Does fragmented cancer care affect survival? Analysis of gastric cancer patients using national insurance claim data

BMC Health Serv Res. 2022 Dec 21;22(1):1566. doi: 10.1186/s12913-022-08988-y.

Abstract

Background: We aimed to investigate the association between fragmented cancer care in the early phase after cancer diagnosis and patient outcomes using national insurance claim data.

Methods: From a nationwide sampled cohort database, we identified National Health Insurance beneficiaries diagnosed with gastric cancer (ICD-10: C16) in South Korea during 2005-2013. We analyzed the results of a multiple logistic regression analysis using the generalized estimated equation model to investigate which patient and institution characteristics affected fragmented cancer care during the first year after diagnosis. Then, survival analysis using the Cox proportional hazard model was conducted to investigate the association between fragmented cancer care and five-year mortality.

Results: Of 2879 gastric cancer patients, 11.9% received fragmented cancer care by changing their most visited medical institution during the first year after diagnosis. We found that patients with fragmented cancer care had a higher risk of five-year mortality (HR: 1.310, 95% CI: 1.023-1.677). This association was evident among patients who only received chemotherapy or radiotherapy (HR: 1.633, 95% CI: 1.005-2.654).

Conclusions: Fragmented cancer care was associated with increased risk of five-year mortality. Additionally, changes in the most visited medical institution occurred more frequently in either patients with severe conditions or patients who mainly visited smaller medical institutions. Further study is warranted to confirm these findings and examine a causal relationship between fragmented cancer care and survival.

Keywords: Fragmented cancer care; Healthcare utilization; Survival; cancer policy.

MeSH terms

  • Humans
  • Insurance*
  • Proportional Hazards Models
  • Retrospective Studies
  • Stomach Neoplasms* / therapy
  • Survival Analysis