Psychoactive substances have major impact on injuries in rural arctic Norway – A prospective observational study

Rural areas have increased injury mortality with a high pre‐hospital death rate. Knowledge concerning the impact of psychoactive substances on injury occurrence is lacking for rural arctic Norway. These substances are also known to increase pre‐, per‐ and postoperative risk. The aim was by prospective observational design to investigate the prevalence and characteristics of psychoactive substance use among injured patients in Finnmark county.


| INTRODUC TI ON
In Norway, every 12th hospital admission is due to injury, and approximately 2500 injured persons die annually. Accidents are the main cause of death for people below the age of 45, with falls, acute poisonings and road traffic accidents representing the majority. 1,2 In a Norwegian study from 2011 a psychoactive substance was detected in blood in 44% of patients admitted due to injury of patients in an urban setting. 3 Between 2003 and 2008, 1/3 of all drivers who died in road traffic accidents had a psychoactive substance in their blood, 1,4 in Finland from 2007-2011 this proportion was 42%. 5 There is an association between preoperative alcohol use and postoperative morbidity and mortality, including infections, length of hospital stay, need for intensive care, 6 and intensive care complications and outcomes. [7][8][9][10] Alcohol can cause delirium tremens, a serious complication 11,12 seen in 24%-33% of alcohol-dependent patients admitted to a somatic hospital for any cause. 13 A recent study showed that about 20% of acutely ill medical patients have a harmful pattern of alcohol use, 14 and only 14%-27% of patients with an alcohol use disorder seek treatment. 15 The American College of Surgeons Committee on Trauma advocates alcohol screening and brief intervention in all trauma centres. 16 Alcohol-related death has been shown to occur significantly more frequently and several years prematurely in patients previously admitted due to a head injury in conjunction with a finding of ethanol in blood. This was true even when the earlier case of head injury was mild, without accompanying signs of traumatic brain injury (TBI) defined as unconsciousness, amnesia, reduced Glasgow Coma Scale (GCS) or neurological/radiological findings. 17 Severe alcohol intoxication is a risk factor for delayed trauma centre admission for TBI, 18 in addition to it complicating the use of level of consciousness for prediction of head injury severity. 19 Pre-existing use of illicit substances, sedatives and opioids also often presents challenges and requires appropriate management in the critical care and surgical setting. [20][21][22] The county of Finnmark, the northernmost part of Norway, has only 75.000 mostly rural inhabitants, covering an area of 48.000 km 2 , an area slightly larger than Denmark. It has one of the country`s highest rates of reported violent crime per capita 23 and long distances to medical services. Two emergency hospitals cover the area, and the nearest regional trauma centre lies 320 km south of the county border.
Rural areas have significantly higher injury and mortality rates, [24][25][26][27][28] and in Finnmark, the death rate after accidents is the highest in the country, 1 almost double that of an urban area. 29 Of those who die following trauma in Finnmark, 86% expire before arriving at hospital, 30 and improved trauma systems have made no difference. 31 In general, even relatively short delays in arrival of pre-hospital care and distance to hospital impact trauma patients` mortality, [32][33][34][35] and prevention seems to be one of few options. Psychoactive substances impact the risk of injury and patients´ subsequent medical treatment, including pre-, per-and postoperative care. 20 Although an injury might have more severe outcomes in rural areas, little is known about the impact of preventable risk factors like psychoactive substance use.
The aim of this study was to investigate the prevalence and characteristics of psychoactive substance use among injured patients in Finnmark county (psychoactive substances being alcohol, prescription-and illicit psychoactive substances). The primary analysis was associations between age, gender, place of injury, general alcohol use, sensation seeking behaviour and psychoactive substance use prior to the injury. Secondary analyses were the association between self-reported use of alcohol and measured levels of ethanol, and characteristics of patients with reduced GCS or head injuries.

| Participants and study design
This was a prospective observational study including all patients ≥18 years of age admitted to the two emergency hospitals in Finnmark county after injuries from January 2015 to August 2016.
The study covered all seasons and all times of every day to account for variations. All types of injuries were included. Non-accidentrelated injuries were excluded (eg achilles tendon rupture during normal exercise). Also excluded were patients who due to pre-existing conditions and permanently reduced mental state could not provide informed consent.
All patients in the study signed a written informed consent, and were under no circumstances included before verbal and written information could be provided to the patient while lucid and able to relevantly consider this consent. Patients unable to consider consent on admission due to incapacitation had a spare blood sample secured in the ED as part of routine blood sampling, and were later contacted and asked for possible delayed consent. Sampling time was registered to allow for evaluation of metabolization of substances. Patients lost to follow-up before consenting, or unable to consent within two months of their admittance were not included in the study and the study blood sample was destroyed. Only after consent was a questionnaire completed by the patient and admitting nurse or study co-ordinator, and the blood sample was sent for analysis in accordance with the study protocol approved by the Norwegian regional ethics committee.

| Sample size
Due to the nature of the study we were unable to perform a formal power calculation, but an earlier Norwegian study indicated that historically predicted numbers of patients would provide good statistical strength (approximately 1000 injury patients over 18 months). 3

| Variables
The admitting ED nurse registered amongst others the reason for admittance, time of injury, time of blood sample and Glasgow Coma Scale (GCS) on admittance (tool for describing mental alertness from coma to fully alert on a scale from 3-15 by the evaluation of verbal response, eye opening and motor response). Efforts were made to ensure that registered times were as exact as possible, and they could be left blank when uncertain. The patient questionnaire contained demographic factors and circumstances surrounding the injury (type of place, whether it was a fall, type of activity if on public road, reason for injury, and specification of whether violence was part of the incident). The term "violence" was defined as being attacked by another person or otherwise being involved in a physical altercation regardless of blame. Patients were also asked to answer whether they had ingested alcohol, sedatives, psychoactive pain medication, illicit substances or hypnotics to aid sleep within the 6 hours before the incident, alcohol screening via AUDIT-C questionnaire (Alcohol Use Disorders Identification Test -Consumption) 39

| Ethics
The study was approved by the Norwegian Regional Ethics Committee (2014/2033/REK south-east A) according to Norwegian regulations/requirements. According to the approval blood samples and questionnaires were solely identified by a study code linked to the patient's name and social security number via the separately secured consent form for the purpose of correction of data. No results were registered in the patient`s medical records, or made available for anything other than research purposes.

| Statistical analyses
IBM® SPSS® Statistics 25 was used for statistical analysis, where bivariate cross tables were used to analyse associations between the presence of psychoactive substances and patient and injury characteristics. Pearson`s Chi 2 statistical analysis was used to assess statistical significance, if necessary Fischer`s exact test, and independent sample t-test was used for the comparison of means. Mean was reported for normally distributed data, median if not. Multivariable logistic regression was used to further define risk factors. The level of significance was set at P < .05. The STROBE guideline was consulted to ensure quality of reporting. 44

| RE SULTS
Six hundred and eighty four consenting injured patients were included in the study, details shown in Figure 1. Eighty-one percent of patients asked consented to participation. The age group > 60 showed the highest rate of declined consent, with lower age groups being similar to each other with approx. 16% declined consent. The F I G U R E 1 Inclusion and exclusion of patients in the study. *These patients were excluded due to reduced ability to consider consent, language barriers, death during admittance, protracted intensive care treatment, failed blood sampling, direct transfer to intensive care or regional trauma centre, and were not approached before leaving hospital, or not being available for follow-up  Amphetamine Note: Psychoactive substances detected in blood samples from 684 patients admitted to hospital after injuries. For alcohol, self-reported use is also shown. Blood was tested for a broad panel of substances including several known new psychoactive substances. a All positive results where administration of a tested psychoactive substance was known to be, or likely to be by healthcare professionals as part of treatment of the injury in question were removed. In some cases, time of injury or time of arrival was not available, therefore the total number of results can be slightly higher than the two categories of "before and after 6 hours" combined. b In addition to blood samples, patients were asked whether they had ingested alcohol during the 6 hours preceding the injury to correct for metabolized ethanol due to long transport time to hospital. c One patient replied "don't know", the blood sample was negative. Bivariate testing and a multivariable logistic regression analysis is presented in Table 3 and 4 and showed a significant association between psychoactive substance use and of having an AUDIT-C score Proportions of positive results by age-and main substance groups are shown in Figure 2. In the oldest age group, alcohol and illicit substances were less often detected (P = .004 and P = .025 respectively). In contrast, psychoactive medicinal drugs (opioids, sedatives and/or hypnotics) were significantly less often found in the youngest age group (P = .003).

Hypnotics
No significant difference was seen between age groups regarding psychoactive substance use as a whole (P = .277). Alcohol was overall the most prevalent substance, identified in 20% of women and 24.9% of men, with no difference between genders (P =.132).

| Self-reported use of substances
In 94.5% of our cases, self-reported use of alcohol was confirmed by blood sample (Table 1). In almost half of cases reporting having used illicit substances, none were found in blood. However, 82.6% who tested positive denied the use of illicit substances.

| D ISCUSS I ON
We found a high rate of psychoactive substance use among injured patients in rural Northern Norway, with no overall difference between genders or age groups. Ethanol was most prevalent, and only patients >64 years of age showed significantly less use of alcohol  range from 26.2% to 62.5%, however, with major caveats regarding methodology. 48 Although results are slightly lower in our study than in the study from Oslo performed in 2007-08, 3 the impact of psychoactive substances in relation to injury is still very concerning, with recent findings of an even higher prevalence of alcohol in the more severely injured (Finland, 54% alcohol-positive). 49 Our findings of a significant association between using a psychoactive substance and falling support similar conclusions in other studies. [50][51][52] In our study, there was little difference in use overall of psychoactive substances between age groups, but the type differed.
Alcohol and illicit substances were significantly less often detected in the oldest age group, indicating that vigilance is warranted for these substances also in middle-age, not only young adults. Psychoactive medicinal drugs were significantly less often identified in the young age group (18-34 years of age), suggesting that an age-adjusted approach could be warranted, but falls in relation to psychoactive substances in general are an important problem in the young as well as the geriatric population. 53,54 A population-based study on severe TBI in Iceland showed the most common mechanism of injury to be falls from low heights with 28% injured under the influence of alcohol and a mean age of 41 years. 55 AUDIT-C has been shown to be sensitive for identifying harmful alcohol use in emergency department populations, 39,40 and our results showed that increased AUDIT-C was associated with testing positive for a psychoactive substance. Findings of a possible harmful use of alcohol in general were quite high, and the AUDIT-C cut-off ≥5, which we used for both males and females was found optimal in TA B L E 4 Logistic regression analysis of patient and incident factors vs odds of detecting a psychoactive substance

F I G U R E 2
Proportions of positive results in each age group by type of substance. Results are shown with and without self-reported use of alcohol ≤6 hours prior to the incident for all-cause total and for ethanol. Number of patients in each age group is noted in the bar-chart itself, significance levels according to Pearson`s Chi 2 analysis are shown to illustrate differences between age groups in each category. *Opioids, sedatives and/or hypnotics a study of admitted trauma patients. 39 Even with a cut-off ≥4 a metaanalysis revealed increased long-term mortality. 56,57 Sub-group analyses of patients admitted with a GCS < 15 or discharged with a diagnosis of head injury or the more serious TBI, show the main substance involved to be alcohol, where this combination is already known to be an indicator of early death. 17 The combination alcohol and TBI accounted for 5.6% of all admittances due to injury in our hospitals. Our results did not support admittance GCS at these levels being an indicator of risk of serious intracranial head injury with only 1 of 26 patients arriving with a reduced level of consciousness ending up being discharged with a serious cerebral diagnosis. 18 These findings should be considered with caution since we did not have any patients with a GCS less than 11, and only six with more serious intracranial pathology than concussion. Patients with serious intracranial pathology precluding consideration of consent could be an issue, but we have no reason to believe that this represents many cases.
We suggest targeted prevention to be recommendable rather than focusing solely on improved treatment of injuries. Screening, with self-reported use of alcohol, AUDIT-C, and blood sampling should be surmountable. 14 A screening program not including blood sampling could be of limited use for substances other than alcohol due to our findings that self-reported use of other substance groups to a lesser degree can be validated by blood sampling. We believe identification and particular care while admitted should precede a plan for follow-up in primary care.

| Limitations
Our study has several limitations, it did not include fatalities, patients transferred to the regional trauma centre or, patients injured to such an extent that they could not consider consent were not registered in their hospital file, neither could they be used for any purpose apart from this research.
Correlation was satisfactory between measured ethanol in blood and self-reported alcohol use during the 6 hours preceding the incident, which supported applying self-reported alcohol use as a pseudomeasure for metabolized ethanol after long transport times to hospital, as previously recommended by a study investigating this challenge. 60 This was less clear for other substances and the cut-off point of 6 hours could at least partly explain the lacking correlation with self-reported use. To further investigate possible associations one could apply cut-off levels accepted to involve impairment, but that was not the aim of this study. Finally, the study design does not allow for evaluation of causality. Note: Table 5 shows how many patients were discharged with a diagnosis of head injury, or more specifically TBI, depending on whether they tested positive for a psychoactive substance on admission or not. a Patients discharged with any ICD-10 diagnosis indicating an external force injury to the head. For this purpose, ICD-10 categories S00-S09 were considered, excluding one patient due to suspected isolated involvement of the eye.

| CON CLUS ION
b Traumatic Brain Injury. c Patients discharged with ICD-10 category S06 or its sub-categories.
Patients in other ICD-10 categories than S06 -intracranial injuries were reviewed (including categories S02, S07 and S09) to ensure that aberrant coding did not exclude relevant cases involving intracranial injury.
consumption, decreased level of consciousness on admittance and head injury were associated with the use of a psychoactive substance, and all age-groups and both genders appear to be at risk. The combination of blood samples and questionnaires was valuable in identifying correlations. Risk-taking behaviour was not found to be associated with the use of psychoactive substances.

ACK N OWLED G M ENTS
Oslo University Hospital, formerly the Norwegian Institute of Public Health initiated the study in partnership with Finnmark Hospital Trust. A special thank you is owed to the many skilled emergency ward nurses involved in gathering data, particularly study nurses and laboratory staff at Hammerfest and Kirkenes hospitals for valuable assistance in blood sampling logistics.

CO N FLI C T S O F I NTE R E S T
No conflicts of interest have been identified.

D I SCL A I M ER
Views regarding future research or possible interventions in this article are those of the authors, and not necessarily official views of the institutions involved.