Temporal dynamics of patient complaints in new hospitals: A dynamic time warping and impulse response function analysis of the South Korean healthcare system.

Background: Patient complaints are critical indicators for improving healthcare quality and patient safety, offering insights into systemic issues within medical institutions. Newly operational hospitals, owing to operational instability and limited resources, are more prone to frequent complaints, reflecting several service delivery problems. Despite the significance of these complaints, systematic analyses of how they evolve and how each complaint domain influences patient harm in new South Korean hospitals have been scarce.

Method: We analysed 1,113 patient complaints collected from March 2022 to June 2024 from a newly operational hospital in South Korea. These complaints were categorised into three major domains: Clinical (Quality, Safety), Management (Environment, Institutional Processes), and Relationship (Listening, Communication, Respect and Patient Rights) using the Healthcare Complaints Analysis Tool. Based on a vector autoregression model, we employed dynamic time warping to examine temporal patterns among the complaints and impulse response function analysis to assess the influence of these factors on patient harm over time.

Results: The results identified that in the early stages of hospital operations, Clinical Safety and Patient Harm were closely clustered, indicating a strong association between clinical factors and patient harm. In the mid-term period (September 2022 to February 2023), Respect and Patient Rights and Environment became major clusters, indicating a shift towards Management and Relationship issues as primary causes of complaints. In the later term (March 2023 to June 2024), Environment became clustered with Listening and Communication, whereas Patient Harm became more independent. Impulse response function analysis showed that Clinical factors had a strong, immediate positive impact (0.91017, p < 0.01 on Day 1) on patient harm that gradually decreased over the subsequent ten-day period. Management factors exhibited a smaller initial effect (0.00756, p < 0.05 on Day 1) but demonstrated a gradual, cumulative pattern over time. Relationship factors initially had a minimal impact (0.04551, p < 0.1 on Day 1) but became more significant in later days (0.06566, p < 0.01 on Day 3).

Conclusion: Clinical safety has an immediate, significant negative impact on patients in new hospitals, requiring prompt attention. As hospitals stabilise, Management and Relationship factors play a significant role in patient safety. Our findings underscore the need for dynamic resource allocation and strategic planning that evolves from focusing on clinical safety to incorporating management improvements and patient-centred communication strategies to enhance patient safety and service quality in new hospitals in South Korea.