Individuals with psychiatric illnesses represent a significant

Transcription

1 CLINICAL INVESTIGATIONS Emergency Medicine and Psychiatry Agreement on Diagnosis and Disposition of Emergency Department Patients With Behavioral Emergencies Amy M. Douglass, John Luo, MD and Larry J. Baraff, MD Abstract Objectives: The objective was to determine the level of agreement between emergency physicians (EPs) and consulting psychiatrists in their diagnosis and disposition of emergency department (ED) patients with behavioral emergencies. Methods: The authors conducted a prospective study at a university teaching hospital ED with an annual census of approximately 45,000 patients. During study hours, each time a psychiatric consultation was requested, the emergency medicine (EM) and consulting psychiatry residents were asked to fill out similar short questionnaires concerning their diagnoses and disposition decisions after they consulted with their attending physicians. EM and psychiatry residents were blinded to the other s assessment of the patient. Residents were asked about their evaluation of patients regarding: 1) psychiatric assessments, 2) if the patients presented a danger to themselves or others or were gravely disabled, and 3) the need for emergency psychiatric hospitalization. Results: A total of 408 resident physician pairs were enrolled in the study. Patients ranged in age from 5 to 92 years, with a median age of 31 years; 50% were female. The most common psychiatric assessments, as evaluated by either EPs, consulting psychiatrists, or both, were mood disorder (66%), suicidality (57%), drug alcohol abuse (26%), and psychosis (25%). Seventy-three percent were admitted for acute psychiatric hospitalization. Agreement between EPs and psychiatrists was 67% for presence of mood disorder, 82% for suicidality, 82% for drug alcohol abuse, 85% for psychosis, and 85% for grave disability. There was 67% agreement regarding patient eligibility for involuntary psychiatric hold. EPs felt confident enough to make disposition decisions 87% of the time; for these patients there was 76% agreement with consulting psychiatrists about the final disposition decision. Conclusions: The 67% agreement between EPs and consulting psychiatrists regarding need for involuntary hold, and 76% agreement regarding final disposition, demonstrate a substantial disagreement between EPs and psychiatrists regarding management and disposition of ED patients with psychiatric complaints. Further studies with patient follow-up are needed to determine the accuracy of the ED assessments by both EPs and consulting psychiatrists. ACADEMIC EMERGENCY MEDICINE 2011; 18: ª 2011 by the Society for Academic Emergency Medicine From the UCLA Emergency Medicine Center (AMD, LJB) and the Department of Psychiatry (JL), David Geffen School of Medicine at UCLA, Los Angeles, CA. Received March 6, 2010; revisions received July 26 and September 8, 2010; accepted September 28, The authors have no relevant financial information or potential conflicts of interest to disclose. Supervising Editor: James R. Miner, MD. Address for correspondence: Larry J. Baraff, MD; LBaraff@mednet.ucla.edu. Reprints will not be available. Individuals with psychiatric illnesses represent a significant and increasing proportion of emergency department (ED) patients. 1 Unfortunately, there is little emergency medicine (EM) research regarding ED management of psychiatric patients. The majority of the literature focuses on the frequency of various behavioral emergencies in ED patients and improved ways to differentiate physical illness from psychiatric disease. 2 5 Since emergency physicians (EPs) are often the only health care providers available to manage patients who present ISSN ª 2011 by the Society for Academic Emergency Medicine 368 PII ISSN doi: /j x

2 ACAD EMERG MED April 2011, Vol. 18, No to the ED with behavioral emergencies, it is important that they be competent to evaluate patients with behavioral emergencies and determine the necessity for psychiatric consultation and for inpatient management. We were only able to identify two previous studies, both published over 10 years ago, that examined the agreement of EPs with consulting psychiatrists assessments of ED patients with psychiatric complaints. 6,7 Garbrick et al. 6 conducted a study at a California county teaching hospital in 1996, and Tse et al. 7 conducted a study in a large acute care facility in Hong Kong in ,7 The patient demographics (i.e., age and sex) in both studies were similar to those in this report. However, the current study has twice as many patients as these prior two and is the first to address agreement between EPs and consulting psychiatrists regarding the diagnosis and disposition of ED patients in a large, U.S. teaching hospital. In many EDs, including our own, the management protocol for most patients with behavioral emergencies is for the EP to perform a medical screening exam and then obtain a psychiatric or mental health consultation. This approach often results in patients with behavioral emergencies having long waits in the ED for psychiatric consultation, which worsens ED overcrowding. If EPs were able to identify those patients with less serious behavioral emergencies who do not need psychiatric consultation in the ED, these patients might be better served by referral to outpatient mental health services. Furthermore, many EDs do not have psychiatric consultation available, making it imperative that EPs know how to manage these patients. 8,9 The purpose of this study was to determine the level of agreement between EPs and consulting psychiatrists in their diagnosis and disposition decisions for patients with behavioral emergencies. METHODS Study Design We conducted a prospective study to determine the level of agreement between EPs and consulting psychiatrists in their diagnoses and disposition decisions of ED patients with behavioral emergencies. The study was approved by the UCLA institutional review board. Written informed consent was obtained from all participating EM and psychiatry residents. Study Setting and Population The study was performed at the Ronald Reagan UCLA ED, a comprehensive Level I trauma center that serves approximately 45,000 patients per year. The Ronald Reagan UCLA Medical Center is a teaching hospital that is affiliated with the David Geffen School of Medicine at UCLA. Consulting psychiatry physicians were associated with the UCLA Department of Psychiatry. Patients who presented to the ED and received psychiatric consultations while EM research assistants were present in the 12-month period between June 15, 2009, and June 18, 2010, were eligible for inclusion in the study. EM and psychiatry residents who evaluated patients receiving psychiatric consultations were asked to participate in the study by completing questionnaires regarding their assessments of the patients. Study Protocol Emergency medicine research assistants were available to present questionnaires to residents from 8AM to midnight, 7 days a week. Each time a psychiatric consultation was requested for a patient in the ED, the EM and consulting psychiatry residents were asked by a research assistant to fill out similar short questionnaires regarding their diagnoses and disposition decisions after they consulted with their attending physicians. Questionnaires were identical except that EM residents were asked how sure they were of their disposition decisions, while psychiatry residents were asked the reasons for their disposition decisions. EM and psychiatry residents were blinded to the other s assessment of the patient prior to completing the questionnaires. Completion of the questionnaires was optional and was not used to evaluate resident performance. The physicians and patients remained anonymous with unique research codes to identify the particular patient. Emergency medicine and psychiatry residents were asked about their evaluation of patients regarding: 1) psychiatric assessments (mood disorder, suicidality, drug alcohol abuse, psychotic disorder, and grave disability); 2) eligibility for an involuntary 72-hour psychiatric hold for patients who were a danger to themselves or others or were gravely disabled; and 3) the need for emergency psychiatric hospitalization. Psychiatry residents who decided to admit patients for psychiatric hospitalization were also asked to provide their chief reasons for admission. Reasons for admission included: patient placed on involuntary hold, needs acute psychiatric treatment, needs acute psychiatric evaluation, has no outpatient follow-up, has no social support, and other. After the EM and consulting psychiatry resident each filled out a questionnaire regarding their evaluation of the psychiatric patient, a research assistant completed a data form, which included information about the patient s age, sex, ethnicity, insurance status, and final disposition decision (i.e., whether the patient was admitted or discharged or transferred to another facility). Data Analysis Research assistants entered data collected from the questionnaires into an Access database (Microsoft Corp., Redmond, WA). Questions that were left blank were noted as missing responses and were not counted as negative responses. Data were exported from the Access database to Excel (Microsoft Corp., Redmond, WA). For continuous variables, including wait times for psychiatric consultations and disposition decisions, means (± standard deviations [SD]) were determined using Excel. Agreement between EM and psychiatry residents regarding psychiatric diagnosis, eligibility for involuntary hold, and disposition decision were calculated using kappa values. Kappa values were calculated using GraphPad software (GraphPad, San Diego, CA; Kappa values were chosen to measure agreement over methods that are designed to express performance or correctness (i.e., sensitivity specificity) because neither EPs nor psychiatrists opinions were considered to be the criterion standard. Previous studies assessing

3 370 Douglass et al. EM VS. PSYCHIATRY AGREEMENT ON DIAGNOSIS AND DISPOSITION agreement between equally weighted physicians diagnoses also used kappa values to evaluate agreement. 6,10 RESULTS During the study period from June 15, 2009, to June 18, 2010, there were a total of 989 patients who presented with behavioral emergencies for whom psychiatric consultations were requested during the hours that research assistants were available to collect questionnaires. Questionnaires were completed by both EM and psychiatry residents for 408 patients. Psychiatry and EM residents varied in stages of training, but regardless of experience, residents were asked to fill out questionnaires after consulting with their attending physicians. Fifty percent of the patients were female. Patients ranged in age from 5 to 92 years, with a median age of 31 years and a mean (±SD) age of 37 (±21) years. Thirteen percent of patients were at least 65 years old. Sixteen percent of patients were under 18 years of age. Child psychiatrists were available for consultation to evaluate pediatric patients. The most common psychiatric diagnoses (as indicated by the EM residents, psychiatry residents, or both) were mood disorder (66%), drug alcohol abuse (26%), and psychosis (25%). A total of 259 of 362 patients (72%) were considered eligible for an involuntary hold, with 232 of 408 (57%) assessed as suicidal and 81 of 408 (20%) assessed as gravely disabled by EM residents, psychiatry residents, or both. Only 362 of the 408 sets of physicians responded to the question regarding eligibility for involuntary hold because it was optional based on applicability to the patient s condition. Of the 408 patients receiving psychiatric consultations, 296 (73%) were admitted for acute psychiatric hospitalization. The most common reasons for psychiatric hospitalization were placement on an involuntary hold (154 of 296, 52%) and need for acute psychiatric treatment (100 of 296, 34%). Agreement between EPs and consulting psychiatrists regarding patient diagnoses was variable (Table 1). There was fair agreement about presence of mood disorder (67% [n = ], j = 0.34, Table 2). Agreement regarding suicidality was good (82% [n = ], j = 0.64, Table 3). Of the 18% that disagreed, in 43 of 73 cases (59%) the EP thought the patient was suicidal, while the consulting psychiatrist did not. Agreement regarding psychosis was moderate (85% [n = ], j = 0.46, Table 4). Agreement regarding drug alcohol Table 1 Summary of Prevalence and Agreement Regarding Diagnoses Diagnosis Assessment Number of Patients* (%) Agreement, Kappa Mood disorder 270 (66) 273 (67) 0.34 Suicidality 232 (57) 335 (82) 0.64 Drug and alcohol abuse 108 (26) 335 (82) 0.38 Psychosis 102 (25) 345 (85) 0.46 Gravely disabled 81 (20) 347 (85) 0.31 n = 408. *As indicated by EM or psychiatry or both. Table 2 Agreement About Mood Disorder EM Assessment Mood Disorder, No Mood Disorder, Mood disorder 135 (33) 68 (17) No mood disorder 67 (16) 138 (34) n = 408; total agreement, 67%; total disagreement, 33%. j = 0.34 (95% CI = 0.25 to 0.43). Table 3 Agreement About Suicidality EM Assessment Suicidal, Not Suicidal, Suicidal 159 (39) 43 (11) Not suicidal 30 (7) 176 (43) n = 408; total agreement, 82%; total disagreement, 18%. j = 0.64 (95% CI = 0.57 to 0.72). Table 4 Agreement About Drug and Alcohol Abuse EM Assessment Drug Alcohol abuse, No Drug Alcohol Abuse, Drug alcohol abuse 35 (9) 27 (7) No Drug alcohol abuse 46 (11) 300 (74) n = 408; total agreement, 82%; total disagreement, 18%. j = 0.38 (95% CI = 0.27 to 0.50). abuse was fair (82% [n = ], j = 0.38, Table 5), as was agreement regarding grave disability (85% [n = ], j = 0.31, Table 6) and patient eligibility for involuntary hold (67% [n = ], j = 0.34, Table 7). In cases where the EPs and consulting psychiatrists disagreed, 59 of 118 (50%) of the time, the consulting psychiatrist thought the patient met the criteria for an involuntary psychiatric hold, while the EP did not. Emergency physicians were asked to indicate whether they thought patients should be admitted and their degree of certainty by checking definitely, probably, unsure, probably not, or definitely not. For 354 of 408 cases (87%), EPs felt confident enough to make disposition decisions, while 13% checked unsure. Of the EPs who felt confident enough to make disposition decisions, there was fair agreement with consulting psychiatrists about the final disposition decision, i.e., whether to admit patients to psychiatry or discharge (76% [269 out of 354], j = 0.36, Table 8). However, agreement varied depending on EP response and was higher when EPs thought the patient

4 ACAD EMERG MED April 2011, Vol. 18, No Table 5 Agreement About Psychosis EM Assessment Psychotic, Not Psychotic, Psychotic 39 (10) 21 (5) Not psychotic 42 (10) 306 (75) n = 408. Total agreement, 85%; total disagreement, 15%. j = 0.46 (95% CI = 0.35 to 0.57). Table 6 Agreement About Grave Disability EM Assessment Grave Disability, No Grave Disability, Grave disability 20 (5) 20 (5) No grave disability 41 (10) 327 (80) n = 408. Total agreement, 85%; total disagreement, 15%. j = 0.31 (95% CI = 0.19 to 0.44). Table 7 Comparison of Involuntary Hold Eligibility Assessment EM Involuntary Hold Decision Psychiatry Involuntary Hold Decision Yes, No, Yes 141 (39) 59 (16) No 59 (16) 103 (28) n = 362. Total agreement, 67%; total disagreement, 33%. j = 0.34 (95% CI = 0.24 to 0.44). Table 8 Comparison of Disposition Decisions EM Admission Decision Psychiatry Disposition Decision Admit, Discharge, Agreement, % Definitely 82 (20) 11 (3) 88 Probably 141 (35) 34 (8) 81 Unsure 33 (8) 21 (5) Probably not 34 (8) 34 (8) 50 Definitely not 6 (1) 12 (3) 67 n =408. ED physicians confident enough to make disposition decision, 87%. Total agreement, 76%; total disagreement, 24%; j = 0.36 (95% CI = 0.25 to 0.47). needed admission. When the EP checked definitely there was 88% (82 of 93) agreement; when EP checked probably there was 81% (141 of 175) agreement; when the EP checked probably not there was 50% (34 of 68) agreement; and when the EP checked definitely not, there was 67% (12 of 18) agreement (Table 8). Table 9 presents the psychiatrists primary reasons for admitting each patient and whether EPs wanted to admit or discharge or were unsure for each reason. Over 80% of patients were admitted for one of two reasons: placement on involuntary hold (154 of 296, 52%) or needing acute treatment (100 of 296, 34%). Agreement between EPs and consulting psychiatrists was higher when the reason for admission was involuntary hold (123 of 140, 88%) rather than needing acute treatment (74 of 89, 83%). The former reason for admission is more serious and would be more likely to result in detrimental effects if the patient was discharged. DISCUSSION Agreement between EPs and consulting psychiatrists regarding patient diagnoses ranged from fair to good (j = 0.31 to 0.64). Agreement regarding patient eligibility for involuntary psychiatric hold was fair (j = 0.34). In 16% of cases, the EP felt that the patient did not require involuntary psychiatric hold while the psychiatrist did. Agreement regarding final disposition decision was fair (j = 0.36). In 11% of cases where EPs felt confident enough to make disposition decisions, the EP felt that the patient could be discharged, while the psychiatrist felt that the patient needed admission. These results indicate that it is possible that patients who were felt to require hold or admission by a psychiatrist could have been discharged if a psychiatry consultation had not been available. A higher proportion of patients were admitted for acute psychiatric hospitalization (73%) in our study than in the study by either Garbrick et al. (56%) 6 or Tse et al. (59%). 7 We found that the most common reasons for admission were placement on involuntary hold and need for acute psychiatric treatment. The other studies did not list reasons for admission. Comparison of the studies chief diagnoses is difficult because each used different categories of psychiatric illness. Garbrick et al. 6 did not report diagnoses, but the most common reasons for presentation were disruptive behavior (28%), overdose (24%), danger to self (23%), selfinflicted trauma (9%), other (9%), domestic violence (3%), and hallucination (3%). The most common diagnoses reported by Tse et al. 7 were schizophrenia (41%), depression (15%), and adjustment disorder (10%). We found higher agreement between EPs and consulting psychiatrists regarding diagnosis and similar agreement regarding disposition of psychiatric patients in the ED than these prior two studies. Garbrick et al. 6 did not report agreement on diagnoses, and Tse et al. 7 reported percent misclassification of diagnoses by EPs, ranging from 17% (schizophrenia) to 100% (somatization disorder), with an average kappa value of We found fair agreement between emergency and consulting psychiatry physicians regarding final disposition decision. Garbrick et al. 6 only reported agreement when the consulting psychiatrist admitted the patients, and Tse et al. 7 did not report agreement regarding disposition decisions. Garbrick et al. 6 found that when psychiatrists admitted patients, EPs agreed 76% of the time.

5 372 Douglass et al. EM VS. PSYCHIATRY AGREEMENT ON DIAGNOSIS AND DISPOSITION Table 9 Psychiatry s Chief Reasons for Admission Versus EM Disposition Decisions Admit EM Disposition Decision Discharge Unsure Agreement on Disposition Psychiatry s Chief Reason for Admission Placed on involuntary hold (is danger to self, % 154 others, gravely disabled) Voluntary admission for acute treatment % 100 (medication change, intensive program, etc.) Voluntary admission for acute psychiatric evaluation % 9 (diagnostic clarification) Voluntary admission for no outpatient follow-up available % 10 Voluntary admission for no social support % 3 Other % 3 Unknown (no response provided) % 17 Total We found that agreement between EPs and psychiatrists was greater when EPs thought patients needed to be admitted than when they thought patients could be discharged, as psychiatrists were more likely to think admission was warranted. We hypothesize that this may have been due to their consideration of additional factors, including social and situational issues, discussions with the patients referring psychiatrist or psychologist, admission for medication adjustment, etc. Although we found slightly higher agreement between EPs and consulting psychiatrists regarding assessment of psychiatric patients in the ED than previous studies, the level of agreement suggests that for some patients, whom EPs might choose to discharge, psychiatric consultation is still warranted. For example, in 7% of patients, the psychiatry resident felt that the patient was suicidal, while the emergency resident did not; for 16% of patients whom the psychiatric consultant placed on involuntary hold, the EP indicated he or she would not. Psychiatric consultants have access to prior psychiatric medical records, which are generally not available to EPs. We hypothesize that the psychiatrists access to prior psychiatric records may account for some of the differences in assessment and disposition decisions. Because EPs are often the only health care providers available to manage patients who present to the ED with behavioral emergencies, it is important that they be competent to determine the necessity for psychiatric consultation. 8 A recent survey of 243 EDs across California found that only 51% have mental health specialists on call. 9 The study also found that 40% of ED directors report that it has become increasingly difficult to transfer psychiatric patients to inpatient psychiatric facilities. A similar study surveying 233 California ED directors found that 73% were not able to admit patients to inpatient psychiatry services in their hospitals, and when transfer was necessary, the average transfer wait time was 7 hours. 10 This study also found that 23% of EDs report discharging patients with suicidal ideation without evaluation by mental health specialists. Because EPs are often responsible for evaluating and managing patients with behavioral emergencies without the aid of psychiatry consultation, it is crucial that they receive adequate training in the management of patients with these conditions. However, a survey of U.S. and Canadian EM residency program directors found that 76% of these programs offer no formal training in the management of psychiatric patients. 11 Only 14% of programs report offering a 1-month psychiatry rotation. The results of our study, in conjunction with the growing lack of availability of mental health care specialists in EDs, and the increasing numbers of ED patients with behavioral emergencies, suggest that additional studies are needed to determine whether more training of EPs in the assessment and disposition of these patients is warranted. LIMITATIONS Our study addressed agreement between EPs and consulting psychiatrists regarding assessment and disposition of ED patients. However, we were not able to assess the accuracy of their assessments and disposition decisions, as there was no criterion standard and no follow-up of patients. When psychiatric consultation is obtained it is generally the psychiatric assessment that determines the diagnosis and disposition decisions. However, psychiatrists assessments and disposition decisions are not an acceptable criterion standard, since previous studies have demonstrated that agreement among psychiatrists regarding assessment of patients in the ED is often poor. 12 This lack of a criterion standard suggests the need for additional studies with patient outcome data to assess the accuracy of these assessments. Although we found substantial disagreement between EPs and consulting psychiatrists regarding need for admission, it is possible that this disagreement may be falsely high, as some admitted patients may not have needed emergency psychiatric hospitalization. A poststudy interview with a director of psychiatric EM revealed that at our institution, psychiatric patients are often hospitalized for reasons other than medical necessity, including social factors and the requests of referring psychiatrists. Another limitation was the small scope of our study. Our sample size was only 408 patients. The study was conducted at a single urban, academic teaching hospital

6 ACAD EMERG MED April 2011, Vol. 18, No that serves a subset of the population and thus may lack external validity. Results would likely differ if the study were conducted at a community hospital. A large portion of potential patients were not included in the study because in approximately half of the cases we were unable to obtain completed questionnaires from the psychiatry or the EM resident. It cannot be determined why questionnaires were not completed. Most commonly, we could not obtain completed questionnaires from the consulting psychiatrists. Consulting psychiatrists were in the ED for a short period of time after evaluating patients and often spoke with their attending physician outside of the ED. Research assistants attempted to contact all consulting psychiatry residents after they evaluated patients; however, they were often unable to contact the psychiatrists after they completed their consultation and before they left the ED. In a smaller percentage of cases, we did not obtain completed questionnaires from the psychiatry or EM residents because participation in the study was optional. As we did not collect patient identifying information, and because we did not have institutional review board approval to review patient records, we are not able to compare enrolled and nonenrolled patients. CONCLUSIONS Agreement between EM and psychiatry residents regarding diagnosis ranged from 67% to 85%, was 67% regarding need for involuntary hold, and was 76% for patient disposition. If EPs were to manage patients without the aid of psychiatric consultation, our results suggest that some patients who were felt to require involuntary hold or admission by psychiatrists would have been discharged. Additional studies with patient follow-up are needed to determine the accuracy of EP diagnosis and disposition decisions for ED patients presenting with psychiatric complaints. The authors thank Guy Merchant, Elizabeth Lancaster, and the UCLA Emergency Medicine Research Associates for distributing questionnaires and compiling data. Without them, this study would not have been possible. References 1. Larkin GL, Claassen CA, Emond JA, et al. Trends in U.S. emergency department visits for mental health conditions, 1992 to Psychiatry Serv. 2005; 56: Pandya A, Larkin GL, Randles R, Beautrais AL, Smith RP. Epidemiological trends in psychosisrelated emergency department visits in the United States, Schizophr Res. 2009; 110: Hazlett SB, McCarthy ML, Londner MS, Onyike CU. Epidemiology of adult psychiatric visits to US emergency departments. Acad Emerg Med. 2004; 11: Sood TR, Mcstay CM. Evaluation of the psychiatric patient. Emerg Med Clin North Am. 2009; 27: Sankaranarayanan J, Puumala SE. Epidemiology and characteristics of emergency departments visits by US adults with psychiatric disorder and antipsychotic mention from 2000 to Curr Med Res Opin. 2007; 23: Garbrick L, Levitt MA, Barrett M, Graham L. Agreement between emergency physicians and psychiatrists regarding admission decisions. Acad Emerg Med. 1996; 3: Tse SK, Wong TW, Lau CC, Yeung WS, Tang WN. How good are accident and emergency doctors in the evaluation of psychiatric patients? Eur J Emerg Med. 1999; 6: Baraff LJ. A mental health crisis in emergency care. Behav Healthcare. 2006; 26: Menchine MD, Baraff LJ. On call physician specialist availability and higher level of care transfers in California emergency departments. Acad Emerg Med. 2008; 15: Baraff LJ, Janowicz J, Asarnow JR. Survey of California emergency departments about practices for management of suicidal patients and resources available for their care. Ann Emerg Med. 2006; 48: Santucci KA, Sather J, Baker MD. Emergency medicine training programs educational requirements in the management of psychiatric emergencies: current prospective. Pediatr Emerg Care. 2003; 19: Way BB, Allen MH, Mumpower JL, Stewart TR, Banks SM. Interrater agreement among psychiatrists in psychiatric emergency assessments. Am J Psychiatry. 1998; 155: