Attributable fractions: fundamental concepts and their visualization

Preventable incidence cases from non-communicable diseases attributable to insufficient physical activity in Chile

OBJECTIVES

Lack of sufficient physical activity (PA) has been associated with an increased risk of several non-communicable diseases (NCDs) and all-cause mortality. This study aimed to estimate the number of preventable incidence cases of NCDs attributable to insufficient PA in the Chilean population.

STUDY DESIGN

Comparative risk assessment modelling study.

METHODS

This study examined data from 5834 participants aged ≥20 years from the Chilean National Survey (2016-2017). PA was assessed by the Global Physical Activity Questionnaire (GPAQ), and metabolic equivalent of tasks (METs) were assigned according to PA intensity. Estimated incidence cases of NCDs in Chile in 2019 were obtained from the Global Burden of Disease study. Relative risks for breast cancer, colon cancer, ischaemic heart disease, diabetes and stroke were obtained from a published meta-analysis and applied to the prevalence of insufficient PA estimates through the potential impact fraction equation.

RESULTS

High levels of PA (≥8000 MET-min/week) could potentially avoid more than 22,000 (64.6 %) incidence NCD cases, ranging from 498 (10.1 %) preventable cases of breast cancer to 5629 (14.7 %) cases of diabetes. Other modelled scenarios also showed to reduce the incidence cases of all five NCDs but to a lesser extent; where at least PA recommendation was achieved, preventable NCDs were reduced by 6522 cases (18.7 %), and where a 10 % relative reduction in insufficient PA level in the population was achieved, preventable NCDs were reduced by 651 (1.8 %) cases.

CONCLUSIONS

The study results provide estimates for the incidence cases of preventable NCDs attributable to insufficient PA, highlighting the important role of PA in NCD prevention in Chile.

The invisible costs of obstructive sleep apnea (OSA): Systematic review and cost-of-illness analysis

BACKGROUND

Obstructive sleep apnea (OSA) is a risk factor for several diseases and is correlated with other non-medical consequences that increase the disease's clinical and economic burden. However, OSA's impact is highly underestimated, also due to substantial diagnosis gaps.

OBJECTIVE

This study aims at assessing the economic burden of OSA in the adult population in Italy by performing a cost-of-illness analysis with a societal perspective. In particular, we aimed at estimating the magnitude of the burden caused by conditions for which OSA is a proven risk factor.

METHODS

A systematic literature review on systematic reviews and meta-analyses, integrated by expert opinion, was performed to identify all clinical and non-clinical conditions significantly influenced by OSA. Using the Population Attributable Fraction methodology, a portion of their prevalence and costs was attributed to OSA. The total economic burden of OSA for the society was estimated by summing the costs of each condition influenced by the disease, the costs due to OSA's diagnosis and treatment and the economic value of quality of life lost due to OSA's undertreatment.

RESULTS

Twenty-six clinical (e.g., diabetes) and non-clinical (e.g., car accidents) conditions were found to be significantly influenced by OSA, contributing to an economic burden ranging from €10.7 to €32.0 billion/year in Italy. The cost of impaired quality of life due to OSA undertreatment is between €2.8 and €9.0 billion/year. These costs are substantially higher than those currently borne to diagnose and treat OSA (€234 million/year).

CONCLUSIONS

This study demonstrates that the economic burden due to OSA is substantial, also due to low diagnosis and treatment rates. Providing reliable estimates of the economic impact of OSA at a societal level may increase awareness of the disease burden and help to guide evidence-based policies and prioritisation for healthcare, ultimately ensuring appropriate diagnostic and therapeutic pathways for patients.

Causes of Outcome Learning: a causal inference-inspired machine learning approach to disentangling common combinations of potential causes of a health outcome

Nearly all diseases are caused by different combinations of exposures. Yet, most epidemiological studies focus on estimating the effect of a single exposure on a health outcome. We present the Causes of Outcome Learning approach (CoOL), which seeks to discover combinations of exposures that lead to an increased risk of a specific outcome in parts of the population. The approach allows for exposures acting alone and in synergy with others. The road map of CoOL involves (i) a pre-computational phase used to define a causal model; (ii) a computational phase with three steps, namely (a) fitting a non-negative model on an additive scale, (b) decomposing risk contributions and (c) clustering individuals based on the risk contributions into subgroups; and (iii) a post-computational phase on hypothesis development, validation and triangulation using new data before eventually updating the causal model. The computational phase uses a tailored neural network for the non-negative model on an additive scale and layer-wise relevance propagation for the risk decomposition through this model. We demonstrate the approach on simulated and real-life data using the R package 'CoOL'. The presentation focuses on binary exposures and outcomes but can also be extended to other measurement types. This approach encourages and enables researchers to identify combinations of exposures as potential causes of the health outcome of interest. Expanding our ability to discover complex causes could eventually result in more effective, targeted and informed interventions prioritized for their public health impact.

A Bayesian approach for estimating the partial potential impact fraction with exposure measurement error under a main study/internal validation design

The partial potential impact fraction describes the proportion of disease cases that can be prevented if the distribution of modifiable continuous exposures is shifted in a population, while other risk factors are not modified. It is a useful quantity for evaluating the burden of disease in epidemiologic and public health studies. When exposures are measured with error, the partial potential impact fraction estimates may be biased, which necessitates methods to correct for the exposure measurement error. Motivated by the health professionals follow-up study, we develop a Bayesian approach to adjust for exposure measurement error when estimating the partial potential impact fraction under the main study/internal validation study design. We adopt the reclassification approach that leverages the strength of the main study/internal validation study design and clarifies transportability assumptions for valid inference. We assess the finite-sample performance of both the point and credible interval estimators via extensive simulations and apply the proposed approach in the health professionals follow-up study to estimate the partial potential impact fraction for colorectal cancer incidence under interventions exploring shifting the distributions of red meat, alcohol, and/or folate intake.

Diet Quality as Measured by the Healthy Eating Index 2015 and Oral and Pharyngeal Cancer Risk

BACKGROUND

Alcohol and tobacco are the major risk factors for oral and pharyngeal cancer, but diet is likely to have a role, too.

OBJECTIVE

The objective was to analyze the relationship between adherence to the 2015-2020 Dietary Guidelines for Americans (DGA), as measured by the Healthy Eating Index 2015 (HEI-2015), and oral and pharyngeal cancer risk. Moreover, this work aimed to quantify the number of avoidable cases under different scenarios of increased adherence to the DGA, with the use of the potential impact fraction. This estimates the proportion of cases that would occur if the distribution of the risk factor in the population followed an alternative distribution.

DESIGN

A multicenter, case-control study was conducted in Italy between 1991 and 2009. Participants' usual diet for the 2 years preceding study enrolment was assessed using a food frequency questionnaire.

PARTICIPANTS AND SETTING

Cases were 946 patients admitted to major hospitals with incident, histologically confirmed oral and pharyngeal cancer. Controls were 2,492 patients admitted to the same hospitals for acute non neoplastic conditions.

MAIN OUTCOME MEASURES

The adherence to the DGA was assessed using the HEI-2015 score (range = 0 to 100), based on 13 components. The outcome was oral and pharyngeal cancer.

STATISTICAL ANALYSES PERFORMED

Odds ratios and the corresponding 95% CIs were estimated using multiple logistic regression models adjusted for tobacco, alcohol, and other relevant covariates. The potential impact fraction was estimated under different scenarios of adherence to the DGA.

RESULTS

In this Italian population the HEI-2015 score ranged from 33.4 to 97.5. A higher HEI-2015 score was associated with a lower risk of oral and pharyngeal cancer, with an odds ratio of 0.70 (95% CI 0.62 to 0.79) for a 10-point increment of the score. The estimated potential impact fraction was 64.8% under the maximum achievable reduction scenario, and it ranged from 9% to 27% following other more feasible scenarios.

CONCLUSIONS

The HEI-2015 score was inversely related to oral and pharyngeal cancer risk in this Italian population. This analysis allowed for the estimation of the fraction of preventable cases, under different feasible scenarios. A share of 9% to 27% of avoidable cases of oral and pharyngeal cancer might be obtained across real-world scenarios of adherence to the DGA as measured by the HEI-2015 score.

The Social And Economic Cost Of Sleep Disorders

STUDY OBJECTIVES

To estimate economic cost of common sleep disorders in Australia for 2019-2020.

METHODS

Costs were estimated for obstructive sleep apnea (OSA), insomnia, and restless legs syndrome (RLS) using prevalence, financial, and nonfinancial data from national databases. These included: (1) financial costs associated with health care, informal care, productivity losses, non-medical accident costs, deadweight loss from taxation/welfare inefficiencies; and (2) nonfinancial costs associated with loss of well-being. They were expressed in US dollars ($).

RESULTS

Estimated overall cost of sleep disorders in Australia in 2019-2020 (population: 25.5 million) was $35.4 billion (OSA $13.1 billion; insomnia $13.3 billion, RLS $9.0 billion). Of this, the financial cost component was $10.0 billion, comprised of: health system costs $0.7 billion; productivity losses $7.7 billion; informal care $0.2 billion; other, mainly non-medical accident costs, $0.4 billion; and deadweight losses $1.0 billion. For moderate to severe OSA syndrome, insomnia unrelated to other conditions and RLS, financial costs represented $16,717, $21,982, and $16,624 per adult with the condition for the year, respectively. The nonfinancial cost was $25.4 billion.

CONCLUSIONS

The economic costs associated with sleep disorders are substantial. The financial component of $10.0 billion is equivalent to 0.73% of Australian gross domestic product. The nonfinancial cost of $25.4 billion represents 3.2% of total Australian burden of disease for the year. Health system costs of these disorders are low relative to those associated with their consequences, suggesting greater expenditure on detection, treatment and prevention is warranted.

Quantification and interpretation of attributable mortality in core clinical infectious disease journals

Attributable mortality is an important metric that mirrors the public health effect of a potentially harmful infection by accounting not only for the effect of infection on mortality, but also for its prevalence within the target population. We did a systematic literature review to identify how attributable deaths were quantified and interpreted in core clinical infectious diseases journals between Jan 1, 2013, and April 6, 2020. Of the 1591 abstracts screened, 234 entered the primary analysis. Our summary of the epidemiological measures used in these articles reveals fundamental shortcomings in the conception of attributable mortality. Because of its importance as a basis for decision making on public health matters, it is essential to correctly quantify and report on attributable mortality. Our recommendation for quantification and reporting of attributable deaths aids clinical researchers in the correct statistical assessment of the burden of infections. Fictional as well as real data is used to illustrate these issues.

Comparative Disease Assessment: a multi-causal approach for estimating the burden of mortality

Background

The Comparative Risk Assessment (CRA) framework comprehensively evaluates the impact of exposure to risk factors on health populations using the counterfactual causal approach.

Methods

We propose a framework, Comparative Disease Assessment (CDA), for assessing the impact of exposure to morbidity from some diseases on health outcomes, particularly death from other (relevant) diseases. This framework has been developed following the ideas of the CRA framework and using the widely accepted concept that exposure to morbidity is usually a risk factor for health outcomes (morbidity/mortality) related to other diseases. Our framework uses a counterfactual and not a categorical approach when attributing the burden of health outcomes to potential causes.

Results

This paper describes the different steps and assumptions required to implement the CDA framework, and an illustrative example is used considering diabetes mellitus morbidity as a risk factor for death from heart diseases.

Conclusions

One advantage of the CDA framework is that it can be applied using multi-causal death registries. Some assumptions are needed to implement it in order to avoid biases, but at least it can provide preliminary estimations of the impact of exposure to diseases as risk factors for deaths from other diseases. Another main advantage is that the burden of deaths is no longer attributed to a single cause, the underlying cause, as it is almost always done. Finally, this framework provides information on the pattern of comorbidity in a (sub)population of subjects who is about to die. These patterns can be used as a reference for alternative patterns of the general population or patterns of other specific subpopulations.

Methods to assess the contribution of diseases to disability using cross-sectional studies: comparison of different versions of the attributable fraction and the attribution method

BACKGROUND

This study aims to illustrate the differences between approaches proposed for apportioning disability to different diseases in a multicausal situation, i.e. the unadjusted attributable fraction (AF), the adjusted AF, the average AF and the attribution method (AM). This information is useful to better interpret results obtained from cross-sectional data and help policy makers decide on public health strategies.

METHODS

Data for 29 931 individuals, representative of the French household population, who participated in the 2008-09 cross-sectional Disability-Health Survey, were included. Disability was defined as any limitation reported with the Global Activity Limitation Indicator. Unadjusted AFs were calculated using Levin's formula. Adjusted AFs were estimated for each disease by calculating predicted probabilities of disability for each individual in the dataset, under the assumption that the individual is unexposed to this specific disease (logistic model). Average AFs are based on the same methodology, but have the additional advantage that the average AFs for different diseases sum to the total AF associated with eliminating all diseases. AM accounts for competing risks and partitions total disability prevalence into additive contributions of different diseases and background disability (additive model).

RESULTS

All methods obtained similar results with respect to the estimates of the disease contribution to disability prevalences and to ranking of the diseases, except unadjusted AFs, as the method ignores multimorbidity. Confounders other than diseases, such as age and gender, should be accurately taken into account.

CONCLUSIONS

Conceptual differences, strengths and limitations of the different approaches were discussed.

Expanding attributable fraction applications to outcomes wholly attributable to a risk factor

The problem central to this document is the estimation of change in disease attributable to an epidemiological exposure variable that stems from a change in the distribution of that variable. We require that both disease and exposure are quantifiable as real numbers, and then ask how to estimate the fraction of disease attributable to exposure, producing the general attributable fraction methodology. After the mathematical framework is in place, we explore the implications of a disease that is wholly attributable to a given risk factor, demonstrate why standard applications of the attributable fractions do not extend, and present general methodological considerations for this case. Finally, we demonstrate the methodology using the example of alcoholic psychoses.

Relative contribution of overweight and obesity to rising public hospital in-patient expenditure in South Australia

Objective Arguments to fund obesity prevention have often focused on the growing hospital costs of associated diseases. However, the relative contribution of overweight and obesity to public hospital expenditure growth is not well understood. This paper examines the effect of overweight and obesity on acute public hospital in-patient expenditure in South Australia over time compared with other expenditure drivers. Methods Annual inflation-adjusted acute public admitted expenditure attributable to a high body mass index was estimated for 2007-08 and 2011-12 and compared with other expenditure drivers. Results Expenditure attributable to overweight and obesity increased by A$45million, from 4.7% to 5.4% of total acute public in-patient expenditure. This increase accounted for 7.8% of the A$583million total expenditure growth, whereas the largest component of total growth (62.4%) was a real increase in the average cost per separation. Conclusions The relatively minor contribution of overweight and obesity to expenditure growth over the time period examined invites reflection on arguments to boost preventive spending that centre upon reducing hospital costs. These arguments may inadvertently detract attention from the considerable health and social burdens of overweight and obesity and from unrelated sources of expenditure growth that reduce opportunities for state governments to fund obesity prevention programs despite their comparative benefits to population health. What is known about the topic? Stand-alone estimates suggest that overweight and obesity are placing a considerable financial burden on the Australian public healthcare system. What does this paper add? Our findings challenge common perceptions about the relative importance of overweight and obesity in the context of rising public in-patient expenditure in Australia. What are the implications for practitioners? Consistent serial estimates of overweight- and obesity-attributable expenditure enable its tracking and comparison with other potentially controllable expenditure drivers that may also warrant attention. Explicit consideration of population health trade-offs in expenditure-related decisions, including in enterprise bargaining, would enhance transparency in priority setting.

Quantifying degrees of necessity and of sufficiency in cause-effect relationships with dichotomous and survival outcomes

We suggest measures to quantify the degrees of necessity and of sufficiency of prognostic factors for dichotomous and for survival outcomes. A cause, represented by certain values of prognostic factors, is considered necessary for an event if, without the cause, the event cannot develop. It is considered sufficient for an event if the event is unavoidable in the presence of the cause. Necessity and sufficiency can be seen as the two faces of causation, and this symmetry and equal relevance are reflected by the suggested measures. The measures provide an approximate, in some cases an exact, multiplicative decomposition of explained variation as defined by Schemper and Henderson for censored survival and for dichotomous outcomes.

The measures, ranging from zero to one, are simple, intuitive functions of unconditional and conditional probabilities of an event such as disease or death. These probabilities often will be derived from logistic or Cox regression models; the measures, however, do not require any particular model.

The measures of the degree of necessity implicitly generalize the established attributable fraction or risk for dichotomous prognostic factors and dichotomous outcomes to continuous prognostic factors and to survival outcomes. In a setting with multiple prognostic factors, they provide marginal and partial results akin to marginal and partial odds and hazard ratios from multiple logistic and Cox regression.

Properties of the measures are explored by an extensive simulation study. Their application is demonstrated by three typical real data examples.

The economic cost of inadequate sleep

Study Objectives

To estimate the economic cost (financial and nonfinancial) of inadequate sleep in Australia for the 2016-2017 financial year and relate this to likely costs in similar economies.

Methods

Analysis was undertaken using prevalence, financial, and nonfinancial cost data derived from national surveys and databases. Costs considered included the following: (1) financial costs associated with health care, informal care provided outside healthcare sector, productivity losses, nonmedical work and vehicle accident costs, deadweight loss through inefficiencies relating to lost taxation revenue and welfare payments; and (2) nonfinancial costs of loss of well-being. They were expressed in US dollars ($).

Results

The estimated overall cost of inadequate sleep in Australia in 2016-2017 (population: 24.8 million) was $45.21 billion. The financial cost component was $17.88 billion, comprised of as follows: direct health costs of $160 million for sleep disorders and $1.08 billion for associated conditions; productivity losses of $12.19 billion ($5.22 billion reduced employment, $0.61 billion premature death, $1.73 billion absenteeism, and $4.63 billion presenteeism); nonmedical accident costs of $2.48 billion; informal care costs of $0.41 billion; and deadweight loss of $1.56 billion. The nonfinancial cost of reduced well-being was $27.33 billion.

Conclusions

The financial and nonfinancial costs associated with inadequate sleep are substantial. The estimated total financial cost of $17.88 billion represents 1.55 per cent of Australian gross domestic product. The estimated nonfinancial cost of $27.33 billion represents 4.6 per cent of the total Australian burden of disease for the year. These costs warrant substantial investment in preventive health measures to address the issue through education and regulation.

Attributable fraction for multiple risk factors: Methods, interpretations, and examples

The attributable fraction is the candidate tool to quantify individual shares of each risk factor on the disease burden in a population, expressing the proportion of cases ascribable to the risk factors. The original formula ignored the presence of other factors (i.e. multiple risk factors and/or confounders), and several adjusting methods for potential confounders have been proposed. However, crude and adjusted attributable fractions do not sum up to their joint attributable fraction (i.e. the number of cases attributable to all risk factors together) and their sum may exceed one. A different approach consists of partitioning the joint attributable fraction into exposure-specific shares leading to sequential and average attributable fractions. We provide an example using Italian case-control data on oral cavity cancer comparing crude, adjusted, sequential, and average attributable fractions for smoking and alcohol and provide an overview of the available software routines for their estimation. For each method, we give interpretation and discuss shortcomings. Crude and adjusted attributable fractions added up over than one, whereas sequential and average methods added up to the joint attributable fraction = 0.8112 (average attributable fractions for smoking and alcohol were 0.4894 and 0.3218, respectively). The attributable fraction is a well-known epidemiological measure that translates risk factors prevalence and disease occurrence in useful figures for a public health perspective. This work endorses their proper use and interpretation.

Pregnancy-Related Risk Factors Are Associated With a Significant Burden of Treated Hypertension Within 10 Years of Delivery: Findings From a Population-Based Norwegian Cohort

BACKGROUND

The association between pregnancy complications and women's later cardiovascular disease has, primarily, been evaluated in studies lacking information on important covariates. This report evaluates the prospective associations between pregnancy-related risk factors (preeclampsia/eclampsia, gestational hypertension, pregestational and gestational diabetes mellitus, preterm delivery, and fetal growth restriction) and pharmacologically treated hypertension within 10 years after pregnancy, while adjusting for a wide range of covariates.

METHODS AND RESULTS

Prepregnancy normotensive women participating in the MoBa (Norwegian Mother and Child Cohort Study) from January 2004 through July 2009 were linked to the Norwegian Prescription Database to identify women with pharmacologically treated hypertension beyond the postpartum period of 3 months. The burden of hypertension associated with pregnancy-related risk factors was evaluated using an attributable fraction method. A total of 1480 women developed pharmacologically treated hypertension within the follow-up among 60 027 women (rate of hypertension, 3.6/1000 person-years). The proportion of hypertension associated with a history of preeclampsia/eclampsia, gestational hypertension, preterm delivery, and pregestational or gestational diabetes mellitus was 28.6% (95% confidence interval, 25.5%-31.6%) on the basis of multivariable analyses adjusting for numerous covariates. The proportion was similar for women with a healthy prepregnancy body mass index (18.5-24.9 kg/m2; attributable fraction (AF)% 25.9%; 95% confidence interval, 21.3%-30.3%), but considerably higher for nulliparous women at baseline within the first 5 years of follow-up. Small-for-gestational age, however, did not increase subsequent hypertension risk in multivariable analyses.

CONCLUSIONS

A structured postpartum follow-up of high-risk women identified through pregnancy-related risk factors would facilitate personalized preventive strategies to postpone or avoid onset of premature cardiovascular events.

Decomposition of the population attributable fraction for two exposures

PURPOSE

The population attributable fraction (AF) is frequently used to quantify disease burden attributable to exposures. AF is interpreted as the fractional reduction of disease events that would occur if exposures were eliminated. This article aims to provide a decomposition of the overall AF for two exposures into AFs for each of two exposures and AF for their interaction, using potential outcomes framework.

METHODS

We provide the decomposition formula with and without confounders. We discuss an estimation method using standard regression models. We also show that these AFs without confounders can be effectively visualized.

RESULTS

By a numerical comparison, we show that our decomposition is different from a previous decomposition, which does not have a causal interpretation if confounding exists. We illustrate the proposed decomposition using a large prospective cohort study data.

CONCLUSIONS

When the primary exposure cannot be modifiable, the interventional interpretation of AF is difficult. Even then, if there exists an interaction between the exposure and another modifiable exposure, our decomposition can show what extent of the effect of the primary exposure can be eliminated by intervening on the modifiable exposure.

Estimation of Mortality Rates, Life Expectancy, and Life Potential of Illegal Drug Users

The study presents proposed methods of estimation for mortality rates of illegal drug users and nonusers and their application in estimation of the life expectancy and the life potential loss of the users in Poland. Mortality rates for male users were in Poland in 2013, as an average, 4 times higher for male illegal drug users than for nonusers, and 7 times as high for females. For male illegal drug users, the loss of more than 12 years of life can be expected, whereas for females the figure stands at about 8 years. Similar losses of life potential of the users can be expected. The applied methods can be particularly useful in estimation of the social costs of illegal drug use. The results strongly underline the negative effect of the drug use which may have a preventive effect.

High birth weight and perinatal mortality among siblings: A register based study in Norway, 1967-2011

BACKGROUND

Perinatal mortality according to birth weight has an inverse J-pattern. Our aim was to estimate the influence of familial factors on this pattern, applying a cohort sibling design. We focused on excess mortality among macrosomic infants (>2 SD above the mean) and hypothesized that the birth weight-mortality association could be explained by confounding shared family factors. We also estimated how the participant's deviation from mean sibling birth weight influenced the association.

METHODS AND FINDINGS

We included 1 925 929 singletons, born term or post-term to mothers with more than one delivery 1967-2011 registered in the Medical Birth Registry of Norway. We examined z-score birth weight and perinatal mortality in random-effects and sibling fixed-effects logistic regression models including measured confounders (e.g. maternal diabetes) as well as unmeasured shared family confounders (through fixed effects models). Birth weight-specific mortality showed an inverse J-pattern, being lowest (2.0 per 1000) at reference weight (z-score +1 to +2) and increasing for higher weights. Mortality in the highest weight category was 15-fold higher than reference. This pattern changed little in multivariable models. Deviance from mean sibling birth weight modified the mortality pattern across the birth weight spectrum: small and medium-sized infants had increased mortality when being smaller than their siblings, and large-sized infants had an increased risk when outweighing their siblings. Maternal diabetes and birth weight acted in a synergistic fashion with mortality among macrosomic infants in diabetic pregnancies in excess of what would be expected for additive effects.

CONCLUSIONS

The inverse J-pattern between birth weight and mortality is not explained by measured confounders or unmeasured shared family factors. Infants are at particularly high mortality risk when their birth weight deviates substantially from their siblings. Sensitivity analysis suggests that characteristics related to maternal diabetes could be important in explaining the increased mortality among macrosomic infants.

Worldwide burden of gastric cancer in 2010 attributable to high sodium intake in 1990 and predicted attributable burden for 2030 based on exposures in 2010

Assessing the impact that patterns of Na intake may have on gastric cancer will provide a more comprehensive estimation of Na reduction as a primary prevention approach. We aimed to estimate the proportion of gastric cancer cases that are attributable to Na intake above the recommendation by the WHO (≤2 g/d) throughout the world in 2010, as well as expected values for 2030. Population attributable fractions (PAF) were computed for 187 countries, using Na intakes in 1990 and 2010 and estimates of the association between Na intake and gastric cancer, assuming a time lag of 20 years. Median PAF ranged from 10·1% in low to 22·5 % in very high Human Development Index (HDI) countries in men (P<0·001) and from 7·2 to 16·6 %, respectively, among women (P<0·001). An increase in median PAF until 2030 is expected in most settings, except for countries classified as low HDI, in both sexes. High Na intakes account for a large proportion of gastric cancer cases, and proportions are expected to increase in almost all of the countries. Intensified efforts to diminish Na intake in virtually all populations are needed to further reduce gastric cancer burden.

Estimating average attributable fractions with confidence intervals for cohort and case–control studies

Chronic diseases tend to depend on a large number of risk factors, both environmental and genetic. Average attributable fractions were introduced by Eide and Gefeller as a way of partitioning overall disease burden into contributions from individual risk factors; this may be useful in deciding which risk factors to target in disease interventions. Here, we introduce new estimation methods for average attributable fractions that are appropriate for both case–control designs and prospective studies. Confidence intervals, derived using Monte Carlo simulation, are also described. Finally, we introduce a novel approximation for the sample average attributable fraction that will ensure a computationally tractable approach when the number of risk factors is large. An R package, [Formula: see text], implementing the methods described in this manuscript can be downloaded from the CRAN repository.

High population attributable fractions of myocardial infarction associated with waist-hip ratio

OBJECTIVE

To estimate population attributable fractions (PAF) of acute myocardial infarction (AMI) associated with anthropometric measures by sex and age.

METHODS

The Cohort of Norway study identified 140,790 participants free of cardiovascular disease, 1994-2003. Participants were followed for AMI through 2009 by record linkages through the Cardiovascular Disease in Norway Project. PAFs were adjusted for age, smoking, systolic blood pressure, diabetes, and the ratio of total cholesterol to high-density lipoprotein cholesterol.

RESULTS

The PAFs associated with a waist-hip ratio (WHR) in the top two quintiles were 26.1% (95% confidence interval, CI 14.6-36.1) for middle-aged women (<60 years, mean of 41 years) and 9.3% (95% CI 3.0-15.1) for similarly aged men after adjustment for body mass index (BMI) and conventional risk factors. However, PAFs associated with anthropometric measures in elderly participants (≥ 60 years, mean of 70 years) were non-significant in multivariable analyses. Also, WHR was a significant predictor of AMI among men and women without an enlarged waist circumference (<102 cm for men and < 88 cm for women) in adjusted analyses.

CONCLUSIONS

WHR measurements could improve identification of at-risk individuals above and beyond that of conventional risk factors, BMI, or an enlarged waist circumference.

Beyond case fatality rate: using potential impact fraction to estimate the effect of increasing treatment uptake on mortality

Background

IMPACT is an epidemiological model that has been used to estimate how increased treatment uptakes affect mortality and related outcomes. The model calculations require the use of case fatality rate estimates under no treatment. Due to the lack of data, rates where treatment is partially present are often used instead, introducing bias. A method that does not rely on no-treatment case fatality rate estimates is needed.

Methods

Potential Impact Fraction (PIF) measures the proportional reduction in the disease or mortality risk, when the distribution of a risk factor changes. Here, we first describe a probabilistic framework for interpreting quantities used in the IMPACT model, and then we show how this is connected with PIF, facilitating its use for the estimation of the relative reduction of mortality caused by treatment uptake increase. We compare the proposed and standard methods to estimate the reduction of cardiovascular disease deaths in Ontario, if utilization of coronary heart disease interventions was increased to the level of 90%.

Results

Using the proposed method, we estimated that increasing treatment to benchmark levels uptake results in a reduction of 22.5% in cardiovascular mortality. The standard method gives a reduction of 20.8%.

Conclusions

Here we present an alternative method for the estimation of the effect of treatment uptake change on mortality. Our example suggests that the bias associated with the standard method may be substantial. This approach offers a useful tool for epidemiological and health care research and policy.

Population attributable risk fraction for selected chronic diseases in India

BACKGROUND

India's current health transition stage poses a critical challenge of dealing with the unfinished agenda of communicable diseases and the steadily rising burden of noncommunicable diseases. A significant burden of chronic diseases in India is attributable to household and individual level health risk factors coupled with socioeconomic conditions. From this perspective, this article made a first time effort to assess disease burden attributable to health risk factors using cross-sectional population health survey data.

METHODS

Population attributable fractions (PAF) were estimated for a cluster of health risk factors that include unsafe water, lack of sanitation, exposure to cooking smoke, tobacco and alcohol use, physical inactivity, and socioeconomic conditions on a set of widely prevalent chronic diseases such as tuberculosis, malaria, diarrhea, diabetes, angina, and asthma. Data from the 2003 World Health Survey was used.

RESULTS

The analysis revealed evidence of a significant contribution of health risk factors to India's escalating chronic disease burden. The contribution of health risk factors toward chronic disease burden varied by residence.

CONCLUSION

Results suggest that promotional health care based policies to deal with health risks should be a major priority in policy agenda to combat with the challenge of emerging noncommunicable disease coupled with the persistent burden of communicable diseases. Disease burden in India could be halved by effectively modifying exposure to the risk factors through promotional health care.

Model-Based Estimation of the Attributable Risk: A Loglinear Approach

This paper considers model-based methods for estimation of the adjusted attributable risk (AR) in both case-control and cohort studies. An earlier review discussed approaches for both types of studies, using the standard logistic regression model for case-control studies, and for cohort studies proposing the equivalent Poisson model in order to account for the additional variability in estimating the distribution of exposures and covariates from the data. In this paper we revisit case-control studies, arguing for the equivalent Poisson model in this case as well. Using the delta method with the Poisson model, we provide general expressions for the asymptotic variance of the AR for both types of studies. This includes the generalized AR, which extends the original idea of attributable risk to the case where the exposure is not completely eliminated. These variance expressions can be easily programmed in any statistical package that includes Poisson regression and has capabilities for simple matrix algebra. In addition, we discuss computation of standard errors and confidence limits using bootstrap resampling. For cohort studies, use of the bootstrap allows binary regression models with link functions other than the logit.

Determining the best population-level alcohol consumption model and its impact on estimates of alcohol-attributable harms

BACKGROUND

The goals of our study are to determine the most appropriate model for alcohol consumption as an exposure for burden of disease, to analyze the effect of the chosen alcohol consumption distribution on the estimation of the alcohol Population- Attributable Fractions (PAFs), and to characterize the chosen alcohol consumption distribution by exploring if there is a global relationship within the distribution.

METHODS

To identify the best model, the Log-Normal, Gamma, and Weibull prevalence distributions were examined using data from 41 surveys from Gender, Alcohol and Culture: An International Study (GENACIS) and from the European Comparative Alcohol Study. To assess the effect of these distributions on the estimated alcohol PAFs, we calculated the alcohol PAF for diabetes, breast cancer, and pancreatitis using the three above-named distributions and using the more traditional approach based on categories. The relationship between the mean and the standard deviation from the Gamma distribution was estimated using data from 851 datasets for 66 countries from GENACIS and from the STEPwise approach to Surveillance from the World Health Organization.

RESULTS

The Log-Normal distribution provided a poor fit for the survey data, with Gamma and Weibull distributions providing better fits. Additionally, our analyses showed that there were no marked differences for the alcohol PAF estimates based on the Gamma or Weibull distributions compared to PAFs based on categorical alcohol consumption estimates. The standard deviation of the alcohol distribution was highly dependent on the mean, with a unit increase in alcohol consumption associated with a unit increase in the mean of 1.258 (95% CI: 1.223 to 1.293) (R2 = 0.9207) for women and 1.171 (95% CI: 1.144 to 1.197) (R2 = 0. 9474) for men.

CONCLUSIONS

Although the Gamma distribution and the Weibull distribution provided similar results, the Gamma distribution is recommended to model alcohol consumption from population surveys due to its fit, flexibility, and the ease with which it can be modified. The results showed that a large degree of variance of the standard deviation of the alcohol consumption Gamma distribution was explained by the mean alcohol consumption, allowing for alcohol consumption to be modeled through a Gamma distribution using only average consumption.

Alcohol-attributable burden of disease and injury in Canada, 2004

OBJECTIVE

This analysis aimed to estimate the burden of disease and injury caused and prevented by alcohol in 2004 for Canadians aged 0-69 years and compare the effects of different magnitudes of adjustment of survey data on these estimates.

METHODS

Alcohol indicators were obtained from the Canadian Alcohol and Drug Use Monitoring Survey 2008 and were corrected to 80% coverage using adult per capita recorded and unrecorded consumption. Risk relations were taken from meta-analyses. Estimates of burden of disease and injury were obtained from the World Health Organization.

RESULTS

In 2004, 4,721 (95% CI 1,432-8,150) deaths and 274,663 (95% CI 201,397-352,432) disability-adjusted life years lost (DALYs) of Canadians 0-69 years of age were attributable to alcohol. This represented 7.1% (95% CI 2.1-12.2%) of all deaths and 9.3% (95% CI 6.8-11.9%) of DALYs for this age range. The sensitivity analysis showed that the outcome estimates varied substantially based on the adjusted coverage rate.

CONCLUSION

More attention to burden of disease and injury statistics is required to accurately characterize alcohol-related harms. This burden is preventable and could be reduced by implementation of more effective policies.

Population-attributable fractions of Axis I and Axis II mental disorders for suicide attempts: findings from a representative sample of the adult, noninstitutionalized US population

OBJECTIVES

We aimed to determine the percentage of suicide attempts attributable to individual Axis I and Axis II mental disorders by studying population-attributable fractions (PAFs) in a nationally representative sample.

METHODS

Data were from the National Epidemiologic Survey on Alcohol and Related Conditions Wave 2 (NESARC; 2004-2005), a large (N = 34 653) survey of mental illness in the United States. We used multivariate logistic regression to compare individuals with and without a history of suicide attempt across Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Axis I disorders (anxiety, mood, psychotic, alcohol, and drug disorders) and all 10 Axis II personality disorders. PAFs were calculated for each disorder.

RESULTS

Of the 25 disorders we examined in the model, 4 disorders had notably high PAF values: major depressive disorder (PAF = 26.6%; 95% confidence interval [CI] = 20.1, 33.2), borderline personality disorder (PAF = 18.1%; 95% CI = 13.4, 23.5), nicotine dependence (PAF = 8.4%; 95% CI = 3.4, 13.7), and posttraumatic stress disorder (PAF = 6.3%; 95% CI = 3.2, 10.0).

CONCLUSIONS

Our results provide new insight into the relationships between mental disorders and suicide attempts in the general population. Although many mental illnesses were associated with an increased likelihood of suicide attempt, elevated rates of suicide attempts were mostly attributed to the presence of 4 disorders.

Attributable fractions for partitioning risk and evaluating disease prevention: a practical guide

INTRODUCTION

The attributable fraction (AF) is used for quantifying the fraction of diseased ascribable to one or more exposures. The methodology and software for its estimation has undergone a considerable development during the last decades.

OBJECTIVES

To introduce methods for: (i) apportioning excess risk to multiple exposures, groups of exposures and subpopulations; (ii) graphical description; and (iii) survival data.

RESULTS

Adjusted, sequential and average AFs are reasonable measures obtainable with standard software. The latter two both sum up to the combined AF for a set of exposures. The average AFs are independent of the exposures' ordering. For an ordered, preventive strategy, scaled sample space cubes illustrate the effects on the risk of disease from stepwise exposure removal. Pie charts illustrate the portions of the total risk ascribed to different exposures or risk-profiles. Attributable hazard fraction, AF before time t, and AF within study incorporate time to disease and interventions.

CONCLUSIONS

The practice of crude calculations of AFs in epidemiology should be abandoned. Further development of methods for AFs with survival data and possibly linking it to causal modelling is of interest.

Avoidable portion of tobacco-attributable acute care hospital days and its cost due to implementation of different intervention strategies in Canada

The impact of four effective population-based interventions, focusing on individual behavioural change and aimed at reducing tobacco-attributable morbidity, was assessed by modeling with respect to effects on reducing prevalence rates of cigarette smoking, population-attributable fractions, reductions of disease-specific morbidity and its cost for Canada. Results revealed that an implementation of a combination of four tobacco policy interventions would result in a savings of 33,307 acute care hospital days, which translates to a cost savings of about $37 million per year in Canada. Assuming 40% coverage rate for all individually based interventions, the two most effective interventions, in terms of avoidable burden due to morbidity, would be nicotine replacement therapy and physicians' advice, followed by individual behavioural counselling and increasing taxes by 10%. Although a sizable reduction in the number of hospital days and accumulated costs could be achieved, overall these interventions would reduce less than 3% of all tobacco-attributable costs in Canada.

Attributable fractions with survival data

Attributable fraction (AF) is an important concept in clinical and epidemiological studies. The concept has mainly been discussed in relation to case-control studies, cross-sectional studies, and follow-up studies of fixed length. Here, we propose and discuss several ways of defining and estimating AFs with right-censored survival data, and thus with varying lengths of follow-up. In particular, we define the attributable hazard fraction, the AF before time t, and the AF within study. These measures have different interpretations and may give different numerical values, as illustrated in an application to real data on time to the first receiving of cash benefits for hearing impairment in children. The results underline the need for careful selection of the type of measure and interpretation when reporting AFs for survival data.

The impact of health behaviour on long term sickness absence: results from DWECS/DREAM

Long term sickness absence (LTSA) is a major public health problem. We examined the impact of four, potentially modifiable, health behaviours, such as smoking, alcohol consumption, leisure time physical activity, and the associated variable of body mass index on the risk of subsequent LTSA. This was done by following a representative population sample of 5,020 Danish employees aged 18-69 for 18 months in a national register on social transfer payments. Risk estimates for onset of LTSA and etiologic fractions were computed. In women, ex-smokers and heavy smokers had an increased risk of LTSA of 1.61 and 2.05 respectively after adjustment for age, family status, socio economic status, school education, physical and psychosocial work environment exposures and diagnosed disease. In men, effect estimates were smaller and only borderline significant in the fully-adjusted model. The etiologic fraction of smoking was 17.4% in men and 25.5% in women.

Average Attributable Fractions: A Coherent Theory for Apportioning Excess Risk to Individual Risk Factors and Subpopulations

The attributable fraction in a population and the attributable fraction in exposed are different epidemiologic measures for quantifying the contribution of a risk factor to the risk of disease. While the attributable fraction in a population depends on both the relative risk of disease and the risk of being exposed in the population, the attributable fraction in exposed depends only on the relative risk. Similar relationships apply to the combined attributable fraction in a population and in exposed, respectively, for quantifying the total contribution of a group of risk factors. Eide and Gefeller (1995) showed how the sequential and average attributable fractions could be applied to quantify the contributions of the individual risk factors to a combined attributable fraction in a population. The present paper shows how this methodology can be extended to the combined attributable fraction in exposed. The resulting average attributable fractions in exposed are compared to other proposed methods. The relationship between the average attributable fractions in a population and in exposed is outlined, thus establishing a coherent theory for apportioning attributable fractions in individuals, groups of individuals and populations, to single risk factors or groups of risk factors like modifiable versus nonmodifiable factors.

A Scaled Sample Space Cube Used to Illustrate Attributable Fractions

A three-dimensional graphic method is proposed for displaying the association structure between multiple explanatory variables and their relation to a categorical response. The method combines the techniques of mosaic displays and scaled Venn diagrams, and is especially useful for illustrating attributable fractions in epidemiology. The primary purpose is to show the reduction of disease risk in a population if the joint exposure distribution or the conditional risk function is modified, and the method can be extended to illustrate the potential effects of successive removal of exposures on the overall risk of disease. The scaled sample space cube may be used for communicating the difficult concept of attributable fraction to statisticians, the medical community and the general public in an easily understandable way. Demonstrations of the method use theoretical models as well as data from the Hordaland study on the effect of smoking and occupational exposure on obstructive lung disease. Also, the general principle of adding a third dimension to a mosaic display, instead of using shading or colouring, to show an attribute of a cell in a multiway contingency table, can be helpful for other purposes, as in residual analysis of a loglinear model fitting.

Avoidable burden of disease: conceptual and methodological issues in substance abuse epidemiology

Determining the proportion of avoidable disease burden attributable to substance use is important for both policy development and intervention implementation. Current epidemiological theory has in principle provided a method to estimate avoidable burden of disease and the available statistical tools can provide first rough estimates. The method described in this paper, and its statistical procedures, are exemplified to estimate avoidable burden of tobacco-related disease in Canada. However, further effort is needed to find solutions in the methodological details, namely exposure measurement, risk factor multidimensionality, estimation of changes in exposure distribution over time, and estimation of risk relationships from multiple exposures changing over time with multiple endpoints (causal webs). The impetus to begin refining methods to obtain better starting points for estimating avoidable burden of disease is obvious and should be carried through in order to see real changes through evidence-based policy and intervention.

The contribution of leading diseases and risk factors to excess losses of healthy life in Eastern Europe: burden of disease study

Background

The East/West gradient in health across Europe has been described often, but not using metrics as comprehensive and comparable as those of the Global Burden of Disease 2000 and Comparative Risk Assessment studies.

Methods

Comparisons are made across 3 epidemiological subregions of the WHO region for Europe – A (very low child and adult mortality), B (low child and low adult mortality) and C (low child and high adult mortality) – with populations in 2000 of 412, 218 and 243 millions respectively, and using the following measures: 1. Probabilities of death by sex and causal group across 7 age intervals; 2. Loss of healthy life (DALYs) to diseases and injuries per thousand population; 3. Loss of healthy life (DALYs) attributable to selected risk factors across 3 age ranges.

Results

Absolute differences in mortality are most marked in males and in younger adults, and for deaths from vascular diseases and from injuries. Dominant contributions to east-west differences come from the nutritional/physiological group of risk factors (blood pressure, cholesterol concentration, body mass index, low fruit and vegetable consumption and inactivity) contributing to vascular disease and from the legal drugs – tobacco and alcohol.

Conclusion

The main requirements for reducing excess health losses in the east of Europe are: 1) favorable shifts in all amenable vascular risk factors (irrespective of their current levels) by population-wide and personal measures; 2) intensified tobacco control; 3) reduced alcohol consumption and injury control strategies (for example, for road traffic injuries). Cost effective strategies are broadly known but local institutional support for them needs strengthening.

The effects of joint misclassification of exposure and disease on the attributable risk

While there is extensive methodological literature analysing the effects of misclassification on the relative risk under various misclassification scenarios, for the attributable risk only the effects of non-differential misclassification either of exposure or disease, and the effects of non-differential independent misclassification of exposure and disease have been discussed for the 2 x 2-situation. The paper investigates the effects of non-differential correlated misclassification of exposure and disease on the attributable risk taking possible correlations of both types of misclassification into account. Furthermore, a comparison with the corresponding effects on the relative risk is drawn. We propose a matrix-based approach to describe the underlying structure of non-differential misclassification. The bias arising from non-differential misclassification in the attributable risk and relative risk is evaluated in four examples assuming under- or overreporting of exposure and disease. In each of the four examples we found scenarios where pronounced differences in degree and, more importantly, in direction of bias occurred. Our results clearly demonstrate the danger lying in the stereotype transfer of findings regarding misclassification effects on the relative risk to other epidemiologic risk measures and underline the necessity of specific analyses of the effects of misclassification on the attributable risk.

Undernutrition as an underlying cause of child deaths associated with diarrhea, pneumonia, malaria, and measles

BACKGROUND

Previous analyses derived the relative risk (RR) of dying as a result of low weight-for-age and calculated the proportion of child deaths worldwide attributable to underweight.

OBJECTIVES

The objectives were to examine whether the risk of dying because of underweight varies by cause of death and to estimate the fraction of deaths by cause attributable to underweight.

DESIGN

Data were obtained from investigators of 10 cohort studies with both weight-for-age category (<-3 SDs, -3 to <-2 SDs, -2 to <-1 SD, and >-1 SD) and cause of death information. All 10 studies contributed information on weight-for-age and risk of diarrhea, pneumonia, and all-cause mortality; however, only 6 studies contributed information on deaths because of measles, and only 3 studies contributed information on deaths because of malaria or fever. With use of weighted random effects models, we related the log mortality rate by cause and anthropometric status in each study to derive cause-specific RRs of dying because of undernutrition. Prevalences of each weight-for-age category were obtained from analyses of 310 national nutrition surveys. With use of the RR and prevalence information, we then calculated the fraction of deaths by cause attributable to undernutrition.

RESULTS

The RR of mortality because of low weight-for-age was elevated for each cause of death and for all-cause mortality. Overall, 52.5% of all deaths in young children were attributable to undernutrition, varying from 44.8% for deaths because of measles to 60.7% for deaths because of diarrhea.

CONCLUSION

A significant proportion of deaths in young children worldwide is attributable to low weight-for-age, and efforts to reduce malnutrition should be a policy priority.

Mortality in England and Wales attributable to any drinking, drinking above sensible limits and drinking above lowest-risk level

AIMS

To quantify mortality attributable to any alcohol consumption, and mortality attributable to consumption above different levels.

DESIGN

We related all-cause mortality to alcohol consumption using cause-specific mortality models from a systematic review and using the distribution of alcohol consumption and causes of death by age and sex in England and Wales in 1997. We estimated the deaths and person-years of life lost to age 65 that were attributable: to any drinking; to drinking above the nadir (the level of alcohol consumption carrying the lowest risk); and to drinking more than the British Royal Colleges' recommended limits of 21 units/week in men and 14 units/week in women.

FINDINGS

Ischaemic heart disease deaths prevented by alcohol consumption (11 276 in men, 4050 in women) roughly balanced other deaths attributable to alcohol consumption (9246 in men, 4216 in women). Overall, 0.8% of all deaths in men were prevented by alcohol consumption (95% confidence interval, 0.2% to 1.3%), while 0.1% of all deaths in women were attributable to alcohol consumption (95% confidence interval, - 0.3% to 0.4%); 2.1% (1.9-2.3%) of all deaths in men and 0.8% (0.6-1.0%) of all deaths in women were attributable to drinking more than the recommended limits, while 2.8% and 1.2% of deaths, respectively, were attributable to drinking above the nadir. Of all person-years of life lost to age 65, 10.3% in men and 5.6% in women were attributable to any drinking; 8.5% and 4.0% were attributable to drinking above the recommended limits; and 12.6% and 6.0% were attributable to drinking above the nadir.

CONCLUSIONS

Although overall mortality risks and benefits of alcohol consumption appear roughly equal, drinking above recommended limits remains responsible for many deaths and a large loss of person-years of life.

Estimates of global and regional potential health gains from reducing multiple major risk factors

BACKGROUND

Estimates of the disease burden due to multiple risk factors can show the potential gain from combined preventive measures. But few such investigations have been attempted, and none on a global scale. Our aim was to estimate the potential health benefits from removal of multiple major risk factors.

METHODS

We assessed the burden of disease and injury attributable to the joint effects of 20 selected leading risk factors in 14 epidemiological subregions of the world. We estimated population attributable fractions, defined as the proportional reduction in disease or mortality that would occur if exposure to a risk factor were reduced to an alternative level, from data for risk factor prevalence and hazard size. For every disease, we estimated joint population attributable fractions, for multiple risk factors, by age and sex, from the direct contributions of individual risk factors. To obtain the direct hazards, we reviewed publications and re-analysed cohort data to account for that part of hazard that is mediated through other risks.

RESULTS

Globally, an estimated 47% of premature deaths and 39% of total disease burden in 2000 resulted from the joint effects of the risk factors considered. These risks caused a substantial proportion of important diseases, including diarrhoea (92%-94%), lower respiratory infections (55-62%), lung cancer (72%), chronic obstructive pulmonary disease (60%), ischaemic heart disease (83-89%), and stroke (70-76%). Removal of these risks would have increased global healthy life expectancy by 9.3 years (17%) ranging from 4.4 years (6%) in the developed countries of the western Pacific to 16.1 years (43%) in parts of sub-Saharan Africa.

INTERPRETATION

Removal of major risk factors would not only increase healthy life expectancy in every region, but also reduce some of the differences between regions. The potential for disease prevention and health gain from tackling major known risks simultaneously would be substantial.

The economics of smoking and cardiovascular disease

The direct medical cost of cardiovascular and circulatory diseases was $151 billion in 1995, approximately 17% of all direct medical care costs in the United States. Incidence and prevalence based estimates indicate that smoking is a major contributing factor for cardiovascular disease and associated costs. Statewide smoking control programs and workplace and public area smoking bans are effective in reducing smoking prevalence. Smoking cessation therapies are very cost-effective interventions for the prevention of cardiovascular disease. Incidence based estimates indicate that smoking cessation control expenditures in the United States have been a cost effective method for reducing the direct medical costs of cardiovascular disease in the past, and may be cost saving in the future. The expected cost of producing an additional ex-smoker has been estimated to be approximately $1,000 to $1,500. Most or all of this cost can be recovered in the short run from savings in avoided heart attacks and strokes alone in healthy quitters. Observational studies of the direct medical costs following cessation in those observed to quit show a reduction utilization, but which may occur only after a lag of three to five years.

Comparative quantification of health risks conceptual framework and methodological issues

Reliable and comparable analysis of risks to health is key for preventing disease and injury. Causal attribution of morbidity and mortality to risk factors has traditionally been conducted in the context of methodological traditions of individual risk factors, often in a limited number of settings, restricting comparability.In this paper, we discuss the conceptual and methodological issues for quantifying the population health effects of individual or groups of risk factors in various levels of causality using knowledge from different scientific disciplines. The issues include: comparing the burden of disease due to the observed exposure distribution in a population with the burden from a hypothetical distribution or series of distributions, rather than a single reference level such as non-exposed; considering the multiple stages in the causal network of interactions among risk factor(s) and disease outcome to allow making inferences about some combinations of risk factors for which epidemiological studies have not been conducted, including the joint effects of multiple risk factors; calculating the health loss due to risk factor(s) as a time-indexed "stream" of disease burden due to a time-indexed "stream" of exposure, including consideration of discounting; and the sources of uncertainty.

Selected major risk factors and global and regional burden of disease

BACKGROUND

Reliable and comparable analysis of risks to health is key for preventing disease and injury. Causal attribution of morbidity and mortality to risk factors has traditionally been in the context of individual risk factors, often in a limited number of settings, restricting comparability. Our aim was to estimate the contributions of selected major risk factors to global and regional burden of disease in a unified framework.

METHODS

For 26 selected risk factors, expert working groups undertook a comprehensive review of published work and other sources--eg, government reports and international databases--to obtain data on the prevalence of risk factor exposure and hazard size for 14 epidemiological regions of the world. Population attributable fractions were estimated by applying the potential impact fraction relation, and applied to the mortality and burden of disease estimates from the global burden of disease (GBD) database.

FINDINGS

Childhood and maternal underweight (138 million disability adjusted life years [DALY], 9.5%), unsafe sex (92 million DALY, 6.3%), high blood pressure (64 million DALY, 4.4%), tobacco (59 million DALY, 4.1%), and alcohol (58 million DALY, 4.0%) were the leading causes of global burden of disease. In the poorest regions of the world, childhood and maternal underweight, unsafe sex, unsafe water, sanitation, and hygiene, indoor smoke from solid fuels, and various micronutrient deficiencies were major contributors to loss of healthy life. In both developing and developed regions, alcohol, tobacco, high blood pressure, and high cholesterol were major causes of disease burden.

INTERPRETATION

Substantial proportions of global disease burden are attributable to these major risks, to an extent greater than previously estimated. Developing countries suffer most or all of the burden due to many of the leading risks. Strategies that target these known risks can provide substantial and underestimated public-health gains.