Physical trigger factors were particularly decisive for the classification as they showed remarkable differences between the three clusters. Cluster 1, labeled high responders, showed high values for the triggers heat and cold, whereas cluster 2, labeled intermediate responders, presented with high values for the trigger heat and low values for cold. The third cluster, labeled low responders, did not react to thermal triggers. The first two clusters showed more divers clinical symptoms especially with regard to dermatological and cardiological complaints. Subsequent association analyses revealed relationships between triggers and clinical complaints: Abdominal discomfort is mainly triggered by histamine consumption, dermatological discomfort by exercise, and neurological symptoms are related to physical exertion and periods of starvation. The reasons for the occurrence of cardiological complaints are manifold and triggers for respiratory complaints still need better identification.

CONCLUSION

Our study identified three distinct clusters on the basis of physical triggers, which also differ significantly in their clinical symptoms. A trigger-related classification can be helpful in clinical practice for diagnosis and therapy. Longitudinal studies should be conducted to further understand the relationship between triggers and symptoms.

Cognitive phenotypes in late-life depression

OBJECTIVE

To identify cognitive phenotypes in late-life depression (LLD) and describe relationships with sociodemographic and clinical characteristics.

DESIGN

Observational cohort study.

SETTING

Baseline data from participants recruited via clinical referrals and community advertisements who enrolled in two separate studies.

PARTICIPANTS

Non-demented adults with LLD (n = 120; mean age = 66.73 ± 5.35 years) and non-depressed elders (n = 56; mean age = 67.95 ± 6.34 years).

MEASUREMENTS

All completed a neuropsychological battery, and individual cognitive test scores were standardized across the entire sample without correcting for demographics. Five empirically derived cognitive domain composites were created, and cluster analytic approaches (hierarchical, k-means) were independently conducted to classify cognitive patterns in the depressed cohort only. Baseline sociodemographic and clinical characteristics were then compared across groups.

RESULTS

A three-cluster solution best reflected the data, including "High Normal" (n = 47), "Reduced Normal" (n = 35), and "Low Executive Function" (n = 37) groups. The "High Normal" group was younger, more educated, predominantly Caucasian, and had fewer vascular risk factors and higher Mini-Mental Status Examination compared to "Low Executive Function" group. No differences were observed on other sociodemographic or clinical characteristics. Exploration of the "High Normal" group found two subgroups that only differed in attention/working memory performance and length of the current depressive episode.

CONCLUSIONS

Three cognitive phenotypes in LLD were identified that slightly differed in sociodemographic and disease-specific variables, but not in the quality of specific symptoms reported. Future work on these cognitive phenotypes will examine relationships to treatment response, vulnerability to cognitive decline, and neuroimaging markers to help disentangle the heterogeneity seen in this patient population.

Using Markov chains and temporal alignment to identify clinical patterns in Dementia

In the healthcare sector, resorting to big data and advanced analytics is a great advantage when dealing with complex groups of patients in terms of comorbidities, representing a significant step towards personalized targeting. In this work, we focus on understanding key features and clinical pathways of patients with multimorbidity suffering from Dementia. This disease can result from many heterogeneous factors, potentially becoming more prevalent as the population ages. We present a set of methods that allow us to identify medical appointment patterns within a cohort of 1924 patients followed from January 2007 to August 2021 in Hospital da Luz (Lisbon), and to stratify patients into subgroups that exhibit similar patterns of interaction. With Markov Chains, we are able to identify the most prevailing medical appointments attended by Dementia patients, as well as recurring transitions between these. To perform patient stratification, we applied AliClu, a temporal sequence alignment algorithm for clustering longitudinal clinical data, which allowed us to successfully identify patient subgroups with similar medical appointment activity. A feature analysis per cluster obtained allows the identification of distinct patterns and characteristics. This pipeline provides a tool to identify prevailing clinical pathways of medical appointments within the dataset, as well as the most common transitions between medical specialities within Dementia patients. This methodology, alongside demographic and clinical data, has the potential to provide early signalling of the most likely clinical pathways and serve as a support tool for health providers in deciding the best course of treatment, considering a patient as a whole.

Tracking clusters of patients over time enables extracting information from medico-administrative databases

CONTEXT

Identifying clusters (i.e., subgroups) of patients from the analysis of medico-administrative databases is particularly important to better understand disease heterogeneity. However, these databases contain different types of longitudinal variables which are measured over different follow-up periods, generating truncated data. It is therefore fundamental to develop clustering approaches that can handle this type of data.

OBJECTIVE

We propose here cluster-tracking approaches to identify clusters of patients from truncated longitudinal data contained in medico-administrative databases.

MATERIAL AND METHODS

We first cluster patients at each age. We then track the identified clusters over ages to construct cluster-trajectories. We compared our novel approaches with three classical longitudinal clustering approaches by calculating the silhouette score. As a use-case, we analyzed antithrombotic drugs used from 2008 to 2018 contained in the Échantillon Généraliste des Bénéficiaires (EGB), a French national cohort.

RESULTS

Our cluster-tracking approaches allow us to identify several cluster-trajectories with clinical significance without any imputation of data. The comparison of the silhouette scores obtained with the different approaches highlights the better performances of the cluster-tracking approaches.

CONCLUSION

The cluster-tracking approaches are a novel and efficient alternative to identify patient clusters from medico-administrative databases by taking into account their specificities.

Qluster: An easy-to-implement generic workflow for robust clustering of health data

The exploration of heath data by clustering algorithms allows to better describe the populations of interest by seeking the sub-profiles that compose it. This therefore reinforces medical knowledge, whether it is about a disease or a targeted population in real life. Nevertheless, contrary to the so-called conventional biostatistical methods where numerous guidelines exist, the standardization of data science approaches in clinical research remains a little discussed subject. This results in a significant variability in the execution of data science projects, whether in terms of algorithms used, reliability and credibility of the designed approach. Taking the path of parsimonious and judicious choice of both algorithms and implementations at each stage, this article proposes Qluster, a practical workflow for performing clustering tasks. Indeed, this workflow makes a compromise between (1) genericity of applications (e.g. usable on small or big data, on continuous, categorical or mixed variables, on database of high-dimensionality or not), (2) ease of implementation (need for few packages, few algorithms, few parameters, ...), and (3) robustness (e.g. use of proven algorithms and robust packages, evaluation of the stability of clusters, management of noise and multicollinearity). This workflow can be easily automated and/or routinely applied on a wide range of clustering projects. It can be useful both for data scientists with little experience in the field to make data clustering easier and more robust, and for more experienced data scientists who are looking for a straightforward and reliable solution to routinely perform preliminary data mining. A synthesis of the literature on data clustering as well as the scientific rationale supporting the proposed workflow is also provided. Finally, a detailed application of the workflow on a concrete use case is provided, along with a practical discussion for data scientists. An implementation on the Dataiku platform is available upon request to the authors.

Social and Metabolic Determinants of Prevalent Hypertension in Men and Women: A Cluster Analysis from a Population-Based Study

Essential hypertension (HTN) has a complex spectrum of pathophysiological determinants and current guidelines provide limited information on high-risk groups that should be targeted for its primary prevention. The objective of our research was to identify clusters of social and metabolic factors associated with prevalent HTN in men and women from a population-based survey in Romania. Of the 1477 participants in the main study, 798 with complete data were analyzed here. Using two-step cluster analysis, one high-risk cluster in women and two high and intermediate risk for prevalent HTN in men were identified. Older age, rural area, lower education, and higher burden of metabolic factors characterized clusters with higher risk, while intermediate risk in men was characterized by a more metabolically healthy phenotype in younger individuals. In logistic regression, men in Cluster 1 vs. those in Cluster 3 had an odds ratio (OR) of 9.6 (95%CI: 4.6; 20.0), p < 0.001 for prevalent HTN, while OR for Cluster 2 vs. Cluster 3 was 3.2 (95%CI: 1.4; 7.4), p = 0.005. In women, the OR for HTN was 10.2 (95%CI: 5.7; 18.5) if assigned to Cluster 2 vs. Cluster 1, p < 0.001. These results pointed out the subgroups and communities that the primary prevention of HTN should be prioritized in.

Evaluating Personality Assessment Inventory Response Patterns in Active-Duty Personnel With Head Injury Using a Latent Class Approach

OBJECTIVE

Previous research has found that among those with brain injury, individuals have a variety of different potential symptom sets, which will be seen on the Personality Assessment Inventory (PAI). The number of different groups and what they measure have varied depending on the study.

METHOD

In active-duty personnel with a remote history of mild traumatic brain injury (n = 384) who were evaluated at a neuropsychology clinic, we used a retrospective database to examine if there are different groups of individuals who have distinct sets of symptoms as measured on the PAI. We examined the potential of distinct groups of respondents by conducting a latent class analysis of the clinical scales. Post hoc testing of group structures was conducted on concurrently administered cognitive testing, performance validity tests, and the PAI subscales.

RESULTS

Findings indicate a pattern of broad symptom severity as the most probable reason for multiple groups of respondents, suggesting that there are no distinct symptom sets observed within this population. Pathology levels were the most elevated on internalizing and thought disorder scales across the various class solutions.

CONCLUSION

Findings indicate that among active-duty service members with remote brain injury, there are no distinct groups of respondents with different sets of symptom types as has been found in prior work with other neuropsychology samples. We conclude that the groups found are likely a function of general psychopathology present in the population/sample rather than bona fide differences.

Correlates of wanting to seek help for mental health and substance use concerns by sexual and gender minority young adults during the COVID-19 pandemic: A machine learning analysis

The COVID-19 pandemic has worsened the mental health and substance use challenges among many people who are Two Spirit, lesbian, gay, bisexual, transgender, queer, questioning, and intersex (2SLGBTQI+). We aimed to identify the important correlates and their effects on the predicted likelihood of wanting to seek help among 2SLGBTQI+ young adults for mental health or substance use concerns during the pandemic. A cross-sectional survey was conducted in 2020-2021 among 2SLGBTQI+ young adults aged 16-29 living in two Canadian provinces (Ontario and Quebec). Among 1414 participants, 77% (n = 1089) wanted to seek help for their mental health or substance use concerns during the pandemic, out of these, 69.8% (n = 760) reported delay in accessing care. We built a random forest (RF) model to predict the status of wanting to seek help, which achieved moderately high performance with an area under the receiver operating characteristic curve (AUC) of 0.85. The top 10 correlates of wanting to seek help were worsening mental health, age, stigma and discrimination, and adverse childhood experiences. The interactions of adequate housing with certain sexual orientations, gender identities and mental health challenges were found to increase the likelihood of wanting to seek help. We built another RF model for predicting risk of delay in accessing care among participants who wanted to seek help (n = 1089). The model identified a similar set of top 10 correlates of delay in accessing care but lacked adequate performance (AUC 0.61). These findings can direct future research and targeted prevention measures to reduce health disparities for 2SLGBTQI+ young adults.

Oxygen debt repayment in the early phase of veno-arterial extracorporeal membrane oxygenation: a cluster analysis

Introduction

Early oxygen debt repayment is predictive of successful weaning from veno-arterial extracorporeal membrane oxygenation (V-A ECMO). However, studies are limited by the patient cohort’s heterogeneity. This study aimed to understand the early state of oxygen debt repayment and its association with end-organ failure and 30-day survival using cluster analysis.

Methods

A retrospective, single-center study was conducted on 153V-A ECMO patients. Patients were clustered using a two-step cluster analysis based on oxygen debt and its repayment during the first 24 h of ECMO. Primary outcomes were end-organ failure and 30-day survival.

Results

The overall mortality was 69.3%. For cluster analysis, 137 patients were included, due to an incomplete data set. The mortality rate in this subset was 67.9%. Three clusters were generated, representing increasing levels of total oxygen debt from cluster 1 to cluster 3. Thirty-day survival between clusters was significantly different (cluster 1: 46.9%, cluster 2: 23.4%, and cluster 3: 4.8%, p = 0.001). Patients in cluster 3 showed less decrement in liver enzymes, creatinine, and urea blood levels. There were significant differences in the baseline oxygen debt and the need for continuous veno-venous hemofiltration (CVVH) between survivors and non-survivors (p < 0.05). Forty-seven patients (34.3%) migrated between clusters within the first 24 h of support. Among these patients, 43.4% required CVVH. Notably, patients requiring CVVH and who migrated to a cluster with a higher oxygen debt repayment showed better survival rates compared to those who migrated to a cluster with a lower oxygen debt repayment.

Conclusions

Oxygen debt repayment during the first 24 h of V-A ECMO shows to correspond with survival, where the baseline oxygen debt value and the necessity for continuous kidney replacement therapy appear to be influential.

Exploration of mood spectrum symptoms during a major depressive episode: The impact of contrapolarity-Results from a transdiagnostic cluster analysis on an Italian sample of unipolar and bipolar patients

Background

Subthreshold hypomania during a major depressive episode challenges the bipolar-unipolar dichotomy. In our study we employed a cross-diagnostic cluster analysis - to identify distinct subgroups within a cohort of depressed patients.

Methods

A k-means cluster analysis— based on the domain scores of the Mood Spectrum Self-Report (MOODS-SR) questionnaire—was performed on a data set of 300 adults with either bipolar or unipolar depression. After identifying groups, between-clusters comparisons were conducted on MOODS-SR domains and factors and on a set of sociodemographic, clinical and psychometric variables.

Results

Three clusters were identified: one with intermediate depressive and poor manic symptomatology (Mild), one with severe depressive and poor manic symptomatology (Moderate), and a third one with severe depressive and intermediate manic symptomatology (Mixed). Across the clusters, bipolar patients were significantly less represented in the Mild one, while the DSM-5 “Mixed features” specifier did not differentiate the groups. When compared to the other patients, those of Mixed cluster exhibited a stronger association with most of the illness-severity, quality of life, and outcomes measures considered. After performing pairwise comparisons significant differences between “Mixed” and “Moderate” clusters were restricted to: current and disease-onset age, psychotic ideation, suicidal attempts, hospitalization numbers, impulsivity levels and comorbidity for Cluster B personality disorder.

Conclusions

In the present study, a clustering approach based on a spectrum exploration of mood symptomatology led to the identification of three transdiagnostic groups of patients. Consistent with our hypothesis, the magnitude of subthreshold (hypo)manic symptoms was related to a greater clinical severity, regardless of the main categorical diagnosis.

Cluster analysis reveals distinct patterns of childhood adversity, behavioral disengagement, and depression that predict blunted heart rate reactivity to acute psychological stress

BACKGROUND

There is considerable evidence documenting associations between early life adversity, behavioral disengagement, and depression with blunted cardiovascular reactivity to acute psychological stress. However, while often examined as independent predictors, it is also likely that a combination of these factors uniquely relate to cardiovascular reactivity.

PURPOSE

The present study employed multivariate cluster analysis to examine if distinct combinations of these outcomes relate to cardiovascular stress reactivity.

METHODS

Participants (N = 467) were predominantly female (60.6%) with a mean age of 19.30 years (SD = 0.82). Measures of early life adversity, behavioral disengagement, and depression were completed; in addition, participants had their blood pressure and heart rate monitored throughout a standardized stress testing session. Cardiovascular reactivity was calculated as the difference between mean stress and mean baseline cardiovascular values.

RESULTS

Analyses revealed two clusters with distinct patterns of exposure to early life adversity, levels of behavioral disengagement and depression, uniquely related to cardiovascular reactivity. In unadjusted models, Cluster 1 that was characterized by greater exposure to early life adversity, higher levels of behavioral disengagement and depression, was associated with lower systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate (HR) reactivity. Cluster 2 was characterized by reactivity values similar to the sample means. In fully adjusted models, Cluster 1 predicted heart rate reactivity to stress.

CONCLUSIONS

The present study identifies a behavioral cluster that is characteristic of a blunted heart rate reactivity profile, significantly extending the research in this area.

Clinical Phenotypes of Atrial Fibrillation and Mortality Risk-A Cluster Analysis from the Nationwide Italian START Registry

Patients with atrial fibrillation (AF) still experience a high mortality rate despite optimal antithrombotic treatment. We aimed to identify clinical phenotypes of patients to stratify mortality risk in AF. Cluster analysis was performed on 5171 AF patients from the nationwide START registry. The risk of all-cause mortality in each cluster was analyzed. We identified four clusters. Cluster 1 was composed of the youngest patients, with low comorbidities; Cluster 2 of patients with low cardiovascular risk factors and high prevalence of cancer; Cluster 3 of men with diabetes and coronary disease and peripheral artery disease; Cluster 4 included the oldest patients, mainly women, with previous cerebrovascular events. During 9857 person-years of observation, 386 deaths (3.92%/year) occurred. Mortality rates increased across clusters: 0.42%/year (cluster 1, reference group), 2.12%/year (cluster 2, adjusted hazard ratio (aHR) 3.306, 95% confidence interval (CI) 1.204−9.077, p = 0.020), 4.41%/year (cluster 3, aHR 6.702, 95%CI 2.433−18.461, p < 0.001), and 8.71%/year (cluster 4, aHR 8.927, 95%CI 3.238−24.605, p < 0.001). We identified four clusters of AF patients with progressive mortality risk. The use of clinical phenotypes may help identify patients at a higher risk of mortality.

Comparison of Unsupervised Machine Learning Approaches for Cluster Analysis to Define Subgroups of Heart Failure with Preserved Ejection Fraction with Different Outcomes

Heart failure with preserved ejection (HFpEF) is a heterogenous condition affecting nearly half of all patients with heart failure (HF). Artificial intelligence methodologies can be useful to identify patient subclassifications with important clinical implications. We sought a comparison of different machine learning (ML) techniques and clustering capabilities in defining meaningful subsets of patients with HFpEF. Three unsupervised clustering strategies, hierarchical clustering, K-prototype, and partitioning around medoids (PAM), were used to identify distinct clusters in patients with HFpEF, based on a wide range of demographic, laboratory, and clinical parameters. The study population had a median age of 77 years, with a female majority, and moderate diastolic dysfunction. Hierarchical clustering produced six groups but two were too small (two and seven cases) to be clinically meaningful. The K-prototype methods produced clusters in which several clinical and biochemical features did not show statistically significant differences and there was significant overlap between the clusters. The PAM methodology provided the best group separations and identified six mutually exclusive groups (HFpEF1-6) with statistically significant differences in patient characteristics and outcomes. Comparison of three different unsupervised ML clustering strategies, hierarchical clustering, K-prototype, and partitioning around medoids (PAM), was performed on a mixed dataset of patients with HFpEF containing clinical and numerical data. The PAM method identified six distinct subsets of patients with HFpEF with different long-term outcomes or mortality. By comparison, the two other clustering algorithms, the hierarchical clustering and K-prototype, were less optimal.

Frontline Service Employees’ Profiles: Exploring Individual Differences in Perceptions of and Reactions to Workplace Incivility

Employee turnover is a big issue in the service industry, which can be significantly affected by job stressors including workplace incivility. This exploratory study aims to identify the frontline service employees’ profiles exploring to what extent individuals may have different perceptions of incivility and social supports at work and showing different reactions (job outcomes). In a cross-sectional study, 291 completed questionnaires from a sample of Norwegian frontline service employees were subjected to correlation analysis, K-means clustering, and post hoc ANOVA analysis with Bonferroni correction. Cluster analysis revealed three distinct clusters of employees with different profiles, which indicated that those who perceived the highest level of workplace incivility and the lowest level of social supports at work showed the highest turnover intention compared to that of others. Moreover, employees with longer tenure and the highest perception of social supports at work coped better with workplace incivility and showed the lowest turnover intention.

Codesign and implementation of an equity-promoting national health literacy programme for people living with inflammatory bowel disease (IBD): a protocol for the application of the Optimising Health Literacy and Access (Ophelia) process

INTRODUCTION

Non-government organisations (NGOs) often represent people who are underserved or experiencing vulnerability. Crohn's & Colitis Australia (CCA) is aware that many Australians with inflammatory bowel disease (IBD) are not reached by current communication and engagement activities. The aim of the CCA IBD project is to implement the Optimising Health Literacy and Access (Ophelia) process over 3 years to collaboratively codesign ways to improve delivery of information, services and resources for people with IBD and their carers.

METHODS AND ANALYSIS

Health literacy and other data for phase 1 will be collected using the Health Literacy Questionnaire, eHealth Literacy Questionnaire, IBD-related questions and qualitative interviews with people with IBD and their carers to ascertain their lived experience. Quantitative data will be analysed using descriptive statistics and cluster analysis. Identified clusters will be combined with qualitative data to develop vignettes (narratives of people's experiences of living with IBD) for stakeholder workshops to generate ideas for useful, accessible and sustainable solutions for identified health literacy needs. Selection and testing of health literacy actions happens in phase 2 and implementation and evaluation in phase 3 (2021-2023). Outcomes of this project include giving voice to people living with IBD, their carers and frontline healthcare practitioners. Genuine codesign informs the development and implementation of what is needed and wanted to improve access to and availability and quality of information and resources that support people to manage their health. There is potential for other NGOs to use the CCA Ophelia model in other health contexts to improve engagement with and understanding of the needs of the people they serve and to reduce health inequalities and improve health outcomes.

ETHICS AND DISSEMINATION

Ethics approval for Ophelia phase 1 has been obtained from the Human Research Ethics Committee of Swinburne University of Technology (Ref: 20202968-4652) and by the South West Sydney Local Health District Research and Ethics Office for the purposes of questionnaire recruitment at Liverpool Hospital (Ref: 20202968-4652). Dissemination of the study findings will be the national codesign process and ownership development across the CCA community and through the genuine engagement of clinicians and relevant managers across Australia. The model and process will be directly distributed to international IBD associations and to other NGOs. It will also be disseminated through publication in a peer-reviewed journal, conference presentations and public reports on the CCA and Swinburne University of Technology website.

Progression of the functional deficit in a group of patients with bipolar disorder: a cluster analysis based on longitudinal data

We aimed to examine the trajectory of psychosocial functioning in a sample of euthymic patients with bipolar disorder (BD) throughout a 5-year follow-up. Ninety-nine euthymic bipolar patients and 40 healthy controls (HC) were included. A neurocognitive assessment (17 neurocognitive measures grouped in 6 domains) was carried out at baseline. The split version of the Global Assessment of Functioning scale (GAF-F) and the Functioning Assessment Short Test (FAST) were used to examine psychosocial functioning at baseline (T1), and after a 5-year follow-up (T2). The statistical analysis was performed through repeated measures ANOVA and hierarchical cluster analysis based on the GAF-F and the FAST scores at T1 and T2. Eighty-seven patients (87.9%) were evaluated at T2. The cluster analysis identified two groups of patients. The first group included 44 patients (50.6%) who did not show a progression of the functional impairment (BD-NPI). The second cluster, which included 43 patients (49.4%), was characterized by a progression of the functional impairment (BD-PI). The BD-PI had a higher number of relapses and a higher number of hospitalizations during the follow-up period, as well as worse neurocognitive functioning than the BD-NPI. The repeated measures ANOVA confirmed that the psychosocial performance of BD-NPI is stable while there was a progression of the functional deterioration in BD-PI. The trajectory of the psychosocial functioning of patients with BD is not homogeneous. Our results suggest that in at least one subset of patients with BD, which might account for half of the patients, the disease has a progressive course.

Co-designing eHealth and Equity Solutions: Application of the Ophelia (Optimizing Health Literacy and Access) Process

Background: The unequal access, challenges and outcomes related to using technology have created the digital divide, which leads to health inequalities. The aim of this study was to apply the Ophelia (Optimizing Health Literacy and Access) process, a widely used systematic approach to whole of community co-design, to the digital context to generate solutions to improve health and equity outcomes.

Methods: This was a mixed method study. A cross-sectional survey was undertaken at 3 health organizations in Victoria, Australia using the eHealth Literacy Questionnaire (eHLQ) as a needs assessment tool. Cluster analysis was conducted to identify subgroups with varying eHealth literacy needs. These data, combined with semi-structured interviews with clients, were used to generate vignettes representing different eHealth literacy profiles. The vignettes were presented at co-design workshops with clients and health professionals to generate solutions for digital health services improvement. Expert validation and proof-of-concept testing was explored through mapping the process against Ophelia guiding principles.

Results: The cluster analyses identified 8 to 9 clusters with different profiles of eHealth literacy needs, with 4 to 6 vignettes developed to represent the eHealth literacy strengths and weaknesses of clients at each of the 3 sites. A total of 32, 43, and 32 solutions across 10 strategies were co-created based on ideas grounded in local expertise and experiences. Apart from digital solutions, non-digital solutions were frequently recommended as a strategy to address eHealth literacy needs. Expert validation identified at least half of the ideas were very important and feasible, while most of the guiding principles of the Ophelia process were successfully applied.

Conclusion: By harnessing collective creativity through co-design, the Ophelia process has been shown to assist the development of solutions with the potential to improve health and equity outcomes in the digital context. Implementation of the solutions is needed to provide further evidence of the impact of the process. The suggested inclusion of non-digital solutions revealed through the co-design process reminds health organizations and policymakers that solutions should be flexible enough to suit individual needs. As such, taking a co-design approach to digital health initiatives will assist in preventing the widening of health inequalities.

The Susceptibles, Chancers, Pragmatists, and Fair Players: An Examination of the Sport Drug Control Model for Adolescent Athletes, Cluster Effects, and Norm Values Among Adolescent Athletes

Although there are few high-profile cases of adolescent athletes being caught doping, up to a third of young athletes may dope. In order to generate a more accurate understanding of why adolescent athletes dope, it is important to validate models that help to explain this behavior. The aims of this study were 3-fold: firstly, to test the Sport Drug Control Model for Adolescent Athletes (SDCM-AA); secondly, to generate athlete profiles that would help quantify the proportion of athletes who are at risk of doping; and thirdly, to create norm values for the Adolescent Sport Doping Inventory (ASDI), which would allow national doping organizations, sporting organizations, and clubs to benchmark the scores of their athletes for key psycho-social variables linked to doping. A total of 2208 adolescent athletes from the United Kingdom, Australia, Hong Kong, and the United States completed the ASDI. The data presented an appropriate fit to the SDCM-AA model, in which 54% of the variance in susceptibility to doping was explained in the model, and 44.8% of attitudes toward doping was accounted for. Four distinct clusters of athletes emerged: the Susceptibles (i.e., identified with the benefits of doping, were willing to cheat, and viewed little threat), the Chancers (i.e., identified with the benefits of doping, scored high on willingness to cheat, and were highly influenced by their reference group, but had an average score for threat, self-esteem, and legitimacy), the Pragmatists (i.e., did not engage with any aspects of doping, but were more susceptible than the fair players), and Fair Players (i.e., high levels of sportspersonship, unwilling to cheat, and viewed doping as a threat). The revised SDCM-AA appears a valid model that helps explain the factors associated with doping attitudes and doping susceptibility. Adolescent athletes can be classified into one of four clusters, in relation to doping. Their cluster group could influence the content of the anti-doping education they receive.

Health service use by patients with heart failure living in a community setting: a cross-sectional analysis in North West London

BACKGROUND

The complex nature of heart failure (HF) management, often involving multidimensional care, is widely recognised, but overall health service utilisation by patients with HF has not previously been described.

AIM

To describe overall health service use by adults with HF living in a community setting.

DESIGN AND SETTING

Cross-sectional analysis of prevalent HF cases from January 2015 to December 2018 using an administrative dataset covering primary and secondary care, and 'other' (community, mental health, social care) services in North West London.

METHOD

Healthcare use of each service was described overall and by individual components of secondary care (such as, outpatient appointments), and 'other' services (such as, nursing contacts). Usage patterns were identified using k-means cluster analysis, using all distinct contacts for the whole study period, and visualised with a heatmap.

RESULTS

A total of 39 301 patients with a prevalent diagnosis of HF between 1 January 2015 and 31 December 2018 were found. Of those, approximately 90% used health services during the study period, most commonly outpatient services, GP consultations, unplanned accident and emergency visits, and community services. Use of cardiology-specific services ranged from around 3% (cardiology-related community care) to around 20% (outpatient cardiology visits). GP consultations decreased by 11% over the study period. Five clusters of patients were identified, each with statistically significantly different care usage patterns and patient characteristics.

CONCLUSION

Patients with HF make heavy but heterogeneous use of services. Relatively low and falling use of GP consultations, and the apparently low uptake of community rehabilitation services by patients with HF, is concerning and suggests challenges in primary care access and integration of care.

Cannabis use in active athletes: Behaviors related to subjective effects

Cannabis use has not been well characterized in athletes. Studies primarily examine problematic use or its categorization by anti-doping bodies as a banned substance. Patterns of use, reasons for use, and responses to cannabis consumption have not been studied in a community-based sample of adult athletes. The Athlete PEACE Survey examined cannabis use patterns and subjective effects to cannabis in a community-based cohort of adult athletes. We used mainly social media and email blasts to recruit and SurveyGizmo to collect data. 1,161 (91.1%) of the 1,274 athletes taking the survey completed it. Current cannabis use was evaluated by asking "In the past two weeks, have you used marijuana (including THC and/or CBD)?" and cannabis type used was assessed by asking "What do you primarily use THC, CBD, or both?". Cannabis benefits and adverse effects (i.e. subjective effects) and patterns of use were reported. 302 athletes (26%) currently use cannabis of whom 301 had complete data for cluster analysis. Cluster analysis was used to determine cannabis user phenotypes and exploratory factor analysis (EFA) was used to create subjective effects factors. Associations between cannabis user phenotype clusters and the subjective effects factors were explored using multivariate analysis. Cluster analysis identified three statistically distinct cannabis user phenotypes: (1) older athletes who primarily use medical CBD, (2) mixed age athletes who use cannabis mainly recreationally with both THC and CBD use, and (3) mixed age athletes who used cannabis the longest with primary THC and CBD use. EFA showed three subjective effects factors: (1) Well-being, (2) Calm, and (3) Adverse. Mean positive subjective were higher than mean adverse subjective effects (p<0.001). The cluster using THC and CBD showed the highest mean scores for all three subjective effects factors (p<0.001). Athletes who use a combination of THC and CBD exhibited the most benefit to well-being and calm with minimal adverse effects. Our methodology can be used to develop real-world evidence to inform future use of medical cannabis products.

Stress, anxiety and depression in clinical nurses in Vietnam: a cross-sectional survey and cluster analysis

Background

Hospital nurses are exposed to various work-related factors that may be associated with increased risk of developing different mental disorders. Empirical evidence on the prevalence and correlates of individual mental health problems such as stress, anxiety and depression is widely reported, while a combined pattern of these conditions is unknown. This study aims to examine the co-occurrence of stress, anxiety and depression among clinical nurses, and to explore socio-demographic characteristics of, and working conditions experienced by, nurses that may be associated with these three mental health conditions.

Methods

A cross-sectional study was implemented in one tertiary hospital in Hanoi city, Vietnam, from May to September 2015. A self-reported questionnaire including a short version of the Depression, Anxiety and Stress scale 21 items and questions on demographic and work-related characteristics was delivered to 787 registered nurses. 600 completed questionnaires was used in the final analysis (76.2% response rate). The two-step clustering analysis was performed to identify sub groups. Chi square test and post hoc ANOVA analysis with Bonferroni correction were used to examine differences in psychological status, demographic characteristics and working conditions among the clusters (two-tailed p < 0.05).

Results

The prevalence of self-reported stress, anxiety and depression were 18.5%, 39.8% and 13.2%, respectively. 45.3% participants reported symptoms of at least one mental disorder, 7.3% had all three. Nurses in the first cluster (high prevalence of mental disorders), had high task demand and conflict at work with low job control and reward. The second cluster nurses (moderate percentage of mental strain) were significantly older and in marital relationship, high task demand and job control, and presence of chronic diseases. The lowest proportion of self-perceived mental disorders were observed in the cluster three who were younger and had fewer years of services, moderate task demand and low job control and better physical health in comparison with those in the other two clusters (p < 0.05).

Conclusions

Stress, anxiety and depression were prevalent among clinical nurses. Heterogeneity in demographic characteristics and working conditions were observed across clusters with different patterns of mental disorders. Institutional effort should be emphasized to support nurses in their career development to reduce psychological strains.

Electronic supplementary material

The online version of this article (10.1186/s13033-018-0257-4) contains supplementary material, which is available to authorized users.