Sickness absence > 14 days following mild traumatic brain injuries from traffic accidents: a Swedish nationwide register study

Abstract

Background

Mild traumatic brain injuries (mTBI), including concussions, following traffic accidents is common. How often these injuries lead to sickness absence (SA) among working aged individuals is however insufficiently studied. Thus, the aim of this study was to examine frequency of new SA following mTBI sustained in a road traffic environment and its associations with sociodemographic and injury-related factors.

Method

Nationwide population-based register study. Working aged individuals (18–63), living in Sweden, who in 2014 to 2016 acquired an mTBI in a traffic accident were included based on in- and specialised out-patient health care records. Information on SA (> 14 days), disability pension, pre-injury factors (age, sex, education, marital status, type of living area, country of birth, income from work) as well as injury-related factors (type of road user, in- or outpatient health care) were used in analyses of risk factors for a new SA-spell. Odds ratios (ORs), both crude and adjusted, with 95% confidence intervals (CIs) were estimated with logistic regression.

Results

6073 individuals were identified. 12% had a new SA spell after injury. Sociodemographic risk factors were female sex, older age and being born outside Sweden. Car occupants had higher ORs for new SA, compared to pedestrians, bicyclists, and other road users, and was also associated with longer duration spells (> 90 days). Having received in-patient health care was associated with an OR of 3.7 for new SA compared to those only receiving out-patient health care (including emergency department visits). Having received in-patient health care was also associated with longer duration spells.

Conclusion

A traffic related mTBI is most often a benign injury, seldom resulting in a new SA spell of longer duration. When it does, it is more likely to involve car occupants, and those who have required in-patient health care.

Background

Traumatic brain injury (TBI) is an “acute brain injury resulting from mechanical energy to the head from external physical forces” [1]. It remains a significant global health concern with an estimated worldwide incidence of 69 million new annual cases [2]. The vast majority of all TBI: s are considered mild (mTBI), amounting to over 90% of all cases [3]. Classification of severity is based on the patients status upon presenting to health care, most commonly by use of the Glasgow Coma Scale (GCS) [4]. Not all mTBI: s are associated with visible signs of intracranial injury on day of injury computerised scans or magnetic resonance imaging, thus a common distinction is often made between complicated and uncomplicated mild TBI [5, 6, 7], the latter synonymous with concussion. The sheer volume of mTBI is frequently referred to as a “silent epidemic” [8, 9, 10], because sequelae is predominantly cognitive and emotional in nature and not always immediately apparent for the surrounding people. Common symptoms after mTBI includes headache, attention problems, irritability and anxiety which usually resolves within days or weeks after injury [11]. However, a substantial minority will continue to report unresolved symptoms for months and even years [12, 13, 14, 15].

One major cause for mTBI is road traffic accidents [16, 17, 18]. Road traffic accidents causing physical injury (of any kind) may lead to a wide range of psychosocial consequences and cause significant distress [19, 20]. In particular, returning to work (RTW) if the injury has caused a temporary sick leave may be the most pressing activity to resume. Very few studies have examined RTW rates for individuals who sustained an mTBI in a road traffic accident specifically. Friedland and Dawson [21] found that by six months after injury only 42% of their mTBI subjects (n = 64) had returned to work. A recent retrospective TBI-study by Thor et al. [22] found that 64% of mTBI patients had returned to work by the time the case files were examined, on average two years after injury. There are two systematic reviews on RTW after mTBI in general that shows better rates. In 2014, the International Collaboration on Mild Traumatic Brain Injury Prognosis (iCOMP) included four studies, deemed scientifically admissible, that reported RTW or employment outcome after mTBI. This review concluded that most workers return to work within three to six months after injury but that a minority of 5 to 20% face persisting problems 1–2 years after injury [23]. In 2018 Bloom et al. included 14 studies in their review and reported an RTW-rate of 50% by one month and 80% by six months after injury [24].

Studies addressing RTW on mTBI patient are faced with numerous challenges, and sometimes fall short. In both the above mentioned systematic reviews [23, 24] several issues are highlighted, such as often unclear definitions of RTW, small sample sizes, only single-site studies, varying compensation systems, use of self-report data, selection bias and lack of control of pre-injury status. Some of these issues could be remediated by use of nationwide registers with data from health care and sickness absence (SA) spells following injury. This will allow for studying groups of mTBI patients of adequate size, selected through registers rather than self-selected samples. Using unbiased registry data will also allow for control of pre-injury status and detailed outcome with regards to SA spells with exact extension of reimbursed days. RTW and SA is not the same. In many compensation systems (including the Swedish), SA can be granted also to people who are unemployed or studying. Registry data on SA thus have the potential to capture a more complete view of the consequences of mTBI.

The aim of the present study was to examine SA for working aged individuals with mTBI caused by a road traffic accident. To get the full extent of road traffic accidents we also included accidents to pedestrians in road environments [25]. Furthermore, we wanted to explore risk factors for SA, both generally and for different lengths of SA by different sociodemographic and injury-related factors.

Method

Study design

We conducted a population-based register study of SA among working aged individuals (18–63), living in Sweden during the years 2014–2016 who sustained a new mTBI in a road traffic accident.

The Swedish context

Swedish laws and regulations regarding SA states that all residents, ≥ 16 years old, who have income from work, unemployment or parental leave can apply for SA benefits from the public Swedish Social Insurance Agency if they have a disease or an injury that has caused a reduction of their work capacity. The first day of a SA-spell is a qualifying day and not reimbursed, self-employed can opt for more qualifying days. If the SA-spell lasts longer than seven days, a medical certificate from a physician is required. For most employee’s reimbursement for SA is provided by the employer for the first 14 days. However, for others (e.g., unemployed), the Social Insurance Agency will provide reimbursements from day 2 of the SA-spell. This means that data on shorter SA spells (i.e., < 14 days) is available for these individuals from the agency’s register. In order to not introduce bias, we chose to only include information on SA spells lasting longer than 14 days. SA can be granted for full time or part time (100, 75, 50, 25%) of ordinary working hours. Disability pension (DP) can be granted individuals in ages 19–64 if their injury or disease have resulted in long term or permanent reduction of their work capacity. DP can also be granted full time, or in part-time increments just as SA. This means that an individual on part-time DP can have a simultaneous SA on part time. Calculation of mean and median net days of SA was summed to whole days (e.g., SA for 50% for two gross days would count as 1 net day).

Data sources

Micro-data from five nationwide registers, from three authorities was used. The Longitudinal integration database for health insurance and labour market studies (LISA) from Statistics Sweden was used to identify all individuals living in Sweden in the years 2014 to 2016. LISA was also used to obtain sociodemographic information (sex, age, educational level, country of birth, type of living area, marital status, and having had income from work). The national patient register, from the Swedish National Board of Health and Welfare, holds information on all inpatient health care as well as all specialised out-patient health care, including visits to emergency departments. Main diagnosis, eventual secondary diagnoses and external cause of morbidity is recorded according to the International Statistical Classification of Diseases and related Health Problems; ICD-10 [26]. The external validity of registered concussion diagnosis in the Swedish in-patient register is excellent with 100% positive predictive value, and 99,8% negative predictive value [27], however, no external validation of registered concussion diagnoses in specialized out-patient settings have been performed. The Cause of Death Register from the same authority was used to identify those who had died within the first 30 days after injury. Finally, The Micro Data for Analysis of the Social Insurance (MiDAS) from the Swedish Social Insurance Agency was used for information on dates, extent, and length of SA and DP. All data were linked at the individual level, using the unique personal identity number assigned to all residents in Sweden [28].

Study population

We began by identifying all patients within the patient registers who had been diagnosed with concussion (S06.0) due to an accident while in transportation (external cause of morbidity code V01-99). Road users were then categorized as pedestrians (code V01-09), bicyclists (code V10-19), or car occupants (code V40-49). To ensure the inclusion of pedestrians injured in fall accidents on roads, we expanded our scope to include accidents that occurred on “street and highway” (W00.4, W01.4, W02.4, W03.4, W04.4, W05.4, W10.4, W15.4, W17.4, W18.4, W19.4), and “striking against or bumped into by another person, street and highway” (W51.4). Finally, all other morbidity codes formed a final category of other road users (e.g., motorcycle riders, bus occupants).

We considered the date of the in- or specialized out-patient healthcare visit as the date of injury, as the actual date of injury is not available in the register data. In total, we identified 7,154 working-age individuals who had received health care for concussion caused by a road traffic accident during 2014–2016 when aged 18 to 63. To be included the individual had to live in Sweden 31 December the year before the injury. To include only new injuries, we excluded patients who had received any in- or specialized outpatient health care for any traumatic brain injury (S06) in the 365 days prior to the injury date (n = 187). We also excluded those who had concussion only as a secondary diagnosis (n = 894).

Sociodemographic variables were measured using register data from the year preceding the injury, and were categorized by sex (women, men), age (18–24, 25–34, 45 − 44, 45–54, 55–63 years), educational level (elementary school including missing, high school 10–12 years, university or college > 12 years), country of birth (Sweden, not Sweden), type of living area (cities, towns and suburbs, and rural areas) based on Eurostat’s classification for degree of urbanization [29], marital status (married, not married), having received income from work (yes, no). The reference groups for the regression analysis were selected based on having the lowest expected proportions of new SA (e.g., male sex) or, in some cases, being the largest group (e.g., individuals born in Sweden). While the proportion of new SA was anticipated to be lowest in the youngest age group (18–24 years), we chose the 25–34 age group as the reference because many individuals in the youngest group are likely to be students and may not even be eligible for SA benefits.

Statistical analysis

Basic demographics (sex, age, education level, country of birth, type of living area, marital status), type of injury and inpatient vs. outpatient received health care were shown by SA and DP status at the time of injury, using descriptive statistics. Odds ratios (OR) with 95% confidence intervals (CI) for starting a new SA in connection with the injury were calculated by binary logistic regression. For these analyses, individuals with an already ongoing SA or full-time DP were excluded, because they were not at risk for a new SA spell. Multinomial logistic regression was used to analyse the association between sociodemographic and injury related factors with different durations of SA (< 30 net days, 30–90 net days, > 90 net days). Crude and mutually adjusted ORs were calculated. The statistical analyses were performed using SPSS (version 28) and SAS (version 9.4). The project was approved by the Regional Ethical Review Board in Stockholm, Sweden.

Results

A total of 6073 individuals between the ages of 18–63 received in- or specialized outpatient healthcare for mTBI sustained in a traffic accident during the years 2014 to 2016. The sex distribution was approximately equal, with a slightly higher proportion of women than men. The mean age of the study population was 35.6 (SD 13.4), and the distribution skewed towards the younger age groups, as shown in Table 1. Most individuals were born in Sweden, unmarried, had at least a high school education, lived in urban areas, and were employed.

Table 1 Sociodemographic and injury characteristics for individuals. Aged 18–63 with mild traumatic brain injury sustained in a traffic accident. By sickness absence (SA) and disability pension (DP) status at the time of injury

Most individuals (78%) did not have a new SA (> 14 days) after their injury and were not on SA or full-time DP at the time of the injury. A total of 12% initiated a new SA spell that lasted more than 14 days, while 10% were already on SA or full-time DP at the time of the injury. The proportion of mTBI was highest among bicyclists, accounting for 33% of cases. In-patient healthcare was received by nearly one-third of the study population.

A total of 5473 individuals who were at risk of a new SA (i.e., excluding individuals with full-time DP or ongoing SA at the time of injury) were included in the analysis of odds ratios (ORs), as shown in Table 2. After adjusting for all factors, several sociodemographic factors were found to be associated with new SA. Women were found to be at higher risk of receiving a new SA compared to men. The older age groups were also found to have a higher risk for new SA, although only the OR for the 45–54 age group was significantly higher than the reference group. Individuals born abroad had an elevated OR for new SA. The youngest age group (18–24) and individuals without income from work had significantly lower odds.

Table 2 Odds ratios (OR) and confidence intervals (CI) for new sickness absence (SA) following a mild traumatic brain injury sustained in a traffic accident, by sociodemographic and injury characteristics among the 5473 injured individuals without already ongoing SA or full-time disability pension

Regarding accident and injury characteristics, individuals who received in-patient health care were more likely to initiate a new SA compared to those who only received specialized out-patient health care. Similarly, car occupants had higher odds of new SA than pedestrians, whereas no significant differences were observed for bicyclists or other road users.

Using multinomial logistic regression, crude and adjusted ORs were calculated for different durations of SA (Table 3). Women had a higher risk of new SA than men for short (< 30 days) and medium (30–90 days) duration SA spells, while there were no significant differences for long duration (> 90 days) spells. All older age groups also had a higher risk of long duration SA compared to the reference age group. Regarding injury-related factors, those who had received in-patient healthcare had a significantly higher risk of short, medium, and long-term SA than those who had only received outpatient health care. Having been injured as a car occupant was associated with a higher risk of SA for medium and long-term duration spells compared to pedestrians.

Table 3 Multinomial logistic regression showing odds ratios (OR) and confidence intervals (CI) for different duration of new SA for individuals with mild traumatic brain injury sustained in a traffic accident, by sociodemographic and injury characteristics

Discussion

This nationwide register study investigated the incidence of new SA spells lasting longer than 14 days among 6073 working-age adults who received medical care for a mTBI resulting from a traffic accident. Of those, 12% had a new SA spell (> 14 days), while 10% were already on SA or full-time DP at the time of injury. Our analysis found that females were more likely to have a new SA spell, but there were no differences in the risk of long-duration SA (> 90 days) between sexes. Older age (> 34 years of age) was associated with new SA in particular with long duration SA. Increasing age was also associated with a higher frequency of ongoing SA and DP at the time of injury, which aligns with general workforce trends [30]. However, the prevalence observed in our study was lower than expected based on official workforce data, suggesting a reduced risk of RTI in this group. This may reflect differences in traffic exposure, as individuals on SA/DP likely travel less frequently than those actively working. We also found a higher risk for new SA among individuals born outside of Sweden.

Regarding injury-related factors, our data was limited to healthcare setting (in- vs. outpatient) and type of road user (pedestrian, bicyclist, car occupant, other). However, we found that both of these factors were significantly associated with the risk of new SA. Receiving in-patient healthcare after injury more than doubled the risk of a new SA and increased the risk for long-duration SA compared to those who were treated in outpatient settings only. Additionally, being injured as a car occupant was associated with a higher risk of new SA and long-duration SA compared to injuries sustained by pedestrians.

Our study indicates that the vast majority of individuals recover relatively quickly, and by three months only a small minority of 3% is still on SA. Compared to studies on smaller clinical samples of road traffic injured mTBI patients [21, 22] this number is significantly lower. The higher number in previous studies probably reflects selection biases inherent in clinical studies. For instance, in the Friedland & Dawson study [21], the sample consisted of patients referred to concussion clinics, which obviously limits generalisation to individuals who quickly recover and declines further contact with health care.

A few register-based studies with a narrower target group have reported findings that are more consistent with our results. Ohlin et al. (2018) [31] studied bicycle crashes during a three-year period. A total of 22 045 injured individuals were included, and of those 1600 (7.3%) sustained a concussion. A new SA spell was recorded in 8.4% of these cases, only 1% had an SA spell exceeding 90 days. Kjeldgård et al. (2019) [32] studied bicycle crashes during a single year (N = 7643). Similarly, they reported 8.2% new SA for those who had sustained a concussion. Highest incidence of new SA (35.9%) was found for traumatic brain injuries other than concussions.

In the present study, only a small minority of 3% of adults with mTBI had a long duration SA (> 90 net days). This can be contrasted with the fact that a much larger percentage of adults with mTBI report persisting (> 3 months) symptoms. A recent systematic review and meta-analysis by Cancelliere et al. (2022) [33] found a 31% prevalence of self-reported post-concussion symptoms after three to six months after injury in adult mTBI patients. Taken together, these findings suggest that the majority of mTBI patients with persisting symptoms return to work without having fully recovered. This situation may adversely affect work capacity and contribute to stress reactions.

Indeed, Silverberg et al. (2018) [34] investigated a clinical sample of mTBI-patients and observed that a significant proportion who had resumed work by three months continued to endorse post-concussion symptoms, depression, anxiety and bodily pain and admitted to getting less work done and doing more mistakes at work. Future studies should examine if there are any long-term negative consequences of premature return to work following mTBI, considering both health-related and economic implications.

Due to limitations in the patient registers used in our study, we were unable to include mTBI patients with intracranial signs of brain injury (such as contusions or haemorrhage) due to the absence of TBI severity defining data. Therefore, our findings do not generalize to all mTBI patients. It is worth noting though that concussions constitute the largest subgroup within the broader mTBI spectrum. Among patients presenting to emergency departments with a perfect GCS score of 15, only around 5% exhibit intracranial signs of brain injury on computerized tomography (CT). This percentage increases to 20% for patients with a GCS score of 14 and further rises to 30% for those with a GCS score of 13, which represents the lower limit for mTBI [35].

Intuitively, one would assume that mTBI patients with visible signs of injury have worse outcome, but studies produce mixed results with regard to cognitive functioning [6] and self-reported symptoms [36, 37, 38]. A slower RTW has been found in some studies [39, 40], but not all [41]. A slower RTW for these patients could also reflect different sick-listing practises among doctors [42].

The main strength of this study is that it is not based on a sample of patients but instead all individuals in Sweden who matched the inclusion criteria during 2014 to 2016, virtually with no drop-outs. We used no self-reported data with its risk of possible biases, instead diagnoses were certified by physicians, and all register data came from nationwide administrative registers of acknowledged high quality [43]. However, it is important to note that the codes for external cause, which indicate the specific circumstances of the injury, may contain some level of errors. The inclusion of outpatient data may introduce bias, as the one-year washout period may not fully exclude individuals with older trauma. Future studies should aim to restrict outpatient data to ER visits only to minimize this potential bias. There are some limitations. First of all, we only utilised data on SA spells longer than 14 days. The full burden of concussion on SA is thus not captured, essentially leaving out all self-reported shorter spells that were not accompanied by a mandatory doctor’s certificate. Second, register data are limited to in- and specialised outpatient health care and do not capture visits to general practitioners, which narrowed our possibility to include all patients with concussion. These limitations can be viewed as potential strengths of the study, restricting the inclusion to the more serious cases. Finally, the accuracy and completeness of the register data are potential sources of uncertainty, and further research is needed to evaluate the validity of outpatient data for concussion diagnoses.

Conclusions

To conclude, using register data we found that for individuals who sustained an mTBI in a road traffic environment, SA (> 14 days) followed in 12% of the cases. Longer duration SA spells (> 90 days) were rare and found in only 3% of the cases at risk (i.e., not including individuals already at SA or disability pension at the time of injury). Car occupants and those who required hospitalisation had a higher likelihood of experiencing prolonged SA.