Pathology Testing For ED Mental Health Emergencies
Overview
This study aimed to assess the utilization and impact of pathology tests in patients presenting to the emergency department (ED) with mental health-related concerns. The primary focus was to determine the proportion of patients undergoing these tests and whether the results influenced clinical management.
A retrospective analysis was conducted, reviewing all mental health presentations to a regional ED between January and June 2021. Data collected included patient demographics, the type of pathology tests performed, subsequent changes in management, associated treatment costs, and time to consultation with Emergency Physicians or Psychiatrists. Statistical analyses, including descriptive and multivariate methods, were used to identify associations between pathology test usage, ED length of stay (LOS), and changes in management.
Out of 37,900 ED visits, 1,462 (3.9%) were for mental health-related issues. Pathology tests were performed in 47.7% of these cases, but only 3% of the tests led to a change in patient management. These tests were associated with significant increases in treatment costs (45% for blood tests and 17% for urine tests) and prolonged ED stays, with an average LOS increase of 48%. Notably, patients with coexisting medical conditions were three times more likely to experience a management change based on pathology findings.
In conclusion, pathology testing should be selectively employed for high-risk patients, guided by clinical judgment. A more strategic approach to test utilization can reduce unnecessary delays and costs, ultimately improving the efficiency and quality of care for mental health presentations in the ED.
Introduction
From 2022 to 2023, approximately 8.8 million emergency department (ED) visits occurred in Australia. This growing demand for emergency care and hospital admissions has exacerbated ED overcrowding, leading to reduced access to care, compromised clinical outcomes, and increased adverse events for patients. Additionally, overcrowding has resulted in patients leaving without receiving treatment, higher risks of iatrogenic harm, increased operational costs, and decreased patient satisfaction. Addressing these challenges requires optimizing ED processes and minimizing low-value practices.
A 2020 study at our hospital revealed that patients presenting with mental health-related complaints experienced significantly longer ED lengths of stay (LOS) compared to other patient groups. This finding aligns with 2018 data from the Australasian College for Emergency Medicine. One contributing factor to the extended LOS is the dual assessment required by both ED and mental health teams for most mental health presentations. Furthermore, the absence of a clear local consensus on “medical clearance” has led to the routine ordering of multiple laboratory tests for these patients.
Current guidelines emphasize that history-taking and physical examination are the most sensitive approaches for identifying co-existing physical illnesses in patients with mental health concerns. The American College of Emergency Physicians (ACEP) advises against routine laboratory testing for patients presenting with acute psychiatric symptoms. Similarly, Australian (NSW) guidelines assert that ED clinicians should avoid unnecessary investigations unless specific symptoms warrant further evaluation. Routine urine drug screens are also discouraged, as they rarely influence clinical management or patient disposition.
Despite these recommendations, many mental health teams require completion of medical evaluations and pathology tests before initiating their assessments. This practice, though widespread, is considered of low diagnostic value in most mental health-related presentations. Research shows that laboratory testing changes medical management in only a small percentage of cases. For example, a 2022 meta-analysis found diagnostic tests altered management in just 1.1% of cases, while a 2018 systematic review noted a negligible impact on patient disposition when protocolized investigations were conducted.
Targeted laboratory testing is recommended for high-risk groups, including individuals with advanced age, substance use disorders, no prior psychiatric history, or co-existing medical conditions. While psychiatric illness is often associated with chronic medical conditions and increased mortality, screening for or managing unrelated chronic issues is generally outside the scope of ED evaluations.
Implementing evidence-based changes in practice requires overcoming several barriers, including presenting localized data to clinicians and policymakers. Conducting local research is crucial, as variations in practice and resources across jurisdictions may necessitate tailored solutions. This paper aims to assess the use of pathology tests in ED patients with mental health-related complaints, evaluate their impact on clinical management, and analyze patient trajectory within the ED. By examining these factors, we seek to identify patient characteristics that warrant targeted diagnostic interventions and inform future evidence-based practices for this population.
Method
This retrospective, cross-sectional study utilized electronic medical records (EMRs) to analyze all documented emergency department (ED) presentations related to mental health between January 1 and June 30, 2021. Conducted at a primary referral center treating 73,000 emergency patients annually, the study was approved by the Illawarra Shoalhaven Local Health District Human Research Ethics Committee (QA103).
The primary aim was to determine the proportion of mental health-related ED presentations where pathology tests were performed and whether these tests prompted a change in patient management. Changes in management were defined as medical treatment for identified pathology, consultations with other services, repeat testing, or hospital admissions under medical/surgical teams instead of mental health services. Secondary objectives included examining patient characteristics linked to management changes (e.g., age, sex, presenting complaint), evaluating ED length of stay (LOS), and assessing associated treatment costs. The completeness of medical documentation was also reviewed against standard expectations, with follow-ups for complications within seven days of discharge.
Data were sourced using the ED EMR system “FirstNet.” This system identified cases with a primary mental health-related complaint, as selected by triage nurses from a predefined list of categories (e.g., anxiety, depression, suicidal ideation, self-harm). Time intervals for various stages of care were extracted, such as time from triage to doctor evaluation and from mental health referral to consultation. Cases involving confusion or overdose were excluded due to their classification as medical emergencies requiring pathology. Additional clinical data, such as comorbidities, drug or alcohol use, current medications, and subsequent ED representations, were documented in a REDCap database by trained staff and medical students. Inter-rater reliability was assessed through double data entry for a sample of 100 cases.
The analysis, conducted using SPSS v26, employed a two-step process. First, bivariate associations between predictors and outcomes were examined using correlation, t-tests, and logistic regression. Predictors meeting a significance threshold of p < 0.02 were advanced to multivariable modeling. In the second step, linear regression was used for continuous outcomes (e.g., ED LOS, treatment costs), while logistic regression was applied to binary outcomes (e.g., whether pathology tests were performed or led to changes in management). Non-significant predictors were systematically excluded until only those meeting the p < 0.05 threshold remained.
Exploratory analyses revealed that ED LOS and treatment costs were skewed, necessitating natural log transformations for regression modeling. Residual analyses confirmed the transformation improved normality. Cohen’s kappa and percentage agreement were calculated for inter-rater reliability, ensuring data quality and consistency.
The study focused on understanding factors influencing management changes for mental health patients in the ED, including the impact of pathology testing, and provided insights into care efficiency and cost implications. By assessing documentation quality and subsequent complications, the research aimed to identify opportunities for improving mental health care in emergency settings.
Results
During the study, a total of 37,900 emergency department (ED) visits were recorded, of which 1,462 (3.9%) were related to mental health concerns. The mean age of the cohort was 35 years (SD = 18), and nearly a quarter (24.8%) of these individuals were detained under the Mental Health Act. Inter-rater reliability was evaluated for various measures, showing high agreement for pathology tests (κ = 0.94, 96.9%), urine drug screens (κ = 0.95, 92.8%), sex (κ = 0.98, 99.0%), and age (κ = 0.99, 98.0%). Moderate agreement was observed for intoxicated ATOR (κ = 0.61, 86%), pathology within seven days (κ = 0.41, 66.0%), drug use (κ = 0.51, 66.0%), alcohol use (κ = 0.60, 72.0%), and comorbidities (κ = 0.46, 58.0%). Fair agreement for changes in management (κ = 0.39, 94.0%) was influenced by a high proportion of “no change” outcomes.
Approximately 24.8% of patients (n = 348) were admitted to mental health wards, with 90.2% having documented prior mental health histories. More than half (53.9%, n = 788) were discharged from the ED, while 16.1% (n = 235) left without being seen. A small percentage (3.0%, n = 58) re-presented to the ED within seven days, but none required intervention for significant missed diagnoses, such as physical illnesses or injuries. Documentation quality in the ED was low, with key details like comorbidities missing in 17.2% of mental health ward admissions, drug use recorded in only 34.5%, alcohol use in 40.5%, and medications in 23%.
Blood pathology tests were ordered for 47.7% (n = 697) of all patients and 84.8% (n = 295) of those admitted under mental health care. Urine drug screens were performed for 31.9% (n = 467) of patients overall and 74.3% (n = 255) of those admitted. Admission to mental health wards significantly increased the likelihood of pathology testing (OR = 9.76, 95% CI = 6.85–13.92), as did intoxication at the time of review (OR = 2.72, 95% CI = 1.81–4.07). Interestingly, 10% of pathology tests were conducted within a week of prior testing. Among 21 patients (3.0%) with primary mental health presentations, pathology tests prompted changes in management, with 13 managed in mental health wards or discharged home. Patients with comorbidities were three times more likely (OR = 3.31, 95% CI = 1.03–10.62) to experience management changes due to test results.
The median cost of ED treatment for patients undergoing pathology tests and discharged home was $1,300 (IQR = $890–$1,918). Blood tests increased costs by 45%, while urine tests raised costs by 17%. Presentations involving behavioral disturbances were associated with a 25% increase in pathology costs compared to self-harm or suicide-related presentations.
The median ED length of stay (LOS) for mental health patients was 331 minutes (IQR = 199–581), 83 minutes longer than the general ED population. LOS increased to 505 minutes (IQR = 331–809) for patients requiring pathology tests, representing a 48% increase associated with blood tests, 15% with urine screens, and 26% with intoxication. Patients detained under the Mental Health Act had 30% longer LOS than those who were not detained or admitted. Disposition decision-making accounted for the largest proportion of LOS, with a median time of 176 minutes (IQR = 103–288) from triage to mental health consult and 130 minutes (IQR = 55.3–228) from referral to mental health review. The median time from mental health consultation to final disposition was 138 minutes (IQR = 59–306).
Conclusion
The retrospective analysis of 1,462 mental health-related emergency department (ED) presentations revealed that nearly half (47.7%) of these cases underwent pathology testing, yielding low diagnostic utility. The study identified a correlation between pathology testing and increased ED length of stay (LOS) as well as treatment costs, aligning with findings from other local and international research. Delays in patient throughput were often due to prolonged waits for test results, specialist consultations, and patient deterioration.
The presence of pre-existing medical conditions was the sole predictor of changes in patient management following pathology testing. To improve diagnostic yield, the study suggests limiting pathology testing to high-risk patients or those with significant findings during history-taking and physical examinations. Risk factors for medical complications included older age (>65 years), abnormal physical examination findings, agitation, altered mental state, substance use, and known or suspected medical conditions, such as immunosuppression or the absence of prior psychiatric symptoms.
The analysis also highlighted that ED LOS was significantly prolonged for mental health-related cases, partly due to delays in mental health team evaluations and disposition decisions. To address this, the authors advocate for early, collaborative patient assessments by ED and mental health teams to expedite disposition decisions. Implementing models such as mental health nurse practitioners and extended hours for mental health liaison services was identified as an effective strategy to improve patient flow and outcomes.
Additionally, the study noted deficiencies in the quality of medical documentation for mental health presentations. The authors recommend standardizing evaluation and documentation practices and addressing barriers to adopting existing evidence-based medical clearance protocols. They also emphasized that pathology testing should not delay mental health evaluations, as reinforced by recent policy updates from the NSW Government.
The study acknowledged limitations, including its retrospective design, potential underrepresentation of mental health presentations due to triage coding, and limited data coverage beyond the health district’s electronic medical record (EMR) system. There was also some subjectivity in determining the impact of pathology tests on management decisions.
In conclusion, pathology testing in mental health-related ED presentations rarely influenced patient management and contributed to extended LOS and higher costs. The authors recommend focusing pathology testing on patients with high-risk features and ensuring mental health evaluations are not delayed by awaiting test results. Concurrent evaluations by ED and mental health teams are essential to streamline patient care and reduce unnecessary costs.
