Article, Emergency Medicine

Developing a standardized measurement of alcohol intoxication

a b s t r a c t

Introduction: We assessed multiple examinations and assessment tools to develop a standardized measurement of alcohol intoxication to aid medical decision making in the Emergency Department. Methods: Volunteers underwent an alcohol challenge. Pre- and post-alcohol challenge, subjects were videotaped performing three standardized clinical examinations: (1) Standardized Field Sobriety Test (SFST) examination, (2) Hack’s Impair- ment Index (HII) examination, and (3) Cincinnati Intoxication Examination (CIE). Emergency clinicians evaluat- ed the level of intoxication using five standardized assessment tools in a blinded and randomized fashion: (1) SFST assessment tool (range 0-18), (2) HII assessment tool (range 0-1), (3) St. Elizabeth Alcohol Intoxication Scale (STE, range 0-17), (4) a Visual Analog Scale (VAS, range 0-100), and (5) a Binary Intoxication Question (BIQ). construct validity was assessed along with inter- and intra-rater reliability. Results: Median scores pre- and post-alcohol challenge were: SFST 6 (interquartile range 5) and 11 (3), respectively; HII 0 (0.05), 0.1

(0.1); STE 0 (1), 1 (2); VAS 10 (22), 33 (31). For BIQ, 59% and 91% indicated intoxication, respectively. Inter- rater reliability scores were: SFST 0.71 (95% confidence interval 0.48-0.86) to 0.93 (0.88-0.97) depending on ex- amination component; HII 0.90 (0.82-0.95); STE 0.86 (0.75-0.93); VAS 0.92 (0.88-0.94); BIQ 0.3. Intra-rater re- liability scores were: SFST 0.74 (0.64-0.82) to 0.87 (0.81-0.91); HII 0.85 (0.79-0.90); STE 0.78 (0.68-0.85); VAS

0.82 (0.74-0.87); BIQ 0.71. VAS reliability was best when paired with the HII and SFST examinations. Conclusions: HII examination, paired with either a VAS or HII assessment tool, yielded valid and reliable measurements of al- cohol intoxication.

(C) 2017

  1. Introduction

Alcohol intoxication and alcohol-related injury place a substantial burden on Emergency Departments (ED). In the United States, up to 9.7% of all ED visits have alcohol involved [1,2]. The Centers for Disease Control and Prevention estimate the two-year annual rate of ED visits related to alcohol is 94 per 10,000 for males and 36 per 10,000 for fe- males [3]. Alcohol intoxication as a primary complaint is responsible for up to 1.2% of all ED visits [4,5].

Delivering safe and efficient care to the intoxicated ED patient is challenging because intoxication confounds clinical evaluation. Intoxi- cation impairs patients’ ability to identify and localize pain, hindering the diagnosis of concomitant metabolic or traumatic derangements [6, 7]. Intoxication also decreases the Glasgow Coma Score, making it diffi- cult to determine if Mental status changes are due to intoxication or

Abbreviations: ED, Emergency Department; BAC, Blood or Breath alcohol concentration; SFST, Standardized Field Sobriety Test; CIE, Cincinnati Intoxication Examination; BIQ, Binary Intoxication Question; HII, Hack’s Impairment Index; ICC, intraclass correlation coefficient; VAS, Visual Analog Scale; STE, St. Elizabeth’s Alcohol Intoxication Scale.

* Corresponding author.

E-mail address: [email protected] (J.L. Benoit).

traumatic brain injury [8,9]. Thus, the level of clinical intoxication is crit- ically important for medical decision making. Examples of this include

(1) determining when the patient is sober enough that history and physical examination will be adequate to assess for concomitant pathol- ogy, (2) ensuring the patient is achieving sobriety at an anticipated rate to preclude the need for diagnostic laboratory or imaging studies, (3) assessing when the patient is safe to be discharged home unaccompa- nied, and (4) mitigating the risk of developing alcohol withdrawal syn- drome in the ED for patients uninterested in treatment for alcohol abuse.

Unstructured, subjective clinical evaluation is inadequate in assessing levels of intoxication [10-13]. Blood/Breath Alcohol Concen- tration (BAC) also correlates poorly with level of intoxication, yet is fre- quently measured in the ED [14-16]. Markedly different levels of intoxication may be observed despite similar BAC between individuals. This is particularly true in patients with chronic alcoholism, who may be clinically sober with a BAC that would cause severe intoxication or death in others [17]. Thus, there exists a critical need for a standardized quantitative measurement of intoxication that is both reliable and sen- sitive to change in mental status. This could be used to evaluate and monitor Intoxicated patients, avoid unnecessary testing, and facilitate safe ED management.

http://dx.doi.org/10.1016/j.ajem.2017.01.009

0735-6757/(C) 2017

However, no well-accepted measurement of intoxication currently exists for use in the ED. Two reported tools quantify the degree of intox- ication: Hack’s Impairment Index (HII) and the St. Elizabeth’s Alcohol Intoxication Scale (STE) [18,19]. The HII quantifies the results of a stan- dardized examination while the STE does not include a standardized ex- amination. Neither has been externally validated, and their degree of inter- and intra-rater reliability is unknown. Another tool, the National Highway Traffic Safety Administration’s Standardized Field Sobriety Test (SFST), has been used by law enforcement in the United States for decades. However, it has only been validated against BAC, not clinical intoxication [20-22].

This study was designed as the first step in developing a standard- ized measurement of intoxication due to alcohol for use in the ED. We evaluated multiple candidate standardized clinical examinations paired with multiple candidate assessment tools, determined their responsive- ness to an alcohol challenge as a measure of construct validity, and cal- culated their inter- and intra-rater reliability.

  1. Methods
    1. Study setting and design

This was a prospective diagnostic study conducted in a controlled setting. Institutional Review Board approval was obtained, and all sub- jects gave informed consent to participate. Subjects were compensated

$300.00 for their time.

Study population

Two groups of subjects were enrolled:

  1. Alcohol challenge subjects: Healthy volunteers with a moderate his- tory of alcohol use underwent an alcohol challenge via a standard- ized protocol to obtain a moderate level of intoxication. Before and after intoxication, they underwent standardized clinical examina- tions comprised of a series of physical maneuvers and cognitive tasks. These examinations were videotaped.
  2. Evaluators: Emergency physicians, nurses, research coordinators and paramedics volunteered to view these recorded examinations and measure the level of intoxication using standardized assessment tools. Evaluators were blinded to the elapsed time since the alcohol challenge.

Inclusion criteria for alcohol challenge subjects were age 21 to 65 with a moderate history of alcohol use, assessed by the National Insti- tute on Alcohol Abuse and Alcoholism Three Question Set [23]. Moder- ate use was defined as drinking alcohol at least 2 to 3 times a month, but no greater than 3 to 4 times a week, have no more than 5 to 6 drinks on a typical day, and have had at least 1 episode of binge drinking in the past year. Persons with a personal history of alcoholism, health problems that could be worsened by drinking alcohol (e.g. alcohol flush reaction and allergy to alcohol), the use of medications that would interact with alcohol (e.g. benzodiazepines and acetaldehyde dehydrogenase in- hibitors), pregnancy, or paid employees of the investigator’s depart- ment were not eligible to participate.

The inclusion criteria for evaluators were current employment as an emergency medicine physician, emergency medicine nurse, emergency medicine clinical research coordinator with nursing background, or paramedic. Alcohol challenge subjects were not permitted to participate as evaluators.

Alcohol challenge

The alcohol challenge occurred in groups of three subjects. Subjects were instructed not to consume alcohol or intoxicating drugs during the 24 h prior to the alcohol challenge and to eat a light breakfast the

morning of the challenge. The target BAC for the challenge was

0.16 g/dL, which is twice the legal definition of intoxication in our region (Cincinnati, OH, USA). The volume of alcohol necessary to achieve this BAC was calculated using a standard formula that accounts for gender, age, height, weight, alcohol concentration, and the duration of ingestion [24,25]. For example, to achieve a BAC of 0.16 g/dL, a 25-year-old male weighing 180 lb needed to ingest 11.6 oz (7.7 drinks) of 80 proof alcohol over 60 min [26]. Subjects were required to consume a quarter of the calculated volume every 15 min. Subjects were allowed to choose the 80 proof liquor (e.g. vodka, rum and whiskey) as well as the fruit juice or carbonated beverage mixer.

Examination

Baseline examinations were performed 30 min prior to the alcohol challenge, and subsequent evaluations were done post-alcohol chal- lenge at 30 min (BAC ~ 0.16 g/dL), 2 h (BAC ~ 0.12 g/dL), 4 h (BAC ~ 0.08 g/dL), and 6 h (BAC ~ 0.04 g/dL). BAC was measured at the start of each examination using a breath test (BACtrack model BT- KC10, San Francisco, CA, USA), which correlates highly to blood mea- surements [27,28]. At each time interval, three examinations were performed:

  1. The National Highway Traffic Safety Administration’s Standardized Field Sobriety Test (SFST) examination [21]
  2. Hack’s Impairment Index (HII) examination [18]
  3. The Cincinnati Intoxication Examination (CIE; standardized exami- nation developed to facilitate scoring of the St. Elizabeth Alcohol In- toxication Scale).

All examinations were videotaped with audio and utilized camera zoom to facilitate visualization of specific examination components (e.g. nystagmus and fine motor control). The same person administered the standardized examinations to all alcohol challenge subjects at all time intervals. Subjects were allowed no more than one minute to com- plete each component. Subjects were not given any opportunities to practice. The instructions read aloud to subjects are provided in the Sup- plemental information.

Evaluation

There were five assessment tools used to measure the level of intoxication:

  1. Binary Intoxication Question (BIQ; “Is the subject intoxicated?”)
  2. Visual Analog Scale (VAS; standard 100 mm scale with anchors of

“no intoxication” and “severe intoxication”)

  1. Standardized Field Sobriety Test (SFST) [21]
  2. Hack’s Impairment Index (HII) [18]
  3. St. Elizabeth Alcohol Intoxication Scale (STE) [19].

For the BIQ, evaluators were told to define alcohol intoxication based on the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental disorders [29]. For the VAS anchor, severe intoxica- tion was defined similarly, but modified to require “stupor or coma” as a sign. The assessment tools are provided in the Supplemental information.

The SFST, HII and STE assessment tools were scored in conjunction with the SFST examination, HII examination and CIE, respectively. Both the BIQ and the VAS assessment tools were scored in conjunction with all of the examinations. This yielded nine different examination- assessment combinations (Table 1).

Evaluators completed a two-hour in-person training session where a detailed explanation of each examination and assessment tool was pro- vided. Evaluators then watched 12 selected examination videos and practiced filling out each assessment tool. This was done as a group

Table 1

Examination-assessment tool combinations.

Examination + assessment tool

Standardized Field Sobriety Test Examination + Standardized Field Sobriety Test
  • Standardized Field Sobriety Test Examination + Binary Intoxication Question
  • Standardized Field Sobriety Test Examination + Visual Analog Scale
  • Hack’s Impairment Index Examination + Hack’s Impairment Index
  • Hack’s Impairment Index Examination + Binary Intoxication Question
  • Hack’s Impairment Index Examination + Visual Analog Scale
  • Cincinnati Intoxication Examination + St. Elizabeth Alcohol Intoxication Scale
  • Cincinnati Intoxication Examination + Binary Intoxication Question
  • Cincinnati Intoxication Examination + Visual Analog Scale
  • Nine different combinations of examinations paired with assessment tools were evaluated.

    with a debriefing after each video to encourage questions and discus- sion. The goal was for evaluators to develop a shared understanding of the examinations and assessment tools, and reduce the effect learning would have on scoring. Results from these training sessions were not in- cluded in the analysis.

    After training, each evaluator independently watched and scored 144 videos. Each video consisted of a single subject performing a single examination at a single point in time paired with either a single assess- ment tool or the BIQ and VAS presented together. Evaluators were instructed to answer the BIQ before scoring the VAS. Videos were pre- sented in random non-chronological order, and evaluators were blinded to the time interval between alcohol challenge and examination. A predetermined number of videos were presented to the same evaluator twice in a blinded fashion. Identical sets of videos were watched by evaluators from the four clinical groups. Evaluators were only given a single opportunity to view and score each video. Evaluators were instructed not to compare scores across different videos. If an evaluator did not see a particular maneuver in the video, they were instructed to assume that portion of the examination was unremarkable.

    Statistical analysis

    There is no gold standard measurement of intoxication due to alco- hol to compare against to determine sensitivity and specificity. Similar- ly, differences in tolerance to alcohol and gender differences were expected to produce different levels of intoxication despite similar BAC in alcohol challenge subjects. Therefore, the analysis focused on de- termining which examination-assessment combinations had (1) con- struct validity as demonstrated by an appropriate change in relative intoxication over time based on alcohol pathophysiology, (2) high inter-rater reliability (i.e. different evaluators gave a single video a sim- ilar score), and (3) high intra-rater reliability (i.e. the same evaluator gave a single video a similar score when viewed twice). When the same video was viewed twice by the same evaluator for the purposes of intra-rater reliability, the first viewing was used for calculations of inter-rater reliability.

    Due to underlying design differences in the assessment tools, scores could not be normalized to a single common scale. The score from the HII was calculated as the sum of its five components, divided by the maximum score possible for the components attempted (range 0-1). The STE was calculated as the sum of its five components (range 0- 17). The three components of the SFST were separated, and each com- ponent score was calculated independently as the sum of “yes” answers. Alcohol challenge subjects unable to complete a component of the SFST were given the maximum score for that component. The three compo- nents of the SFST were the walk and turn test (range 0-8), the one leg stand test (range 0-4), and horizontal gaze nystagmus (range 0-6). The VAS was measured in millimeters using a standard ruler (range 0-100). See the Supplemental information for more details.

    Inter-rater reliability for continuous variables was quantified using a two-way, random, absolute average measure Intraclass Correlation

    Coefficient (ICC) [30]. Intra-rater reliability for continuous variables was quantified using a one-way, random ICC. The ICC was interpreted as poor for values b 0.4, fair for values between 0.4 and 0.59, good for values between 0.6 and 0.74, and excellent for values between 0.75 and 1.0 [30]. Inter- and intra-rater reliability for categorical variables were quantified using Light’s kappa for m raters or Cohen’s kappa. Inter- and intra-rater reliability was assessed overall and for each eval- uator subgroup. Statistical analyses used SPSS 22.0 (IBM Corporation, Armonk, NY), R base (Version 3.1.1), and R packages irr, psych and beeswarm [31-34].

    1. Results

    Twelve subjects participated in the alcohol challenge during May and June of 2015. Mean age was 24 (standard deviation 2), all were Cau- casians, one was Hispanic, and nine were female. Range of alcohol vol- ume ingested was 7.34 to 14.10 oz. BAC measured at all time points are shown in Fig. 1. There were no adverse events from the alcohol challenge.

    Evaluators were emergency physicians (n = 4), emergency medi- cine nurses (n = 4), clinical research coordinators (n = 4) and para- medics (n = 4). A total of 2304 assessments were completed by these 16 evaluators in June of 2015. There were 9 assessments that were ex- cluded due to missing data or being incorrectly completed, leaving 2295 assessments for analysis.

    Construct validity

    The distribution of scores at each time point for the VAS, HII, STE, and the three components of the SFST are shown in Fig. 2. All assessment tools identified 0.5 and 2 h post-alcohol challenge as times when sub- jects were most intoxicated, which is consistent with alcohol patho- physiology. The HII and STE had the narrowest range of scores at each time point, while the three components of the SFST had the widest ranges. Overall, the BIQ had the worst performance, and incorrectly cat- egorized 59% of subjects as intoxicated at baseline. The BIQ categorized 91%, 88%, 54%, and 48% of subjects intoxicated at 0.5, 2, 4, and 6 h post- alcohol challenge, respectively.

    Fig. 1. Distributions of Breath Alcohol Concentration at each time point.

    Data are represented by box plots. Boxes represent the interquartile range with the median indicated by the middle line. Whiskers represent data within 1.5 times the interquartile range from the upper and lower quartile. Each dot represents a single subject. Baseline is 30 min before the alcohol challenge, while 0.5, 2, 4, and 6 h are times post-alcohol challenge.

    Image of Fig. 2

    Fig. 2. Distributions of assessment tool scores at each time point.

    The y-axis displays the total range possible for each assessment tool. Data are represented by box plots. Boxes represent the interquartile range with the median indicated by the middle line. Whiskers represent data within 1.5 times the interquartile range from the upper and lower quartile. Baseline is 30 min before the alcohol challenge, while 0.5, 2, 4, and 6 h are times post-alcohol challenge. SFST: Standardized Field Sobriety Test.

    The performance of the VAS assessment tool, stratified by examina- tion, is shown in Fig. 3. The VAS matched alcohol pathophysiology most closely when scored after the HII or SFST examinations. The BIQ per- formed poorly regardless of examination (data not shown).

    Inter-rater reliability

    The overall ICC for inter-rater reliability of the VAS, HII, STE, and SFST are shown on the left in Table 2. All assessment tools demonstrated ex- cellent agreement, except for the horizontal gaze nystagmus compo- nent of the SFST, which had good agreement. The BIQ did not perform

    as well as the continuous measures. Kappa was 0.30, demonstrating poor agreement. The ICC was also calculated stratifying by the clinical background of evaluators (Table 2). For the BIQ, kappa was 0.51, 0.15, 0.15, and 0.50 for physicians, nurses, research coordinators, and para- medics, respectively.

    Intra-rater reliability

    The overall ICC for intra-rater reliability of the VAS, HII, STE, and SFST are shown on the right in Table 2. All assessment tools demonstrated ex- cellent agreement, except for the walk and turn component of the SFST,

    Fig. 3. Visual Analog Scale scores for each time point stratified by examination.

    The y-axis displays the total range possible for the Visual Analog Scale. Median scores are shown. Baseline is 30 min before the alcohol challenge, while 0.5, 2, 4, and 6 h are times post-alcohol challenge. HII: Hack’s Impairment Index examination; CIE: Cincinnati Intoxication Examination; SFST: Standardized Field Sobriety Test examination.

    Table 2

    Inter- and Intra-rater reliability of assessment tools overall and stratified by clinical back- ground of evaluator.

    Inter-rater reliability ICC (95% CI)

    Intra-rater reliability ICC (95% CI)

    Visual Analog Scale

    Overall

    0.92 (0.88 to 0.94)

    0.82 (0.74 to 0.87)

    Physician

    0.83

    (0.73 to 0.89)

    0.80

    (0.70 to 0.87)

    Nurse

    0.59

    (-0.10 to 0.82)

    0.73

    (0.60 to 0.82)

    Research coordinator

    0.71

    (0.50 to 0.82)

    0.93

    (0.89 to 0.96)

    Paramedic

    0.90

    (0.83 to 0.94)

    0.89

    (0.83 to 0.93)

    Hack’s Impairment Index

    Overall

    0.90

    (0.82 to 0.95)

    0.85

    (0.79 to 0.90)

    Physician

    0.81

    (0.55 to 0.92)

    0.88

    (0.75 to 0.95)

    Nurse

    0.52

    (-0.23 to 0.81)

    0.92

    (0.83 to 0.97)

    Research coordinator

    0.58

    (-0.14 to 0.84)

    0.89

    (0.77 to 0.95)

    Paramedic

    0.80

    (0.39 to 0.92)

    0.68

    (0.39 to 0.85)

    St. Elizabeth Intoxication Scale

    Overall

    0.86

    (0.75 to 0.93)

    0.78

    (0.68 to 0.85)

    Physician

    0.68

    (0.28 to 0.86)

    0.87

    (0.72 to 0.94)

    Nurse

    0.46

    (-0.24 to 0.79)

    0.72

    (0.46 to 0.87)

    Research coordinator

    0.79

    (0.52 to 0.91)

    0.89

    (0.75 to 0.95)

    Paramedic

    0.48

    (-0.14 to 0.78)

    0.55

    (0.19 to 0.78)

    SFST Horizontal Gaze Nystagmus Test

    Overall

    0.71

    (0.48 to 0.86)

    0.81

    (0.73 to 0.87)

    Physician

    0.49

    (-0.09 to 0.77)

    0.90

    (0.78 to 0.95)

    Nurse

    0.17

    (-0.17 to 0.51)

    0.70

    (0.42 to 0.86)

    Research coordinator

    0.47

    (-0.22 to 0.77)

    0.77

    (0.54 to 0.89)

    Paramedic

    0.60

    (-0.13 to 0.85)

    0.61

    (0.28 to 0.81)

    SFST One Leg Stand Test

    Overall

    0.93

    (0.88 to 0.97)

    0.87

    (0.81 to 0.91)

    Physician

    0.87

    (0.68 to 0.94)

    0.96

    (0.90 to 0.98)

    Nurse

    0.75

    (0.43 to 0.89)

    0.82

    (0.64 to 0.92)

    Research coordinator

    0.83

    (0.61 to 0.93)

    0.86

    (0.71 to 0.94)

    Paramedic

    0.69

    (0.25 to 0.87)

    0.84

    (0.67 to 0.93)

    SFST Walk and Turn Test

    Overall

    0.81

    (0.67 to 0.91)

    0.74

    (0.64 to 0.82)

    Physician

    0.51

    (-0.08 to 0.79)

    0.90

    (0.78 to 0.95)

    Nurse

    0.41

    (-0.38 to 0.75)

    0.70

    (0.42 to 0.86)

    Research coordinator

    0.62

    (0.12 to 0.83)

    0.77

    (0.54 to 0.89)

    Paramedic

    0.50

    (-0.03 to 0.80)

    0.61

    (0.28 to 0.81)

    ICC: Intraclass Correlation Coefficient; CI: Confidence Interval; SFST: Standardized Field Sobriety Test.

    which had good agreement. The BIQ had good agreement with a kappa of 0.71. As above, the ICC was calculated stratifying by the clinical back- ground of evaluators (Table 2). For the BIQ, kappa was 0.70, 0.27, 0.80, and 0.62 for physicians, nurses, research coordinators, and paramedics, respectively.

    1. Discussion

    This study was the first step in a program of research to evaluate both novel and existing tools that measure the level of intoxication due to alcohol. As such, the goal was not to prove that one examination or assessment tool was superior, but rather to provide comparative data to understand better the strengths and weaknesses. Ultimately, we hope to propose a standardized and validated approach to measuring intoxication, which will facilitate the safe and efficient care of this diffi- cult ED patient population, as discussed in the introduction. Based on our results, the HII examination combined with either the HII or VAS as- sessment tool warrants further investigation. Both examination-assess- ment combinations demonstrated construct validity, as they exhibited sufficient sensitivity to detect changes post-alcohol challenge, and their score curves matched expected alcohol pathophysiology. In addition, they exhibited excellent inter- and intra-rater reliability, and were simple enough for use by providers with different clinical backgrounds. Given these tools can be administered rapidly and require no additional technology, widespread use in the ED would be pragmatic.

    The HII and VAS assessment tools have different strengths and weaknesses. Unlike the controlled setting of this study, the ED is noisy and chaotic. As such, the detailed, rigid structure of the HII assessment tool may yield more stable results when compared to the VAS. On the other hand, the VAS may be able to detect subtle or different signs of in- toxication that are too numerous to account for in a formal scoring sys- tem (e.g. giggling, stumbling between maneuvers, the time it takes to perform the maneuver “correctly”). This may provide more flexibility for the VAS to adapt to different scenarios. Prior studies have demon- strated that the HII assessment tool is capable of quantifying the entire spectrum of intoxication; these data do not yet exist for the VAS [18]. Importantly, both assessment tools have means to address non-partici- pation by poorly cooperative patients. The VAS is scored independent of specific examination components, and the denominator of the HII is based only on the components the patient actually attempts.

    Significant problems exist with other examination-assessment com- binations. The SFST was designed to be a sensitive test of BAC >= 0.08 g/dL, the legal definition of intoxication in our region. As such, we have con- cerns that a ceiling effect will be seen when applied to severely intoxi- cated ED patients. For example, the horizontal gaze nystagmus component consistently approached the maximum score despite an av- erage BAC lower than that commonly seen in the ED [16]. Thus, the SFST may have reduced utility to quantify intoxication at higher levels. In ad- dition, the SFST examination may be challenging for some ED patients to perform correctly even when sober. For example, two of the three com- ponents require significant lower extremity dexterity, which may be impossible for patients with peripheral neuropathy from alcoholism, leg dysfunction from prior stroke, or lower extremity amputation from diabetes or trauma. In contrast, the CIE-STE combination demonstrated low sensitivity at moderate levels of alcohol intoxication. Many compo- nents of the STE are designed to assess severe levels of intoxication, re- ducing its utility to quantify all levels of intoxication. The STE also has no means to deal with non-participation.

    Importantly, the BIQ performed poorly. Clinicians using this tool

    blinded to time since alcohol challenge had difficulty determining who was sober even prior to the alcohol challenge. Given that the BIQ is most similar to current clinical practice, this further emphasizes the need to develop a standardized examination and assessment tool for level of intoxication to facilitate medical decision making.

    Limitations

    The main limitation of this study is spectrum bias, as subjects were less intoxicated than many ED patients. However, one of the long- term goals of this research is the development of a tool that facilitates clinician decision-making regarding whether a patient can be clinically evaluated without alcohol intoxication confounding patient perfor- mance. At severe levels of intoxication, the question of patient perfor- mance is less ambiguous. As such, we focused our preliminary efforts on developing a tool that performs well at lower levels of intoxication. In addition, it is challenging to expose healthy volunteers to the levels of alcohol needed to produce severe intoxication. The second limitation is that this study was conducted in a controlled environment that may not be generalizable to the ED. Similarly, a Hawthorne effect may have occurred, as subjects were aware they were being video recorded. Third, successive cycles of examination and assessment tool use may not have resulted in independence of events. We attempted to maintain independence for alcohol challenge subjects by minimizing the total number of examinations and preventing practice opportunities. For evaluators, we instructed them not to use information from prior videos to interpret subsequent scoring, and a detailed initial training session was used to reduce any learning effect. Finally, our data do not assess whether these tools can differentiate intoxication due to alcohol from other etiologies of intoxication or mental status change. Thus, future studies should evaluate the performance and utility of these tools when applied to actual ED patients.

    1. Conclusions

    This systematic evaluation of standardized examinations and assess- ment tools to measure the level of intoxication due to alcohol demon- strated that the HII examination combined with either the HII or VAS assessment tool produced a valid measurement of intoxication. Inter- and intra-rater reliability was excellent, and these tools may be suitable for a wide range of clinicians.

    Funding and disclosures

    This work was supported by Aldea Pharmaceuticals, which was not involved in study design, implementation, analysis, or the decision to publish. Authors report no personal conflicts of interest.

    Appendix A. Supplementary data

    Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ajem.2017.01.009.

    References

    1. Cherpitel CJ, Ye Y. Trends in alcohol- and drug-related emergency department and primary care visits: data from four U.S. National Surveys (1995-2010). J Stud Alco- hol Drugs 2012;73(3):454-8.
    2. Li G, Keyl PM, Rothman R, Chanmugam A, Kelen GD. Epidemiology of alcohol-related emergency department visits. Acad Emerg Med 1998;5(8):788-95.
    3. Centers for Disease Control and Prevention. National Ambulatory Medical Care Sur- vey. Rate of emergency department visits for alcohol-related diagnoses by sex: Unit- ed states, 2001-2002 to 2009-2010. http://www.cdc.gov/mmwr/preview/ mmwrhtml/mm6235a9.htm. (Accessed February, 2015).
    4. Verelst S, Moonen PJ, Desruelles D, Gillet JB. Emergency department visits due to al- cohol intoxication: characteristics of patients and impact on the emergency room. Alcohol Alcohol 2012;47(4):433-8.
    5. Pletcher MJ, Maselli J, Gonzales R. Uncomplicated alcohol intoxication in the emer- gency department: an analysis of the national hospital ambulatory medical care sur- vey. Am J Med 2004;117(11):863-7.
    6. Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E, DeBlieux PM. Indications for computed tomography in patients with Minor head injury. N Engl J Med 2000; 343(2):100-5.
    7. Hoffman JR, Wolfson AB, Todd K, Mower WR. Selective Cervical spine radiography in blunt trauma: methodology of the national emergency X-radiography utilization study (NEXUS). Ann Emerg Med 1998;32(4):461-9.
    8. Rundhaug NP, Moen KG, Skandsen T, et al. Moderate and severe traumatic brain in- jury: effect of blood alcohol concentration on Glasgow Coma Scale score and relation to computed tomography findings. J Neurosurg 2015;122(1):211-8.
    9. Shahin H, Gopinath SP, Robertson CS. Influence of alcohol on early Glasgow Coma Scale in head-injured patients. J Trauma 2010;69(5):1176-81 [discussion 1181].
    10. Mahler SA, Pattani S, Standifer J, Caldito G, Conrad SA, Arnold TC. Clinical sobriety as- sessment by emergency physicians in blunt trauma patients with acute alcohol ex- posure. J Emerg Med 2010;39(5):685-90.
    11. Gentilello LM, Villaveces A, Ries RR, et al. Detection of acute alcohol intoxication and chronic alcohol dependence by trauma center staff. J Trauma 1999;47(6):1131-5 [discussion 1135-9].
    12. Bond J, Witbrodt J, Ye Y, Cherpitel CJ, Room R, Monteiro MG. Exploring structural re- lationships between blood alcohol concentration and signs and clinical assessment of intoxication in alcohol-involved injury cases. Alcohol Alcohol 2014;49(4):417-22.
    13. Touquet R. Clinical assessment compared with breathalyser readings in the ER: con- cordance of ICD-10 Y90 and Y91 codes. Emerg Med J 2005;22(10):695.
    14. Olson KN, Smith SW, Kloss JS, Ho JD, Apple FS. Relationship between blood alcohol concentration and observable symptoms of intoxication in patients presenting to an emergency department. Alcohol Alcohol 2013;48(4):386-9.
    15. Miner JR, Gaetz A, Biros MH. The association of a decreased level of awareness and blood alcohol concentration with both agitation and sedation in intoxicated patients in the ED. Am J Emerg Med 2007;25(7):743-8.
    16. Cherpitel C, Bond J, Ye Y, et al. Clinical assessment compared with breathalyser read- ings in the emergency room: concordance of ICD-10 Y90 and Y91 codes. Emerg Med J 2005;22(10):689-95.
    17. Redmond A. BAC and Conscious level. Alcohol Alcohol 1983;18:89-91.
    18. Hack JB, Goldlust EJ, Gibbs F, Zink B. The H-Impairment Index (HII): a standardized assessment of alcohol-induced impairment in the emergency department. Am J Drug Alcohol Abuse 2014;40(2):111-7.
    19. Volz TM, Boyer KS. The development of a behaviorally-based alcohol intoxication scale. J Emerg Nurs 2014;40(4):330-5.
    20. Tharp V, Burns M, Moskowitz H. National highway traffic safety administration: de- velopment and field test of psychophysical tests for DWI arrest. http://www.nhtsa. gov/people/injury/alcohol/sfst/contents.htm. (Accessed February, 2015).
    21. Stuster J, Burns M. National highway traffic safety administration: validation of the standardized field sobriety test battery at BACs below 0.10 percent. http://www. nhtsa.gov/people/injury/alcohol/sfst/contents.htm. (Accessed February, 2015).
    22. Rubenzer S. Judging intoxication. Behav Sci Law 2011;29(1):116-37.
    23. National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism. Recommended alcohol questions. http://www.niaaa.nih.gov/research/guidelines- and-resources/recommended-alcohol-questions. (Accessed February, 2015).
    24. Watson PE, Watson ID, Batt RD. Prediction of blood alcohol concentrations in human subjects. Updating the widmark equation. J Stud Alcohol 1981;42(7):547-56.
    25. Curtin JJ, Fairchild BA. Alcohol and cognitive control: implications for regulation of behavior during response conflict. J Abnorm Psychol 2003;112(3):424-36.
    26. National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism. What is a standard drink? http://www.niaaa.nih.gov/alcohol-health/overview-alco- hol-consumption/standard-drink. (Accessed February, 2015).
    27. Sebbane M, Claret PG, Jreige R, et al. Breath analyzer screening of emergency depart- ment patients suspected of alcohol intoxication. J Emerg Med 2012;43(4):747-53.
    28. Jones AW, Andersson L. Comparison of ethanol concentrations in venous blood and end-expired breath during a controlled drinking study. Forensic Sci Int 2003;132(1): 18-25.
    29. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disor- ders. 5th ed. Washington, DC: American Psychiatric Publishing; 2013.
    30. Hallgren KA. Computing inter-rater reliability for observational data: an overview and tutorial. Tutor Quant Methods Psychol 2012;8(1):23-34.
    31. R Core Team. R: a language and environment for statistical computing. http://www. R-project.org. (Accessed June, 2015).
    32. Gamer M, Lemon J, Singh IFP. irr: various coefficients of interrater reliability and agreement. http://CRAN.R-project.org/package=irr. (Accessed June, 2015).
    33. Revelle W. psych: procedures for personality and psychological research. http:// CRAN.R-project.org/package=psych. (Accessed June, 2015).
    34. Eklund A. beeswarm: the bee swarm plot, an alternative to stripchart. http://CRAN. R-project.org/package=beeswarm. (Accessed June, 2015).

    Leave a Reply

    Your email address will not be published. Required fields are marked *