Differences in the causes of death of HIV-positive patients in a cohort study by data sources and coding algorithms
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
CONCISE COMMUNICATION
Differences in the causes of death of HIV-positive
patients in a cohort study by data sources and
coding algorithms
Victoria Hernandoa,b, Paz Sobrino-Vegasa,b, M. Carmen Burrielc,
Juan Berenguerd, Gemma Navarroe, Ignacio Santosf,
Jesús Reparazg, M. Angeles Martı́nezh, Antonio Antelai,
Félix Gutiérrezj, Julia del Amoa,b, for CoRIS cohort
Objectives: To compare causes of death (CoDs) from two independent sources: National
Basic Death File (NBDF) and deaths reported to the Spanish HIV Research cohort [Cohort
de adultos con infección por VIH de la Red de Investigación en SIDA CoRIS)] and
compare the two coding algorithms: International Classification of Diseases, 10th revision
(ICD-10) and revised version of Coding Causes of Death in HIV (revised CoDe).
Methods: Between 2004 and 2008, CoDs were obtained from the cohort records (free
text, multiple causes) and also from NBDF (ICD-10). CoDs from CoRIS were coded
according to ICD-10 and revised CoDe by a panel. Deaths were compared by 13
disease groups: HIV/AIDS, liver diseases, malignancies, infections, cardiovascular,
blood disorders, pulmonary, central nervous system, drug use, external, suicide, other
causes and ill defined.
Results: There were 160 deaths. Concordance for the 13 groups was observed in 111
(69%) cases for the two sources and in 115 (72%) cases for the two coding algorithms.
According to revised CoDe, the commonest CoDs were HIV/AIDS (53%), non-AIDS
malignancies (11%) and liver related (9%), these percentages were similar, 57, 10 and
8%, respectively, for NBDF (coded as ICD-10). When using ICD-10 to code deaths in
CoRIS, wherein HIV infection was known in everyone, the proportion of non-AIDS
malignancies was 13%, liver-related accounted for 3%, while HIV/AIDS reached 70%
due to liver-related, infections and ill-defined causes being coded as HIV/AIDS.
Conclusion: There is substantial variation in CoDs in HIV-infected persons according
to sources and algorithms. ICD-10 in patients known to be HIV-positive overestimates
HIV/AIDS-related deaths at the expense of underestimating liver-related diseases,
infections and ill defined causes. CoDe seems as the best option for cohort studies.
ß 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins
AIDS 2012, 26:1829–1834
Keywords: AIDS, causes of death, hepatitis, HIV, malignancies, mortality
a
Red de Investigación en Sida, Centro Nacional de Epidemiologı́a, Instituto de Salud Carlos III, Madrid, bCIBER de Epidemiologı́a y
Salud Pública (CIBERESP), Madrid, cServicio de Vigilancia y Salud Pública, Servicio de Información e Investigación Sanitaria,
Zaragoza, dHospital Universitario Gregorio Marañon, Madrid, eHospital Parc Tauli, Sabadell, fHospital Universitario La Princesa,
Madrid, gHospital de Navarra, Pamplona, hHospital Universitario San Cecı́lio, Granada, iHospital Universitário de Santiago de
Compostela, Santiago de Compostela, and jHospital Universitário de Elche, Elche, Spain.
Correspondence to Victoria Hernando, Red de Investigación en Sida, Centro Nacional de Epidemiologı́a – ISCIII, Avda. Monforte
de Lemos 5, 28029 Madrid, Spain.
Tel: +34 91 8222921; fax: +34 91 3877513; e-mail: vhernando@isciii.es
Received: 24 October 2011; revised: 4 January 2012; accepted: 16 February 2012.
DOI:10.1097/QAD.0b013e328352ada4
ISSN 0269-9370 Q 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins 1829
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.1830 AIDS 2012, Vol 26 No 14
Introduction text variables. For the purpose of this study, we coded
CoDs according to two coding systems:
Accurate information on causes of death (CoDs) is hard to
obtain and is subject to selection and information biases. (1) ICD-10 – we assigned a code to each death with the
Most countries have national death registries that compile assistance of an ICD-10 coding expert familiar with
information from death certificates and code CoD by coding rules.
international classification rules. In Spain, the National (2) Revised CoDe – we assigned a code to each death
Statistics Institute collects data on CoD according to the following CoDe protocol assisted by ART-CC.
International Classification of Diseases, 10th revision
(ICD-10). ICD-10 provides a set of rules to extract the
maximum information from the death certificate, usually National Statistics Institute: National Basic
based on a single CoD [1]. ICD-10 was implemented in Death File
1999 and introduced specific codes for HIV/AIDS which The NBDF provides information on the date and CoDs
were not present in ICD-9 [2]. Cohort studies of HIV- of all persons dying in Spain coded by the underlying
positive patients have not traditionally used the ICD-10. CoDs in accordance with the ICD-10. The data are
Other algorithms, such as the Coding Causes of Death in obtained from civil registries and from the National
HIV (CoDe) protocol are used instead [3,4]. CoDe was Statistics Institute itself through the Statistical Bulletin on
developed in 2004 by the Copenhagen HIV Programme Deaths which is compiled from death certificates.
(CHIP) (http://www.cphiv.dk) and collects extensive
information on CoD which is examined in a centralized Period of analysis
review process. CoDe is used by the Data Collection on Data from the two sources were cross-matched in the first
Adverse events of Anti-HIV Drugs (D:A:D) and quarter of 2010 for individuals who had died between
EuroSIDA, among others [5–7]. Unfortunately, as it is January 2004, when CoRIS was created, and December
not always possible to collect detailed information, a 2008, date of the last available update of the NBDF.
simplified version of CoDe selection rules has been used
by the Antiretroviral Therapy Cohort Collaboration Data analysis
(ART-CC) in what we call the ‘revised CoDe’ (http:// Thirteen clinical categories were created for the CoD in
www.art-cohort-collaboration.org). HIV-positive persons. These are described in Table 1,
which also shows the individual ICD-10 and revised
CoRIS, the Cohort of the Spanish Network of CoDe codes included in each category.
Excellence on HIV/AIDS Research, collects information
on multiple CoDs from reporting physicians. In 2008, we The CoD from each source – those reported to CoRIS
obtained CoDs for deceased cohort members from the and those included in the NBDF, and from each
National Statistics Institute, which provides single CoD algorithms – ICD-10 rules and revised CoDe, for
coded with ICD-10. We hypothesize that CoD identified CoD reported to CoRIS– were compared according
from each of source will vary substantially due either to categories of diseases.
sources and/or the coding rules. Our objectives are to
compare the CoD obtained from two independent
data sources using ICD-10: the National Basic Death Results
File (NBDF) and reports to CoRIS provided directly by
the physicians between 2004 and 2008. We also compare Overall, 4687 patients, 11 491.28 person-years follow-up
deaths reported in CoRIS using two different coding and 160 deaths were observed from January 2004 to
algorithms: the ICD-10 and the revised CoDe. December 2008, yielding a crude mortality rate of 1.39
per 100 person-years [95% confidence interval (CI)
1.19–1.62]. Of these 160 deaths, 87% (n ¼ 139) were
Methods men, median age was 44 years [interquartile range (IQR)
37–51], 36% (n ¼ 57) were injecting drug users, 38%
Sources of information (n ¼ 60) heterosexuals, 18% (n ¼ 29) MSM and 9%
CoRIS (n ¼ 14) had others or unknown categories. The median
CoRIS is an open multicentre cohort of HIV-positive CD4 cell count within 6 months of death for the 55%
patients naive to antiretroviral therapy from 32 centres patients with data available was 117 cells/ml (IQR 39–
from 12 of the 17 autonomous regions of Spain. The 259) and 67% received antiretroviral treatment.
project was approved by the Institutional Ethics Review
Boards and each participant is required to sign an Comparing cause of death in National Basic
informed consent form [8]. The variables collected Death File and CoRIS using International
include socio-demographic, epidemiological, clinical, Classification of Diseases, 10th revision
treatment and mortality data; vital status, date of death Comparing the CoD for the 13 categories from each
and underlying CoD, as well as contributing causes as free source yielded concordance in 111 (69%) cases (k ¼ 0.48;
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.Coding causes of death in cohort studies Hernando et al. 1831
Table 1. International Classification of Diseases, 10th revision and revised Coding Causes of Death in HIV codes included in each disease group
category.
Disease groups ICD-10 code CoDe code
HIV/AIDS-associated diseases A02, A073, A15–A19, A30, A31, A812, 01 (AIDS)
B00, B20–B24, B25, B371, B383–B389,
B393–B399, B451–B459, B582, B588–B589,
C46, C53, C83, C857, C859, R75
Liver disease B15–B19, K65, K70–K77, K922 03 (chronic viral hepatitis),
14 (liver failure)
Tumours C00–C52, C54–C82, C84–D48 04 (malignancy)
Infectious diseases A00–A019, A03–A072, A078–A09, A20–A28, 02 (infection)
A32–A563, A568–A811, A813–A99, B01–B09,
B25–B370, B372–B382, B39–B392, B40–B450,
B46–B581, B583, B99, G00–G02, J12–J18
Cardiovascular disease I00–I45, I47–I99 08 (ischaemic heart disease), 09 (stroke),
24 (heart or vascular disease)
Diseases of the blood D50–D89 20 (haematological disease)
Pulmonary diseases J40–J99 11 (primary pulmonary hypertension),
12 (lung embolus), 13 (chronic
obstructive lung disease),
25 (respiratory disease)
Central nervous system diseases G048-G99 23 (CNS disease)
Drug abuse F192, X40–X44, T36–T50, Y10–Y15 19 (substance abuse)
External causes X00–X30, X45–X49, X50–X59, X85–X90, V01–Y98 16 (accident or violent death)
Suicide X60–X84 17 (suicide)
Other diseases Any other code Any other code
Ill defined and unknown causes R092, R95, R960–R961, R98, R99 91 (unclassifiable cause),
92 (unknown)
CoDe, Coding Causes of Death in HIV; CNS, central nervous system; ICD-10, International Classification of Diseases, 10th revision.
CI 95% 0.36–0.59). However, only 29 of the 82 HIV/ known to be HIV infected. When the CoD is pneumonia
AIDS-associated deaths (35%) were classified with exactly in an HIV-positive person, the assigned ICD-10 code is
the same ICD-10 code in each source, so concordance for B20, B20.9 or B24 (HIV/AIDS related), whereas revised
individual codes was very low, although moderate for the CoDe considers CD4 cell counts prior to death: if CD4
13 categories. Of the 49 (31%) discordant CoD, 19 (39%) cell count is below 100 cells/ml, it gets coded as 01
had a completely different CoD, but 14 (29%) would have (AIDS), but if it is unknown or above 100 cells/ml, it gets
been concordant if the information about HIV-positive coded as 02 (infection).
status had been recorded in the Bulletin on Death. For
example, if CoD is cirrhosis and HIV status is unknown, Distribution of the cause of death using different
according to ICD-10, it gets coded as K74.6 (other and data sources and coding algorithms
unspecified cirrhosis of liver), but if HIV status is positive, ‘HIV/AIDS-associated causes’ was the most frequent
it gets coded as B23.8 (HIV disease resulting in other CoD irrespective of the source and coding algorithm,
specified conditions). accounting for 91 (57%) of the causes in the NBDF, 112
(70%) in CoRIS_ICD-10 and 85 (53%) in CoRIS_
Comparing cause of death in CoRIS using CoDe. The second most common category was ‘non-
International Classification of Diseases, 10th AIDS-defining tumours’: 16 (10%) in the NBDF, 21
revision and revised Coding Causes of Death in (13%) in CoRIS_ICD-10 and 18 (11%) in CoRIS_
HIV CoDe. ‘Liver diseases’ was the third most common cause
We compared the CoD according to ICD-10 rules and in CoRIS_CoDe, accounting for 15 (9%) cases, as well as
the revised CoDe algorithm and found concordance in in the NBDF, accounting for 12 (8%) cases. In contrast,
115 (72%) cases for the 13 categories (k ¼ 0.54; CI 95% ‘liver diseases’ accounted for only five (3%) cases in
0.43–0.64). Of the 45 discordant causes, 33 (73%) were CoRIS_ICD-10 (Fig. 1).
classified as ‘HIV/AIDS associated causes’ by ICD-10,
whereas 15 of these 33 (45%) are coded as poorly defined
by revised CoDe, 10 (30%) are classified as liver-related
and eight (24%) as infectious diseases. This is due to an Discussion
ICD-10 coding rule that converts into B24 codes (HIV
disease) all CoD which are unknown or ill defined in There are substantial differences in the CoD of HIV-
anyone who is known to be HIV infected. ICD-10 also positive people according to data sources and classifi-
converts into B23.8 codes (HIV disease) deaths from cation algorithms. HIV/AIDS-associated deaths were the
cirrhosis of viral cause or of unknown cause in people most frequent in our cohort for the 2004–2008 period,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.1832 AIDS 2012, Vol 26 No 14
80
70
60 NBDF
CoRIS_ICD10
Percentage
50
CoRIS_CoDe
40
30
20
10
0 es
s
rs
es
s
d
es
es
e
es
e
s
es
se
se
se
oo
us
id
ou
as
as
as
as
us
us
ic
ea
ea
a
ab
bl
m
u
se
se
se
se
ca
se
ca
S
is
is
e
Tu
s
di
di
di
di
i
th
rd
rd
d
al
ed
g
ru
ed
us
y
em
rn
er
of
ve
la
in
ar
D
te
th
tio
at
cu
ef
s
Li
on
st
Ex
O
se
ci
-d
c
as
sy
lm
fe
so
ea
Ill
ov
In
us
Pu
as
is
di
o
D
S-
ar
rv
ID
ne
C
/A
l
tra Cause of death
IV
H
en
C
Fig. 1. Distribution of the 160 deaths in the CoRIS cohort by disease group category as classified by the National Basic Death
Files (NBDF) and by two coding systems: CoRIS International Classification of Diseases, 10th revision (CoRIS_ICD-10) and
CoRIS Coding Causes of Death in HIV (CoRIS_CoDe).
followed by non-AIDS-defining tumours, for both data 1000 person-years of follow-up in 1997 to 4.4 per 1000
sources and both classification algorithms. The third most person-years in 2002 [10]. The ART-CC found that 50%
common CoD was liver disease according to the NBDF of deaths from 1996 to 2006 were associated with AIDS,
and revised CoDe in CoRIS but was not when cohort followed by non-AIDS malignancies (12%) and non-
data was coded using ICD-10. Applying ICD-10 to AIDS infections (8%). Overall, a decline in HIV/AIDS-
cohort data overestimates HIV/AIDS-associated deaths associated mortality was observed with increased duration
largely at the expense of labelling as HIV/AIDS- of antiretroviral therapy [6].
associated ill defined causes and liver diseases because
deaths from cirrhosis of viral cause or of unknown cause In this study, data from the NBDF, which were coded by
in people known to be HIV infected are assigned to HIV/ ICD-10, seem to be more concordant with those from
AIDS-related causes. This explains why in CoRIS_ICD- CoRIS_CoDe than with CoRIS_ICD-10. This is
10 liver-related causes are in fourth position, why there because the largest source of misclassification bias we
are virtually no deaths attributed to infectious diseases, found was derived from the rule that deaths in persons
none is labelled as ill defined, and HIV/AIDS-associated known to be HIV-positive should be coded as HIV/
causes account for 70% of all deaths. In fact, Garcia- AIDS-related deaths. As information on HIV status is
Fulgueiras et al. [9], in a study of hepatitis B and C more likely to be missing from death certificates than in
mortality in Spain, included an estimation of the CoRIS, a cohort of people with confirmed HIV
attributable fraction of mortality due to hepatitis virus infection, lack of information seems to result in a picture
among cases coded as AIDS. which is closer to reality. These results call for caution
when comparing different studies, as this rule is often not
Applying CoDe to CoRIS shows that just over half of the taken into account.
deaths are due to AIDS, which seems to be more in line
with recent data from similar settings which show a One of the main limitations of NBDF – given that it
reduction in HIV/AIDS-associated deaths compared depends on death certificates – is the poor quality of the
with previous years [5,6,10,11]. In the Swiss HIV Cohort latter. The importance of correct completion of death
Study, HIV/AIDS-associated deaths decreased to 15% of certificates has been well described, together with the
all CoD between 2005 and 2009; 85% of deaths were due common errors that can be minimized with adequate
to non-AIDS-defining conditions, and non-AIDS- training [12,13]. Also, when deaths occur outside the
defining malignancies were the most important CoD. hospital or the doctor certifying the death does not know
In our cohort, non-AIDS-defining malignancies were the the deceased, death certificates may miss important
second CoD [11]. Likewise, data from EuroSIDA show a information [12,14] and HIV infection may be missing
decrease in the incidence rate of AIDS from 118.3 per because of the stigma associated. However, for some
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.Coding causes of death in cohort studies Hernando et al. 1833
conditions such as suicide, which trigger legal notifica- y Marı́a Ángeles Muñoz. Fieldwork, data management
tion, the proportion of deaths was exactly the same with and analysis: Ana Maria Caro Murillo, Paz Sobrino Vegas,
all three strategies. Santiago Pérez-Cachafeiro, Victoria Hernando Sebastián,
Belén Alejos Ferreras, Débora Álvarez del Arco, Susana
The main limitation of this study is the low number of Monge Corella, Inma Jarrı́n Vera. Committee for coding
deaths, which may have introduced random error in some cause of death: Victoria Hernando Sebastián, Paz Sobrino
of our estimates and comparisons. However, this does not Vegas, Carmen Burriel, Roberto Muga, Félix Gutiérrez,
apply to systematic error and the relatively small number Santiago Moreno, Julia del Amo. BioBanco: M Ángeles
of deaths permitted a more comprehensive exploration of Muñoz-Fernández, Isabel Garcı́a-Merino, Coral Gómez
the misclassification bias which would have been Rico, Jorge Gallego de la Fuente y Almudena Garcı́a
unmanageable with larger numbers. Although both Torre. Participating centres: Hospital General Universi-
coding algorithms establish clear rules, we must take tario de Alicante (Alicante): Joaquı́n Portilla Sogorb,
into account that different reviewers can give different Esperanza Merino de Lucas, Sergio Reus Bañuls, Vicente
codes for a given death. Therefore, it is important to Boix Martı́nez, Livia Giner Oncina, Carmen Gadea
establish a centralized review system in order to check and Pastor, Irene Portilla Tamarit, Patricia Arcaina Toledo.
correct these discrepancies. Hospital Universitario de Canarias (Santa Cruz de
Tenerife): Juan Luis Gómez Sirvent, Patricia Rodrı́guez
We can conclude that the revised CoDe classification is Fortúnez, Marı́a Remedios Alemán Valls, Marı́a del Mar
the best way to classify CoD in a cohort study of HIV- Alonso Socas, Ana Marı́a López Lirola, Marı́a Inmaculada
positive patients. Linkage with external registries such as Hernández Hernández, Felicitas Dı́az-Flores. Hospital
the national death index to obtain CoD in people lost to Carlos III (Madrid): Vicente Soriano, Pablo Labarga,
follow-up or with an unknown cause may introduce bias Pablo Barreiro, Carol Castañares, Pablo Rivas, Andrés
given that the ICD-10 overestimates HIV/AIDS-related Ruiz, Francisco Blanco, Pilar Garcı́a, Mercedes de Diego.
deaths. Hospital Universitario Central de Asturias (Oviedo):
Victor Asensi, Eulalia Valle, José Antonio Cartón.
Hospital Clinic (Barcelona): José M. Miró, Marı́a
Acknowledgements López-Dieguez, Christian Manzardo, Laura Zamora,
Iñaki Pérez, M Teresa Garcı́a, Carmen Ligero, José Luis
V.H.S. and J.d.A. conceived the study and wrote the first Blanco, Felipe Garcı́a-Alcaide, Esteban Martı́nez, Josep
draft of the manuscript. All authors contributed to the Mallolas, José M. Gatell. Hospital Doce de Octubre
final draft. V.H.S. and M.C.B. checked and coded data for (Madrid): Rafael Rubio, Federico Pulido, Silvana
all patients. V.H.S. and P.S.-V. extracted the data from Fiorante, Jara Llenas, Violeta Rodrı́guez, Mariano
CoRIS and did the analyses. F.G., J.B., G.N., I.S., J.R., Matarranz. Hospital Donostia (San Sebastián): José
M.A.M. and A.A. contributed to data interpretation. Antonio Iribarren, Julio Arrizabalaga, Marı́a José
Aramburu, Xabier Camino, Francisco Rodrı́guez-
Conflicts of interest Arrondo, Miguel Ángel von Wichmann, Lidia Pascual
This study would not have been possible without the Tomé, Miguel Ángel Goenaga, M Jesús Bustinduy,
collaboration of all the patients, medical and nursing staff Harkaitz Azkune Galparsoro. Hospital General Univer-
and data managers who have taken part in the project. sitario de Elche (Elche): Félix Gutiérrez, Mar Masiá,
The RIS Cohort (CoRIS) is funded by the Instituto de José Manuel Ramos, Sergio Padilla, Andrés Navarro,
Salud Carlos III through the Red Temática de Investiga- Fernando Montolio, Yolanda Peral, Catalina Robledano
ción Cooperativa en Sida (RIS C03/173). Garcı́a. Hospital Germans Trı́as i Pujol (Badalona):
Bonaventura Clotet, Cristina Tural, Lidia Ruiz, Cristina
The ART-Cohort Collaboration (ART-CC), which Miranda, Roberto Muga, Jordi Tor, Arantza Sanvisens.
provided a simplified version of the CoDe protocol used Hospital General Universitario Gregorio Marañón
in this study, is funded by the UK Medical Research (Madrid): Juan Berenguer, Juan Carlos López Bernaldo
Council, grant number G0700820. de Quirós, Pilar Miralles, Jaime Cosı́n Ochaı́ta, Matilde
Sánchez Conde, Isabel Gutiérrez Cuellar, Margarita
This work has been partially funded by grant from FIS Ramı́rez Schacke, Belén Padilla Ortega, Paloma Gijón
(Spanish Networks for Research on AIDS and Public Vidaurreta. Hospital Universitari de Tarragona Joan
Health), 04/0900 and RIS (Spanish HIV Research XXIII, IISPV, Universitat Rovira i Virgili (Tarragona):
Network for excellence), RD06/006. We are grateful for Francesc Vidal, Joaquı́n Peraire, Consuelo Viladés, Sergio
funds provided by CIBERESP (Ciber de Epidemiologia y Veloso, Montserrat Vargas, Miguel López-Dupla,
Salud Pública). Montserrat Olona, Joan-Josep Sirvent, Alba Aguilar,
Antoni Soriano. Hospital Universitario La Fe (Valencia):
Centres and investigators involved in CoRIS: Executive José López Aldeguer, Marino Blanes Juliá, José Lacruz
committee: Juan Berenguer, Julia del Amo, Federico Rodrigo, Miguel Salavert, Marta Montero, Eva Calabuig,
Garcı́a, Félix Gutiérrez, Pablo Labarga, Santiago Moreno Sandra Cuéllar. Hospital Universitário La Paz (Madrid):
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.1834 AIDS 2012, Vol 26 No 14
Juan González Garcı́a, Ignacio Bernardino de la Serna, This work has been presented as Oral Communication
José Marı́a Peña Sánchez de Rivera, Marta Mora Rillo, at the 13th European AIDS Conference, 12 – 15
José Ramón Arribas López, Marı́a Luisa Montes October 2011, Belgrade, Serbia.
Ramı́rez, José Francisco Pascual Pareja, Blanca Arribas,
Juan Miguel Castro, Fco Javier Zamora Vargas, Ignacio References
Pérez Valero. Hospital de la Princesa (Madrid): Ignacio de
los Santos, Jesús Sanz Sanz, Johana Rodrı́guez, Ana Salas 1. ICD-10. International Statistical Classification of Diseases and
Related Health Problems. http://www.who.int/classifications/
Aparicio, Cristina Sarriá Cepeda. Hospital San Pedro- icd/ICD-10_2nd_ed_volume2.pdf. 2011.
CIBIR (Logroño): José Antonio Oteo, José Ramón 2. 1993 revised classification system for HIV infection and
Blanco, Valvanera Ibarra, Luis Metola, Mercedes Sanz, expanded surveillance case definition for AIDS among adoles-
cents and adults. MMWR Recomm Rep 1992; 41:1–19.
Laura Pérez-Martı́nez. Hospital San Pedro II (Logroño): 3. Causes of death in HIV-1-infected patients treated with anti-
Javier Pinilla Moraza. Hospital Universitario Mutua retroviral therapy, 1996–2006: collaborative analysis of 13
de Terrassa (Terrassa): David Dalmau, Angels Jaén HIV cohort studies. Clin Infect Dis 2010; 50:1387–1396.
4. Copenhagen HIV Program. Protocol Coding Causes of Death in
Manzanera, Mireia Cairó Llobell, Daniel Irigoyen Puig, HIV (CoDe). http://www.cphiv.dk/Portals/_default/pdf_folder/
Laura Ibáñez, Queralt Jordano Montañez, Mariona code_protocol_ver_1.0.pdf. [Accessed on 19 April 2011]
Xercavins Valls, Javier Martinez-Lacasa, Pablo Velli, 5. Kowalska JD, Mocroft A, Ledergerber B, Florence E, Ristola M,
Begovac J, et al. A standardized algorithm for determining the
Roser Font. Hospital de Navarra (Pamplona): Julio Sola underlying cause of death in HIV infection as AIDS or non-
Boneta, Javier Uriz, Jesús Castiello, Jesús Reparaz, Marı́a AIDS related: results from the EuroSIDA study. HIV Clin Trials
2011; 12:109–117.
Jesús Arraiza, Carmen Irigoyen, David Mozas. Hospital 6. Kowalska JD, Friis-Moller N, Kirk O, Bannister W, Mocroft A,
Parc Taulı́ (Sabadell): Ferrán Segura, Marı́a José Sabin C, et al. The Coding Causes of Death in HIV (CoDe)
Amengual, Eva Penelo, Gemma Navarro, Montserrat Project: initial results and evaluation of methodology. Epide-
miology 2011; 22:516–523.
Sala, Manuel Cervantes, Valentı́n Pineda. Hospital 7. Smith C, Sabin CA, Lundgren JD, Thiebaut R, Weber R, Law M,
Ramón y Cajal (Madrid): Santiago Moreno, José Luis et al. Factors associated with specific causes of death amongst
Casado, Fernando Dronda, Ana Moreno, Marı́a Jesús HIV-positive individuals in the D:A:D Study. AIDS 2010;
24:1537–1548.
Pérez Elı́as, Dolores López, Carolina Gutiérrez, Beatriz 8. Caro-Murillo AM, Castilla J, Perez-Hoyos S, Miro JM, Podzamc-
Hernández, Marı́a Pumares, Paloma Martı́. Hospital zer D, Rubio R, et al. Spanish cohort of naive HIV-infected
Reina Sofı́a (Murcia): Alfredo Cano Sánchez, Enrique patients (CoRIS): rationale, organization and initial results.
Enferm Infecc Microbiol Clin 2007; 25:23–31.
Bernal Morell, Ángeles Muñoz Pérez. Hospital San 9. Garcia-Fulgueiras A, Garcia-Pina R, Morant C, Garcia-Ortuzar
Cecilio (Granada): Federico Garcı́a Garcı́a, José V, Genova R, Alvarez E. Hepatitis C and hepatitis B-related
Hernández Quero, Alejandro Peña Monje, Leopoldo mortality in Spain. Eur J Gastroenterol Hepatol 2009; 21:895–
901.
Muñoz Medina, Jorge Parra Ruiz. Centro Sanitario 10. Kowalska JD, Mocroft A, Phillips A, Reiss P, Ledergerber B,
Sandoval (Madrid): Jorge Del Romero Guerrero, Gatell J, et al. Risk of cause-specific deaths over calendar time
Carmen Rodrı́guez Martı́n, Teresa Puerta López, Juan and according to cumulative exposure to combination anti-
retroviral therapy (cART) [abstract]. In: Proceedings of the
Carlos Carrió Montiel, Paloma Raposo. Hospital 18th Conference on Retroviruses and Opportunistic Infections;
Universitario Santiago de Compostela (Santiago de February 2011; Boston, USA.
Compostela): Antonio Antela, Arturo Prieto, Elena 11. Ruppik M, Ledergerber B, Rickenbach M, Furrer H, Battegay M,
Cavassini M, et al. Changing patterns of causes of death in
Losada. Hospital Son Dureta (Palma de Mallorca): Melchor the Swiss HIV Cohort Study (SHCS) 2005–2009 [abstract]. In:
Riera, Javier Murillas, Maria Peñaranda, Maria Leyes, M Proceedings of the 18th Conference on Retroviruses and
Angels Ribas, Antoni Campins, Concepcion Villalonga. Opportunistic Infections; February 2011; Boston, USA.
12. Johansson LA, Westerling R. Comparing hospital discharge
Hospital Universitario de Valme (Sevilla): Juan Antonio records with death certificates: can the differences be
Pineda, Eva Recio Sánchez, Fernando Lozano de León, explained? J Epidemiol Community Health 2002; 56:301–
Juan Macı́as, José del Valle, Jesús Gómez-Mateos, Rosario 308.
13. Villar J, Perez-Mendez L. Evaluating an educational interven-
Mata. Hospital Virgen de la Victoria (Málaga): Jesús Santos tion to improve the accuracy of death certification among
González, Manuel Márquez Solero, Isabel Viciana Ramos, trainees from various specialties. BMC Health Serv Res 2007;
Rosario Palacios Muñoz. Hospital Universitario Virgen 7:183.
14. Johansson LA, Westerling R. Comparing Swedish hospital dis-
del Rocı́o (Sevilla): Pompeyo Viciana, Manuel Leal, Luis charge records with death certificates: implications for mor-
Fernando López-Cortés, Mónica Trastoy. tality statistics. Int J Epidemiol 2000; 29:495–502.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.You can also read
