Alcohol Misuse and Report of Recent Depression Symptoms among Emergency Department Patients

Abstract

Objective

This study examines the magnitude of association between alcohol misuse with recent depressive symptoms.

Methods

A cross-sectional study of 412 randomly selected patients ≥18 seeking ED care.

Results

51.0% reported depressive symptoms. At-risk drinking was reported by 26.0% of the sample, 28.2% scored positive for RAPS4. Alcohol abuse, and Binge drinking were reported by 25.1% and 28% of the patients, respectively. According to our results at-risk drinking (OR 2.49; CI 1.47–4.20; P ≤ .001), problem drinking (OR 2.11; CI 1.27–3.51; P ≤ .004), drinking abuse (OR 2.58; CI 1.51–4.40; P < .001), and binge drinking (OR 1.89; CI 1.13–3.15; P < .001) were all related to manifestation of depression symptoms.

Conclusion

The findings of this study yield information that could be used by emergency department health care practitioners and health educators to educate ED patients at-risk for alcohol misuse and depression.

INTRODUCTION

Background

It is well documented that depression is a risk factor for number of physical, functional and social impairments [1]. However, some evidence suggest depressed patients may not meet diagnostic criteria for depression and instead present vague somatic complaints [2] that preclude clinical or self assessment of mental health illness. These subsyndromal depressive symptoms nonetheless could leads to serious distress [3].

While alcohol use is common in the United States there are many forms of excessive drinking that lead to significant medical and mental health problems. They include high level drinking each day, repeated episodes of drinking to intoxication (binge drinking), drinking that result in physical or mental harm (alcohol abuse), and drinking that results in chemical dependence to alcohol (alcohol dependence)[4].

Longitudinal studies have been inconclusive in identifying alcohol as a risk factor for depression [5]. Some studies have documented an elevated frequency of depressive disorders in alcoholic clinical and community populations [1, 6, 7], young adults with hazardous and harmful consumption [8], and elderly population with heavy drinking, and binge drinking. Even among former alcohol users, specific associations exists between prior alcohol dependence and current or recent major depression [10].

Studies which have documented the relationship between the symptomatology of depression and level of alcohol use [11, 14–16], Graham, et al. suggests that lack of consistency in relationship between alcohol consumption and depression could be in part related to way these variables are measured. For example depression has been measured as recent depressive symptoms using Epidemiolgical Studies Depression Scale (CES-D) [11] or as a psychiatric diagnoses according to DSM-III criteria, such as the one used by Diagnostic Interview Scheduled (DIS) [12]. Studies using the impact of these measures on alcohol have shown various findings [13, 14]. Similarly, studies which have measured alcohol consumption as volume of consumption [15, 16] frequency of drinking [9] or quantity per occasion [15, 17] have shown varying results.

Studies of ED populations have been especially important in discerning the relationship between the level of alcohol use and recent depression symptoms, given the high prevalence of alcohol use among ED patients [18, 19], and the lack of access to primary care to receive necessary treatment [20]. In fact primary care providers often are the sole contacts for more than 50% of patients with mental illness. Furthermore, numerous studies point to significant episodes of under-recognized and untreated depression in African American and Hispanic communities {Kessler RC, 2003 #3307; Minsky S, 2003 #3308; Chung H, 2003 #3309}.

The overall goal of this study is to find out if the association between alcohol misuse and recent symptoms of depression is sensitive to the level of alcohol consumption. The specific aim of this study is to examine the association between four levels of alcohol misuse including: at-risk drinking, problem drinking, alcohol abuse, and binge drinking with recent depressive symptoms measured by CES_D among a random sample of Emergency Department (ED) patients receiving care from an inner city hospital. The study null hypothesis is that there is no statistically significant association between levels of alcohol misuse and recent symptoms of depression.

METHODS

Setting

Data for this study was collected continuously for 24 hours per day over 35-day period between March–April, 2001, in an inner-city emergency department (King/Drew Medical Center Emergency Department (KDMC-ED). This hospital is the primary provider of healthcare services for the 1.7 million residents in South Central Los Angeles with an annual census of 50,000–60,000 ED visits. South Central Los Angels is one of the most heavily populated and ethnically, socially, and economically diverse counties in the United States. The community serviced by (KDMC-ED) is predominantly Latino (60.0%) and African-American (23%). Approximately 32.0% of the population lives below the federal poverty level, and 36% of adults are uninsured.

Selection of participants

Patients were considered eligible for the study if they were 18 years of age or older, and were in the KDMC-ED to receive medical care. Patients were considered ineligible for the study if they showed any sign of cognitive impairment, if they spoke a language other than English or Spanish, or if they were in police custody. Patients who required immediate medical attention as determined by the attending doctor were approached to participate in the study following their treatment. Interviewers delayed their work with sampled patients who showed signs of intoxication. patients who gave written informed consent to participate were included. Hispanic respondents were given a choice of being interviewed in English or in Spanish. KDMC-ED computerized logs were used to select study sample. This log reflected consecutive patients who arrived and registered in the ED triage area. This list was continuously generated and updated. Every other consecutive patient was selected as a potential candidate. Every other consecutive patient was selected to avoid overload of potential candidates on study staff. A total of 412 eligible patients provided informed consent and completed the 45-minute face-to-face study survey. This study received full review and approval by the institutional review board of our institution.

Study Measures

The main outcome variable for this study was “recent depressive symptoms” which was measured using the Center for Epidemiologic Studies Depression Scale (CES-D) [11] providing an estimate of depressive symptomatology prevalence over the last 7 days such as frequency of experiencing loss of appetite, feeling happy or depressed, lack of energy, and having crying spells. Numerous studies have found the scale to be valid and reliable (Coefficient alpha = .80) for this purpose [21]. The final score was constructed from a list of 20 items scored from 0 to 3 based on frequency of occurrence of the symptom (“rarely”= 0; “most of the time”= 3) with the possible range of 0 to 60. A response for each item was summed. Respondents with the overall sum score 15 or less were classified as “Not having depressive symptoms” = 0 and were set as the reference group, and respondents whose overall sum values were equal or more than 16 were classified as “Having symptoms of depression” = 1.

Alcohol Measures

Five validated and widely used measures of the alcohol consumption were used to depict a continuum of alcohol misuse in the last 12 months. They include: At-risk drinking was measured by AUDIT (Alcohol Use Disorders Identification Test) which is a 10-item index with validated psychometric properties [22]. Items are scored from 0–4 with the composite index ranging from 0–40. A cutoff point of 8 is set to reflect at-risk drinking.

RAPS4 [23] [24] is a screening test that measures problem drinking using four items with the score of 0 = No or 1= Yes. Positive scores on any one of the questions is considered positive on RAPS4 for problem drinking.

Alcohol-Abuse symptoms were measured by six items reflecting significant negative physical, social, legal, and psychosocial consequences of drinking [25]. These items operationalized DSM-IV criteria for alcohol abuse [24]. They included the following: (1) Have you continued to drink even though you knew it was causing trouble with your family or friends? (2) Did you drive a car, motorcycle, truck, boat, or other vehicle after having too much to drink? (3) Did you get into a situation while drinking or after drinking that increased your chances of getting hurt, like swimming, using machinery, or walking in a dangerous area or around heavy traffic? (4) Did you get arrested or held at a police station because of your drinking? (5) Did you get drunk or have a hangover when you were supposed to be doing something important like being at work, school, or taking care of your home or family? (6) Did you get drunk or have a hangover when you were actually doing something important like being at work, school, or taking care of your home or family? Each item was scored as zero vs. one. Positive on one or more of the items was considered positive for alcohol abuse.

Binge drinking was measured by asking respondents to report the number of standard drinks they usually take at one time/occasion. Those reporting drinking 5 or more drinks at one time during the last 12 months were classified as binger drinkers = 1, with the reference group being those reported 4 or less drinks at one time during the same period = 0.

Socio-demographic Measures

Demographic variables used in this study are listed below with the category specified as the comparison group shown in italic. Gender (female vs. male); age (coded as three categories: 18–35, 36–55, > 56); education (less than high school diploma vs. high school diploma or more); marital status (married or living with someone vs. single, separated or divorce, widowed); employment (Employed fulltime/Part-time/Self vs. unemployed). Ethnicity was categorized as African American VS. Latino. Thirteen (13) patients who identified themselves as White, Asian, Middle Eastern and Native American Indian were excluded from the analysis in addition to four (4) who refused answering this question.

Primary Data Analysis

Descriptive statistics were generated for the demographic and main variables in the study. Bivariate associations between each of the four alcohol measures were calculated through the phi-coefficient. This statistic can be interpreted as a correlation coefficient for variables in a 2×2 table where values near 0 indicate little association and values near 1 indicate perfect predictability. To assess the impact of each alcohol measure on depression symptoms, Pearson Chi-Square tests were performed. These analyses establish whether or not the independent variables are related to the main outcome. A series of multiple logistic regression analyses were conducted to predict the relationship between depression symptoms and alcohol predictors controlling for socio-demographic characteristics. A variable was retained in the model as a confounder if it altered the odds ratio (OR) of one or more variables in the model by ≥10% [26] or remained a significant independent risk factor at the 0.10 level upon entry into the model. All analyses were performed using SPSS Inc. version 12.0, (2004). A p-value ≤0.05 was considered statistically significant.

RESULTS

From a sample of 579 eligible patients, 412 consented to be interviewed, representing a 71% completion rate. Among the eligible patients, the primary reason for non-response was refusal (23%, n= 133) due to discomfort, hearing difficulties, and medical condition interfering with ability to talk. No statistically significant difference was detected between gender, ethnicity and age of those interviewed and those not interviewed.

Characteristics of study subjects

Table 1 includes the overall characteristics of the sample as well as results of Bivariate associations between independent variables and depression. The study sample included, 58.9 % female and 41.1% male, and the majority were not married (61.0%). Latino and African American participants were 53.2% and 46.8%, respectively. Nearly forty six percent (45.9%) of the participants were between the ages of 18–35, and 64.0% were unemployed, with nearly half of the sample (49.8%) having greater than a high school education.

Table 1.

Overall Sample Characteristics and its Bivariate Association with Depression (N = 412).

Variable	Categories	F	(%)	Not Dep. < 16 (n = 201)		Depressed ≥ 16 (n = 208)
				F	%	F	%
Ethnicity	Latino	217	53.2	104	47.9	113	52.1
	AA	191	46.8	96	50.3	95	49.7
Marital Status	Married	159	39	83	52.2	76	47.8
	N. Married	249	61	117	47.0	132	53.0
Gender	Female	243	59	126	52.3	115	47.7
	Male	169	41	75	44.6	93	55.4
Age	18–35	187	45.9	92	49.2	95	50.8
	36–55	165	40.5	77	46.7	88	53.3
	55+	55	13.5	31	56.4	24	43.6
Education*	≥HS	203	49.8	110	54.2	93	45.8
	< HS	205	50.2	90	43.9	115	56.1
Employment*	No	261	64	117	44.8	144	55.2
	Yes	147	36	83	56.5	64	43.5
Depression	No	201	49.1	--	--	-	--
	Yes	208	50.9	--	--	--	--
At-Risk Drinking*	> 8	268	73.6	146	54.5	122	45.5
	≥ 8	96	26.4	33	35.1	61	64.9
RAPS4*	> 1	260	71.4	140	53.8	120	46.2
	≥ 1	104	28.6	39	38.2	63	61.8
Alcohol Abuse*	> 1	271	74.5	148	54.6	123	45.4
	≥ 1	93	25.5	31	34.1	60	65.9
Binge Drinking*	No	263	72.3	138	52.5	125	47.5
	Yes	101	27.7	41	41.4	58	58.6

^*Variable Statistically Significant p < 0.05

A large percentage of the sample reported depressive symptoms (51.0%). Among the alcohol users (n= 364, 88.0%), at-risk drinking was reported by 26.0% of the sample, while 28.2% scored positive for RAPS4. Furthermore, alcohol abuse and binge drinking at least once during the past 12 months were reported by 25.1% and 28% of the respondents, respectively.

Characteristics of sample by depression

Results of Chi-Square tests (Table 1) revealed that patients with depressive symptoms (CES_D ≥ 16) were more likely to be at-risk drinker based on their AUDIT score (64.9%), and problem drinkers based on their RAPS4 scores (61.8%), (P ≤ .001). Furthermore, patients with depressive symptoms were more likely to report abusing alcohol (65.9%) and binge drinking (47.5%) (P ≤ .005) (Table 1). Among socio-demographic variables, we found statistically significant association between education, employment, and symptoms of depression. Slightly over fifty six percent (56.1%) of the patients with less than high school degree, and 55.2% of patients with unemployment status reported symptoms of depression (P ≤ .05).

Associations among the Alcohol Measures

The phi coefficient was used to measure the degree of association between each of the alcohol measures as shown in Table 2. The values range from 0.51 to 0.76 indicating an appreciable magnitude of association between each of the alcohol measures. Although these values don’t indicate perfect predictability it does provide moderately strong support that a person identified as meeting criteria for alcohol misuse on one measure will also be meeting the criteria for alcohol misuse on the remaining alcohol measures.

Table 2.

distribution of Phi-Coefficient Among the Four Alcohol Measures

	AUDIT	RAPS4	Alcohol Abuse	Binge-Drinking
AUDIT	1.00	.0698	0.764	0.673
RAPS4	-	1.00	0.661	0.518
Alcohol Abuse	-	-	1.00	0.636
Binge Drinking	-	-	-	1.00

Recent Depression and At-risk Drinking

Result of the adjusted regression analysis with AUDIT as the main independent alcohol measure along with other socio-demographic variables in the study is reported in Table 3, Model 1. According to this table at-risk drinking (OR 2.49; CI 1.47–4.20; P ≤ .001), male gender (OR 1.86; CI 1.16–2.97; P ≤ .009), having less than high school degree (OR 1.72; CI 1.05–2.82; P ≤ .031), unemployment, (OR .554; CI .345–.891; P ≤ .015), and younger age (OR 2.34; CI 1.24–4.88; P ≤ .023) all are independent predictors of report of depression symptoms.

Table 3.

Four Multiple Logistic Regression Models Presenting Association between Alcohol Misuse, and Self Report of Depression Symptoms.

	Model 1		Model 2		Model 3		Model 4
	AUDIT†		RAPS4†		Alcohol Abuse†		Binge Drinking
	ORab	95% CIc	ORab	95% CIc	ORab	95% CIc	ORab	95% CIc
Alcohol variable	2.49	1.47–4.20	2.11	1.27–3.51	2.58	1.51–4.40	1.89	1.13–3.15
Age in years	2.49	1.12–4.88	2.5	1.21–5.22-	2.31	1.11–4.82	2.43	1.17–4.96
Gender	1.86	1.16–2.97	1.83	1.14–2.91	1.87	1.17–2.98	1.83	1.14–2.94
Ethnicity	NS	--	NS	--	NS	--	NS	--
Education	1.72	1.05–2.82	1.66	1.01–2.72	NS	--	1.73	1.06–2.83
Marital status	NS	--	NS	--	NS	--	NS	--
Employment	.554	.345–.891	.579	.362–.927	.563	.352–.907	.563	.351–.900

^aOdds Ratio displayed if value statistically significant (p < 0.05)

^bExpected β-value

^c95% Confidence Interval

Recent Depression and Problem Drinking

Model 2 in Table 3 shows result of the regression analysis with RAPS4 as the main independent alcohol variable along with other socio-demographic variables. These results reveal a statistically significant association between problem drinking and report of depression symptoms while controlling for other variables in the model (OR 2.11; CI 1.27–3.51; P ≤ .004). Similar to the previous model being male (OR 1.83; CI 1.14–2.91; P ≤ .011),), having less than high school degree (OR 1.66; CI 1.01–2.72; P ≤ .042), unemployment, (OR .579; CI .362–.927; P ≤ .023), and younger age (OR 2.51; CI 1.21–5.22; P ≤ .013) remained to be independent predictors of symptoms of recent depression.

Recent Depression and Alcohol Abuse

Results for the independent association between alcohol abuse and reports of recent depression is reported in model 3, Table 3 This results reveal that alcohol abuser are two and half times more likely to report symptoms of depression than non- abusers (OR 2.58; CI 1.51–4.40; P < .001). Furthermore, being male (OR 1.87; CI 1.17–2.98; P < .009), unemployment, (OR .565; CI .352–.907; P < .018), and younger age (OR 2.31; CI 1.11–4.82; P < .025) remained to be independent predictors of symptoms of depression.

Recent Depression and Binge Drinking

Last model in Table 3 reveals association between binge drinking and depression while controlling for other socio-demographic variables. This model also points to a statistically significant association between binge drinking and report of depression symptoms. Binge drinkers are nearly two times more likely to repot symptoms of depression vs. non-binge drinkers (OR 1.89; CI 1.13–3.15; P < .001). Furthermore, being male (OR 1.83; CI 1.14–2.94; P < .014), having less than high school education (OR 1.73; CI 1.06–2.83), unemployment, (OR .563; CI .351–.900; P < .017), and younger age (OR 2.43; CI 1.17–5.04; P < .014), remained to be independent predictors of symptoms of depression.

Discussion

In this study we have tested if different measures of alcohol misuse impacts recent depression symptomotology in a similar manner among a sample of ED patients. Our findings indicate a high prevalence of recent depression symptoms, (51%) in this sample. This rate is twice as high as depressive symptoms rate in adult population (24%) suggesting that diagnosis of depressive symptoms in ED patients may present unique challenges for healthcare delivery systems, health providers and patients.

Across measures of alcohol misuse, the highest prevalence of alcohol misuse is found for problem drinking based on RAPS4 (28.6%) with an appreciable association among the four measures. Additionally, our findings reveal a statistically significant association between alcohol misuses as measured by the four different screening tools (RAPS4, AUDIT, DSM4-Abuse, binge drinking). and recent symptoms of depression. Furthermore, we observe an overlap between 95% CI’s of the four measures’ (Table 3) suggesting a common magnitude of association between alcohol misuse and recent depression. This is expected since we observed a moderately strong association between all the four alcohol measures.

In general our results are consistent with current findings pertaining to the role of alcohol in presentation of some level of depression, even though the data for this study is 6 years old. In a sample of 14,063 Canadian residents using four types of alcohol measures (frequency, usual and maximum quantity per occasion, volume, and heavy episodic drinking) and two types of depression measures (major depression, and recent depressed affect) Graham et al, 2007 revealed that major depression was primarily related to drinking large quantities of alcohol per occasion, less related to drinking volume and unrelated to drinking frequency. In another study Blow et al [27] reported that individual’s at-risk and problem drinking elevated the risk of depressive symptoms. Results of a fourteen years longitudinal study revealed that at the baseline and 14-year follow-up alcohol consumption was linearly and positively associated with depressive symptoms; prevalence of symptoms increased with greater alcohol consumption [28]. Golstein, using a community survey, showed that the prevalence of depression increased significantly across drinking groups (i.e., 25% among minimal drinkers, to 30% moderate drinkers, to 44% among heavy drinkers) [29]. Furthermore, results from the 2005 National Survey on Drug Use and Health: National findings point to the associations between mental disorders and alcohol use behaviors that do not meet the criteria for substance use disorders such as heavy drinking (i.e., drinking five or more drinks on the same occasion) [30].

It is difficult to compare our findings with those in the literature since we examined a different population and used different measures of alcohol and depression. However, we can conclude that there is statistically significant evidence of an association between alcohol misuse and depression symptomotoloy and the magnitude of this association is consistent across our measures of alcohol consumption (i.e. RAPS, AUDIT, DSM IV- Abuse, and binge drinking.

Results of this study cautions ED health professionals not to overlook evaluation of depressive symptoms among ED patients who experience problems related to alcohol misuse but do not meet criteria for alcohol dependence. ED care providers should be wary about the mental outcomes of at-risk drinking as much as they should be concern about problem drinking, beige drinking, or drinking abuse. This is especially important among their patients since they are more likely to be at higher risk for alcohol misuse and depression [31–33] however are unwilling to disclose these problems to their physicians for fear of being stigmatized [34]. For example, African Americans have significantly lower rates of treatment seeking and are less likely than Whites to find the use of conventional antidepressant medication acceptable [35]. Therefore, screening for alcohol misuse in the ED may expose patients to early evaluation, prevention, and treatment of depression which otherwise may go undetected for these patients [32, 36, 37]. Shared responsibilities of ED sector regarding patients’ mental health needs may curtail the disparity gap that exist in depression treatment due to lack of access to health care services [38, 39].

In this study we were also able to show the differences between ED patients who reported symptoms of depression and those who didn’t report such symptoms in respect to age, gender, education, and employment status. Overall, our findings revealed that among this sample of ED patients those who were male, younger, had less than high school education, and were unemployed were more at risk for symptoms of depression. Previous studies have highlighted female gender, younger age, and low socio-economic status (e.g., unemployment, lower education) to be among other numerous other risk factors for depression.

There is an opportunity and need for ED health care providers to ensure access to different mental treatment modalities among patients that would benefit. Findings of this study regarding the co-existence of alcohol misuse and depressive symptoms for ED population will have implications for prevention and treatment for this population. Results of this study may also elucidate some of the inconsistencies issues that exist in the literature regarding expression of alcohol misuse on depression symptomotology.

One major limitation of this study has to do with the inheriting nature of cross-sectional study, which doesn’t allow separation of cause from effect, and only allows the discussion of association between two variables, without respect to their direction. In addition, our study has the following limitations: self-report, bias related to respondent recall, and social desirability of responses. This study utilized standardized measures and intensive training of research assistants to minimize any such bias. In addition to the above, the study instruments included several alcohol related measures for assessing subject’s alcohol consumption to guard against acquiescence bias (subject’s tendency to express agreement or disagreement toward a statement regardless of its content) [40]. Reliability of alcohol intake based on self-report is based upon evidence for a close correlation between biological markers and self-report of alcohol intake.[41, 42].

Conclusion

This paper attempts to evaluate the role of various levels of alcohol misuse in depression symptomotology among ED patients using a variety of widely-used measures of alcohol intake. We conclude that there is statistically significant evidence of an association between alcohol misuse and depression and the magnitude of this association is consistent across all measures of alcohol consumption (i.e. RAPS, AUDIT, DSM IV- Abuse, and binge drinking). The findings of this study yield information that could be used by emergency department health care practitioners and health educators to educate ED patients at-risk for alcohol misuse and depression. Cautioning ED providers that the role of at-risk drinking, binge problem drinking and drinking abuse in depressive symptoms should not be underestimated. The focus on minorities, systematic sampling, collection of data on a 24-hour basis, and achieving high rates of participation are among other factors that strengthen the results of this study.