Retrospective chart review of substance abuse in patients with psychiatric emergencies in an emerging urban county

Abstract

Objectives

Substance abuse is common in patients with psychiatric emergencies. To further understand the connection between substance abuse and psychiatric disorders, a retrospective chart review was done that included positive drug screens among patients with psychiatric emergencies and to determine whether there was an association between substances used and the psychiatric diagnosis.

Methods

A retrospective chart review of patients seen in an emergency department with psychiatric emergencies was conducted. The review comprised 1000 charts with diagnoses of anxiety, depression, schizophrenia, attention‐deficit/hyperactivity disorder (ADHD), bipolar disorder, alcohol abuse, or schizoaffective disorder. Data collected included patient demographics, tobacco abuse, chief complaint, arrival mode, voluntary versus involuntary status, suicide attempt on presentation, psychiatric diagnoses, urine drug screen, and ethanol results. Chi‐square statistical analysis was conducted to examine the relationship between substances of abuse and psychiatric diagnoses.

Results

Approximately 58% of patients with a history of psychiatric illness had a positive urine drug screen. Of 245 patients with schizoaffective disorder, 69 (28%) were positive for tetrahydrocannabinol (THC) and 48 (20%) were positive for cocaine. Of 225 patients with depression, 59 (29%) were positive for THC and 33 (15%) were positive for cocaine. Cannabis was the most commonly reported substance used among patients with depression, schizophrenia, anxiety, schizoaffective disorder, and bipolar disorder, and ethanol was most common in patients with ADHD. No significant correlations were found between psychiatric diagnosis and positive drug screens. A statistically significant secondary end point was found that White people using cannabinoids were more likely to attempt suicide than were African American people (P = 0.02).

Conclusions

Positive drug screens were common among patients presenting to an ED with psychiatric emergencies. Cannabis was the most commonly reported substance used among patients independent of diagnosis. Ethanol was the most common in patients with ADHD. Urine drug screens are unlikely to provide insights into relationships between specific substance use and psychiatric emergencies.

1. INTRODUCTION

1.1. Background

Emergency department (ED) visits for behavioral health emergencies and substance abuse emergencies are increasing in the United States. ¹ , ² , ³ Dual diagnosis of substance abuse and psychiatric illness are a common and challenging problem. ⁴ , ⁵ , ⁶ , ⁷ Possible reasons for the high prevalence of dual diagnosis patients include self‐treatment, poor judgment, and the possibility that drug use can lead to chronic mental disorders. ⁸ Knowledge of psychotropic drug use is needed to properly diagnosis and treat behavioral health emergencies. Obtaining a urine drug screen on patients presenting with psychiatric emergencies can provide information regarding use of psychotropic drugs in this patient population, but it also increases costs and length of stay while not changing referral patterns. ⁹ , ¹⁰ , ¹¹ In addition, urine drug screens have limitations and do not always correlate with acute intoxication for some substances; however they have utility in determining changing patterns of drug use in a community ¹² and can be an adjunct to history. Some psychiatric consultants request that urine drug screen be routinely performed on patients presenting to the ED if psychiatric consultation and intervention is requested. As a preliminary to further study, a retrospective review was conducted of the relationship between urine drug screen results and psychiatric diagnoses in patients presenting to an academic ED with psychiatric emergencies.

1.2. Importance

Urine drug screens remain controversial and do not help with ED throughput, change length of stay, or change referral patterns. Psychiatric consultants find them helpful in determining treatment plans and interventions. Whether there is a correlation between specific drug use and specific psychiatric emergency, psychiatric diagnosis, or demographics can potentially be of benefit in understanding the increasing numbers of psychiatric visits.

1.3. Goals of this investigation

A retrospective chart review was conducted to determine if urine drug screens correlate with psychiatric diagnoses and other data.

2. METHODS

2.1. Study design

An observational retrospective cohort study of the patients presenting to an ED with psychiatric illness or substance abuse was conducted for the calendar year 2019.

2.2. Setting

The study reviewed charts from a Level 1 academic ED that sees >100,000 patients a year with 10 or more visits a day for psychiatric emergencies. Subjects were medically cleared for a psychiatric consultation by the ED staff.

2.3. Selection of participants

Participants were eligible for inclusion if they had an International Classification of Diseases, Tenth Revision (ICD‐10) code related to psychiatric illness or substance abuse as defined in Table S1. The top 6 codes that contained the largest number of charts were selected for review. A random selection of charts was reviewed for each diagnosis. Duplicate charts were removed. The final number of charts reviewed was 1000. Patients who were incarcerated or persons arrested were excluded from the study because they constitute a vulnerable population for whom additional protections are warranted. This resulted in a total sample size of 873.

2.4. Interventions

There were no interventions for this retrospective chart review.

2.5. Measurements

Demographic information collected for each patient included age, sex, marital status, home address, homelessness, insurance status, sexual orientation, and ethnicity. Health information collected for each patient included smoking status, chief complaint, date of ED visit, mode of arrival to ED for that visit, voluntary or involuntary commitment status (if involuntary who took it out), history of previous suicide attempts (and means of the attempts), and all psychiatric diagnoses. Results of urine drug screens performed using the Abbott C8000 automated analyzer were collected. Urine drug screen measured the presence of barbiturates, benzodiazepines, cocaine, opiates, amphetamines, cannabinoids, phencyclidine, 3,4‐Methylenedioxymethamphetamine (MDMA), and fentanyl. Serum levels of acetaminophen, salicylates, and ethanol level, if measured, were collected. Any ethanol level >0 was marked as positive. If the patient had a positive drug screen to a drug administered by the ED staff, it was recorded as negative. Smoking history was examined in patients with a positive urine drug screen. According to preexisting policy, every patient presenting to the ED with a psychiatric emergency received a urine drug screen in order to appropriately diagnosis and treat the emergency. This study was reviewed and approved by the University and Medical Center Institutional Review Board.

2.6. Outcomes

Relationships between use of each substance of abuse and a psychiatric diagnosis (retrieved form patient charts) were examined. Psychiatric disorders used as measurement were based on physician diagnosis notated in medical records retrieved from ED records.

2.7. Statistical analysis

IBM SPSS Program 26.0 was used to conduct statistical analyses. A chi‐square test was run to examine the relationship between substances of abuse and psychiatric diagnoses along with analyzing the difference among ethnicity groups for different substances of abuse and psychiatric diagnosis. Here, we assigned exposure as the psychiatric diagnosis and the outcome variable a positive drug screening from urine samples. Additionally, patients were separated into 2 groups: those with positive drug screens and those without positive screens. This was done to aid in researchers in this pilot study in their effort to determine potential relationships between substance use and persons with psychiatric disorders. A logistic regression was used to determine the probability of positive drugs screens for specific psychiatric diagnoses.

3. RESULTS

The sample consisted of 873 persons of whom 49.8% were men, 50.1% women, and 0.01% classifying as other. The median age was 34 years old with range from 6 years to 101 years. In this sample 57.8% of patients with a history of psychiatric illness had a positive urine drug screen. Of the drugs screened, cannabis was the most commonly reported substance used among patients with depression, schizophrenia, anxiety, schizoaffective disorder, and bipolar disorder. Ethanol was the most commonly reported substance used in patients with attention‐deficit/hyperactivity disorder (ADHD) (see Table 1). Overall, no statistically significant correlations were found between psychiatric diagnosis or suicide attempts and positive drug screens. Among patients testing positive for drugs, 56.2% were also current smokers and 15.7% were previous smokers, whereas only 24% of patients testing positive for drugs were non‐smokers (x ²(1) = 39.4, P < 0.001) (Table 2). We performed a logistic regression to estimate the likelihood of whether individuals with a specific psychiatric diagnosis also had positive drug screens. The model was statistically significant, P < 0.001, explaining 67.3% of cases. A psychiatric diagnosis of depression was a positive and significant (P < 0.001, 95% CI: 2.56, 6.89) predictor of a person having a positive drug screen. Persons with depression were 4.20 times more likely to have positive drug screen than persons presenting with other disorders. When researchers looked at specific positive screens, they found persons with depression were 7‐fold (adjusted odds ratio 7.7, 95% CI: 2.489, 24.020) more likely to test positive for ethanol. Additionally, psychiatric diagnoses of schizoaffective (P = 0.002, 95% CI: 1.43, 4.59) and bipolar (P = 0.032, 95% CI: 1.13, 15.92) were also positive and significant predictors. The likelihood of individuals having a positive drug screen with these disorders were 2.56 and 4.24, respectively. Although having a diagnosis of ADHD was a positive predictor, it was not significant (P = 0.71, 95% CI: 0.90, 13.92) (Table 3). There was no significant difference in overall suicide attempts between ethnicities (P = 0.364, see Table 4). Because African American and White people made up the 2 largest ethnic groups in the sample, these groups were compared. Results from this comparison indicated African American individuals were 60% more likely to attempt suicide than White individuals (P = 0.02, F = 17.36, df 529). When looking at patient substance use and stratified by psychiatric diagnosis, researchers found no significant difference in suicide attempts between White and African American patients using cannabinoids. However, there was a significant difference between African American and White patients diagnosed with schizophrenia or schizoaffective disorder who used cocaine and suicide attempts. Here, African American individuals were more likely to have a reported suicide attempt (P < 0.001, F = 11.012, df 500). It is important to note that ethnicity may be an effect modifier and this may be a limitation of the study.

TABLE 1.

Percentage of patients by ethnicity with selected psychiatric disorder and associated positive drug test.

		Selected psychiatric disorders
		Depression	Schizophrenia	ADHD	Anxiety	Schizoaffective	Bipolar	Total
Drug(s) Test Positive	Drug(s) Test Positive		n (%)
Amphetamines	White	16 (64)	0	8 (72.2)	0	5 (45.5)	5 (71.4)	34 (63)
	African American	6 (24)	0	2 (18.2)	0	6 (54.5)	2 (28.6)	16 (29.6)
	Hispanic	1 (4)	0	1 (9)	0	0	0	2 (4)
	Mixed	0	0	0	0	0	0	0
	Other	2 (8)	0	0	0	0	0	2 (4)
	Total	25 (7.4)	0 (0)	11 (8.1)	0 (0)	11 (5.1)	7 (8.2)	54 (6.5)
Barbituates	White	6			1	1		8 (50)
	African American	1		2		5		8 (50)
	Hispanic							0
	Mixed							0
	Other							0
	Total	7 (2.2)	0 (0)	2 (1.6)	1 (4.2)	5 (2.5)	1 (1.3)	16 (1.9)
Benzos	White	35	2	12	1	11	7	68 (68.7)
	African American	8	2	5		14	1	30 (30.3)
	Hispanic							0
	Mixed							0
	Other	1						1 (1)
	Total	44 (12.9)	4 (8.7)	17 (13.7)	1 (2.6)	25 (12.1)	8 (10.4)	99 (11.9)
Canabis	White	41	2	10	6	10	16	85 (39.4)
	African American	40	13	1	5	39	9	107 (49.5)
	Hispanic	3	0	13	1	2		19 (8.8)
	Mixed	1	0	1				2 (0.9)
	Other	2	0			1		3 (1.4)
	Total	87 (25.6)	15 (32.6)	25 (20.2)	12 (30.8)	54 (26.1)	23 (29.9)	216 (25.9)
Cocaine	White	27	1	4	0	9	8	49 (40.2)
	African American	15	4	13	3	31	2	68 (55.7)
	Hispanic	0		1	1	2		4 (3.3)
	Mixed	0						0
	Other	1						1 (0.8)
	Total	43 (12.6)	5 (10.9)	19 (15.3)	2 (5.1)	43 (20.8)	10 (13.0)	122 (14.6)
EtOH	White	34	1	19	4	15	13	86 (43)
	African American	34	6	21	0	36	7	104 (52)
	Hispanic	1	1	3	0	1		6 (3)
	Mixed	1	0			0	1	2 (1)
	Other	1	0			1		2 (1)
	Total	71 (20.9)	8 (14.4)	43 (34.7)	4 (10.3)	53 (25.6)	21 (27.3)	200 (24.1)
Fentanyl	White	4	0	0	0	0	1	5 (55.6)
	African American	1		1		1		3 (33.3)
	Hispanic							0
	Mixed							0
	Other							0
	Total	5 (1.5)	0 (0)	1 (0.8)	1 (2.6)	1 (0.5)	1 (1.3)	9 (1.1)
MDMA	White			1	1			2 (66.7)
	African American	1						1 (33.3)
	Hispanic							0
	Mixed							0
	Other	1				1	1	3
	Total	2 (0.6)	0 (0)	1 (0.8)	1 (2.6)	1 (0.5)	1 (1.3)	6 (0.7)
Opiates	White	36	1	8	2	5	11	63 (63)
	African American	14		7	0	13	2	36 (36)
	Hispanic							0
	Mixed							0
	Other	1						1 (1)
	Total	51 (15.0)	1 (2.2)	15 (12.1)	2 (5.1)	18 (8.7)	13 (16.9)	100 (12.0)
Phencyclidine	White	0	0	0	0	1	0	1 (33.3)
	African American	1				1		2 (66.7)
	Hispanic							0
	Mixed							0
	Other							0
	Total	1 (0.3)	0 (0)	0 (0.0)	0 (0)	2 (1.0)	0 (0)	3 (0.4)
Salicylates	White	1	0	0	0 (0)	0	0 (0)	1 (33.3)
	African American			1		1		2 (66.7)
	Hispanic							0
	Mixed							0
	Other							0
	Total	1 (0.3)	0 (0)	1 (0.8)	0 (0)	1 (0.5)	0 (0)	3 (0.4)
Tylenol	White	3	0 (0)	1	0	0	0 (0)	4 (80)
	African American					1		1 (20)
	Hispanic							0
	Mixed							0
	Total							0
	Other	3	0	1	0	1	0	5 (0.6)
	* Total	340 (40.8)	46 (5.5)	124 (14.9)	39 (4.7)	207 (24.4)	77 (9.2)

^*Data derived from chart review; may exceed 100% because some patients were taking more than 1 drug, and psychiatric disorders had multiple diagnosis. Therefore, values and percentages are considered crude estimates.

Abbreviations: ADHD, attention‐deficit/hyperactivity disorder; EtOH, ethanol; MDMA, 3,4‐Methylenedioxymethamphetamine.

TABLE 2.

Smoking frequency among people testing positive for drugs.

	n	%
Current smoker	283	56.2
Previous smoker	79	15.7
Non‐smoker/never smoked	121	24
Unknown	21	4.2
Total	504	100

TABLE 3.

Percentage of patients with suicidal ideations testing positive for drugs for selected diagnosis.

	Depression		Schizophrenia		Schizoaffective		Anxiety disorder		Total
	Yes	No	Yes	No	Yes	No	Yes	No
White	56.9	54.0	0.0	23.5	17.5	32.8	0.0	57.9	43.1
African American	38.9	42.5	100.0	70.6	75.0	64.2	0.0	36.8	52.6
Hispanic	2.8	2.3	0.0	5.9	5.0	3.0	0.0	5.3	3.3
Mixed	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Other	1.4	1.1	0.0	0.0	2.5	0.0	0.0	0.0	1.0
Total	23.5	28.4	1.3	5.6	13.1	21.9	0.0	6.2	100.0

TABLE 4.

Percentage of patients with suicidal ideations testing positive for drugs for selected diagnosis.

	Depression		Schizophrenia		Schizoaffective		Anxiety disorder		Total
	Yes	No	Yes	No	Yes	No	Yes	No
White	56.9	54.0	0.0	23.5	17.5	32.8	0.0	57.9	43.1
African American	38.9	42.5	100.0	70.6	75.0	64.2	0.0	36.8	52.6
Hispanic	2.8	2.3	0.0	5.9	5.0	3.0	0.0	5.3	3.3
Mixed	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
Other	1.4	1.1	0.0	0.0	2.5	0.0	0.0	0.0	1.0
Total	23.5	28.4	1.3	5.6	13.1	21.9	0.0	6.2	100.0

The Bottom Line

Dual diagnosis patients, that is, patients suffering from psychiatric illness and substance abuse disorders, constituted the majority of patients presenting to emergency departments with psychiatric emergencies, as evidenced by data collected on urine drug screens.

4. DISCUSSION

Urine drug screens increase the costs and lengths of stay of ED visits for psychiatric emergency and do not change referral patterns. ⁹ , ¹⁰ , ¹¹ They do not reflect the whole story about substance and psychiatric emergencies but are often required by psychiatric consultants for their evaluations and treatments. Positivity does not correlate with acute toxicity because screens can remain positive after acute toxicity has resolved. Further, they do not rule out substance use because levels can fall below the limit of sensitivity by collection time. However, they can provide information about the prevalence of dual diagnosis patients who have a substance abuse disorder concurrent with a psychiatric illness. Previous studies have shown that patients with mental illness have higher rates of substance use than the general population. ⁸ The Epidemiologic Catchment Area study found a lifetime prevalence of substance use disorder to be 16.7%, whereas the prevalence of substance use disorders in patients with mental illness was nearly doubled at 29%. ¹³ Our study found that a majority of patients with a history of mental illness tested positive (57.8%) for illicit substances. Previous literature has found an association between high levels of cannabis use in adolescents and an increased likelihood of being diagnosed with depression, anxiety, and schizophrenia later in life. ¹⁴ Our study in adults leads to a similar association, as the most commonly reported substances used by patients with a history of depression, schizophrenia, anxiety, schizoaffective disorder, and bipolar disorder was cannabis. A meta‐analysis ¹⁵ found that children with ADHD were 1.7 times more likely to later develop alcohol use disorder than children without ADHD. Our results found this association in adults, as the most commonly reported substance used by patients with a history of ADHD was ethanol. Further study is needed to learn the efficacy of addressing ethanol use in patients with ADHD in an effort to prevent the development of alcohol use disorder. A previous meta‐analysis suggests that cannabis use in young adults is associated with higher rates of suicidality. ¹⁶ An incidental finding in this chart review was a statistically significant association that African American individuals using cannabinoids were more likely to attempt suicide than White individuals using cannabinoids. Further study is needed to determine if this association holds true in a broader context, and if so, there is an underlying risk factor predisposing African American individuals using cannabinoids to attempt suicide relative to White individuals.

5. LIMITATIONS

Limitations of the study include a possible selection bias as complete randomization was not achieved due to the large number of charts meeting our selection criteria based on the ICD‐10 codes. Additionally, there are psychiatric disorders that have overlap (ie, schizoaffective disorders with depression, etc.), hence it is possible this affected our results. Our results may also be skewed by the small sample size of certain substances as we had low rates of patients testing positive for barbiturates, fentanyl, MDMA, and phenylcyclohexyl piperidine (PCP). Not all abused substances, such as inhaled solvents, are included in routine drug screens or addressed in histories obtained. Subjects with levels below the sensitivity of the assays would have been counted as non‐users. Additionally, because this study was conducted retrospectively, potential gaps in documentation of patient information in the chart may lead to misrepresentation of data. Our results may also be limited to regional rates of substance abuse and mental illness as this study was conducted with patients at 1 institution. Urine drug screens capture only recent use so the positivity rates underestimate substance abuse among patients with psychiatric emergencies.

6. CONCLUSIONS

Positive urine drug screens and blood alcohol levels were common among patients presenting with a psychiatric emergency, with cannabis and ethanol being the most commonly found. As this was a pilot initiative, further study through a prospective investigation could provide better insight into the relationship between substance abuse and psychiatric emergencies. Urine drug screens are unlikely to provide insights into relationships between specific substance use and psychiatric emergencies. This study found that urine drug screens did not give insight to the root causes of mental illness, except possibly an association with alcohol and ADHD that needs further investigation.

AUTHOR CONTRIBUTIONS

Conceptualization: William J. Meggs. Data collection: Simran Koura, Joshua Masdon, Avian White. Data analysis: Avian White. Critical review and evaluation of results: Simran Koura, William J. Meggs, Avian White, Joshua Masdon, Kori L. Brewer, Jennifer L. Parker‐Cote. Primary authorship of the paper: Simran Koura. Review and editing of the paper: Simran Koura, William J. Meggs, Avian White, Joshua Masdon, Kori L. Brewer, Jennifer L. Parker‐Cote. Study supervision: William J. Meggs. Procurement of grant or other funding: Simran Koura, William J. Meggs, Kori L. Brewer.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to disclose.

Supporting information

Supporting information

Biography

Simran Koura, MD, is a resident at Loyola Medicine, Illinois.

Koura S, White A, Masdon J, Brewer KL, Parker‐Cote JL, Meggs WJ. Retrospective chart review of substance abuse in patients with psychiatric emergencies in an emerging urban county. JACEP Open. 2023;4:e13028. 10.1002/emp2.13028