Search

Background: Studies of cardiovascular disease risk prediction by machine learning algorithms often do not assess their ability to generalize to other populations and few of them include an analysis of the interpretability of individual predictions. This manuscript addresses the development and validation, both internal and external, of predictive models for the assessment of risks of major adverse cardiovascular events (MACE). Global and local interpretability analyses of predictions were conducted towards improving MACE's model reliability and tailoring preventive interventions., Methods: The models were trained and validated on a retrospective cohort with the use of data from Ribeirão Preto Medical School (RPMS), University of São Paulo, Brazil. Data from Beth Israel Deaconess Medical Center (BIDMC), USA, were used for external validation. A balanced sample of 6,000 MACE cases and 6,000 non-MACE cases from RPMS was created for training and internal validation and an additional one of 8,000 MACE cases and 8,000 non-MACE cases from BIDMC was employed for external validation. Eight machine learning algorithms, namely Penalized Logistic Regression, Random Forest, XGBoost, Decision Tree, Support Vector Machine, k-Nearest Neighbors, Naive Bayes, and Multi-Layer Perceptron were trained to predict a 5-year risk of major adverse cardiovascular events and their predictive performance was evaluated regarding accuracy, ROC curve (receiver operating characteristic), and AUC (area under the ROC curve). LIME and Shapley values were applied towards insights about model interpretability., Findings: Random Forest showed the best predictive performance in both internal validation (AUC = 0.871 (0.859-0.882); Accuracy = 0.794 (0.782-0.808)) and external one (AUC = 0.786 (0.778-0.792); Accuracy = 0.710 (0.704-0.717)). Compared to LIME, Shapley values suggest more consistent explanations on exploratory analysis and importance of features., Conclusions: Among the machine learning algorithms evaluated, Random Forest showed the best generalization ability, both internally and externally. Shapley values for local interpretability were more informative than LIME ones, which is in line with our exploratory analysis and global interpretation of the final model. Machine learning algorithms with good generalization and accompanied by interpretability analyses are recommended for assessments of individual risks of cardiovascular diseases and development of personalized preventive actions., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2024 Shimizu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Background: Health complexity includes biological, psychological, social, and health systems. Having complex health needs is associated with poorer clinical outcomes and higher healthcare costs. Care management for people with health complexity is increasingly recommended in primary health care (PHC). The INTERMED complexity assessment grid showed adequate psychometric properties in specialized settings. This study aimed to evaluate INTERMED's validity and feasibility to assess health complexity in an adult PHC population., Method: The biopsychosocial health care needs of 230 consecutive adult patients from three Brazilian PHC services were assessed using the INTERMED interview. Participants with a total score >20 were classified as "complex". Quality of life was measured using the World Health Organization Quality of Life BREF (WHOQOL-BREF); symptoms of anxiety and depression using the Hospital Anxiety and Depression Scale (HADS); social support using the Medical Outcomes Study-Social Support Survey (MOS-SSS); comorbidity levels using the Charlson Comorbidity Index (CCI). We developed two questionnaires to evaluate health services use, and patient perceived feasibility of INTERMED., Results: 42 participants (18.3%) were classified as "complex". A moderate correlation was found between the total INTERMED score and the total scores of WHOQOL-BREF (rho = - 0.59) and HADS (rho = 0.56), and between the social domains of INTERMED and MOS-SSS (rho = -0.44). After adjustment, the use of PHC (β = 2.12, t = 2.10, p < 0.05), any other health care services (β = 3.05, t = 3.97, p < 0.01), and any medication (β = 3.64, t = 4.16, p < 0.01) were associated with higher INTERMED scores. The INTERMED internal consistency was good (ω = 0.83), and the median application time was 7 min. Patients reported satisfaction with the questions, answers, and application time., Conclusion: INTERMED displayed good psychometric values in a PHC population and proved promising for practical use in PHC., Competing Interests: The authors have declared that no competing interests exist.

Middle-income countries are facing a growing challenge of adequate health care provision for people with multimorbidity. The objectives of this study were to explore the distribution of multimorbidity and to identify patterns of multimorbidity in the Brazilian general adult population. Data from 60202 adults, aged ≥18 years that completed the individual questionnaire of the National Health Survey 2013 (Portuguese: "Pesquisa Nacional de Saúde"-"PNS") was used. We defined multimorbidity as the presence of two or more chronic conditions, including self-reported diagnoses and responses to the 9-item Patient Health Questionnaire for depression. Multivariate Poisson regression analyses were used to explore relationship between multimorbidity and demographic factors. Exploratory tetrachoric factor analysis was performed to identify multimorbidity patterns. 24.2% (95% CI 23.5-24.9) of the study population were multimorbid, with prevalence rate ratios being significantly higher in women, older people and those with lowest educational level. Multimorbidity occurred earlier in women than in men, with half of the women and men aged 55-59 years and 65-69 years, respectively, were multimorbid. The absolute number of people with multimorbidity was approximately 2.5-fold higher in people younger than 65 years than older counterparts (9920 vs 3945). Prevalence rate ratios of any mental health disorder significantly increased with the number of physical conditions. 46.7% of the persons were assigned to at least one of three identified patterns of multimorbidity, including: "cardio-metabolic", "musculoskeletal-mental" and "respiratory" disorders. Multimorbidity in Brazil is as common as in more affluent countries. Women in Brazil develop diseases at younger ages than men. Our findings can inform a national action plan to prevent multimorbidity, reduce its burden and align health-care services more closely with patients' needs., Competing Interests: The authors have no competing interests to report.