Risk Factors of Suicide After Bladder Cancer Diagnosis and More Information to Development Nomogram

Thanks for your attention on our work in the relationship between suicide and bladder cancer diagnosis and we appreciate your positive comment (Li, Ma, and Li 2024) for our study (Liu et al. 2024). In our study (Liu et al. 2024), we developed and verified a nomogram for suicide risk after bladder cancer diagnosis. Data were extracted from the Surveillance Epidemiology and End Results (SEER) database, which provided the most accurate information regarding cancer incidence and survival covering approximately 28% of the population in 18 population groups (Xu et al. 2020). First, we identify some independent risk factors for suicide in bladder cancer patients. Besides, we built and validated a nomogram based on above suicide risk factors. The aim was identified high-risk suicide bladder cancer patients early, timely referral to psychiatry specialist experts for early intervention, and ultimately reduced the suicide rates.

Firstly, the term suicide is used to describe suicidal ideation, preparatory acts towards imminent suicide, suicide attempt and completed suicide (Guo et al. 2021). There are many factors that can lead to suicide, and it is extremely difficult to establish a predictive model of suicide risk. The strongest risk factor associated with suicide was mental illness, with others such as stressful life events and interpersonal difficulties also contributing. Here, we only explore the risk factors for suicide in tumour patients. In our study (Liu et al. 2024), we found four independent risk factors of suicide in patients with bladder cancer by multivariate Cox regression analysis. They were sex (male vs. female, HR = 11.419, 95% CI = 4.661–27.970, p < 0.001), marital status (married vs. single, HR = 0.592, 95% CI = 0.381–0.919, p = 0.020), local tumour extent (T2 vs. Ta/T1, HR = 1.758, 95% CI = 1.137–2.718, p = 0.011) and lymph node invasion, respectively (NX vs. N0, HR = 2.671, 95% CI = 1.166–6.122, p = 0.020). And we built and validated a nomogram based on above suicide risk factors. However, this study was a retrospective study, and lacking of crucial suicide-related information, including psychiatric disorders and previous suicide attempts. Many other studies found that age at diagnosis, gender, ethnicity, surgery, radiotherapy and annual household income were also the tumour-related suicide risk factors (Zhou et al. 2024; Katayama et al. 2023). Due to the limitation of SEER database, some information such as a history of self-harm or mental disorder was lacking, which lead to a limited number of variables in our nomogram.

Secondly, there are various ways in variable screening including logistic regression model, Cox proportional-hazards model, competing-risks model, Poisson regression model, regression using the least absolute shrinkage and selection operator (LASSO), artificial intelligence and joinpoint regression model. Logistic regression model is one of the most extensively used prediction models in medicine to predict the occurrence of one or more clinical events, such as disease, its recurrence, mortality, or recovery. It is frequently used the study for the short-term survival analysis (Lorimer et al. 2019). The Cox proportional hazards regression is one of the most commonly used regression methods for survival analysis, used to correlate multiple risk factors or exposure types with survival times (He et al. 2022). As SEER provides long-term follow-up outcome data, it is ideal for Cox regression analysis. Therefore, we chose the Cox proportional hazard regression model to screen meaningful variables.

Finally, every study has some limitations, of course. Many researchers have paid attention to the situation of suicide among tumour patients and have conducted a series of studies (Fillon 2023; Chen et al. 2024).

In conclusion, we provide a nomogram which could help urological specialists identify bladder cancer patients at risk of suicide and timely referral of psychiatric specialists to avoid unnecessary deaths. We hope more oncology specialists to refine the prediction model together and improve its predictive performance.

Liang Liu: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, writing – original draft, writing – review and editing. Jun-Hui He: data curation, formal analysis, investigation, writing – original draft. Yu Xiao: data curation, formal analysis, investigation, writing – original draft.

The authors have nothing to report.

The authors have nothing to report.

The authors declare no conflicts of interest.