Detecting Conditional Dependence Using Flexible Bayesian Latent Class Analysis

Background: Epidemiological research has indicated higher polysubstance use among individuals who misuse opioids in the United States. Polysubstance use is also associated with multiple emergency department (ED) engagements in intervention services. Less is known about polysubstance use among individuals who use opioids or were engaged in an ED setting as part of an opioid recovery intervention program. Examining emerging patterns of polysubstance use among individuals who use opioids may help shape policy and practice around developing drug markets and social norms. The objective of this study was to understand subpopulations who demonstrated distinct patterns of polysubstance use among participants in an ED Opioid Overdose Recovery Program (OORP) in New Jersey. Method: This study investigated patterns of polysubstance use among participants in a New Jersey OORP using latent class analysis (LCA; N = 1690; 70.2% male; 84.7% White non-Hispanic; meanage = 35.72 ± 11.95). Multinomial logistic regression analyses examined latent class (LC) membership based on clinically relevant indicators. Results: Five latent classes were identified (Model fit: L 2 = 33.76, BIC = 4482.69, AIC = 4245.01, and the bootstrap L 2 p-value =0.27; standard R 2 value of 0.85). Multinomial logistic regression analyses showed that subgroups with the greatest odds of prior OORP engagement related to those participants in subgroups with heroin and polysubstance use. Conclusion: It is critical to examine the heterogeneity among people engaging in prescription opioid use and polysubstance use in the United States. This study represents an important contribution in identifying polysubstance use LC membership groups and associations with clinically relevant indicators among those engaged in an opioid recovery program. Results can be valuable in understanding the nature of the larger epidemic and how it can be addressed.

We aimed to estimate latent classes of concurrent polysubstance use and test for sexual orientation differences in latent class memberships with representative data from adolescents living in 19 U.S. states. We also tested whether sex, race/ethnicity, and age moderated the sexual identity differences in polysubstance use class memberships. We analyzed data from 119,437 adolescents from 19 states who participated in the 2015 Youth Risk Behavior Survey. Latent class analysis characterized polysubstance use patterns based on self-reported frequency of lifetime and past-month use of alcohol (including heavy episodic drinking), tobacco (cigarettes, cigars, and smokeless tobacco), and marijuana. Multinomial logistic regression models tested differences in latent class memberships by sexual identity. Interaction terms tested whether sex, race/ethnicity, and age moderated the sexual identity differences in polysubstance use class memberships. A six-class model of polysubstance use fit the data best and included nonusers (61.5%), experimental users (12.2%), marijuana-alcohol users (14.8%), tobacco-alcohol users (3.8%), medium-frequency three-substance users (3.6%), and high-frequency three-substance users (4.1%). Gay/lesbian- and bisexual-identified adolescents had significantly higher odds than heterosexual-identified adolescents of being in all of the user classes compared with the nonuser class. These sexual identity differences in latent polysubstance use class memberships were generally larger for females than for males, varied occasionally by race/ethnicity, and were sometimes larger for younger ages. Compared with their heterosexual peers, gay/lesbian and bisexual adolescents-especially females-are at heightened risk of engaging in multiple types of polysubstance use. Designing, implementing, and evaluating interventions will likely reduce these sexual orientation disparities.

Health related quality of life in patient with type 2 diabetes: The role of household food insecurity on latent class membership

Understanding farmers’ mindsets is important to improve antimicrobial stewardship in the dairy industry. This cross-sectional study aimed to determine farmers’ knowledge, attitude, and practices with respect to lactational intramammary antimicrobial use (AMU) and antimicrobial resistance (AMR) in Swiss dairy herds. Based on their approach towards subclinical mastitis (SCM) and non-severe cases of clinical mastitis (CM), subgroups of farmers were identified and compared regarding their knowledge, attitude and practices towards AMU and AMR. After conducting qualitative interviews to develop a questionnaire, an online survey was sent to 1296 randomly selected Swiss dairy farmers. Information was gathered on demographic data and farmers’ knowledge, attitude, and practices towards AMU and AMR. A latent class analysis was performed to identify subgroups of farmers based on management of SCM and non-severe CM cases. Based on the results of 542 completed questionnaires, poor knowledge with respect to AMU and AMR was identified, as well as discrepancies between farmers’ perceptions and their actual practices. Farmers approached cows with SCM and non-severe CM similarly, indicating they perceived both mastitis states as the same disease. Intramammary antimicrobial products containing cefquinome, which is a highest priority critically important antimicrobial, were among the 3 most commonly applied intramammary antimicrobials. Five latent classes of farmers were identified based on their management towards SCM and non-severe CM. One group of farmers (18.5% of respondents) indicated that they did not treat those mastitis cases, one group only treated SCM cases (13.8% of respondents), one group only treated non-severe CM cases (28.6% of respondents) and the largest group treated both mastitis states (39.1% of respondents). The latter group was subdivided into a latent class of farmers following guidelines for AMU/AMR (25.5% of respondents) and a group of farmers that were not strictly following these guidelines (13.7% of respondents). Regional differences between farmers, according to altitude and language region, explained some of the variation in latent class membership. Latent class membership was associated with farmers’ attitude to use antimicrobials as little as possible and with using antimicrobials only after performing bacteriological and susceptibility testing. This study gave detailed insight into Swiss farmers’ knowledge, attitude, and practices regarding AMU and AMR and provides opportunities to improve antimicrobial stewardship in Swiss dairy herds. The identified groups of farmers, based on their management practices regarding SCM and non-severe CM, may help to design tailored intervention strategies for improving prudent AMU in the heterogeneous population of dairy farmers in Switzerland.

The Latent Class Structure of Exercise and Eating Behavior

BackgroundHeterogeneity in patients with low back pain (LBP) is well recognised and different approaches to subgrouping have been proposed. Latent Class Analysis (LCA) is a statistical technique that is increasingly being used to identify subgroups based on patient characteristics. However, as LBP is a complex multi-domain condition, the optimal approach when using LCA is unknown. Therefore, this paper describes the exploration of two approaches to LCA that may help improve the identification of clinically relevant and interpretable LBP subgroups.MethodsFrom 928 LBP patients consulting a chiropractor, baseline data were used as input to the statistical subgrouping. In a single-stage LCA, all variables were modelled simultaneously to identify patient subgroups. In a two-stage LCA, we used the latent class membership from our previously published LCA within each of six domains of health (activity, contextual factors, pain, participation, physical impairment and psychology) (first stage) as the variables entered into the second stage of the two-stage LCA to identify patient subgroups. The description of the results of the single-stage and two-stage LCA was based on a combination of statistical performance measures, qualitative evaluation of clinical interpretability (face validity) and a subgroup membership comparison.ResultsFor the single-stage LCA, a model solution with seven patient subgroups was preferred, and for the two-stage LCA, a nine patient subgroup model. Both approaches identified similar, but not identical, patient subgroups characterised by (i) mild intermittent LBP, (ii) recent severe LBP and activity limitations, (iii) very recent severe LBP with both activity and participation limitations, (iv) work-related LBP, (v) LBP and several negative consequences and (vi) LBP with nerve root involvement.ConclusionsBoth approaches identified clinically interpretable patient subgroups. The potential importance of these subgroups needs to be investigated by exploring whether they can be identified in other cohorts and by examining their possible association with patient outcomes. This may inform the selection of a preferred LCA approach.

Estimation of prevalence and diagnostic test accuracy in tuberculosis (TB) prevalence surveys suffer from reference standard and verification biases. The former is attributed to the imperfect reference test used to bacteriologically confirm TB disease. The latter occurs when only the participants screening positive for any TB-compatible symptom or chest X-ray abnormality are selected for bacteriological testing (verification). Bayesian latent class analysis (LCA) alleviates the reference standard bias but suffers verification bias in TB prevalence surveys. This work aims to identify best-practice approaches to simultaneously alleviate the reference standard and verification biases in the estimates of pulmonary TB prevalence and diagnostic test performance in TB prevalence surveys. We performed a secondary analysis of 9869 participants aged ≥15 years from a community-based multimorbidity screening study in a rural district of KwaZulu-Natal, South Africa (Vukuzazi study). Participants were eligible for bacteriological testing using Xpert Ultra and culture if they reported any cardinal TB symptom or had an abnormal chest X-ray finding. We conducted Bayesian LCA in five ways to handle the unverified individuals: (i) complete-case analysis, (ii) analysis assuming the unverified individuals would be negative if bacteriologically tested, (iii) analysis of multiply-imputed datasets with imputation of the missing bacteriological test results for the unverified individuals using multivariate imputation via chained equations (MICE), and simultaneous imputation of the missing bacteriological test results in the analysis model assuming the missing bacteriological test results were (iv) missing at random (MAR), and (v) missing not at random (MNAR). We compared the results of (i)-(iii) to the analysis based on a composite reference standard (CRS) of Xpert Ultra and culture. Through simulation with an overall true prevalence of 2.0%, we evaluated the ability of the models to alleviate both biases simultaneously. Based on simulation, Bayesian LCA with simultaneous imputation of the missing bacteriological test results under the assumption that the missing data are MAR and MNAR alleviate the reference standard and verification biases. CRS-based analysis and Bayesian LCA assuming the unverified are negative for TB alleviate the biases only when the true overall prevalence is <3.0%. Complete-case analysis produced biased estimates. In the Vukuzazi study, Bayesian LCA with simultaneous imputation of the missing bacteriological test results under the MAR and MNAR assumptions produced overall PTB prevalence of 0.9% (95% Credible Interval (CrI): 0.6-1.9) and 0.7% (95% CrI: 0.5-1.1) respectively alongside realistic estimates of overall diagnostic test sensitivity and specificity with substantially overlapping 95% CrI. The CRS-based analysis and Bayesian LCA assuming the unverified were negative for TB produced 0.7% (95% CrI: 0.5-0.9) and 0.7% (95% CrI: 0.5-1.2) overall PTB prevalence respectively with realistic estimates of overall diagnostic test sensitivity and specificity. Unlike CRS-based analysis, Bayesian LCA of multiply-imputed data using MICE mitigates both biases. The findings demonstrate the efficacy of these advanced techniques in alleviating the reference standard and verification biases, enhancing the robustness of community-based screening programs. Imputing missing values as negative for bacteriological tests is plausible under realistic assumptions.

Evaluation of three serological tests for diagnosis of bovine brucellosis in smallholder farms in Pakistan by estimating sensitivity and specificity using Bayesian latent class analysis

IntroductionType 2 diabetes (T2D) is a chronic metabolic disorder that is associated with reduced sleep quality and anxiety, and can cause a decrease in quality of life in these patients. Despite previous studies investigating these factors, few studies have examined their co-occurrence in these patients. To address this research gap, the present study aimed to determine the subgroups of patients with type 2 diabetes based on quality of life, sleep quality, and anxiety in the subgroups of using latent class analysis (LCA).MethodsThis cross-sectional study was conducted using multistage random sampling. A total of 308 patients with type 2 diabetes were randomly selected from health centers in Ardabil. All participants completed four sets of checklists and questionnaires, (Demographic characteristics, 12-item Short Form survey, the Pittsburgh Sleep Quality Index and Generalized Anxiety Disorder 7-item). Data analysis was performed using Analysis of Variance (ANOVA), chi square and latent class analysis.ResultsThree latent classes were identified: The first class (good status) included 56.4% of the participants. Also, the second (moderate status) and third (poor status) classes described 16.5% and 27.1% of the participants, respectively. In latent class 1, the probability of having good quality of life and good sleep quality was higher. In latent class 2, the probability of having moderate quality of life and poor sleep quality was higher. However, these patients revealed no anxiety. Those with third latent class membership were more likely to have moderate quality of life, poor sleep quality, and severe anxiety.ConclusionThis study showed that sleep quality and anxiety is positively related to quality of life in patients with type 2 diabetes. In addition, this study indicated the co-occurrence of sleep quality and anxiety in these patients. Based on these findings, effective and targeted interventions can be designed to improve the health status and quality of life of these patients, taking into account sleep quality and anxiety.

BackgroundAdverse childhood experiences (ACEs) are prevalent, impact long-term physical and mental health, and are associated with eating disorders (EDs) in adulthood. The primary objectives of the current study were: (1) to examine and compare ACEs between two samples: treatment-seeking adults, and a nationally representative sample of adults, (2) to characterize ACEs items and total scores across demographic and diagnostic information in adults seeking treatment for an ED, (3) to statistically classify ACEs profiles using latent class analysis, and (4) to examine associations between ACEs profiles and diagnosis.MethodsThis cross-sectional study assessed patients with a DSM-5 ED receiving treatment between October 2018 and April 2020 at the inpatient, residential, or partial hospitalization levels of care at one of two private ED treatment facilities. ACEs were assessed with the Adverse Childhood Experiences Survey at admission. Generalized linear models and Welch’s t-tests were used to compare ACEs in the current sample with national estimates. A latent class analysis was conducted to examine subgroups of ACEs responses, and differences in these classes by ED diagnoses were examined with multinomial logistic regression.ResultsPatients with EDs had significantly higher ACEs scores (M = 1.95, SD = 1.90) than the nationally representative sample (M = 1.57, SD = 4.72; t = 6.42, p < .001). Within patients with EDs, four latent classes of ACEs item endorsement were identified. Patients with other specified feeding or eating disorder (OSFED) and binge eating disorder (BED) were more likely to fall into the “Household ACEs” and “Abuse ACEs” groups, respectively, compared to anorexia nervosa—restricting subtype (AN-R).ConclusionsPatients with EDs reported more ACEs than the nationally representative sample, and differences in total ACEs and latent class membership were found across ED diagnoses. The current study can inform the development of trauma-informed care for patients with EDs.

Externalizing and internalizing behavioral problems occurs at a high rate among children. However, this has rarely been examined among Japanese children using a person-oriented method. Hence, this study aims to explore its subtypes and clarify their association with family-based group activities. We conducted a cross-sectional survey in a typical community-based suburban area for all families with primary school children in Japan. We investigated children’s internalizing and externalizing behaviors based on the Japanese version of the Strengths and Difficulties Questionnaire (SDQ), and different types of activities that family members frequently engaged in. Data from 206 families were collected and used for the analysis. The subtypes were explored using latent class analysis (LCA). The relationship between family activities and latent class membership was analyzed using a logistic regression model. Moreover, three latent class models and their probabilities were identified, namely, risk group (31.3%), moderate group (44.9%), and normal group (23.8%). Frequent family activities including play sports, traveling or hiking, watching TV and communicating, cooking or making a dessert, and doing housework, which were significantly related to the normal group. These results would add evidence to potential types of children’s behavioral problems and preventive childcare practices needed in the primary gate of families.

Metabolic syndrome is known as a lifestyle disease that results from poor health behaviors. Yet, little is known about the subgroups of midlife women with distinct health behavior profiles who are at risk for developing metabolic syndrome. This study aims to identify latent subgroups of midlife women with distinct health behavior profiles (physical activity, alcohol, diet, and smoking), to describe the characteristics of latent subgroups, and to examine the association between latent class membership and future development of metabolic syndrome. This is a secondary data analysis using baseline and follow-up data from years 1, 3, 5, and 7 (N = 3,100) from the Study of Women's Health Across the Nation (SWAN). Latent class analysis was used to identify latent subgroups of midlife women based on their distinct health behavior profiles. Bivariate and multiple logistic regression was conducted to examine the individual characteristics of each latent subgroup and its association with the future development of metabolic syndrome. A 4-class model was selected: Class 1 (Healthy), Class 2 (Healthy except alcohol), Class 3 (Healthy except diet), and Class 4 (Unhealthy). Significant differences in individual characteristics were found among the four latent classes (p < .001). The regression analysis found that Class 2 had lower odds of developing metabolic syndrome at all future visits with statistical significance reached at visit 3 (p < .05) while Class 4 had higher odds of developing metabolic syndrome at all visits except visit 3 when both compared to Class 1. Clinicians should use the study findings to offer personalized approach to promote healthy behaviors and to guide future development of health promotion programs for midlife women.

Evidence demonstrates sexual orientation disparities in physical and sexual victimization and alcohol misuse and consequences among female youth; however, most extant research has used variable-centered approaches. The current study used latent class analysis (LCA), a person-centered approach, to: (1) model female youths' physical and sexual victimization; (2) examine sexual orientation disparities in physical and sexual victimization latent class membership; (3) and use physical and sexual victimization latent class membership to predict female youths' engagement in alcohol misuse and related consequences. Participants were 7,185 youth assigned female sex at birth (77.0% heterosexual, 12.8% bisexual, 2.3% gay or lesbian; 46.8% racial minority) in grades 9 - 12 who completed the 2017 Youth Risk Behavior Survey - a school-based, cross-sectional survey. LCA uncovered four classes: (1) "Poly-Victimization Class," (2) "No Victimization Class," (3) "Past-Year Sexual Victimization Class," and (4) "Lifetime Rape Class." Sexual orientation emerged as a significant predictor of class membership. Latent classes 3 and 4 were uniquely associated with alcohol misuse and consequences (i.e., binge drinking, riding with a drinking driver, and drinking and driving) among female youth. Our findings help to elucidate the patterns of physical and sexual victimization regarding timeline, tactic used, and context among female youth; sexual orientation disparities in latent class membership; and alcohol misuse correlates of class membership. Prevention efforts aimed at reducing physical and sexual victimization may ultimately reduce female youths' risk of binge drinking, riding with drinking drivers, and drinking and driving.

Autistic youth experience high rates of adverse childhood experiences (ACEs), yet little is known about how different ACEs are associated with mental and physical health outcomes. The aims of this study were to identify latent classes of ACEs in autistic youth and examine how ACEs, measured via cumulative risk or latent class approaches, were associated with mental and physical health outcomes in autistic youth. We conducted a secondary analysis of the 2021-2022 National Survey of Children's Health, a U.S. cross-sectional dataset. The sample included 1,332 autistic youth aged 12-17. A latent class analysis (LCA) of 12 ACE indicators was conducted. Logistic regressions examined associations between ACEs and mental health (anxiety or depression diagnosis) and physical health (poor/fair vs. good/excellent health), controlling for youth characteristics. For each outcome, two regressions were conducted: one using cumulative ACE count (0-12) and one using latent class membership. A 3-class LCA model best fit the data: (1) Low ACEs, (2) Bullying and Discrimination, and (3) Household Disruption and Community Violence. For each outcome, cumulative ACE count and latent classes predicted similar variability. Compared to the Low ACEs class, youth in Class 2 and Class 3 were > 120% more likely to have a diagnosis of anxiety or depression; youth in Class 2 were > 75% more likely to have poor physical health. Findings highlight distinct patterns of ACEs among autistic youth and their links to health outcomes. Tailoring supports to specific types of adversities experienced by youth may foster resilience and promote well-being.

BackgroundBreast cancer patients commonly report using >1 form of complementary and alternative medicine (CAM). However, few studies have attempted to analyze predictors and outcomes of multiple CAM modalities. We sought to group breast cancer patients by clusters of type and intensity of complementary and alternative medicine (CAM) use following diagnosis.MethodsDetailed CAM use following breast cancer diagnosis was assessed in 2002–2003 among 764 female residents of Long Island, New York diagnosed with breast cancer in 1996–1997. Latent class analysis (LCA) was applied to CAM modalities while taking into account frequency and intensities.ResultsFour distinct latent classes of CAM use emerged: 1) “Low-dose supplement users” (40 %), who used only common nutritional supplements; 2) “Vitamin/mineral supplement users” (39 %), using an abundance of supplements in addition to other practices; 3) “Mind-body medicine users” (12 %), with near-universal use of supplements, mind-body medicine techniques, and massage; and 4) “Multi-modality high-dose users” (9 %), who were highly likely to use nearly all types of CAM. Predictors of membership in classes with substantial CAM use included younger age, more education, higher income, Jewish religion, ideal body mass index, higher fruit and vegetable intake, higher levels of physical activity, receipt of adjuvant chemotherapy, and prior use of oral contraceptives.ConclusionsLCA identified important subgroups of breast cancer patients characterized by varying degrees of complementary therapy use. Further research should explore the reproducibility of these classes and investigate the association between latent class membership and breast cancer outcomes.