"When symptoms interact and change": longitudinal network analysis of dynamic symptom changes in oral cancer patients.

PURPOSE: Oral cancer patients experience a heavy and fluctuating symptom burden during the perioperative period, which may compromise recovery and quality of life (QoL). This study aimed to characterize dynamic symptom networks to identify core symptoms and their trajectories across surgical stages.

METHODS

In this longitudinal study, 597 oral cancer patients were recruited from a tertiary hospital in China between September 2023 and October 2024. Symptoms were assessed using the MD Anderson Symptom Inventory for Head and Neck Cancer (MDASI-H&N) at admission (T1), postoperative day 3 (T2), day 7 (T3), and 1 month after surgery (T4). Network analysis was applied to identify core symptoms and clusters, and cross-lagged panel modeling was used to examine temporal inter-symptom relationships.

RESULTS

A total of 527 patients were analyzed. Across the perioperative course, up to four clusters emerged; oral function-related and psychosomatic clusters consistently dominated. At T1-T2, the strongest cross-lagged effect was observed between distress and sleep disturbance (βCL = 0.13), with restless sleep showing the greatest strength (rIS = 2.38). From T2-T3, dental/gum problems predicted swallowing/chewing difficulty (βCL = 0.15), which had the highest in-strength (rIS = 0.96). QoL was most negatively associated with fatigue at T1 (r = -0.13) and with swallowing/chewing difficulty at T4 (r = -0.40).

CONCLUSION

Core symptom clusters remained stable, but specific core and bridging symptoms shifted by stage. Symptom complexity peaked on postoperative day 3, identifying a critical intervention window. These findings provide evidence to guide precise, stage-specific symptom management and improve outcomes in oral cancer care.