Changing health behaviors to improve health outcomes after angioplasty: a randomized trial of net present value versus future value risk communication

Changing health behaviors to improve health outcomesafter angioplasty: a randomized trial of net present value

versus future value risk communication

M. E. Charlson1*, J. C. Peterson1, C. Boutin-Foster1, W. M. Briggs2,G. G. Ogedegbe3, C. E. McCulloch4, J. Hollenberg1, C. Wong5 and

J. P. Allegrante1,6

Abstract

Patients who have undergone angioplasty expe-rience difficulty modifying at-risk behaviors forsubsequent cardiac events. The purpose of thisstudy was to test whether an innovative ap-proach to framing of risk, based on ‘net presentvalue’ economic theory, would be more effectivein behavioral intervention than the standard ‘fu-ture value approach’ in reducing cardiovascularmorbidity and mortality following angioplasty.At baseline, all patients completed a health as-sessment, recieved an individualized risk profileand selected risk factors for modification. Theintervention randomized patients into two vary-ing methods for illustrating positive effects ofbehavior change. For the experimental group,each selected risk factor was assigned a numericbiologic age (the net present value) that approx-imated the relative potential to improve currenthealth status and quality of life when modifying

that risk factor. In the control group, risk reduc-tion was framed as the value of preventing futurehealth problems. Ninety-four percent of patientscompleted 2-year follow-up. There was no differ-ence between the rates of death, stroke, myocar-dial infarction, Class II–IV angina or severeischemia (on non-invasive testing) between thenet present value group and the future valuegroup. Our results show that a net present riskcommunication intervention did not result insignificant differences in health outcomes.

Introduction

Patients who have undergone percutaneous trans-

luminal coronary angioplasty (PTCA) must make

behavioral changes, such as stopping smoking, in-

creasing physical activity, decreasing cholesterol

and reducing weight, in order to reduce the risk of

subsequent cardiac events. Cessation of smoking,

for example, has been shown to result in a lower

recurrence rate, with less angina, fewer limitations

of physical activity and increased survival [1–5].

Other lifestyle changes, such as taking lipid-lowering

medicine, have been shown to reduce the recur-

rence of the disease. At least two studies have

shown that intensive lipid-lowering therapy can re-

sult in regression of disease [6, 7].

Despite the risk of recurrence, most patients

with coronary heart disease who have undergone

PTCA do not make major changes in behaviors

[8]. Because patients find it difficult to adopt and

maintain behavioral changes after angioplasty

[9, 10], we sought to investigate a novel approach

1Center for Complementary and IntegrativeMedicine,Weill

Cornell Medical College, New York, NY 10065, USA,2Department of Mathematics, Central Michigan University,

Mount Pleasant, 48859, 3Department of Medicine, College of

Physicians and Surgeons, Columbia University, New York,

NY 10032, USA, 4Department of Epidemiology and

Biostatistics, School of Medicine, University of California,

San Francisco, CA 94143, USA, 5Department of Medicine,

Weill Cornell Medical College New York, New York, NY

10021, USA and 6Department of Health and Behavior

Studies, Teachers College and Department of Sociomedical

Sciences, Mailman School of Public Health, Columbia

University, New York, NY 10027, USA

*Correspondence to: M. E. Charlson.

E-mail: [email protected]

� The Author 2007. Published by Oxford University Press. All rights reserved.For permissions, please email: [email protected]

doi:10.1093/her/cym068

HEALTH EDUCATION RESEARCH Vol.23 no.5 2008

Pages 826–839

Advance Access publication 19 November 2007

by guest on February 18, 2016http://her.oxfordjournals.org/

Dow

nloaded from

http://her.oxfordjournals.org/

to communicating risk in an effort to motivate

patients to make changes in health behavior. The

current approach to communicating risk informa-

tion largely helps patients to understand how mak-

ing behavioral changes will reduce ‘future risk’.

Our intervention approach was designed to ap-

proach the communication of risk by framing the

value of behavior modification in terms of the ‘net

present value’ (what patients can gain now) of the

change. Thus, our approach was designed to pro-

vide patients and their physicians with a personally

relevant means by which they could prioritize life-

style changes in terms of a tangible value associated

with the change in relative risk associated with

a change in behavior. This approach and its poten-

tial application to understanding risk communica-

tion in health behavior change have theoretical

origins in net present value economic theory [11].

In economic theory, the concept of net present

value involves taking the future value of a good or

investment and relating it to what it is worth in

present dollar terms by discounting based on

expectations of interest rate, inflation and demand.

For the study, this theoretical perspective was ap-

plied in communicating the benefit of changing

health behavior in terms of reduced risk. To do

so, the health effects of a potentially beneficial be-

havioral change were portrayed in present value or

present value health terms that have been calculated

within a short (3-year) time frame and for which the

change in mortality can be clearly and understand-

ably related to the change in risk at any given age.

Thus, as Allegrante and Roizen [11] have pointed

out: ‘. if smoking a pack {of cigarettes} a day

increases the risk for the average 50-year-old Cau-

casian male to that of the average 58-year-old

{counterpart}, net-present value of smoking today

is 8 years of aging (or, in effect, a loss of 8 years)’.

We have outlined in detail the theoretical frame-

work and intervention methods underlying our ap-

proach to communicating future health risks in the

PTCA patient population in a previously published

paper [12].

This paper reports a randomized trial whose ob-

jective was to evaluate the efficacy of a behavioral

intervention that communicated risk based on a net

present value approach versus a standard ‘future

value’ approach in reducing cardiovascular morbid-

ity and mortality following angioplasty or stenting.

In addition, we examined the impact of the inter-

vention on stages of change operationalized from

the Transtheoretical Model of Stages of Change

[13, 14] and behavior-specific self-efficacy [15, 16].

Methods

Patient eligibility criteria, screening andrecruitment

This study protocol (#0302005991) was approved

by the Institutional Review Board of Weill-

Cornell Medical College on 6 July 1998 and has

been reapproved every year since. The study was

conducted from 1999 to 2003. All patients under-

going coronary artery catheterization who had at

least single-vessel stenosis and revascularization

with PTCA or coronary artery stenting were eligible

for enrollment if they were (i) not enrolled in an-

other trial designed to modify behavior, (ii) verbally

fluent in English and (iii) able to provide informed

consent within 1 week after the procedure. We

identified and screened for potential eligibility a to-

tal of 2022 patients through a systematic, daily re-

view of the cardiac catheterization schedule at the

New York-Presbyterian Hospital—Weill-Cornell

Medical College during an 18-month period.

Patients were enrolled during hospitalization fol-

lowing PTCA or coronary artery stenting.

Various factors contributed to the ineligibility of

screened patients (Fig. 1). In total, 38% of patients

who were screened met eligibility requirements. In

total 53% of eligible patients were enrolled. A total

of 660 patients provided written informed consent

and were randomized.

Baseline demographic and clinical history

Patient demographic status was documented for

age, sex, marital status, ethnicity, occupation, edu-

cation and employment. The mean age of all

patients was 62 years, and 27% were female.

Twenty-one percent were African-American or La-

tino. Sixty-two percent were married. Forty-three

Changing health behaviors after angioplasty

827


Dow

nloaded from


Fig. 1. CONSORT model: flow of patients through the study.

M. E. Charlson et al.

828


Dow

nloaded from


percent were college graduates, and 16% had not

completed high school. Forty-three percent were

working full time and 36% were retired.

The duration and history of patients’ symptoms

and treatments were noted, as well as history of

myocardial infarction, unstable angina, congestive

heart failure, previous catheterization or coronary

artery bypass grafting. Prior to angioplasty, 58%

of patients had unstable angina and 25% had a his-

tory of myocardial infarction. The mean ejection

fraction 50%. At baseline, 17% had undergone

prior bypass surgery, while 36% had prior angio-

plasty/stenting. Overall, 57% had hypertension,

26% diabetes and 14% obesity.

The Charlson comorbidity scale [17] was used to

evaluate the burden of comorbid disease. Baseline

medications were also recorded, especially the use of

aspirin, nitrates, beta adrenergic blockers, calcium

channel blockers, digitalis, diuretics, angiotensin-

converting enzyme inhibitors, other anti-hypertensive

agents, insulin or oral hypoglycemic agents, sedative/

hypnotics, anti-depressants and steroids. Height and

weight were recorded, and body mass index was cal-

culated. Total cholesterol, high density lipoprotein

(HDL) and low density lipoprotein (LDL) levels were

also recorded. Pertinent data from the angioplasty in-

cluded the number of vessels treated, the number and

type of stents, the ejection fraction and complications

of the procedure, if any. Procedures were performed

on an average of two vessels.

Patients completed a health assessment that eval-

uated 13 cardiac risk factors, including physical

activity, smoking, diet and medications. The dietary

questionnaire was based on the semi-quantitative

food frequency questionnaire, a reproducible and

valid measure of intake of many nutrients, which

has been used longitudinally to assess the dietary

habits of older adults and patients with chronic and

cardiac disease [18–21]. The physical activity ques-

tionnaire comprised a modification of the Minne-

sota leisure time activity questionnaire [22]. The

risk assessment provided all patients with their

own individualized risk factor profile in which

the behavior was highlighted for change, if an

individual patient’s status was outside normal or

recommended ranges. The possible areas for car-

diovascular risk behavior change can be broadly

cast in five major domains:

Physical activity: increase physical activity, in-

crease aerobic exercise or increase strength training.

Smoking: stop smoking (or continue not to

smoke).

Diet/weight: reduce weight, reduce red meat,

reduce dietary intake of saturated fat and choles-

terol, increase fiber-rich food, increase flavonoid-

rich food, increase folic acid intake.

Blood pressure, diabetes: control/reduce blood

pressure, control diabetes.

Medications: take beta blockers/reduce heart rate.

All patients received feedback about their status

on the selected list of 13 cardiovascular risk factors.

Recommendations for behaviors to target for

change were based on pre-specified cutoffs [12].

Patients were then asked to choose two to three risk

factors to target for lifestyle change. Although more

risk factors could have been chosen, we decided

that only two to three risk factors could be feasibly

managed by patients. Both the patient and cardiol-

ogist received a copy of the risk profile and behavior

change choices. All patients were then asked to

consult with their cardiologist prior to implementing

any new health behavior change activity.

Stage of change is a measure of readiness for

initiating changes in each risk factor and included

five possible stages: pre-contemplation, contempla-

tion, preparation, action and maintenance. The

stage of change model has been utilized in assessing

behavior change with various health-related behav-

iors, including diet [23], weight [24], exercise [25]

and changing more than one behavior at a time [26].

Researchers have evaluated rates of the stages of

change and how they differently impact behavior

change likelihood [27].

Experimental and control intervention

The intervention theory and methods utilized in this

study have been described in extensive detail in an

earlier report [12]. Patients were randomized to one


829


Dow

nloaded from


of two groups, which determined the specific type

of feedback that they would receive: the experimen-

tal net present value (n = 329) or the control stan-

dard future value intervention (n = 331). The study

biostatistician prepared a randomization schedule

from random numbers prior to the enrollment of

the first patient. The assignments were placed in

sealed opaque numbered envelopes prior to the on-

set of the study. Randomization was stratified for

diabetes, gender and minority status in order to

achieve balancing of those factors. None of the de-

mographic or clinical characteristics differed statis-

tically between the two groups (Table I). Once

adjusting for multiple comparisons among each of

the characteristics, there was no significant differ-

ence for any of the patient characteristics at baseline.

Patient education about each risk factor in an in-

dividualized profile was framed using either a future

risk approach or a net present value approach,

where each factor is weighted according to its rel-

ative risk. To help patients in both groups get

started, we provided them with behavior change

‘tip sheets’ for each of the risks they selected to

change, as well as referrals for behavior change

programs and other community resources. The tip

sheets provided practical advice about how to

change behavior. For example, the physical activity

tip sheet suggested parking a car at the far end

of a parking lot in order to walk further to one’s

destination.

In both groups, stages of behavioral change [13]

and behavior-specific self-efficacy [15, 16] for each

of the behaviors of interest—increasing physical ac-

tivity, stopping smoking, increasing or decreasing

certain dietary elements, losing weight, controlling

blood pressure and taking medications—were also

assessed. The stage of change construct was

assessed by a single-item rating of five stages (1 =

pre-contemplation, 2 = contemplation, 3 = prepara-

tion, 4 = action and 5 = maintenance). We assessed

stage of change for all patients at baseline and again

at 12- and 24-month follow-ups for each of the cho-

sen risk factor behaviors. To measure the discrete

stage of change, we asked the individual whether

they were seriously considering changing a risk fac-

tor behavior of interest within the next 6 months.

Similarly, behavior-specific self-efficacy was

assessed with a single-item confidence rating on

a scale of 1–10 (1 = not at all confident, 10 = very

confident) for each chosen behavioral risk factors,

at baseline and at 12- and 24-month follow-ups. For

example, to measure self-efficacy with regard to

overall physical activity, we asked: ‘How confident

are you that you will be able to start doing more

regular physical activity (starting with 10 min a day

for at least 3 days a week; and gradually working up

Table I. Selected demographic and clinical characteristics of

patients in the net present value and future value groups at

baseline (by percent, unless otherwise indicated)

Characteristic Net present

value

Future

value

Demographics

Age (mean age in years 6s.d.) 62.7 6 11.6 62.8 6 1.5

Female 27 27

African-American 12 12

Latino 8 10

Married 61 64

High school graduate 43 39

College graduate 39 46

Employed full time 43 44

Clinical

Prior myocardial infarction 24 43

Prior angioplasty 35 38

Congestive heart failure 6 5

Hypertension 56 58

Depression (mean CES-D

score 6s.d.)

11.2 6 11 11.6 6 11

Average Body Mass Index in

kg/m2

28.0 6 5.3 28.0 6 5.3

Ejection fraction 51 51

Charlson comorbidity score

0 34 34

1–2 39 41

>3 27 25

Peripheral Vascular Disease 18 20

Diabetes 26 27

Cancer 12 14

Stroke 8 5

Medications

Calcium channel blockers 9 9

Beta blockers 54 46

Nitrates 26 22

Diuretics 11 12

Aspirin 76 76

Number of vessels treated 2.3 6 2.0 2.2 6 1.9

Left main disease 6 4


830


Dow

nloaded from


to longer periods for 4–5 days a week) within the

next month?’

Net present value approach (experimentalgroup)

The experimental group received feedback about

their cardiovascular risk factors framed in a com-

mon metric—‘biologic age’. Biologic age was cal-

culated with a research version of the RealAge�software, which was modified for use in this study.

RealAge combines a person’s historical data and

habits to provide a posterior state of knowledge

risk. In the modified version, at the outset an indi-

vidual’s biologic age is the same as his or her chro-

nologic or calendar age. Patients were told their

estimated biologic age (calculated using data from

the baseline health risk assessment). The data for

each risk factor were communicated in terms of the

amount of ‘biologic age reduction’ one could

achieve within 3 months and within 2 years, if each

risk factor was changed. From this individual pro-

file, patients were asked to choose two to three of

these risk behaviors to begin changing ‘now’ be-

cause doing so ‘will decrease your biologic age’.

For each risk factor, they were given a set of action

steps, i.e. individualized guidelines for each risk

factor that emphasized the present value of making

a behavioral change.

Future value or standard approach(control group)

The control group received feedback about their

cardiovascular risk factors in terms of deviation

from normative values. From their individual pro-

files, patients were asked to choose two to three of

these risk factors to begin changing now because

doing so ‘will increase your lifespan’. For each risk

factor, they were given a set of action steps, i.e.

individualized guidelines for each risk factor that

emphasized the future value of behavioral change.

Follow-up

For 2 years, both groups of patients were contacted

and interviewed by telephone at 3-month intervals.

The telephone contact provided motivational sup-

port to patients in both groups, using a specific

script that was based loosely on the principles of

motivational interviewing developed by Emmons

and Rollnick [28], to assist patients in adopting

and maintaining behavioral changes. In addition,

each call included an updated assessment of interval

clinical events and changes in cardiovascular risk

profile, as well as an opportunity for the individual

patient to change or add new risk factors.

The follow-up was designed to capture all rele-

vant interval events, including behavioral change

and new outcomes or cardiac procedures or related

hospitalizations. In total, 88% of patients had their

clinical status assessed at 3–6 months, 76% at 9–12

months, 63% at 15–18 months and 93% at 24

months. In total, 86% of patients had detailed risk

factor profiles updated at 3–6 months, 71% at 9–12

months, 53% at 15–18 months and 47% at 21–24

months. There was no difference in follow-up rates

between the net present value and future value

groups. Seventy-six percent of the patients com-

pleted four or more follow-up assessments.

Primary outcomes

Patients completed their final evaluation for major

post-operative clinical outcomes at ;24 months

(68 weeks). The primary outcome was freedom

from death, myocardial infarction, stroke, Class

II–IV angina or severe asymptomatic ischemia at

24 months. The time course of freedom from death,

myocardial infarction, stroke, Class II–IV angina or

severe asymptomatic ischemia was also docu-

mented, with the primary end point being the time

to first event. Revascularization was not included as

an end point because indications vary widely.

End point review was conducted by a cardiolo-

gist who was blinded to randomization group and

included:

Mortality: all deaths from any cause.

Myocardial infarction: new persistent Q waves of

>0.03 msec and >1 mV in depth were required

in two contiguous leads on a standard 12-lead

electrocardiogram in the absence of a new con-

duction abnormality or a marked change in the

QRS axis.


831


Dow

nloaded from


Angina: the absence of angina was defined as

freedom from Class II–V anginal symptoms.

Severe ischemia on non-invasive testing: non-

invasive testing was not routinely performed on

patients enrolled in this trial. Whether or not

patients had non-invasive testing was decided

entirely by the patient’s own physician. This

end point was designed to capture those pa-

tients who were without anginal symptoms,

but who were found on non-invasive testing to

have sufficiently severe ischemia to warrant

revascularization.

Stroke: this was defined as the occurrence of

a new major focal neurologic deficit, which per-

sisted >24 hours.

Power and data analysis

We have estimated the occurrence of the principal

outcome defined above (mortality, myocardial in-

farction, stroke, Class II–IV angina or severe is-

chemia) is 30%. (Total event rates after

angioplasty range from 30—to 45%, and the event

rates after stents are somewhat lower; 30% is taken

as the event rate in the control arm.) Hence, the

expected outcome rate in the control arm, Pc, is

taken to be 0.30. A drop in the incidence of any of

the adverse outcomes to a level of 0.20 would con-

stitute a clinically important difference between the

two management strategies. Therefore, d (where d =

Pc � Pe) is taken to be 0.10. Based on the above

values of Pc, d and a two-sided alpha of 0.05, with

a b error of 0.2 for 80% power, the sample size is

estimated to be a total of 590 patients or 295 in each

group.

Sample size was based on 80% power for the

occurrence of any one of the following out-

comes: death, myocardial infarction, stroke, Class

II–IV angina or severe ischemia on non-invasive

testing. The level for testing the significance of

the principal outcome was set at P = 0.05. All

data were entered into an integrated series of data-

bases and then transported to SAS for windows

v.8.02 (Cary, SC) for statistical analyses. After 24

months of follow-up, we analyzed the data for main

effects between groups, efficacy and safety accord-

ing to the intention-to-treat principle.

Results

Risk factors at baseline

Overall, there were no statistically significant dif-

ferences between randomization groups in the spe-

cific risk factors chosen for change, as described

below. Patients had an average of six risk factors

recommended for behavioral change at baseline.

Patients in both groups were asked to identify risk

factors they wished to select for change, and, on

average, they chose three risk factors for change.

Table II shows the percentage of patients at baseline

for whom a risk factor was recommended for

change and the percentage of patients who chose

a risk factor for change, if recommended. Increas-

ing physical activity was recommended for change

in 85% of patients and reducing cholesterol was

recommended in 86%; quitting smoking was rec-

ommended in the 25% of patients who were cur-

rent smokers. The most common risk factors

chosen at baseline for change were overall physi-

cal activity (66%), smoking (62%) and weight loss

(59%).

Table II shows the proportion of patients whose

stage of change was preparation as well as ratings of

perceived self-efficacy. Most patients who chose to

change a behavior were at the preparation stage for

change. Patients had high self-efficacy or confi-

dence in their ability to modify health behaviors

immediately after angioplasty, with average self-

efficacy scores of between 7 and 9. For either stage

of change or self-efficacy, there were no statistically

significant differences between randomization

groups.

At baseline, patient biologic ages were calcu-

lated, with biologic age being an identical mean

of 68 years in both groups. On average, the biologic

age of patients participating in the study was ;6

years older than their actual chronologic age. Only

patients in the net present value group were given

feedback in terms of biologic age.


832


Dow

nloaded from


Table II. Risk factors for patients that were recommended for change and those that were chosen for change at baseline for the net

present value and future value groups (n = 660)

Risk factors Recommended

for change

(% yes)a

Chosen for change

if recommended

(% yes)b

Stage of

change was

preparation (%)c

Self-efficacy

(Mean and SD)d

Physical activity

Overall physical activity

Present value (experimental) 87 66 73 8.2 6 2.1

Future value (control) 84 66 72 8.2 6 2.0

Aerobic exercise



Strength-building exercise



Smoking

Quit or reduce smoking



Diet/weight

Lose weight



Reduce red meat consumption



Reduce cholesterol through diet or medical treatment



Eat more flavonoid-rich foods



Eat more fiber

Present value (experimental) 52 48 86

Future value (control) 50 49 68

Increase folic acid



Other

Control heart rate



Control blood pressure



Control diabetes



SD, standard deviation.aPercent of total sample that received this recommendation in their risk factor profilebPercent of sample that chose to work on this risk factor when they received this recommendation in their risk factor profilecIf the risk factor was chosen for change, the possible range of stages included pre-contemplation, contemplation, preparation, actionand maintenancedIf the risk factor was chosen for change, the range of the score for self-efficacy was 1–10, 1 = not at all confident, 10 = very confident.


833


Dow

nloaded from


Overall patterns of behavior change

At baseline, patients chose an average of 2.8 risk

factors for change. This increased to a total of 5.1

over the course of the 2-year follow-up interval.

Table III shows the percentage of risk factors cho-

sen for change at any follow-up compared with

those chosen at baseline for both groups. Also

shown is the percent of patients who achieved ac-

tion or maintenance on any behavior at any time

point compared with action or maintenance at the

last assessment for the behavior. Over the period of

follow-up, patients in the net present value group

reached action on 1.5 risk factors, identical to the

rates of action in the future value group. On aver-

age, patients in both groups were able to maintain

action at the last assessment on 1.1 risk factors,

most often physical activity, weight reduction or

lowering cholesterol. Overall, 37% of patients

made no health behavior changes, 34% changed

one health behavior and 29% changed two or more

health behaviors.

Main trial outcomes

In total, we obtained major trial outcomes at the 2-

year follow-up on 622 of the 660 patients (94%)

enrolled in the study (Fig. 1). Of the 622 patients

who completed a 2-year follow-up, 6 patients had

died before the end of the study. Thirty-eight

patients were lost to follow-up at 24 months, though

six of these patients had outcomes prior to loss to

follow-up.

At the 2-year follow-up, there was no difference

in the principal outcome of the trial or in the com-

bined rates of death, stroke, myocardial infarction,

Class II–IV angina and severe ischemia on non-in-

vasive testing by randomization group. The differ-

ence in outcome rate between the two groups was

not statistically significant, assessed with a test for

difference in proportions P = 0.23; with a 95%

confidence interval for the difference of the net

present value percentage minus the future value

percentage of �13 to 2.9%. Using logistic regres-

sion to control for baseline and demographic vari-

ables such as sex, age, comorbidity score and

clinical status (ejection fraction, etc.) did not

change the results: there was no difference between

the net present value and future value groups. This

was not unexpected as the two groups were so

closely matched (Table I). Table IV shows that

the rates of each outcome were similar in both

groups, and the overall rate of combined outcomes

was 39.1% in the net present value group and

34.2% in the future value group.

When we analyzed all patients regardless of ran-

domization group, the greater the number of health

behaviors successfully changed, the less likely it

was that an adverse event would occur for an in-

dividual patient.

Discussion

Cardiac rehabilitation following coronary revascu-

larization seeks to assist patients who have under-

gone PTCA with necessary lifestyle changes.

Recent American and European studies [29–34]

of cardiac rehabilitation that includes comprehen-

sive lifestyle modification have shown promise in

reducing and even reversing coronary artery dis-

ease. Early studies by Ornish et al. [29] and Haskell

et al. [30]demonstrated that a program of intensive

multifactor risk reduction reduced luminal narrow-

ing in coronary arteries of men and women with

coronary artery disease. In addition, those who

made changes in lifestyle and took lipid-lowering

medication over a 4-year period were found to have

decreased rates of hospitalization for cardiac-

related clinical events [30]. In a later study, Ornish

et al. [31] showed that an intensive lifestyle modi-

fication program of dietary fat restriction, stress

management, exercise, smoking cessation and

group psychosocial support over a 5-year period

could reduce coronary atherosclerosis without the

use of lipid-lowering drugs. Koertge et al. [32]

demonstrated that a multicomponent lifestyle

change program for diet, exercise, stress manage-

ment and social support can produce significant

improvements in diet, exercise and stress manage-

ment practices, which can be maintained over a

12-month period. In another study, Pasquali et al.[33] evaluated the impact of cardiac rehabilitation


834


Dow

nloaded from


Table III Behaviors chosen for change at baseline or at any time, and whether or not the patients achieved action or maintenance in

both the net present value and future value groups (n = 660)

Risk factors Chosen for

change at

baseline (%)

Behavior

chosen for

change at

any time (%)

Achieved action

or maintenance

of behavior at

any point (%)

Action or

maintenance of

behavior at last

assessment (%)

Physical activity

Overall physical activity

Present value (experimental) 56.8 83.6 38.6 21.3

Future value (control) 55.2 83.7 31.4 20.9

Aerobic exercise



Strength-building exercise



Smoking

Quit or reduce smoking



Diet/weight

Weight loss



Reduce red meat

Present value (experimental) 25.9 38 5.2 4.3


Reduce cholesterol through diet or medical treatment



Eat more flavonoid-rich foods

Present value (experimental) 23.9 42 5.5 3.9

Future value (control) 24.3 47.4 6 3.6

Eat more fiber



Increase folic acid



Other

Control blood pressure



Control heart rate



Control diabetes




835


Dow

nloaded from


on functional outcomes among patients who had

been revascularized and found significant improve-

ments in physical functioning and secondary pre-

ventive behaviors at 6 months, particularly among

men; similar findings have been reported in a 2003

study by the Norwegian Vestfold Heartcare Study

Group [34]. While such studies show the promise of

cardiac rehabilitation, most patients with coronary

heart disease do not easily make necessary lifestyle

changes following angioplasty, and maintaining

such changes in the long term still remains chal-

lenging. Recent surveys conducted by EUROAS-

PIRE I and II Group [35] show that lifestyle change

among people with coronary artery disease is a con-

cern. Moreover, McDonald et al. [36] have reported

that even when effective in the long term, behav-

ioral interventions designed to improve adherence

with medication often involve complex and inten-

sive combinations of information and various other

behavioral techniques.

The objective of this randomized trial was to

evaluate whether feedback of individualized risk

profiles framed as the opportunity to reduce one’s

biologic age was more effective after 2 years in

reducing mortality and major cardiovascular mor-

bidity among patients who had just undergone

PTCA or coronary artery stenting, as compared

with the standard risk reduction approach. The

novel theoretical concept underlying this behavior

change intervention was framing risk reduction as

the opportunity to reduce one’s present biologic age

(net present value approach) contrasted with the

standard motivation to reduce risk, which was

framed as one’s opportunity to reduce future risk

(future value approach).

In addition to the concept of communicating risk

using net present value (biologic age), our interven-

tion incorporated intervention concepts from well-

established models of behavior change, including

those of individualized feedback [37, 38], stages

of change [13, 14] and self-efficacy [15, 16]. We

also scheduled frequent telephone follow-up by

trained research interventionists who provided

a minimal level of motivational support to all

patients throughout the study. Our overall goal in

the experimental arm of the trial was to convey to

patients the immediate health impact of specific

behaviors on their biologic age as well as the im-

mediate impact of changing them. We hypothesized

that armed with the value of how much biologic age

reduction one could achieve by, for example, stop-

ping smoking (a difficult behavior, but one that

yields greater immediate reduction in biologic age

than, for example, taking folate), the pattern in the

choice of behaviors to change would be different in

the two groups. Although patients most often chose

physical activity, weight reduction or lowering cho-

lesterol as the behaviors of interest for change, there

was no difference in pattern of choice in the two

groups. Moreover, there was no difference in the

main trial outcomes at the end of the study. For

example, the confidence intervals calculated for

the difference in net present and future value

showed that this difference would probably not be

different by >13% and that net present value could

not be better than future value by >2.9%, which is

only a negligible improvement.

Why were no differences in pattern of choice for

health behavior change or health outcome observed

at the end of the trial? There are several plausible

reasons—both theoretical and methodological—

why the intervention failed to produce the hypoth-

esized changes. First, patients in both arms of

the trial received multiple telephone contacts

throughout a 2-year follow-up period during which

they were interviewed about their health and given

motivational support. As a consequence, the at-

tention to the control group may have diluted any

Table IV. Major trial outcomes by intervention groups

Outcomes Net

present

value (%)

Future

value

(%)

P values

Mortality 4.1 4.4 0.83 (chi square)

Myocardial

infarction

4.2 4.4 0.8774 (chi square)

Class II–IV

angina

33.0 28.2 0.1787 (chi square)

Severe ischemia 1.0 1.6 0.7249 (Fisher’s exact test)

Stroke 1.3 0.6 0.4475 (Fisher’s exact test)

Total 39.1 34.2 0.23 (chi square)


836


Dow

nloaded from


potential differential impact of the presentation of

biologic age in the experimental net present value

group.

Second, our use of an understandable and per-

sonally relevant metric—biologic age—was

intended to be consistent with a rational model of

risk communication. We believed that the value of

communicating risk in terms of biologic age was in

revealing to the patient the increment of reduction

in biologic age that could be gained by changing

each individual risk factor. However, we did not

assess our study participants’ belief in the concept

prior to the study, so it is possible that patients in

the intervention group simply did not believe in the

concept of biologic age reduction or that they could

successfully reduce their biologic age by changing

their health behaviors (the outcome expectancy).

Moreover, many people do not understand concepts

of probabilistic thinking, survival or the distinctions

between harmful events occurring within popula-

tions and those within individuals or absolute

versus relative risk [39, 40]. For example, when at-

tempting to communicate the potential reduction in

risk for coronary artery disease that can be attrib-

uted to taking lipid-lowering drugs, the difference

between reporting relative or absolute risk is

likely to influence the perception of risk by the

patient [41].

Third, consistent with the Transtheoretical Model

of Stages of Change, telling people that they can

reduce their biologic age by changing their behavior

constitutes an intervention largely aimed at raising

risk awareness and increasing positive outcome

expectancies. However, in the context of most theo-

ries of health behavior change, risk awareness and

outcome expectancies are typically specified as pre-

dictors of the non-action phases or behavioral inten-

tions. Thus, because of the ‘intention-behavior-gap’

that has been observed in previous research [42],

targeting such cognitive factors alone is not suffi-

cient for achieving behavior change.

Fourth, had we assessed potential mediating var-

iables such as risk awareness and outcome expec-

tancies, both of which theoretically might have

been influenced by our intervention, we may have

been able to gain additional insight into why our

intervention did not work and why [43]. For exam-

ple, it is possible that our intervention was not ef-

fective in influencing critical mediating variables or

that our intervention ‘did’ influence important medi-

ators that we did not assess, but, although doing so,

did not make any difference in the main outcomes.

Finally, our study was not only aimed at reducing

those behavioral risk factors that were potentially

modifiable but also showing an improvement in

clinical outcomes. Even though patients may rea-

sonably be expected to initiate behavior changes,

maintaining such changes over a 2-year period of

time is difficult without an intensive behavioral in-

tervention that was beyond our interventional at-

tempt to motivate patients through a novel risk

communication message. It is possible that we

might have observed the hypothesized difference

between the two groups had we designed the in-

tervention for a shorter time period. Moreover, the

dose-response rate of health behavior change and

clinical outcomes needs to be considered. Although

people might be able to increase their physical activ-

ity from 0 to 1 time a week, this might not be suffi-

cient to influence the clinical outcomes we studied.

While these results are disappointing, identifying

an effective means of motivating and supporting

behavioral changes in patients with coronary artery

disease who have undergone PTCA remains a chal-

lenge. Although the communication of risk to such

patients is of great interest, communicating risk

alone may not be sufficient to stimulate the neces-

sary behavior change in the face of treatment that

produces such dramatic results in the relief of

symptoms. Because patients undergoing PTCA

and stenting do get dramatic symptomatic relief

it is not surprising that patients—and the doctors

who perform such procedures—view the treatment

as a veritable magic bullet. In light of this, perhaps

the interventional cardiologist needs to reinforce

both the importance and urgency of behavioral

change rather than telling the patient that the prob-

lem ‘has been fixed’. To do otherwise may leave the

patient with a false sense of cure for the underlying

causes of a disease that invariably requires persis-

tence at changing behavior if the benefits of PTCA

are to be maintained over time.


837


Dow

nloaded from


Funding

National Heart, Lung, and Blood Institute (R01

HL62161091).

Acknowledgements

A research version of the RealAge software was

made available to the investigators, without charge,

by RealAge. We thank Paula McKinley and Can-

dace Young for their contributions in the early

phases of the study and the anonymous reviewers

for their insightful comments and useful sugges-

tions for revision.

Conflict of interest statement

None declared.

References

1. Cavender JB, Rogers WJ, Fisher LD et al. Effects of smok-

ing on survival and morbidity in patients randomized to

medical or surgical therapy in the Coronary Artery Surgery

Study (CASS): 10-year follow-up. CASS Investigators.

J Am Coll Cardiol 1992; 20: 287–94.2. Galan KM, Deligonul U, Kern MJ et al. Increased frequency

of restenosis in patients continuing to smoke cigarettes after

percutaneous transluminal coronary angioplasty. Am J Car-

diol 1988; 61: 260–3.3. Myler RK, Shaw RE, Stertzer SH et al. Recurrence after

coronary angioplasty. Cathet Cardiovasc Diagn 1987; 13:

77–86.4. Ramanathan KB, Vander Zwaag R, Maddock V et al. In-

teractive effects of age and other risk factors on long-term

survival after coronary artery surgery. J Am Coll Cardiol

1990; 15: 1493–9.5. Reis GJ, Kuntz RE, Silverman DI et al. Effects of serum

lipid levels on restenosis after coronary angioplasty. Am J

Cardiol 1991; 68: 1431–5.6. Watts GF, Lewis B, Brunt JN et al. Effects on coronary

artery disease of lipid-lowering diet, or diet plus cholestyr-

amine, in the St Thomas’ Atherosclerosis Regression Study

(STARS). Lancet 1992; 339: 563–9.7. Brown G, Albers JJ, Fisher LD et al. Regression of coronary

artery disease as a result of intensive lipid-lowering therapy

in mean with high levels of apolipoprotein B. N Engl J Med

1990; 323: 1289–98.

8. McKenna KT, Mass F, McEniery PT. Coronary risk factorstatus after percutaneous transluminal coronary angioplasty.Heart Lung 1995; 24: 207–12.

9. Gaw BL. Motivation to change life-style following PTCA.Dimens Crit Care Nurs 1992; 11: 68–74.

10. Gulanick M, Bliley A, Perino B et al. Recovery patterns andlifestyle changes after coronary angioplasty: the patient’sperspective. Heart Lung 1998; 27: 253–62.

11. Allegrante JP, Roizen MF. Can net-present value economictheory be used to explain and change health-related behav-iors? Health Educ Res 1988; 13: i–iv.

12. Charlson ME, Allegrante JP, McKinley PS et al. Improvinghealth behaviors and outcomes after angioplasty: using eco-nomic theory to inform intervention. Health Educ Res 2002;17: 606–18.

13. Prochaska JO, DiClemente CC, Norcross JC. In search ofhow people change. Applications to addictive behaviors. AmPsychol 1992; 47: 1102–14.

14. Prochaska J, Redding C, Evers K. The transtheorecticalmodel and states of change. Health Behavior and HealthEducation. New York, NY: Jossey Bass, 1997

15. Bandura A. Social Learning Theory. Englewood Cliffs, NJ:Prentice-Hall, 1977

16. Strecher V, DeVellis B, Becker M et al. The role of self-efficacy in achieving health behavior change. Health Educ Q1986; 13: 73–91.

17. Charlson ME, Pompei P, Ales KL et al. A new methodof classifying prognostic comorbidity in longitudinal stud-ies: development and validation. J Chronic Dis 1987; 40:373–83.

18. Rimm EB, Giovannucci EL, Stampfer MJ et al. Reproduc-ibility and validity of an expended self-administered semi-quantitative food frequency questionnaire among malehealth professionals. Am J Epidemiol 1992; 135: 1114–26discussion 1127–36.

19. Liu S, Stampfer M, Hu F et al. Whole-grain consumptionand risk of coronary heart disease: results from the Nurses’Health Study. Am J Clin Nutr 1999; 70: 412–9.

20. Liu S, Lee I, Ajani U et al. Intake of vegetables rich incarotenoids and risk of coronary heart disease in men: thePhysicians’ Health Study. Int Epidemiol Assoc 2001; 30:130–5.

21. Sesso H, Liu S, Gaziano J et al. Dietary lycopene, tomato-based food products and cardiovascular disease in women.J Nutr 2003; 133: 2336–41.

22. Taylor HL, Jacobs DR, Schucker B et al. A questionnaire forthe assessment of leisure time physical activities. J ChronicDis 1997; 31: 741–55.

23. Ni Mhurchu C, Margetts B, Speller V. Applying the stages-of-change model to dietary change. Nutr Rev 1997; 55: 10–6.

24. Jeffery R, French S, Rothman A. Stage of change as a pre-dictor of success in weight control in adult women. HealthPsychol 1999; 18: 543–6.

25. Costakis C, Dunnagan T, Haynes G. The relationship be-tween the stages of exercise adoption and other healthbehaviors. Am J Health Promot 1999; 14: 22–30.

26. Boyle R, O’Connor P, Pronk N et al. Stages of change forphysical activity, diet, and smoking among HMO memberswith chronic conditions. Am J Health Promot 1998; 12:170–5.


838


Dow

nloaded from


27. Nigg C, Burbank P, Padula C et al. Stages of change acrossten health risk behaviors for older adults. Gerontologist1999; 39: 473–82.

28. Emmons KM, Rollnick S. Motivational interviewing inhealth care setting. Opportunities and limitations. Am J PrevMed 2001; 20: 68–74.

29. Ornish D, Brown SE, Scherwitz LW et al. Can lifestylechanges reverse coronary heart disease? The Lifestyle HeartTrial. Lancet 1990; 336: 129–33.

30. Haskell WL, Alderman EL, Fair JM et al. Effects of in-tensive multiple risk factor reduction on coronary athero-sclerosis and clinical cardiac events in men and womenwith coronary artery disease. The Stanford CoronaryRisk Intervention Project (SCRIP). Circulation 1994; 89:975–90.

31. Ornish D, Scherwitz L, Billings J et al. Intensive lifestylechanges for reversal of coronary heart disease. J Am MedAssoc 1998; 280: 2001–7.

32. Koertge J, Weidner G, Elliott-Eller M et al. Improvement inmedical risk factors and quality of life in woman and menwith coronary artery disease in the Multicenter LifestyleDemonstration Project. Am J Cardiol 2003; 91: 1316–22.

33. Pasquali S, Alexander K, Coombs L et al. Effect of cardiacrehabilitation on functional outcomes after coronary revas-cularization. Am Heart J 2003; 145: 445–51.

34. Group VHS. Influence on lifestyle measures and five-yearcoronary risk by a comprehensive lifestyle intervention pro-gramme in patients with coronary heart disease. Eur J Car-diovasc Prev Rehabil 2003; 10: 429–37.

35. Group, E.I.a.I. European Action on Secondary Prevention

by Intervention to Reduce Events. Clinical reality of coro-

nary prevention guidelines: a comparison of EUROSPIREI

and II in nine countries. Lancet 2001; 357: 995–1001.36. McDonald HP, Garg AX, Haynes RB. Interventions to en-

hance patient adherence to medication prescriptions: scien-

tific review. J Am Med Assoc 2002; 288: 2868–79.37. Champion V, Foster J, Menon U. Tailoring interventions for

health behavior change in breast cancer screening. CancerPract 1997; 5: 283–8.

38. Prochaska J, DiClemente C, Velicer W et al. Standardized,

individualized, interactive, and personalized self-help pro-

grams for smoking cessation. Health Psychol 1993; 12:

399–405.39. Alasewski A, Horlick-Jones T. How can doctors communi-

cate information about risk more effectively? Br Med J2003; 327: 729–31.

40. Alasewski A. A person-centered approach to communicat-

ing risk. PLoS Med 2005; 2: e41.41. Skolbekken J. Communicating the risk reduction achieved

by cholesterol reducing drugs. Br Med J 1998; 316: 1956–8.42. Sheeran P. Intention-behavior Relations. A Conceptual and

Empirical Review. Chichester, England, 2002, 1–36.43. Michie S, Abraham C. Interventions to change health behav-

iors: evidence based or evidence-inspired? Psychol Health2004; 19: 29–49.

Received on April 21, 2006; accepted on September 4, 2007


839


Dow

nloaded from


Changing health behaviors to improve health outcomes after angioplasty: a randomized trial of net present value versus future value risk communication

Documents