Our study revealed that among suicidal jumpers who survive the jump, the in-hospital days were higher for several years after the index injury compared with levels preceding the event. The in-hospital days were mainly due to psychiatric reasons. This is a new finding, as no other previously published study has presented long-term follow-up data in this patient group on use of healthcare resources.

A somewhat surprising finding in this study was that the in-hospital days spent began to increase already 1 year before the suicidal jump in patients with a pre-existing psychiatric diagnosis. Plausible explanations may include that the psychiatric or somatic morbidity that eventually leads to the suicide attempt begins to manifest a few years before the jump. This is seen in the results as increased in-hospital days in the year preceding the jump. A previous study revealed that the risk of suicide is higher among cancer patients than in patients with another somatic morbidity [19]. Previous studies have also established that the patient’s risk of suicide is several-fold higher during the first few months following psychiatric inpatient care [20] and this pattern may be observed in our results as an increase in in-hospital treatment in the years preceding the jump. This emphasizes the importance of suicide prevention during and following psychiatric inpatient care. However, the significance of this finding in our study is uncertain due to the relatively short pre-trauma follow-up period of only 2 years. More variation in the parameters could occur with a longer pre-injury follow-up period. The use of healthcare resources in the month preceding death by any suicide method in the elderly population was recently reported [21]. Only 54% of the elderly patients had a healthcare contact in the month preceding death and only 28% had a contact due to psychiatric problems.

The high NISS score indicates that these patients were severely injured. Median age was lower than in the entire blunt trauma population in HTR (29 vs. 53 years) [18]. These findings are consistent with previous studies published on suicidal jumpers. We therefore conclude that suicidal jumpers that survive to hospital are younger than average trauma patients and the in-hospital treatment is demanding for these patients and for the healthcare system [5, 7, 8, 11, 12, 14, 22,23,24,25,26].

Four previous studies on suicidal jumpers only reported in-hospital mortality that varied from 7.5 to 21.4% [7, 8, 11, 12]. These studies were heterogenous, as the total number of patients included ranged from 40 to 1070. Three of the study populations were smaller (40–64 patients) and did not report ISS inclusion criteria [7, 11, 12]. Topp et al. had the largest study population of 1040 suicidal jumpers and included patients with ISS > 9 (mean 31.8) [8]. They reported an in-hospital mortality of 21.4% [8]. In our study, the 30-day mortality was 14.2%. Variability in in-hospital mortality could be explained by the differences between patient demographics, since our study seemed to have significantly younger patients (median age 29 years) when compared with the patients in Topp et al. (mean age 39.6 years) [8]. Trauma patients with pre-existing psychiatric disorder or suicide attempt have higher risk for in-hospital mortality and longer hospital stay [27]. In the current study, the mortality was similar in both groups but the patients with pre-existing psychiatric disorder had higher number of in-hospital days in the first year after the index injury.

Only one previous study on suicidal jumpers only reported longer-term mortality, which was 17.5% at 1 year [12]. The 1-year mortality was 15.7% in this study. The mortality between group 1 and 2 increased in the long-term, with a 14% higher 5-year mortality in patients with a pre-existing psychiatric diagnosis. Our findings on long-term mortality are new, as no other previous study on suicidal jumpers only has presented mortality data up to 5 years. A previous study on patients with major blunt trauma revealed a 5-year mortality of approximately 30% [18], which is somewhat higher than that observed in the present study. However, the patients were significantly older in the previous study; median age was 53 years compared with the median age of 29 years in the current study. The previous study also included isolated head injuries, which probably explains at least some of the difference in mortality rates. Furthermore, only 28% of the population in this study had suffered a severe traumatic brain injury.

In previous publications, a large proportion of jumpers had contact with psychiatric services or had a psychiatric diagnosis in their records before the suicidal jump [5, 6, 23]. This study is consistent with these findings. Almost half of the patients in this study did not have any contact with healthcare professionals before the suicidal jump due to psychiatric reasons. One possible explanation is that for some of these patients, the suicide attempt is the first manifestation of psychotic illness and psychosis [6]. Mood disorders and psychotic illnesses appear to be the most common psychiatric morbidities among the suicidal jumpers [22, 24, 28]. It is possible that in many of these cases, there might have been an opportunity to prevent major catastrophic events, such as deliberate jump from a height, as the healthcare system had a prior contact with these patients. Limited resources and the problem of identifying risk factors for suicidal acts may be the most substantial challenges in preventing these acts. A recent study showed that young adults who had been treated in-hospital for a fracture had a 6-fold higher risk of death in a 10-year follow-up compared with the general population [29]. Suicide (28%) was the leading cause of death for these patients. Fractures requiring in-hospital treatment in young adults may indicate suicidal behavior in later life. One study [30] reported long-term (6 ± 3 years) outcomes on the patients (n = 35) with pre-existing psychiatric disorder after suicide attempt by jumping from height or leaping in front of train or car. Good or satisfactory quality of life was observed in half of the patients at the end of the follow-up.

It is currently established that if the psychiatric disorder escalates to the point of suicide attempt, the effects are catastrophic for the patients and relatives and are expensive for healthcare systems. Therefore, it would be reasonable to invest more resources towards prevention, such as identifying these patients. However, since almost half of the patients did not have any pre-existing psychiatric diagnosis before the suicide attempt, other measures are also necessary. Some concrete evidence exists on the measures that can be taken to prevent specifically deliberate jumps from a height; bridges are known destinations of suicidal jumpers, probably due to their popularity and well-known locations [27, 31, 32]. Addition of protective fences to bridges that make access to a jumping location more difficult reduced the amount of suicidal jumps from the bridges in question [28].

The strengths of this study include the fairly large number of patients from a 12-year time-period, with complete medical records obtained from three different registries. This was also the first study with a 5-year follow-up of healthcare resource utilization after suicidal jumps from a height.

The nature of registry-based research can be considered both a limitation and a strength of our study. The HTR included only patients with NISS ≥ 16, hence suicidal jumpers with more minor injuries were not included in the analysis. However, the trauma unit of the HUH is the only hospital in the catchment area that treats patients with major blunt trauma, and we were able to capture a fairly accurate picture of this patient group as the HTR includes all patients admitted to our hospital within the registry inclusion criteria. The HTR has been routinely maintained for several years and the data quality of the HTR has been considered as excellent and therefore our data can be considered reliable [16].

However, it is possible that the intent behind the patient’s jump from height is not always accurately recorded in the medical e-records due to human errors in data imput or, for example, the patient’s reduced level of consciousness during treatment. Therefore, some cases may be missing from the study.

Regarding patient grouping, it is possible that a patient may have had a pre-injury psychiatric disorder without any visits to healthcare providers in 2 years and therefore was not included in the group 2.

Our study included only patients who attempted suicide by jumping from a height. Therefore, our results cannot be generalized to patients who attempt suicide with other methods.