Complete text PDF - Semes

BRIEF REPORTS
Diagnostic agreement between a comprehensive,
fast-track community hospital emergency department
and the corresponding referral hospital
FRANCISCO JOSÉ MELLADO VERGEL 1 , NORBERTO DÍAZ RICOMÁ 1 , DIEGO MENA PARRA 1 ,
FERNANDO ROSELL ORTIZ 2 , ANTONIA MARÍA CONTRERAS RODRÍGUEZ 1 , LAURA LEÓN RUIZ 3
1 Hospital de Alta Resolución El Toyo. Empresa Pública Hospital de Poniente de Almería. Consejería de Salud
de la Junta de Andalucía. Almería, Spain. 2 Empresa Pública Emergencias Sanitarias. Almería, Spain. 3 Complejo
Hospitalario Torrecárdenas. Almería, Spain.
CORRESPONDENCE:
Dr. Francisco José Mellado Vergel
C/Grecia, 110
040009 Almería, Spain.
E-mail:
franciscojose.mellado@ephpo.es
fjmelladov@hotmail.com
RECEIVED:
7-5-2010
ACCEPTED:
11-5-2010
CONFLICT OF INTERES:
None
Introduction
Diagnostic error is one of the main adverse
events that can occur in hospitalized patients.
Emergency departments, with increasingly high
demand for services in recent years 1-5 are one of
the high-risk areas 6 . Surveillance systems to monitor
these errors are notably absent in most centres,
nor are there standard values establishing
what constitutes an acceptable rate of misdiagnosis
in emergency departments 7 .
The emergency department (ED) of the “Hospital
de Alta Resolución El Toyo”, a comprehen-
Objectives: To assess between-hospital agreement on diagnoses for patients transferred
from a "high resolution" (comprehensive, fast-track) community hospital to the
corresponding referral hospital, analyzing the rate of diagnostic error, epidemiologic
factors related to error, reasons for error, and influence on in-hospital mortality.
Methods: Diagnoses for patients transferred in 2009 were studied. Agreement was
analyzed in 2 categories: 1) diagnosis within the correct syndrome grouping, whether
strictly coincident or not, and 2) incorrect diagnosis. Variables compared were age, sex,
the hospital department ordering the transfer, and diagnosis. The main reason for error
was investigated.
Results: A total of 1494 patients were transferred to the referral hospital. The mean (SD)
age of transferred patients was 41.6 (30) years; 56% were men. Ninety-one percent
were sent from the emergency department. The most frequent diagnostic groups were
trauma (27%), internal medicine (25.8%), surgery (20.2%), and pediatrics (9.7%). There
was diagnostic agreement in 91.8% of the cases. The most common reasons for error
were insufficient clinical evaluation (69.7%), and misinterpretation of radiographs
(21.3%). Age was associated with lack of diagnostic agreement (P=.003). No deaths
occurred as a result of lack of agreement.
Conclusions: In general, there was diagnostic agreement; the error rate was 8.2%. Sex,
origin of the transfer order, and diagnostic group were unrelated to error; younger
patient age was related to lack of agreement. The main reason for error was inadequate
clinical evaluation. There were no deaths among patients for whom there was diagnostic
disagreement. [Emergencias 2010;22:282-285]
Key words: Diagnostic agreement. Diagnostic error. Mortality. Hospital emergency
health services.
sive, fast-track community hospital, is staffed by
family physicians trained in accident and emergency
medicine, with a laboratory, radiology and
CT service, and a radiologist (conventional physically
present on weekdays, telemedicine at other
times), without any other specialist being physically
present. The service consists of an ED with
consultation and rapid treatment room, and a
multipurpose area that functions as a short stay
unit. For specialized assessment and/or other
diagnostic tests, the patient is transferred to their
referral hospital (Torrecárdenas Hospital Complex).
282 Emergencias 2010; 22: 282-285

DIAGNOSTIC AGREEMENT BETWEEN A COMPREHENSIVE, FAST-TRACK COMMUNITY HOSPITAL EMERGENCY DEPARTMENT
We studied the group of patients attended at
our ED and then transferred to the referral hospital
in the year 2009, in order to determine the
degree of agreement between our ED diagnosis
and the definitive discharge diagnosis, and analyzed:
a) the rate of misdiagnosis, b) possible relationship
between misdiagnosis and epidemiological
variables, c) the primary reasons for error, and
d) possible mortality associated with these errors.
Method
We performed a prospective, observational, cohort
study. The study population consisted of patients
referred from our ED to the reference hospital
during the year 2009. From the emergency
medical history we gathered data on age, sex, origin
of transfer and ED diagnosis. We then compared
this diagnosis with that in the report from
the referral hospital concerned. The diagnoses
were coded according to International Classification
of Diseases, 9th Edition, Clinical Modification.
We considered diagnostic agreement as being the
same diagnosis or, if not exactly the same, the
syndrome was correctly identified, requiring action
and/or tests not available at our hospital. Otherwise,
the diagnosis was considered erroneous. In
cases of misdiagnosis, we reviewed in detail the
medical history and laboratory tests to determine
if the reason for the error lay primarily in: a) inadequate
clinical assessment (background, symptoms
or signs, physical examination), b) radiological interpretation,
c) laboratory analytical interpretation)
electrocardiographic interpretation (in this group
we analyzed cases to assess the impact of such interpretation
on the treatment administered).
We compared the variables analyzed for the
group with and without diagnostic agreement.
Descriptive analysis was performed using univariate
and multivariate binary logistic regression.
Results
During the study period, a total of 50,461
emergency patients were attended; 1,598 (3.2%)
were transferred, of whom 1,494 (93%) were
transferred. The remaining 104 patients were
mainly those who used their own means and subsequently
consulted a specialist (ophthalmology,
otolaryngology and pediatrics), but also those erroneously
identified (mainly foreigners) and patients
who ultimately decided not to attend the
referral hospital. The reasons for transfer were to
undergo specialized assessment and/or other diagnostic
tests (imaging tests, invasive procedures
or specific analytical microbiology).
The mean age was 41.6 ± 30.0 years, 56%
were male. As for the origin of transfer, the ED
transferred 1,359 (91%) and the multi-purpose
area 135 (9%). Diagnostic groups were distributed
as follows: trauma 403 (27%), internal
medicine 386 (25.8%), surgery 302 (20.2%), pediatrics
145 (9.7%), gynecology 118 (7 9%), ophthalmology
104 (7%) and otolaryngology 36
(2.4%). There was diagnostic agreement in 1,372
(91.8%) of the cases (87.2% coincided and
12.8% had correct identification of the syndrome),
and 122 (8.2%) discordant diagnoses.
The reasons for diagnostic error were inadequate
clinical assessment in 85 cases (69.7%), radiological
interpretation in 26 (21.3%), analytical interpretation
in 7 (5.7%) and electrocardiographic interpretation
in 4 (3.3%) cases.
Table 1 lists the most common discordant diagnoses,
grouped according to type of error.
There were no deaths in this group of patients
without diagnostic agreement.
Univariate analysis of the groups with and
without diagnostic agreement is shown in Table
2. The group without diagnostic agreement was
younger, most commonly referred from the ED,
and belonging to the pediatric diagnostic group.
Table 1. Most frequent suspected diagnoses categorized according to reasons for error in the 122 patients without final diagnostic
agreementwith the referral hospital
Error in clinical assessment Radiological error Analytical error ECG error
N = 85 (69,7%) N = 26 (21,3%) N = 7 (5,7%) N = 4 (3,3%)
Soft tissue injury (11) Limb fracture (19) Respiratory infection (2) Arrhythmias (4)
Infection in children (7) Vertebral fracture (3) Miscellaneous (5)
Acute abdomen (6)
Upper gastrointestinal bleeding (4)
Acute coronary syndrome (4)
Gastrointestinal disorders in children (4)
Stroke (3)
Ear disease (3)
Miscellaneous * (43)
Miscellaneous (4)
*By specialties: surgery (12), internal medicine (8), pediatrics (7), ophthalmology (7), trauma (3), gynecology (3) otolaryngology (3). ECG: electrocardiography.
Emergencias 2010; 22: 282-285 283

F. J. Mellado Vergel et al.
Table 2. Univariate analysis in patient groups with and without diagnostic agreement
Variables DA (N = 1.372) No DA (N = 122) p OR (IC95%)
Age, mean ± SD 42.2 ± 24.7 35.2 ± 26.8 < 0.05
Sex, n (%) 0.24
Male 761 (55.5) 75 (61) 1 (reference)
Female 609 (44.5) 48 (39) 0.8 (0.548-1.167)
Origin of transfer, n (%) < 0.05
ED 1.241 (90.5) 118 (95.9) 1 (reference)
Multipurpose area
Diagnostic groups, n (%)
130 (9.5) 5 (4.1) 0.404 (0.162-1.008)
Traumatology 367 (26.8) 36 (29.3) 0.55 1.132 (0.754-1.700)*
Internal Medicine 357 (26) 29 (23.6) 0.55 0.876 (0.568-1.352)*
Surgery 280 (20.4) 22 (17.9) 0.50 0.849 (0.525-1.371)*
Pediatrics 124 (9) 21 (17.1) < 0.05 2.070 (1.250-3.429)*
Gynecology 114 (8.3) 4 (3.3) 0.46 0.371 (0.134-1.022)*
Ophthalmology 97 (7.1) 7 (5.7) 0.56 0.793 (0.360-1.747)*
Otolaryngology 32 (2.3) 4 (3.3) 0.52 1.407 (0.489-4.04)*
DA: diagnostic agreement. OR: odds ratio. CI: confidence interval. *The reference category is constituted by other patients.
Binary logistic regression with these three variables
was performed, and only age remained a
significant variable, so the group of patients without
diagnostic agreement were younger.
Discussion
With nearly 1,500 patients, the cohort analyzed
is the largest published to date in Spain. It also has
the peculiarity of including a wide range of diagnostic
categories, covering many different specialties,
and an age range from childhood to old age.
A comparison of the diagnoses between our
ED and the referral hospital showed a misdiagnosis
rate of 8.2%. There is no established value
against which to measure the quality of diagnosis
of our center. The only reference values available
are from previous studies that used different samples
and methods of analysis, with a rate of misdiagnosis
that varies widely, from a low 0.6% 11
4.6% 7 , 10.7% 5 , 14.6% 12 , and up to 19.7% 4 .
A very important finding was that there were
no deaths among any of the patients with no
diagnostic agreement, a key factor in assessing
health outcomes after medical errors.
According to our results, sex, origin of transfer
and diagnostic groups are not related to erroneous
diagnoses, only age remained associated
with no diagnostic agreement, a fact not found in
other studies, where age was unrelated 7 , or increased
age was associated with misdiagnosis, as
in the work of Miret et al. 4 , but in this study the
population studied presented advanced middle
age (73 years). As for the different diagnostic
groups, several studies have linked error with a
wide variety of diagnoses 11-16 .
As regards the reasons for error in our series,
the main cause (69.7%) was inadequate clinical
assessment, as found in other similar studies 4,7,11
with percentages between 40 and 55%. However,
as indicated by Tudela et al. 7 , to interpret these
results we must also consider the possibility that
the initial clinical assessment was not incorrect,
but in some cases the evolution of the disease involved
increasing expressiveness and the final diagnosis
only became evident after admission to
the referral hospital. Errors of radiographic interpretation
17 were the second most common cause
of error, and in our series they occurred predominantly
when the diagnosis was trauma. As for errors
in electrocardiographic interpretation, the
four errors in our study were merely in classifying
the type of arrhythmia, with no impact on the
treatment administered.
With these results it seems reasonable to suggest
that a systematic clinical and radiological assessment
and careful diagnosis could improve accuracy.
The combination of experienced clinical professionals
with more time for patient care and less
ED workload could help to improve diagnostic
agreement 7 .
Certain limitations of our study should be noted;
first, 7% of patients transferred to the referral
hospital were lost to follow up, but on review it
was found that these were cases of low morbidity
and attended the referral hospital or other specialists
using their own means. Second, the gold
standard was the final diagnosis of the referral
hospital, supported by the main criteria for referral,
but this could occasionally be incorrect. Third,
the limited number of cases in the group with no
diagnostic agreement could have influenced some
of our results.
Finally, further studies on diagnostic agreement
are required to establish acceptable limits, taking
into account the peculiarities of each centre. Such
284 Emergencias 2010; 22: 282-285

surveillance systems serve to improve the relationship
between different levels of care, undergo
joint training activities and improve the quality of
patient care 4 .
References
DIAGNOSTIC AGREEMENT BETWEEN A COMPREHENSIVE, FAST-TRACK COMMUNITY HOSPITAL EMERGENCY DEPARTMENT
1 Millá J. Urgencias médicas: algo más que una serie televisiva. Med
Clin (Barc). 2001;117:295-6.
2 Sieck S. The evolution of a new standard of hospital care. Paradigmshift
to the emergency department and the role of point of care
testing. Point of Care. 2006;5:2-5.
3 Sánchez López A, Bueno Cavanillas A. Factores asociados al uso inadecuado
de un servicio de urgencias hospitalario. Emergencias. 2005;17:138-44.
4 Mirete C, López-Bayo L, Blázquez JC, Rodríguez-Rodriguez P, Serrano
MI, De Teresa L. El error diagnóstico en el área de ingresos de un
hospital de media-larga estancia. Estudio de concordancia diagnóstica.
Rev Esp Geriatr Gerontol. 2005;40:195-8.
5 Estella A, Pérez-Bello L, Sánchez JI, Toledo MD, Del Águila D. Actividad
asistencial en la unidad de observación de un hospital de segundo
nivel. Emergencias. 2009;21:95-8.
6 Weingart SN, Wilson R, Gibberd RW, Harrison B. Epidemiology of
medical error. BMJ. 2000;320:774-7.
7 Tudela P, Mòdol JM, Veny A, Tor J, Bonet M, Rego MJ. Estudio de la concordancia
diagnóstica entre el área médica de urgencias y la de hospitalización
en un hospital general. Med Clin (Barc). 2002;119:531-3.
8 Rodríguez-Caravaca G, Villar del Campo I. Concordancia diagnóstica
entre atención primaria y atención especializada tras consulta urgente.
Aten Primaria. 2000;25:292-6.
9 Fernández-Cotrina JM, García-Ruiz AJ, Márquez-Castilla M, Jiménez-
Guerrero O, Reina-Caballero E, Bocanegra-Pérez A, et al. Acuerdo
diagnóstico entre equipos de emergencias y servicios hospitalarios.
Gac Sanit. 2004;18:458-63.
10 Carrasco JL, Jover L. Métodos estadísticos para evaluar la concordancia.
Med Clin (Barc). 2004;122(Supl 1):28-34.
11 Chellis M, Olson JE, Augustine J, Hamilton GC. Evaluation of missed
diagnosis for patients admitted from the emergency department.
Acad Emerg Med. 2001;8:125-30.
12 Plaza S, Álvarez-Sala JL, García T, Montero J,González J, Montero E,
et al. El error en el diagnóstico de urgencia. Estudio retrospectivo de
1.233 historias clínicas. An Med Intern. (Madrid). 1987;4:171-6.
13 Tudela P, Davant E, Monreal M, Segura A, Valencia J, Carreres A.
Análisis clínico de la tromboembolia pulmonar no sospechada en el
servicio de urgencias. Med Clin (Barc). 2000;114:292-3.
14 Mòdol JM, Tudela P, Veny A, Sahuquillo JC, Tor J, Giménez M. Indicaciones
y rendimiento diagnóstico de la punción lumbar en urgencias.
Med Clin (Barc). 2002;118:10-2.
15 Balaguer JV, Gabriel F, Braso JV, Núñez C, Catalá T, Labios M. El papel de
la autopsia clínica en el control de calidad de los diagnósticos clínicos en
una unidad de urgencias. An Med Intern. (Madrid). 1998;15:179-82.
16 Famularo G, Salvini P, Terranova A, Gerace C. Clinical errors in emergency
medicine: experience at the emergency department of an italian
teaching hospital. Acad Emerg Med. 2000;7:1278-81.
17 Urrutia A, Bechini J, Tor J, Olazábal A, Rey-Joly C. Evaluación de la interpretación
de la radiografía de tórax por los médicos de guardia
en un hospital universitario. Med Clin (Barc). 2001;117:323-33.
Concordancia diagnóstica entre el servicio de urgencias de un hospital de alta resolución
y su hospital de referencia
Mellado Vergel FJ, Díaz Ricomá N, Mena Parra D, Rosell Ortiz F, Contreras Rodríguez AM, León Ruiz L
Objetivos: Conocer la concordancia diagnóstica de los pacientes trasladados desde las urgencias de un hospital de alta
resolución a su hospital de referencia, y analizar el índice de error diagnóstico, sus características epidemiológicas, los
motivos fundamentales de error y su influencia en la mortalidad intrahospitalaria.
Métodos: Se analizaron los pacientes trasladados durante el año 2009. Se establecieron 2 niveles de concordancia
diagnóstica: 1º coincidente o bien no coincidente pero sindrómicamente correcta, y 2º errónea. Se compararon las variables:
edad, sexo, origen de traslado y diagnóstico. Se determinó el motivo fundamental de error.
Resultados: Se trasladaron al hospital de referencia 1.494 pacientes. La edad media fue de 41,6 ± 30 años, y el 56%
fueron hombres. Desde urgencias se trasladaron el 91%. Los grupos diagnósticos más frecuentes fueron: traumatología
27%, medicina interna 25,8%, cirugía 20,2% y pediatría 9,7%. La concordancia diagnóstica fue del 91,8%. Los motivos
de error más frecuentes fueron una valoración clínica inadecuada (69,7%) y la interpretación radiológica (21,3%).
La variable asociada al grupo de pacientes sin concordancia diagnóstica fue la edad (p = 0,003). No hubo ningún fallecimiento
en los pacientes sin concordancia diagnóstica.
Conclusiones: La concordancia diagnóstica fue mayoritaria con un índice de error del 8,2%. El sexo, el origen de traslado
y los grupos diagnósticos no están relacionados con el error diagnóstico, sólo los pacientes de menor edad se
asocian significativamente a la no concordancia diagnóstica. El motivo de error principal fue una inadecuada valoración
clínica. No se registró mortalidad en el grupo de pacientes sin concordancia diagnóstica. [Emergencias
2010;22:282-285]
Palabras clave: Concordancia diagnóstica. Error diagnóstico. Mortalidad. Urgencias hospitalarias.
Emergencias 2010; 22: 282-285 285