Artificial intelligence (AI) is revolutionizing healthcare by providing tools for disease prediction, diagnosis, and patient management. This review focuses on two key AI methodologies in healthcare: predictive modeling and causal inference. Predictive models excel in identifying patterns to forecast outcomes but are limited in explaining the underlying causes. In contrast, causal inference focuses on understanding cause-and-effect relationships, which makes effective medical interventions possible. Although randomized controlled trials (RCTs) are the gold standard for causal inference, they face limitations including cost and ethical concerns. As alternatives, emulated RCTs and advanced machine learning techniques have emerged for estimating causal effects, bridging the gap between prediction and causality. Additionally, Shapley values and Local Interpretable Model-Agnostic Explanations improve the interpretability of complex AI models, making them more actionable in clinical settings. Integrating prediction and causal inference holds great promise for advancing personalized medicine, enhancing patient outcomes, and optimizing healthcare delivery. However, careful application of AI tools is crucial to avoid misinterpretation and maximize their potential.

Artificial intelligence (AI) is revolutionizing healthcare by providing tools for disease prediction, diagnosis, and patient management. This review focuses on two key AI methodologies in healthcare: predictive modeling and causal inference. Predictive models excel in identifying patterns to forecast outcomes but are limited in explaining the underlying causes. In contrast, causal inference focuses on understanding cause-and-effect relationships, which makes effective medical interventions possible. Although randomized controlled trials (RCTs) are the gold standard for causal inference, they face limitations including cost and ethical concerns. As alternatives, emulated RCTs and advanced machine learning techniques have emerged for estimating causal effects, bridging the gap between prediction and causality. Additionally, Shapley values and Local Interpretable Model-Agnostic Explanations improve the interpretability of complex AI models, making them more actionable in clinical settings. Integrating prediction and causal inference holds great promise for advancing personalized medicine, enhancing patient outcomes, and optimizing healthcare delivery. However, careful application of AI tools is crucial to avoid misinterpretation and maximize their potential.

Artificial intelligence (AI) is revolutionizing healthcare by providing tools for disease prediction, diagnosis, and patient management. This review focuses on two key AI methodologies in healthcare: predictive modeling and causal inference. Predictive models excel in identifying patterns to forecast outcomes but are limited in explaining the underlying causes. In contrast, causal inference focuses on understanding cause-and-effect relationships, which makes effective medical interventions possible. Although randomized controlled trials (RCTs) are the gold standard for causal inference, they face limitations including cost and ethical concerns. As alternatives, emulated RCTs and advanced machine learning techniques have emerged for estimating causal effects, bridging the gap between prediction and causality. Additionally, Shapley values and Local Interpretable Model-Agnostic Explanations improve the interpretability of complex AI models, making them more actionable in clinical settings. Integrating prediction and causal inference holds great promise for advancing personalized medicine, enhancing patient outcomes, and optimizing healthcare delivery. However, careful application of AI tools is crucial to avoid misinterpretation and maximize their potential.

Artificial intelligence (AI) is revolutionizing healthcare by providing tools for disease prediction, diagnosis, and patient management. This review focuses on two key AI methodologies in healthcare: predictive modeling and causal inference. Predictive models excel in identifying patterns to forecast outcomes but are limited in explaining the underlying causes. In contrast, causal inference focuses on understanding cause-and-effect relationships, which makes effective medical interventions possible. Although randomized controlled trials (RCTs) are the gold standard for causal inference, they face limitations including cost and ethical concerns. As alternatives, emulated RCTs and advanced machine learning techniques have emerged for estimating causal effects, bridging the gap between prediction and causality. Additionally, Shapley values and Local Interpretable Model-Agnostic Explanations improve the interpretability of complex AI models, making them more actionable in clinical settings. Integrating prediction and causal inference holds great promise for advancing personalized medicine, enhancing patient outcomes, and optimizing healthcare delivery. However, careful application of AI tools is crucial to avoid misinterpretation and maximize their potential.

We present a case of continuous ambulatory peritoneal dialysis peritonitis caused by Citrobacter freundii complicated by a fungal infection with abscess formation. A 34-year-old woman was admitted to our hospital with abdominal pain. Isolate cultures were confirmed as Citrobacter freundii by DNA sequencing of the 16s ribosomal ribonucleic acid (RNA). Antibiotic therapy was ineffective and Candida tropicalis was isolated in follow-up blood cultures. We administered an antifungal agent and removed the peritoneal catheter. A sudden fever developed, and abdominal computed tomography showed intra-abdominal abscesses. Percutaneous drainage was performed, but no bacteria were cultured. After draining the abscesses, the patient recovered. Citrobacter species are unusual pathogens in peritonitis, and fungal peritonitis is a serious complication of bacterial peritonitis. Indwelling catheters should be removed and appropriate antibiotic therapy provided. Suspicion of a fungal infection combined with bacterial peritonitis will improve the prognosis of patients on peritoneal dialysis.
Abdominal Abscess ; Abdominal Pain ; Abscess ; Adult ; Bacteria ; Candida tropicalis ; Catheters ; Catheters, Indwelling ; Citrobacter ; Citrobacter freundii* ; Drainage ; Female ; Fever ; Follow-Up Studies ; Humans ; Peritoneal Dialysis ; Peritoneal Dialysis, Continuous Ambulatory* ; Peritonitis* ; Prognosis ; RNA ; RNA, Ribosomal, 16S* ; Sequence Analysis, DNA

Background/Aims: Hypertension is considered a risk factor in immunoglobulin A nephropathy (IgAN). However, after IgAN diagnosis, the relationship between early blood pressure control and renal prognosis remains unclear. This study aimed to analyze the association between the prognosis of IgAN patients and a controlled status of hypertension within the first year of IgAN diagnosis. Methods: We retrospectively analyzed 2,945 patients diagnosed with IgAN by renal biopsy. The patients were divided into ‘normal,’ ‘new-onset,’ ‘well-controlled,’ and ‘poorly-controlled’ groups using blood pressure data from two consecutive measurements performed within a year. The Kaplan-Meier survival analysis and Cox proportional-hazards regression model were used to survey the independent association between recovery from hypertension and the risk of IgAN progression. The primary endpoint was IgAN progression defined as the initiation of dialysis or kidney transplantation. Results: Before IgAN diagnosis, 1,239 patients (42.1%) had been diagnosed with hypertension. In the fully adjusted Cox proportional-hazards models, the risk of IgAN progression increased by approximately 1.7-fold for the prevalence of hypertension. In the subgroup analyses, the ‘well-controlled’ group showed a statistically significant risk of IgAN progression (hazard ratio [HR], 3.19; 95% confidence interval [CI], 1.103 to 9.245; p = 0.032). Moreover, the ‘new-onset’ and ‘poorly-controlled’ groups had an increased risk of IgAN progression compared to the ‘normal’ group (HR, 2.58; 95% CI, 1.016 to 6.545; p = 0.046 and HR, 3.85;95% CI, 1.541 to 9.603; p = 0.004, respectively). Conclusions Although hypertension was well-controlled in the first year after IgAN diagnosis, it remained a risk factor for IgAN progression.

Several studies have reported that depression is prevalent in patients with diabetes or chronic kidney disease. However, the relationship between weight changes and the risk of depression has not been elucidated in patients with diabetic kidney disease (DKD). Methods: From the Korean National Health Insurance Service database, we selected 67,866 patients with DKD and body weight data from two consecutive health examinations with a 2-year interval between 2009 and 2012. Weight change over 2 years was categorized into five groups: ≥–10%, <–10% to ≥–5%, <–5% to <5%, ≥5% to <10%, and ≥10%. The occurrence of depression was monitored via the codes of International Statistical Classification of Diseases, 10th revision through the end of 2018. Results: During the 5.24-year follow-up, 17,023 patients with DKD developed depression. Weight change and the risk of depression had a U-shaped relationship: patients with ≥–10% weight change (hazard ratio [HR], 1.12) and those with ≥10% weight change (HR, 1.11) showed higher HRs for depression than those with <–5% to <5% weight change, even after adjusting for several confounding factors. In the subgroup analyses, the risk of depression tended to increase as weight gain or weight loss increased in all subgroups. Conclusion: Both weight loss and weight gain increased the risk of depression in patients with DKD.

Background/Aims: The neutrophil-to-lymphocyte ratio (NLR) has a prognostic value in cardiovascular disease, infection, inflammatory disease, and several malignancies. Therefore, the NLR has a possible predictive value in patients with chronic kidney disease (CKD), but this predictive value has not been validated. Here, we aimed to investigate the possibility of NLR as a predictor of CKD progression. Methods: This retrospective observational study included 141 patients with non-dialysis CKD. The participants were divided into terciles (T1, T2, and T3) according to NLR. The primary outcome was defined as a composite kidney event, which included a decline in the estimated glomerular filtration rate (eGFR) of at least 50% or initiation of renal replacement therapy during the follow-up period. Results: The mean follow-up duration was 5.45 ± 2.11 years. The mean NLRs were 1.35 ± 0.05 in T1 (n = 47), 2.16 ± 0.04 in T2 (n = 47), and 4.29 ± 0.73 in T3 (n = 47). The group with the highest NLR (T3) had higher baseline CKD and serum creatinine and lower eGFR levels than the group with the lowest NLR (T1). The cumulative incidence rate of composite kidney events was significantly higher in T3 compared with T1 (p < 0.001, log-rank test). Cox regression analysis revealed that high NLR was associated with the risk of composite kidney events (adjusted hazard ratio, 3.33; 95% confidence interval, 1.43–7.76). Conclusions A higher NLR reflects the more advanced stage of CKD and suggests a role for NLR as a biomarker for predicting CKD progression.