The Effect of Marital Status on Suicide Risk Among Patients with Breast Cancer: A Population-Based sIPTW Competing Risk Analysis

Association Between Marital Status and Increased Suicide Mortality Risk in Breast Cancer Patients

As breast cancer survival rates improve globally, the medical community is shifting its focus. The focus moves from acute clinical management to the complexities of long-term survivorship. While the biological battle against a tumor is well-documented, the psychological toll remains a crisis. Survivors often face depression, anxiety, and profound loneliness.

Recent data suggest that suicide is a critical concern among these survivors. Suicidal ideation occurs in roughly 10% of women with breast cancer. Existing research has pointed toward certain risk factors. However, the specific influence of social support structures remains poorly quantified. Researchers have lacked a way to distinguish between those who die from cancer and those who die by suicide. This distinction is vital for understanding the true impact of social isolation. A new population-based study addresses this gap. It reveals that unmarried or non-partnered status is associated with a 34% higher risk of suicide mortality among breast cancer patients (sHR = 1.34, 95% CI: 1.12–1.60). Notably, this risk is significantly mitigated in high-income environments.

The blind spot in mortality modeling

To understand why this study is necessary, one must understand the mathematical trap of "competing risks." In traditional epidemiology, researchers often use Cox proportional hazards models to estimate the risk of an event. However, a fundamental problem arises when studying suicide in a cancer population. A patient might die from their cancer before they ever have the chance to die by suicide.

If a researcher treats these cancer deaths as simple "censoring" (the statistical practice of treating a subject as having dropped out of a study), they introduce bias. They essentially ignore the fact that the cancer death "competed" with and prevented the suicide from occurring. Previous studies on marital status and breast cancer have mostly focused on all-cause mortality. They have left the specific intersection of social isolation and suicide mortality under-explored. Furthermore, earlier studies often failed to account for "effect modifiers." These are variables that change the strength of the relationship between marriage and survival.

Balancing the cohort through sIPTW

The authors approached this problem using a massive dataset from the SEER (Surveillance, Epidemiology, and End Results) Program. This dataset encompasses 825,047 patients diagnosed with primary breast cancer between 2000 and 2022 .

Married and unmarried patients often differ in their baseline characteristics. They may differ in age, race, and tumor stage. A direct comparison would therefore be scientifically unsound.

To resolve this, the researchers employed a three-step methodology:

1. sIPTW (Stabilized Inverse Probability of Treatment Weighting): This technique assigns weights to individuals. It creates a "pseudo-population" where the married and unmarried groups are perfectly balanced. For example, if unmarried patients tend to be younger, the model adjusts the weights. It gives more weight to older unmarried patients and younger married patients. This effectively neutralizes age as a confounding variable. The authors report that after weighting, all baseline covariates reached a Standardized Mean Difference (SMD) of $\leq$ 0.01 [Table 1]. An SMD below 0.10 indicates a negligible imbalance between groups.
2. Fine–Gray Competing-Risk Modeling: Instead of ignoring cancer deaths, the authors used the Fine–Gray model. This model treats non-suicide deaths as "competing events." This allows for a more accurate estimation of the subdistribution hazard ratio (sHR). The sHR represents the relative risk of suicide specifically, accounting for other deaths in the population.
3. Landmark Analysis: The authors wanted to ensure the findings were not just a fluke of the immediate post-diagnosis period. They performed "landmark" analyses. They looked only at patients who survived at least 1, 3, and 5 years. They reset the clock at each interval to see if the risk persisted .

Evidence of a persistent social buffer

The results of the Fine–Gray model confirm a stark disparity in risk. Unmarried/non-partnered patients face a significantly higher risk of suicide mortality (sHR = 1.34, 95% CI: 1.12–1.60) .

This means the risk is 34% higher than for partnered patients.

The study decomposes this risk across several dimensions: * Demographics: Male sex was associated with a much higher risk (sHR = 2.46). Older age ($\geq$65) and non-White race appeared to be protective factors. * Clinical Profile: Patients with estrogen receptor (ER)-negative tumors showed an increased risk (sHR = 1.49). * Temporal Persistence: The landmark analyses demonstrated that this is not a transient spike. The elevated risk for unmarried patients remained significant at the 1-year (sHR = 1.39), 3-year (sHR = 1.46), and 5-year (sHR = 1.34) marks .

One striking finding involves the "Social Buffering Framework." This theory suggests that social support mitigates the stress of a diagnosis. The authors found that macro-level wealth acts as a substitute for this personal buffer. In the lowest income quartile (Q1), the risk for unmarried patients was high (sHR = 1.87). However, in the highest income quartile (Q4), this excess risk was almost entirely neutralized (sHR = 0.71, P = 0.63) .

This suggests that in affluent areas, structural advantages can compensate for the absence of a spouse. These advantages may include better access to professional counseling or financial liquidity.

Limitations in the longitudinal view

While the study is methodologically rigorous, it has constraints. First, the SEER database captures marital status only at the time of initial diagnosis. This is a static snapshot. It cannot account for changes in relationship quality or widowhood during survivorship. A patient who becomes widowed three years later might enter a different risk category.

Second, as a cohort study, the researchers can identify associations but cannot prove causality. We cannot say that being unmarried causes higher suicide risk. We can only say the two are strongly correlated. Finally, the actual number of suicide deaths (529) is relatively small. This can limit the stability of estimates for very specific, rare subgroups of tumor types.

The verdict: A mandate for psychosocial screening

The evidence suggests that social connection is a vital component of cancer survivorship. The finding that the "marital buffer" disappears in wealthy environments is telling. It suggests that social isolation is both a personal burden and a systemic vulnerability.

For clinicians, the takeaway is actionable. Suicide prevention cannot be a "one-size-fits-all" protocol. Screening efforts should be prioritized for younger patients, males, and those in lower-income regions. Integrating routine psychosocial assessments into survivorship care is essential. This is a necessary component of mortality prevention.