Prediction models using artificial intelligence and longitudinal data from electronic health records: a systematic methodological review.

OBJECTIVE: To describe and appraise the use of artificial intelligence (AI) techniques that can cope with longitudinal data from electronic health records (EHRs) to predict health-related outcomes. METHODS: This review included studies in any language that: EHR was at least one of the data sources, collected longitudinal data, used an AI technique capable of handling longitudinal data, and predicted any health-related outcomes. We searched MEDLINE, Scopus, Web of Science, and IEEE Xplorer from inception to January 3, 2022. Information on the dataset, prediction task, data preprocessing, feature selection, method, validation, performance, and implementation was extracted and summarized using descriptive statistics. Risk of bias and completeness of reporting were assessed using a short form of PROBAST and TRIPOD, respectively. RESULTS: Eighty-one studies were included. Follow-up time and number of registers per patient varied greatly, and most predicted disease development or next event based on diagnoses and drug treatments. Architectures generally were based on Recurrent Neural Networks-like layers, though in recent years combining different layers or transformers has become more popular. About half of the included studies performed hyperparameter tuning and used attention mechanisms. Most performed a single train-test partition and could not correctly assess the variability of the model's performance. Reporting quality was poor, and a third of the studies were at high risk of bias. CONCLUSIONS: AI models are increasingly using longitudinal data. However, the heterogeneity in reporting methodology and results, and the lack of public EHR datasets and code sharing, complicate the possibility of replication. REGISTRATION: PROSPERO database (CRD42022331388).

Polypharmacy Patterns in Multimorbid Older People with Cardiovascular Disease: Longitudinal Study.

(1) Introduction: Cardiovascular disease is associated with high mortality, especially in older people. This study aimed to characterize the evolution of combined multimorbidity and polypharmacy patterns in older people with different cardiovascular disease profiles. (2) Material and methods: This longitudinal study drew data from the Information System for Research in Primary Care in people aged 65 to 99 years with profiles of cardiovascular multimorbidity. Combined patterns of multimorbidity and polypharmacy were analysed using fuzzy c-means clustering techniques and hidden Markov models. The prevalence, observed/expected ratio, and exclusivity of chronic diseases and/or groups of these with the corresponding medication were described. (3) Results: The study included 114,516 people, mostly men (59.6%) with a mean age of 78.8 years and a high prevalence of polypharmacy (83.5%). The following patterns were identified: Mental, behavioural, digestive and cerebrovascular; Neuropathy, autoimmune and musculoskeletal; Musculoskeletal, mental, behavioural, genitourinary, digestive and dermatological; Non-specific; Multisystemic; Respiratory, cardiovascular, behavioural and genitourinary; Diabetes and ischemic cardiopathy; and Cardiac. The prevalence of overrepresented health problems and drugs remained stable over the years, although by study end, cohort survivors had more polypharmacy and multimorbidity. Most people followed the same pattern over time; the most frequent transitions were from Non-specific to Mental, behavioural, digestive and cerebrovascular and from Musculoskeletal, mental, behavioural, genitourinary, digestive and dermatological to Non-specific. (4) Conclusions: Eight combined multimorbidity and polypharmacy patterns, differentiated by sex, remained stable over follow-up. Understanding the behaviour of different diseases and drugs can help design individualised interventions in populations with clinical complexity.