In this study, we investigated the HR-QoL of seriously injured patients and determined the associations between various patient- and injury-specific factors and HR-QOL. As previously reported in the literature [3, 23], the impact of a serious injury on the patient's HR-QoL is significant and HR-QoL after injury remains lower compared to the population reference values. In the present study, however, the severity of the injury (measured by ISS and NISS) had no significant association with the decrease in HR-QoL. Among the different types of injuries, spine and spinal cord injuries in particular had a negative impact on HR-QoL, whereas the impact of head injuries on HR-QoL appeared to be less. A longer stay in intensive care seems to be associated with lower HR-QoL. In addition, a low level of education also seems to be associated with lower HR-QoL.

In the literature, the relationship between injury severity scores and HR-QoL varies [3, 18]. Contrary to what one might expect, higher injury severity scores have not always led to lower quality of life [3]. It is likely, therefore, that the scoring systems (i.e., ISS, NISS), which were originally created to predict the mortality of trauma patients, are not suitable for measuring HR-QoL. For example, although a life-threatening hemorrhage in the abdomen and a fracture of the spine can produce the same score, recovery from these injuries may be completely different. Indeed, it is likely that the person injured in the abdominal area may have very few symptoms or be even asymptomatic after a year has passed, whereas the person with a spinal fracture may still need significant help with their daily activities. For this reason, the clinician should not use injury severity scores alone to evaluate a patient's future HR-QoL. Instead, the expected consequences of the injury should first be considered. The differences in how the scores are calculated may also produce heterogeneity in the existing literature. For example, although both ISS and NISS scores are derived from the AIS scoring system, the ISS adds only the most severe body region, while the NISS can add up to three injuries per body region. Therefore, an injury to the head with a skull fracture, subdural hemorrhage, and contusion will provide significantly different severity scores, depending on whether it was evaluated by the ISS or NISS. However, in the current study, neither the ISS nor the NISS appeared to be good predictors of HR-QoL.

Regarding the types of injuries, our study is in the line with earlier reports that injuries to the spine and especially to the spinal cord have the worst negative effect on HR-QoL [8]. Spinal cord injuries are often associated with many kinds of permanent disabilities, and the consequences for the patient are often life changing. Chronic pain is common and up to 77% of patients with spinal cord injury report chronic pain [24]. The level of spinal cord injury also matters, and patients with tetraplegia frequently report more problems with physical functioning and bodily pain than patients with paraplegia [25]. Patients who sustained head injuries reported lower pre-injury HR-QoL but eventually made a better recovery than patients without head injuries. Compared to other trauma patients, patients with mild head injury have been found to report more problems with mental health prior to injury. Although this may explain the worse baseline situation, it does not prevent the patients' mental health from improving later [26]. The number of physical symptoms caused by a head injury has also been found to decrease over time, [27, 28], although the post-injury HR-QoL of patients with head injury generally remains reduced compared to population norms [11].

A low level of education was associated with a larger decrease in HR-QoL. This finding is in line with a study by Haider who noted that although information on level of education is not routinely collected in trauma registries, a low level of education is the most predictive variable of worse long-term outcome after severe trauma [29]. Although no differences were observed in pre- or post-injury HR-QoL, we noticed that the higher the patient's education level, the smaller the subsequent decrease in HR-QoL. It may be, therefore, that a high level of education acts as a protective factor. In addition to physical qualities, recovering from a serious injury requires that patients have good mental skills, i.e., resilience, coping skills, and self-efficacy [24, 25, 29], all of which are likely to be greater in patients who are more highly educated. Returning to working life is also a recognized factor for quality of life [5, 28]. It maybe that returning to working life after injury is more successful for a better educated person than, for example, for a less educated person whose work profile is likely to be more physically demanding.

Low socioeconomic status has previously been found to be linked to a reduction in HR-QoL [18]. The perceived HR-QoL was the lowest among the unemployed and older populations, which may well be due to a previous physical or mental illness, such as depression or anxiety symptoms. In our data, however, we found no differences in HR-QoL between different socioeconomic groups during recovery from injury. This finding may be influenced by the current practice in Finland that all severe injuries are treated under universal health care, where the patient's wealth or income level does not affect where the patient is treated and what the treatment includes. It should be noted that the definition of socioeconomic status or group varies between different studies and can be based on, for example, current annual income or residential area [30]. In our study, data on occupation were used for definition purposes. Although a different job title does not necessarily directly mean a higher socioeconomic status, it is reasonable to presume that a working person with perhaps a higher education has better access to economic resources and social position in relation to others than an unemployed person or pensioner. The effects of having a higher socioeconomic status are likely to be similar to those of having a higher level of education, which are linked in many ways in Finland [31].

Younger patients perceived their pre-injury HR-QoL to be better than older patients, which is a common finding found in many population-level studies. An exception was made by patients older than 65 years whose pre-injury HR-QoL even exceeded the population-level reference values. It is possible that this may have been caused by recall bias. However, it may be that the patients in our study were, in general, more active than older people in previous studies, or they had perhaps adapted better to previous minor ailments. Like in some previous studies, the change in HR-QoL between different age groups was not statistically different, suggesting equal recovery regardless of patient age [9, 11, 12].

As in many studies on HR-QoL, various psychological biases (i.e., recall bias) can affect the results. Although the HR-QoL data were prospectively collected as soon as possible in the ICU, it may still be that some patients rate their pre-injury HR-QoL to be better or worse than it was [32, 33]. However, as data collection started early in the ICU/HDU, the risk for these biases is likely reduced. Even though the patients were contacted by phone, the response rate remained moderate. When evaluating the response rate, it should be borne in mind that the study cohort consisted of only seriously injured patients. After a severe injury, the first-year mortality rate is 10% to 15%, and the consequences of the injury can act as a barrier to reaching the patient. It may be, for example, that a patient has been subjected to long-term institutional care, has no family members that could have answered the phone, or has simply lost interest in the study. It should be noted, however, that the response rate in our study does not differ from most other HR-QoL studies, especially when the data consist of seriously injured patients. There were some differences between responders and non-responders: the responders were older (mean age 53.5 vs. 47.0 years, p<0.001), had a higher level of education, were of a higher socioeconomic group, and the rate of spine injuries was higher (33% vs. 24%, p=0.024). We believe, therefore, that the study still provides a reliable picture of these patients. It should be noted that in present study multiple variables were tested for statistical significancy. Therefore, the possibility that significant findings occur by change is increased. However, we argue that all the tests were planned in advance and all variables had sound reason to be tested for significancy due to known associations from earlier studies.

The definitive strength of this study is the study population which consists of all severely injured trauma patients treated at single major trauma center responsible for all major trauma in its catchment area during the study period. The data on HR-QoL were collected prospectively and the data collection on HR-QoL started at the earliest possible time in the ICU/HDU, which is likely to have reduced the risk for recall bias. To complement the analysis, additional information on injury profile and diagnostics test results were added retrospectively from the electronic medical records.