HEALTH EDUCATION RESEARCH Vol.17 no.5 2002Theory & Practice Pages 606–618

Improving health behaviors and outcomes afterangioplasty: using economic theory to inform

intervention

Mary E. Charlson, John P. Allegrante1, Paula S. McKinley2, Janey C. Peterson,Carla Boutin-Foster, Gbenga Ogedegbe and Candace R. Young

Abstract

Patients who have been relieved of cardiacsymptoms following angioplasty may not besufficiently motivated to initiate behaviorchanges that can reduce risk of subsequentcardiac events. Finding an effective means tohelp patients modify their behavior thus pre-sents a unique challenge. This paper describesan innovative behavioral intervention whosetheoretical underpinning is net-present valueeconomic theory. This intervention is beingevaluated in a randomized controlled trial inwhich all patients complete a computerizedbaseline health assessment of 14 cardiovascularrisk factors. Each patient is presented with anindividualized risk-factor profile and asked tochoose risk factors for modification. In theexperimental group, each risk factor is pre-sented with a corresponding numerical biologicage value that represents the relative potentialto benefit from modifying each risk factor. Riskreduction for these patients is framed as theopportunity to reduce present biologic age (thenet-present value), and improve current healthstatus and quality of life. In the control group,

Division of General Internal Medicine and Center forComplementary and Integrative Medicine, Weill MedicalCollege of Cornell University, New York, NY 10021,1National Center for Health Education, and Departmentof Health and Behavior Studies, Teachers College,Columbia University, New York, NY 10027 and2Behavioral Medicine Program, Columbia-PresbyterianMedical Center, New York, NY 10032, USACorrespondence to: J. P. Allegrante

© Oxford University Press 2002. All rights reserved

risk reduction is framed in the standard risk-factor approach as the value of preventingfuture health problems. We hope to demonstratethat economic theory is a plausible perspectivefrom which to design interventions aimed atcommunicating risk and facilitating change inhealth behaviors.

Introduction

Data from clinical trials and statewide registriesof percutaneous transluminal coronary angioplasty(PTCA) show that between 30 and 40% ofangioplasty patients experience recurrent angina,myocardial infarction (MI), death or repeatprocedures by 2 years of follow-up [BARI et al.,1990; Hampton et al., 1993; Rodriguez et al.,1993; Hamm et al., 1994; King et al., 1994;Rogers et al., 1995; The Bypass AngioplastyRevascularization Investigation (BARI) Investig-ators, 1996; Zhao et al., 1996; Hannan et al., 1997;Writing Group for the BARI Investigators, 1997].The high recurrence rates occur because of re-stenosis of the target lesions or because of progres-sion of the disease that results in new coronaryartery lesions. Patients who are male or who haveunstable or recent onset angina have a greaterlikelihood of recurrence in the target vessel(Holmes et al., 1984; Blackshear et al., 1987;Rupprecht et al., 1990). In addition, patients withlong or complex, Grade C lesions are at higherrisk for re-stenosis than those with Grade B lesions[Hamm et al., 1994; Rogers et al., 1995; TheBypass Angioplasty Revascularization Investi-gation (BARI) Investigators, 1996; Writing Groupfor the BARI Investigators, 1997]; diabetics also

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

Using economic theory to inform intervention

have worse long-term outcomes than non-diabetics(Galan and Hollman, 1986; Klugherz et al., 1996;Laham et al., 1996).

Given the high recurrence rates in this popula-tion, an expert committee of the WHO hasrecommended that rehabilitation should be pro-vided for all patients undergoing PTCA (Wenger,1991). Clinicians have expressed concern, how-ever, that PTCA patients compared to othercardiovascular patient populations may be toooptimistic about their future health status.Angioplasty is a procedure that is relatively non-invasive and yields rapid symptom relief. Thesebenefits of PTCA may, paradoxically, interferewith motivation to undertake changes in behaviorsthat contributed to the development of coronaryartery disease (Kimble, 1998). Even when recom-mended changes are adopted, they are not main-tained over the long term (Morocutti et al., 1999).As a result, patients who have undergone angio-plasty may underestimate their risk of future coron-ary artery disease and the potential benefits—bothin the present and future—of modifying their riskbehaviors. Thus, in spite of consensus that behaviorchange is vital for PTCA patients, post-procedurerehabilitation efforts may prove to be especiallydifficult in this population. This may be the reasonwhy risk assessment and behavioral interventionhave not been reliably integrated into post-PTCApractice (Agren et al., 1989; Wright and Strang,1997).

There is, in fact, evidence that patients whohave undergone PTCA tend to be either poorlyinformed about health risk behaviors or unlikelyactually to change their behaviors (Fletcher, 1986;Hansen, 1988; Faris and Stotts, 1990; Gaw, 1992;Gaw-Ens and Laing, 1994; Gulanick et al., 1998).For example, Gaw reported that only 50% ofpatients interviewed could identify specific cardiacrisk factors thought to be related to their cardiovas-cular problems (Gaw, 1992). In a review of 19studies, Gentz found that during the post-PTCArecovery period patients reported the need for moreeducation about risk factors, lifestyle changes andsurvival management (Gentz, 2000).

607

While the necessity for PTCA alone couldbe considered a potential motivator for behaviorchange, patients tend not to alter their healthrisk behaviors without an active intervention. Forexample, patients 1 year after PTCA showed onlysmall reductions in cholesterol (12 mg/dl), smallreductions in body mass (0.41 kg/m2) and anincrease in prevalence of smoking from 5 to 13%(McKenna et al., 1995). These researchers didreport a significant increase in the proportion ofpatients who were engaged in moderate exercise(from 32 to 65%) and suggested that relief fromischemic symptoms after angioplasty may haveenabled patients to be more active.

Thus, in light of PTCA patients’ reticence toadopt and maintain health behavior change, thispopulation presents a unique challenge: can a moreeffective means of communicating future healthrisks and motivating PTCA patients to initiate andmaintain health behavior changes be developed,and what theoretical approach should inform thedesign of such an intervention?

We have taken the perspective originally pro-posed by Allegrante and Roizen that the studyof human economic behavior—specifically net-present value economic theory—holds promise asa theoretical underpinning for behavioral interven-tions applied to health (Allegrante and Roizen,1998). This paper describes a randomized con-trolled intervention trial of PTCA patients thattests whether this approach will enhance motivationto make behavioral changes. By demonstrating thatprinciples underlying behavior in the economicrealm can serve as the theoretical basis for interven-tion, this project promises to link the work ofbehavioral psychology with that of economics.

Previous intervention approaches

The majority of behavioral interventions forpatients with coronary artery disease have targetedsubjects after surgical coronary artery bypass graft(CABG) or MI (Hedback and Perk, 1987). Multi-factorial cardiac risk reduction interventions thathave used cardiac endpoints generally have beensuccessful in lowering event rates in the treatment

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

M. E. Charlson et al.

groups (Brown et al., 1990; Buchwald et al., 1990;Ornish et al., 1990; Schuler et al., 1992; Wattset al., 1992). In spite of this success, there is littleresearch targeting multifactorial lifestyle behaviorchange in patients following angioplasty.

Most studies targeting angioplasty outcomeshave either been epidemiological assessments ofrisk factors for re-stenosis or studies of the effectsof medical but not behavioral interventions onclinical outcomes (Hollman et al., 1989; Aroraet al., 1990; Benchimol et al., 1990, 1993). Todate only two major randomized controlled trialshave assessed the efficacy of multifactorialbehavioral intervention programs for PTCApatients.

In one of these trials a Swedish team developedan intervention program targeting stress manage-ment, and changes in diet, exercise and smokingbehaviors (Hofman-Bang et al., 1999; Lissperset al., 1999). The intervention consisted of anintense residential program of health education,behavior change training and planning for futurelifestyle changes. The second phase was an 11-month maintenance program with regular contactsbetween the patient and a nurse acting as a personalcoach. As a key component of the intervention,the coaches developed an individualized programfor each patient based on his or her pre-PTCAhealth behaviors and stress profile. Patients in thecontrol group were asked to stay in contact withtheir own physician and were neither encouragednor discouraged to modify lifestyle risk factors.

Immediately after the 12-month intervention,there were significant improvements in body massindex, cholesterol levels, exercise capacity andfrequency, and dietary knowledge and habits; butthere were no group differences in clinical events.At 1-year follow-up the intervention group hadmaintained their more positive health behaviorsand had had fewer hospital admissions for cardiacevents compared to the control group. There wasno group difference, however, in the percentage ofpatients who had returned to work.

In the other study, the Stanford Coronary RiskIntervention Project (SCRIP), the interventioninvolved an intensive, long-term multiple risk-

608

reduction program targeting several factors,including diet, exercise, weight loss, smoking andmedications among patients with PTCA or CABG(Haskell et al., 1994). Like the Swedish group’sstudy, this program also featured an individualizedintervention. Staff nurses and dieticians instructedparticipants individually on risk factors includingbody weight, blood pressure, cholesterol level,dietary fat intake, sodium intake, smoking andphysical activity. Standard program goals were setfor each risk factor and applied to all participants.Emphasis was placed on each patient in the studyreaching the maximum program goal for each riskfactor, while minimum and intermediate goalswere applied selectively to enhance motivation andprogram adherence.

Risk-reduction participants returned to the clinicevery 2–3 months for progress reports and clinicalmeasures. Medical and risk factor evaluations werecompleted at baseline and yearly for 4 years.After 4 years patients in the treatment grouphad decreased narrowing of coronary arteries, asconfirmed by arteriograms, and fewer hospitaliza-tions for clinical cardiac events.

Neither of these two research teams describesspecifically the theoretical bases of their programs.Both programs included intervention methods thatappear to have been informed by at least twopopular behavioral theories: Social LearningTheory (Bandura, 1977), and the TranstheoreticalModel and Stages of Change (Prochaska et al.,1992). Some of the intervention techniquesincluded both education about risk factors andpractical training in how to achieve behaviorchange. Individualization, a key feature of effectivebehavior change programs, was also used. Hence,both these programs tailored certain features ofthe intervention to individuals’ risk profiles andbehavior change preferences. It is unclear, however,which components of these programs were mosteffective and why. While the intervention effectsof these studies were modest, such results arepromising.

Thus, previous approaches to intervention havefocused on presenting risk-factor information topatients that helps them to understand how making

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

Using economic theory to inform intervention

behavioral changes will reduce future risk. Suchapproaches have not provided a means by whichthe patient can prioritize lifestyle changes in termsof relative risks and such approaches have failedto give the patient or his or her physician a wayto place a tangible value on the change in relativerisk. Given the challenges health practitioners facein convincing angioplasty patients of the need tomake immediate changes to stem the progressionof the disease, it would appear that such patientsare in need of a new intervention approach thatframes the value of behavior modification in termsof present and direct benefit.

The Healthy Behaviors Trial

The Healthy Behaviors Trial is a randomizedcontrolled trial designed to compare the efficacyof two intervention strategies—a standard careapproach versus an experimental biologic-ageapproach—of motivating behavioral change toimprove health behavior and health outcomes (i.e.2-year survival free from MI, stroke, Class II–IVangina or severe ischemia on non-invasive testing)among coronary artery disease patients who haveundergone angioplasty with or without stents. Theobjective of this study is to promote behaviormodification in this patient population throughindividualized risk profile feedback, and to testwhat we believe is an innovative and potentiallyefficacious motivational approach to risk factoreducation.

At baseline all participants complete a com-puterized health assessment that evaluates 14cardiac risk factors, including such factors asphysical activity, smoking, diet, blood pressure andmedications. Table I lists the risk factors targetedfor intervention. Based on the results of theirassessment, all patients choose two to three factorsfor modification. Patient education about each riskfactor in an individualized profile is framed usingeither a standard approach or an experimentalbiologic-age approach where each factor isweighted according to its relative risk. To getpatients started, we provide them with behavior-change ‘tip sheets’ for each of the risks they have

609

Table I. Risk factors targeted for behavior change

Physical activity1. increase physical activity2. increase aerobic exercise3. increase strength training

Smoking4. stop smoking (or continue not to smoke)5. increase servings of fish6. reduce red meat7. reduce dietary intake of cholesterol and saturated fat8. increase fiber-rich food9. increase flavonoid-rich food

10. increase folic acid intake

Blood pressure11. control/reduce blood pressure

Medications12. take aspirin13. take β-blockers/reduce heart rate14. take statins

selected to change, as well as referrals for behavior-change programs and other community resources.Both groups of patients are followed and inter-viewed by telephone every 3 months for 2 years.The telephone contact provides motivationalsupport to patients in both groups, and we usebasic principles of motivational interviewing thathave been developed by Emmons and Rollnick(Emmons and Rollnick, 2001) during the callto assist patients in adopting and maintainingbehavioral changes. In addition, each call includesan updated assessment of interval clinical eventsand changes in cardiovascular risk profile, as wellas an opportunity for the individual patient tochange or add new risk factors. We also assessstages of behavioral change and self-efficacy dur-ing these calls. A total of 660 patients will beenrolled.

The trial is thus designed to investigate theefficacy of a novel motivational approach in whicha computerized health risk assessment presentsindividuals with the explicit net-present value offuture health actions. Using this approach, a com-mon health risk currency, which we believe to bean important, culturally relevant metric of benefitthat people can understand—biologic age—has

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

M. E. Charlson et al.

been derived from the change in life expectancyin Kaplan–Meyer (Cox, 1972) survival curves andfrom studies indicating that changes in lifestylefactors, in turn, can alter survival curves. However,unlike previous first-generation health-riskappraisal and computerized risk assessments, ourapproach utilizes a larger database of epidemiologicevidence to arrive at an estimation of biologic age,which can then be contrasted with the patient’schronologic age.

We believe this is worth investigating becauseonly a few previous studies have attempted tomodel age-specific and population-specific benefitexpectations that result from risk-factor modifica-tion. Benefit in terms of prevention of disease-related morbidity and mortality has been measuredas gains in life expectancy (Tsevat et al., 1991;Welch et al., 1996; Wright and Weinstein, 1998).In particular, Wright and Weinstein found gains inlife expectancy of 6.2 months from exercise and28–34 months from quitting cigarette smoking inpopulations at elevated risk for cardiovasculardisease (Wright and Weinstein, 1998). Roizen andGoetz have reported using the concept of whatthey call RealAge©, a measure of the physiologicimprovement that results from interventionaffecting health status, in order to translate thepresent value of healthy choices and motivatebehavior change (Roizen and Goetz, 1998). Theybelieve that patients respond to the idea of slowingdown the aging process or becoming physiologic-ally younger, and that this model allows bothpatients and physicians to make more rationalhealth-related choices based on the preference ofpatients to be as young as they can be. The studydescribed here is the first to apply the concept ofbiologic age to a sick population (M. F. Roizen,pers. commun.).

The experimental condition: the biologic-age approach

In the experimental group, patients receive theircurrent biologic age based on their risk profile andare able to contrast this with their chronologic age.Patients are presented with an individualized risk-factor profile listing potential areas for behavior

610

change, each of which carries a differential benefitfor reducing biologic age. Risk reduction is framedas the opportunity to reduce one’s present biologicage and improve one’s current quality of life. Forexample, in an otherwise healthy 50-year-old malewho has been smoking a pack of cigarettes eachday for 25 years, this lifestyle risk makes hisbiologic age 54 years old. Thus, the net-presentvalue of such a man quitting smoking is approxi-mately a 4-year reduction in his biologic age.

Patients in the experimental intervention groupare presented with the numerical value representingthe relative potential to benefit from modifyingeach corresponding behavior. These patients areshown the potential impact that adopting eachbehavior can have on lowering their biologic agevalue and are then asked to choose two or threebehaviors for modification in order to reduce theirbiologic age. In addition, patients receive standard-ized educational materials as do patients enrolledin the control group; however, patients in theexperimental arm of the trial receive a biologic-age equivalent for each factor. Thus, although thecontent of the information about risk is identicalfor both groups, the presentation of the potentialreduction in biologic age that is associated withsuccessful modification of each individual riskfactor is unique to our experimental group.

The health risk assessment used in this study isa modified version of the RealAge© computerizedrisk assessment developed and popularized byMichael Roizen (Roizen, 1998), and a team ofscientists and software developers. RealAge©

utilizes the concept of biologic age that was firstintroduced by Sadusk and Robbins (Sadusk andRobbins, 1968). The first-generation of health risk(or hazard) appraisals was developed by the USCenters for Disease Control in the 1970s. Later,the Carter Center, Control Data Corporation andMedical Datamation were among those whodeveloped commercial versions of these earlyhealth risk appraisal instruments for large-scaleapplications in population screening. Severalreports (Wagner et al., 1982; Schoenbach et al.,1987) have sought to assess the utility of healthrisk appraisal. Although these early health risk

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

Using economic theory to inform intervention

appraisal instruments all used some form of bio-logic age as the outcome of health risks, theseinstruments were based on databases of relativerisk of all-cause mortality that were derived largelyfrom observational epidemiologic studies ratherthan randomized controlled trials on modifyingrisks.

RealAge© measures a large number of healthrisk factors, ranging from cardiovascular risk fac-tors such as smoking, diet and exercise, to variedother risk factors, such as family health history,driving habits, social support and stressful lifeevents, all of which can be used to predict mortality.The algorithm for calculating biologic age is basedon the use of relative risk of all-cause mortalityvalues from Kaplan–Meyer survival curves foreach factor assessed. Roizen (Roizen, 1998) andcolleagues (Rhodes et al., 2001; Roizen and Goetz,1998) report having drawn these relative-riskvalues from randomized, controlled studies of theeffects of the various risk factors on life expectancyand the benefit of changes in lifestyle factorson survival.

Similar to other health risk assessments, theRealAge© software calculates biologic age by com-bining data from a person’s health history andcurrent health habits to estimate his or her riskrelative to the person’s age cohort. Before anindividual answers a single question, his or herbiologic age is the same as his or her calendarage. The response categories for each health riskfactor in the program are weighted using a relative-risk statistic drawn from published research. Eachtime the individual provides information regardinghealth status, these relative risks modify the currentestimate of the individual’s biologic age.

With permission from Dr Roizen, we modifiedthe assessment instrument for this study in severalways. First, the assessment of risks was reducedto assess only those modifiable risk factors targetedfor intervention in this study, as well as relevantunmodifiable risk factors such as medical comor-bidities and family medical history. Table II showsthe risk factors and other variables of interestbeing assessed in this trial, how each variable ismeasured, and the behavior change goal associated

611

with the risk factors. Second, we capped themaximum biologic age for each chronologic ageby a constant conversion factor (C. McCulloch,pers. commun.). This change prevents the estimateof biologic age from being inflated in our studypopulation due to the high incidence of priorcardiovascular events and other medical co-morbidities. Third, we updated the nutritionalvalues of all foods listed in the dietary section ofthe assessment using values drawn from publishedjournal articles [e.g. (Hertog et al., 1992, 1993)]and the unpublished databases of Rimm (E. B.Rimm, pers. commun.).

The control condition: the standard careapproach

At baseline, patients in the control condition com-plete an identical survey of cardiovascular andlifestyle risk factors as those in the experimentalgroup. In the control group, the results of patientsurvey responses are compiled into an indivi-dualized list of risk factors. Patients are then givenrecommendations to make behavioral modifica-tions based on this list of cardiovascular riskfactors. Risk profile information is framed andpresented in the standard risk reduction approachas the potential to reduce risk of future negativeoutcomes. For example, a patient is told that bystopping smoking, he or she may reduce his or herfuture risk of subsequent events. Patients are thenasked to choose two to three areas of risk to targetfor behavioral modification. In our study, we givepatients a set of standard patient educationalmaterials tailored for each risk factor that empha-sizes the future value of the specific behavioralchange.

The theoretical approach: net-presentvalue economic theory

A variety of theoretical models have guided thedevelopment of behavioral and educational inter-ventions in clinical and community settings, includ-ing cardiac rehabilitation programs. These includethe Health Belief Model (Rosenstock, 1966;Becker, 1974), Theory of Planned Behavior (Ajzen

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

M. E. Charlson et al.

Table II. Variables under study, measurement and behavior change goals targeted in the intervention

Variable Measurement Behavior change goal

Risk factors1. overall physical total weekly kilocalories of exercise total weekly kilocalories of

activity exercise2. aerobic exercise total weekly minutes of aerobic exercise activities �30 min/week3. strength exercise total weekly minutes of resistance exercise activities �30 min/week4. smoking total pack-years to date; current cigarettes per day; time since quitting stop smoking or continue not

smoking to smoke5. dietary fish number of dietary 3-oz servings/week �1 serving/week6. dietary red meat number of dietary 3-oz servings/week �1 serving/week7. dietary fiber daily dietary fiber intake (g) �23.2 g/day8. dietary flavonoids daily dietary flavonoid intake (mg) �30 mg/day9. dietary folic acid daily dietary intake of folic acid (folate) from diet and vitamin 600 µg/day

(folate) supplements (µg)10. blood pressure (BP) baseline: current BP in hospital chart; follow-ups: self-report BP (mmHg) �120/80 mmHg11. low-dose aspirin daily aspirin intake (mg) men only: 1 aspirin (325 mg)

or baby aspirin(81 mg)/day

12. heart rate average heart rate based on medical chart (baseline) or self-report �75 b.p.m.(follow-ups) ( b.p.m.)

13. blood sugar control for hemoglobin A1C level or self-report estimate of low-average-high blood tight control: � hemoglobindiabetics sugar A1C

14. dietary cholesterol and daily dietary intake from all foods (mg) male �3 mg/day; femalesaturated fat �52 mg/day

daily dietary intake per 1000 kilocalories (mg) �9 mg/day15. serum cholesterol baseline: medical chart; follow-ups: self-report; 24-month follow-up: age �70: CHL � 160, HDL

physical exam �59/age � 70: CHL �180,HDL �59

16. dietary polyunsaturated daily dietary polyunsaturated fat per day from all foods (mg) �7.5 mg/dayfat intake

17. overweight status body mass index (BMI) from self-report height and weight; and (for male BMI � 25.3/femalewomen only) whether weight has ever cycled �10% within 5-year BMI � 26.99interval

Mediators1. comorbidity Charlson Comorbidity Index none2. stages of change single-item rating of five stages, which is obtained via interview, and is none

assessed for each of the targeted risk factor behaviors listed above3. self-efficacy single-item confidence rating for each behavior, which is obtained via none

interview, and is assessed for each of the targeted risk factor behaviorslisted above

4. depression CESD (Center for Epidemiologic Studies of Depression) none

Outcomes (at 24-month follow-up)1. mortality all deaths of any cause prior to 24 months of follow-up will be none

documented2. myocardial infarction resting ECG; definite myocardial infarction prior to 24 months of follow- none

up, based on the 1968 revision of the Minnesota code3. angina class II–IV anginal symptoms based on the Rose criteria none4. severe ischemia severe Ischemia during non-invasive stress testing none5. stroke occurrence of a new major focal neurologic deficit that persisted more none

than 24 h and includes cortical blindness, persistent coma, hemiplegia,hemiparesis, aphasia, sensory changes or hemianopsia

6. functional status MOS SF-36 (Medical Outcomes Study Short Form-36) Health Survey, nonewhich evaluates seven domains of function: physical, social, role, energy,mental, pain and general health

7. subsequent clinical any subsequent cardiac catheterization, angioplasty/stents or coronary noneprocedures artery surgery during the 2 years of follow-up

The ‘behavior change goal’ values are based on the values used in the survey as cutoffs for ‘success’. They are embedded in the databasesthat guide the program in making decisions about which recommendations are generated for the individual.CHL, cholesterol; HDL, high-density lipoporotein.

612

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

Using economic theory to inform intervention

and Fishbein, 1980; Ajzen and Madden, 1986),Social Learning Theory (Bandura, 1977) and theTranstheoretical Model of Stages of Change(Prochaska et al., 1992). All of these theories arecognitive-behavioral models of behavior and eachis grounded in the value-expectancy formulationof social and behavioral psychology that wasoriginally developed by Lewin (Lewin, 1935). Asuseful as these theories have proved to be inguiding intervention efforts, one notable limitationis that none of the theories presents an easilyunderstood and culturally relevant metric by whichpeople can evaluate the utility of making behavioralchanges. We believe that without such a metric,people cannot operationalize value expectancy ina clinically or personally meaningful way. It is forthis reason that we have turned to a theory ofeconomic behavior to gain new and potentiallyuseful insight into what might better motivatehealth behavior changes.

The notion of economics as a potentially influ-ential factor in behavioral motivation andbehavioral change has only recently caught theinterest of health scientists. Several reports havenow examined economics as a factor in drugaddiction (Marlatt and Kilmer, 1998; Bretteville-Jensen, 1999) and cigarette smoking (Bickel andMadden, 1999; Suranovic et al., 1999). The bulk ofthis work has been based on Becker and Murphy’s(Becker and Murphy, 1988) theory of rationaladdiction. The principal concepts from behavioraleconomics that have been identified by Camerer(Camerer, 1999)—expected utility and risk,exponential discounting of future costs and bene-fits, maximization of personal utility versus socialutilities, and equilibrium—have all been shown inthis work to be critical to making choices. More-over, because such economic concepts havepsychological foundations in the behavioral prin-ciples of adaptation, loss-aversion, generalizedreinforcement and the human preferences forimmediacy, they need to be taken into accountwhen designing an intervention to change healthbehavior. Thus, integrating economic and psy-chosocial models could be useful to understand

613

cigarette smoking, drug use and other health-related behaviors (Montoya et al., 2000).

At the core of Becker and Murphy’s model isthe work on net-present value economic theory ofthe University of Chicago economist Gary Becker(Becker, 1976). Becker has demonstrated that thereis convergent empirical evidence that net-presentvalue economic theory has a strong motivationalinfluence on individuals’ economic decisions andother behaviors. Conceptually, net-present valueessentially takes the future value of a product orinvestment and relates it to what it is worthin present-dollar terms by discounting based onexpectations of interest rate, inflation and demand.Another University of Chicago economist, MiltonFriedman (Friedman, 1963), was one of the firstto recognize the implications of motives and prefer-ences in economic behavior. He suggested thatwith longer delays, rewards could eventually bediscounted into insignificance. This means that themaximum planning horizon for rational investorswould not exceed 3 years.

When these concepts are applied to other areasof decision making, they may help to explainwhy individuals who must make decisions aboutundertaking complex changes in health behaviortend to undervalue the potential future gain of thebehavior change in present terms. For example,the individual who has had angioplasty and mustcontrol high blood pressure to avoid subsequentevents may experience side effects of anti-hypertensive medication today. As a result ofthese immediate side effects, that individual maydiscount the value of controlling his high bloodpressure to zero because the perceived benefit—especially in light of the more immediate sideeffects of treatment—is a distant benefit. Thus, itbecomes critical to explain the health benefits ofa recommended behavioral change in present valueor present health terms. Moreover, the benefits ofa medical treatment or behavior change mustbe perceived as being achievable in the nearfuture term.

The novel theoretical concept in our interventionis the economic perspective on human behaviorthat we have chosen to take in motivating and

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

M. E. Charlson et al.

facilitating health behavior change. The interven-tion we have developed attempts to motivatebehavior change by framing risk reduction as theopportunity to reduce one’s present biologic age.This approach is contrasted with the standardmotivation to avoid some future risk of events.Our intervention incorporates proven conceptsfrom well-established models of behavior change,including those of individualized feedback, stagesof change and self-efficacy. Unlike other similarinterventions, however, it standardizes the valueof behavior change based on what economic theoryhas to say about what motivates people in makingdecisions about the future. Unlike a previousgeneration of health risk appraisals, we believe themetric by which we are attempting to communicaterisk and the potential benefit of reducing risk—theconcept of biologic age—has powerful connota-tions in a culture that values and reinforces theeternal search for the ‘fountain of youth’.

Figure 1 presents our representation of the logicmodel for the hypothesized intervention effects weexpect our approach to yield. The model suggeststhat knowing and comparing the biologic agevalues attached to various health risk behaviorswill be effective motivators for behavior change.The age values will demonstrate to patients inunderstandable terms the relative detriment ofvarious risk behaviors and relative worth ofchanging them. We have hypothesized that patientswill compare the potential age reduction of varioushealth behavior changes and choose to changebehaviors with the greatest payoff, i.e. those withthe highest relative risk for morbidity and mortality.Patients in the control group will have no informa-tion with which to compare the potential benefitof various changes, so they are likely to choosebehaviors they feel are easier to change. Overall,then, the control group’s behavior change choicesare expected to be more varied and less effective.As a result, the effect of knowing one’s biologicage on clinical health outcomes will be mediatedby more effective health behavior changes.

The behavior change choices and successes ofboth groups will be moderated by the known impactof self-efficacy, stages of change and medical

614

comorbidities, but we expect that the biologic-ageapproach will significantly enhance self-efficacyand willingness to advance to an ‘action’ or ‘main-tenance’ stage of change. Patients’ expectationsthat they can re-gain some lost vitality and feel‘younger’ soon, as opposed to extending theirlifespan at the end of life, will also serve tomotivate maintenance of health behaviors longerfor the experimental group, thereby mediatingimproved long-term health outcomes.

Conclusion

While numerous theoretical approaches have beenutilized to change health behavior in various patientpopulations, the application of constructs frombehavioral economics to the design of a behavioraland educational intervention in cardiac rehabilita-tion is novel. This is the first randomized controlledtrial to operationalize an intervention approach thatapplies economic concepts of human behavior toa model of motivating health behavior change ina physiologically compromised population. Thesecond-generation computerized expert-systemhealth risk assessment that is being utilized in thiswork draws on established principles of healthrisk assessment to assess diet, physical activity,smoking and other risk factors of the patientpopulation. Unlike previous approaches, however,it calculates and then presents individuals with theunique, explicit net-present value of their futurehealth actions through the application of theconcept of biologic age.

There are two particularly novel aspects of thiscalculation compared to other health risk appraisalinstruments. First, the health risk assessment pre-sents the relative-risk effects on predicted mortalityin terms of current age units. For example, a 50-year-old person with several risk factors may havea current biologic age of 54. Other health riskappraisals essentially provide the same informa-tion, but present it as years added to life, expectedage of death or some other metric that too oftencan be interpreted by the individual at risk ashaving consequences too far into the future to have

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

Using economic theory to inform intervention

Fig. 1. Logic model for the hypothesized intervention effect of biologic age on health behavior change.

impact on his or her current health choices. Second,the biologic-age assessment also calculates theexpected change in biologic age within a 3-yearhorizon for each risk factor, if an individualchanged their risk status on that factor throughbehavior change. Based on net-present valueeconomic theory (and unlike previous approachesto health risk appraisal), these short-term horizonpredictions give the individual an achievable goalthat has current motivational value. The net-present

615

value approach, thus, attempts to motivate theindividual to make health choices based on whateconomists have learned about human behaviorand rational choices by taking the future gain insurvival outcome and translating it into net-presentyears of age. It is in this way that we hypothesizethat it helps motivate people by informing eachindividual of explicit and realistic estimates ofthe present value of future gains afforded bybehavioral change.

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

M. E. Charlson et al.

We believe that this project promises to providean empirical demonstration of the potential ofnet-present value economic theory as a plausibletheoretical perspective from which to design inter-ventions aimed at communicating risk, andmotivating the adoption and maintenance of com-plex health behaviors. Such a demonstration couldhelp build an important conceptual and scientificbridge between behavioral psychology, social cog-nition and learning—all of which have comprisedthe dominant theoretical perspectives underlyingprevious efforts to facilitate behavior change incardiac rehabilitation and other health promotionprograms—and behavioral economics.

Acknowledgements

The authors thank Lynn Burrell and Ray Marksfor their assistance in preparing the manuscript.This work is funded by a grant from the NationalHeart, Lung, and Blood Institute, 1R01 HL62161,and is one of the projects of the NIH BehaviorChange Consortium.

References

Agren, B., Olin, C., Castenfors, J. and Nilsson-Ehle, P. (1989)Improvements of the lipoprotein profile after coronary bypasssurgery: additional effects of an exercise training program.European Heart Journal, 1, 451–458.

Ajzen, I. and Fishbein, M. (1980). Understanding Attitudesand Predicting Social Behavior. Prentice-Hall, EnglewoodCliffs, NJ.

Ajzen, I. and Madden, T. J. (1986) Prediction of goal-directedbehavior: attitudes, intentions, and perceived behavioralcontrol. Journal of Experimental Social Psychology, 22,453–474.

Allegrante, J. P. and Roizen, M. F. (1998) Can net-presentvalue economic theory be used to explain and change health-related behaviors? Health Education Research, 13, i–iv.

Arora, R. R., Konrad, K., Badhwar, K. and Hollman, J. (1990)Re-stenosis after transluminal coronary angioplasty: a riskfactor analysis. Catheterization and CardiovascularDiagnosis, 19, 17–22.

Bandura, A. (1977) Social Learning Theory. Prentice-Hall,Englewood Cliffs, NJ.

BARI, CABRI, EAST, GABI and TRIA (1990) Coronaryangioplasty on trial. Lancet, 335, 1315–1316.

Becker, G. S. (1976) The Economic Approach to HumanBehavior. University of Chicago Press, Chicago, IL.

Becker, G. S. and Murphy, K. M. (1988) A theory of rationaladdiction. Journal of Political Economy, 96, 675–700.

616

Becker, M. H. (1974) The Health Belief Model and personalhealth behavior. Health Education Monographs, 2 (SpecialIssue).

Benchimol, D., Benchimol, H., Bonnet, J., Dartigues, J. F.,Couffinhal, T. and Bricaud, H. (1990) Risk factors forprogression of atherosclerosis six months after balloonangioplasty of coronary stenosis. American Journal ofCardiology, 65, 980–985.

Benchimol, D., Bonnet, J., Benchimol, H., Drouillet, F., Duplaa,C., Coufinhal, T., Desgranges, C. and Bricaud, H. (1993)Biological risk factors for re-stenosis after percutaneoustransluminal angioplasty. International Journal ofCardiology, 38, 7–18.

Bickel, W. K. and Madden, G. J. (1999) A comparison ofmeasures of relative reinforcing efficacy and behavioraleconomics: cigarettes and money in smokers. BehaviouralPharmacology, 10, 627–637.

Blackshear, J. L., O’Callaghan, W. G. and Califf, R. M.(1987) Medical approaches to prevention of re-stenosis aftercoronary angioplasty. Journal of the American College ofCardiology, 9, 834–848.

Bretteville-Jensen, A. L. (1999) Addiction and discounting.Journal of Health Economics, 18, 393–407.

Brown, G., Albers, J. J., Fisher, L. D., Schaefer, S. M., Lin, J.T., Kaplan, C., Zhao, X. Q., Bisson, B. D., Fitzpatrick, V. F.and Dodge, H. T. (1990) Regression of coronary arterydisease as a result of intensive lipid-lowering therapy in menwith high levels of apolipoprotein B. New England Journalof Medicine, 323, 1289–1298.

Buchwald, H., Varco, R. L., Matts, J. P., Long, J. M., Fitch, L.L., Campbell, G. S., Pearce, M. B., Yellin, A. E., Edmiston,W. A., Smink, R. D., Jr, Sawin, H. S., Jr, Campos, C. T.,Hansen, B. J., Tuna, N., Karnegis, J. N., Sanmarco, M. E.,Amplatz, K., Castaneda-Zuniga, W. R., Hunter, D. W.,Bissett, J. K., Weber, F. J., Stevenson, J. W., Leon, A. S.,Chalmers, T. C. and POSCH Group. (1990) Effect of partialileal bypass surgery on mortality and morbidity from coronaryheart disease in patients with hypercholesterolemia—reportof the Program on the Surgical Control of the Hyperlipidemias(POSCH). New England Journal of Medicine, 323, 946–955.

Camerer, C. (1999) Behavioral economics: reunifyingpsychology and economics. Proceedings of the NationalAcademy of Sciences, 96, 10575–10577.

Cox, D. R. (1972) Regression models and life tables. Journalof the Royal Statistical Society, 34, 187–220.

Emmons, K. M. and Rollnick, S. (2001) Motivationalinterviewing in health care settings. Opportunities andlimitations. American Journal of Preventive Medicine, 20,68–74.

Faris, J. A. and Stotts, N. A. (1990) The effect of percutaneoustransluminal angioplasty on quality of life. Progress inCardiovascular Nursing, 5, 132–140.

Fletcher, G. F. (1986) Rehabilitation after coronaryangioplasty—is it effective? Archives of Physical Medicineand Rehabilitation, 67, 517–519.

Friedman, M. (1963) Windfalls, the ‘horizon’, and relatedconcepts in the permanent-income hypothesis. In Christ,C. F., Friedman, M., Goodman, L. A., Griliches, Z.,Harberger, A. C., Liviatin, N., Mincer, J., Mundlak, Y.,Nerlove, M., Patinkin, D., Telser, L. G. and Theil, H.(eds), Measurement in Economics: Studies in Mathematical

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

Using economic theory to inform intervention

Economics and Econometrics in Memory of Yehuda Grunfeld.Stanford University Press, Stanford, CA, pp. 3–28.

Galan, K. M. and Hollman, J. L. (1986) Recurrent of stenosesafter coronary angioplasty. Heart and Lung, 15, 585–587.

Gaw, B. L. (1992) Motivation to change lifestyle followingPTCA. Dimensions of Critical Care Nursing, 11, 68–74.

Gaw-Ens, B. and Laing, G. P. (1994) Risk factor reductionbehaviors in coronary angioplasty and myocardial infarctionpatients. Canadian Journal of Cardiovascular Nursing, 5,4–12.

Gentz, C. A. (2000) Perceived learning needs of the patientundergoing coronary angioplasty: an integrative review ofthe literature. Heart and Lung, 29, 161–172.

Gulanick, M., Bliley, M., Perino, B. and Keough, V. (1998)Recovery patterns and lifestyle changes after coronaryangioplasty: the patient’s perspective. Heart and Lung, 27,253–262.

Hamm, C. W., Reimers, J., Ischinger, T., Rupprecht, H. J.,Berger, J. and Bleifeld, W. (1994) A randomized study ofcoronary angioplasty compared with bypass surgery inpatients with symptomatic multivessel coronary disease.German Angioplasty Bypass Surgery Investigation (GABI).New England Journal of Medicine, 331, 1037–1043.

Hampton, J. R., Henderson, R. A., Julian, D. G., Parker, J.,Pocock, S. J., Sowton, E., Wallwork, J., Chamberlain, D. A.,Dark, J. F., Joy, M. D., Seed, P., Youard, B., Yates, A. K.,Curry, P. V. L., Deverall, P. B., Jackson, G., Pumphrey,C. W., Parker, D. J., Treasure, T., Pepper, J., Smith, J., Ward,D. E., Brooks, N., Moussalli, H., Bennett, D., Bray, C.,Campbell, C., Deiraniya, A. K., Jones, M., Lawson, R.,Rahman, A., Ward, C., Balcon, R., Magee, P., Layton, C.,Timmis, A. D., Wright, J. E. C., Swanton, R. H., Pugsley,W., Hubner, P. J. B., Firmin, R. K., Gershlick, A. H., Spyt,T., Okeffe, D. B., Okane, H., Cleland, J., Gladstone, D. J.,Morton, P., Murtagh, J. G., Scott, M. E., Cobb, S. M.,Wheatley, D. J., Dunn, F. G., Hutton, I., Lorimer, A. R.,Rae, A. P., Hills, W. S., Dymond, D. S., Edmondson, S. J.,Banim, S. O., Davies, D. W., Nathan, A. W., Reece, G.,Spurrell, R. A. J., Cumberland, D. C., Smith, G. H., Oakley,G. D. G., Jewitt, D. E., Keates, J. R. W., Bucknall, C. A.,Forsyth, A. T., Wainwright, R. J., Shabbo, F. P., Oriorden, J. B.,Bloomfield, P., Cameron, E. W. J., Debono, D., Rothman,M. T., Wood, A. J., MacDonald, A. H., Mills, P. G.,Lewis, T., Foale, R. A., Stanbridge, R., Fox, K. A. A. andBreckenridge, I. M. (1993) Coronary angioplasty versuscoronary-artery bypass-surgery—The Randomized Interven-tion Treatment of Angina (RITA) Trial. Lancet, 341, 573–580.

Hannan, E. L., Racz, M., Ryan, T. J., McCallister, B. D.,Johnson, L. W., Arani, D. T., Guerci, A. D., Sosa, J. andTopol, E. J. (1997) Coronary angioplasty volume–outcomerelationships for hospitals and cardiologists. Journal of theAmerican Medical Association, 277, 892–898.

Hansen, V. L. (1988) Compliance with risk factor reductionamong post coronary bypass surgery and post coronaryangioplasty patients. Applied Nursing Research, 1, 94.

Haskell, W. L., Aldeman, E. L., Fair, J. M., Maron, D. J.,Mackey, S. F., Superko, H. R., Williams, P. T., Johnstone, I.M., Champagne, M. A. and Krauss, R. M. (1994) Effectsof intensive multiple risk factor reduction on coronaryatherosclerosis and clinical cardiac events in men and women

617

with coronary artery disease. The Stanford Coronary RiskIntervention Project (SCRIP). Circulation, 89, 975–990.

Hedback, B. and Perk, J. (1987) 5 year results of acomprehensive rehabilitation programme after myocardialinfarction. European Heart Journal, 8, 234–242.

Hertog, M. G. L., Hollman, P. C. H. and Katan, M. B. (1992)Content of potentially anticarcinogenic flavonoids of 28vegetables and 9 fruits commonly consumed in theNetherlands. Journal of Agricultural and Food Chemistry,40, 2379–2383.

Hertog, M. G. L., Hollman, P. C. H. and Van de Putte, B.(1993) Content of potentially anticarcinogenic flavonoids oftea infusions, wines, and fruit juices. Journal of Agricultureand Food Chemistry, 41, 1242–1246.

Hofman-Bang, C., Lisspers, J., Nordlander, R., Nygren, A.,Sundin, O., Ohman, A. and Ryden L. (1999) Two-year resultsof a controlled study of residential rehabilitation for patientstreated with percutaneous transluminal coronary angioplasty:a randomized study of a multifactorial programme. EuropeanHeart Journal, 20, 1465–1474.

Hollman, J., Badhwar, K., Beck, G. J., Franco, I. andSimpfendorfer, C. (1989) Risk factors for recurrent stenosisfollowing successful coronary angioplasty. ClevelandClinical Journal of Medicine, 56, 517–523.

Holmes, D. R., Vlietstra, R. E., Smith, H. C., Vetrovec, G. W.,Kent, K. M., Coweley, M. J., Faxon, D. P., Gruentzig, A.R., Kelsey, S. F., Detre, K. M., Vanraden, M. J. and Mock,M. B. (1984) Restenosis after percutaneous trans-luminalcoronary angioplasty (PTCA): a report from the PTCARegistry of the National Heart, Lung, and Blood Institute.American Journal of Cardiology, 53, 77C–81C.

Kimble, L. P. (1998) Cognitive appraisal and cardiac riskreduction behavior following coronary angioplasty. WesternJournal of Nursing Research, 20, 733–744.

King, S. B., Lembo, N. J., Weintraub, W. S., Kosinski, A. S.,Barnhart, H. X., Kutner, M. H., Alazraki, N. P., Guyton, R.A. and Zhao, X. Q. (1994) A randomized trial comparingcoronary angioplasty with coronary bypass surgery. EmoryAngioplasty versus Surgery Trial (EAST). New EnglandJournal of Medicine, 331, 1044–1050

Klugherz, B. D., DeAngelo, D. L., Kim, B. K., Herrmann, H.C., Hirshfeld, J. W. and Kolansky, D. M. (1996) Three-yearclinical follow-up after Palmaz–Schatz stenting. Journal ofthe American College of Cardiology, 27, 1185–1191.

Laham, R. J., Carrozza, J. P, Berger, C., Cohen, D. J., Kuntz,R. E. and Baim, D. S. (1996) Long-term (4- to 6-year)outcome of Palmaz–Schatz stenting: paucity of late clinicalstent-related problems. Journal of the American College ofCardiology, 28, 820–826.

Lewin, K. (1935) A Dynamic Theory of Personality. McGraw-Hill, New York.

Lisspers, J., Sundin, O., Hofman-Bang, C., Nordlander, R.,Nygren, A., Ryden, L. and Ohman A. (1999) Behavioraleffects of a comprehensive, multifactorial program forlifestyle change after percutaneous transluminal coronaryangioplasty: a prospective, randomized, controlled study.Journal of Psychosomatic Research, 46, 143–154.

Marlatt, G. A. and Kilmer, J. R. (1998) Consumer choice:implications of behavioral economics for drug use andtreatment. Behavior Therapy, 29, 567–576.

McKenna, K. T., Maas, F. and McEniery, P. T. (1995) Coronary

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from

M. E. Charlson et al.

risk factor status after percutaneous transluminal coronaryangioplasty. Heart and Lung, 24, 207–212.

Montoya I. D., Atkinson, J. S. and Trevino, R. A. (2000)Economics as a factor in models of behavioral motivationand change. Journal of Substance Use and Misuse, 35,329–346.

Morocutti, G., Tuniz, D. and Fioretti, P. M. (1999)Comprehensive cardiac rehabilitation: an issue to bereaddressed. European Heart Journal, 20, 1448–1450.

Ornish, D., Brown, S. E., Scherwitz, L. W., Billings, J. H.,Armstrong, W. T., Ports, T. A., McLanahan, S. M., Kirkeeide,R. L., Brand, R. J. and Gould, K. L. (1990) Lifestyle changesand heart disease. Lancet, 336, 741–742.

Prochaska, J. O., DiClemente, C. C. and Norcross, J. C. (1992)In search of how people change: applications to the addictivebehaviors. American Psychologist, 47, 1102–1114.

Rhodes, K. V., Lauderdale, D. S., Stocking, C. B., Howes, D.S., Roizen, M. F. and Levinson, W. (2001) Better healthwhile you wait: a controlled trial of a computer-basedintervention for screening and health promotion in theemergency department. Annals of Emergency Medicine, 37,284–291.

Rodriguez, A., Boullon, F., Perez-Balino, N., Paviotti, C., SosaLiprandi, M. I., Pacalios, I. F., Blasksley, E., Newell, J. B.,Pujadas, G., Paolasso, E., Baudino, C. and Madoery, R. (1993)Argentine randomized trial of percutaneous transluminalangioplasty versus coronary-artery bypass-surgery inmultivessel disease (ERACI): in-hospital results and 1-yearfollow-up. Journal of the American College of Cardiology,22, 1060–1067.

Rogers, W. J., Alderman, E. L., Chaitman, B. R., DiSciascio,G., Horan, M., Lytle, B., Mock. M. B., Rosen, A. D., SuttonTyrrell, K., Weiner, B. H. and Whitlow, P. L. (1995) BypassAngioplasty Revascularization Investigation (BARI):baseline clinical and angiographic data. American Journalof Cardiology, 75, 9C–17C.

Roizen, M. F. (1998) Real Age: Are You as Young as You CanBe? Harper-Collins, New York.

Roizen, M. F. and Goetz, A. (1998) Gains in life expectancyfrom medical interventions [Comment]. New EnglandJournal of Medicine, 339, 1943.

Rosenstock, I. M. (1966) Why people use health services.Milbank Memorial Fund Quarterly, 44, 94–124.

Rupprecht, H. J., Brennecke, R., Bernhard, G., Erbel, R., Pop,T. and Meyer, J. (1990) Analysis of risk factors for restenosisafter PTCA. Catheterization and Cardiovascular Diagnosis,19, 151–159.

Sadusk, J. R., Jr and Robbins, L. C. (1968) Proposal for health-hazard appraisal in comprehensive health care. Journal ofthe American Medical Association, 203, 1108–1112.

Schoenbach, V. J, Wagner, E. H. and Beery, W. L. (1987)Health risk appraisal: review of evidence for effectiveness.Health Services Research, 22, 553–580.

618

Schuler, G., Hambrecht, R., Schlierf, G., Niebauer, J., Hauer,K., Neumann, J., Hoberg, E., Drinkmann, A., Bacher, F. andGrunze, M. (1992) Regular physical exercise and low-fatdiet. Effects on progression of coronary artery disease.Circulation, 86, 1–11.

Suranovic, S. M., Goldfarb, R. S. and Leonard, T. C. (1999)An economic theory of cigarette addiction. Journal of HealthEconomics, 18, 1–29.

The Bypass Angioplasty Revascularization Investigation(BARI) Investigators (1996) Comparison of coronary bypasssurgery with angioplasty in patients with multivessel disease.New England Journal of Medicine, 335, 217–225.

Tsevat, J., Weinstein, M. C., Williams, L. W., Tosteson, A. N.A. and Goldman, L. (1991) Expected gains in life expectancyfrom various coronary heart disease risk factor modifications.Circulation, 83, 1194–1201.

Wagner, E. H., Beery, W. L., Schoenbach, V. J. and Graham.(1982) An assessment of health hazard/health risk appraisal.American Journal of Public Health, 72, 347–352.

Watts, G. F., Leis, B., Brunt, J. N. and Lewis, E. S. (1992)Effects on coronary artery disease of lipid lowering diet ordiet plus cholestyramine in the St Thomas ArteriosclerosisRegression Study. Lancet, 39, 563–569.

Welch, H. G., Albertsen, P. C., Nease, R. F., Bubolz, T. A. andWasson, J. H. (1996) Estimating treatment benefits for theelderly: the effect of competing risks. Annals of InternalMedicine, 124, 577–584.

Wenger, N. K. (1991) Rehabilitation after cardiovascular diseasewith a special emphasis on developing countries:recommendations of a World Health Organization ExpertCommittee (October 1991). Heart Beat: Journal of theInternational Society and Federation of Cardiology, 1, 2–3.

Wright, J. and Strang, J. R. (1997) Reducing the risk aftercoronary artery bypass surgery: documentation of risk factorsand communication between hospital and general practice.Public Health, 111, 157–160.

Wright, J. C. and Weinstein, M. C. (1998) Gains in lifeexpectancy from medical interventions: standardizing dataon outcomes. New England Journal of Medicine, 339,380–386.

Writing Group for the BARI Investigators (1997) Five yearclinical and functional outcome comparing bypass surgeryand angioplasty in patients with multivessel coronary disease.Journal of the American Medical Association, 277, 715–721.

Zhao, X. Q., Brown, B. G., Stewart, D. K., Hillger, L. A.,Barnhart, H. X., Kosinski, A. S., Weintraub, W. S. and King,S. (1996) Effectiveness of revascularization in the Emoryangioplasty versus surgery trial. A random comparison ofcoronary angioplasty with bypass surgery. Circulation, 93,1954–1962.

Received on February 14 2001; accepted on May 22, 2001

by guest on June 4, 2013http://her.oxfordjournals.org/

Dow

nloaded from