Article Text

Download PDFPDF
Stair-related falls in the USA: traumatic brain injury and the role of alcohol intoxication
  1. Bart Hammig,
  2. Sydney Haldeman
  1. Public Health, University of Arkansas Fayetteville, Fayetteville, Arkansas, USA
  1. Correspondence to Professor Bart Hammig, Community Health Promotion, University of Arkansas Fayetteville, Fayetteville, AR 72701, USA; bhammig{at}uark.edu

Abstract

Introduction Falls are a leading cause of head injury in the USA. Stair-related fall injuries are common and often more serious than same level falls. Alcohol is a known contributor to unintentional injuries, and often associated with fall-related injuries, specifically falls occurring on stairs. The objective was to examine the association between alcohol use and head injuries (traumatic brain injuries (TBIs) or craniomaxillofacial (CMF) injuries) among persons aged 15–64 presenting to an emergency department (ED) with an injury resulting from a fall on stairs.

Methods Using the 2019 National Electronic Injury Surveillance System, US ED records were examined. Injuries due to falling on stairs were retrieved, with the role of alcohol in the outcome of TBI and CMF injuries examined. Prevalence ratios (PR), adjusted for covariates (blood alcohol level screening, patient demographics, drug use, disposition) were obtained from average marginal predictions derived from logistic regression models.

Results An estimated 687 902 patient visits related to falls on stairs occurred during the study period. Patients who presented with alcohol intoxication had a higher prevalence of TBI (PR 2.7 95% CI 2.3 to 3.1) and CMF injuries (PR 2.5; 95% CI 2.3 to 2.8). PRs were more pronounced among patients with blood alcohol concentration ≥0.1, as was hospital admission.

Conclusions Stair-related falls represent a common cause of ED visits for falls in the USA. Alcohol intoxication had a detrimental effect on the prevalence of TBI and CMF. A multimodal treatment approach may be beneficial given the complex interrelationship between the injury type and alcohol.

  • epidemiology
  • accidents
  • substance abuse

Data availability statement

Data are available in a public, open access repository. Numerical data are publicly for download at https://www.cpsc.gov/cgibin/neissquery/home.aspx with no condition of reuse.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Falls represent a common cause of injury, especially among older adults. Falls on stairs have been found to increase the propensity for head injury, since many involve falling from a height.

WHAT THIS STUDY ADDS

  • Alcohol as a contributing factor to stair-related falls is less understood and has implications for younger adults. This study examines this issue in the context of blood alcohol level and type of head injury, namely traumatic brain injury or a more general classification of maxillofacial injury.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • Implications of the findings include point of care issues arising in the emergency department among injured patients presenting with alcohol intoxication, inclusive of referring for treatment of alcohol use disorders, diagnostic difficulties of intoxication making signs and symptoms associated with TBI more difficult to decipher, and insight into more strategic prevention efforts.

Introduction

Falls are a leading cause of injury in the USA,1 and account for a large proportion of traumatic brain injuries (TBIs) and craniomaxillofacial injuries each year,2 3 which are typically the most serious consequence of a fall. Falls from heights tend to result in more severe injuries, namely head injuries,4 and older persons have been shown to be at high risk of trauma resulting from falls.5 Since falls are ubiquitous and result from a wide variety of circumstances, it becomes important to investigate fall-related injuries in a more strategic manner. Falls on staircases contribute significantly to the burden of fall-related injuries, with an annual estimate in excess of 1 million emergency department (ED visits annually in the USA.6 While much of the focus on staircase fall injuries has been on the older adult population,7–9 a few studies have examined younger populations and the role of alcohol use.10–15 Alcohol has been found to be associated with TBI among persons who were injured falling on stairs.10 12 14 15 However, these studies used trauma centre or hospitalised samples, which make generalisation difficult. Moreover, few examined blood alcohol concentration or blood alcohol levels (BAC/BAL) among injured patients.11 13 15 To better understand the role of alcohol use in stair-related falls, the objective of the current study was to examine the association between alcohol use and head injuries, either TBI or craniomaxillofacial injuries (CMF), among persons aged 15–64 presenting to an ED with an injury resulting from a fall on stairs.

Materials and methods

Using data from 2019, we analysed the National Electronic Injury Surveillance System (NEISS) for falls on stairs. The NEISS is a publicly available dataset that represents injury visits to EDs and uses a nationally representative sample of US hospitals.16 Consisting of a stratified probability sample of 5000 US hospitals that contain a 24-hour ED and have greater than 6 beds, medical charts are reviewed by trained intake staff at all hospitals at urban, suburban, rural and children’s hospitals. Sample weights were applied to each case patient based on the inverse probability of selection, and used to calculate national estimates of patient visits. If the patient visited multiple times for the same reason, only the first visit was recorded.

Measures

Cases were defined as anyone presenting to an ED for an injury related to a fall on stairs (product code 1842). The NEISS does not use International Classification of Diseases, 10th revision, Clinical Modification (ICD-10-CM) coding. Rather, the Consumer Product Safety Commission uses its own coding system. Evaluation of comparability of the NEISS and ICD coding scheme has shown favourable comparability between the two systems, particularly when case narratives are incorporated.17 To evaluate head injuries comprehensively and to estimate the burden on the healthcare system, two types of injuries were evaluated, TBI and craniomaxillofacial (CMF) injury. A diagnosis of TBI was defined as a diagnosis of concussion; the body part injured was head and the diagnosis was internal organ injury; or the body part injured was head and the diagnosis was coded as fracture. This definition was in accordance with prior research applying the ICD-9-CM definition to the NEISS data set, which found a sensitivity of 91.0%.17 18 A CMF was defined based on the body part injured, irrespective of diagnosis, hence a much broader definition than TBI. A case was defined as CMF if the main body part injured included any of the following areas: face, eye, head, mouth, ear or neck.

To examine the role of alcohol in TBI or CMF injuries, we examined all patients who presented with a stair fall injury and identified those patients for whom alcohol use was suspected or confirmed. To identify alcohol-related injuries, personnel were directed to examine the ED intake reports and indicate whether the patient ‘consumed alcohol prior to or during the incident,’ and to indicate confirmed or suspected alcohol consumption. If alcohol consumption was reported, the intake staff were further instructed to indicate ‘yes, and include the patient’s blood alcohol level/concentration (BAL or BAC) in the comment. If there is no BAL/BAC available, this should be stated in the comment. If the ED record indicates that alcohol consumption is suspected, but there is no BAL/BAC available, then select 1-Yes and include this information in the comment.’ Next, among those patients who were coded ‘yes’ for alcohol, we examined narrative text for BAC/BAL levels. Narrative text were examined using STATA’s screening command among the subset of patients who had been coded as using alcohol. A listing of possible terms (eg, BAL NOT, NO BAL, BAL NS, BAL) were used to identify whether or not the patient had a BAL/BAC screen completed. In an effort to gain a deeper understanding of the role of alcohol in the injury event, we examined the alcohol variable in two ways. First, among those patients coded as confirmed or suspected alcohol use, narrative text entries were examined and dichotomised into ‘BAC yes’ as identified by the specified BAC/BAL level, or ‘BAC no’ as identified in the narrative. ‘BAC yes’ indicated that the patient did consume alcohol per the BAC/BAL screen, whereas ‘BAC no’ was indicative of suspected alcohol use, since the patient was coded as having consumed alcohol, but no BAC/BAL screen was conducted. These patients were then compared with the non-alcohol group. In the second analysis, we selected only those patients who had a positive BAC/BAL screen and dichotomised them by BAC/BAL level, either <0.1 or ≥0.1. These patients were then compared with the non-alcohol group.

Next, we examined the narratives to ensure that the reason for the ED visit was due to an injury. Less than 1% of the cases identified were considered non-injury related (eg, chest pain) and were omitted.17 After reviewing the narrative text entries, it became apparent that very few alcohol-related falls on stairs occurred under the age of 15. Given the potential exacerbation of mobility limitations, balance issues and prescription drug interactions, we did not examine injuries among persons aged 65 and older as it would be difficult to fully understand the contribution of alcohol.7 Inter-rater reliability for coding open-ended narrative text ensued as follows. Two coders were trained to identify whether or not BAC/BAL measures were used. First, this prescreening served to improve the inter-rater reliability coefficients, which were assessed using Gwet’s AC kappa statistic.19 Reliability coefficients for coding of BAC/BAL yes or no (Gwet’s AC1 kappa: 0.91, 95% CI 0.85 to 0.94) were deemed acceptable. Next, the subset of patients who had a BAC/BAL completed were then dichotomised into BAC <0.1 and BAC ≥0.1 (Gwet’s AC1 kappa: 0.85, 95% CI 0.81 to 0.92) . Discrepancies between coders were examined and resolved by the principal investigator.

The TBI and CMF prevalences by alcohol-related factors were estimated by fitting logistic regression models. Prevalence ratios (PR), adjusted for covariates of age, sex, race, other psychoactive drug use, (coded as ‘yes’ if intake personnel specified a drug was involved and if the narrative text mentioned a specific psychoactive substance as being present, (eg, heroin, cocaine, marijuana)); disposition, BAC screening (yes/no) and BAC level, were then obtained from STATA’s margins command.20 21

In an effort to more fully understand the scope of the association, PRs were calculated to examine if TBI was associated with (1) any alcohol use, dichotomised using not alcohol related as the reference; (2) whether or not the patient was screened for alcohol use, resulting in three categories: alcohol associated stair fall screened, alcohol associated stair fall not screened, and stair fall not associated with alcohol, using this latter category as the reference and (3) a model examining BAC/BAL levels stratified into <0.1 BAC level or ≥0.1 BAC level. This also resulted in three categories: alcohol-related stair fall with BAC/BAL ≥0.1, alcohol-related stair fall with BAC/BAL <0.1 and stair fall not associated with alcohol, using this latter category as the reference. Those patients who were coded as having an alcohol-related fall on stairs, but not screened, were omitted from this latter model. The same models were constructed to examine the association between CMF injury and alcohol. Data were analysed using STATA MP 13, with sample weights and design variables employed to produce national estimates.

Results

During the study period, a total of 687 902 patients presented to the ED with an injury related to falling on stairs. Among these patients, 6.8% had consumed alcohol as reported by intake personnel. Among those who consumed alcohol, the majority were male (65.9%) and aged 45–64 (52.1%), with 6.3% having also consumed another drug in addition to alcohol. Among those who fell down stairs and had consumed alcohol, 55.8% were diagnosed with a CMF and 25.3% met the more specific criteria for a diagnosis of TBI. With regard to disposition, 25.5% were admitted to the hospital. Conversely, among the patients who had not consumed alcohol (93.2%), the majority were female (65.9%) with 1.1% having consumed a drug other than alcohol. Within the non-alcohol consumption group, 16% were diagnosed with a CMF, 7.6% with a TBI and 7.6% were admitted to the hospital (see table 1).

Table 1

Characteristics of alcohol-related fall injuries on stairs among patients treated in the US emergency departments, 2019 (N=687 902)

Table 2 presents analyses examining PRs between alcohol-related indicators and TBI injuries and CMF injuries resulting from falling on stairs when controlling for other covariates. In examining TBI, those patients who were injured by falling on stairs and had consumed alcohol had a significantly higher prevalence of TBI (28.3%) when compared with those who were injured but did not consume alcohol (10.4%), resulting in a statistically significant PR of 2.7 (95% CI 2.3 to 3.1). With regard to BAC screening, among those patients who had consumed alcohol prior to or during the injury event, those patients who were not screened had almost double the prevalence of TBI diagnosis (20.1 vs 10.3; PR=1.9; 95% CI 1.6 to 2.3), while those who were screened had a TBI prevalence of 34.4 compared with a prevalence of 10.3 among those who presented without alcohol exposure, resulting in a PR of 3.3 (95% CI 2.7 to 4.0). Based on these results, a more than threefold higher prevalence of TBI was associated with BAC screening. Finally, among those patients who were BAC/BAL screened, those with a BAC/BAL <0.1 (prevalence=20.9) did not show a significantly higher prevalence of TBI diagnosis when compared with those who presented without alcohol use (20.9 vs 10.2, PR 2.0.; 95% CI 0.3 to 3.8). However, among those whose BAC/BAL was ≥0.1, a significant PR of 3.5 (95% CI 2.9 to 4.1) was found when compared with those patients injured on stairs who did not consume alcohol. This was indicative of a three and half time higher prevalence of presenting with a TBI among the patient population who had a BAC/BAL ≥0.1.

Table 2

Adjusted prevalence ratios between alcohol-related indicators and traumatic brain injury diagnosis or craniomaxillofacial injuries among persons treated for injuries due to falling on stairs (N=687 902)

In examining CMF injuries, among those patients who were injured by falling on stairs and had consumed alcohol, 68.2% were diagnosed with a CMF injury; compared with 26.9% of those who were injured but did not consume alcohol, resulting in a PR of 2.5 (95% CI 2.3 to 2.8). With regard to BAC screening, among those patients who had consumed alcohol prior to or during the fall event, those patients who had consumed alcohol but were not BAC screened had more than double the prevalence of CMF injury diagnosis when compared with those who presented without alcohol use (PR 2.6; 95% CI 2.3 to 3.0); while those who were BAC screened had a PR of 3.3 (95% CI 3.0 to 3.7) when compared with those who patients who had not consumed alcohol. Among those patients who were BAC/BAL screened, those with a BAC/BAL <0.1 had a PR of 2.1 (95% CI 1.2 to 3.0), while those whose BAC/BAL was ≥0.1 had a PR of 3.4 (95% CI 3.1 to 3.8) when compared with those injured patients who did not consume alcohol.

Discussion

Most studies on fall related injuries either focus on older adults due to the higher prevalence of injurious falls associated with this group, or falls on same levels or fall from heights. Recent research has pointed to stair-related falls as a unique subset of risk for fall-related injury.6 10 12 While prevention of stair related fall injuries among older adults may revolve around coordination and muscle strengthening interventions, lighting and marking of stairs,9 prevention of these injuries among younger age groups, especially if they involve alcohol, will be difficult.

Falls on stairs may increase the likelihood of injury when compared with falling on a level surface due to the uneven surface, potentially hard edges and possibly falling down stairs, resulting in a fall from a height. While many studies have been conducted examining alcohol-related falls in general, and alcohol-related falls from heights, we chose to focus on stair related falls given the aforementioned factors, along with the likelihood that they would be more commonly encountered in everyday life. The addition of alcohol may increase the risk of not only falling on stairs, but also having the fall result in a more severe injury, given that alcohol negatively impacts motor coordination and reaction time. This may result in more difficulty navigating stairs, thus increasing the likelihood of a fall, and subsequently decreasing the likelihood of effectively bracing the fall, resulting in head injuries.11 Our findings did indicate that the prevalence of TBI and CMF injuries were higher among those injured patients who consumed alcohol. In concurrence with previous literature, alcohol-related falls in this study were predominant among males.10 While we did control for age and sex in this study, we did not investigate the potential modification of these variables as it relates to injury outcomes. Future studies may warrant examination of these factors to determine if they modify injury risk. When examined by BAC/BAL level, the prevalence of both injuries were higher among the groups that screened ≥0.1 BAC/BAL. With regard to TBI, those who screened <0.1 BAC/BAL did not have a significantly higher PR when compared with those who had not consumed alcohol, though this may have been due to the low cell size. This finding is not surprising, as ED personnel may be less likely to conduct a BAC/BAL screen if their observation of the patient indicates that they are not demonstrating alcohol-related impairment, even if the patient reported consuming some level of alcohol prior to or during the event. Conversely, those patients who do display effects of alcohol impairment were probably more likely to be screened, with the resulting BAC/BAL finding ≥0.1.

A higher proportion of patients whose injuries were associated with alcohol were admitted to the hospital. In this study, the reason for the decision to admit cannot be determined. We did find that alcohol-effected patients who fell on stairs had injuries consistent with more direct and severe impact to the head area as indicated by TBI diagnosis, findings which have been observed in other fall-related injury studies.10 11 22–24 Also, clinicians may admit head injured patients based on precaution of intracranial bleeds. However, in a study examining patients who fell on stairs and were treated in a level 1 trauma unit,15 the authors found a prevalence of 83.1% of patients presented with TBI and that those patients who had consumed alcohol had lower injury severity scores when compared with non-alcohol using patients. While counter to prior studies, the difference in patient populations between the studies makes direct comparisons difficult. Alternatively, it may be that those patients who present with high levels of alcohol are admitted to the hospital for their own safety, regardless of injury severity, so as to prevent further injury events or to screen further for alcohol use disorder. Clearly, further research is needed to understand the complex interaction of alcohol, falls and head trauma.

A strength of this study was the analysis of BAC/BAL levels of the injured patients, which to date has been uncommon in the fall-related injury literature.11 13 15 Also, the NEISS is a large, population-based sample which allows for more specific analyses and is less likely to be influenced by referral patterns seen in studies of single trauma units. This study is also subject to several limitations. Since only ED visits were examined, precluded were injured persons who did not seek treatment, sought treatment elsewhere or died. NEISS case narratives are subject to human error, and may lack specific aspects of the injury circumstance. In some instances, such as race/ethnicity, incomplete documentation on the intake form resulted in missing data for some estimates. In addition, alcohol consumption was identified by observation of the patient by intake personnel. Therefore, it is not known how recently the person had consumed alcohol or what their BAC/BAL may have been at the time of the injury prior to being transported to the ED. This issue was likely further complicated by presenting with a head injury. The likely result of this misclassification would be a bias in the PR towards the null. ED visits do not capture deaths due to the injury manner in question. Therefore, the most severe outcomes are not captured in this data set. Lastly, no data on socioeconomic status was captured, thus, we were unable to account for disparities based on this measure.

Conclusion

This study highlights the complex nature of injuries related to falling on stairs, especially in conjunction with alcohol use. Alcohol use results in a higher prevalence of head injuries among persons falling on stairs, which may result in more sequelae and increased treatment costs. Prevention of injuries related to falling on stairs among the age group represented in the current study differs from that of older adults whose falls are more likely to result from such factors as frailty, muscle weakness, eyesight or medication side effects. Prevention of alcohol-related falls presents a challenge to the medical and public health communities, with trauma units and EDs serving as a gateway for these injuries. Referral for alcohol use disorder may be warranted among these patients. In addition, diagnostic procedures to identify intracranial injury may be challenging with alcohol intoxicated patients. A multidisciplinary approach to these injured patients may be warranted given the complex interrelationship between the injury type and alcohol.

Data availability statement

Data are available in a public, open access repository. Numerical data are publicly for download at https://www.cpsc.gov/cgibin/neissquery/home.aspx with no condition of reuse.

Ethics statements

Patient consent for publication

Ethics approval

Ethical approval for the current study was conducted by the US Consumer Product Safety Commission. NEISS is a public anonymised database and was therefore exempt from approval by the local Institutional Review Board (IRB).

References

Footnotes

  • Contributors BH devised the study, conducted analyses and wrote portions of all sections; SH wrote portions of the introduction and discussion, aided in analyses. BH is the guarantor of the study.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.