OBJECTIVES: Incomplete or incorrect causal theories are a key source of bias in machine learning (ML) algorithms. Community-engaged methodologies provide an avenue for mitigating this bias through incorporating causal insights from community stakeholders into ML development. In health applications, community-engaged approaches can enable the study of social drivers of health (SDOH), which are known to shape health inequities. However, it remains challenging for SDOH to inform ML algorithms, partially because SDOH variables are known to be interrelated, yet it is difficult to elucidate the causal relationships between them. Community-based system dynamics is a community-engaged methodology that can be used to cocreate formal causal graphs, called causal loop diagrams, with patients. MATERIALS AND METHODS: We used community-based system dynamics to create a causal graph representing the impacts of SDOH on the progression of chronic kidney disease, a chronic condition with SDOH-driven health disparities. We conducted focus groups with 42 participants and a day-long model building workshop with 11 participants. RESULTS: Our model building workshop resulted in a final graph comprising 16 variables, 42 causal links, and 5 subsystems of semantically related SDOH variables. CONCLUSION: This final graph, representing the causal relationships between social variables relevant to chronic kidney disease, can inform the development of clinical ML algorithms and other technological interventions.

OBJECTIVES: The inclusion of social drivers of health (SDOH) into predictive algorithms of health outcomes has potential for improving algorithm interpretation, performance, generalizability, and transportability. However, there are limitations in the availability, understanding, and quality of SDOH variables, as well as a lack of guidance on how to incorporate them into algorithms when appropriate to do so. As such, few published algorithms include SDOH, and there is substantial methodological variability among those that do. We argue that practitioners should consider the use of social indices and factors-a class of area-level measurements-given their accessibility, transparency, and quality. RESULTS: We illustrate the process of using such indices in predictive algorithms, which includes the selection of appropriate indices for the outcome, measurement time, and geographic level, in a demonstrative example with the Kidney Failure Risk Equation. DISCUSSION: Identifying settings where incorporating SDOH may be beneficial and incorporating them rigorously can help validate algorithms and assess generalizability.

OBJECTIVE: There are over 363 customized risk models of the American College of Cardiology and the American Heart Association (ACC/AHA) pooled cohort equations (PCE) in the literature, but their gains in clinical utility are rarely evaluated. We build new risk models for patients with specific comorbidities and geographic locations and evaluate whether performance improvements translate to gains in clinical utility. MATERIALS AND METHODS: We retrain a baseline PCE using the ACC/AHA PCE variables and revise it to incorporate subject-level information of geographic location and 2 comorbidity conditions. We apply fixed effects, random effects, and extreme gradient boosting (XGB) models to handle the correlation and heterogeneity induced by locations. Models are trained using 2 464 522 claims records from Optum©'s Clinformatics® Data Mart and validated in the hold-out set (N = 1 056 224). We evaluate models' performance overall and across subgroups defined by the presence or absence of chronic kidney disease (CKD) or rheumatoid arthritis (RA) and geographic locations. We evaluate models' expected utility using net benefit and models' statistical properties using several discrimination and calibration metrics. RESULTS: The revised fixed effects and XGB models yielded improved discrimination, compared to baseline PCE, overall and in all comorbidity subgroups. XGB improved calibration for the subgroups with CKD or RA. However, the gains in net benefit are negligible, especially under low exchange rates. CONCLUSIONS: Common approaches to revising risk calculators incorporating extra information or applying flexible models may enhance statistical performance; however, such improvement does not necessarily translate to higher clinical utility. Thus, we recommend future works to quantify the consequences of using risk calculators to guide clinical decisions.

Zero-shot transfer learning for multi-domain dialogue state tracking can allow us to handle new domains without incurring the high cost of data acquisition.This paper proposes new zero-short transfer learning technique for dialogue state tracking where the in-domain training data are all synthesized from an abstract dialogue model and the ontology of the domain.We show that data augmentation through synthesized data can improve the accuracy of zero-shot learning for both the TRADE model and the BERT-based SUMBT model on the MultiWOZ 2.1 dataset.We show training with only synthesized in-domain data on the SUMBT model can reach about 2/3 of the accuracy obtained with the full training dataset.We improve the zero-shot learning state of the art on average across domains by 21%.