Background
Severe somatic illness is associated with the risk of suicidal behavior. This information is important in relation to the treatment of severe somatic illness, as patients may express suicidal behavior at some point in time. There is limited knowledge about the temporal relationship between the time when the life-threatening diagnosis is given and when the suicidal behavior begins.
Aim
To investigate whether a diagnosis of severe somatic disease abruptly increases the risk of suicide and deliberate self-harm in the Danish population, and to assess when the risk is greatest.
Methods
Using the Danish nationwide health registries, we identified all individuals diagnosed with cancer between 2012 and 2021. We included all types of cancer except 'other and unspecified malignant neoplasm of skin'. Each individual was followed from 2 years before their cancer diagnosis until 5 years after. If individuals died or emigrated, they were censored from the study at that point. The time unit was quarters of a year, and for each quarter, incidence rates (IR) and incidence rate ratios (IRR) for deliberate self-harm and suicide were calculated using generalized estimating equation (GEE) models in Stata.
Results
In a population of around 6 million Danish citizens, 383,462 were diagnosed with cancer during the period 2012-2021. The median age was 69 years old, and 50.2% were male. Among demographic characteristics, 58.1% were married, 33% had a low level of education, 66.1% were retired, and the general level of income was low. A total of 9,352 deliberate self-harm events and 290 suicides were observed during the follow-up period. We found a clear pattern with a very high risk for suicide and deliberate self-harm in the first quarters after being diagnosed with cancer. In the first quarter, the incidence rate (IR) for suicide was 60 per 100,000 person-years, which lowered to 35 in the next quarter. The cancer population had a significantly lower IR for deliberate self-harm prior to the date of diagnosis compared to the background population, but the IR rose significantly to 320 per 100,000 person-years in the quarter after diagnosis. Similar to suicide, the IR for self-harm decreased in the next quarter and continued to decrease as time moved on.
Conclusions

The risk of suicidal behavior is very high in the first quarter after a cancer diagnosis. We recommend that clinicians be aware of the significant increase in risk for patients recently diagnosed with cancer and communicate this risk to the patient. It is important for clinicians to know how to handle patients at risk of suicidal behavior and whom to contact to facilitate the necessary help and treatment.