Auditing secondary prevention of ischemic heart disease in rural areas of Spain: an opportunity for improvement. José M. Turón 1 , Julián Librero 1,2 , Gabriel J. Diaz, Juan J. García, Carme Beltral and Francisco Abal, on behalf of the REDIMER study group (semFYC). 1. Instituto Aragonés de Ciencias de la Salud 2. Fundación Instituto de Investigación en Servicios de Salud Correspondence to José M. Turón; E mail: [email protected]. REDIAPP Aragón. Paseo Sagasta 52, planta 6ª, 50008 Zaragoza (Spain) Telephone number 699 827850
Transcript
Auditing secondary prevention of ischemic heart disease in rural areas of Spain: an opportunity for improvement.
José M. Turón1, Julián Librero1,2, Gabriel J. Diaz, Juan J. García, Carme Beltral and Francisco Abal, on behalf of the REDIMER study group (semFYC).1. Instituto Aragonés de Ciencias de la Salud 2. Fundación Instituto de Investigación en Servicios de
Salud
Correspondence to José M. Turón; E mail: [email protected] . REDIAPP Aragón. Paseo
AUDITING SECONDARY PREVENTION OF ISCHEMIC HEART DISEASE IN RURAL AREAS OF SPAIN: AN
OPPORTUNITY FOR IMPROVEMENT
AimTo describe the standard of secondary prevention received by individuals
with a history of ischemic heart disease (IHD) in Spanish rural areas, and the factors associated with the low standards of quality.
Methods Medical audit. Patients with a history of IHD, whose data were
contributed by 72 rural physicians in 12 autonomous regions (1030 cases). Quality criteria were used based on international guidelines. Multivariate analysis was employed to assess the variables associated with poor quality attention.
Results30.9% of patients gave target Low Density Lipoprotein (LDL) readings and 68.1% reached the target Blood Pressure (BP). 48.5% took beta blockers. 29% of patients had not visited a specialist within the previous year. The fact that patients had visited a specialist within the previous year was associated with their having followed the types of treatments recommended in the guidelines (p< 0.01) and with obtaining target LDL and BP readings (p<0.05). Patients from the smallest villages had the lowest probability of having LDL controls and also of receiving hypolipidaemic therapy (p<0.01). People with antecedents of isolated angina had lower probability of being treated with antiaggregants than those who had experienced previous acute myocardial infarction (p<0,01). ConclusionsThe quality of secondary prevention for these patients shows there is room for improvement. Problems of accessibility exist for some collectives, which may be improved with the involvement of rural primary health care teams.
Keywords. Rural Health Services, medical audit, cardiovascular diseases, Spain, Quality Assurance, Health Care
Cardiovascular diseases make up the leading cause of death in Spain1. Despite the
fact that patients with antecedents of ischemic heart disease form the collective
with the highest cardiovascular risk and are considered a priority group in
preventive cardiology2, the process of secondary prevention (SP) of ischemic heart
disease has never traditionally formed part of the services offered in primary health
care in our country. It is carried out by specialists, with variable involvement from
GPs.
Several studies3,4,5,6, pointed out numerous problems in the quality of attention
received by these patients, also in Spain7.
In this context, our group, the rural research network (REDIMER) – an
organism dependent on the Rural Medicine Work Group of the Spanish Society for
Family and Community Medicine – made up of Spanish doctors working in rural
areas-, set about the task of describing specific aspects related to the quality of SP
of ischemic heart disease in rural and semi-rural areas in Spain, and analysing
those factors associated with the differences in its management.
Methods
Observational study of quality standards and variables potentially associated with
shortcomings in quality.
Participating doctors.
78 GPs working in rural and semi-rural areas in from 12 Spanish Autonomous
Regions. According to the Spanish National Institute of Statistics criteria,
villages/towns with fewer than 2000 inhabitants are considered as rural, and towns
with between 2000 and 10000 inhabitants as semi-rural.
The participating doctors constituted a convenience sample; they were recruited by
members of the Rural Group at national conferences, local meetings or in their
workplaces. They were offered the opportunity to participate in a network whose
aim is to bring to light information from the rural environment. 78 interested doctors
were sent a letter inviting them to participate in this first project and all accepted.
Study Units.
Patients under 85 years of age, with diagnosed Ischemic Heart Disease, from the
practices of the doctors were identified by registering those that came to renew
their prescriptions for cholesterol-lowering drugs, antiaggregants-anticoagulants,
beta-blockers, calcium-channel blockers or nitrates between September and
December, 2002. Included in the study were cases, entered by the participating
doctors, which fulfilled the following criteria:
1. Existence in the patient’s medical records of a hospital report diagnosing
Ischemic Heart Disease prior to 1 December 2001.
2. Patient under 85 years of age
3. Patient did not die or move in the year 2002
Attention rendered between 1 December 2001 and 1 December 2002 was
analysed.
Criteria for quality assessment.
The following were used as fulfilment criteria for intermediate result targets. The
time frame used was the previous 12 months.
- The mean of the last two BP readings of the year was to be lower than
140/90 mmHg.
- The last LDL reading of the year was to be lower than 100 mgrs/dl (2.6
mmol/L).
- The last body mass index (BMI) reading of the year was to be lower than 25.
- Patients were to walk at least two hours per week. This was ascertained by
telephone survey using a Likert scale of 5 categories (from always to never)
Exceptions: physical or psychological disability.
- Patients on the programme were not to smoke. This was ascertained by
telephone survey.
Cases of LDL <130-mgrs/dl (3.4 mmol/L) - and likewise total cholesterol< 200
mgrs/dl (5.2 mmol/l)- and BMI<30 were also considered. Fulfilment criteria were
inspired by international guidelines on SP, SP in elderly patients, and chronic
angina 8,9,10. We chose the target BP reading 140/90 mmHg in order to be able to
compare results with previous studies3,5,11.
Fulfilment of the above-mentioned criteria was valued (fulfils/does not
fulfil/exception). The criterion of Sedentarism “Do you walk at least two hours per
week?” was considered fulfilled if the reply was “always” or “nearly always”.
Process variables.
Dichotomic variables, created with the aim of standardising the intensity of the
patient monitoring from the health centre, were formulated with the consensus of
the professionals from the Rural Group. We estimated the existence of an entry in
the patients’ medical records within the previous six months of :
- At least two visits with blood pressure readings.
- At least one LDL Cholesterol reading. Patients with
hypertriglyceridaemia were excluded.
- At least two visits with weight annotation. Excluded : Patients who
received visits in their homes.
Other variables studied:
- Age
- Gender
- Diabetes
- Population of the patient’s town/village of residence.
- Distance, in minutes, from the patient’s village/town of residence to:
o the town where the referred cardiologist’s surgery is located
o the town/village where the health centre is located
In cases where the patient resided in the same locality as the health centre/referred
cardiologist’s surgery a value of 0 was given.
- Visits made to the referred cardiologist the previous year (question in
telephone survey) (Yes, No)
- Type of ischemic heart disease: record of acute myocardial infarction (AMI)
(Yes, No) or of angina (Yes, No)
- Consumption of prophylactic drugs:
o antiaggregants-anticoagulants
o beta blockers
o angiotensin-converting enzyme (ACE) Inhibitors
o angiotensin II antagonist (ARBs)
o cholesterol lowering drugs
Source of data:
- Medical records for all variables except for physical activity, smoking, and
visits made to referred cardiologist in the previous year which were obtained
by telephone survey.
Analysis:
The percentage of fulfilment of the referred quality criteria was determined. We had
to accept lack of data for some criteria results (BP, LDL and BMI) for certain
patients owing to the fact that medical records were used as the source of data.
Data was missing from 198 cases (19%) in the BP in range criterion, 243 cases
(23%) in objective LDL<100, and the same figure for BMI. A descriptive study was
made of the characteristics of the missing case data in the three criteria.
For the bivariate analysis of the factors associated with standard of assistance
(patient characteristics and place of residence), chi-square, student-T, ANOVA,
Mann Whitney-U and Kuskall Wallis tests were used.
In the study of the association between diverse individual factors and the poor
quality of assistance, the odds ratio of each variable was calculated, adjusted for
the rest, by means of logistic regression. Firstly, independent variables were: sex,
age group (under 65, over 65), population of place of residence distributed into
terciles (fewer than 294, 295-3400, 3401 and over), distance from patient’s
town/village of residence to that where the PC health centre is located, distance
from patient’s town/village of residence to the locality where the cardiologist’s
surgery is found, diabetes (Yes or No), visits to specialist in the previous twelve
months (Yes or No), history of AMI (Yes or No) and history of Angina (Yes or No).
The dependent variable was the existence or not of a quality assistance process,
adjusted to the model as “0” if it fulfilled the criterion, and “1” if it did not. Also by
means of logistic regression, factors associated with the consumption of drugs
recommended by the guidelines (antiaggregants-anticoagulants, beta blockers,
hypolipidaemic agents and ACE Inhibitors) have been explored, using the same
independent variables.
In all of the analyses, the regression model was adjusted for possible dependence
between observations of individuals belonging to the same physician (physician
cluster effect).
Results
There were 1030 patients of whom 69.6 % were males. Mean age : 70.3 +
9.7 SD (Range 33-85). 75.6% were over 65 years of age. 55.6% had a history of
AMI and 62% of angina. Telephone surveys could not be conducted with 12
From among the target range criteria (Table 2), those that were met least were
the targets of LDL<100 (30.9%) and BMI<25 (19.6%). 51.5% of cases did not use
beta blockers and 34.7% did not use any hypolipidaemic agents at all (table 2).
In the three variables most affected by lack of information (BP<140/90, LDL<100
and obesity) the missing cases showed, in relation to the study total, a lower
proportion of diabetic patients (18, 15 and 21% respectively) and of patients visited
by a specialist (39%, 39% y 37%).
Factors associated with non-fulfilment of quality criteria
Tables 3–4 give the factors that have shown a statistically significant
association (p<0,05) with the studied criteria. The variables that were more
frequently associated with quality criteria were the fact of having visited a specialist
in the previous year, diabetes, age, and population of patient’s town/village of
residence.
With regard to pathology type, patients with a history of AMI had greater
probability of being treated with antiaggregants (ORa-multivariate adjusted Odds
Ratio- 2.41), hypolipidaemic agents (ORa1.8) or ACE inhibitors (ORa1.7) than
those who had only experienced angina.
The distance from patient’s town/village of residence to that of specialist’s
surgery or health centre variable was not associated with either having visited the
specialist or fulfilment of process variables, respectively.
Discussion
The results presented are subject to limitations derived from the representativeness
of the doctors and patients. At the outset of the study, few primary care (PC)
practices had computerized records (11%) and there were problems with access to
the computerized diagnostic records of district hospitals. For these reasons we
decided to create new diagnostic records with the collaboration of volunteer
doctors. The new register was made up of patients coming to health centres to
renew prescriptions, taking into account the fact that in the Spanish National Health
Service almost all prescriptions for chronic disease are issued in PC. The sample,
therefore, is not necessarily representative of IHD management in the rural
environment. On the other hand, this study deals with the largest number of cases
and from the greatest number of locations of those carried out to date in PC, rural
or urban, in Spain. In any case our initial hypothesis was that the problems in
quality that we would encounter in patients of these doctors would not be less
important than those encountered by the general collective of rural physicians.
The inclusion criteria for patients who came to renew their prescriptions in a
period of 4 months may have excluded the less regular or non-attending patients.
This could cause an overestimation in the consumption of medication and, on the
other hand, introduce selection bias if the inverse care law12 were fulfilled and ex-
cluded patients had a higher burden of CVD risk factors. However, we consider
this bias, which would again underestimate quality problems, to be minimal in a
public health system, such as ours, that requires compulsory registration with one
doctor, and which subsidises the full cost of medication for pensioners. With the
exception of one, all SP quality studies carried out in PC in Spain following hos-
pital discharge are cross-sectional7.
As no tradition exists of specific SP programmes in PC, the quality of follow up is
greatly influenced by actions taken by specialists and hospitals. Preceding studies
have focused on very localized catchment areas and therefore may be influenced
by the quality of attention in one particular hospital. This is not the case in our study
which given that it takes in patients from 12 Autonomous Regions with a number of
different tertiary referral hospitals.
Comparison with other studies (table 5)
We have also consulted prospective studies3,6 closest in time to our own.
The range of BP figures are higher than those obtained in all of the studies
consulted 3, 6,11 including the prospective ones. Our study did not take in the use of
antihypertensives, a factor which might help to explain the differences.
Target LDL fulfilment is unsatisfactory, although it is better than results
obtained in the consulted cross-sectional studies carried out in Spain 11,13 .
The percentage of smokers obtained is much lower than reflected in studies
where a breath carbon monoxide monitor was used. There was likely to have been
bias in the gathering of this information as it was the patient’s doctor who
conducted the telephone survey.
Our results in recommended drug prescription are in line with concurrent studies in
Spain (table 5)6,11,13,14 , with moderately higher ACE inhibitor/ARB consumption. The
Euroaspire study3 makes note of the infrequent use of beta blockers in Spanish
patients, data that are corroborated in all of the studies carried out in Spain, as well
as in ours. From the results of our study for LDL it can also be perceived that an
added effort should be made to extend the use of cholesterol-lowering treatments.
Discussion of associations ascertained
We found that the patients from the smallest villages had the least probability of
having recorded LDL readings, unlike those from intermediate and larger towns.
Although no association was found for the variable “distance (in minutes by car)
from patient’s town/village of residence to the town/village where the health centre
is located” (place where blood samples were taken), this paradox could be
explained by the fact that a large number patients had no access to a private
vehicle, or to public transport to reach the place where their health centre is located
at times when extractions take place, factors not considered in the study. People
resident in the smallest villages also had a lower probability of being on cholesterol-
lowering treatments.
For older patients (over 65), it so happened that there was a lower
probability of being on hypolipidaemic therapy, yet they reached target LDL levels
of <100 mgrs/dl better than younger patients. It is possible that here there is a bias
owing to survival.
Women appeared to be more sedentary and obese than men, who reached
target blood pressure on more occasions than women. On the contrary, women
visited the health centre more often. These associations have been also pointed out
by another study in Spain 11.
We also found in the multivariate analysis that diabetic patients receive a
more regular monitoring than the remainder of the patients in this study (table 3).
We believe this could be due to the fact that it is a disease whose monitoring is
programmed in primary health care. Nevertheless, this greater frequency of
controls in the health centre has not resulted in any association at all with the
fulfilment of target criteria. This data could be related to what has been termed in
previous studies15 record of activities without clinical purpose, i.e., the frequent
orientation of monitoring programmes for chronic patients centred on the carrying
out and recording of activities, with scant emphasis on obtaining target results.
Patients who did not visit their specialist in the previous year satisfied to a
lesser degree target figures included in the criteria, with worse fulfilment of target
LDL and BP. They also had greater probability of not following prophylactic
treatments, and went to fewer check-ups in PC. While we are aware of the
restraints set by the cross-sectional design of our study, we cannot rule out this
information being an example of the inverse care law12, as it is commonly observed
that there is no direct connection between the use of services and necessity.
CONCLUSIONS
Our data show that there is room for improvement, particularly in a number of
areas. It has been said that one of the reasons for the current validity of the inverse
care law16 within a public health system, is that specific groups within the population
are less represented in studies that involve the general population owing to the diffi-
culty in obtaining good rates of response. This, in turn, has consequences within
the process of policy making. This could be the case of the rural sector. The
present work, with all the outlined limitations, aims to bring to light information with
which to improve this situation.
We need to find an effective and feasible organizational strategy which will allow
SP to be introduced with participation of Primary Care in rural areas. There are suc-
cessful experiences in the literature17, 18,19 based on the involvement of Primary
Health Care. A strategy based on process management and the wider use of in-
formation technology in the rural environment would be a good starting point.
AcknowledgmentsREDIMER group : José María Turón Alcaine, Gabriel J. Diaz Grávalos, Rafael Alonso Roca, Jose Luis Rodríguez Yeste, Josep Cañellas Isern, Ana Vazquez Torguet, Beatriz Solans Aisa, Teofilo Lorente Aznar Mª José González Bouzo, José Luis Romero Limia, Luis A. Vázquez Fernandez, Gerardo Palmeiro Fernández, Ana Rodríguez Fernández, Celso Sánchez González, José Manuel Quintáns, Margarita Arandia García, Mª Adoración Juiz Crespo, Enrique López Vázquez, , Isabel Monreal Aliaga, Lourdes Enciso Ciriano, Javier Citoler Perez, José María Millat Medina, Ana Carmen Gimenez Baratech, Ana Mª Fernandez García, Salvador Gestoso Gayá, David Medina i Bombardó, Laura Romero Fernandez, Jasone Basterretxea Oiarzabal, Daniel Domínguez Tristancho, Mercedes Martinez Gonzalez, Inmaculada Casado Gorriz, Jaume Banqué Vidiella, Juan José García Díaz, Tomás García Martinez, Esteban Gracia Gil, Rosana Arribas Garcia, José Tomás Gomez Saenz, Maria Salud Hernández Juanes, José Antonio Morales Ruiz, Emiliano Rodríguez Sánchez, Maria José Gamero Samino, Eva Gomez Iglesias, Miguel Angel Mercader Mercadé, Pere Farrás Serra, Rodolfo Montoya Barquet, Jose Manuel Lopez de Goicoechea, Francisco Abal Ferrer Joaquín Cuetos Alvarez, Aurora Suarez Reguera,Pablo Belderrain Belderrain, Lina Belenguer Carreras, Lucía Sierra Santos, José María Rubio Ruiz, Carlos Del Valle Hernández, Javier Martín Fuertes, Sagrario Lausín Marín, José Alcubierre Cura, Eva Rua Portu, Antón Aldanondo Gabilondo, Jesús Ubalde Saenz, Francisco José Fagundez Santiago, Pilar Del Rio, Luis García Burriel, Sara Fanlo Abellá, Javier Marzo Arana, Luis Angel Florez, Amelia Rojas Marín, Mercé Fuentes Pujol, Dolors García i Perez, Antonio Morales Jiménez, Albert Clapés Roca, Carme Beltral Lopez, Ramón Barberá Reus, Jordi Espinás Boquet, Rafael Gracia Ballarin, Maria Angeles Gutierrez Stampa and Esteban Gonzalez Lopez.We are grateful to the Aragonese node of REDIAPP and the Instituto Aragonés de Ciencias de la Salud for their collaboration on this project.
References
1. Ministerio de Sanidad y Consumo. Plan Integral de Cardiopatía Isquémica 2004-2007.
2. De Backer G, Ambrosioni E, Borch-Johnsen K, Brotons C, Cifkova R, Dallongeville J, et
al. European guidelines on cardiovascular disease prevention in clinical practice.
European Heart Journal 2003; 24(17):1601-1610.
3. Euroaspire II Study Group. Lifestyle and risk factor management and use of drug
therapies in coronary patients from 15 countries. Principal results from EURASPIRE II
Euro Heart Survey Programme. European Heart Journal (2001) 22, 554-572.
4. Brady AJB, Oliver MA, Pittard, JB. Secondary prevention in 24431patients with coronary
14. Romera I., Salinero M.A., Del Río P. Prevención secundaria de la cardiopatía isquémica
tipo angor y post infarto agudo de miocardio en Atención Primaria. Semergen. 2002; 28
(7): 359-62
15. Saturno, P.J.; Gascón, J.J.; Bueno, J.M., Alcaraz, J.; Martinez, P. El diagnóstico de las
dislipemias en Atención Primaria : un servicio a mejorar. Resultados de una evaluación
multicéntrica. Aten Primaria1999; 25 (2): 82-88.
16. Watt G. The inverse care law today. Lancet 2002; 360: 252-254
17. Campbell, NC; Ritchie, LD; Thain, J.; Deans, H G; Rawles, J M; Squair JL. Secondary
prevention in coronary heart disease: a randomised trial of nurse led clinics in primary
care. Heart. 1998; 80 (5): 447-452.
18. Secondary prevention clinics for coronary heart disease: four year follow up of a
randomised controlled trial in primary care. Murchie P., Campbell N., Ritchie D., Simpson
JA, Thain J. BMJ.2003. January 11; 326(7380):84
19. Randomised trials of secondary prevention programmes in coronary heart disease:
systematic review. McAlister, Finlay A; Lawson, Fiona M E; Teo, Koon K; Armstrong,
Paul W. BMJ. 2001.323 (7319), 957-962.
Table 1.Sample characteristics
Median (range) or n (%)1 Number of inhabitants of town/village of residence2 Usual Travelling time in minutes from town/village of residence to the town/village of the health centre (HC) or the specialist’s surgery (SS), by car. 3Patients who did not visit their specialist in the previous twelve monthsAMI- Acute Myocardial Infarction
Age (>65)
779 (75.6%)
Inhabitants 1 792 (4-15.000)Time to HC (min) 2 5 (0-45)Time to SS (min)2 30 (0-180)
Gender (male) 716 (69.6%)
Diabetes 282(27.4%)
AMI 571(55.6%)
Angina 642(62.0%)
No specialist visit3 297(28.9%)
Fulfilment of criteria %BP < 140/90 mmHg 68.1LDL < 100 mgrs/dl (2,6 mmol/L) 30.9LDL < 130 mgrs/dl (3,4 mmol/L) 72 BMI < 30 62.5BMI < 25 19.6non-sedentary 69.7non_smoking 92.2 2 BP previous six months 69.7 1 LDL checks previous six months 59.52 weight checks previous six months 43.6
Prophylactic drug consumptionAntiaggregrant/anticoagulants 85Hypolipidaemic agents 65.3Beta blockers 48.5ACE inhibitors 37.7Angiotensin II antagonist 8.1Table 2. Results
BP, Blood pressure; LDL, Low density lipoprotein; BMI, Body mass index; ACE: Angiotensin-converting enzyme
ORa (CI95)
BP in range Gender: Male Visits to specialist
1.62(1.15-2.28)†1.47(1.02-2.13)*
LDL <100 mgrs/dl
Age over 65 Visits to specialist
1.76(1.17-2.65)†1.57(1.03-2.38)†
LDL < 130 mgrs/dl Visits to specialist 1.58(1.08-2.31)*
BMI > 25 Diabetes Age over 65
1.73(1.07-2.78)†0.57(0.34-0.94)*
BMI > 30 Diabetes Gender: Male
1.68(1.19-2.38) †0.67(0.47-0.95)*
Sedentarism Gender: Male 0.36(0.25-0.50) ‡
Smoking Gender: Male Age over 65 Visits to specialist
Diabetes Age over 65 Inhabitants nucleus population Town/Village>3400 inhabitants (Basal: Village<294) Town/Village 294-3400 (Basal: Village < 294)
Visits to specialist
1.95 (1.38-2.76)‡0.63(0.45-0.90)*
2.02(1.13-3.61)*1.80(1.18-2.76)†
1.57(1.13-2.20)†
Two weight checks six months
Diabetes Age over 65 Visits to specialist
2.91(1.97-4.30) ‡1.57(1.05-2.33)*1.57(1.05-2.33)*
Table 3. Factors associated with fulfilment of quality criteria (multivariate analysis). ORa(CI95): Odds ratio adjusted (Confidence Interval 95%)* p<0.05 † p<0.01 ‡ p<0.001Factors analysed: gender, age group, population of place of residence distributed into terciles (fewer than 294, 295-3400, 3401 and over),distance from patient’s town/village of residence to locality of health centre and specialist’s surgery, diabetes, visits to specialist, history of AMI and history of Angina.Abbreviations: BP, Blood pressure; LDL, Low density lipoprotein; Inhabitants nucleus population, No of inhabitants of patient’s town/village of residence ; BMI, Body mass index.
ORa (CI95)
Hypolipidaemic agent
Diabetes AMI Age over 65 Inhabitants nucleus population1 Town/ Village>3400 inhabitants (Basal : Village <294) Visits to specialist
Table 4. Factors associated with prescription of recommended drugs (multivariate analysis). ORa(CI95): Odds ratio adjusted (Confidence Interval 95%)* p<0.05 † p<0.01 ‡ p<0.001 Factors analysed: gender, age group, population of place of residence distributed into terciles (fewer than 294, 295-3400, 3401 and over), distance from patient’s town/village of residence to localities of health centre village and specialist’s surgery, diabetes, visits to specialist, history of AMI and history of Angina.
1Inhabitants nucleus population: No of inhabitants of patient’s town/village of residence ; 2ACE: Angiotensin-converting enzyme
