Association between information overload, asthenopia, and depressive and anxiety symptoms in patients with lung cancer: A computer-simulated network analysis.

OBJECTIVE

To examine symptom-level associations among cancer information overload (IO), asthenopia, depressive symptoms, and anxiety symptoms in patients with lung cancer and to compare model-based node-perturbation rankings with centrality rankings from a Gaussian graphical model (GGM).

METHODS

In this cross-sectional study, 566 patients with lung cancer completed assessments of IO, asthenopia, anxiety, and depressive symptoms. A 25-node binary Ising model was estimated. Network Intervention Response Analysis (NIRA) was used to evaluate model-estimated changes in network sum scores after hypothetical node-specific threshold perturbations. A GGM based on the corresponding non-dichotomized variables was used as a sensitivity analysis, and network stability was evaluated using bootstrap procedures.

RESULTS

The Ising network showed prominent conditional associations between IO4 (perceived impossibility) and Patient Health Questionnaire (PHQ-9) item 2 (sad mood), asthenopia and PHQ-9 item 3 (sleep problems), and IO1 (decision overload) and Generalized Anxiety Disorder scale (GAD-7) item 2 (uncontrollable worry). Under the hypothetical alleviating perturbation, perturbing "Sad mood" produced the largest estimated decrease in the network sum score, from an unperturbed baseline of 12.71 to 10.85. Under the hypothetical aggravating perturbation, perturbing "Trouble relaxing" produced the largest estimated increase, from an unperturbed baseline of 12.74 to 14.63. In the GGM, information skepticism and sleep problems were among the most central nodes. The Ising and GGM absolute edge-weight patterns showed broad correspondence (Mantel r = 0.8385, P < 0.001), although asthenopia was operationalized differently across models.

CONCLUSIONS

This cross-sectional study characterized model-estimated associations among IO, asthenopia, depressive symptoms, and anxiety symptoms in patients with lung cancer. The perturbation findings are hypothesis-generating and identify candidate symptoms for future prospective testing; they do not establish clinical effectiveness, therapeutic benefit, or causal effects.