Estimation of the burden of cardiovascular disease attributable to modifiable risk factors and cost-effectiveness analysis of preventative interventions to reduce this burden in argentina

Rubinstein et al. BMC Public Health 2010, 10:627http://www.biomedcentral.com/1471-2458/10/627 Estimation of the burden of cardiovasculardisease attributable to modifiable risk factors andcost-effectiveness analysis of preventativeinterventions to reduce this burden in Argentina Adolfo Rubinstein1,2*, Lisandro Colantonio1, Ariel Bardach1,3, Joaquín Caporale1,4, Sebastián García Martí1,2,Karin Kopitowski2, Andrea Alcaraz1, Luz Gibbons1, Federico Augustovski1,2, Andrés Pichón-Rivière1 Background: Cardiovascular disease (CVD) is the primary cause of mortality and morbidity in Argentinarepresenting 34.2% of deaths and 12.6% of potential years of life lost (PYLL). The aim of the study was to estimatethe burden of acute coronary heart disease (CHD) and stroke and the cost-effectiveness of preventativepopulation-based and clinical interventions.
Methods: An epidemiological model was built incorporating prevalence and distribution of high blood pressure,high cholesterol, hyperglycemia, overweight and obesity, smoking, and physical inactivity, obtained from theArgentine Survey of Risk Factors dataset. Population Attributable Fraction (PAF) of each risk factor was estimatedusing relative risks from international sources. Total fatal and non-fatal events, PYLL and Disability Adjusted LifeYears (DALY) were estimated. Costs of event were calculated from local utilization databases and expressed ininternational dollars (I$). Incremental cost-effectiveness ratios (ICER) were estimated for six interventions: reducingsalt in bread, mass media campaign to promote tobacco cessation, pharmacological therapy of high bloodpressure, pharmacological therapy of high cholesterol, tobacco cessation therapy with bupropion, and a multidrugstrategy for people with an estimated absolute risk > 20% in 10 years.
Results: An estimated total of 611,635 DALY was lost due to acute CHD and stroke for 2005. Modifiable risk factorsexplained 71.1% of DALY and more than 80% of events. Two interventions were cost-saving: lowering salt intake inthe population through reducing salt in bread and multidrug therapy targeted to persons with an absolute riskabove 20% in 10 years; three interventions had very acceptable ICERs: drug therapy for high blood pressure inhypertensive patients not yet undergoing treatment (I$ 2,908 per DALY saved), mass media campaign to promotetobacco cessation amongst smokers (I$ 3,186 per DALY saved), and lowering cholesterol with statin drug therapy(I$ 14,432 per DALY saved); and one intervention was not found to be cost-effective: tobacco cessation withbupropion (I$ 59,433 per DALY saved) Conclusions: Most of the interventions selected were cost-saving or very cost-effective. This study aims to informpolicy makers on resource-allocation decisions to reduce the burden of CVD in Argentina.
annual deaths around the world, constituting 11% of esti- Chronic diseases are increasing in developing countries mates for the global burden of disease. It is estimated and cardiovascular diseases account for 17.7 million that mortality due to coronary heart disease (CHD) andstroke will increase by approximately 145% among menand women from 1990 to 2020 in Latin America, com- * Correspondence: 1 pared with a 28% increase for women and a 50% increase Institute for Clinical Effectiveness and Health Policy (IECS), Buenos Aires, for men over the same period in developed countries Full list of author information is available at the end of the article 2010 Rubinstein et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License ), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.
Rubinstein et al. BMC Public Health 2010, 10:627 In Argentina, chronic non-communicable diseases populations, address high mortality and morbidity dis- account for more than 50% of the overall morbidity and eases, and include multi-level integrated efforts. Inter- mortality. In fact, the age-adjusted mortality rate of car- diovascular disease, including CHD and stroke was cardiovascular risk subjects, are also cost-effective but 206.4 per 100,000 (265.4 for men and 161.8 for women), usually require clinical involvement and more resources.
representing 34.2% of deaths and 12.6% of years of Moreover, recent studies have consistently shown potential life lost [Adjusted mortality rates for non- the cost-effectiveness of interventions that lower communicable diseases, as well as Potential Years of the burden of cardiovascular disease in developing Life Lost (PYLL) have declined steadily since 1987, while mortality rates of communicable, maternal, perina- The aims of this study were 1) to develop an analytical tal and nutritional conditions have remained relatively model to estimate the burden of acute CHD and stroke constant in the same 20-year period. Still, the adjusted attributable to modifiable cardiovascular disease risk fac- rate for non-communicable chronic diseases has been tors in Argentina, 2) to explore the costs of major cardi- five to six times the rate of communicable diseases in ovascular events, and 3) to calculate the cost- Argentina, and the absolute number of deaths is increas- effectiveness of different population-based and clinical ing due to the increasingly elderly population interventions in order to inform local policy makers on In common with many other Latin American coun- resource-allocation priority setting.
tries, Argentina falls into an intermediate mortalitygroup where the main risk factors for disease are hyper- tension, an elevated body mass index (BMI), alcohol We conducted a population-level comparative risk abuse and smoking [Elevated BMI is due to excess assessment for seven modifiable cardiovascular risk fac- calories and insufficient activity, and a large proportion tors to be included in a model to assess their impact on of hypertension is due to these same lifestyle risks in major cardiovascular events: acute myocardial infarction addition to a poor diet quality. Primary data describing (AMI), other non-infarction ischemic events and stroke.
the prevalence and distribution of cardiovascular risk We also estimated the individual and aggregate effect of factors in Argentina has recently been obtained through population-based and clinical interventions that might two different population-based sources: the 2005 Minis- modify the risks associated to these risk factors. These try of Health National Risk Factor Survey [and the interventions are supported by evidence in the literature Cardiovascular Risk Factor Multiple Evaluation in Latin for clinical efficacy and population effectiveness esti- America (CARMELA) [. There is strong evidence that mates that take into account detection and patient com- a 50% reduction in cardiovascular deaths can be attribu- pliance. Cardiovascular risk factors and interventions table to the reduction of just three modifiable risk fac- were modeled for the adult population over 35 years old tors, namely tobacco use, high blood pressure and since they are the usual target for most clinical interven- elevated cholesterol [In Latin America, the majority tions. Finally, cardiovascular events, Disability Adjusted of cardiovascular risk could be explained by tobacco Life Years (DALY) and interventions costs were derived.
use, abnormal lipids, abdominal obesity and high bloodpressure as shown in the recently published INTER- Selection of Risk Factors HEART Latin American study [. Most cardiovascular We selected specific risk factors that fulfilled the follow- diseases are preventable and there is evidence that sup- ing criteria: (1) Sufficient evidence was available on the ports the effectiveness of interventions to reduce the presence and magnitude of likely causal association with burden of cardiovascular disease through strategies that CHD and stroke from high-quality epidemiological stu- reduce risk factors. Unfortunately, strategies to manage dies, (2) available interventions existed to modify asso- cardiovascular conditions have been largely developed ciated risk, (3) data on risk factor prevalence was for high-income countries which may not be affordable available from the First Argentinean Survey of Risk Fac- to most of the developing world Although there tors (FASRF) or other nationally representative surveys has been widespread recognition of the benefit of cost- not subjected to selection bias.
effectiveness evaluation to inform national health sys- The seven modifiable risk factors selected were: 1) tems of priority settings, its potential has not been rea- high blood pressure (HBP), 2) high cholesterol, 3) over- lized in the vast majority of countries, including weight and obesity, 4) elevated fasting glucose level and Argentina . Nevertheless, cost-effective interventions type-2 diabetes mellitus, 4) tobacco smoking, 5) physical to prevent cardiovascular disease in developing countries do exist, but have not been widely applied. Specifically, Unfortunately, consumption of vegetables and fruits population and community-based interventions appear was ill-defined in the FASRF since the daily quantity of to be highly cost-effective when they reach large servings was not specified, for which we had to exclude Rubinstein et al. BMC Public Health 2010, 10:627 this measure for further analysis. Other specific indivi- reduction in disease-specific deaths that would occur if dual dietary factors that would meet criteria for causal risk factor prevalence had been reduced to zero. This is effects, such as intake of trans fat, low marine omega-3 known as the population-attributable risk (PAR) and (seafood), and low polyunsaturated fat (exchanged for measures the total effects of a risk factor (direct as well saturated fat) were also excluded because of lack of reli- as mediated through other factors). In order to estimate able data on their respective prevalences, after a thor- the PAR of each risk factor, we developed an epidemio- ough search of local surveys.
logical simulation model in Microsoft Excel(r), contain-ing the prevalence and distribution of risk factors according to each age and sex strata as observed in the Risk Factors exposure FASRF []. In this way, this matrix of 41,392 registries We obtained risk factor prevalence and distribution for from the FASRF, representing the Argentine population, each individual enrolled in the FASRF is a nationally was split into all possible combinations of risk factors.
representative survey including 41,393 subjects from all Additional risk for each combination was assumed to be districts of the country sampled through a probabilistic the product (multiplication) of the relative risk of the multi-stage process The prevalence of risk factors was risk factors involved [Finally, the baseline absolute obtained from self-reports obtained during an in-person risks for both fatal and non-fatal events for people with- interview that was subsequently validated with direct mea- out any of the selected risk factors were derived consid- sures in one district. For those subjects who reported not ering the overall risk, prevalence and additional risk to have ever measured her/his blood pressure (11.94%), associated to each combination. The global risk of death serum cholesterol (43.25%) or glycemia (23.49%) we con- (across all combinations and age groups) was calibrated sidered them as not having the risk factor in the survey.
against the overall number of deaths due to CHD and As this assumption could have underestimated their pre- stroke corresponding to Argentina in the year 2005 valence and population-attributable risk, we developed a Finally, the number of non-fatal events for each death logistic regression model to estimate the odds and prob- from CHD and stroke was extrapolated using the lethal- abilities for a subject with a certain demographic and risk ity rate from the Public Hospital registry corresponding factor profile to have an abnormal value in each of these to the year 2000 three risk factors. These new set of risk factors prevalence In addition to the estimation of the prevalence of were used as an alternative scenario in the sensitivity ana- cardiovascular risk factors and their associated relative lysis. STATA v8.0 was used to run these models.
risk, the spreadsheet contained the cost and disutility Etiological effects of risk factors on disease-specific associated with each event in order to obtain a deter- ministic estimate of the burden of disease, expressed in We obtained the relative risk for CHD and stroke attri- DALY and overall costs. A DALY is a summary mea- butable to each risk factor for each exposure category sure that combines years of life lost due to premature (since all risks were measured in categories in the death and years of life lived with disability One FASRF), based on published observational studies, sys- DALY can be thought of as one lost year of healthy tematic reviews or meta-analyses of epidemiological stu- life. DALYs were calculated based on the model devel- dies. In previous observational studies used for effect oped by Murray et al. sizes, the majority had adjusted for potential confound- The duration of disability was estimated by using the ing factors. Each relative risk used in our analysis repre- software DISMOD II Disability weights were sents the best judgment of the evidence for the effect of obtained from two Australian studies on burden of dis- risk factor exposure on disease-specific mortality. The ease []. For the calculation of years of life lost due etiological effect sizes along different age-strata and gen- to premature death, we used a life expectancy at birth der are shown in Table of 80 and 82.5 for men and women, respectively, as Disease-specific deaths recommended for global comparisons in the Global Bur- The number of deaths by CHD (ICD-10 codes I20×, I24× den of Disease study . Finally, years of life lost due to and I25× for non-infarction ischemic events and I21× premature death were obtained from National death and I22× for AMI) and stroke (ICD-10 codes I60-I61, registries and years of life lived with disability were I63-I64) were obtained from the National Directorate of obtained by multiplying the estimated number of non- Health Statistics of the Argentine Ministry of Health fatal events by each disability weight, for each age gen-der strata Estimating mortality and disability attributable to risk In order to estimate the PAR associated to each risk factor, a new estimation of deaths, non fatal events, For each risk factor and for each disease causally asso- DALY and costs of CHD and stroke were calculated.
ciated with its exposure, we computed the proportional These estimations were obtained multiplying the basal Rubinstein et al. BMC Public Health 2010, 10:627 Table 1 List of relative risks included into the model Relative risk for coronary heart disease High blood pressure (m) High blood pressure (w) High glycemia (m) High glycemia (w) High cholesterol (m) High cholesterol (w) Current smoker (m) Current smoker (w) Former smoker (m) Former smoker (w) Non-sedentary life style (m) Non-sedentary life style (w) Relative risk for stroke High blood pressure (m) High blood pressure (w) High glycemia (m) High glycemia (w) High cholesterol (m) High cholesterol (w) Current smoker (m) Current smoker (w) Former smoker (m) Former smoker (w) Non-sedentary life style (m) Non-sedentary life style (w) m: men; w: women.
absolute risk by the product of the relative risks involved language , in which we performed 1,000 iterations of in each combination stratum, assuming a relative risk the prevalence for each combination of risk factors equal to 1 for the index risk factor, weighted by its assuming a binomial distribution. Therefore, a new abso- respective prevalence. Overall deaths, non fatal events, lute risk was obtained in each iteration, and new estima- DALY and costs between the estimation for Argentina in tions of total deaths, non fatal events, DALY and costs 2005 and the new estimation without the index risk fac- were obtained. Finally, we used the empirical PAR distri- tor, was assumed to be the PAR attributable to that parti- bution to estimate the 95% confidence interval (95%CI) cular risk factor. We programmed a macro using Python using the percentile method.
Rubinstein et al. BMC Public Health 2010, 10:627 Definition and Selection of Interventions selected cities in Argentina to make bakers reduce salt Different population-based and clinical interventions to in bread by using special salt dispensers . This inter- reduce cardiovascular disease burden were explored con- vention could imply a population-wide reduction of 1.33 sidering not only the evidence of efficacy and effectiveness mmHg of systolic blood pressure per person and 1% of but also the feasibility to be implemented in Argen- the PAR of CHD and stroke tina. Relative risk reductions of the interventions were Mass Media Campaign to promote tobacco cessation adjusted by population effectiveness measures taking into This program of the National Ministry of Health account target population coverage as well as patient com- involves an annual campaign through four TV spots, six pliance. All interventions have a time horizon of 5 years radio spots and written material in major newspapers, after which maximum population effectiveness is assumed.
magazines and public spaces. Costs were retrieved from The evidence about population effectiveness of mass data from previous campaigns of the National Ministry media campaign targeted to the promotion of physical of Health. This intervention would reduce the preva- activity , salt reduction in food control of lence of smoking by 7% overweight and obesity and promotion of healthy Individual (clinical) interventions habits was non-conclusive, and hence these inter- Treatment of high blood pressure Interventions ventions were not included in the model. On the other involved lifestyle change promotion and pharmacological hand, evidence on the effectiveness of media campaigns therapy to achieve blood pressure control (SBP/DBP less against smoking was generally strong and local programs than 140/90). Target population was composed of adults had already been implemented [. Efficacy of inter- over 35 years old with the diagnosis of high blood pres- ventions were modeled as a relative risk reduction or by a sure and no treatment (over 1.3 millions of Argentine reduction on risk factor prevalence. Effect sizes and joint population representing 8.2% people older than 35 years effect of interventions used in the analysis were based on old), estimating for this intervention a relative risk systematic reviews of randomized trials and meta-analysis, reduction of 44% for CHD and 49% for stroke ]. We when possible. Intervention effects with their correspond- assumed that 40% of the population would take one ing relative risks estimates are shown in Table drug, 40% two drugs and 20% three or more drugs. The drugs and daily doses evaluated were hydrochlorothia- Lowering salt intake in the population through redu- zide (25 mg), atenolol (50 mg), enalapril (10 mg), and cing salt in bread A program involving the cooperation amlodipine (10 mg), and the treatment mix was 50% of between the Government, consumer associations and the population taking thiazides, 20% atenolol, 20% the Bakery Chambers in an effort to reduce 1 gram of angiotensin-converting enzyme inhibitor and 10% amlo- salt per 100 grams of bread. Argentina has an average dipine . The same efficacy for each drug category individual consumption of 12 grams of salt per day, 3.4 was also assumed. Analysis indicated that these inter- grams coming from bread. Local experiences showed ventions, with a 50% rate of disease detection and 50% that it is possible to reduce the amount of salt in bread drug compliance as indicated by the Canadian Hyper- without being detected as less palatable. At present, tension Guidelines would reduce PAR of cardiovas- there is a pilot training program implemented in cular disease and stroke by 8%.
Table 2 Effectiveness of selected interventions Population based interventions Mass Media Campaign promote tobacco cessation Reduction of current smoker Reducing salt in bread Clinical interventions Bupropion treatment for tobacco cessation Annual cessation rate: 28%.
Pharmacological high blood pressure treatment* For CHD: RR = 0.66 For stroke: RR = 0.51 Pharmacological high cholesterol treatment with atorvastatin For CHD: RR = 0.77 For stroke: RR = 0.81 Treatment with four drugs (Polypill strategy) for people with an absolute cardiovascular risk of For CHD: RR = 0.34 more than 20% at 10 years For stroke: RR = 0.32 CHD: Coronary Heart Disease, RR: relative risk.
* Include: atenolol, enalapril, amlodipine and hidroclorothiazide.
** Include: aspirine, enalapril, amlodipine and atorvastatin.
Rubinstein et al. BMC Public Health 2010, 10:627 Treatment of high cholesterol This intervention therapeutic procedures) for AMI, other non-infarction involved promotion of low-cholesterol diet and use of ischemic events such as unstable angina and stroke were statins (atorvastatin 10 mg, 20 mg and 40 mg for 50%, first identified. For each event the quantities and unit 40% and 10% of the target population, respectively), prices of inputs were retrieved from hospital databases according to local estimates and assumptions. Target and other local sources as well as expert opi- population was adults over 35 years old with high choles- nion when necessary. The quantities of each input iden- terol and no treatment (almost one million people repre- tified were assessed and multiplied by the unit price of senting 5.2% of people older than 35 years old).
each item to obtain the unit cost of each resource.
Achieving a cholesterol target of less than 240 mg/dl, (6.2 Finally, the total cost of the acute event resulted from mm/l) would provide an estimated reduction of 8% of the the addition of all of the identified consumed resources PAR of CHD and stroke with a 50% detection and 50% in each category.
drug compliance rate according to ATP III Costs of interventions Tobacco cessation therapy Motivational interventions Costs included program-level expenses associated with from health professionals and drug therapy with bupro- management of the interventions (i.e. administration, pion for a 2-month period (300 mg per day) would training and information, dissemination by multiple result in an estimated reduction of 4% of the PAR of media sources) and patient-level costs (i.e. primary care CHD and stroke In most studies with bupropion visits, ancillary tests and drugs). The quantities of each for tobacco cessation, the annual quitting rate of smo- input required were assessed and multiplied by the unit kers was 28% vs. 12%, as compared to placebo price of each input for the 5 year intervention imple- According to a recent national survey of tobacco preva- mentation period. The quantity of patient-level resource lence, only 11% of total smokers in Argentina were will- inputs for each intervention (i.e. inpatient hospital days, ing to quit smoking and therefore were considered the doctor visits, tests, drugs) were identified from local or target population for this intervention []. According international published data if available or expert opi- to these estimates, the spontaneous annual cessation nion should the former not be available. Costs of drugs rate would be 1.32% (12% of the 11% of smokers willing were calculated using a mix of blood pressure lowering to quit) that would raise to 3.08% with bupropion (28% drugs composed of 50% hydrochlorothiazide, 20% ateno- × 11%), since we would expect a prevalence reduction lol, 20% enalapril and 10% amlodipine, according to a of 1.76% (3.08%-1.32%).
published local study Cost of blood pressure lower- Treatment based on a population absolute risk ing drugs, atorvastatin and bupropion as well as other approach (Polypill strategy) Since the "Polypill" is not input costs and expense data were extracted from local yet in Argentine markets, we designed a pharmacologi- sources [. Other cost data were obtained from the cal therapy with 4 pills (hydrochlorothiazide 25 mg, ena- Health Care Costs Database from the Institute of Clini- lapril 10 mg, atorvastatin 10 mg and aspirin 100 mg), cal Effectiveness and Health Policy ]. A list of costs prescribed to people older than 35 years old with an and sources of the interventions and selected health estimated combined risk of a cardiovascular event over events is depicted in Table the next decade above 20%, based on the data from the Cost of clinical interventions included, in addition to FASRF. This intervention would imply a relative risk their specific costs of visits, tests and drugs, 290 coun- reduction of CHD of 66% (RR = 0.34) and of stroke of trywide training workshops on cardiovascular risk detec- tion, assessment and control targeted to 8,639 general Assuming that at least 50% of the target population is practitioners from the public and private health sector, reached by the intervention, a 50% patient compliance along the 5 year period of the intervention, with peri- rate with treatment for this group, and 70% of provider odic boosters through email and postal mail. Except compliance due to a presumed raised awareness of risks when explicitly stated, costs related to labor, equipment, for both subjects, the Polypill strategy would result in a capital, overhead or joint costs were regarded as exist- population effectiveness of 17,5%. Relative risks for ing, ongoing, or common to all interventions and there- CHD and stroke for individuals from this high-risk sub- fore were excluded in the calculation. We also excluded group were estimated by using the beta coefficients costs of accessing health interventions that would from the Framingham Heart Study include the resources used by patients and their familiesto obtain an intervention (transport costs) as well as Estimating costs of acute cardiovascular events and productivity gains or losses, as the study was conducted from a purchaser perspective. All costs were calculated Costs of acute events in Argentine pesos for the year 2007, requiring in some Cost categories (i.e. inpatient hospital days, doctor visits, cases the use of Health and General CPI to adjust for tests, drugs and ancillary services, and diagnostic and annual inflation, and finally converted and expressed Rubinstein et al. BMC Public Health 2010, 10:627 Table 3 Interventions and related health events summary then compared to the estimation without the interven- tion. In addition, the annual cost of the intervention was Event cost per hospital admission imputed for the year analyzed. For each intervention, Coronary Heart Disease the Incremental Cost-Effectiveness Ratio (ICER) of the interventions compared to no intervention was mea- sured as cost per averted DALY. Effect sizes and joint Mass Media Campaign promote tobacco cessation* effect of interventions used in the analysis were based Reducing salt in bread† on systematic reviews of randomized trials and meta- Individual interventions (yearly cost per person‡) analysis, when possible.
Pharmacological high blood pressure treatment To translate changes in the risk of age and sex specific Pharmacological high cholesterol treatment cardiovascular disease events into changes in population Bupropion treatment for tobacco cessation health quantified in terms of DALY, we used a standard Modified Polypill strategy methodology described elsewhere There is no universal criterion that defines a threshold I$: international dollars. PPP conversion rate (2007) 1.55 Argentinean peso = 1I$, cost-effectiveness ratio, above which an intervention * Ten years duration of campaign, with discounting (3% annual rate).
would not be considered cost-effective. We chose to use † Assuming 54 meeting of 30 bakers each (around 800-600 bakers), with guidelines specifically intended for international compar- discounting (3% annual rate).
isons, as proposed by the Commission on Macroeco- ‡ Includes health center visits, drug and lab test costs. Programmatic costswere not included (I$ 1,194,067.52).
nomics and Health, which defines interventions with an Note: Cost of blood pressure lowering drugs, atorvastatin and bupropion as ICER that is less than three times Gross Domestic Pro- well as other input costs and charges data were extracted from local sourcesOther cost data were obtained from the Health Care Costs Database duct per capita as a "cost-effective" intervention and as by the Institute of Clinical Effectiveness and Health Policy "very cost-effective" if ICER is less than the GDP percapita [. Argentina's GDP per person in 2007 was into international dollars using the Purchase Power Par- estimated in I$ 13,255.09 ity conversion rate AR$ 1.55 = 1 I$ . The discount-ing of long term costs was performed at a 3% rate.
Uncertainty and sensitivity analysis We also did a probabilistic, multivariate sensitivity ana- Since Argentina's healthcare system consists of a multi- lysis using Monte Carlo simulation of 1,000 ran- tier system divided in three large sectors: public, social domly selected sets of variables, to assess the effects of security and private, we incorporated the perspective of uncertainty in the prevalence of risk factors, population the whole Argentine healthcare system as a purchaser of attributable risk and effect sizes of interventions. In health services.
addition, an undiscounted scenario was considered forcosts and DALY, and a non age-weighted scenario was Calculating cost-effectiveness of interventions also analyzed for DALY.
Cost-effectiveness analysis considers the costs andeffects of adding new interventions to current practice or the cost of replacing an existing intervention with We estimated a lethality rate of 11.9% in men and 18% another targeting the same condition. In order to esti- in women; and 17.4% in men and 18.9% in women, for mate the reduction in disease burden related to the CHD and stroke, respectively [According to these reduction of cardiovascular disease, we built a model to estimates, about 263,025 annual acute CHD and stroke predict the burden associated with specific diseases or events would be expected, representing an annual cost risk factors to develop disease. We calculated the effect of I$ 1,036,506,958. More than 60% of total events and of interventions in our model, assuming that all reduced costs are accounted by men. Table shows the estima- the relative risk associated with the presence of each tion of the overall number of annual cardiovascular cardiovascular risk factor. In the case the effect of the events in Argentina, burden of disease and costs of intervention was a reduction of the prevalence (i.e.: events. As observed, more than 600,000 DALYs and tobacco cessation), a new relative risk was estimated as almost 400,000 YPLL were lost in 2005 due to CHD a proportional combination of the relative risk asso- ciated with the risk factor (for the proportion of peoplethat were still smokers) and the relative risk of those Burden of Disease attributable to modifiable that no longer had that risk factor (i.e.: former smokers).
cardiovascular Risk Factors Finally, the model translated these changes into a new Population attributable risks, costs of events and DALY estimation of cardiovascular events, overall costs and lost to cardiovascular disease for the overall risk factors DALY lost, specific for age and sex. This estimation was and for each single modifiable risk factor selected, can Rubinstein et al. BMC Public Health 2010, 10:627 Table 4 Estimation of total cases, costs and burden of costs and costs per beneficiary, health effects in terms of disease of acute CHD and stroke DALY averted (non age-weighted and 3% discounted), percent of DALY saved due to cardiovascular disease Total AMI events [%] and average cost-effectiveness ratio for each in I$ per DALY averted. Two interventions were cost-saving: low- Total non-infarction events [%] ering salt intake in the population through reducing salt in bread and treatment targeted to persons with an Total stroke events [%] absolute risk above 20% in 10 years (modified polypill strategy). Moreover, the implementation of the polypill strategy was also associated with almost a 2% decrease Total costs* I$ [%] 667,728,147 368,778,811 1,036,506,958 in DALY lost to cardiovascular disease. On the other hand, the impact of reducing salt in bread was more Total DALY† [%] limited (0.11% of decrease of DALY lost) due in part to the lower extension and magnitude of this intervention.
Total PYLL‡ [%] Two interventions had very acceptable ICER: 1) drug therapy for high blood pressure in hypertensive patients 95%CI: 95% confidence interval, AMI: acute myocardial infarction, DALY: not yet under going treatment with an ICER of I$ 2,908 disability-adjusted life years, I$: international dollars, PYLL: Potential Years ofLife Lost.
per DALY saved and an annual reduction of 2.3% of * Only direct medical costs by hospitalization were considered. Costs are cardiovascular disease burden; and 2) mass media cam- measured in 2007 international dollars (I$).
paign to promote tobacco cessation amongst smokers, † With discounting (3% annual rate), without age weight.
with an ICER of I$ 3,186 per DALY saved (0.11% of car- ‡ With discounting (3% annual rate).
diovascular disease burden). An additional intervention, be seen in Tables and respectively. All risk factors lowering cholesterol with statins (I$ 14,432 per DALY together explained 75% of fatal and non-fatal acute saved), was considered cost-effective according to the CHD and stroke events, 82,4% of acute CHD events guidelines mentioned above. Finally, one intervention, (84.0% in women) and 62.4% of strokes (66,6% in men).
tobacco cessation with bupropion (I$ 59,433 per DALY Similarly, modifiable risk factors explained 75,5% of saved) was not found to be cost-effective. This is in part costs of acute events and 70.7% of DALY lost. The most because bupropion is much more expensive than blood important single risk factor was high BP, explaining 37% pressure lowering drugs and also because, as it is not of all CHD and strokes and one-third of all DALY lost currently covered in the public sector, the government in 2005. The rest of the risk factors have similar attribu- does not usually exert its purchasing power to get lower table burden in term of CV events, ranging between prices. Following local surveys, we assumed that only 13,9% (high glycemia) to 18,1% (physical inactivity). (see 11% of the population of smokers would be willing to quit smoking each year and consequently start on a pro-gram, the population impact of tobacco cessation ther- Cost-effectiveness of selected interventions apy was much smaller than expected.
Table summarizes the results of economic evaluation Figure shows the ICER of the six distinct interven- of the 6 distinct interventions giving their total annual tions along the cost effectiveness plane with their Table 5 Proportional Burden of Disease and costs attributable to all cardiovascular risk factors potentially modifiable Population-attributable Fraction % (95%CI) Total AMI events (both fatal and non-fatal) Total non-infarction events (both fatal and non-fatal) Total stroke events (both fatal and non-fatal) Total events (both fatal and non-fatal) Results of 1,000 iterations, both sexes.
95%CI: 95% confidence interval, AMI: acute myocardial infarction, DALY: disability-adjusted life years, PYLL: Potential Years of Life Lost.
* Only direct medical costs by hospitalization were considered.
† With discounting (3% annual rate), without age weight.
‡ With discounting (3% annual rate).
Rubinstein et al. BMC Public Health 2010, 10:627 Table 6 Proportional Burden of Disease and costs attributable to each cardiovascular risk factor potentially modifiable Population-attributable Fraction % (95%CI) Total AMI events (both 22.5 (22.2 - 22.8) 20.9 (20.4 - 21.3) 38.5 (37.9 - 39.1) 25.1 (24.4 - 25.8) 13.9 (13.2 - 14.5) 20.9 (20.6 - 21.2) fatal and non-fatal) Total non-infarction events 18.5 (18.1 - 18.8) 18.6 (18.2 - 19.0) 40.9 (40.1 - 41.5) 26.2 (25.2 - 27.0) 14.8 (14.0 - 15.6) 21.6 (21.2 - 22.0) (both fatal and non-fatal) Total stroke events (both 10.8 (10.5 - 11.2) 11.7 (11.3 - 12.0) 32.7 (32.0 - 33.3) 13.2 (12.3 - 14.0) 13.0 (12.7 - 13.3) fatal and non-fatal) Total events (both fatal 16.9 (16.5 - 17.2) 16.7 (16.3 - 17.0) 37.0 (36.3 - 37.5) 18.0 (17.4 - 18.5) 13.9 (13.1 - 14.6) 18.1 (17.7 - 18.3) 17.3 (17.0 - 17.6) 17.1 (16.7 - 17.4) 37.3 (36.6 - 37.9) 18.9 (18.3 - 19.5) 13.9 (13.2 - 14.7) 18.4 (18.1 - 18.7) 16.1 (15.7 - 16.4) 13.8 (13.5 - 14.0) 36.6 (36.1 - 37.0) 13.4 (13.0 - 13.7) 13.6 (13.0 - 14.2) 15.5 (15.2 - 15.7) 16.6 (16.3 - 16.9) 15.1 (14.8 - 15.4) 37.5 (36.9 - 38.0) 16.6 (16.1 - 17.1) 13.9 (13.2 - 14.6) 16.9 (16.6 - 17.1) Total AMI events (both 26.7 (26.2 - 27.1) 24.7 (24.1 - 25.2) 37.0 (36.4 - 37.6) 19.0 (18.3 - 19.6) 12.4 (11.8 - 13.0) 20.2 (19.8 - 20.6) fatal and non-fatal) Total non-infarction events 23.1 (22.6 - 23.5) 22.3 (21.8 - 22.8) 40.1 (39.3 - 40.8) 19.3 (18.6 - 20.0) 13.5 (12.8 - 14.2) 20.6 (20.2 - 21.2) (both fatal and non-fatal) Total stroke events (both 14.2 (13.7 - 14.7) 16.7 (16.0 - 17.3) 35.2 (34.3 - 36.0) 13.5 (12.6 - 14.4) 12.2 (11.7 - 12.6) fatal and non-fatal) Total events (both fatal 21.6 (21.1 - 22.0) 21.4 (20.9 - 21.9) 37.4 (36.7 - 38.1) 14.7 (14.1 - 15.2) 13.1 (12.4 - 13.8) 17.8 (17.4 - 18.2) 22.0 (21.6 - 22.5) 21.7 (21.2 - 22.2) 37.6 (36.8 - 38.2) 15.2 (14.7 - 15.8) 13.1 (12.4 - 13.8) 18.1 (17.7 - 18.6) 22.1 (21.6 - 22.6) 20.5 (20.1 - 20.9) 38.7 (38.1 - 39.3) 12.1 (11.7 - 12.6) 13.0 (12.3 - 13.7) 15.6 (15.3 - 16.0) 21.6 (21.1 - 22.1) 20.7 (20.2 - 21.1) 38.7 (38.1 - 39.3) 13.4 (12.9 - 13.8) 13.1 (12.4 - 13.8) 16.5 (16.2 - 16.9) Total AMI events (both 11.9 (11.5 - 12.3) 11.0 (10.5 - 11.6) 42.4 (41.1 - 43.5) 41.0 (39.2 - 42.7) 17.7 (16.2 - 19.1) 22.6 (22.2 - 23.1) fatal and non-fatal) Total non-infarction events 10.5 (9.9 - 11.2) 42.5 (40.7 - 44.0) 41.1 (38.4 - 43.1) 17.7 (15.8 - 19.6) 23.7 (23.1 - 24.2) (both fatal and non-fatal) Total stroke events (both 29.9 (28.9 - 30.8) 12.8 (11.5 - 14.2) 14.0 (13.6 - 14.3) fatal and non-fatal) Total events (both fatal 36.2 (34.9 - 37.3) 23.8 (22.4 - 24.9) 15.2 (13.8 - 16.8) 18.6 (18.1 - 19.0) 36.8 (35.5 - 37.9) 25.5 (24.1 - 26.7) 15.5 (14.0 - 17.0) 19.0 (18.6 - 19.5) 10.5 (10.1 - 10.9) 34.5 (33.8 - 35.2) 14.5 (13.9 - 15.1) 14.1 (13.2 - 15.1) 15.3 (15.0 - 15.6) 10.4 (10.0 - 10.8) 36.0 (35.0 - 36.9) 20.7 (19.7 - 21.6) 14.8 (13.6 - 16.0) 17.3 (16.9 - 17.6) Results of 1000 iterations, basal case, both sexes.
95%CI: 95% confidence interval, AMI: acute myocardial infarction, DALY: disability-adjusted life years, PYLL: Potential Years of Life Lost.
* Only direct medical costs by hospitalization were considered.
† With discounting (3% annual rate), without age weight.
‡ With discounting (3% annual rate).
respective probability distribution. The shaded area cor- In all circumstances, the ranking as well as the magni- responds to the cost-saving interventions.
tude of each intervention remained the same. Finally,the results of the probabilistic sensitivity analyses to Sensitivity Analyses estimate the uncertainty surrounding the central esti- We examined the effect of a change in the PAR estimate mates of each intervention is expressed through the 95% of the overall risk factors selected along a reasonable CI showed in Table range of probabilities, by creating alternative scenarioswith different prevalence and distributions of risk fac- tors. We also explored undiscounted and age-weighted Our study analyzed the FASRF at individual level to esti- DALY as compared to the base case scenario with mate the burden of cardiovascular disease in Argentina DALY discounted at 3 percent and non age-weighted.
attributable to modifiable risk factors in order to Rubinstein et al. BMC Public Health 2010, 10:627 Table 7 Cost effectiveness analysis of selected interventions Net Total costs * Reducing salt in bread Treatment targeted to persons with an absolute risk above 20% in 10 years (polypill strategy Pharmacological therapy for high blood pressure Mass Media Campaign to promote tobacco cessation Pharmacological therapy of high cholesterol Therapy with Bupropion for tobacco cessation 95%CI: 95% confidence interval, I$: international dollars - PPP conversion rate (2007) 1.55 Argentinean peso = 1 I$, DALY: disability adjusted life years, ICER:incremental cost-effectiveness ratio. The ICERs express the results of 1,000 iterations.
* Net Total costs are calculated as Total costs minus the corresponding averted event costs. All costs are measured in 2007 International dollars (I$).
‡ Derived from bootstrapping techniques.
model the impact of some preventive interventions to factors explained between 14% and 18%. WHO reducing this burden, as well as estimating their cost- recently addressed the importance of chronic disease effectiveness. Based on our data, the PAR of all the prevention as a neglected health issue in LMIC; risk factors analyzed explained more than 75% of the achievement of the global goal to reduce chronic dis- acute CHD events and strokes in men and women.
ease death rates by 2% every year would avert 36 mil- Only high blood pressure explained more than one- lion deaths between 2005 and 2015 Achieving third of the events while each one of the other risk this target would also save almost 10% of the expected Figure 1 Interventions along the cost-effectiveness plane. Costs are expressed in International dollars (I$, 2007).
Rubinstein et al. BMC Public Health 2010, 10:627 loss in national income in these settings . Consider- above, partly because we only included a series of one- ing the growing burden of cardiovascular disease and off meetings with bread makers from large cities, and costs in developing countries, especially for transitional also because we used a lower effect size.
countries like Argentina, this study is critical to pro- In regards to the intervention oriented to reduce high vide local decision-makers with information about car- blood pressure and high cholesterol our ICER were diovascular disease burden. Furthermore, by comparing remarkably higher than those reported by Murray et al.
the relative costs and health effects of interventions for ]. The causes of this apparent discrepancy are two- preventing cardiovascular disease, we can focus policy fold: firstly, the counterfactual scenario designed by debate concerning the trade-offs or opportunity costs Murray, based on the WHO-CHOICE methodology of financing one intervention over another.
entails lifting the constraints of the current mix of inter- Establishing the cost-effectiveness of preventive inter- ventions, using a null scenario of no costs and no inter- ventions for cardiovascular disease in developing coun- ventions as a starting point, as opposed to our try contexts is not straightforward, due to both the assumption that almost half of Argentine population paucity of existing evidence, and because there is no were already receiving treatment; based on the data of universally agreed threshold for considering the cost- the FASRF; and secondly, our cost estimates are consid- effectiveness of an intervention to be ‘too high' or erably higher, which reflects the fact that key interven- ‘right'. What is acceptable to health and finance deci- tion resource inputs in Argentina (including human sion-makers depends largely on the country context.
resources, secondary care and drugs) are much more The Disease Control Priorities Project (DCPP), has iden- expensive than the regional average. The addition of tified several chronic disease interventions as cost-effective individual-level interventions with a multi-drug regimen at a cost of below US$1,000 per DALY ]. However, on the basis of opportunistic contact with the health the affordability of interventions will vary significantly service, as reported by Gaziano et al. [has been esti- across countries, even among a group of interventions mated at US$ 2.93 per capita in a country like Argen- believed to be cost-effective in the global sense. More- tina, but would save a further 50,000 lives over a over, sensitivity analysis done as part of the cost effec- 10-year period.
tiveness analysis modeling for the DCPP showed that Our analyses have shown that the multidrug regimen the cost-effectiveness of public education campaigns at of four highly effective drugs (polypill strategy, with an the population level could be very good or far less annual cost of I$ 101 or I$ 32 per capita in 2007) could favorable depending on how much it cost to reach peo- lead to cost-saving prevention and treatment for sub- ple using a reasonable range of costs. In addition, even jects with an absolute risk above 20% in 10 years, with a very inexpensive intervention might not be worth 2% of reduction in DALY lost to cardiovascular disease implementing if it targets a chronic disease with low even considering a population effectiveness of less than prevalence in a given country or region.
20% the potential targeted population. Other treatment In an earlier analysis, Murray et al. modeled cutpoints for this intervention, where ICERs would likely selected population-based and individual health inter- be far higher, were not evaluated, including subjects ventions to lower high blood pressure and high choles- with lesser CV risk or subjects over age 55 as originally terol in the epidemiological contexts of developing proposed by Wald and Law (53), The study of Gaziano countries. The authors found that all interventions were et al about the cost-effectiveness of the Polypill regimen highly cost-effective in the sub-region of the Americas modeled an ICER of less than US$ 900 for a similar to which Argentina belongs.
risk population in the Latin American region.
More recently, Asaria et al., assessed the financial According to the threshold adopted by World Health costs and health effects of a voluntary reduction in the Organization, an intervention that saves one DALY for salt content of processed foods by manufacturers plus a less than three times the gross domestic product (GDP) mass media campaign to encourage dietary change in 23 per capita is considered cost-effective, while one that selected low and middle income countries, including saves a DALY for less than one GDP per capita is Argentina. They estimated that a 15% reduction in diet- deemed very cost-effective . As Argentina's GDP per ary salt intake in Argentina would save 60,000 lives over person in 2007 was I$13,255 (US$ 6,644) esti- the period 2006-2015 at a cost of US$ 0.14 per capita mated ICER of all interventions analyzed except tobacco (equivalent to AR$ 16.7 million for a population of cessation with bupropion (I$ 59,433 per DALY saved) Argentina (38 millions in 2005) fell well within the ‘very cost-effective' or cost-effective As compared to these previous studies [our category. In fact, two interventions - reducing salt intake intervention to decrease salt intake in bread was cost- in bread and the absolute risk approach therapy with saving, although both our health impact and cost esti- four drugs - were cost-saving. In an earlier study of mates were appreciably lower than those summarized cost-effectiveness of cardiovascular interventions in Rubinstein et al. BMC Public Health 2010, 10:627 Buenos Aires, in which we used a counterfactual sce- any benefit from the pharmacological interventions in nario of no costs and no interventions, most of our the population that is already receiving treatment, as if interventions were very cost-effective [Should we they were appropriately controlled, which is not true.
have used a counterfactual scenario based on what the No matter that this is aligned with our conservative public health sector was actually spending on the care estimates, the ICER of high blood pressure or high cho- of cardiovascular disease we would have obtained much lesterol therapy look less attractive in terms of reduc- lower ICER or even cost-saving interventions like we tion of disease burden or cost-effectiveness; 6) as in all have found in this study.
modeling studies, our study synthesized data from many In addition, the potential budget impact of the imple- sources and used several assumptions in the design of mentation of the four cost saving or cost-effective inter- the model. As real life decision making tools, these ventions mentioned above was in the range of I$ 194 types of model-based studies are explicit analyses to million in 2005. This expenditure would be partly offset help health priority setting, and are not a "search of the by the savings obtained through avoided cardiovascular ultimate truth"; and 7) some inputs, as it is also com- acute events. Moreover, the financing of these interven- monplace in modeling studies, were derived from inter- tions, even considering low population effectiveness national sources. This was done mainly with relative according to our conservative scenario, could reduce at measures such as relative risks of different cardiovascu- least 7% the cardiovascular disease burden with its con- lar risk factors, or relative effects of interventions, which sequent health, economic and social impact.
on the other hand, are widely thought to be more gen- Some limitations of the present work are important to eralizable from setting to setting.
be acknowledged. 1), the risk factors included in themodel were limited to those that were specifically addressed in the national survey as they were specifi- Overall, evidence exists to conclude that there are cally defined. In this regard, concerning the intake of important clinical as well as economic consequences of fruits and vegetables, we were bound to the two defined cardiovascular disease, consequences that are not only options as posed in the specific question of the survey: important to the individual and his/her family but also more or less than five servings a week (rather than to the economy at large. At the same time, there are more or less than five servings a day), which is clearly severe gaps in the evidence that call for more research inappropriate based on WHO recommendations into the avoidable burden of cardiovascular disease, in This limitation prompted us to exclude this risk factor particular for developing countries. Despite the increas- for further analysis. Other risk factors related to diet ing burden of cardiovascular disease in Argentina, rank- such as trans fat, low marine omega-3, and low polyun- ing first over the last decades as a cause of mortality saturated fat were also excluded due to lack of popula- and morbidity, national health programs and policies are tion based data stressing the importance of obtaining still focused on interventions aimed to tackle commu- future national-level data on these and other dietary nicable diseases or perinatal or childhood conditions, risk factors for future analysis; 2) since the prevalence overlooking actions and programs targeted to lifestyle of risk factors was obtained from self-reports of partici- and nutritional changes in the population at large or pants and not from direct measures, they were defined pharmacological interventions to reduce cardiovascular dichotomously or categorically (as having or not having disease burden in high risk people.
the risk factor) for the calculation of the PAR. This In conclusion, most of the interventions selected were implies that the risk of a particular risk factor behaves cost-saving or very cost-effective according to WHO like an "all or none" phenomenon, which is obviously standards. Moreover, the financing of these interven- not true given the continuous nature of this risk in all tions could reduce at least 5% the cardiovascular disease of the selected risk factors. In this regard, estimating burden with its consequent health, economic and social the theoretical minimum risk exposure distribution impact. This study aims to inform policy makers on would be a more appropriate method should this had resource-allocation decisions to reduce the burden of been possible; 3) we have just modeled interventions CVD, especially for middle-income developing countries that either had been tested in pilot studies (i.e.: reducing like Argentina.
salt in bread) or were considered key data to model theintervention (i.e.: just 11% of total smokers in Argentina as potential quitters to model the impact of tobacco 95%CI: 95% confidence interval; AMI: acute myocardial infarction; CEAC: cost- cessation with bupropion); 4), we did nor model poten- effectiveness acceptability curves; CHD: coronary heart disease; DALY: tial side effects of the multidrug intervention. Ignoring disability-adjusted life years; DCPP: Disease Control Priorities Project; GDP:Gross Domestic Product; I$: international dollars; ICER: Incremental Cost- side effects in the analysis could overestimate the ICER Effectiveness Ratio; LMIC: low and middle-income countries; PAF: Population of the polypill strategy. 5) our study does not assume Attributable Fraction; PYLL: Potential Years of Life Lost Rubinstein et al. BMC Public Health 2010, 10:627 Rubinstein A, Belizan M, Discacciati We would like to thank Daniel Comandé, our librarian, for his invaluable contribution to this work.
Lisandro Colantonio is Fogarty International Clinical Research Fellow 2009- International journal of technology assessment in health care 2007, 2010 through the International Clinical Research Fellows Program at Vanderbilt Institute for Global Health, Vanderbilt University in Nashville, TN Murray CJ, Lauer JA, Hutubessy RC, Niessen L, Tomijima N, Rodgers A, (R24 TW007988).
Lancet 2003, 361:717-725.
1Institute for Clinical Effectiveness and Health Policy (IECS), Buenos Aires, Gaziano TA, Opie LH, Weinstein MC: Argentina. 2Division of Family and Community Medicine, Hospital Italiano de Buenos Aires, Buenos Aires, Argentina. 3Programa de Prevención del Infarto Lancet 2006, 368:679-686.
en Argentina (PROPIA), Universidad Nacional de La Plata, Buenos Aires, Nissinen A, Berrios X, Puska P: Argentina. 4Centro de Endocrinología Experimental y Aplicada (CENEXA), Universidad Nacional de La Plata, Buenos Aires, Argentina.
Bull World Health Organ 2001, 79:963-970.
Estadísticas vitales. Información básica - 2005: Buenos Aires: Dirección de Authors' contributions Estadísticas e Información de Salud. Secretaría de Políticas, Regulación y AR conceived the study, coordinated the teamwork and participated in its Relaciones Sanitarias. Ministerio de Salud y Ambiente de la Nación 2006.
design and analysis. He also led the writing of the manuscript. LC carried Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB: out the modeling in Excel and Python, and participated in literature search and drafting of the manuscript. AB participated in the literature search, Circulation 1998, 97:1837-1847.
carried out the assessment of measures of effect for conditions and risk Egresos de establecimientos oficiales según variables seleccionadas.
factors, and for effectiveness of interventions. AA and KK also participated in República Argentina - Año 2000: Buenos Aires: Secretaría de Políticas, the literature search and abstraction of measures of effects. APR and SGM Regulación y Relaciones Sanitarias. Ministerio de Salud 2003.
participated in the design of the study and made substantial statistical contributions. LG participated in the modeling design in Python. JC took Bull World Health Organ 1994, 72:429-445.
responsibility for assessing economic disease impacts and costing of the Barendregt JJ, Van Oortmarssen GJ, Vos T, Murray CJ: interventions, and helped in the cost effectiveness analysis. FA made substantial contributions in the CEA. All authors read and approved the final Popul Health Metr 2003, 1:4.
Victorian Burden of Disease Study. Mortality and morbidity in 2001.
Book Victorian Burden of Disease Study. Mortality and morbidity in 2001 City: Competing interests Public Health Group, Rural and Regional Health and Aged Care Services This study was funded by an independent grant from the Comision "Salud Division. Victorian Government Department of Human Services 2005.
Investiga" of the Argentine Ministry of Health (Becas Carrillo-Oñativia).
Mathers C, Vos T, Stevenson C: The burden of disease and injury in The authors declare that they have no competing interests.
Australia. Book The burden of disease and injury in Australia City: AustralianInstitute of Health and Welfare 1999.
Received: 14 January 2010 Accepted: 20 October 2010 Bassi S: PLoS Comput Biol Published: 20 October 2010 2007, 3:e199.
Sellers DE, Crawford SL, Bullock K, McKinlay JB: Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Soc Sci Med 1997, 44:1325-1339.
Budaj A, Pais P, Varigos J, Lisheng L: Cavill N, Bauman A: Lancet 2004, 364:937-952.
Sports Sci 2004, 22:771-790.
Argentina: indicadores básicos 2008. Buenos Aires, Argentina: Ministerio Finlay SJ, Faulkner G: de Salud 2009.
Ministerio de Salud de la Nación, Dirección de Prevención y Protección de 2005, 40:121-130.
la Salud: Boletín de Vigilancia. Enfermedades No Transmisibles y Factores Marshall AL, Owen N, Bauman de Riesgo. Buenos Aires: Ministerio de Salud de la Nación, Dirección de Prevención y Protección de la Salud 2009.
J Sci Med Sport 2004, Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJ: Kahn EB, Ramsey LT, Brownson RC, Heath GW, Howze EH, Powell KE, Lancet 2006, 367:1747-1757.
Stone EJ, Rajab MW, Corso P: Primera Encuesta Nacional de Factores de Riesgo: Buenos Aires: Ministerio Am J Prev Med 2002, de Salud y Ambiente de la Nación, First 2006.
Schargrodsky H, Hernandez-Hernandez R, Champagne BM, Silva H, He FJ, MacGregor GA: Vinueza R, Silva Aycaguer LC, Touboul PJ, Boissonnet CP, Escobedo J, Cochrane Database Syst Rev 2004, CD004937.
Pellegrini F, et al Hooper L, Bartlett C, Davey SG, Ebrahim S: Am J Med 2008, 121:58-65.
Cochrane Database Syst Rev Abegunde DO, Mathers CD, Adam T, Ortegon M, Strong 2004, CD003656.
Lancet 2007, 370:1929-1938.
Obes Res 1995, 3(Suppl 2):283s-288s.
Lanas F, Avezum A, Bautista LE, Diaz R, Luna M, Islam S, Yusuf S: Miles A, Rapoport L, Wardle J, Afuape T, Duman Circulation 2007, 115:1067-1074.
Murray CJL, Lopez AD: The global burden of disease: a comprehensive Health Educ Res 2001, 16:357-372.
assessment of mortality and disability from diseases, injuries, and risk Gepkens A, Gunning SL: Interventions to reduce socioeconomic health factors in 1990 and projected to 2020. Cambridge, MA: Harvard University differences: A review of the international literature. Eur J Pub Health 1996, The World Health report: Reducing Risks, Promoting Healthy Life. Geneva: Foerster SB, Kizer KW, Disogra LK, Bal DG, Krieg BF, Bunch World Health Organization 2002.
Rubinstein et al. BMC Public Health 2010, 10:627 Am J Prev Med 1995, Am J Health Promot 1990, 4:435-440.
Blanco P, Gagliardi J, Higa C, Dini A, Guetta J, di Toro D, Botto F, Heimendinger J, Chapelsky D: Sarmiento RA: Infarto agudo de miocardio. Resultados de la Encuesta Adv Exp Med Biol 1996, 401:199-206.
SAC 2005 en la República Argentina. Rev Argent Cardiol 2007, 75:163-170.
Puska P, Tuomilehto J, Nissinen A, Salonen JT, Vartiainen E, Pietinen P, Elizari MV, Martinez JM, Belziti C, Ciruzzi M, Perez dela Hoz R, Sinisi A, Koskela K, Korhonen Carbajales J, Scapin O, Garguichevich J, Girotti L, Cagide A: Farquhar JW, Fortmann SP, Flora JA, Taylor CB, Haskell WL, Williams PT, Maccoby N, Wood PD: European heart journal 2000, 21:198-205.
Gurfinkel EP, Bozovich GE, Dabbous O, Mautner B, Anderson F: Socio 1990, 264:359-365.
economic crisis and mortality. Epidemiological testimony of the financial Asaria P, Chisholm D, Mathers C, Ezzati M, Beaglehole R collapse of Argentina. Thrombosis journal [electronic resource] 2005, 3:22.
Rojas JI, Zurru MC, Patrucco L, Romano M, Riccio PM, Cristiano Lancet 2007, 370:2044-2053.
2006, 66:547-551.
Sposato LA, Esnaola MM, Zamora R, Zurru MC, Fustinoni O, Saposnik G: J Public Health Manag Pract 2007, 13:296-306.
Stroke 2008, 39:3036-3041.
Sposato LA, EM M, Cirio JJ, Zurru MC, Rey RC, Domínguez R, Lepera S, Rotta Public Health Rep 1987, 102:398-403.
Escalante R, Herrera G, Abiusi G, et al: Acute ischemic stroke treatment in Macaskill P, Pierce JP, Simpson JM, Lyle DM: Argentina. ReNACer. Argentinian Stroke Registry. 2006.
Pichón-Rivière A, Regueiro A, Souto A, Augustovski F: Base de datos de Public Health 1992, 82:96-98.
costos sanitarios Argentinos. Book Base de datos de costos sanitarios Secker-Walker RH, Gnich W, Platt S, Lancaster T: Argentinos City: Instituto de Efectividad Clínica y Sanitaria (IECS) 2004.
Cochrane Database Syst Rev 2002, Ministerio de Salud: Resolución 372/2001. Aranceles de los Hospitales Públicos de Gestión Descentralizada. Book Resolución 372/2001. Aranceles Apro N, Gil GP, Rodríguez J, Puntieri MV, Ferreyra VA, Gulivart VL, Freile GE, de los Hospitales Públicos de Gestión Descentralizada City: Superintendencia Gambarotta L, Blasco R, Aguilar V, et al: Relevamiento del uso de sal en de Servicios de Salud 2001, 12, vol. Resolución 372/2001. pp. 12.
los productos de panaderías artesanales de la República Argentina e Ministerio de Salud: Resolución 488/2002. Normas y módulos para los implementación de acciones de desarrollo, tecnológicas, de asistencia Hospitales Públicos de Gestión Descentralizada. Unidad Hospital Público técnica y extensión con el objeto de bajar su utilización y consumo.
(UHP). Book Resolución 488/2002. Normas y módulos para los Hospitales Promoción de la Salud cardiovascular y la alimentación saludable Estudios y Públicos de Gestión Descentralizada. Unidad Hospital Público (UHP) City: experiencias Buenos Aires: Ministerio de Salud 2005.
Superintendencia de Servicios de Salud 2002, vol. Resolución 488/2002.
He FJ, MacGregor GA: Manual Farmacéutico On Line. Hypertension 2003, 42:1093-1099.
Comisión Nacional de Medicamentos Remediar (personal communication): Ferrante D, Levy D, Peruga A, Compton C, Romano Buenos Aires: Ministerio de Salud 2007.
Compras de medicamentos, años 2004-2006 (personal communication): Rev Panam Salud Publica 2007, 21:37-49.
Neuquén: Subsecretaría de Salud 2007.
Gimpel NE, Schoj V, Rubinstein A: Quality management of hypertension in World Economic Outlook Database, April 2009. primary care: do physicians treat patients' blood pressure level or cardiovascular risk? Quality in Primary Care 2006, 14:211-217.
Fox-Rushby JA, Hanson K Hemmelgarn BR, McAlister FA, Grover S, Myers MG, McKay DW, Bolli P, Health policy and planning Abbott C, Schiffrin EL, Honos G, Burgess E, et al 2001, 16:326-331.
Sachs JD: Macroeconomics and Health: Investing in Health for Economic Development. Report of the Commission on Macroeconomics and Can J Cardiol 2006, 22:573-581.
Health. Geneva: World Health Organization 2001.
Third Report of the National Cholesterol Education Program (NCEP) Baltussen RMPM, Adam T, Tan-Torres Edejer T, Hutubessy RCW, Acharya A, Expert Panel on Detection, Evaluation, and Treatment of High Blood Evans DB, Murray CJL: Making choices in health : WHO guide to cost- Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation effectiveness analysis. Geneva: World Health Organization 2003.
World Bank, International Comparison Program database. Hughes J, Stead L, Lancaster T Cochrane Database Syst Rev 2004, CD000031.
Hurt RD, Sachs DP, Glover ED, Offord KP, Johnston JA, Dale LC, PharmacoEconomics 2006, 24:1043-1053.
Khayrallah MA, Schroeder DR, Glover PN, Sullivan CR, et al: Halpern EF, Weinstein MC, Hunink MG, Gazelle GS: Engl J Med 1997, 337:1195-1202.
Med Decis Making 2000, 20:314-322.
Swan GE, McAfee T, Curry SJ, Jack LM, Javitz H, Dacey S, Bergman K: World Health Organization, Department of Chronic Diseases and Health Promotion: Preventing chronic diseases : a vital investment : WHO global Arch Intern Med 2003, report. Geneva: World Health Organization 2005.
Strong K, Mathers C, Leeder S, Beaglehole R: Guía Nacional de Tratamiento de la adicción al tabaco. Book Guía Lancet 2005, 366:1578-1582.
Nacional de Tratamiento de la adicción al tabaco City: Ministerio de Salud y Jamison DT: Chapter 1. Investing in Health. In Disease Control Priorities in Ambiente de la Nación 2005.
Developing Countries. Edited by: Jamison DT, Breman JG, Measham AR, Alleyne G, Claeson M, Evans DB, Jha P, Mills A, Musgrove P. Washington, BMJ 2003, 326:1419.
DC: Oxford University Press and The World Bank; , Second 2006:.
Lim SS, Gaziano TA, Gakidou E, Reddy KS, Farzadfar F, Lozano R, Rodgers A: 2002, 360:2-3.
Baigent C, Keech A, Kearney PM, Blackwell L, Buck G, Pollicino C, Kirby A, Sourjina T, Peto R, Collins R, Simes Rubinstein et al. BMC Public Health 2010, 10:627 Population 2006. Total Gross Domestic Product (GDP) 2006. Rubinstein A, Garcia Marti S, Souto A, Ferrante D, Augustovski F:Cost Eff ResourAlloc 2009, 7:10.
World Health Organization: Prevención de las enfermedadescardiovasculares : guía de bolsillo para la estimación y el manejo delriesgo cardiovascular. Ginebra: Organización Mundial de la Salud 2008.
Lewington S, Clarke R, Qizilbash N, Peto R, Collins Lancet2002, 360:1903-1913.
Miura K, Daviglus ML, Dyer AR, Liu K, Garside DB, Stamler J, Greenland P:ArchIntern Med 2001, 161:1501-1508.
Lloyd-Jones DM, Evans JC, Levy JAMA 2005,294:466-472.
Huxley R, Barzi F, Woodward BMJ 2006, 332:73-78.
Bogers RP, Bemelmans WJ, Hoogenveen RT, Boshuizen HC, Woodward M,Knekt P, van Dam RM, Hu FB, Visscher TL, Menotti A, et al: Archives ofinternal medicine 2007, 167:1720-1728.
Ezzati M, López AD: Smoking and oral tobacco use. In ComparativeQuantification of Health Risks: Global and Regional Burden of DiseaseAttributable to Selected Major Risk Factors. Edited by: Ezzati M, López AD,Rodgers A, Murray CJL. Geneva, Switzerland: World Health Organization;2004:883-957.
CDC SAMMEC. CPS-II. Unpublished estimates provided by AmericanCancer Society (ACS). See Thun MJ, Day-Lally C, Myers DG, et al. Trendsin tobacco smoking and mortality from cigarette use in CancerPrevention Studies I (1959 through 1965) and II (1982 through 1988). In:Changes in cigarette-related disease risks and their implication forprevention and control. Smoking and Tobacco Control Monograph 8.
Bethesda, MD: US Department of Health and Human Services, PublicHealth Service, National Institutes of Health, National Cancer Institute1997; 305-382. NIH Publication no. 97-1213. In Book CDC SAMMEC. CPS-II.
Unpublished estimates provided by American Cancer Society (ACS). Edited by:See Thun MJ, Day-Lally C, Myers DG. Bethesda, MD: US Department ofHealth and Human Services, Public Health Service, National Institutes ofHealth, National Cancer Institute; 1997:305-382, Trends in tobacco smokingand mortality from cigarette use in Cancer Prevention Studies I (1959through 1965) and II (1982 through 1988). In: Changes in cigarette-relateddisease risks and their implication for prevention and control. Smoking andTobacco Control Monograph 8.
Hu G, Tuomilehto J, Silventoinen K, Sarti C, Mannisto S, Jousilahti Archives of internal medicine 2007, 167:1420-1427.
Amarenco P, Labreuche J, Lavallee P, Touboul PJ: Submit your next manuscript to BioMed Central
Stroke 2004, 35:2902-2909.
and take full advantage of:
Lee CD, Folsom AR, Blair Stroke 2003, 34:2475-2481.
• Convenient online submission
Pre-publication history • Thorough peer review
The pre-publication history for this paper can be accessed here: • No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
Cite this article as: Rubinstein et al.: Estimation of the burden ofcardiovascular disease attributable to modifiable risk factors and cost- • Research which is freely available for redistribution
effectiveness analysis of preventative interventions to reduce thisburden in Argentina. BMC Public Health 2010 10:627.
Submit your manuscript at

Source: https://www1.hospitalitaliano.org.ar/multimedia/archivos/noticias_archivos/49/pdf/49_135II.pdf