How does Linkage to the National Death Index Affect Population-Based Net Survival Estimates for Women with Cervical Cancer in Saudi Arabia?

BACKGROUND: Population-based cancer survival is a key metric for the assessment of cancer control strategies. Accurate estimation of cancer survival requires complete follow-up data for all patients. AIM: To explore the impact of linking national cancer registry data to the national death index on net survival estimates for women diagnosed with cervical cancer in Saudi Arabia during 2005-2016. METHODS: We acquired data on 1,250 Saudi women diagnosed with invasive cervical cancer during the 12- year period 2005-2016 from the Saudi Cancer Registry. These included the woman's last known vital status and the date of last known vital status, but this was restricted to information from clinical records and death certificates that mention cancer as a cause of death ("registry follow-up"). We submitted available national ID numbers to the National Information Center (NIC) of the Ministry of Interior, to ascertain the date of death, from any cause of death, for women who had died up until 31 December 2018 ("NIC follow-up"). We estimated age-standardised 5-year net survival using the Pohar-Perme estimator under five different scenarios using the two sources of follow-up, and censoring at the date of last contact with the registry versus extending survival until the closing date if no information on death was obtained. RESULTS: 1,219 women were eligible for survival analysis. Five-year net survival was lowest when using NIC followup only (56.8%; 95%CI 53.5 - 60.1%), and highest when registry follow-up only was used and survival time was extended until closure date for those with no information on death (81.8%; 95%CI 79.6 - 84%). CONCLUSION: Reliance solely on information from deaths certified as due to cancer and clinical records leads to a high proportion of missing deaths in the national cancer registry. This is probably due to low quality of certification of the cause of death in Saudi Arabia. Linkage of the national cancer registry to the national death index at the NIC identifies virtually all deaths, providing more reliable survival estimates, and it eliminates the ambiguity in determining the underlying cause of death. Therefore, this should become the standard approach to estimating cancer survival in Saudi Arabia.