Mental Health Emergency Department Visits: An Exploration of Case Definitions in North Carolina

Abstract

Background:

Mental health (MH) disorders comprise a high disease burden and have long-lasting impacts. To improve MH, it is important to define public health MH surveillance.

Methods:

We compared MH related definitions using ICD-10-CM codes: The Council of State and Territorial Epidemiologists’ (CSTE) surveillance indicators for all MH, mood or depressive, schizophrenic, and drug/alcohol-induced disorders; and North Carolina’s (NC) syndromic surveillance system’s definition for anxiety/mood/psychotic disorders, and suicide/self-harm. We compared code definitions and frequent codes in 2019 emergency department (ED) data for those age ≥10 years.

Results:

CSTE’s definition resulted in over one million MH-related visits (23% of all ED visits) and NC’s definitions in 451,807 MH-related visits (9% of all ED visits). Using CSTE’s broadest definition, nicotine use was the most common visit type; using NC’s definitions, it was major depressive disorder.

Conclusions:

Standardizing population-level MH indicators benefits surveillance efforts. Given its prevalence, efforts should focus on documenting MH to improve treatment and prevention.

Introduction

The World Health Organization states that “there is no health without mental health (MH)”[1]. MH is an integral part of well-being, yet MH disorders comprise the highest burden of disease in the United States (US) and often have long-lasting impacts [2]. MH disorders also frequently co-occur, with around 20% of those with MH disorders also having substance use (SU) disorders and vice versa [3]. Yet, the way MH and co-occurring MH/SU disorders are defined and documented is inconsistent and falls under the umbrella of behavioral health (BH), which is also ill-defined in population surveillance. This lack of consistency and clarity creates challenges in consistent public health surveillance and response efforts. Emergency departments (ED’s) are safety net providers for all crises, including for BH and ED boarding of patients has become a significant challenge nationwide [4–5]. To reduce the burden of BH visits on EDs, accurate BH surveillance is critical to ensure efficient distribution of resources to communities [5]. Adding to the complexity are varied definitions for both independent MH and interrelated MH/SU disorders. To develop adequate public BH surveillance, it is important to first develop agreed upon terminology and case definitions for MH, SU, and co-occurring disorder presentations to emergency departments.

The burden of BH disorders has increased during the COVID-19 pandemic, as ED visits related to MH crises, alcohol consumption – associated with MH issues such as anxiety and depression – and overdoses have been on the rise [4–6]. These visits are all directly and/or indirectly related to MH, yet there is no standard definition as to what comprises a MH ED visit versus a SU visit or what defines a co-occurring disorder (e.g., whether an ED visit for alcohol intoxication is a MH disorder or a co-occurring MH/SU disorder), leading to variable estimates of MH’s independent and co-occurring burdens. The increase in BH alongside COVID-19 provides an example of the need for timely surveillance of syndemic effects [9]. This timely surveillance is not only uniquely enabled by responsive syndromic surveillance systems, but for many, the ED may be their only primary care setting for BH, including during emergencies [10]. The Centers for Disease Control and Prevention (CDC) reports also indicate that the proportion of MH ED visits has increased during the COVID-19 pandemic, but their definition is not clearly stated [11]. Additionally, public health surveillance systems have created agreed upon definitions for overdose (OD) events but not created standard syndromic surveillance case definitions to signal potential upticks in MH disorders [12]. Furthermore, OD events are not the only presentation of a primary SU disorder and thus underestimate the prevalence of SU disorders. These population health surveillance definitions differ from the Diagnostic and Statistical Manual (DSM-V) definitions used by clinical settings to diagnose both mental illness and SU disorders at the individual level [13]. For improving public health surveillance, intervention targeting, resource allocation, and ultimately reducing the burden that EDs face in seeing MH/SU emergencies, it is vital that the medical and public health communities have a common set of surveillance case definitions for MH disorders and co-occurrence with other conditions, particularly SU disorders and OD events. Our objective was to begin standardizing the definition of a MH disorder by comparing commonly used definitions for MH disorders in North Carolina.

Methods

The Council for Territorial and State Epidemiologists (CSTE) provided broad definitions for categories of MH disorders in 2019 using International Classification of Disease-10-Clinical Modification (ICD-10-CM) codes – one for all mental, behavioral, and neurodevelopmental disorders, and specific definitions for drug and alcohol-induced disorders, mood or depressive disorders, and schizophrenic disorders [14]. Thus, CSTE includes some SU disorders in their MH disorder definitions. On the other hand, North Carolina’s Disease Event Tracking and Epidemiologic Collection Tool (NC DETECT), NC’s syndromic surveillance system, has several MH definitions available to local data users, including one for anxiety, mood, and psychotic disorders, and one for suicide and self-harm [15]. To explore how overlapping or disparate the CSTE and NC DETECT definitions are and to begin creating a standardized definition for MH-related ED visits, we compared eight specific and combined MH definitions from CSTE (five definitions) and NC DETECT (three definitions) using ICD-10-CM codes (Table 1) using any mention of these definitions across an array of NC’s 37 diagnosis fields. These definitions were run for all NC ED visits among NC residents ages 10 and up in 2019. This included 4.9 million ED visits across 125 EDs. To compare definitions, we tracked overlap in ICD-10-CM codes and compared the most frequent MH codes used in the full ED data set. All analyses were conducted in R version 4.0.3.

Table 1:

ICD-10-CM Code Comparison for Mental Health Related Emergency Department Visits

			CSTE					NC DETECT
Code Group	ICD-10-CM Codes	Description	Combined CSTE Definitions	Drug & Alcohol	All Except Drug & Alcohol	Schizophrenic Disorders	Mood or Depressive Disorders	Combined NC DETECT Definitions	Anxiety, Mood, & Psychotic Disorders	Suicide & Self-Harm
F	01–09	Mental disorders due to known physiological conditions
	10–19	Mental and behavioral disorders due to psychoactive substance use
	20–29	Schizophrenia, schizotypal, delusional, and other non-mood psychotic disorders
	30–39	Mood [affective] disorders
	40–49	Anxiety, dissociative, stress-related, somatoform and other nonpsychotic mental disorders
	50–59	Behavioral syndromes associated with physiological disturbances and physical factors
	60–69	Disorders of adult personality and behavior
	70–79	Intellectual disabilities
	80–89	Pervasive and specific developmental disorders
	90–99	Behavioral and emotional disorders with onset usually occurring in childhood
T	14.91	Suicide Attempt
	36–50	Poisoning by, adverse effect of and underdosing of drugs, medicaments and biological substances
	51–65	Toxic effects of substances chiefly nonmedicinal as to source
	71	Asphyxiation, intentional self-harm
X	71–83	Intentional self-harm
Z	04.6	Encounter for general psychiatric examination, requested by authority
	13	Screening for disorder
	65.8	Other specified problems related to psychosocial circumstances
	72–73	Problems related to lifestyle; life-management difficulty

Results

Besides the CSTE and NC DETECT combined definitions, the CSTE specific definition for any MH except drug and alcohol is the broadest, followed by NC DETECT’s anxiety, mood, and psychotic disorders definition (Figure 1 Panels A and B). By CSTE’s any MH disorders definition, we found over one million MH-related ED visits in NC in 2019 – making up 23% of ED visits in the state. By using NC DETECT’s combined definitions as described above, we found 451,807 MH-related ED visits in NC in 2019 – making up 9% of visits. This discrepancy is largely due to ED visits captured only by the CSTE definition that include drug and alcohol codes. There were an additional 19,044 visits only captured by NC DETECT’s anxiety, mood, and psychotic disorder definition and 2,811 visits only captured by NC DETECT’s suicide and self-harm definition. However, there is overlap between visits (Figure 1 Panel C). For example, most visits with a suicide or self-harm indicator (N=8,650) also have at least one other mental health code and 428,902 visits have codes for both NC DETECT anxiety, mood, and psychotic disorders and one of the CSTE definitions. That said, the CSTE definitions miss many suicide-related, self-harm, and other mental health codes despite its broad definition while capturing intellectual and developmental disabilities, which are qualitatively different from other MH visits. When including drug and alcohol codes (F10–19), nicotine dependence (F17.210) was the most common MH code, which is not typically the cause of a MH-related ED visit. Excluding drug and alcohol codes, a single episode of a major depressive disorder (F32.9) was the most common code used for MH visits in the dataset.

Figure 1.

CSTE MH-Related Definitions (Panel A), NC DETECT MH-Related Definitions (Panel B), and Case Definition Overlap (Panel C) in 2019 North Carolina Emergency Department Visits of NC residents age 10 and older.

*This is not a formal NC DETECT case definition. We combined the anxiety, mood, and psychotic disorder and suicide and self-harm definitions to assess overlap.

Discussion

The CSTE and NC DETECT definitions used in this analysis are broad. While there is overlap between the ED visits they capture, there are also large discrepancies in what ED visits are included as MH visits. These discrepancies make it difficult to conduct public health surveillance work, characterize MH burden, and prevent future MH disorders. The discrepancies also suggest useful MH categorizations may be context dependent. This work also highlights the difficulty in categorizing ED visits as co-occurring MH/SU visits. For example, if someone has a code for nicotine dependence (captured in CSTE’s MH definition) and an overdose code, that may not be a co-occurring MH/SU visit, only a SU visit. Prior work has highlighted this issue, categorizing some F1 codes for alcohol use as SU rather than MH [16].

MH ED visits only capture one aspect of MH – primarily acute MH crises or when individuals lacking access to community-based MH treatment seek care in ED settings [17]. These visits range from self-harm to psychosomatic panic, severe anxiety, psychosis interfering with ability to care for self, or behavior that prompts an individual to be brought to the ED by family members or law enforcement. The lack of access to community-based care contributes to prolonged wait-times in already overcrowded EDs and the use of scarce ED resources [18]. Psychiatric boarding, waiting in hallways or other ER areas for inpatient beds, has become an increasingly prevalent problem for those presenting to the ED with a MH crisis. It is important that our definitions of what these MH disorders are clear for surveillance, proper resource allocation to healthcare systems, and so that MH disorders can be properly triaged and treated in the community [5].

There are important limitations to note in our descriptive analysis and multiple suggestions for future research. First, our analysis is limited to NC and has limited generalizability. Second, each ED visit is an encounter rather than an individual and repeat visits cannot be distinguished. Third, there is not a way to identify an individuals’ primary diagnosis, as the first-listed code is often related to billing rather than primary health importance. Relatedly, if there is not a billable charge associated with an ICD-10-CM code, it may be used less frequently (e.g., suicide attempt) even if relevant to the visit. Lastly, additional definitions for MH/SU visits should be considered in future work. These include the Agency for Healthcare Research and Quality’s definition for anxiety, trauma, and stress-related disorders, the Center for Medicaid Service’s definition for depression, and the National Syndromic Surveillance Program’s Electronic Surveillance System for the Early Notification of Community-based Epidemics definition for a MH disaster-related ED visit [12–14]. Future work may also incorporate case note key word searches into syndromic definitions and consider how frequency of codes used may vary based on billing priorities. Additional future directions include deciphering whether the ED visit was for a MH crisis, and the severity of that crisis, by considering whether the ED visit led to a hospital admission and consulting triage notes. Lastly, after establishing a MH case definition, there is a need to look at meaningful sub-categories (e.g., mood disorders) MH co-occurrence with OD events and, more broadly, SU disorders.

Standardizing population-level MH indicators is necessary for surveillance efforts to characterize its burden to efficiently allocate funding and inform treatment and prevention programming. Furthermore, these varied definitions have implications for not only MH prevalence, but also for the prevalence of MH/SU disorders and other co-occurring health issues. Given that prevalence depends upon the definition being used, future efforts should focus on creating revised, standardized MH, SU and combined MH/SU definitions to conduct adequate public BH surveillance.