Inappropriate Use of an Accident and Emergency Department ...

ADMINISTRATION/ORIGINAL CONTRIBUTION

Inappropriate Use of an Accident and Emergency Department: Magnitude, Associated Factors, and Reasons—An Approach With Explicit Criteria

From the Servicio de Urgencias, Hospital General Universitario, Elche*; Fundación Instituto de Investigación en Servicios de Salud, València‡; and Servicio de Urgencias, Hospital Morales Meseguer, Murcia,§ Spain. Received for publication April 10, 2000. Revision received September 13, 2000. Accepted for publication October 24, 2000. Presented in part at the XIX Meeting of Health Economics, Zaragoza, Spain, June 1999. Address for reprints: Salvador Peiró, MD, PhD, Fundación Instituto de Investigación en Servicios de Salud, Pl. María Beneyto 2, 10–46008 València, Spain; +34-96-3920574; E-mail [email protected]. Copyright © 2001 by the American College of Emergency Physicians. 0196-0644/2001/$35.00 + 0 47/1/113464 doi:10.1067/mem.2001.113464

Teresa Sempere-Selva, MD, PhD* Salvador Peiró, MD, PhD‡ Pilar Sendra-Pina, MD* Consuelo Martínez-Espín, MD§ Inmaculada López-Aguilera, MD*

See related article, p. 580, and editorial, p. 629. Background: We evaluate the appropriateness of medical visits to the accident and emergency department (A&ED) of a university hospital using an instrument based on explicit and objective criteria, analyze the association between inappropriate visits and certain factors, and identify reasons for inappropriate use. Methods: This concurrent review of a random sample of 2,980 adult medical patients’ visits to the A&ED of the hospital of Elche uses the Hospital Urgencies Appropriateness Protocol, an instrument based on explicit criteria. We analyze the association between inappropriate use and specific factors, and provide a descriptive analysis of reasons for inappropriate use assigned by A&ED staff. Results: Of the total number, 882 (29.6%) of the visits were evaluated as inappropriate. Inappropriate use was associated with younger patients, use of own means of transportation, referral by the hospital, certain months of the year, and certain diagnostic groups of lesser severity. The most frequent reasons for inappropriate use were the patients’ greater trust in the hospital than primary care (451 [51.1%]), inappropriate use of services by patients (160 [18.1%]), and inappropriate referrals by primary care physicians (142 [16.1%]). Conclusion: Inappropriate use represents an important percentage of use of the A&ED. Many reasons contribute to it, although foremost among them is patient preference (and the convenience and accessibility) of these services compared with primary care. [Sempere-Selva T, Peiró S, Sendra-Pina P, Martínez-Espín C, López-Aguilera I. Inappropriate use of an accident and emergency department: magnitude, associated factors, and reasons—an approach with explicit criteria. Ann Emerg Med. June 2001;37:568-579.]

5 6 8

ANNALS OF EMERGENCY MEDICINE

37:6 JUNE 2001

INAPPROPRIATE USE OF THE EMERGENCY DEPARTMENT Sempere-Selva et al

INTRODUCTION

M AT E R I A L S A N D M E T H O D S

For years, not only the health care community but society at large have been concerned about overuse of hospital accident and emergency departments (A&EDs). Hospital A&EDs are being used with increasing frequency (in Spain visits increased from 9.2 million in 1984 to 15.3 million in 19941), and a large portion of this increase is attributed to inappropriate or nonurgent visits. Reviews of more than a hundred studies conducted in different countries,2-5 including Spain,5 place the volume of inappropriate use of A&EDs between 20% and 80% of total visits. Thus, A&EDs currently handle many cases that could be alternatively attended by primary care. Although there is no consensus about the true impact of this situation, a series of problems often are considered to be associated with overuse of A&EDs: (1) lack of continuity or follow-up of treatment of patients who replace primary care visits with visits to the A&ED, (2) diversion of necessary resources from life-threatening situations, (3) adverse effect of the work overload in the A&ED on hospital organization, and (4) the supposedly higher costs involved in A&ED care than in primary care. Most studies of inappropriate use of A&EDs have examined practices in a single hospital and have used physicians’ subjective criteria to define the “urgent” nature of a case (at times supported by guidelines based on generic criteria). Studies in the Spanish setting6-22 indicate that the percentage of inappropriate visits ranges from 25%, when appropriateness is established on the criterion of administering diagnostic tests, up to 80%, when only cases subsequently admitted to the hospital were considered true emergencies. Apart from local factors and differences in the populations studied, this discrepancy is primarily the result of the lack of consensus concerning the definition of “emergency” and, even more relevant, the absence of operational criteria to decide adequacy of care for each situation.23-27 Without such tools, each situation must be evaluated subjectively, with or without guidelines to aid decisionmaking. Because of these factors, it is impossible to compare one hospital with another or monitor the performance of any given hospital over time. This study was undertaken to evaluate the appropriateness of medical visits to the A&ED of a university hospital, using an instrument based on explicit criteria. We also sought to analyze the association between inappropriate use and specific factors of the visit, and identify the reasons motivating misuse of the A&ED.

This crossover-designed study describes the use of the A&ED of a university hospital through the review of a random sample of 2,980 visits. Research was conducted in the A&ED of a 416-bed public university hospital that covers the health care needs of the 230,000 inhabitants of a Health District in the Valencia Region (Spain). The hospital belongs to the Valencia Regional Government health care system, a network with more than 8,000 beds (approximately 80% of beds available in the region) that is responsible for caring for the 3.9 million inhabitants of the region. Among the system’s features are universal coverage, the cost-free status of care, and hospital and primary care funding by a general budget. Physicians are compensated by salary and have semi–civil servant status. The network is organized in Health Zones (10,000 to 30,000 inhabitants cared for by 1 primary care team) that are grouped in Health Districts (100,000 to 250,000 inhabitants), which are served by 1 general hospital. During working hours (all day in rural zones), emergencies are treated by the primary care teams, and after hours and in urban zones, by an out-ofhospital emergency service, although patients are free to go directly to the hospital A&ED. The Elche hospital’s A&ED is organized in 2 well-differentiated areas: obstetrics-gynecology and general emergencies. The general emergencies area, in turn, has a separate section for pediatrics and another area for emergencies in trauma, surgery, internal medicine, and specializations. In 1996, the general emergencies service had 62,286 cases (94.1% corresponded to medical patients). The study population included adult medical patients who presented to the hospital A&ED between May 1996 and April 1997. The sample size was calculated to estimate the proportion of inappropriate visits assuming infinite populations, a rate of inappropriate use of approximately 30%, based on an earlier validation study,28 and precision of ±2% for the 95% confidence interval (CI). This sample, which included approximately 2,000 cases, was increased by another 1,000 to guarantee the availability of cases in the analysis of variables with multiple categories. Because use of the A&ED is subject to several time variations but it is difficult to employ reviewers 24 hours a day year-round, we established a random sampling of 84 eight-hour shifts (8 AM to 3 PM, 3 PM to 10 PM, 10 PM to 8 AM). For this purpose, 21 numbers between 1 and 91 or 92 were generated randomly, depending on the number of days of each quar-

JUNE 2001

37:6


5 6 9


ter; the first 7 were assigned to morning shifts, the next 7 to afternoon shifts, and the next 7 to night shifts. The process was repeated for the 4 quarters, and all the patients who were seen during these shifts were reviewed. The field work was done by members of the medical staff of the A&ED and by residents working in the A&ED. All participants were previously trained in the use of the study protocols and worked under the direct supervision of the research team. The working shifts of the participating physicians were scheduled to coincide with the days and shifts that had been selected randomly for the sample. Outcome measures were (1) appropriateness of the visit to the A&ED, (2) motivation to consider a visit as appropriate, and when applicable, (3) reasons for inappropriate use of these services. The Hospital Urgencies Appropriateness Protocol (HUAP) was used to identify inappropriate visits. HUAP is an instrument designed around explicit criteria that defines as inappropriate cases treated in the A&ED that could have been resolved in a similar fashion outside the hospital setting.28 This definition does not question the appropriateness of the medical care provided in the A&ED, and therefore assumes that cases in which the A&ED provided emergency care were appropriate, even when they were not true emergencies. The HUAP (Appendix 1) comprises 5 sections including criteria for severity, treatment, diagnostic tests, some outcomes, and a section applicable only to patients in the A&ED who were not referred by a physician. Criteria 1.1 through 1.8 assess the severity of the case. They record the stability of the patient’s physiologic systems (pulse, blood pressure, temperature, electrolytic balance, blood gas values), and register the sudden loss of functionality of an organ or system (including fractures, hemiplegia, and so on). The second set of criteria (2.1 through 2.4) covers treatment administered in the A&ED assuming the case is severe, and assuming that such treatment would be difficult to provide in the primary care setting (still a possibility at this point). Except for procedures performed in the operating room and setting bones with plaster casts, this section includes the administration of oxygen or of any drug or fluids intravenously when they are indicated for any reason other than maintaining a preventive pathway. The third group of criteria (3.1 through 3.4) covers diagnostic tests that are performed in the A&ED, indicating the need for a rapid diagnosis. The next series of criteria (4.1 through 4.2) identifies patients admitted or kept for a prolonged time in the A&ED, suggesting that this was an appropriate case. The last set of criteria (5.1 through

5 7 0

5.8) refers to self-referrals and attempts to determine whether the patient was justified in bypassing primary care treatment. Sections are provided for subjective criteria pertaining both to patients referred by a physician (section 4.9) and self-referred patients (section 5.9). For a visit to be assessed as appropriate by the HUAP, at least 1 of the criteria in any section must be met. Cases that met none of the criteria are therefore considered inappropriate. When the HUAP was validated in an earlier study,28 in which the prevalence of inappropriate use of A&ED services was 27%, the instrument demonstrated excellent intrareviewer and interreviewer reliability, although its validity, compared with the judgments of experts, proved to be moderate. Appendix 2 summarizes the validation study of the HUAP. The HUAP is complemented with a list of possible reasons for inappropriate visits. This list was compiled on the basis of a literature review and incorporates suggestions made by clinicians in an effort to make it as extensive and detailed as possible. Its purpose is to help the attending clinician assign a reason for inappropriate A&ED use as part of the concurrent review of each case. All information analyzed in the study was compiled concurrently during the patient’s visit and was recorded by the attending physician. In many cases, 1 of the authors was also present. When a visit was classified as inappropriate, 1 of the researchers (always a member of the A&ED staff) interviewed the patient, following a nonstructured format, and assigned the reasons for inappropriate use (Appendix 1, list of reasons). The following variables and definitions were used: date and hour of arrival, the patient’s sex and age (79 years), residence (health care zone, also taking into account patients temporarily living outside their own health care zones and patients brought for evaluation by law enforcement agencies), the physician or institution who referred the patient (variables grouped as self-referral, referred by a physician from the hospital, or referred by the hospital itself, referred by the primary care physician or any other physician outside the hospital setting, either with or without a referral slip, referred by a judge or a law enforcement agency, referred by the out-of-hospital emergency service, and transferred from another hospital), means of transportation (ambulance, bus, car driven by patient, car driven by other person, walking, taxi), time necessary to reach the hospital, and whether the patient lived alone. The study was authorized by the hospital directorate (observational studies do not need the ethical commit-


37:6 JUNE 2001


tee’s approval) and did not include any data that would reveal the identity of the patient. Only the file number was retained to retrieve and review the case records at a later date. All necessary measures were implemented to bar access from third parties to this information. The first analysis describes the characteristics of the A&ED patients. The next analysis details the HUAP criteria met by the patients. The next step establishes possible associations between the patients’ characteristics and the percentage of inappropriate visits. Differences that were possibly significant were evaluated by 95% CIs. The final analysis describes the reasons for inappropriate visits on the basis of the evaluation made by the attending clinician after the open-format interview with the patient. All calculations were made with the STATA statistical program (version 5.0; Stata Corporation, College Station, TX). R E S U LT S

Table 1 describes the study population. Of the total, 79% of the patients went to the A&ED with no referral, 12% were referred by their primary care physician, and 4% were referred by the out-of-hospital emergency service; the total number of patients from outside the hospital’s district was less than 10%. A breakdown by disease category shows that 15.8% of the patients sought care for diseases of the central nervous system and sense organs, 14.9% for respiratory diseases, and 12.6% for circulatory system diseases. Of the total number of visits to the A&ED, 29.6% (95% CI 27.9 to 31.2) were evaluated as inappropriate and the remaining 70.4% as appropriate (Table 2). Table 2 also shows the percentage of patients who met criteria for appropriateness: 16.1% of the patients presented at least 1 criterion indicating a severe condition, 24.6% received 1 or more of the treatments included in the HUAP, 63.6% received at least 1 diagnostic test (90.3% of the patients evaluated as appropriate had at least 1 diagnostic test), and 19.4% of the cases were considered appropriate because they met at least 1 criterion under the section of the HUAP reserved for patients who visited the A&ED with no medical referral. A greater proportion of inappropriate visits was associated with younger patients, patients arriving in their own vehicles, referrals from the hospital, residence in specific health care zones, patients temporarily outside their own health care districts, and patients referred by law enforcement agencies (Table 3). No significant differences were found for sex, living alone, or the distance to the hospital,

JUNE 2001

37:6


Table 1.

Patient characteristics.* Characteristic Age (y) 79 Sex Male Female Lives alone Yes No Mode of transportation Ambulance Drives/bus/on foot Car/taxi Others Distance from hospital 21 min Referral Self-referral Primary care physician Out-of-hospital emergency service Hospital Others Residence Health care zone 1 Health care zone 2 Health care zone 3 Health care zone 4 Health care zone 5 Health care zone 6 Health care zone 7 Rural zone Outside own zone Court orders Others Diagnostic groups (ICD-9) Infectious diseases Neoplasias Endocrine diseases Diseases of the blood Mental diseases Diseases of the CNS and senses Diseases of circulatory system Diseases of respiratory system Diseases of digestive system Genitourinary diseases Diseases of skin Musculoskeletal diseases Poisoning and injuries Others Total

No. (%)

95% CI

1,299 (44.1) 766 (26.0) 675 (22.9) 206 (7.0)

42.2–45.9 24.4–27.6 21.4–24.5 6.1–7.9

1,479 (49.6) 1,501 (50.4)

47.8–51.4 48.5–52.1

139 (4.7) 2,822 (95.3)

3.9–55.1 94.5–96.0

288 (9.6) 395 (13.2) 2,226 (74.6) 71 (2.4)

8.6–10.7 12.0–14.5 73.0–76.2 1.8–2.9

1,296 (48.7) 1,039 (39.9) 326 (12.3)

46.8–50.6 37.1–40.9 11.0–13.5

2,358 (79.3) 353 (11.9) 123 (4.1) 121 (4.1) 17 (0.6)

77.8–80.7 10.7–13.0 3.4–4.9 3.3–4.8 0.3–0.9

1,151 (39.9) 326 (11.3) 230 (8.0) 346 (12.0) 164 (5.7) 112 (3.9) 264 (9.2) 74 (2.6) 80 (2.8) 14 (0.5) 121 (4.2)

38.1–41.7 10.1–12.5 7.0–9.0 10.8–13.2 4.8–6.5 3.2–4.6 8.1–10.2 2.2–3.4 2.0–3.2 0.3–0.8 3.4–5.0

71 (2.4) 44 (1.5) 42 (1.4) 19 (0.6) 188 (6.3) 470 (15.8) 374 (12.6) 443 (14.9) 276 (9.3) 288 (9.7) 31 (1.0) 176 (5.9) 135 (4.5) 423 (14.2) 2,980 (100.0)

1.8–2.9 1.0–1.9 1.0–1.8 0.3–0.9 5.4–7.1 14.4–17.0 11.3–13.7 13.5–16.1 8.2–10.3 8.6–10.7 0.6–1.4 5.0–6.7 3.7–5.2 12.9–15.4 —

ICD-9, International Classification of Diseases, ninth revision; CNS, central nervous system. * Cases without data: age, 34; lives alone, 19; distance, 319; referral, 8; residence, 8.

5 7 1


or between self-referred patients (30.4% of inappropriate visits) or patients referred by a physician (26.5%). There were fewer inappropriate visits in March and October; April had the highest rate (40%), but there were no significant differences between the days of the week and the time of day. Diseases of the blood and circulatory system, neoplasias, and respiratory and genitourinary diseases

Table 2.

Inappropriate visits and criteria for classification as appropriate. Criteria on the HUAP Severity 1.1 Loss of consciousness, disorientation, … 1.2 Sudden loss of sight or hearing 1.3 Altered pulse (140 beats/min) 1.4 Altered blood pressure 1.5 Alterations in blood electrolytes and gases 1.6 Persistent fever (5 days) not controlled with… 1.7 Active hemorrhage… 1.8 Sudden loss of functional capacity Subtotal criteria for severity Treatment 2.1 Intravenous medication/fluids 2.2 Administration of oxygen 2.3 Plaster casts 2.4 Surgical intervention/procedure Subtotal criteria for treatment Tests 3.1 Monitoring of vital signs/take… 3.2 Radiology of any type 3.3 Laboratory test, except… 3.4 ECG, except… Subtotal criteria for diagnostic tests Others 4.1 >12 h in A&ED 4.2 Admitted to hospital/transferred/dead 4.9 Others Subtotal other criteria Patient 7.2 Has had accident… 7.3 Symptoms suggesting vital emergency… 7.4 Pathology known to patient… 7.5 Patient told to come if… 7.6 Requires primary care and hospital is closest 7.9 Others Subtotal criteria for patients Total appropriate visits Total inappropriate visits Total

No. (%)

95% CI

207 (6.9) 59 (2.0) 41 (1.4) 39 (1.3) 110 (3.7) 28 (0.9) 97 (3.2) 46 (1.5) 481 (16.1)

6.0–7.8 1.5–2.5 0.9–1.8 0.9–1.7 3.0–4.4 0.6–1.3 2.6–3.9 1.1–2.0 14.8–17.5

728 (24.4) 213 (7.1) 7 (0.2) 69 (2.3) 732 (24.6)

22.9–26.0 6.2–8.1 0.1–0.4 1.8–2.8 23.0–26.1

62 (2.1) 1,569 (52.6) 1,518 (50.9) 710 (23.8) 1,894 (63.5)

1.6–2.6 50.8–54.4 49.1–52.7 22.3–25.3 61.8–65.3

24 (0.8) 438 (14.7) 24 (0.8) 468 (15.7)

0.5–1.1 13.4–16.0 0.5–1.1 14.4–17.0

11 (0.4) 384 (12.9) 190 (6.4) 42 (1.4) 24 (0.8) 10 (0.3) 577 (19.4) 2,098 (70.4) 882 (29.6) 2,980 (100.0)

0.1–0.6 11.7–14.1 5.5–7.2 1.0–1.8 0.5–1.1 0.1–0.5 17.9–20.8 68.8–72.0 27.9–31.2 —

The subtotals in the group of “appropriate” patients refer to the percentage of patients who met at least 1 criterion in the respective section of the protocol. However, because a patient may have met several criteria simultaneously, subtotals may not necessarily correspond to the sum of the respective columns. See Appendix 1 for a complete item description.

5 7 2

Table 3.

Inappropriate use of hospital A&ED by patient characteristics and monthly distribution. Characteristic Age (y)*