Effects of mandatory screening labs in directing the disposition of the apparently healthy psychiatric patient in the emergency department.
When the American military entered combat operations in Afghanistan and Iraq over 14 years ago, military medical departments and the American government were woefully unprepared to handle the psychological strains that would be placed on some 2.2 million troops (and their families) who were deployed over the years. A 2008 report by the RAND Corporation estimated that more than 26% of troops have returned from these combat operations with mental health issues. (2) Along with the active duty service member, family members are seeking care for their own mental health issues in record numbers. In a 2002 study, Calhoun et al (3) determined that partners of Veterans diagnosed with posttraumatic stress disorder (PTSD) experience more caregiver burden and have poorer psychological adjustment than partners of Veterans without PTSD. According to the Veteran's Administration, partners of Veterans with PTSD reported lower levels of happiness and less satisfaction in their lives with about half reporting having felt "on the verge of a nervous breakdown." (4) The psychological stress and ongoing struggle for adequate treatment for these service members and their families have led to a strain on the mental health care system at many of our military bases. Several recently published reports attest that service members interested in accessing mental health care often face long wait lists. (5) Although these wait times can vary considerably from one behavioral health clinic to another; the Department of Defense Mental Health Task Force noted that delays of 30 days for an initial mental health appointment are not uncommon. This gap in care leads many service members and their families to seek help for their mental health concerns at local EDs.
For adult psychiatric cases that present to the ED and are otherwise medically stable, mandatory screening labs add unnecessary costs, use limited resources, and extend emergency department stays without changing the final disposition of the patient. The screening labs that are required for psychiatric admission are often based on expert opinion and medical screening performed by psychiatrists, who are not trained to evaluate emergency patients for medical conditions, rather than emergency providers. (6) In a retrospective study of 212 patients, Korn et al (7) concluded that patients with a psychiatric chief complaint and normal physical exam findings may be safely referred to psychiatric services without the use of ancillary testing in the ED. In a similar study, Amin and Wang found that patients presenting to the ED with psychiatric chief complaints and benign histories and physical exams have a low likelihood of clinically significant laboratory findings. (1)
We conducted this study to determine if mandatory psychiatric admission laboratory tests yield results that change the disposition of a psychiatric patient from admission to a psychiatric service to admission to a medical service if that patient has presented with no medical complaints and no abnormal physical exam findings following examination by a trained emergency provider. We believe that if these mandatory laboratory results do not change the patient's admitting service, the use of laboratory testing for psychiatric patients should be performed only at the discretion of the evaluating emergency medicine provider, thus increasing patient throughput in the department, reducing risks to patients and staff, and decreasing costs in the healthcare system.
We conducted a single center retrospective cohort chart review study.
This study was conducted at a Level III trauma center supporting more than 42,000 military personnel and more than 145,000 family members, retirees, and civilian emergencies. In 2012, the ED evaluated 77,403 patients, with 3,686 (4.8%) of those patients being psychiatric patients. The department has a full time staff of 8 licensed social workers that evaluate all psychiatric complaints and assist the emergency clinicians with the assessment and placement of these patients. The medical center has a 15-bed inpatient psychiatric ward and a 5 bed ICU, as well as the ability to admit psychiatric patients to 3 offsite inpatient psychiatric treatment facilities.
Selection of Participants
We conducted this study using a database maintained by the medical center Department of Social Work of all emergent psychiatric patient evaluations and their dispositions. The database records a patient's name, age, military unit, chief complaint, and disposition. Using SPSS software (IBM Inc, Armonk, NY) and the general rule for multivariate analysis, we powered our study for 30 subjects enrolled per independent variable. With 9 independent variables, we determined that we would have to enroll at least 270 patients to adequately power our study. We performed a post hoc analysis to ensure adequate power. We trained an emergency medicine second-year resident to apply our specific inclusion and exclusion criteria to all patients in the database. In addition to her training, our research resident was blinded to our hypothesis and was not involved in the development of our study to avoid data collection bias. We instructed her to pull paper medical records if electronic medical records were not complete and informed her that we would perform quality checks on her data collection and recording. She spent 3 months enrolling study participants that met criteria. She began by narrowing our search to all patients aged 18-65 years who presented with a psychiatric chief complaint at the ED and were admitted to a psychiatric or medical service from the ED during the period December 1, 2011 through December 1, 2013. From this patient group, she reviewed each medical chart for any physical complaints, vital sign abnormalities, or abnormal history and physical exam findings recorded by the nurse or emergency department provider. Patients who were not strictly psychiatric patients were eliminated. After eliminating all patients with concurrent medical complaints or physical exam or vital sign abnormalities, 682 patients remained enrolled in the study. After data was collected, the primary investigator randomly reviewed 3 patient charts from each month to ensure our data collector had enrolled appropriate patients and collected accurate data.
Our primary outcome measure was the percentage of dispositions that changed from psychiatric admission to medical admission as a result of clinically significant laboratory abnormalities in the mandatory screening laboratory tests. To determine the percentage, we divided the total number of patients who underwent each mandatory laboratory test by the total number of patients whose disposition was changed to a medical admission based on an abnormality in each laboratory test. We also tabulated the total number of clinically significant laboratory abnormalities, defined as laboratory results that would require treatment per our hospital protocols, prior to disposition, but did not change the disposition of the patient. Our hypothesis was that the mandatory psychiatric admission laboratory tests do not yield results that change the disposition of a psychiatric patient from admission to a psychiatric service to admission to a medical service.
Method of Measurement and Data Collection
After finding the 682 patients that met our inclusion and exclusion criteria, the data research resident removed personal identifications from the data and recorded the laboratory results for each patient into an Excel spreadsheet. Normal laboratory results were recorded as "normal," and abnormal laboratory results were recorded as the actual laboratory value. If the research resident encountered missing electronic medical records, the paper hardcopy was pulled for review. Using this method, we had no incomplete charts or missing data points. The results of the mandatory psychiatric screening laboratory tests for the 682 study participants (complete blood count (CBC), comprehensive metabolic panel (CMP), thyroid stimulating hormone (TSH), acetaminophen (APAP), aspirin (ASA), blood alcohol level (BAL), urinalysis (UA), urine pregnancy test (UHCG), urine drug screen (UDS)) were reviewed separately by 2 independent board certified emergency physicians for abnormalities that were clinically significant enough to require treatment in the department according to the hospital psychiatric admission protocols. The reviewing physicians were blinded to any change in disposition that resulted from the laboratory tests and were not involved in the development or execution of this study in any way. This study was conducted under a protocol reviewed and approved by the Institutional Review Board and in accordance with the approved protocol.
Six hundred eighty-two subjects met our criteria and were enrolled in the study. All of the patients presented with a chief psychiatric complaint (ie, "life stress," suicidal ideation, homicidal ideation, depression, anxiety), denied concurrent physical ailments, and had normal vital signs and nursing and provider assessments. Descriptive statistics of the subjects are presented in Figures 1 and 2.
Of 5,606 laboratory tests ordered and results received for these patients (682 UA, CMP, BAL, TSH, ASA, CBC, UDS, APAP, and 150 UHCG), as shown in Table 1, there were 59 that were considered clinically significant abnormal results by our 2 reviewing emergency physicians. These were laboratory values that, according to our hospital psychiatric admission criteria, would require treatment prior to psychiatric admission.
Thirty patients had abnormal UAs and were started on treatment for asymptomatic urinary tract infections prior to admission to psychiatry. Five patients had hypokalemia on their CMP and were treated with K-Dur and admitted to psychiatry. The other 2 abnormal CMPs showed elevated creatine for which the patients received IV fluids and the repeat laboratory values returned to normal prior to psychiatric admission. The 7 patients with elevated BAL had repeat laboratory tests drawn until levels were under 100, meeting admission standards to psychiatry. Thyroxine free (free T4) levels were obtained on the 3 patients with abnormal TSH and results were normal, eliminating concern for acute thyroid abnormalities. These patients were cleared for psychiatric admission with recommendations for repeat laboratory work upon discharge. Finally, the patient with elevated ASA had serial ASA levels drawn in the emergency department to confirm a downward trend prior to psychiatric admission.
Of 682 psychiatric patients, only one received a final disposition to a medical service based upon abnormal laboratory studies, a probability of 1/682=0.1% (95% CI: 0.0% to 0.9%). That patient was a 43-year-old male who presented without any medical complaints but a chief complaint of "suicidal ideation." This patient had an abnormal CMP (elevated glucose) and UA (glucose and ketones present) and was found, through his laboratory work, to have diabetic ketoacidosis (DKA). Upon diagnosis, he received IV fluids and insulin in the emergency department and was transferred to the intensive care unit. The probability that an abnormal laboratory test will result in a change in patient disposition is 1/682=0.1% (95% CI: 0.0% to 0.9%).
Using Fisher's Exact Test, we found that there was a statistically significant difference in the change in patient disposition between an abnormal and normal CMP (P=.013) and an abnormal and normal UA (P=.047).
Because many of our laboratory tests had no abnormal results (CBC, UDS, APAP, UHCG), we decided to investigate if the probability of an abnormal finding was truly "zero" and, if not, what the estimated probability is for an abnormal result in one of those laboratory tests in the future. Using the Modified Wald Equation and the rule of three, among the tests for which there were no abnormal results, the estimated probability of an abnormal laboratory finding in the future was 0.4% to 0.7% for the CBC, UDS, and APAP, and 2.0% to 3.1% (high limit of the 95% CI) for UHCG. This means that even though our data had some laboratory tests with no abnormal results, there is still a small probability that there will be abnormal results in future patients. The analysis results are presented in Table 2.
There were also laboratory tests for which there were abnormal results but no change in disposition (ASA, BAL, and TSH), and we estimated the probability of future abnormal results in those laboratory tests as well. Using the same Modified Wald Equation, we estimate the probability of a future abnormal result in the laboratory tests that already had abnormal results was up to 1.8% (high limit of the 95% CI), presented in Table 3. This means that there is continued probability of these laboratory tests resulting abnormally, but this does not change the probability that these abnormal results will alter disposition in the future.
Due in part to the demand placed on emergency departments and the prevailing evidence that mandatory psychiatric laboratory tests do not alter a patient's disposition, the American College of Emergency Physicians (ACEP) created a clinical policy in 2005 that provided a Level B and two Level C recommendations (recommendations for patient management that may identify a particular strategy that reflect moderate clinical certainty) against the routine testing of clinically stable adults that present to the emergency department with psychiatric symptoms. 8 Despite these recommendations and the results from prior studies, clinical practice has not changed in most emergency departments which is, in part, due to poor communication between the emergency medicine providers and the accepting psychiatrists.
Our results validate the findings from prior studies, such as Korn et al, (7) Parmar et al, (6) and Amin and Wang, (1) in the military population. Our study demonstrates that these mandatory screening laboratory tests change the disposition in psychiatric patients with no physical complaint or abnormal physical exam findings less than 1% of the time (high end of 95% CI). With only one exception, none of the performed tests altered a patient's disposition. Our study confirms that extensive screening laboratory tests for medically stable psychiatric patients are very unlikely to change the patient's disposition. Medicine accepts a reasonable miss-rate order to avoid over-testing and unnecessary treatments. Many decision-making rules, such as pulmonary embolism rule-out criteria, accept the risk of missing a positive at anything below 2%. We believe that our results, which show a less than 1% probability in a change of disposition, demonstrate that forgoing mandatory screening laboratory tests in these stable psychiatric patients is allowing an acceptable level of risk. This risk is further mitigated due to these patients being admitted to a hospital psychiatric ward where they will continue to be monitored and can be transferred or reevaluated quickly if their presentation changes.
In analyzing the individual laboratory tests for clinically significant abnormal results, we found many laboratory results are often returned as "normal." However, we know that "zero abnormal results" does not mean "zero risk," so we determined the future probability of an abnormal result in those laboratory tests with no abnormal results (CBC, APAP, UDS, UHCG) in our study and found that, with the exception of the UHCG, there was less than a 1% estimated probability. This is a miss rate of an abnormal finding that we in the medical field can be comfortable with. This means that we can, essentially, stop ordering these tests in stable psychiatric patients with a less than 1% chance of missing a clinically significant abnormal result in the future. In the laboratory studies with abnormal results that did not alter a patient disposition (BAL, TSH, ASA), there is a less than 2% estimated probability of a future abnormal result. This is a miss rate that we can also be comfortable with, as none of these abnormalities ultimately change the patient disposition. This study suggests that these 6 laboratory tests can be safely eliminated from our battery of mandatory tests, as none of them altered the patient disposition and there is less than 2% estimated risk for any clinically significant abnormal results in future testing. The urine pregnancy test was only ordered on female patients and a positive result did not alter the disposition at our hospital as all inpatient psychiatric facilities are capable of treating pregnant patients. This laboratory result may direct a treatment plan for the psychiatrist but does not alter the disposition of our medically stable psychiatric patient. We suggest that the UHCG can be ordered by the psychiatrist once the patient is admitted rather than in the emergency department where it does not alter the patient course.
Again, focusing on the utility of individual laboratory tests, the only ones with abnormalities that required a change in disposition were the CMP and the UA. Both of these laboratory tests were positive for elevated glucose, and the results initiated the workup for DKA in the only patient with a disposition change. The reviewing physicians believe the CMP results were cause enough to alter the disposition, but the UA results did not, on their own, result in the disposition change. This patient was a new-onset diabetic with no previous workup or diagnosis of diabetes and he denied any physical symptoms throughout his ED course, even after he was told of his diagnosis and began emergency treatment. These 2 laboratory results were significant and initiated a formal DKA workup due to the hyperglycemia found in each. These laboratory tests may be considered redundant or may be replaced by the finger-stick glucose test, which is much less expensive and faster to rule out DKA in the stable psychiatric patient. The non-life-threatening diagnoses found by these 2 laboratory tests were mild hypokalemia and asymptomatic urinary tract infections that required treatment prior to psychiatric admission, but add to our concerns for overtreatment of laboratory results in asymptomatic patients. From our results, we believe that the CMP and UA can be replaced by the finger-stick glucose testing to screen for hyperglycemia. With the elimination of these 2 laboratory tests, the probability of a disposition change will remain less than 1% and risk continues to be mitigated as these patients are all admitted to the hospital and monitored for symptom changes.
Our results, which would suggest that we only need a finger-stick glucose test to ensure no disposition change in stable psychiatric patients, differ from those of Parmar et al (6) in which the laboratory results caught an acetaminophen overdose in a medically cleared patient. As a result, they recommended keeping the APAP level as a screen for suicidal patients.
With these results, a dialogue should begin within the hospital system, between the emergency and psychiatric departments, in the hopes of eliminating completely or reducing the number of mandatory tests required for a psychiatric admission. Because there is no standardized medical clearance for a psychiatric patient, it is important that each hospital system review the efficacy of the system they have in place. This study shows that extensive mandatory screening laboratory tests do not change the disposition of psychiatric patients presenting without medical complaint to a military hospital. In an era of fiscal limitations within the military and an emphasis on decreasing emergency department wait times, the departments of emergency medicine and psychiatry could use this study results, in conjunction with the 2005 guideline from ACEP and the results of prior studies on the matter, to streamline the process of psychiatric admissions within the military and reduce the medical screening laboratory tests required for medical clearance.
We hypothesized that a thorough history and physical exam would be sufficient to find an abnormality in a psychiatric patient without any medical complaints and that mandatory laboratory tests would not change the disposition. Our results show that, with one exception, history and physical exam alone can appropriately provide the disposition of medically stable psychiatric patients. Emergency providers and psychiatrists should work together to create appropriate testing strategies and admission criteria for psychiatric patients.
Our study site had characteristics distinctive to the military in that our study population was a mean age of 28 years and over 75% male. The population consisted of a relatively healthy active duty population and their family members. Our study site was also unique in that it had multiple inpatient psychiatric facilities that were able to readily accept our patients, including the inpatient psychiatric ward of the study hospital. This made admitting a psychiatric patient relatively less difficult. This population and psychiatric disposition resources may not reflect all other institutions and may be skewed to the military hospitals across the country, however the ability for an emergency provider to appropriately evaluate psychiatric patients for an underlying medical condition should be universal to emergency departments. We believe that our observations, with regard to mandatory testing of psychiatric patients with no physical complaint or abnormal physical exam finding, are applicable to multiple emergency department settings. We performed a retrospective cohort chart review study in which we used a second year emergency department resident to review all of the patient charts and record data over the course of 3 months. We believe that by having a single data collector performing this task over a short period of time, we decreased the possibility of multiple interpretations of the data collection process and data recording. Our data collector was blinded to our hypothesis to decrease data collection bias. She was also informed that we would perform random inspections of her data to ensure accuracy. (9) We pulled paper records of all charts that were not complete in the electronic medical records to ensure complete data collection. We also reviewed records that spanned a 2-year time period, attempting to get a better long-term picture and increase our data points.
Because this was a retrospective review and we only had written documentation of the patient encounter, we had to trust that documentation as fact. In reviewing provider and nursing notes, occasional documentation errors were found with regards to normal vital signs and laboratory findings. Abnormal laboratory results were recorded as normal and abnormal vital signs were recorded as normal. This brings to light the limitations of a retrospective study and, while our data collector did her best to fully evaluate patient documentation, we must acknowledge that due to human error, not all documentation made in patient charts may be accurate. We could see when abnormal results were recorded as normal but we could not see if abnormal physical exam findings were recorded as normal; this is always a possibility and a risk when trusting retrospective charts. Finally, in reviewing retrospective data, we could not see the order in which the patient evaluation process occurred. In our department, these mandatory laboratory tests may be ordered in triage, prior to an emergency provider's evaluation, or by the provider after the bedside evaluation. If these laboratory tests are ordered in triage and results obtained prior to the physical evaluation, those results may guide some providers to a more focused and detailed history and physical examination with the possibility of discovering abnormalities, thereby excluding that patient from our study. Ideally, to truly test the emergency provider's ability to find a medical abnormality in a stable psychiatric patient, the patient would be evaluated without the provider's knowledge of the laboratory results so as not to guide the history and exam.
Finally, in reviewing retrospective data, we could not see the order in which the patient evaluation process occurred. In our department, these mandatory laboratory tests may be ordered in triage, prior to an emergency provider's evaluation, or by the provider after the bedside evaluation. If these laboratory tests are ordered in triage and results obtained prior to the physical evaluation, those results may guide some providers to a more focused and detailed history and physical examination with the possibility of discovering abnormalities, thereby excluding that patient from our study. Ideally, to truly test the emergency provider's ability to find a medical abnormality in a stable psychiatric patient, the patient would be evaluated without the provider's knowledge of the laboratory results so as not to guide the history and exam.
A prospective validation study, ideally with a more varied patient population, is recommended to more broadly apply our findings to other practice centers. A large prospective multicenter study at military hospitals is needed to derive and validate a clinical decision-making rule that can be used across the Department of Defense and in similar populations for screening of psychiatric patients with no physical complaint.
Of 682 patients evaluated with a psychiatric chief complaint and no medical complaint of abnormal physical exam findings, one patient disposition was changed from psychiatric admission to medical admission based on laboratory results alone. Our study validates the findings of previous studies and indicates that these mandatory screening laboratory tests rarely alter a patient's disposition after evaluation by an emergency trained provider. By eliminating these laboratory tests we accept a less than 1% miss rate. We believe this is further mitigated due to these patients all being admitted to the hospital.
(1.) Amin M, Wang J. Routine laboratory testing to evaluate for medical illness in psychiatric patients in the emergency department is largely unrevealing. West J Emerg Med. 2009;10(2):97-100.
(2.) Tanielian TL, Jaycox LH, eds. Invisible wounds of war: Psychological and Cognitive Injuries, Their Consequences, and Services to Assist Recovery. Santa Monica, CA: Rand Corporation; 2008.
(3.) Calhoun PS, Beckham JC, Bosworth HB. Caregiver burden and psychological distress in partners of veterans with chronic posttraumatic stress disorder. J Trauma Stress. 2002;15(3):205-212.
(4.) Kulka RA, Schlenger WE, Fairbank JA, et al. Trauma and the Vietnam War Generation: Report of Findings from the National Vietnam Veterans Readjustment Study. New York: Brunner/Mazel Inc; 1990.
(5.) Johnson SJ, Sherman MD, Hoffman JS, et al. The Psychological Needs of U.S. Military Service Members and Their Families: A Preliminary Report. Washington, DC: American Psychological Association Presidential Task Force on Military Deployment Services for Youth, Families, and Service Members; 2007.
(6.) Parmar P, Goolsby C, Udompanyanan K, et al. Value of mandatory screening studies in emergency department patients cleared for psychiatric admission. West J Emerg Med. 2012;13(5):388-393.
(7.) Korn CS, Currier GW, Henderson SO. "Medical clearance" of psychiatric patients without medical complaints in the emergency department. J Emerg Med. 2000;18:173-176.
(8.) Lukens TW, Wolf SJ, Edlow JA, et al. Clinical Policy: Critical Issues in the Diagnosis and Management of the Adult Psychiatric Patient in the Emergency Department. Ann Emerg Med. 2006;47(1):79-84.
(9.) Worster A, Haines T. Advanced statistics: understanding medical record review (MRR) studies. Acad Emerg Med. 2004;11(2):187-192.
MAJ Karyn E. Kagel, SP, USA
CPT Meghan Smith, MC, USA
CPT Ilya Y Latyshenko, MC, USA
MAJ Christopher Mitchell, MC, USA
LTC Andrew Kagel, MC, USA
MAJ Kagel, LTC Kagel, and CPT Smith are with the Department of Emergency Medicine, Carl R. Darnall Army Medical Center, Fort Hood, Texas.
CPT Latyshenko is with the Department of Emergency Medicine, Womack Army Medical Center, Fort Bragg, North Carolina.
MAJ Mitchell is Regimental Surgeon, 3rd Cavalry Regiment, Fort Hood, Texas.
Table 1. Clinically significant abnormal laboratory tests and results that changed disposition. Laboratory Test UA CMP BAL TSH ASA Number of laboratory tests ordered 682 682 682 682 682 Number of normal laboratory tests 651 674 675 680 681 Number of clinically 31 8 7 2 1 significant abnormal laboratory tests Number of dispositions changed due to 1 1 0 0 0 laboratory results Laboratory Test CBC UDS APAP UHCG Number of laboratory tests ordered 682 682 682 150 Number of normal laboratory tests 682 682 682 150 Number of clinically 0 0 0 0 significant abnormal laboratory tests Number of dispositions changed due to 0 0 0 0 laboratory results UA indicates urinalysis; CMP, complete metabolic panel; BAL, blood alcohol level; TSH, thyroid stimulating hormone; ASA, aspirin; CBC, complete blood count; UDS, urine drug screen; APAP, acetaminophen; UHCG, urinary human chorionic gonadotropin. Table 2. Estimated probability of a future significant abnormality in laboratory tests with no abnormal results. N Abnormality Normal Probability of an Abnormality Occurring in the Future CBC 682 0 682 0.0% APAP 682 0 682 0.0% UDS 682 0 682 0.0% UHCG 150 0 150 1.3% 95% CI Rule of Three CBC 0.4% (-0.1% to 0.7%) 0.4% APAP 0.4% (-0.1% to 0.7%) 0.4% UDS 0.4% (-0.1% to 0.7%) 0.4% UHCG 1.8% (-0.5% to 3.1%) 2.0% CBC Indicates complete blood count; UDS, urine drug screen; APAP, acetaminophen; UHCG, urinary human chorionic gonadotropin. Table 3. Estimated probability of a future significant abnormality in laboratory tests with abnormal results. N Abnormality Normal Probability of an Abnormality Occurring in the Future BAL 682 7 675 1.0% (Modified Wald) TSH 682 2 680 0.3% (Modified Wald) ASA 682 1 681 0.1% (Modified Wald) 95% CI BAL 0.8% (0.3% to 1.8%) TSH 0.6% (0.0% to 1.2%) ASA 0.5% (-0.1% to 0.9%) BAL Indicates blood alcohol level; TSH, thyroid stimulating hormone; ASA, aspirin. Figure 1. Number of subjects by age group. Age Group (years) 18-25 350 26-33 204 34-41 87 42-49 34 50-57 9 58-65 6 Note: Table made from bar graph. Figure 2. Percentages of subjects by gender. Female 22% Male 78% Note: Table made from pie chart.
|Printer friendly Cite/link Email Feedback|
|Author:||Kagel, Karyn E.; Smith, Meghan; Latyshenko, Ilya Y.; Mitchell, Christopher; Kagel, Andrew|
|Publication:||U.S. Army Medical Department Journal|
|Date:||Jul 1, 2017|
|Previous Article:||Low prevalence of carbapenem-resistant enterobacteriaceae among wounded military personnel.|
|Next Article:||Resilience-enhancing relationships: what we can learn from those with a history of adverse childhood experiences.|