Screening for Depression, Anxiety, and Suicidality: Insights From Electronic Health Record Data in an Academic Medical Center

Abstract

Introduction:

In 2023, 14.6 million adult Americans had serious mental illness, including Major Depressive Disorder (MDD) and anxiety disorders. Further, 12.3 million adult Americans had serious suicidal thoughts and behaviors. This report examined mental health screening in non-psychiatry clinics to identify gaps in screening practices using electronic health record (EHR) data.

Methods:

De-identified data were accessed from the Epic EHR from 2016 to 2022. All participants with a recorded encounter with a mental health screening within 5 years of the first encounter date were included. The screening tools included Patient Health Questionnaire – 9 (PHQ-9), Generalized Anxiety Disorder – 7 (GAD-7), and Concise Health Risk Tracking Self-Report (CHRT-SR₇).

Results:

The frequency of self-report use (% of total encounters) increased from 2016 to 2022. During this period, the internal medicine department consistently screened for anxiety using the GAD-7 scale in over 35% of total encounters and for suicidality using the CHRT-SR₇ scale in over 20% of total encounters. Other non-psychiatry departments used these scales less frequently.

Conclusions:

This report highlights the increase in the use of key mental health questionnaires by non-psychiatric practitioners over time after implementation. EHR data enables the identification of gaps in screening for depression, anxiety, and suicidality across departments in a medical center.

Introduction

In 2023, 58.7 million adult Americans reported having a mental illness, including mental, behavioral, or emotional disorders, in the past year. ¹ Major Depressive Disorder (MDD), one of the most prevalent mental health disorders and one of the leading causes of disability worldwide,^2-5 is a chronic/recurrent illness affecting 8.5% of adults in the United States in 2023. ¹ MDD is characterized by a sad mood, a loss of interest or inability to enjoy or derive pleasure from activities, changes in body weight, insomnia or hypersomnia, fatigue, feeling worthless, a lack of concentration, or suicidal ideation persistently for at least 2 weeks. MDD is diagnosed if an individual reports at least five of these symptoms, with one being sad mood or loss of interest in activities (Diagnostic and Statistical Manual of Mental Disorders (DSM-5).^6,7 Major depression often co-occurs with anxiety, which is associated with severe symptoms and treatment resistance, significantly affecting quality of life.^8,9 Depression severity is associated with a progressive worsening of work productivity, ¹⁰ psychosocial functioning, ¹¹ health-related quality of life,^8,9 and life satisfaction. ¹² In addition, the prevalence of depression is higher in patients with chronic medical conditions, including chronic kidney disease (CKD; 26.5%), ¹³ diabetes (28%), Parkinson’s disease (23%), myocardial infarction (29%), stroke (18%), and cancer (16%).^14,15 Comorbid depression in patients with chronic diseases increases healthcare services utilization.^16,17

Despite the higher prevalence of depression in patients with chronic medical conditions, many depression cases remain undiagnosed and untreated. Among those screened for depression in the World Health Organization (WHO) study on global AGEing and adult health (Sage) Wave 1, 6% presented with clinical depression. However, 87% of these screen-positive individuals reported no history of diagnosis or treatment at study enrollment. ¹⁸ Other studies have observed the prevalence of undiagnosed depression in the general population.^18-21 Undiagnosed and untreated depression exacerbates existing medical conditions, including diabetes and cancer. ²² Untreated depression in individuals with medical conditions is associated with poor quality of life, disease-related complications, increased risk of functional disability, and increased risk of all-cause mortality^22-26 and untreated depression during pregnancy is associated with preterm birth, perinatal complications, and postpartum depression. ²⁷ With this growing appreciation of undetected and untreated depression and the associated adverse health outcomes, the U.S. Preventive Services Task Force (USPSTF) recommended routine screening for depression in 2009. ²⁸ In 2011, the Centers for Medicare & Medicaid Services (CMS) started to cover annual screening for depression for Medicare beneficiaries in primary care settings. The current recommendations by the USPSTF include depression and suicide risk screenings for adults, including pregnant and postpartum women and older adults over 65 years.^22,29 Despite these efforts, depression, anxiety, and suicide risk screening rates remained low in primary care settings.^30-32 Further, little is known about depression, anxiety, and suicide risk screening practices in specialty departments. Identifying gaps in screening practices in primary care settings and specialty departments will help develop strategies to improve mental health screening and service utilization.

Electronic health records (EHR) are widely used in the United States’ healthcare systems to record and track an individual’s utilization of healthcare services. The HIPAA (Health Insurance Portability and Accountability) Act of 1996 requires that medical records be readily available to patients. ³³ This resulted in the development of patient portals, providing patients with easy EHR access. These applications provide a direct and confidential channel for patients to receive and convey information to their doctors. ³⁴ EHR systems effectively improve documentation and data availability, streamline order entry, reduce medical errors, and offer reminders for recommended screenings, including depression and anxiety.^30,35 MyChart, an electronic patient portal, provides a secure online portal for patients to access their patient health records within Epic’s EHR system. Each division or department can customize the health screening surveys system for activation in the MyChart portal before a healthcare visit. Many of these screenings are used to rule out symptoms, assign assessments based on the patient population, or assign additional assessments, such as the Patient Health Questionnaire (PHQ), based on prior responses.

This report examined screening practices in non-psychiatry departments in a large academic medical center and associated health resource systems to identify where patients are being screened for depression, anxiety, and suicidal ideation. This study is the first step toward identifying opportunities to advocate for widespread depression, anxiety, and suicide risk screening in non-psychiatry clinical settings and to integrate evidence-based training in non-psychiatric clinical areas.

Methods

Data Source

The medical center’s Clinical Research Information System (utCRIS) data warehouse, modeled after the i2b2 system at the National Center for Biomedical Computing Partners HealthCare, incorporates extensive inpatient and outpatient data from the Epic Electronic Medical Record. Access to the MyChart data stored in the utCRIS data warehouse was approved by the Institutional Review Board (IRB; STU092015-01). The inclusion criteria for these analyses were: (1) adults (age ≥ 18 years old) who had engaged with the healthcare system and completed assessments on MyChart between January 1, 2016, and December 31, 2022. (2) adults who completed at least one mental health assessment within 5 years of the visit date. Data extraction included encounter dates, provider and department information, annual depression screening, results, and additional ordered and completed screening measures. Before analyses, all data were de-identified.

Clinical Measures

Three self-report measures were used in this analysis:

Patient Health Questionnaire — 9 (PHQ-9) is a self-administered measure of depression severity, with nine items corresponding to the nine symptoms of depression. ³⁶ Each item inquires about the frequency of depressive symptoms over the past 2 weeks, with scores ranging from 0 (“not at all”) to 3 (“nearly every day”). The PHQ-9 has a total score between 0 and 27.

Generalized Anxiety Disorder — 7 (GAD-7) is a 7-item self-report measure of anxiety symptoms, and a total score ≥ 8 represents probable generalized anxiety disorder. ³⁷

Concise Health Risk Tracking Self-Report (CHRT-SR₇). consists of 7 self-report measures used to detect and monitor suicidal risk. The responses range from 0 (strongly disagree) to 4 (strongly agree). ³⁸

Data Analyses

MyChart data were first segmented by encounter visit year, and the data were summarized by number of unique patients, number of unique encounters, and total number of encounters. Due to the large number of clinical departments, departments were summarized into eight categories: internal medicine, specialties, physical medicine, surgery, oncology, other clinics, family medicine, and OB/GYN. The number of encounters where patients completed either PHQ-9, GAD-7, or CHRT-SR7 was summarized by encounter year and department category.

Results

The analytical sample included MyChart data from 82 943 individuals and their unique encounters from 2016 to 2022. Individuals were predominantly female, white, and non-Hispanic (Table 1). The total number of encounters where self-report questionnaires were administered to assess depression (PHQ-9), anxiety (GAD-7), or suicidal risk (CHRT-SR₇) increased from 9798 encounters in 2016 to 106 775 encounters in 2022 (Table 1). The use (% of total encounters) of PHQ-9 as a screening tool for depression increased from 2016 to 2022 (Figure 1). The use (% of total encounters) of CHRT-SR₇ and GAD-7 increased from 2016 to 2019, decreased in 2020, and increased until 2022 (Figure 1).

Table 1.

Demographics of the Study Population.

Characteristics of the Study Population	2014	2015	2016	2017	2018	2019	2020	2021	2022
Sex
Male	2950	11 345	7757	8408	9763	10 407	11 800	12 101	11 087
Female	4981	22 473	16 414	16 220	18 011	18 709	21 967	23 366	23 159
Unknown	0	2	1	1	2	2	6	5	3
Race
White	5801	24 478	17 837	18 350	20 201	21 105	24 108	24 916	23 725
Black	944	3664	2707	2818	3444	3631	4472	4991	5150
Asian	351	1479	1045	1028	1267	1314	1561	1892	1829
American Indian	25	100	74	72	69	85	97	115	107
Hawaiian/Pacific Islander	6	28	25	17	15	12	15	17	26
More than one race	804	4071	2484	2344	2780	2971	3520	3541	3412
Ethnicity
Hispanic	565	2665	1834	1969	2416	2567	3147	3551	3469
Non-Hispanic	6931	28 914	20 998	21 360	23 851	24 945	28 747	30 153	29 064
Unknown	435	2241	1340	1300	1509	1606	1879	1768	1716
Total number of encounters	9798	68 489	49 690	52 921	62 554	67 520	88 938	105 418	106 775

Figure 1.

Total number of encounters registered in EPIC during the study window for data collection. A graphical representation of the total encounters where mental health self-reports were administered between 2016 and 2022. Abbreviations: CHRT-SR₇, Concise Health Risk Tracking Self-Report; GAD-7, generalized anxiety disorder-7; PHQ-9, Patient Health Questionnaire — 9.

PHQ-9 use (% of total encounters) increased over time (less than 5% of total encounters) in physical medicine, surgery, oncology, family medicine, OB/GYN departments, and other clinics from 2016 to 2022 (Figure 2). During this time, the internal medicine department consistently used the GAD-7 tool for screening anxiety in patients in over 35% of total encounters (Figure 3). Other departments used GAD-7 in under 20% of total encounters. The internal medicine department consistently used CHRT-SR₇ in over 30% of total encounters from 2016 to 2022. Other departments used CHRT-SR₇ in under 25% of total encounters (Figure 4).

Figure 2.

Total number of encounters that measured depression severity. The percent of total encounters where depression severity was assessed using the Patient Health Questionnaire-9 (PHQ-9) across different medical departments over time.

Figure 3.

Total number of encounters that measured anxiety symptoms. The percent of total encounters where depression severity was assessed using the generalized anxiety disorder-7 (GAD-7) self-report questionnaire across different medical departments over time.

Figure 4.

Total number of encounters that measured suicidality. The percent of total encounters where depression severity was assessed using the Concise Health Risk Tracking Self-Report (CHRT-SR₇) across different medical departments over time.

Discussion

The bidirectional associations between depression and physical health are well-documented.^39,40 Comorbid depression in patients with chronic medical conditions is associated with poor quality of life, poor health outcomes, ⁴¹ increased healthcare utilization, ¹⁷ increased risk of premature death,^42,43 and increased economic burden. ² Thus, depression screening as a routine practice in non-psychiatry departments will help identify patients most at risk of poor health outcomes. Between 2016 and 2022, screening rates for depression using PHQ-9 in non-psychiatric clinics increased in frequency of use across total encounters in the system. The first step in depression screening in clinical practice was administering PHQ-2, which assesses anhedonia and the frequency of depressed mood. PHQ-9 is then administered to those patients with PHQ-2 scores equal to or over 3. The frequency of PHQ-9 administration across departments was relatively low. In contrast, anxiety screening using the GAD-7 questionnaire and suicidality screening using CHRT-SR₇ were much higher. Notably, screening rates were over 35% of total encounters in the Department of Internal Medicine and remained stable and consistently under 20% of all encounters in other departments. Overall, the systematic introduction and clinical availability of these self-report questionnaires increased mental health screening across the system.

To understand how these data fit into the broader clinical context, we reviewed previous reports on using EHR data in clinical systems with a longer history of collecting comparable data. EHR data has been used to determine depression prevalence, ⁴⁴ depression and suicidality screening rates in primary care settings,^45,46 suicide risk rates, 1-year follow-up care in a large pediatric primary care system, ⁴⁷ depression and medical comorbidities, ⁴⁸ and study medical comorbidity and healthcare utilization in patients with treatment-resistant depression. ¹⁶ EHR data integrated with molecular-genetic tools have been used to study polygenic liability to depression in a large population^49,50 and duration of buprenorphine treatment ⁵¹ previously. Medication prescription patterns for anxiety and depressive symptoms in pediatric patients ⁵² and antidepressant use^53,54 were monitored using EHR systems previously. EHR data has also been used to discern depression subphenotypes based on medical comorbidities and medication use,^55,56 and predict postpartum depression.^57,58 Combined, these studies highlight the utility of EHR data in evaluating mental health screening practices.

Mental health screening questionnaires are typically administered during the precheck process for medical appointments. In the case of a particular patient who has been screened using these tools, their EHR data can be used to follow up with patients during the appointments or after the first encounter to establish continuity of care across system-wide encounters. Patient access to their MyChart health account also allows them to track their activities and could lead to more routine remote screening sessions that could be integrated into the standard of care. Thus, incorporating mental health screening in primary and specialty care settings will help identify those requiring mental health services, improve health outcomes, and reduce the burden of mood disorders.

Several limitations exist in the current analyses using EHR data. The EHR data extraction did not include PHQ-2 data, which limited our understanding of the lower frequency of PHQ-9 use across the departments, particularly in internal medicine and OB/GYN departments. As the PHQ-2 is administered as the first step in depression screening in clinical practice, recording this information in EHR systems would help identify gaps in screening practices. Additionally, clinics may have administered these measures in paper format without recording the data in the patients’ EHRs (especially if the patients screened negative), thus introducing data collection bias.

Our current analysis using MyChart data enables the identification of gaps in screening for depression, anxiety, and suicidal ideation across departments in a medical center. It provides an opportunity to introduce education services, to increase screening, and to provide clinical guidelines for treatment. Implementation of routine screening tests for depression, anxiety, and suicidal ideation in non-psychiatry departments is vital in standardizing mental healthcare across specialties. In large healthcare systems, clinician training in the use of mental health scales and innovative strategies is an excellent strategy to reach individuals who may not directly seek out care for mental health issues.