CONTENTS

Introduction

Characteristics of a Controlled Community Study

Theoretical Background
  Mass Communication
  Community Organization
  Environmental Change

Cessation Studies
  Stanford Three-Community Study
  Australian North Coast Program
  Swiss National Research Program
  The North Karelia Project
  Other Large-Scale Studies
  Related Studies of Cessation

Prevention Studies
  Stanford Study
  The North Karelia Youth Project
  Other Prevention Studies

Methodological Issues

Directions for Future Studies

Summary and Conclusions

References

501


Introduction
Community studies of smoking cessation and prevention are
defined as research in which geographically defined populations or
age cohorts are selected for experimental intervention or as control
or comparison groups. In this chapter, four major studies of cessation
are described, and related research findings are briefly considered.
Three major studies of prevention are also reviewed, with less
extensive presentations of other recent research. The theoretical
background for these studies is outlined, methodological issues are
discussed, and directions for future research are suggested. It is
concluded that community studies represent a significant emerging
paradigm for public health research.
The first public health "revolution," the recently achieved control
of major infectious disease, evolved in two phases: (1) recognition
that certain aspects of the environment are associated with disease,
and (2) establishment of sanitation facilities and services, such as
refuse collection and plumbing systems, for entire communities of
people. Basic public health engineering that was revolutionary a
century ago is now taken for granted by society. Today we emphasize
the second public health "revolution," the current efforts to control
chronic disease through behavior change. This development has also
evolved in two stages: (1) recognition of behaviors, such as cigarette
smoking, as being associated with disease, and (2) development of
new services to change ihose behaviors (Wynder and Hoffman 1979).
As might be expected, this most recent history of public health is no
less turbulent than that which preceded it.
Various factors have hindered progress. After the basic causes of
smoking-related diseases were recognized, there has been a tendency
to invest scarce resources in increasingly intricate studies of disease
causation processes. In addition, once funding is secured for inter-
vention to reduce smoking, efforts often focus on individual-level,
clinically oriented interventions (Lichtenstein 1982). If these inter-
ventions do not succeed, society is inclined to "blame the victim,"
much as the poor were held responsible for the microbes in their
water by 19th century social conservatives. The socioeconomic
gradient in rates of smoking and smoking-related disease may also
slow society's response at a community level to this modern,
noncommunicable disease epidemic. The greatest factor inhibiting
progress, however, is the cost of prevention. Such costs are com-
pounded by political obstacles stemming from the tremendous
influence of the tobacco industry, which employs thousands of people
and regularly delivers a substantial portion of tax revenues (Breslow
1982; Fritschler 1975; Sapolsky 1980). On a national scale, dramatic
changes in tobacco consumption have occurred in response to
successive measures (Warner 1977; Warner and Murt 1982). but
these have been limited largely to higher income groups. Few


examples of bold community or regional efforts comparable to those
involved in the control of infectious disease have been witnessed, and
the prevalence of smoking has declined much less dramatically
among women and has even increased in some minorities.

Characteristics of a Controlled Community Study

For this discussion, a controlled community study is defined
according to the scope of intervention and quality of research design,
with the essential feature being identification of natural, location-
based aggregations of individuals as well as formal and informal
social systems. In a community study, the entire population of a
geographic area is considered, so that a church or worksite is not a
community itself, but one of many systems constituting the total
network of interactions. The term "community" originally referred
to small systems numbering no more than several thousand persons.
This discussion, however, includes research on towns, counties, or
other relatively independent zones with up to several hundred
thousand inhabitants. Because educational systems represent all age
cohorts of youth, school-based studies are also reported. Studies of
entire States or nations are only briefly considered.

Because the population size to be addressed is a limiting factor in
any social program, the large numbers of people involved in a
community study dictate selection of intervention methods. Clinical
or other people-oriented approaches that typify behavioral research
on smoking cessation and prevention (Bernstein 1969; Bernstein and
McAlister 1976; Pechacek and McAlister 1980; Lando and McGovern
1982; Lichtenstein 1982) are not feasible for programs directed
toward many thousands of people. Community studies instead
emphasize large-scale delivery systems such as the mass communica-
tion media. Because community participation is now considered
essential for success, such studies also include community organiza-
tion programs seeking to stimulate interpersonal communication in
ways that are feasible on a large-scale basis. Community studies also
may involve environmental change, such as programs to modify the
purchase price or availability of consumer products or to sanction
public behaviors.

Because the emphasis herein is on controlled community research,
attention is limited to studies in which valid inferences can be made
concerning the effects of intervention on smoking rates in an entire
population. The essential elements are use of adequate measures of
smoking behavior applied over time in order to estimate long-term
trends, and equally important, the inclusion of control or reference
areas for the purpose of comparison. There are, of course, many
questions and controversies regarding the usefulness and validity of
large-scale experimental or quasi-experimental research (Campbell

504


and Cook 1979). Social policies such as those needed to sharply
reduce smoking are not likely to be introduced without experimental
trials, however, and the studies reported herein probably represent
the best currently attainable compromise between external and
internal validity. Given the very small number of studies meeting
even the minimal methodological criteria, it would be unwise to
restrict this review t,o the standards required by laboratory or
clinica! st.udies.

Theoretical Background

Effective mass communication, community organization, and
environmental change require a theoretical basis for planning. Most
community studies of health promotion and disease prevention are
based on fundamental theories and concepts from the behavioral
sciences. The most important of these are briefly outlined below.

Mass Communication

Theories on mass media effect have changed during the recent
history of communication research, and several clear stages have
been identified (Klapper 1960; Griffiths and Knutson 1960; Atkin
1979; Flay et al. 1980; Wallack 1981). Media were initially considered
nearly omnipotent in directly altering behavior, but it was later
discovered that they are incapable of producing effects independent
of other, more powerful social forces. The most recent view is that
mass media may have effects, but that they are smail and largely
dependent on facilitation from interpersonal influences and favor-
able environmental circumstances. Notwithstanding these limita-
tions, shifts of a few percentage points in consumer preferences may
be very significant in product marketing, while similar reductions in
chronic disease-promoting behaviors may have enormous absolute
significance in a population of several millions. One mass media
effect that is agreed upon by most communication scientists is
termed the "agenda-setting function" (McCombs and Shaw 19721, in
which the media powerfully influence topics generated in formal and
informal social gatherings. Media communication can also inform
and teach simple skills (Bandura 1977). But the manner in which
people actually behave with regard to a particular topic of discus-
sion, and whether or not information or skills are actually used,
depends more upon interpersonal forces than upon the medin
messages themselves.

Community Organization

The theories and concepts that underlie communit,y organization
are less well developed than those applied to media planning.
Although there is broad agreement that the effects of the media are

505


,.;:y: <!`<`l s! '     ,,,' z.., ,I' ;;..a.:: ;-... i : here is DO clear consensus on the

, )/,_ : ,,:~ :      `~ \`:.- \:I` ho:&- they can be feasibly modified in
~`Il111`c i';Li :,;.: j:.e. : .4.~t.f`;.~i principle is derived from Bandura's
! 1977 i ciistinct `lln br:\veen factors influencing acquisition of new
behaviors and : ~11 )se influencing performance of new behaviors.
Media communications can model new behaviors so that they are
learned (acquired) on a cognitive level (the person knows how to
perform the behavior). However, cueing and feedback (direct social
reinforcement: are usually needed for behavioral learning, or actual
performance of the new behavior. Numerous studies of learning via
media commumcation show that when a complex beha?rior is being
learned, effectiveness is sharply enhanced by providing supplemen-
tary interpersonal communication for encollragement, feedback, and
reinforcement (Bandura 1977). To create feedback and reinforce-
ment in a communiLy setting, organizations must be involved to
provide roles and structure for interpersonal communication. Where
formal so&: networks are not involved, communication and influ-
ence will be diffused through families and other informal systems
(Meyer ct al. 1977). The effectiveness of interpersonal communica-
tion can be greatly enhanced, however, by organizing formal or
semiformal structures, such as learning groups using leaders trained
to lead discussions, answer questions, and provide encouragement
and followup. Various campaigns in agricultural development illus-
trate these principles (Green 1970; Rogers and Shoemaker 1971).

A related factor is the notion of generalized social support (Caplan
et al. 19751, which refers not to the differential social reinforcement
of specific behaviors, but to the general extent and quality of
interpersonal relationships. Social relations appear to be generally
helpful, probably because of their "stress-buffering" effects. Social
ties within the family probably enhance cessation and prevention of
tobacco use. For example, spousal support leads to higher successful
quit rates (West et al. 1977; Mermelstein et al. 19831, and lower rates
of teenage smoking occur in families in which neither parent smokes
(National Institute on Education 1979). The general enhancement of
interpersonal support networks is, of course, a primary objective of
religious groups and social work and most other helping professions.

Environmental Change

There are also theories and concepts from which environmental
changes can be planned (Bandura 1977; Craik 1973), the basic
principle of which is to modify the availability and cost of products or
behaviors (such as by limiting supply or prohibiting behaviors in
public settings). Not all such measures can be applied by communi-
ties as defined herein. For example, regulations on mass media
advertising can probably be controlled in most cases only at the
Federal level, although billboard advertising may be amenabie to

506


more local control. Other promising interventions such as taxation
&wit et al. 1981; Fugi 1980) can be applied within fairly small
localities, but the risk of "black market" competition is lessened
when economic controls are fairly uniform across larger geographic
units. Product availability and regulation of behavior can be
achieved by towns or countries, but restrictive regulations almost
invariably arouse opposition unless the public is yvilling to self-
enforce the restrictions. Therefore, it is legitimate to favor voluntary
restraints over those that require formal policing. Syme and Alcalay
(1982) call for intervention and prevention efforts at the community
level using a public education agenda Such programs will seek to
increase public awareness of the health consequences or smoking,
create an atmosphere in which smoking is recognized as a minority
behavior, influence public policy, and increase antismoking adver-
t.isements.

Breslow (1982) has recently reviewed the environmental and
public policy approaches to smoking control and calls for a "compre-
hensive strategy that will mobilize all available resources most
effectively" (p. 149). He advocates Federal, State, and local legisla-
tion as the most important forms of social action directed t,oward
action alternatives as well as research. He states that "protection
and advancement of economic Interests wili gene;ally follow prevail-
ing ideology. Finding ways of cutting through the economic barriers,
as always, will pose a challenge to public health. Whi!e co.mprcmises
will be necessary, the objective of steady movement toward the g:naT
now seems attainable" :p. 149).

Cessation Studies

Controlled community studies otl smoking ressI.::on are stil!
relatively scarce. However, community trials for carciiovascuiar
disease prevention, in which cigarette smoking is ill+' major risk
factor, are providing son12 of the best examples of I 2 ". .: 1; in tl;is
area. These studies have tended to focus on cessation among adu!ts
because primary outcomes of the trials include possib!e short-tltrm
(5 to lo-year) effects on c,,lrdiovascular mortality and morbidity
rates that could hypothetically result from widespread adult cessa-
tion. These studies were reviewed ir. depth in the 1983 Report of the
Surgeon General T/z Health C'onsc:~u~nces of` ,!nlc;king (USDHHS
1983).

Stanford Three-Community Study

The most well known U.S. cardiovascular community study was
conducted in California by Farquhar, Maccob:r, and colleagues at
Stanford University (Farquhar et al. 1977; Maccoby and Alexander
1980; Meyer et al. 1980). Beginning in 1972 and ending in 1976, the

507


study was supported through the National Institutes of Health
research grant program and involved three small communities
(population of each was approximately 20,000) nonrandomly as-
signed to control, media-only, or media and face-to-face programs.
The three towns are all within 100 miles of Stanford University. The
control town, Tracy, is located in an inner valley and not exposed to
media sources common to the other two towns, Gilroy and Watson-
ville. Gilroy, assigned to the media-only condition, is situated in a
coastal valley, and Watsonville, receiving a small additional inter-
personal communication program, is on the coast. In Watsonville, a
cardiovascular high risk group including many smokers (n= 169) was
identified, and 113 cases were randomly assigned at a 2:l ratio for
face-to-face intervention. The three towns are demographically
similar, although the proximity of Watsonville and Gilroy to the
larger cities of Santa Cruz and San Jose gives them more cosmopoli-
tan features than Tracy. Nevertheless, the research design was
probably the best balance of feasibility and external and internal
validity that could have been achieved in that setting given the
limited available resources.

In each town, multistage probability samples of households were
contacted and invited to a survey station where questionnaires and
physiological measures were administered. These surveys were
applied in the autumn months, beginning in Watsonville in Septem-
ber and ending in Tracy in November. Approximately 600 persons
aged 35 to 59 were sampled at each location. The measurements
included questions about smoking, and serum samples from high risk
participants were analyzed to estimate thiocyanate concentrations
as a check on inaccurate reporting of smoking status (Meyer et al.
1980). These measurements were taken annually for 4 years,
yielding a picture of 3-year smoking trends among the survey
participants in the three communities.

The program of media communication was conducted over 3 years
(1973 to 1975), with greatest intensity in the first and second years of
work. Television, newspapers, radio, billboards, and direct mail
advertising were designed to provide information and to model
attitudes and skills that would promote behavioral changes associ-
ated with lowered cardiovascular risk, such as weight reduction,
lowered fat consumption, and increased exercise. To encourage
cessation of smoking, information about its harmful effects was
given, along with advice on how to stop smoking. In booklets mailed
to the sampled households in Gilroy and Watsonville, instructions
were provided for simple self-control skills (Meyer et al. 1980). In
brief television and radio communications, actors were shown
recommending or modeling cessation of smoking in a variety of
authoritative and entertaining ways.

508


The face-to-face, intensive instruction program was provided for
113 randomly assigned high risk participants in Watsonville, of
whom 107 started treatment, and 77 continued for the second annual
examination. Activities were based on principles of behavioral
psychology and group dynamics and were designed to reinforce and
train skills for behavior change (Meyer et al. 1980). The first-year
program consisted of classes and home visits, mostly during the
summer of 1973. During the summer of the following year, aggres-
sive followup activities were conducted to reinforce smokers who
reported cessation and to encourage and train those who were not
yet able to quit. This maintenance program included training in
stress management and other intensive, individual counseling for
those who consented to continuing contact. In the third year, the
activities were gradually reduced to telephone contacts and a small
"reunion" in the summer of 1975.

The results of the program among high risk participants in each of
the four surveys are displayed in Figure 1. Over the 3 years of study,
the prevalence of smoking decreased markedly among the group
receiving media and face-to-face communication. The group receiv-
ing media intervention showed an initial decline as compared with
the control group, but the change did not differ from the modest
reduction observed in Tracy over the entire 3-year period. Because
the high risk samples included most of the older smokers in the
survey samples from each community, the data on cessation in the
complete samples corresponded closely to that of the high risk group.
The serum thiocyanate tests indicated very slight overreporting of
cessation (Farquhar et al. 19771, but self-reported cessation rates
were not adjusted for thiocyanate findings (Kasl 1980). There was
some attrition in this longitudinal study, but not enough to account
for the clear differences in cessation rates between the intensively
instructed and the other participants. Adjust.ing for attrition, 32
percent reported sustained cessation in the intensive intervention
group (Meyer et al. 1980). This research supports the hypothesis that
face-to-face communication is a necessary part of a successful
community program to reduce smoking. Farquhar and his colleagues
(Farquhar et al. 1981) conclude that the question of how such
communication can be feasibly and cost-effectively provided on a
widespread basis remains to be answered by further studies, and this
conclusion will be noted in a later section. A number of critical
comments can be made in regard to the Stanford three-community
study, and these have been thoroughly discussed in numerous
publications (Leventhal and Cleary 1980; Meyer et al. 1980). The
primary shortcomings concern the quasi-experimental research
design and the inability to generalize from the longitudinally
followed study group to the entire community.

509


GIlroy-media only

3

FIGURE l.-Comparison of cessation rates among smokers
      in communities subjected to varied intensities
      of education, the Stanford three-community
       study
SOURCE Farquhar et al f 1981 I

Australian North Coast Program

A trial very similar to the Stanford three-community study was
conducted by the regional Health Department of New South Wales,
Australia (Egger et al. 19831, beginning in 1978 and, as in the
Stanford study, using small community populations of 12,000 to
27,000 persons. The towns of Tamworth, Coffs Harbor, and Lismore
were assigned to one of three conditions: control, media only, or
combined media and community programs. The three towns were all
between 300 and 400 miles from the research center in Sydney.
Coffs Harbor, the media-only town, was approximately half the size
of the other two communities. Tamworth was not served by the
regional administrative center in Sydney.

In each town, a series of random sample surveys was conducted.
Interviews and physiological examinations were requested from up
to two adults in randomly sampled households. Those who refused
appointments were given self-report questionnaires to be completed
at home. Separate samples were drawn in 1978, 1980, and 1981 with
the objective of measuring 600, 1,200, and 1,200 different persons in
the 3 respective years. Smoking behavior, attitudes, and knowledge

510


regarding smoking were measured, and in 1980 and 1981, serum
thiocyanate values were determined for a randomly selected 5
percent subsample. Measures were also made of other cardiovascular
disease risk factors in accordance with the program's primary goal.

The media program included output from a television station, a
radio station, and several small newspapers, supplemented by other
materials such as stickers, posters, T-shirts, and balloons. The three
stages of the media campaign were designed to sequentially raise
awareness, provide information, and stimulate action. Advertising
time was purchased to insure presentation of television commercials
during peak viewing periods. Beginning in October 1979, the print
advertisements were suspended for several months because of a
complaint to the Media Council of Australia. The overall campaign
continued for approximately 1 year. The community program.that
was applied in Lismore was varied. Several different kinds of groups,
clinics, workshops, and other interpersonal support systems were
organized; physicians were also involved. A total of 386 smokers
participated in these activities, most (150) joining a lday workshop.
The 3-month success rate in smoking cessation, measured by
telephone interview, was 16 percent for the workshop participants.
The highest success rate (48 percent) was found for the 40 persons
who received help kits from physicians. The community program
also included other programs (e.g., for physical fitness and stress
management), which may have facilitated the cessation of smoking.

The results of the project were based on the three independent
sample surveys. The authors point out that "there was substantial
confounding owing to age and sex differences between towns" (p.
1127). Analyzing results according to a multiple logistic model
controlling for age and sex differences, a significant treatment effect
(p=O.O5) was observed. Thiocyanate analyses showed no difference
between towns in the small estimated invalidity of self-reports (3
percent). Among men and women in various age groups, consistently
different rates of change in prevalence of smoking were found. In
Lismore, absolute percentage reductions ranged from 6 to 15
percent. In the media-only town (Coffs Harbor) absolute 6 to 11
percent reductions were estimated. In Tamworth (the control)
absolute reductions of only 2 to 5 percent were found. The research-
ers found no evidence of effects on knowledge or attitudes.

Many criticisms can be made from the standpoint of a clinical
scientist accustomed to dealing with individual subjects in highly
controlled settings. If limitations to inference are clearly acknowl-
edged, however, the Australian study provides a practical illustra-
tion of what may be achieved through community intervention. The
findings are based on independent samples, and thus represent
changes that seem to have occurred on a communitywide basis. The
magnitude of the apparent effect was particularly encouraging

511


among rhe youngest age groups I 18 to 25) where absolute reductions
in prevalence of smoking showed a threefold difference between the
maximrlm intervention (15.6 percent) and control towns (5.0 per-
cent 1. The most significant limitation arises out of the nonrandom
assignment of communities and problems with their comparability.

Swiss National Research Program

Another important community study was conducted in Switzer-
land (Autorengruppe Nationales Forschungsprogramm 1984; Gutz-
willer and Schweizer 1983). Four communities of 12,000 to 16,000
inhabitants were selected, two each from the German-speaking and
the French-speaking parts of the country. A fifth community in the
Italian-speaking region was also studied, but only for epidemiological
purposes. French-speaking and German-speaking pairs were ran-
domly assigned to intervention (Nyon) or regular care (Solothurn
and Vevey) conditions. This project was conducted over a 4-year
period with research support from the Swiss National Science
Foundation. A baseline assessment was made in late 1977 and early
1978 by stratified random sampling and examination of 2,000
persons aged 16 to 69 in each community. With attrition of
approximately 30 percent, this sample was resurveyed at the end of
1980, at which time another independent sample was drawn and
surveyed. A questionnaire was used to determine smoking behavior,
and plasma thiocyanate was measured on a subsample of respon-
dents. Other health-related factors such as blood pressure, lipid
fractions, exercise tolerance, and psychosocial adjustment were also
measured.

The community interventions were conducted over 2.5 years in
Nyon, with guiding principles of "active local participation" and
"integration into existing local health and social services." The
central feature was the establishment of a Citizen Health Action
Committee in each of the two towns, with a coordinator assigned to
guide local planning and implementation. Media and community
organizations were combined to promote a variety of programs in
each location, including classes, self-help groups, and meetings to
discuss topics such as environmental regulation via public non-
smoking areas.

The results of this study are moderately encouraging. Within the
sample population surveyed at baseline and followup, 26 percent of
the regular smokers in the intervention communities reported
cessation. In the control areas, 18 percent of the corresponding group
reported cessation. The investigators also report effects on other
cardiovascular disease risk factors, although plasma cholesterol
reductions were significant only for women in the German-speaking
region. A cost-benefit estimation model has been applied to the data
and results indicate a twofold cost-benefit ratio. The use of indepen-

512


dent surveys and the random assignment of communities represent
significant methodological strengths as compared with the Stanford
and Australian studies. However, the rural communities in this
study were very small. and only four were included. The final report
has been presented to formal decisionmakers to determine whether
broader national efforts are warranted.

The North Karelia Project

The best documented long-term community study is being conduct-
ed in Finland by Puska and colleagues (Puska et al. 1979, 1981,
1983a; Puska and Koskela 1983; McAlister et al. 19821. Also
beginning in 1972 and continuing to the present, the research is
comparing changes in cardiovascular disease risk factors in two
neighboring counties of eastern Finiand, Kuopio and North Karelia.
Both are large rural areas with numerous small farming, lumber. or
mining communities and a single major town. North Karelia is
representative of eastern Finland as a whole in having one of the
world's highest rates of cardiovascular disease (Pyorala 1974; World
Health Organization 1975). Financial support for the intervention
came from the Finnish Ministry of Health, following a formal
request from leaders in North Karelia for help in reducing the high
mortality and morbidity levels. Research funds were awarded by the
Academy of Finland. The neighboring county of Kuopio was selected
as a reference or control location. In North Karelia, a broad program
was implemented to provide new services, education, and training
through community health centers, the mass media, and a variety of
community organizations. During the second 5 years of the project,
media programs were carried out on a national level, with special
organizing and support for activities in North Karelia.

In both counties, independent samples of households were drawn
in 1972,1977, and 1982, with approximately 5,000 persons aged 25 to
59 in 1972 sampled in each area in the first two surveys and about
4,000 aged 25 to 64 sampled in the 1982 measurement. Response
rates were generally excellent, and nearly 90 percent of the sample
participated by attending local survey centers in the spring of each
of the survey years (1972, 1977, 1982). Self-reported cigarette-smok-
ing behavior was measured in all three surveys. In the 1977 survey,
serum thiocyanate values were estimated for a subsample, and in the
1982 survey, for all participants. For the 1977 sample, there was 99
percent agreement in smoking status (smoker/nonsmoker), and
when classified by intervals of 5 or 10 cigarettes, the agreement
between results was 93 and 97 percent, respectively (Puska et al.
1979). The age-adjusted partial correlation between daily reported
number of cigarettes and serum thiocyanate in 1982 was approxi-
mately 0.7 among men and women in both areas. In the years when
the larger surveys were not conducted, smaller samples (1,200 to

513


3,500) were selected yearly for a postal survey of North Karelia.
Together, these measurements provide a comparative view of trends
over 10 years in the populations of both counties and a year-to-year
picture of the changes in North Karelia.

The program of service, education, and training was very broad in
scope. Initially, an intensive educational campaign was conducted
for reduction of cigarette smoking with cooperation from the news
media. Physicians and public health nurses staffing community
health centers were provided special training and were encouraged
to recommend cessation to all patients visiting the centers. Tens of
thousands of leaflets and posters were distributed to encourage
nonsmoking. With assistance from Heart Association volunteers in
each small community, informal restrictions on smoking and point-
of-purchase advertising were adopted. During 1976 and 1977, these
measures became part of a package of national legislation that
increased cigarette tax revenue, directed health services to provide
information services, limited public smoking, and banned tobacco
advertising.

During the second 5-year period, an effort was made to provide
nationwide applications, while maintaining intensive community
work in North Karelia. A series of programs was broadcast national-
ly on television to demonstration how "average" people stop smoking
(Puska et al. 19811. In association with these broadcasts, special
organizing campaigns were conducted in North Karelia to increase
social support for people attempting to quit. These activities
occurred in the winters of 1978, 1979, 1980, and 1982 and were seen
nationally by a majority of the population, with higher viewership in
North Karelia. During the first broadcast series, an effort was made
to encourage the formation of informal volunteer-led self-help
groups to view the broadcasts together. Very few volunteers succeed-
ed in establishing groups, however, and focus in subsequent years
has been on training volunteers to provide even less formal cueing,
reinforcement, and support in their incidental, day-today contacts
with cigarette smokers (Puska et al. 1981).

The results to date (Puska and Koskela 1983) are presented in
Table 1. Over the lo-year period, self-reported numbers of cigarettes
smoked per day fell by more than one-third among men in North
Karelia. In the control or reference area, a less than 10 percent
reduction was observed. Changes in prevalence of smoking account
for most of this difference. No evidence of an effect occurred among
women, with rates of smoking going up in both areas. In Figure 2,
the year-to-year data for 25- to 29-year-old men and women in North
Karelia show an interesting pattern, with the sharpest declines
among men associated with the first year of work and with the
television broadcasts and associated activities in 1978 to 1980 and in
1982. Since 1978, when new antismoking laws were passed, the

514


proportion of male smokers aged 15 to 64 has changed from 44 to 31
percent in North Karelia, and from 39 to 35 percent in the rest of the
country, a difference of 9 percent in absolute rates of change over
that 4-year period. Although the effect of changes in smoking cannot
be separated from the effects of new hypertension services and other
measures to prevent cardiovascular disease, there is some early
indication that mortality rates may have been influenced by the
program: a 24 percent decline in cardiovascular deaths has been
observed in North Karelia, compared with a 12 percent decline
nationally in Finland (Puska et al. 1983a, bj.

The North Karelia Project has received much attention, and
various points of controversy have been widely discussed. The
methodology of the study does not compare with that achieved in
controlled clinical trials, but it may have been the optimal design
that was feasible in the circumstances. Puska and his colleagues
point out that "it is easy to say that the North Karelia Project was
successful because of the unique historical background and because
the conditions in North Karelia were favorable for the program.
However, at the planning stage, great concern was expressed
because the area was rural, of low socioeconomic status with high
unemployment, and so forth." Because a large number of indepen-
dent units were not randomly assigned to experimental and control
conditions, the Finnish study cannot be taken as a conclusive test of
the effects of community programs, but it does provide a promising
illustration and evaluation of what can be achieved through broad
and vigorous intervention to reduce smoking behavior. Its major
strengths are the relatively large number of different communities
that were studied and the lo-year followup interval.

Other Large-Scale Studies

There have been a number of other large-scale controlled studies,
single and multifactor clinical trials, and worksite trials that provide
a context for the consideration of the community-level intervention
studies described above. These studies were discussed in the 1983
Report of the Surgeon General The Health Consequences of Smoking
and include the London Civil Servants Smoking Trial (Rose and
Hamilton 1978; Rose et al. 1980, 19821, the Gijteborg (Sweden) study
(Werko 1979; Wilhelmsen 1981; Wilhelmsen et al. 19721, the Oslo
(Norway) study (Hjermann et al. 1981; Holme et al. 19811, the World
Health Organization European Collaborative Trials (WHO European
Collaborative Group 1974; Kornitzer et al. 1980a, b) and the Multiple
Risk Factor Intervention Trial (MRFIT) (Hughes et al. 1981;
MRFIRG 1982). In the sole American study (MRFIT), 12,866 men, 35
to 57 years old and at high risk for coronary heart disease (CHD),
were entered into a randomized clinical trial designed to test the
effect of a multifactor intervention program on CHD morbidity and

515


TABLE l.-Mean amount of reported daily smoking (*SD) in North Karelia and the reference area, in
     independent baseline (1972) and byear (1977) and IO-year (1982) followup survey samples,
      by sex and age

Sex and age

(Y-d

1972

North Karelia

  1977

1982

1972

Reference area

   1977

1982

Net difference, percentf@e '

197P1977   19721982

Men
3049
40-49
w.59
Total

Women
30-3
4@49
5049
Total

10.6 + 11.7    9.0 f 12.5    7.2 + 10.7
10.2 * 11.2    6.5 + 11.6    7.0 + 10.3
9.0 t 10.7    7.9 + 11.5    5.7 +  9.7
10.0 + 11.3    8.5 + 11.9    6.6 2 10.2



1.4 +  4.2    1.6 +  5.1    2.6 zk  5.8
1.2 +  4.0    1.1 +  4.0    1.5 *  4.4
0.7 +  3.2    0.7 +  3.4    0.9 f  3.6
1.1 +  3.8    1.1 +  4.2    1.7 +  4.7

8.5 + 11.0    9.0 t 11.9    8.6 + 11.5             16       33
9.0 + 10.7    8.6 ~fr 11.4    8.0 + 11.7             10       21
7.7 +-  9.9    7.7 + 10.5    6.5 +  9.6             16       23
6.5 f 10.6    8.5 + 11.4    7.8 f 11.1             13       28

1.5 f  4.5    2.0 +  4.6    2.8 k  6.2             15        2
1.0 +  3.2    1.2 +  5.0    1.7 -t  4.9             23       36
1.0 +  3.6    0.8 +  3.2    1.2 *  4.0             43       41
1.2 *  3.8    1.3 k  4.4    1.9 +  5.1              8       14

Four-way ANOVA'       19721977

Area            P<O.O5
Time               n.a.
Area-time          p<O.Ol
Area-tim~X        p<o.m

19721982

".8.
p<O.OOl
p<O.ool
p<O.ool

ANOVA for linear
trend 1972. 1977, 1982

   Men
   Women

North Karelia

p<O.OOl
p<O.all

Reference area

n.a.
p<o.au

`(North Karelia. 1982/1977-1972) - (Reference area, 1982/1977-1972)
~ANOVA=Ady~isofvariance.
SOUKCE: Puska and Kmkela (19a7).


FIGURE 2.-Prevalence of smoking among men and
     women, aged 25-29, North Karelia, 19721982
SOURCE- Puska and Kcakela f 19831

mortality. They were followed for an average of 7 years. At intake,

517


TABLE 2.-Results of community studies compared with
      other large-scale studies

Study

Years of   Net percent reduction
study      in smcking '

Strengths

Stanford Three
Community study

3

l&20

Matched
communities

Panel study
OtIlY

Australian
North Coast
study

3

15

Independent
samples

Problems with
comparability
of groups

Swiss National
Research
swam

3

8

Randomization,
independent
S4%lllpl@3

Small size and
number of
study sites

North Karelia
project

10

25'

Numerous
communities,
independent
samples

Nonrandom
assignment

Other large
scale studies 3

2-10

.w5

Internal
validity

EXtHlk31
validity

I Difference between percent reductmn in proportion of smokers m the maximum interventnn verws control
condlLrons.
*Difference between percent reductum m the mean number of cigarette smoked per day among men.
`Clmical and work-site tnals: the London CIVI! Servants Smokmg Trial, the Gotebarg study. the Oslo study, the
WHO Collaborative Trml. and the Multiple Rlak Factor Intervention Trml.

the self-reported prevalence of smoking was 64 percent. The 6,428
men randomized into the Special Intervention (SI) group received
intensive group sessions, personal instruction, and other followup to
encourage and support cessation of smoking, as well as intensive
assistance to alter other CHD risk factors. There were 6,438 men in
the group randomly assigned to Usual Care WC). At 6-year followup,
self-reported cessation rates were 43 percent in the SI group and 25
percent in the UC group. Thiocyanate analyses showed a necessity
for small adjustments in these figures to 42 percent and 24 percent,
respectively, a statistically significant difference (p < 0.01). Compari-
son of the results from the four community studies with those from
the large-scale studies, as shown in Table 2, indicates distinct
similarities. The individual studies each have somewhat different
weaknesses, but all indicate that absolute reductions in smoking
prevalence in intervention communities are about 12 percent
greater than reductions in comparison communities.

Several ambitious community studies of cardiovascular disease
prevention are currently in progress with support from the National
Heart, Lung, and Blood Institute, but no additional data on
community-level intervention are available. At Stanford University,
a large study of two intervention and three control towns with
several hundred thousand residents has been in progress since 1978
and is projected to continue for at least 10 years (Farquhar I978). A

51r


similar study, with three matched pairs of variously-sized towns and
cities, has been in progress in Minnesota since 1979 cRla:sk'aurn et
al., in press; Blackburn, in press; Blackburn and Pechacek, in press).
A smaller study of one Rhode Island town a.ld one town in a
neighboring State began in 1980 (Lasater 1983). Because it has
become clear that community studies are the most natural and cost-
effective method for testing new public health services to reduce
chronic disease (Farquhar 1978; Puska et al. 1983a. b), more of these
efforts may be useful, particularly if costs can be reduced and, where
possible, absorbed into existing services.

Related Studies of Cessation

A number of recent efforts have increased understanding of and
confidence in the methods employed in large-scale community
studies. Research on methods of stimulating interpersonal support
for mass media programming is particularly relevant (Colletti and
Brownell 1982). In quasi-experimzntal studies, McAlister et al. (1980)
and Puska et al. (1981) have reported methods for facilitating the
effectiveness of televised smoking cessation classes in Finland.
Formal self-help groups appear difficult to organize, but reorganiza-
tion of less formal social reinforcement in natural interaction
sett,ings appears feasible and effective. Related studies of television
and other media-based methods are described by Danaher et al.
(1983), Best (1980), Lerlthar (1981), and others. Dubren (1977a, b),
Brengelmann (1976), and the American Cancer Society (1981) found
that effects of media programs may be enhanced if a telephone
hotline is offered. Flay et al. (1983b) used a school-based, family-
oriented prevention program that included a five-segment television
Tomponent to be aired the following week to encourage participation
of cigarette-smoking parer& Overall, parents of students in exwri-
mental groups were over three times more likely to view the
cessation segment than parents of control students, with similar
differences observed for the proportions of successful parental
attempts to quit. Within the experimental groups, teacher training
had a significant effect (p< .Cl) on raising participation rates by
parents of program students. Considering only homes with smokers,
51 percent of the students with trzined teachers reported that at
least one cigarette smoking adult viewed one or more cessation
segments, compared with 37 percent of students with untrained
teachers. Furthermore, 38 percent of parents with trained teachers
attempted to quit smoking, compared with 24 percent of parents of
students with untrained teachers !p< .OOl). At l-year followup, the
cessation rates in the two groups were 19 and 13 percent, respective-
ly, according to children's reports of parental behavior. The validity
of the indicators of adult smoking behavior is at issue; student
reports could contain bias, which will be estimated in further

519


analyses. Nevertheless, these results suggest that children can
enhance the effectiveness of a media program in encouraging
parents to stop smoking, and that organized social reinforcement is
important in mass media smoking cessation programming (Flay et
al. 1983b). More research is needed on these and other methods of
large-scale social reinforcement and support for cessation of smok-
ing.

Some attention has been given to environmental changes that
might contribute to the cessation of cigarette smoking. Evidence
indicates that smoking can be regulated by counteradvertising,
restrictions on advertising, warning labels, and symbolic or govern-
mental actions such as the Surgeon General's Report of 1964
(Warner 1977). Complete prohibition of smoking is not an acceptable
alternative, but restrictions on smoking locations may be helpful for
people attempting to quit voluntarily (Horwitz et al. 1982). Taxation
of cigarettes has provided an effective deterrent in some population
groups, particularly among younger men (Lewit et al. 1981). In
Finland, cigarette taxes have been increased and a portion (0.5
percent) of the funds are dedicated to support services to reduce the
prevalence of smoking (Puska and Koskela 1983). Similarly, integrat-
ed environmental and educational interventions may be useful in
further large-scale efforts.

Prevention Studies

Although evidence increasingly indicates that it is more cost
effective to prevent the onset of smoking among young people than
to change the dependent behavior of adults, smoking prevention has
received far less attention than cessation in controlled community
research. This can be attributed to the complex processes involved in
the onset of smoking (e.g., Evans 1976; Leventhal and Cleary 19801.
to the extended timespan required for proper evaluation of preven-
tion studies, and to the overall tendency of the health sciences to
focus more on individual-level research than on primary prevention
studies in which whole populations must be followed. Research on
the prevention of smoking onset was considered in detail in the 1982
Report of the Surgeon General The Health Consequences of Smoking:
Cancer (USDHHS 1982), and only selective studies are reviewed
herein. Although there are no currently available published reports
on controlled community studies of smoking onset prevention in
whole populations, a number of school-based studies have been
conducted recently that involve large proportions of particular age-
grade cohorts in discrete communities.

520


Stanford Study

In one such study McAlister et al. (1979, 1980b) assigned the 1978
seventh grade cohorts in two similar suburban California towns to
control (receiving Usual Car: health education) or to a special peer
prevention program to deter the onset of smoking. The special
intervention program consisted of 7 classroom hours distributed over
the school year in which 15 to 17-year-old peer leaders led Socratic
dialogues, role plays, and simple contests to reinforce smoking
avoidance behavior among 12-year-old students and to promote the
social desirability of nonsmoking. Role plays and simple contests
were used to help students learn assertive ways of declining offers to
smoke, emphasizing counterarguments to the kinds of perceived peer
social pressures that may be associated with the onset of smoking
(McAlister et al. 1979). Several followup sessions were conducted
when the study cohort reached the eighth grade.

At baseline and three followup points, students in the control and
the environmental groups provided anonymous self-reports of smok-
ing through classroom surveys. Participation rates were close to 100
percent in all surveys, and the measures were supplemented by
collecting samples of exhaled breath at followup. These repeated
surveys provided good measures of smoking prevalence in the age-
grades studied in the two similar suburban communities, and
although attrition and replacement may have threatened inference,
the findings in Figure 3 indicate that the special intervention may
have had a preventive effect. Note that the dependent variable is the
percentage of participants who reported smoking during the previ-
ous week. Definitions of smoking in adolescents often differ, making
results difficult to compare. Over 2 years of initial followup, the rate
of self-reported smoking onset diverged sharply in the two age-grade
cohorts, and differences persisted over an additional year of followup
(Telch et al. 1982).

The North Karelia Youth Project

Stemming from the North Karelian project described in an early
section, a youth program was initiated in six rural communities in
eastern Finland (Vartiainen et al. 1983). Two urban and two rural
communities were assigned to intensive (direct contact with the
experimenter) or countywide (teacher-led) intervention programs in
North Karelia. Two matched communities were selected as controls
in the neighboring province of Kuopio, and subjects were seventh
grade students (average age of 13 years) in the autumn of 1978. From
a population of 897 available students, nearly all participated in the
baseline survey; 95 percent participated in a 2-year followup; and 88
percent participated in a second followup 30 months after the
program began. The survey included measurements of self-reported
smoking and related variables as well as serum thiocyanate levels.

521

480-144 0 - 85 - 18


20



15

10

Months of study

FIGURE 3.--Long-term effects of a peer group training
       course on smoking behavior of B-year-old
        students
SOURCE T&h et al (1981).

The special intervention spanned the seventh and eighth grade
years and included 10 hours of instruction in the intensive program
and 5 hours in the countywide program. The methods were based on
those employed by McAlister et al. (1980), as reported in the
preceding study. Peer leaders, students 1 to 2 years older than the
subjects, were the primary agents used to deliver the antismoking
message. The program used role play and other active learning
techniques to teach skills for resisting the social and psychological
pressures to smoke and to reinforce negative attitudes toward
smoking.

The results indicated a preventive effect from the special interven-
tion, as shown in Figure 4. Among boys, the followup rates of
smoking at least once a month were significantly lower in the four
intervention schools (21 to 24 percent) than in the reference area
communities (39 percent). Results for girls were less conclusive, but
the overall difference between groups at followup was significant: a
17 percent rate of smoking at followup in intervention schools and a
24 percent rate in control schools. Analysis of the serum thiocyanate
samples showed that 2 to 3 percent of the professed nonsmokers gave
inaccurate self-reports, but the slight underreporting was greater in
the control schools-indicating that inferences from the reported
findings were not significantly threatened by self-report biases.

Other Prevention Studies

In a more rigorous study, Flay et al. (1983a) assigned 22 schools in
Canada, some randomly, to receive special interventions or to serve
as controls. The smoking prevention curriculum consisted of three
major components, using the social psychological model delivered in

522


Autumn               Autumn Spnng
1970                 1983   1981

FIGURE 4.-Percentage of children who reported smoking
       at least once a month in baseline survey (1978)
      and two followup surveys (1980 and 1981)
SOURCE Vamamen et al (1983)

six l-hour weekly sessions during grade six. The social psychological
model aims at developing future attitude and behavior changes and
acquiring social skills, and involves eliciting information from
children rather than providing it for them. The three components
focused on smoking consequences and reasons for smoking, social
influences promoting smoking, and decisionmaking and public
commitment. Two booster sessions were delivered in both grades
seven and eight. Short-term findings show a preventive effect,
manifested in grade seven by an increased level of experimental
smoking in the control group compared with the experimental
group. In another study, Worden et al. (1983) tracked changes in
smoking rates in two towns receiving a mass media campaign and a
high-intensity program of adult communication skills related to
adolescent smoking prevention. Over 1 year of followup, onset rates
were markedly greater among a set of comparison towns receiving
only the media campaign and some low-intensity community inter-
vention.

523


?;lcAlister ;lYP:j: randomly assigned members of five matched
pairs of community and neighborhood schools to serve as controls or
to receive an experimental prevention program. The study included
students in the f'irst year of secondary school (sixth or seventh grade)
in 1979 in coastal and inland California towns and inner-city and
suburban locations near Boston, Massachusetts. Over 2 school years
of initial followup. sharply lower smoking onset rates were observed
in some of the schools receiving the experimental programs. The
findings are difficult to interpret and suggest that variability among
highly diverse community age-grade cohorts, inconsistent implemen-
tation. and other related factors threaten inferences that may be
drawn from studies of relatively small numbers of unique age-grade
cohorts. Preventive effects on smoking onset have been reported by
other investigators in Houston (Evans 19761, Minnesota (Hurd et al.
1980), New York (Botvin et al. 1980), and elsewhere; they were
reviewed extensively in the 1982 Report of the Surgeon General The
Health Consequences of Smoking (USDHHS 1982). These studies
have tended to show positive results for the nontraditional health
education methods, particularly those using peer teaching and role
plays of saying no or resisting social pressures toward smoking.
Currently, large-scale research on the prevention of smoking is in
progress in the community studies of cardiovascular disease preven-
tion that were cited in a previous section, as well as in other research
centers in the United States and abroad. Internationally, Sweder!
has declared the creation of a "Smoke-Free Generation," and other
countries have taken various steps to deter the onset of smoking
(Wake et al. 1982).

Although studies were not sufficient to confirm the hypothesis. it
is probable that young people not strongly dependent on tobacco are
most sensitive to the various "environmental" policy options for
smoking reduction. For example, the price elasticity of decisions to
smoke regularly (smoking status) is -1.4 for males aged 20 to 25, but
lower for older men (Lewit et al. 1981). This indicates that a 9
percent increase in the price of cigarettes might yield a 15 percent
decrease, in the proportion of male smokers in the younger age group.
Environmental changes related to the marketing, price, or availabili-
ty of cigarettes tend to be implemented in whole States or nations
and are not amenable to controlled demonstration research. How-
ever, as local authorities play an increasing role in various matters,
opportunities for innovative community level research may be
available.

Methodological Issues

In view of the high costs of clinical trials, such as those incurred by
the Multiple Risk Factor Intervention Trial, a strong argument can

524


be made for the cost effectiveness and generalizability of community
studies of chronic disease prevention (e.g.. Farquhar 19'iS~. However,
there are methodological problems with the community studies that
deserve careful consideration. In order to ensure strict adherence to
assumptions of the statistical theories supporting experimental
inference, independent units of observation must be sampled.
Because the behavior and disease rates of people within a communi-
ty are obviously not independent, the data from geographic units
must be aggregated at various levels, such as family, neighborhood,
community, and region. Thus, for example, smoking rates in three
communities assigned to three different experimental conditions
must be treated as three discrete observations, but having only one
observation per condition does not permit use of traditional statisti-
cal procedures for hypothesis testing. By assigning more than one
community to each condition, between-community variance can be
estimated to provide more valid tests of program effect. As the
statistical theory guiding community studies becomes more devel-
oped (Flay and Cook 19811, future research may be expected to
involve more sites, with fewer cases sampled in each site.
Another problem concerns the comparability of groups. Unless a
large number of communities are random1.y assigned to conditions,
the strict methodologist can identify obvious threats to experimental
validity. For example, it might be natural to expect a bias toward the
application of experimental programs in settings favorable to the
adoption of innovation, while using "less interested" communities
for the control group. This inevitably raises questions about the
comparability of communities with regard to socioeconomic status,
cosmopolitan features, or other hard-to-measure social characteris-
tics. If the experimental group has a favorable predisposition at
baseline, inferences about changes in health-related variables are
severely threatened. When a high degree of demographic similarity
between communities can be demonstrated, confidence in inferential
statistics is enhanced. If possible, the communities to be compared
should be assigned randomly to experimental groups or to control
groups.
Problems with comparability are also introduced by the possibili-
ties of experimental contamination or confounding effects of compet-
ing experimental programs. This issue must be thoroughly analyzed
with respect to the North Karelia Project, where it appears that
program results for dietary change and control of hypertension have
been diluted by program spillover and the establishment of new
health services in the reference (control) area.
Related to the issue of comparability of groups within a study is
the problem of generalization to broader populations. For example,
some groups that fall into the lower socioeconomic strata have not
followed the general population trend toward smoking cessation and


may also be at increased risk for smoking-related disease from
concurrent industrial exposures. Results of trials involving such
populations (WHO European Collaborative Trials) should be exam-
ined for overall outcome as well as evaluation of programmatic
elements wherever possible. Differential effectiveness of interven-
tion techniques with varying populations remains to be established.

Another methodological question is raised by studies that rely
primarily on self-reports that may be biased by intervention
programs (Evans et al. 1977; Benfari et al. 1977; Pechacek et al., in
press). Physiological indicators of cigarette smoking are expensive in
studies involving large numbers of individual measurements. Most
researchers take physiological measurements from a subsample of
the group providing self-reports. If there is no evidence of self-report
bias between groups, experimental comparisons can be based on the
self-report data. However, the usefulness of this procedure depends
upon the statistical power of the test comparing relationships
between self-reports and physiological measures in the different
experimental groups. Tests based on very small subsamples will
almost certainly show no statistically significant evidence of self-
report bias, but only because they lack the statistical power to detect
the relatively small differences that might confound inference.
Numerous other methodological points are pertinent to the review of
community studies. For example, cohort studies that track individu-
als over time may be much less generalizable than those that involve
repeated independent surveys, but are critical for studying develop
ment of certain behaviors, such as smoking onset. A great need
exists, therefore, for more focused awareness on the various method-
ological concerns that limit the interpretation of community studies.

Directions for Future Studies

There is a clear need for further research on community-level
intervention to reduce smoking. The challenge is to develop relative-
ly inexpensive methods that can be easily implemented on a large-
scale basis. This will involve refinements in three broad activity
categories: (11 education and instruction related to smoking, smoking
cessation, and smoking prevention; (2) social reinforcement in
support for nonsmoking behavior; and (3) environmental changes
related to cigarettes and cigarette smoking.

Education and instruction methods are needed to convey informa-
tion, attitudes, and skills more effectively as they relate to the
cessation and prevention of smoking. For example, as the factors
contributing to the process of smoking cessation and prevention are
identified (DiClemente and Prochaska 1982; McAlister 1983; Leven-
thal et al. 1980), they can be modeled via television or other forms of
mass communication. Although the schools have an obvious role in

526


smoking prevention, the kinds of educational activities that appear
to produce results are not easily adopted by traditional educators.
Innovative education and training programs can be marketed to
those willing to pay for therapeutic or consultative services related
to smoking cessation and prevention, but less costly methods need to
be developed for communitywide application. The establishment of
smoking cessation programs within existing health services and the
integration of chronic disease prevention with mental health promo-
tion are also needed to effect broad-scale societal education and
change.

Growing evidence indicates that the social reinforcement and
support provided in formal therapies can be effectively evoked at far
less expense by self-help groups and natural helping networks. In the
North Karelia Project, community volunteers were taught to rein-
force learning of smoking cessation skills from television (Puska et
al. 1981). Children may also be powerful, natural sources of social
reinforcement (Flay et al. 1983b). Related methods are being applied
in other ongoing community studies to harness the influence of
natural social networks for antismoking campaigns (Pechacek et al.,
in press). In addition to reinforcing specific behaviors and attitudes
related to smoking prevention and cessation, social environments
can give more general support and assistance to people trying to cope
with stress, strain, and conflict. Given the relationship between
chronic smoking, stress, and alienation, it is reasonable to expect a
positive effect from interventions that reduce stress, improve coping,
and increase social support (Colletti and Brownell 1982).

Other community efforts are being made in noncontrolled con-
texts, as in the following examples. First, the American Cancer
Society has introduced an "Adopt A Smoker" program to involve ex-
smokers in the annual Great American Smokeout Day. Second,
community physicians are being urged to increase the frequency and
intensity of cessation advice given to smokers. Finally, the television
and newspaper media have become involved in the antismoking
campaign, frequently including programming and articles that
feature behavioral scientists discussing smoking behavior and the
techniques of quitting.

Options for environmental change and public policy are varied and
complex (Farquhar et al. 1981). The outright prohibition of smoking
is not feasible, but limited prohibitions on smoking in public and
some private places can have desirable effects in shifting negative
attitudes. Restrictions on marketing are difficult on a local level, but
constraints on advertising or availability may have a powerful effect.
As indicated in a previous section, new taxes or other broad
environmental changes are more likely to affect young smokers
among whom dependence is not firmly established. Thus, they may
be expected to have cumulative, long-term effects on future smoking

527


rates. There is a role for research in clarifying optimal forms of
restriction, but is has not been implemented, even in worksites
where the effects of restrictions could be evaluated (Orleans and
Shipley 1982).

Future studies must also develop an integrated model for combin-
ing cessation and prevention activities. Although the processes
involved in the adoption of smoking are clearly different from those
involved in its discontinuance, there are commonalities to be
explored to find more efficient strategies for smoking reduction. The
research of Flay et al. (1983a) is particularly promising in this
regard. Any program that changes perceptions of social norms or
other environmental factors may influence nonadoption or discon-
tinuance of cigarette smoking, but effects will probably be greatest
on young people not dependent on tobacco or on older age groups
where the adverse health effects of smoking are already becoming
apparent.

Summary and Conclusions

1. Community studies of smoking cessation and prevention are
becoming an established paradigm for public health action
research. Such studies emphasize large-scale delivery systems,
such as the mass media, and include community organization
programs seeking to stimulate interpersonal communication in
ways that are feasible on a large-scale basis.

2. Although there, are methodological limitations to nearly all
communitywide studies, the results yield fairly consistent
positive results, indicating that large-scale`programs to reduce
smoking can be effective in whole populations. Person-to-
person communication appears to be a necessary part of a
successful community program to reduce smoking.

3. Further research is needed, with both improved methodology
and more emphasis on low socioeconomic status groups that
have not yet shown population trends toward reduced smoking.
4. Several promising directions for research are clear, but the
most important future trends will be toward the establishment
of smoking reduction programs within existing health services,
the combination of chronic disease prevention with mental
health promotion via mass media and community intervention,
and the development of social policy to establish integrated
strategies for smoking cessation and prevention.

528


References

AMERICAN CANCER SOCIETY. A Study of the Impact of the 1980 Great American
Smokeout. Summary Report. National program sponsored by Gallup Organization,
Inc., and analyzed by Lieberman Research, Inc., January 1981,7 pp.
ATKIN, C. Research evidence on mass mediated health communication campaigns.
In: Nimmo, D. (Editor). Communication Yearbook III. New Brunswick, New Jersey,
Transaction Books, 1979.

AUTORENGRUPPE NATIONALES FORSCHUNGSPROGRAMM. Wirksamkeit der
Gemeindeorientierten Pruvention Kardiovascularer Krunkheiten. [Effectiveness of
Community-Oriented Prevention of Cardiovascular Disease.] Bern, Hans Huber,
  1984.

BANDURA, A. Social Learning Theory. Englewood Cliffs, New Jersey, Prentice-Hall,
  1977.

BENFARI, R.C., MCINTYRE, K., BENFARI, M.J.F., BALDWIN A., OCKENE, J. The
use of thiocyanate determination for indication of cigarette smoking status.
Evaluation Quarterly l(4): 629438, November 1977
BERNSTEIN, D.S. Modification of smoking behavior: An evaluative review. Psycho-
logical Bulletin 71(6): 418-440, June 1969.

BERNSTEIN, D.A., McALISTER, A.L. The modification of smoking behavior:
Progress and problems. Addictive Behaviors l(2): 89-102, 1976.
BEST, J.A. Mass media, self-management, and smoking modification. In: Davidson,
P.O., Davidson, SM. (Editors). Behavioral Medicine: Changing Health Lifestyles.
New York, Brunner/Mazel, 1980, pp. 371-390.
BLACKBURN, H. Research and demonstration projects in community cardiovascular
disease prevention. Journal of Public Health Policy, in press.
BLACKBURN, H., LUEPKER, R., KLINE, F.G., BRACHT, N., CARLAW, R.,
JACOBS, D., MI'ITELMARK, M., STAUFFER, L., TAYLOR, H.L. The Minnesota
Heart Health Program: A demonstration project in cardiovascular disease
prevention. In: Weiss, S. 0Qlitor). Settings for Health Promotion in Behavioral
Health: A Handbook for Health Enhancement and Disease Prevention. Silver
Spring, Maryland, John Wiley & Sons, in press.
BLACKBURN, H.. PECHACEK, T. Smoking Cessation and the Minnesota Heart
Health Program. In: Nosbakken, D. (Editor). Proceedings of the Fifth World
Conference on Smoking and Health, in press.
BOTVIN, G.J., ENG, A., WILLIAMS, C.L. Preventing the onset of cigarette smoking
through life skills training. Preventive Medicine 9(l): 135-143, January 1980.
BRENGELMANN. J.L. Experimente zur Behandlung des Rauchens. [Experiments in
the Treatment of Smoking.] Stuttgart, Verlag W. Kohlhammer, 1976.
BRESLOW, L. Control of cigarette smoking from a public policy perspective. In:
Breslow, L., Fielding, J.E., Lave. L.B. &litors). Annual Review of Public Health.
Volume 3. Palo Alto, California, Annual Reviews Inc., 1982, pp. 129-151.
CAMPBELL, D.T., COOK, T.D. Quasi-Experimentation: Design and Analysis for Field
Settings. Chicago, Rand McNally College Publishing Company, 1979.
CAPLAN, R.D., COBB, S., FRENCH, J.R.P., Jr. Relationships of cessation of smoking
with job stress, personality, and social support. Journal of Applied Ps-ychology
6of2): 211-219, April 1975.

COLLE'ITI, G., BROWNELL, K.D. The physical and emotional benefits of social
support: Application to obesity, smoking, and alcoholism. In: Hersen, M., Eislee,
R.M., Miller, P.M. (Editors). Progress in Behavior Modification. Volume 13. New
York, Academic Press, 1982, pp. 109-178.
CRAIK, K.H. Environmental psychology. Annual Reuierc, of Ps.vchology 24: 403422,
  1973.

DANAHER. B.G., BERKANOVIC, E., GERBER, B. Smoking and television: Review of
extant literature. Addwtit,e BehaLllors 812): 173-182. 1983

529


DiCLEMENTE, CC., PROCHASKA, J.O. Selfchange and therapy change of smoking
behavior: A comparison of processes of change in cessation and maintenance.
Addictioe Behaviors 7(2): 133-142, 1982.

DUBREN, R. Evaluation of a televised stopsmoking clinic. Public Health Reports
92(l): 81-84, January-February 1977a.

DUBREN, R. Self-reinforcement by recorded telephone messages to maintain
nonsmoking behavior. Journal of Consulting and Clinical Psychology 45(3): 35%
360, June 1977b.

EGGER, G., FITZGERALD, W., FRAPE, G., MONAEM, A., RUBINSTEIN, P.,
TYLER, C., MCKAY, B. Result of large scale media antismoking campaign in
Australia: North Coast "Quit For Life" Programme. British Medical Journal 286:
112~1128,1983.

EVANS, R.I. Smoking in children: Developing a social psychological strategy of
deterrence. Preventive Medicine 5(l): 122-127, March 1976.
EVANS, RI., HANSEN, W.E., MITTLEMARK, M.B. Increasing the validity of self-
reports of smoking behavior in children. Journal of Applied Psychology 62(4): 521-
523, April 1977.

FARQUHAR, J.W., WOOD, P.D., BREITROSE, H., HASKELL, W.L., MEYER, A.J.,
MACCOBY, N., ALEXANDER, J.K., BROWN, B.W., Jr., McALISTER, A.L.,
NASH, J.D., STERN, M.P. Community education for cardiovascular health. Luncet
l(8023): 1192-1195, June 4,1977.

FARQUHAR, J.W. The community-based model of life style intervention trials.
American Journal ofEpidemiology 108(2): 103-111, August 1978.
FARQUHAR, J.W., MAGNUS, P.F., MACCOBY, N. The role of public information
and education in cigarette smoking controls. Canadian Journal of Public Health
72(6): 412-420, November-December 1981.

FLAY, B.R., COOK, T.D. The evaluation of public communication campaigns. In: Rice,
R., Paisley, W. (Editors). Media Campaigns. Beverly Hills, Sage, 1981.
FLAY, B.R., d'AVERNAS, J.R., BEST, J.A., KERSALL, M.W., RYAN, K.B. Why
young people do it and ways of preventing it. In: McGrath, P., Firestone, P.
(Editors). Pediatric and Adolescent Behauioml Medicine. Springer Series on
Behavior Therapy and Behavioral Medicine, Volume 10. New York, Springer
Verlag, 1983a, pp. 132-183.

FLAY, B.R., DITECCO, D., SCHLEGEL, R.P. Mass media in health promotion. Health
Education Quarterly 7: 127-143,198O.

FLAY, B.R., HANSEN, W.B., JOHNSON, CA., SOBEL, J.L. Involvement of Children
in Motiuating Smoking Parents to Quit Smoking With a Television Progmm. Paper
presented at the Fifth World Conference on Smoking and Health, Winnipeg,
Manitoba, July 1983b.

FRITSCHLER, A.L. Smoking and Politics: Policy-Making and the Federal Bureau-
cracy. Second Edition). Englewood Cliffs, New Jersey, Prentice-Hall, 1975.
FUJI, E. The demand for cigarettes. Applied Economics 12: 479-489,1980.
GREEN, L.W. Identifying and overcoming barriers to the diffusion of knowledge
about family planning. Advances in Fertility Control X2): 21-29, 1970.
GRIFFITHS, W., KNUTSON, A.L. The role of maas media in public health. American
Journal of Public Health 50(4): 515-523, April 1960.
GUTZWILLER, F., SCHWEIZER, W. Intervention on smoking: An individual and
collective challenge. In: Schettler, F.G., Gotto, A.M., Middelhoff, G., Habenicht,
A.J.R., Jurutka, K.R. (Editors). Atherosclerosis: Proceedings of the Sixth Znterna-
tionnl Symposium. New York, Springer Verlag, 1983.
HJERMANN, I., VELVE BYRE, K., HOLME, I., LEREN, P. Effect of diet and
smoking intervention on the incidence of coronary heart disease. Report from the
Oslo study group of a randomized trial in healthy men. Lancet 2(82591: 1303-1310,
December 12,1981.

530


HOLME, I., HELGELAND, A., HJERMANN, I., LEREN, P. The Oslo study: Social
indicators, risk factors and mortality. In: Bostrom, H., Ljungstedt, N. (Editors).
Medical Aspects of Mortality Statistics. Skandia International Symposia. Stock-
holm, Sweden, Almquist and W&sell International, 1981, pp. 165-181.
HORWITZ, M.B., HINDI-ALEXANDER, M., WAGNER, T.J. Psychosocial Mediators
of Long-Term Abstinence Following Smoking Cessation. Paper presented at the
99th meeting of the American Psychological Association, Washington, D.C.,
August 2491982.

HUGHES, G.H., HYMOWITZ, N., OCKENE, J.K., SIMON, N., VOGT, T.M. The
Multiple Risk Factor Intervention Trial (MRFIT). V. Intervention on smoking.
Preuentiue Medicine lO(4): 476-500, July 1981.
HURD, P.D., JOHNSON, CA., PECHACEK, T.F., BAST, L.P., JACOBS, D.R.,
LUEPKER, R.V. Prevention of cigarette smoking in seventh grade students.
Journal of Behauioml Medicine 3(l): 15-28, March 1980.
KASL, S.V. Cardiovascular risk reduction in a community setting: Some comments.
Journal of Consulting and Clinical Psychology 48(2): 143-149, April 1980.
KLAPPER, J.T. The Effects ofMass Communication. New York, Free Press, 1960,302
  PP.

KORNITZER, M., DeBACKER, G., DRAMAIX, M., THILLY, C. The Belgian Heart
Disease Prevention Project: Modification of the coronary risk profile in an
industrial population. Circulation 61(l): 18-25, January 198Oa.
KORNITZER, M., DRAMAIX, M., KI'ITEL, F., DeBACKER. G. The Belgian Heart
Disease Prevention Project: Changes in smoking habits after two years of
intervention. Preuentiue Medicine 9(4): 496-503, July 198Ob.
LANDO, H.A., McGOVERN, P.G. Three-year data on a behavioral treatment for
smoking: A follow-up note. Addictiue Behaviors 7(2): 77-181, 1982.
LASATER, T. The Pawtucket Heart Program. Paper presented at ADAMHA Commu-
nity Prevention Research Technical Review, June 1983.
LEATHAR, D.S. The use of mass media health education campaigns in Scotland.
Journal of International Health Education 19(4): 122129,1981.
LEVENTHAL, H., CLEARY, P.D. The smoking problem: A review of the research and
theory in behavioral risk reduction. Psychological Bulletin B&2): 370-405, 1980.
LEVENTHAL, H., CLEARY, P., SAFER, M., GUTMANN, M. Cardiovascular risk
modification by community-based programs for life-style change: Comments on the
Stanford Study. Journal of Consulting and Clinical Psychology 48(2): 159-158,
April 1980.

LEWIT, E.. COATE, D., GROSSMAN, M. The effects of government regulation on
teenage smoking. Journal of Law and Economics 24: 545573, December 1981.
LICHTENSTEIN, E. The smoking problem: A behavioral perspective. Journal of
Consulting and Clinical Psychology W6): 804-819, 1982.
MACCOBY, N., ALEXANDER, J. Use of media in lifestyle programs. In: Davidson,
P.O., Davidson, SM. (Editors). Behavioral Medicine: Changing Health Lifestyles.
New York, Brunner/Mazel, 1980.

McALISTER, A. Social-psychological approaches. In: National Institute on Drug
Abuse. Preventing Adolescent Drug Abuse. Intervention Strategies. NIDA Research
Series, Volume 47. U.S. Department of Health and Human Services, Public Health
Service, Alcohol, Drug Abuse, and Mental Health Administration, DHHS Publica-
tion No. (ADM)83-1280,1983, pp. 36-50.

McALISTER, A.L., PERRY, C., MACCOBY, N. Adolescent smoking: Onset and
prevention. Pediatrics 63(4): 65&658, April 1979.
McALISTER, A., PUSKA, P., KOSKELA, K., PALLONEN, U., MACCOBY, N.
Psychology in action: Mass communication and community organization for public
health education. American Psychologist 35(4): 375-379, April 1980.

531


McALISTER, A.. PUSKA, P, SALONEN, J.T., TUOMILEHTO, J., KOSKELA, K.
Theory and action for health promotion: Illustrations from the North Karelia
Project. American Journal of Public Health 72~1): 43-50, January 1982.
McCOMBS, M.E., SHAW, D.L. The agenda-setting function of mass media. Public
Opinion Quarterly 36(2): 176187, Summer 1972.
MERMELSTEIN, R., LICHTENSTEIN, E., MCINTYRE, K. Partner support and
relapse in smoking-cessation programs. Journal of Consulting and Clinical
Psychology 51(3): 465-466, June 1983.

MEYER, A., MACCOBY, N., FARQUHAR, J. The role of opinion leadership in a
cardiovascular health education campaign. In: Ruben, B.D. (Editor). Communica-
tion Yearbook I. New Brunswick, New Jersey, Transaction Books, 1977.
MEYER, A.J., NASH, J.D.. McALISTER, A.L., MACCOBY, N., FARQUHAR, J.W.
Skills training in a cardiovascular health education campaign. Journal of
Consulting and CIznical Psychology 48(2): 129-142, April 1980.
MULTIPLE RISK FACTOR INTERVENTION RESEARCH GROUP. Multiple Risk
Factor Intervention Trial. Risk factor changes and mortality results. Journal of the
American Medical Association 248(12): 1465-1477, September 24,1982.
NATIONAL INSTITUTE ON EDUCATION. Teenage Smoking: Immediate and Long-
Term Patterns. US. Department of Health, Education, and Welfare, National
Institute on Education, November 1979,267 pp.
OCKENE, J.K., NUTALL, R., BENFARI, R.C., HURWITZ, I., OCKENE, I.S. A
psychosocial model of smoking cessation and maintenance of cessation. Preventive
Medicine lo(5): 623-638, September 1981.

ORLEANS, C.S., SHIPLEY, R.H. Worksite smoking cessation initiatives: Review and
recommendations. Addictive Behaviors 7(l): l-16, 1982.
PECHACEK, T.F., FOX, B.H., MURRAY, D.M., LUEPKER, R.V. Review of techniques
for measurement of smoking behavior. In: Mataraxzo, J.D., Miller, N.E., Weiss,
S.M., Herd, J.A. (Editors). Behauioral Health: A Handbook of Health Enhancement
and Disease Preuention. New York, John Wiley and Sons, in press.
PECHACEK, T.F., McALISTER, A. Strategies for the modification of smoking
behavior: Treatment and prevention. In: Ferguson, J.M., Taylor, C.B. (Editors).
Extended Applications and Issues. The Comprehensive Handbook of Behavioral
Medicine, Volume 3. Jamaica, New York, Spectrum Publications, 1980, pp. 257-
  298.

PUSKA, P., KOSKELA, K. Community-based strategies to fight smoking: Experiences
from the North Karelia Project in Finland. New York State Journal of Medicine
83(13): 1335-1338, December 1983.

PUSKA, P., McALISTER, A., PEKKOLA, J., KOSKELA, K. Television in health
promotion: Evaluation of a national programme in Finland. International Journal
of Health Education 24(4): 238250, October-December 1981.
PUSKA, P., NISSINEN, A., SALONEN, J.T., TUOMILEHTO, J. Ten years of the
North Karelia Project: Results with community-based prevention of coronary
heart disease. Scandinavian Journal of Social Medicine ll(3): 65-68,1983a.
PUSKA, P., SALONEN, J.T., NISSINEN, A., TUOMILEHTO, J., VARTIAINEN. E..
KORHONEN, H., TANSKANEN. A., RONNQVIST, P., KOSKELA, K., HU'ITU-
NEN, J. Coronary risk factor changes during ten years of a preventive community
programme, the North Karelia Project. British Medical Journal 287: 184G1844,
  1983b.

PUSKA. P., TUOMILEHTO. T., SALONEN, J., NEIITAANMAKI, L., MAKI, J.,
VIRTAMO, J., NISSINEN, A., KOSKELA, K., TAKALO, T. Changes in coronary
risk factors during comprehensive five-year community programme to control
cardiovascular diseases (North Karelia Project). British Medical Journal 2(6198):
1173-1178, November 10,1979.

PYRORALA, K. The epidemiology of coronary heart disease in Finland. A review.
Duodecim 90: 1605-1622.1974.

532


ROGERS, E., SHOEMAKER. F. Communication of Innotjations: A Cross-Cultural
Approach. New York, Free Press, 1971.

ROSE. G., HAMILTON, P.J.S. A randomised controlled trial of the effect on middle-
aged men of advice to stop smoking. Journal of Epidemiology and Communit:
Health 32(4): 275281, December 1978.

ROSE, G., HAMILTON, P.J.S., COLWELL, L., SHIPLEY, M.J. A randomised
controlled trial of anti-smoking advice: iLLyear results. Journal of Epzdemiology
and Community Health 36l21: 102-108. June 1982.
ROSE, G., HELLER, R.F., PEDOE. H.T., CHRISTIE, D.G.S. Heart disease prevention
project: A randomized controlled trial in industry. British Medzcal Journal 116216):
747-751, March 15,198O.

SAPOLSKY, H.M. The political obstacles to the control of cigarette smoking in the
United States. Journal of Health Politics. Polic? and Lnu 512): 277-290. Summer
  1980.

SYME, S.L., ALCALAY, R. Control of cigarette smoking from a social perspective. In:
Breslow, L., Fielding, J.E.. Lave, L.B. (Editors). Annual ReLiew of Public Health.
Volume 3. Palo Alto, California, Annual Reviews Inc.. 1982, pp. 179-199.
TELCH, M.J., KILLEN, J.D., McALISTER, A.L.. PERRY. C.L., MACCOBY. N. Long-
term follow-up of a pilot project on smoking prevention with adolescents. Journal
of Behavioral Medicine 5(1 I. l-8.1982.

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Conse-
quences of Smoking: Cancer. A Report of the Surgeon General. U.S. Department of
Health and Human Services, Public Health Service. Office of the Assistant
Secretary for Health, Office on Smoking and Health. DHHS Publication No.
(PHS)82-50179,1982,302 pp.

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES. The Health Conse-
quences of Smoking: Cardioaascular Disease. A Report of the Surgeon General. U.S.
Department of Health and Human Services, Public Health Service, Office of the
Assistant Secretary for Health, Office on Smoking and Health. DHHS Publication
No. (PHS)84-50204,1983,388 pp.

VARTIAINEN, E., PALLONEN. U., McALISTER, A., KOSKELA, K., PUSKA, P.
Effect of two years of educational intervention on adolescent smoking (The North
Karelia Youth Project). Bulletin of the World Health Organization 61(31: 529-532.
  1983.

WAKE, R., McALISTER, A.. NOSTBAKKEN. D. A Manual on Smoking and Children
Technical Report Series, Volume 73. Geneva, International Union Against Cancer,
  1982.

WALLACK, L.M. Maas media campaigns: The odds against finding behavior change.
Health Education Quarterlv 8(3): 209-259, Fall 1981.
WARNER, K.E. The effects of the antismoking campaign on cigarette consumption.
American Journal ofPublic Health 67(7): 645-650, July 1977.
WARNER, K.E., MURT, H.A. Impact of the anti-smoking campaign on smoking
prevalence: A cohort analysis. Journal of Public Health Policy 3(4): 374-390.
December 1982.

WERKO, L. Prevention of heart attacks. A multifactoral preventive trial in
Gothenberg, Sweden. Annals of Clinical Research 1112): 71-79, April 1979.
WEST, D.W., GRAHAM, S., SWANSON, M., WILKINSON, G. Five year follow-up of a
smoking withdrawal clinic population. American Journal of Public Health 67~61:
536-544, June 1977.

WILHELMSEN, L. Risk factors for disease according to population studies in
ateborg, Sweden. Paper presented at Skandia International Symposia, September
2%25, 1980. In. Bostrom, H.. Ljungstedt. H. IEditors1. Medical Aspects of Mortaltt?
Statutics. Stockholm, Sweden, Almquist and Wiksell International. 1981, pp. 7%
  88.

533


WILHELMSEN, L.. TIBBLIN, G., WERKO, L. A primary preventive study in
Gothenburg, Sweden. Preuentiue Medicine lfl-2): 15%160, March 1972.
WORDEN, J.K., FLYNN, B.S., BRISSON, SF., McCAULIFFE, T.L. Adult Communi-
cation Skills Training to Preuent Adolescent Smoking. Paper presented at the
Annual Meeting of the American Public Health Association, Dallas, November 15,
  1983.

WORLD HEALTH ORGANIZATION. Vital Statistics and Causes of Death. 1972,
Volume 1. Geneva, 1975.

WORLD HEALTH ORGANIZATION EUROPEAN COLLABORATIVE GROUP. An
international controlled trial in the multifactorial prevention of coronary heart
disease. International Journal of Epidemiology X3): 219-224, September 1974.
WYNDER, E.L., HOFFMANN, D. Tobacco and health: A societal challenge. New
England Journal of Medicine 3OOf16): 894-903,1979.

534