Document 5035

The quality of this microfiche is heavily dependent upon the
qiairty of the wigtrwl thesis submitted for microfihning. Every
efforthasbeenmadetoecrsuretheh~quaktyoBreproductiort possible
S'ii manque des pages. veuiliez communiquer avec I'univer-
Some pages may have indistinct print specally if the ongaral
pages were typed with a p a x typewritef ribbon or ifthe Imiversity sent us an mferior photocopy.
Camdim CoWrigM Act. R.S.C. 1970, c. G30. Piease read
. .
m which accompany this thesis.
theauthorrtatKx\ f
i n a Preventive Hedicine Clinic
B.A.. Sinwul Fraser ttntvergity, 1979
i n the Departaent
@ Jacqttel ine Douglas 1983
June 1983
All rights reserved. This thesis nay mot be
reproduced i n whole or i n part, by photocopy
or other rtreans, w3thout p e m l s s b n o f the author.
Jacquel ine &Anne
Master o f Arts
T i t l e o f thesis:
Program Research:
Attitudes and Syraptom
Reduction i n a Preventi-ve Medicine C l i n i c
Examining Conwittee:
Senior Supervisor
External Examiner
V i s i ti ~g Professor
kpartrrPnt of Psychology
Simn Fraser U n f m s i t y
Oate ,Approved:
Septwhar 26, 1983
I hereby grant to S i m Fraser Unf v e ~ stiy the right to I-d
y thesis or d t r g e r t r t f m (the t i t l e of which i s shawtr bdw) to users
of the Simon Fraser Unlverslty Library, and t o make p a r t i a l o r singlap
copies only f o r susir users o r i n response t o a request fras the l i b r a r y
m u l t i p l e copying o f t h i s thesis f o r scholarly purposes may be granted
by me o r the Dean o f Graduate Studies.
I t I s understood t h a t copying
publication o f t h i s thesis f o r f i n a n c i a l gain s h a l l not be allowed
thout my w r i t t e n permission.
T i t l e o f Thesls/Dissertation:
entive P
wC l f n f c
Jacauel f ne 3 0 - h e h & l as
Tkrapeutfc prograns can supply [email protected](cbrsw i t h scarce ct5nfcaT
populations , large sl.pTes and long-term &st4 ng opportunities.
i t can be d i f f i c u l t to acammhte both sxprhmtat rigow& the
advantage o f the above fact~?%within a Preventive M i c i n e program
Syarptorn reduction and attitudes toward personal responsfbility f o r health
were invkstigated:
233 61ients who had completed a Cornel 1 Hedical Index upow entering
such a program were retested a f t e r 2 years o r .more. A significant
decrease i n symptomatology was i d e n t i f i e d
related t o sex or nt,@ey;of
In- subject
visits for
which was not
the sample,
subsample, scores on the Krantz HeaftnrOplnion
Survey and the Wallston Health Locus o f Control f o r Health were not
- --
related t o symptom reduction, nor were age, sex arereducation. Stnoking
and alcohol consumption, while also not related t o symptom reductkm,
should be retested since base measures w
ep very lay.
selection bias i n Ule sample, these results cannot be generalized t o the
overall population..
shrinkage, too few measuring f n s t ~ t s poorly
validated measures, lack
o f invol v g e n t with the program and i t s staff, no c o n t r o f gmup and other
genralizability. m i l e i t i s iqortant for resbrchelIs to regain
flexible in Uleir approach to &fmg program evallatfms and o
resea-tzh i n cl inicat settings, 4n this case the obtained results and
efforts expended i n obtaini-ng these data.
.............................................. *..*..*'*.*.w
...................r,a..................f ..................6
Responsibility as an Issue i n ~ & ! b t i o n . . ....................I*...'13
The Present Study ...............................................
The Setting ......................................................... 19
Applicability of the Settfng
Data Set B......
The Process?..........................................................54
The OutAttl-
Phase..................................................... 58
and Smta
.'List o f Tables
Table 1:
Wonnative Data f o r the Krantz Health Optnfon Survey........32
Table 2: Wonetlve bata f o r the ~ a l k t o nHealth Locus of ControT..
. Table 3:
fable 4:
Table 6:
Table 7:
Table 8:
Table 9:
Table 10:
Analysis o f Variance Betneen Drop-outs and Opm Files..
Logistic Regressfen f o r Group bbership..
R e l i a b i l i t i e s f o r CHI: Data Set A.. ...................... ..'M
... "
Stepwise h l t i p l e &gresTm to
Descrfptivc Statistics f o r Data Set 8.1. ................. ..49
Hagnitude o f E f f e c t f o r Attitude and Demographics.. ........48
Descriptive Statistics f o r Data Set A,.
Table 11: Testing kam for Data Set 0: CMI Scores, Age & Education..%)
Table 12: 7-tests Between Hales and Females f o r A11 kasures:
............,. .................................51
Data Set
Table 13: R e l a t h s h i p BeBeken
Sex and Group W e n h i p .
Figure 1: Details o f Saple Shrinkage for Data Set A
8.. ............. -67
Table A:
Praportion o f Sarrgle Decrease: Data Set A.,
Table 0:
e l c Dcnslse: Data &t
Data Set 8..66
ABpndix B *
Table C:
Table 8:
Table E:
............ ..A8
F r s q u e n d a for OII Sawrs: Data kt .B :
.,. ........ ,
Frequencies f o r FkC and flDS xores: wtta St B ...........
Freqwncies for Cnf Scores: Data Set A,.
freqtm€ly have d i f f i c u l t y @kining access t o c l i n i c a l populations,
s h l d be based
researched principles and techniques.
a s s u p t i o k underlying a proposed treatment plan can then be recognized.
researdr can be sooght
tk support
o i refute these assusptions, and
necessary slodificaiions can be made before the c l i e n t i s treated.
# i s way. needless.constmptim of the a d possibly himfu'l aspects of
the proposed intervention can be Prinirfzed or eliminated,
In m n i n g a set o f &elmtiours to identify areas where changes
- - t a n ~ ~ ~ t t ~ ~ ~ u a r p ~ ~ O t f e n I r a acausal
e a D links
o u r or
physfological and'psychosocial factors.
within a behavioural systew provides the =st j u s t i f i a b l e basis'for
planning effective interventions.
These behavioural systg?r are
usually colaplex, and clinicians are expected t o &sign effective treatrrents i n the presence o f m t t i p l e , interactive factors.
cases, the intervention
service delivery.
I n sbsre
be integrated into a larger program of
I n order t o maximize tke expectation that
interventions w i l l be sucessful, many c l i n i c f a m study oledJcal and
psychological 1iterature i n search o f infomation that can k directly
appl f e d t o a trea-t
A coaprehensi w working knarledge of the w o u s system o r of
behavioural research soaretimes yields a useful intervention strategy.
b u t very often i t does not.
Why is i t so caAlraan for t h e b e s t
e x p e r i m t a l minds t o spend t k i r working lives researchingljn a
particutar area of befwiourih $nvest+gation, yet their many
discoveries are of limited use t o the clinician? ' I t may be that the
clinician i s often dealing w i t h behaviours and disorders on a very
- -
-- -
- - ----
- -
-- - -
In the
different level than .S s usually investigated by researchers.
interest of attain ng experiwental rigour, experimentalists often must
narrow the question o r field of investigation, while practicing
clinicians are usually concerned w i t h a much broader segbent of a
physiological or psychosocial system.
or acute strebs, defined somewhat loosely, here as, 'a generai concept
a l l circmstances, good and bad, that require bodily
adaptation by the autonomic and endocrine systems," (Buck, 1973, pg.30),
behave very differently than under normal conditions, or severe organic
If this is sa, i t i s not unreasonable t o suggest that
reseakh based on well functioning people or on those suffering from
serious organic i n s u l t may not be o"fch
- - ----
help i n treating the large
group of cl tents seen by clinical psychologists today for stress induced
motivational system disorders.
Animal research provides an understanding of the nervous system
- ---
and the physiological aspects of such disorders as anorexia neyosa
(Bemis, 19781, obesity (Crisp, 1978) and other disorders (Bloom, Segal
& Gqi 1lemin, 1976).
Althqugh other aspects of behaviour disorders ar&
studied i n animals a s uetl, i t can be d i f f i c u l t t o relate these t o a
therapeutic situation with human clients.
Clinical reseiirch w i t h humans has provided a large b d y of
knowledge about t h e aetiology, s y & t a ~ t o l oand
~ ~ treatment of
disorders (Bemis , 1978; Crisp & Stonehi 11, 1971 ; Jenkins, 1979;
& Brady, 1979; Selye, 1956; Vigersky, 1977; Yakelihg,
- -
- -
- pp -
-- - --
DeSouza b Beardwood, 1977). However, only some of the conditions i n
an actual intervention setting are reproduceable i n the laboratory, and
i t can be d i f f i c u l t t o generalize laboratory results t o operating
programs of intervention.
Much of the research done w i t h i n existing
programs is d i f f i c u l t to apply, as we1 l . Frequently, the need for
experimental rigour + n & & h e r - - ~ t ~ i considera
g ~ l
tktr~-(e.g.~~------time and monetary constraints) determine the nature of the research
T h i s can result i n a restricted study or a change i n the
In t h i s situation ,/the type of treatment actually
quekions asked.
investigated bears l i t t l e resemblance t o that received by the
program's clients.
T h i s leaves an important gap between research and
Intervention teEhniques must be studied under tightly control led
- -
However, there i s also a need to investigate both t h e
- -
nature and the outcome of t r e a t m n t as i t actually occurs.
Related t o t h i s , are the issues o f cooperation w i t h , program
- ----
managers, and gaining access t o long-term data.
- -
- - ---
I t has been suggested
(Cowen, 1978: Hackler, 1979) that program evaluation and p g r a m
related thegretical research should be done non-intrusively w i t h i n
existing programs.
Cwen states that the effectiveness of com&i ty
service programs w i 11 ultimately be assessed by combining results
f m many poorly designed studies. This is because a number o f
obstacles t o doing well designed outcome,research arise from
differences i n values and objectives between evaluators and program
Hackler recomnds t h a t traditional scientific procedures be
reversed, so that questions are limited by the data a t hand.
than beginning with theories and designing research to t e s t them, he
suggests that researchers use available data to evaluate programs. He
further suggests g i v i n g aid to program staff and government departments
i n csllecting a large body of data, as a f i r s t step.
Rigorous methods
and complex issues could then be slowly apprlo_ached, as mesearchers
and program directors become more aware of one another's needs and
m r e trusting of one another's intentions.
Hadkler's views have been criticized on a number of points
(Corrado, 1981). Hackler suggests that val i d experimnts should not
be done ( a t least i n i t i a l l y ) , because they are difficult t o carry
out w i t h i n politically and organizationally sensitive programs
and because their frequently negative results could exacerbate the
situation by causing marginally useful programs to be cancelled.
He further concl ~ d & - ~ t h a t t h e d i f f i c ' u ~imp1
* f ementati on and the
frequency o f negative results (Corrado 1abels this the "nothing
works" position) together contribute to the scarcity of valid
- -
experiments i n t h e s e mttings.
Corrado suggests that Hackler's
conclusions are both contradictory and based on invalid assumptions.
He ,offers alternative explanations for the paucity of experiments
and for the frequency of negative results.
Corrado also gives
examples of interpretation errors i n negative evaluation reports and
points out that experimental studies have found some programs either
partially o r who1 l y effective.
These criticisms severely weaken the
val i d i t y of Hacl ker ' s recomnendations . Whether o r not these
recomnendations are justifiable on the basfs gf his evidence, some of
the issqes raised shou*
be investigated.
I t i s important to deal
w i t h the issue of doing valid research w i t h i n the program setting.
There i s also a need for research that can be directly applied to
the design of intervention plans.
I n the foregoing discussion, hJo problem areas haw t e n
First, i t i s dften difficult for practicing clinicians t o
use of experimental findings.
Second, i t has been suggested
hat i t would be well to make, use of existing data, because
professionals are often loathe to have researchers "interfering" w i t h
their clients.
In conjunction w i t h this, there are often large samples
and long-term data available w i t h i n ongoing therapeutic programs that
are not easily accessible to researchers from other sources.
In effect, i t i s suggested (Hackler, 1979) that i n order to be
relevant to the treatment situation, to maximize cooperation from
practitioners, ,and to use existing data well, r e s e a r c e s should be
t a i l o r i n g projects t o f i t established programs and data bases.
It i s
Vurther suggested t h a t t h i s be done i n a non-instrusivi manner f o r
both program eval uat ion and other types o f research.
Pre'ventive intervention i s one area where these issues can be
Experimental data on the aetiology and e a r l y treatment of
disorders are p a r t i c u l a r l y relevant t o t h i s type o f intervention.
Preventive treatment pfans generally deal with comparatively large
numbers of pe
e, making programs based on t h i s model p o t e n t i a l l y
useful as a way t o spread the expertise o f a few professionais m n g
many recipients.
Some aspects o f the intervention can be allocated t o
non-professionajs o r t o the c l i e n t .
Preventive programs can'reduce the
r i s k o f severe problems developing, while encouraging c l i e n t s t o
p a r t i c i p a t e i n t h e i r own health care.
I n the mental health f i e l d , three levels o f prevention have been
defined (Caplan, 1964) according t o the point a t whfch intervention
occurs and the type of target population.
Primary prevention occurs
n the process of disorder development.
This i s l e a s t closely
related t o the work tradf t i o n a l l y done by mental health specialists i n
t h e i r efforts t o eliminate disorders.
This type o f prevention
involves intervening before sm&
occur so t h a t a disorder i s
prevented f m i3evelopTngF Giiiry prevention-hasbeen defined 'as a
comuni ty concept which involves lowering the r a t e of new cases in a
poputatlon over time.
Its takget i s not the individual.
I t s aim
rather, is to reduce the risk of developing 'illness f o r a whole
population so that'feuer people w i l l become ill.
Secondary prevention seeks to lower the incidence o f a particular
disorder w i t h i n an "at rjsknkpopulation. T h i s is usually attained by
treatment of existing cases to
sen their s e w r t t y and duration.
the relevant factor i n secondary prevention
identified cases.
is decreasing
Specificity is important here i n terns of disorder,
so #I& pr&[email protected] m idestiiie8 ea4+hR8eidkzdsxH&ee
epidemic or more severe w i t h i n the target group.
Tertiary prevention can be Wewed as primarily a patch-up masure.
Here, intervention occurs well after a disorder has developed, and its
focus i n on the individual.
Cosmunity wide rehabilitation is attempted
by individual treatment. T h j s type of intervention i s closest t o the
- -
- -
practises of most comglnity mental health and medical treatment centres,
i n that problems are seldom attacked until they areAwelt'entbriched.
A1 though Cap1an ' s three level &fin4 tion of prevention was
developed explicitly t o f i t a c m u n i t y wide concept o f prevention
which is distinct from the conventional psychiatric practise of cme-onL
one intervention, its concept of levels can also be applied to
individual treatment.
, el *hate
fx t o r s
In t h i s way, lifestyle changes designed to
whi ctr m t d ' f e a ~ d 3 s e a s ~ c z 1 n 3-ifc:*das
of i
development of a syndrome can be seen as secondary prevention, and
forsts of d i r e c t
e# well es*-l isMdiscmks em
be viewed as tertiary t r e a m t on an individual level.
For the
purpose of prevention, only primary and secondary intervention are of
genuine interest. Tertiary intervention can be viewed as preventing
disease progression o r preventing mortality.
However, these are not
geneqal l y the goals of prevention programs,
Preventive Medicine
I n p r e k t i ve medicine programs, there is an interface between
psychology and medicine.
In s t r i v i n g t o prevent the development of
degenerative disease, physical complaints a p addressed before they
- - -
- --
-- - -
occur. Thi s i s general I y acconrpl i shed through attempts sto reduce risk
Some of the risk factors involved, such as hypertension and
obesity , are of interest t o physicians as predi sposers t o degenerative
disease, and they are widely studied by psycho1 ogi s t s , as we1 1. These
complaints are particularly intiresting to those who study behavioural
laedicine (Pomerleau 4 Brad-y, 19791,
- - -
In the past, programs of prevention have been aimed primarily a t
reducing risk factors w i t h i n a fairly large segment of the population.
Three related methods of intervention have been widely used t o
counteract the spread of degenerative disease.
Personal heal
t h e
services such as uaccination, educational measures Bnd envi rotmental
action have a l l been found useful t o various degrees.
A t present,
several large scale intervention t r i a l s are underway i n Europe 'and the
- -
United States (Breslow, 1978; Hall, R o b h s 8 Gesner, 1972; Williams &
Arnold, 1977). The goals of these programs are to t e s t the feasibility
of reducing risk factors and, ultimately, to reduce disease.
Multiple R i s k Factor Intervention Trials (WFIT) used group therapy t o
- -
f a c i l i t a t e l i f e s t y l e changes
in 12.866 people a t r i s k o f coronary heart
These meastires wwe specifically directed a t n u t r i t i o n ,
Those who readied a c r i t e r i o n r i s k factor
smoking and hypertens-ion.
reduction were subsequently placed on iba'intenance program (Breslow.
A more cost effective pethod was used by the f k n f o r d University .
Heart Disease Prevention Program.
I n t h i s study, two C a l i f o r n i a
u m n u n h e s were the target of extensive mass-media campaigns over ,a
two year period.
I n one
of these,JndiniduaJ c o u ~ e l l ~ w a
_alsas -
-- -
provided for a small subsample with a t h i r d comnunity used as a
comparison group.
The treatment conrnunit i e s received information via
television,' radio, billboards, posters -and the mail.
They showed a
sustained decrease i n r i s k factors over two years, whereas the
"untargeted" comparison commnity increased r i s k f a c t o r levels.
Evaluators concTuded t h i t these &thod< can-perFudip
ti ch%je
t h e i r habits and effectively decrease the r i s k o f heart disease a t
a reasonable cost i n dollars (Breslow, 1979).
program were not reported by Breslaw.
Actual costs o f bhe
However, a cursory comparison
of the Stanford and MRFIT programs suggests that the Stanford technique,
emphasizing large scale advertising, would be less costly on a
per capita basis than the
HRFIT which offered more personal attention.
disease, especially among middle aged males. The goals
o f t h i s project were t o lessen r i s k factors and t o provide tested f i e l d
nethods f o r nationwide use i n the control of general _heal+h problems,
designed t o run over six years,
I t employed
of techniques adtniniStered by individuals or institutions.
wethods included W i a campaigns, environmental changes such as
restriction of public smoking, training health care personnel, and
providing health information t o the public.
Findings a t the end o f
four and one ha1f years indicated enthusiastic pub1 ic cgperation,
decreased smoking among middle aged males, increased use of l& f a t
m i 1k , and- decreases i n blood pressure.
A decl ine i n the incidence of
Several other programs are s t i 11 i n progress. The American Health
Foundatiws "Know Your Body" programs amng New York school children
i s monjtoring and trying t o decrease raised blood cholesterol, high
blood pressure, smoking, obesit; and other risk factors ( ~ iliains
Arnold , 1977) , There is a1so a program i n Swi tzerl and s e t up to en1i s t
comnunfty resources i n .the reductron o f c a r d i o E u 1a r - r l s k f G t o i F i n t h e
slow, 1978), and the Health Hazard Apprai sal Program
which was developed a t the Indianapolis Methodist Hospital is used
throughout, North America t o evaluate indi v i dual risk 1eve1 (Hal l ,
Robbins & Gesner, 1972). Several of these progr*ms (for example the
Stanford and North Karelia projects) concentrate their efforts toward
.envirdnmental change on comnuni ty-wide education.
a more individual level
approach prevention on
Others, such as MRFIT,
There i s already some evidence that t h e Stanford and North Karelia
projects have had a t least some effect on cardiovascular risk factors,
a1though some of the necessary control comparisons have
not -been made.
programs is the possibility of confounding by a general trend toward
lowered incidence of such risk factors as smoking and hypertension i n
control groups.
T h i s effect has been attributed t o a general increase
i n health awareness over the past decade (Breslow, 1979).
I t may be
difficult to separate out haw much of t h i s awareness is related t o t h e
existence of p r m $ f v e pmgritlrts a d government interest in such matters.
The WFIT project has not yet produced clear evaluation results,
available from the various studies w i t h i n this relatively new area of
intervention, the feasibiltty of preventive programs w i 11 be more
easily evaluated.
Evaluation of these programs will provide data
on successful preventive techniques, and could make preventive
channelled into those techniques and programs that have been found most
Ineffective measures could be dropped o r modified.
Furthermore, i f the apparent trend- toward this type of intervention
holds, efforts t o maintain health w i 11 1i kely focus increasingly on risk
Decisions concerning
factors rather than on establ i shed diseases.
governmental support for such a trend w i l l , hopefully, depend on the
results of evaluative studies w i t h i n existing programs.
--- - --
There i s a
- - --- ---- - -- -
clear need for both process and outcow evaluations i n this area toprovide infomation on which to base these decisions.
Evidence is rapidly accunulating that a decrease i n risk factors
-- -
%d smm of i ? l health i s due, largely, t o p a r t i c u l a r aspects o f
l i f e s t y l e change ( W a y 8 Presley, 1975). A c t i v i t y level, nutrition,
I t has
smoking and drug use affect both general health and longeuity.
been stated/l8feslaw.
1979) that health maintenance requires a posltlve
(active) strategy t o prevent disease and extend l i f e .
This could be
attained by developing &a1 ~ f u 1lifestyler, tmpmuing #te
e n v i r o m n t and turning t h e focus o f aedicine t w a r d health maintenance.
The type o f preventive e f f o r t s now being advocated by B r e s l w and
encompasi additional ~ e f f o r t s ~ ~ ~ o s e _ ph1oPsXLWEU
- ---
t r a d i t i o n a l physician's o f f i c e .
People must also take steps. t o protect
t h e i r own health.
Yhile post-illness medical care can usuallyvdeal
w i t h syisptans on a f a t y l y 1~mediate1eve1 , preventive medicine extends
i n t o a l l facets o f l i f e .
This requi,r&
types o f r i s k factors i n h m w ~illness.
cholesterol levels).
poor nutrftion.
systematically attacking two
One i s the body changes which
Another i s personal habits such as smoking and
Breslow suggests t h a t packages o f health care should
be geared t o the specific needs of peopl;
as they age.
Toward t h i s end,
the medical profession has become increasingly interested i n the
practice of preventive medicine.
It has been suggested (~chman, 1979)
t h a t during the 1980's there w i l l be an increase i n b q h the qua1i t y
- --
and quantity o f preyentiye
ite pqular--rt
. IS
. l.Ckeaj(
t h a t private bhysicians w i l l base t h e i r practices on t h i s concept rather
than on the concept of cure.-
-- -
- -
-- -
- -
Respomibili t y f o r decreastng r i s k factors rests, ultimately, w i t h
the individual.
lilany o f the steps cons1dered necessary, such as
decreasing health threatening habits o r removing harmful factors from
one's iainediate e n v i m t , occur through changes i n behaviour.
Persona1 responsibility f o r individual health care i s one aspectof the
larger issue o f personal responsibility f o r one's
orsn well being.
fnvolving the c l i e n t i n his/her own health care has been used t o
aid i n lowerim-risk
factors, a
One type o f intervention which involves both self-involvement and
, .
personal control with respect t o w i n g preventive achon i s
contiflgency or behavioural contracting.
This f o m o f i n w e n t i o n was
introduced by H a r m , Csanyi and Rechs (1969) as a device f o r altering
classroom behaviour, and has since been used i n the treatment o f
a1cohcil isll (Hi 11er. 1972; M i 1le< t&oi
& ~i-. 1974) , drug abuse
(Boudi n, l972), marital problems (3acobson, 1977; Stuart, 1969) ,
sumking reduction ( E l l i o t t & Tighe, 1968; Winnett, W 3 ) , and i n
self-treatment of a number o f problem behaviours (Kanfer, 1980).
Of particular interest, here, i s the effect . o f personal
involvemnt and personal control on therapeutic outcome.
these factors as 'outcome
I n looking a t
ctors, Vani c e l l i (1979) found a negative
controlling alcohol abuse ana aftercare p a r. t l c w . - To # e d e g r e
that cooperation i n Phe form o f voluntary post-treatment contact can 'be
considered .an outcane, mere self-in"fl~8nent i n d e f i n i n g ~ k _ ~ : m t & A
resulted i n a more positive outconle (continued cooperation) for t h i s
S e l f - i n v r r l v ~ ti n treatment 4nf tiation also affects outcome.
,Using attendance rates a t alcohol treatment centres as an outcome
Rosenberg and tif t i k (1976) found that 1nvol untary referrals
were significantly wore littely t o attend than were voluntary referrals.
Another study [Davis & D i m * l96@ f o d no Qffference h-the eutcw!
measure of attendance.
O t h e r interventfon outccli~emeasures also s k d no df fference
mre orl~~~felf-faualuearent,lfc~--*=
as an indicator of self-involvenient, no difference was found i n length
of abstinence during the f i r s t year post-treatnmt (Aharan, 1967).
Simi lar results were found (Voegler, 1976) using absolute alcohol
consuaption as an outcome criterion. However, t h i s finding did not
hold across a11 studies itsing decision to seek help as an indicator of
sel f-invol v k n t .
Wexberg (1953) f w d that sel f-referrals were Mce
a s likely to have &wed alcohol conslraption six mnths a f t e r
treatment began, and were more likely t o experience self-repbrted
"social improvement" than were those coerced i n t o treatment.
I n their revie* of t h i s researchbunha. L Lss (1982).con+luded
that the studies contained several mthodologi cal problems, especially
selection bias.
Dunha and Mass attempted t o clarify the smewht
referral for trea-nt
and .various
okipation, prior t r e a t m t ) , on "stab1e abstinence" after
treatment ceased.
They found t h a t coercion resulted i n s i g n i f i c a n t l y
higher success rates than d i d s e l f r e f e r r a l f o r m s t patient types.
One interesting subgroup label T&
" i n s t i t u t i o n a l l y dependenta, was'
found most successful i n t h i s treatment program. Group members bere
referred by agencies f o r treatment, and were judged highly dependent
A s i g n i f i c a n t d i f f e m c e i n outcome between these and the
group i d e n t i f i e d as independent ( s e l f referral group), who were 1east
, was
This indicated a personal it y r e l p e d
control i n a t least one phase o f treatment.
Lack o f responsibility
f o r decision making ( t o the degree that thi-s can be considered the
relevant variable i n r e f e r r a l by an external source) seems t o r e s u l t in
a more positive outcame f o r dependent individuals.
The opposite effect
i s fotmd f o r independent personality types w i t h i n the alcohol abusing
Along w i t h t h e i r effects on c l i e n t cooperation and on other outcome
measures, personality variables are related t o a variety o f physical
Sow personality factors that have been shown t b r e l a t e t o
physical symptomtology are self concept ( G o t t m n & Lewis, 1982).
state-trai t anxiety (Auerbach , 1973; Spiel berg, 1973) and assertiveness
(Keane, Martin, Berber, Yooten, Fleece & Williams, 1982).
whether i t is used as a treatment outcome measure, i t is of interest
i n a study of preventive intervention.
I t is especjal ty applicable
where the decision t o carry out preventive measures is pr$marily. up
t o the individual, as i t is i n a small scale private program.
i t may be important both i n treatment i n i t f ation and i n carrying out
suggested 11fes tyl e changes.
A number of factors have been found to influence a client's use of
Belief Model o f preventive health behaviour.
This mdel "states that
cooperation w i t h preventive regimens can be predicted by susceptibility
t o a particular disease, degree of sever3t y OF disease contracted,
believed efficacy of preventive action, barriers to action such as
physical or financial difficulties etc., and presence of -a cue to action
that enhances the client's awareness of hSs/her feelings about the
condition. This model has been used t o predict participation i n a
number of preventive programs (Ekcker, Kaback, Rosenstock & Ruth, l975),
and iaswnization of various types ( C m i n g s , Jette, Brock & Haefner,
In addition t o health beliefs, intention has also been found
important i n predicting cucperati on w i t h preventive programs.
1967) and tuberculosis screening participation (Uurtel e, Roberts, &
-mz)have &I
A strong&-
o ft a
outcomes other than compliance has also been well supported (Ajzen &
Fishbein, 1977; B)agozri, 1981; ~ i s h b e i n& Ajzen, 1974; Weigel &
In addition) t o its relationship to o u t c m s of various sorts,
attitude has been found t o relate t o personal involvement.
Cacioppo &
(1981) concluded that under conditions o f high
personal re1 evance, attitudes are prinari ly influenced by qua1 i ty of
Under ton
involvement conditions, the source is more
In the preventive health care field, therefore, i t may be
that information i s more important for some client subgroups (e.g.,
those who perceive the issue as one of high personal relevance) ,whi 1e
the actual setting and i t s staff may be m r e salient for others (e.g.,
those who perceive the issue of l i t t l e personal relevance).
If this i s
each group toward changes i n lifestyle.
The Present Study
clearly,- the relationship of aaitudes to health related behaviour
and. its o u t c m s i s not a simp1e one. Personality variables, intention
and compliance are stme of the factors involved i n t h i s complicated
One of the many questions needing further study is the
A well established private preventive medicine clinic provided a
setting within which t h e following issues could be addressed:
Investigation of the process involved i n seeking clear,
useable outcome data frwr a past-hoc, non-intrusive evaluation,
as measured by symptom reduction.
Advantages and problems in such an undertaking.
'Does the
quality of results support the use of these techniques?
Investigation of a psychological question under the above
condj tions.
I s there a relationship between a t t i t u d e toward
health care, feelings of personal control over health, and
degrke of symptom reduction?
Met hod
The Setting,
The Vancouver Preventive Medicine Centre. This program was
established i n response to a perceived need i n Vancouver for a positive
program of health maintenance. Two physicians founded the centre i n
February, 1977 w i t h five doctors now on staff. There has been a number
of newspaper reviews and several other types o f media coverage since
the centre opened. This coverage has a l l been positive, and public
opinion i s clearly i n favour of such a program.
I t i s s t i l l to be
determined, however, whether i t i s effective.
The centre operates on a self-referral basis, w i t h the aim of
identifying and decreasing risktfactors for the development of disease.
The primary intent is t o induce clients to modify harmful aspects of
their lifestyle.
For each patient, risk factors are determined through
interviews, physical examination and appropriate laboratory tests.
Plasma triglycerides, cholesterol, glucose levels and uric acid are
Doctors counsel improved eating, habits i n an attampt t o b r i n g
elevated levels of the above f a c t ~ r st o acceptable levels.
risk analysis i s calculated based on laboratory tests-and various
personal habits, and expressed as a numerical value.
T h i s score is
calcutate$ i n such a way t h a t i t correlates positijely w i t h degree of
risk of coronary problems.
Further to this, a Health Hazard Appraisal
(Mi l s m , 1978) is done. This computer scored, and the client's
"effective age" i s calculated.
Effective age i s an expression of
physical degeneration, and i t o f t e n d i f f e r s markedly from chronological
Where t h i s occurs, information i s given on how t o a t t a i n a more
desireable e f f e c t i v e age.
Since elevated e f f e c t i v e age i s a strong
r i s k f a c t o r f o r a number o f diseases (Milsum, 1 9 n ) , p a t i e n t s are
encouraged t o make concrete e f f o r t s toward decreasing i t
The p a t i e n t ' s s t r e s s l e v e l i s a l s o monitored u s i n g the Social
Readjustment Rating Scale (Holmes & Rahe, 1967)-
Those experiencjng
h i g h scores on t h i s measure are informed t h a t they a r e a t r i s k o f
developing problems i n b o t h p h y s i c a l and psychological health* They
are i n s t r u c t e d i n m e t h d s o f decwasing stress f a c t o r s i n t h e i r
environment, e i t h e r through m o d i f i c a t i o n o f t h e i r own h a b i t s o r o f
s t h e i r environments.
A computerized d i e t a r y analysis reveals s p e c i f i c
n u t r i t i o n a l d e f i c i t s and i n s t r u c t i o n s on improved d i e t are given where
I n a d d i t i o n t o a l l o f t h i s , a personal data sheet i s kept
t o m n i t o r each p a t i e n t ' s progress.
During the i n i t i a l interview,
past and present symptoms are moqQored using t h e Cornell Medical
Index Health Questionnaire (Brodman , Erdmann , Lorge, & Wol ff , 1949).
T h i s information i s used i n i n i t i a l diagnosis, as w e l l as t o manitor '
general l e v e l s o f health.
Physicians a t t h e centre were very cooperative i n p r o v i d i n g data
and access t o a l a r g e subject pool f o r the study.
As a r e s u l t o f the
apparent success of t h e i r efforts, they were a n x h s to kave,tk+r
program eval uatgd,
H m e e r , since these doctors feel [email protected] that-
p a t i e n t s should not be inconvenienced and since they do n o t have time
o r funds t o c o n t r i b u t e t o t h e research, the s e t t i n g presents a
particularly challenging situation for doing 'Vesearch.
established treatment programs frequently present similar situations,
i t would be interesting to explore the possibilities offere& by this
There have been few evaluative studies of preventive medicine
of any description.
Of the several government sponsored
health care programs now running, only a few have been i n
enough to yield any evaluation data.
Furthermore, there have been no
evaluation studies found on preventive medicine i n private practice.
As the medical profession's interest i n this aspect of medicine
i t s comnitmnt of time and resources are increasing (Geyman, 1979;
Sloane, 1979), there i s a real need for evaluation of preventive
medicine i n family practice.
Applicability of the Setting. The issue of personal responsibility for
health care i s also of primary importance in this p-ogra&t, because
preventive lifestyle changes are ultimately l e f t up t o the patient.
Evidence i s rapidly acc,mulating t o support the idea that eating,
smoking, exercise and other lifestyle changes result i n decreased risk
factors for disease.
Activity level , nutrition, smoking and drug use
affect both general health and longevity, and these can only be
control led, barring legislative changes, by the individual (Breslow,
1978; HcCamy & Presley, 1975). ~es'pmsibilityfor decreasing risk
factors rests, finally , w i t h the individual s i n c e r i s k factor reduction
i s accompljshed through behaviaur change. Researchers i? the area of
psychosomatics have long been aware of the intimate connection between
psychosocia 1 and physi olli ml factmx i n t
Attitudes W a r d personal responsibility for health care is one
psychosocial variable that may affect both compliance w i t h treatment and
sel f-referral
These actfons, i n t u r n , may affect symptom reduction.
T h i s setting provides the opportunity to look a t the relationship
between individual attitudes toward health care, feel ings of personal
responsibility for health care, and physical symptoms. Additionally,
provides an opportunity to examine t h e p s i bi 1i t i e s for answering
research questions w i t h i n an ongoing intervention program, usfng'
non-intrusive data collection methods.
Cornel 1 Medical Index Health Qoestionnaire.
The Cornel 1 Medical
Index (CMI.) was chosen as a symptom checklist and as a future
evaluation tool by t h e medical practitioners a t the formation of ihe
Vancouver Preventive Medicine Centre. This i nstrwnt had already been
i n use for three years prior to the beginning of this study.
provided the only useable long-tern pretest data for identifying a
change i n symptomatology for the Centre's client popu'lation.
The WI is a sel f-admini sterid. sex-speci f i c questionnal r e given
to people older than 13 years of age. Administration time i s estimated
a t 10 to 30 minutes for the 125 items, each answered "yes" or *nom.
The questions are informally mrded to make them s u b b l e - f o r
administration t o a varied population, and their seleclicm
on questions generally asked by physicians i n a comprehensive medical
history (Brodman, e t a1 ., 1949).
The questions are clustered according
- -
to particular organ systems, into subgroups.for the following content
areas: A
eyes and ears; B
- resp$ratory;
- cardigvasuclar; :-
- digestive; E - musculoskeletal; F - s k i n ; G - nervous system;
H - genitourinary; I - fatigueability; 3 - frequency of illness;
K - rniscellaneous diseases; L - habits; M - inadequacy; N -. depression;
0 - anxiety; P - sensitivity; Q - anger; R - tendon. Analysis of t e s t
results are often further divided into sections A-L f o r 'various s m t i c
symptoms and M-R for psychological symptoms.
In an attempt tn p w i & coateat ualiditj~.B r o w Erxhm,
Lorge L Uolff (1951) showed 94% agreement w i t h hospital examinations
on general diagnoses and 87% on specific .i 1lness categories.
In later
studies , Brodman , Erdmann , Lorge , Gershenson , and Wo1 f f ( 1952) and
Brodman; van Woerkm, Erdmann, and Goldstein (1959) were abie t o
discriminate between a ntrmber of groups 'using CMI patterns of
In addition t o its usefulness as an indicator of physical i l l
heaith, i t has been used to indicate emotional disturbance.
In this
regEd, the original studies carried out by the developers of the t e s t
were not convincing.
Using a sample of 5,121 medical and surgical
patients f m New York City Hospital , a 526 patient subsample of
Fhese who were diagnosed as neurotic during subsequent examination,
?wl&??Fjt s e k t e d #ew Ye& t.esMeRts,
t - - P l 311
l ;m
certain patterns of response on the CMI could be an indication of
emotional disorder.
BroQRan and his colleagues found that 30 o r m r e
nyes" responses were given by 76% of psychiatric outpatients, 65%.?f
females i n the neurotic group and 52% of males i n the neurotic g m u b
Excluding the 5,121 hospital patients group from.which the neurotic
group was selected, the next highest proportion w i t h a score of
30 or more occurred for females i n the "normal" New York residents.
These proportions were, appakently, compared subjectively w i t h
proportions of these groups who had scores of 10, 20, 40, 50, 60 and
- -
For the category of "20 yes responses",
70 "yes" responses on the CMI.
the differences betweqn the " n o m l u groups and the emotionally
disturbed groups were
30" category.
bt as subjectively striking as
i n the "score of
Only the hospital employees group occurred markedly
less often t h a n the other groups i n the "20 yes response" category
(5%males; 13%females~, However, s h c e no s t a t i s t i c a l 'comparisons
were reported, these scoring categories may not be a valid.
-discriminator amng groups, based on these data.
This group of researchers also reports that emotional disturbance
can be assumed fm answering both 'yes" and "non on 3 or more
questions. o m i t t i n g 6 o r more questions, o r ' from adding 3 or mre
However, this conclusion i s
remarks or question modifications.
based on frequency of occurrence w i t h i n the 5,121 hospital patjents,
- -- -
w i t h no other group used for comparison. Since no comparisons were
reported, the concl usions are unwarranted.
Other researchers have found differences among general medical
patients and neurotics.
and Fry (1962) found that the mean total
--- - - -
- -
of *yesn %sponses f o r two samples of subjects taken f r o . general
practitioners' patient loads were 15.4 and 17.4, while two neurqtic
sampl M a s defined by elevated scores on Eysenck I
had mean totals of 29.8 and 41.3.
neuroti ci sm s c a l d
In another study, Desroches and
Larson (1963) reported total scores of 45.3, 58.0 and 69.7 for
dmicil iary, general medical and psychiatric patients respectively.
These 'subjects were selected from patients attending a veteran's
Scores for a l l groups i n thi4s study were well above those
found by Brwn and Fry t o separate psychologigal disorders from other
-- -
types of complaint.
Reasons for t h i s are not clear.
- -
However, these
f i n d i n g s do bring into question the validity of either the samples used
or of the CMI as a discriminator of psychological disorder.
Further studies have correlated either the total CMI scores or
various subsection scores.with established tests of emotional disorder
(Dudley, 1976; McDonald, 1967; Harks, 1967; V-hese,
1970; Weiss,
1969). A typical criterion 'for emotional disorder has been the 30 or
more "yesn ksponses on the total CMI found by Brocfiwn e t a l . (1952),
a1though others have been identified.
Ryle and Hamil ton (1962). i n
screening &wried couples for neurosis, defined CMI scores o f 0-15 as
normal, 16-30 as intermediate, and m r e than 30 as highly neurotic.
Pond, Ryle, and H a m i l ton (1963) found that mo,m than 16 'yesm
indicate neurosis i n a working class population.
A total score of 30
or more, or a score of 10 or m r e on sections H-R was used as a
criterion for n e u f o t i ~ i s .i n 234 general hospital patients (Johns, 1972).
- --
M d k a ~ - a f m s e a r c breparts have accepted raised CNI ~(3b~res
as an lndicatiod of emtional disturbance, the validity of these
assumptions has been frequently criticized.
Some researchers
(Abramson, 1966; Desroches & Larson, 1963) argue for the importance of
local n o m .
Others (Amoff, Strough I S e p u r , 1956) s t a t e that the
tool for gathering valuable infomation i n a standardized,
usefulness of the CHI is not on a statistical level, but rather as a
and objective manner t o aid clinicians i n medical examination.
The CMI has some face validity as a symptom checkljst and as a
- -
conserver of physician hours, b u t it i s not feasible as a quantT=ve
measure of psychological adjustment.
-- -
- -
Levitt (1974) suggests that when
only a total score is r e q u i r e d there are a number of measures
available which are mre easily and quickly acbnini stered.
when used as a symptom checklist the CnI can provide jnfomtion about
1 - 5
patients' total health i n a manner @at i s easily accessed b ~ b u s y
medical practitioners.
b s t objections t o the MI as a valid
quantitative masure of psychological adjustment seem t o be, based on
the simplicity of design form, subjective basis of original f ters
selection and lack of normative data.
With regard t o its lack of
norarative data, most research has been concerned w i t h emotionally
o r chronically physically IF1 patients, or for very select subgroups of
1954; Cal h e c k - b a n
, 1x6;tawton,
1959; Richman, St ade
1966; Stout, Wight & Bruhn, 1969; White, Reznikoff & Evelv, 1958).
Scores are affected by childhood experiences and marital adjustment
(Ryle & Hamilton, 1962), by age and possibly by social class.
Reportedly, the CHI has shuwn l i t t l e predictive validity for specific
physi ca1 disorders, b u t does correlate fairly we1 1 (correlation val ue
unreported) w i t h physf cians ' appraisals of emotional heal t h (Abramson,
Terespolsky, B m k & Kark, 1965). Sex. education and ethnic g~oup
appear to have littTe or no effect on its validjty.
In order t o deal w i t h s a w of the criticisms of the CMI and t o
supply more essential nodtive
data, Seymour (1976) studied 1046 male
that those diagnosed with other measures as neurotic scored higher on
the CMI than those diagnosed normal.
In addition, there were several
patterns of response which distinguished a group of "high responders"
from "1ow responders*.
Overall, Seymour concluded that the' CMI
provided substantial predictive validity i n the context of -health
. - care
for this sample. However, the skewed distribution of responses to
different subsectinns indicated that equal weights cannot be given t o
each response.
He stated that i t is unreasonable to consider seven
"yesu responses on sections J (frequency of illness) , C (cardiovascular),
and D (digestive) t o be equivalent.
suggest that i t
~ h e s eresults and conclusions
not be wise t o consider the number of "yes"
responses, alone, as an indicator of emotional disorder.
be a poor basis fojdistinguiishing b
I t may also
e ~ ~ ~ u ~ w i t h igeneral
i t h e
pojW'Gation. Rather, t h e pattern of response may be more i m p o r t a n t a s
an indicator.
Seymur tested s&&Zs IWce, with a 14-l6 weekTn€erval between
t e s t s , i n an attempt t o provide v i l i d i ty data.
The 041 was administered
i n i t i a l l l , and another ( o m unreported) physician's check1ist was
given after the interval.
Because there was significant correlation
between the scores on these two measures ,'i t was concluded that sane
measure of reliability was given for the CHI. This i s an extremely
Certainly; a clearer
weak argument for reliability of this measure.
picture would have emerged had i t also been administered a f t e r the
14-16 week interval.
tkm k k y dET a
These data indicate more about symptom stability
M t h e elin 4t-ei%itrt'fi€y. H M r , 3 n i
no 6tWr
reliability data have been located for this measure, i t has been
reported here.
Another study (Clum, Flag & Holberg, 1970) investigated the
differential stability of individual CMI items during Marine recruit
While there was a Jack of stability .for historical~itemsi n
the Index, i t was concluded that subjects were under stress and were
therefore, highly atypical.
This, then, was thought to indicate that
the lack of re1 iabil i t y demonstrated on this area of the questionnaire
may not generalize to other situations. However, since no reliability
was demonstrated for parti cular subsections either , the measure does not
appear t o be quantifying a stable t r a i t .
Short interval test-retest reliability i s also lacking i n the WI
f5urjevi ch ., m6). fi a sample of miEPmiTftary-outjiiitients ,-theem
Based on the above studies, i t must be concluded that t h i s measure
i s rat4er wtstab'te over time. However, a symptom checklist i s expected
t h i s purpose.
Generally, t h i s t e s t has gone out of fashion i n recent
years a s a research instrument because of i t s unwieldiness and the
avai l a b i ~ 4 t of
l shorter and l e s s subjectively derived t e s t s .
reviewer strongly questions i t s use ds a psychometric i n s t r u m e n t
( ~ ~ k k e n1972).
Although i t has been shown as a poor measlire o f most
psychological disorders and is not a re1 iable measure of any t r a i t t h a t
can be monitored over time, i t is s t i l l used as a symptom checklist by
physicians i n medical practice.
Since i t apparently has face and
content val i d i t y f o r the purpose of identifying ma1 functioning organ
systems and particular physical disorders, i t s continued use i n t h i s
s e t t i n g is understandable.
However, its use as a research tool i n any
b u t the broadest sense is indefensible.
The only parameter t h a t this
i n s t r u m e n t can be assumed t o indicate w i t h any measure of certainty is
number of symptoms of i l l health being experienced a t time of
For the purpose of this study, the scores reported f o r subjects o
the CMI will be interpreted only as number of reported symptoms. I t
cannot be assumed t h a t a decrease i n overall symptom number-actually
indicates improved health.
That kind of interpretation would require
careful scrutiny of both number and type of item by a qualified medical
The "meaning" of the responses given i n this research
w i l l , ther fore, not be sought.
Since t h i s i n s t r u m e n t has been used by the Preventive Medicine
- -
-- - -
Centre a t patient intake interviews from the program's i v e p t i o n , i t
o f f e r s a useful wuge of increase o r decrease i n nmber of symptom of
i l l health before and a f t e r treatment a t the Centre.
The change i n
quantity of symptoms is considered t o be a useful measure of overall
patient discomfort (Brodman e t a l . , 1951). Where the change is large,
i t may even indicate a change of s t a t e from psychological disturbance
t o an undisturbed s t a t e ( f o r example, a change fm 30 "yes" responses
t o 5 "yes" responses, overall).
Therefore, the CMI will be considered
but indicative of general di scomfort.
Krantz Health Opinion Survey. T h i s t e s t was developed and validated
i n 1980 as an attempt t o provide a measure of individual a t t i t u d e s
toward d i f ferent treatment approaches ( Krantz , Baum & Wi11i ams , 1980).
Since previous research-suggested t h a t patients prefer various Fevets
of active participation i n t h e i r own health care, Krantz and his
colleagues reasoned t h a t a 'valid measure of patient
related t o treatment outcomes.
would be
Scales were constructed t o measure two
important aspects of self-involvement.
These a r e acceptance of
information offered by medical practitioners and active involvement
i n s e l f care.
The sixteen item t e s t was developed from a pool of 40 original
questions, using a 359 subject sample.
involve men^
Factor analysis yielded a
- --
- -
Subscale (11 items) and an Information Subscale
( 7 items).=' High scores represent positive attitudes toward both s e l f
directed and informed treatment.
Items i n both subscales r e f e r t o
routine aspects -of health; care, rather severe i l l n e s s .
this reason, i t was d e e d particularly suitable f o r use i n t h i s study.
since preventive health care as practiced a t the Vancouver- Preventive
Medicine Centre deals primarily with routine health improvement
measures and changes i n l i f e s t y l e , rather than w i t h i l l n e s s , per se.
The t o t a l Krantz Heilth Opinion Survey (HOS) showed a Kuder\
Richardson Re1 iabi l i t y of .77, .74 and "above .74" when admi n n t e r e d
t o three separate college student samples.
Test-retest re1 iabi t i t y was
Validity of t h i s t e s t was established through administration t o
unselected residents of a college dormitory, students reporting t o a
college infirmary f o r minor i l l n e s s and students enrolled i n a medical
The three samples which represented extremes i n
preference for different types of treatment scored as predicted, -and
the t e s t s discriminated between these groups.
Norms a r e provided i n
Table 1, The t e s t correlated .31 w i t h the Wallston Health Locus of
Control Scale, which measures expectancies of a b i l i t y t o control one's
orm health.
I t showed "only modest" correlation w i t h repression-
sensitization (value unreported), and low or near zero correlation with
the Crowne-Harlowe Social Desirability Scale and w i t h the .Hypochondriasis
Scale of the MPI . Ov~ra'll, the t e s t was found t o be a r e l i a b l e , val id
measure of patient a t t i t u d e s toward self-involvement i n health care.
- -
- 7 4 f o r the t o t a l HOS Scale over a seven week period.
self-help course.
- -
Table 1
Normative Data f o r the Krantz Health Opinion Survey
Mean Score
College Dormitory Residents
Users of Col lege f nfi rmary
Students i n Self-help Class
a Standard Deviation.
Wallston Health Locus of Control Scale. This eleven item s c a l e was
developed as an area-speci f i c measure of expectancies regarding 1ocus
of control in health issues (Wallston, Kaplan & Maides, 1976).
I t has
been shown t o p;ovide discriminant validity from R o t t e r ' s (1966)
Locus o f Control Scale.
I t was a l s o found t o ' b e almost t o t a l l y f r e e of
social d e s i r a b i l i t y b i a s , a s shown by a correlation of -0.01 with the
Crowne-Marlowe scale.
Furthermore, "internals" a s measured by t h i s
instrument who value health highly sought more health related
.information than other groups.
Weight control patients reported
k i n g more s a t i s f i e d w i t h the program when i t was consistent w i t h t h e i r
Health Locus of Control type.
Norms a r e -provided f o r primari 1y black, hypertensive outpatients,
college dormitory residents and older college students who do not
1ive on campus (see Tab1 e 2) .
_Table 2
Nonnative Data f o r the Wallston Health Locus o f Control Scale
Mean Score
College studentsb
College Students
C m u n i ty Res idents
Hypertensive Outpatients
a standard Deviation.
Two separate groups o f College students were tested.
Data Set A.
T e s t - r e t e s t data were obtained from 233 of the 729
p a t i e n t s who attended the Vancouver Preventive Medicine Centre, 1743
West 10th Ave.,
Vancouver, B.C.,
between August, 1977 and August 1979.
The primary data source was p a t i e n t f i l e s kept by the Centre, f o r
c l i e n t s who had completed a C M I during t h e i r f i r s t v i . s i t .
Some o f the
f i l e s had been closed because p a t i e n t s no longer maintained contact,
w h i l e others were s t i l l open.
For the open f i l e s , a c t i v e involvement
Of t h e 729 p a t i e n t s ' f i l e s f o r which a C M I had been
administered a t the i n t a k e interview, 20.9%,(153 p a t i e n t s ) had not
continued a f t e r the i n i t i a l v i s i t .
This group was l a b e l l e d "drop-outs".
An attempt was made t o obtain r e t e s t data from both drop-outs and
p a t i e n t s w i t h continued contact (sample shrinkage shown i n Appendix A ) .
For t h i s phase o f the i n v e s t i g a t i o n , drop-outs were l a b e l l e d
" c o n t r o l group" and those who had v i s i t e d t h e centre a t l e a s t twice
were the experimental group.
t h e Centre a number o f times.
Most o f t h e experimental group had v i s i t e d
Howewr, actual number o f v i s i t s was
n o t recorded on most o f these a c t i v e f i l e s .
These two groups were
separated i n an attempt t o i d e n t i f y a treatment versus a non-treatment
These two groups were subsequently combined under t h e l a b e l
"responders" (N=231) f o r f u r t h e r analyses.
were dropped.
A t t h i s p o i n t , two male -
One-had completed t h e female version of Section H, and
t h e other had completed both male and female v e r s i o ~ so f t h i s section.
A second group o f p a t i e n t s were selected who had been given
t h e CHI. during t h e i r i n t a k e interview, b u t who had n o t returned the
second (mai 1ed) q u e s t i o h a i re.
This group o f 236 "non-responders" was
randomly selected from the t o t a l Vancouver Preventive Medicine Centre
p a t i e n t population, excluding those used i n the f i r s t phase o f t h e
The number was chosen t o be approximately equal t o
t h e responder sample.
This group, along w i t h 229 o f the responder
sample comprised Data Set A.
Two more o f the responder group were
dropped a f t e r r e l i a b i l i t i e s f o r t h e CMI had been calculated.
These two
subjects were' found t o be out1 i e r s w i t h i u a .subsample of ;this gpoup
(Data Set B) and were, therefore, dropped from both samples.
Appendix A
explains sample shrinkage.
:Data Set B. The o r i g i n a l sample f o r t h i s section o f t h e i n v e s t i g a t i o n
consisted of 126 (54%) o f the o r i g i n a l 233 member wresponc&rugroup used
i n the f i r s t phase of the project.
These 126 .subjects represent a l l
of those people i n the responder sample for whom data on smoking,
alcohol consumption, miles driven per year and seat belt use had been
collected a t their original v i s i t t o the Centre.
Since smoking and
alcohol consumption were to be monitored for changes and included i n the
analyses, only the 126 subjects were used.
Testable data were obtained from 60 of these (21 males ,239 females),
representing 47,6% of the original su8sanple.
high scores on the CMI.
Two male s u b k c t s had very
Since these scores were important variables i n
subsequent regression analyses w i t h i n t h i s sample, these two subjects '
results were judged t o unnecessarily bias the sample.
These out1 iers
were, therefore, dropped from b o t h Data Set A and Data Set B.
shrinkage for Data Sets A and B are detailed i n Appendix A.
A1 1 patients who had come t o the Vancouver Preventive Medicine
Centre (WC) and completed a CHI prior t o August, 1979 were sent a
second copy by mail.
Both active and inactive f i l e s
Are used.
covering l e t t e r briefly explai~edthe purpose of the & I , absured the
patient of confidentiality, and made a plea for cooperation.
A s t a y e d sex f-&drsseci e~vefepwas i ~ c f u c k d~ 4 t htke +wstkmrxtire
t o f a c i l i t a t e return rates;
The f i r s t mail in4 consisted of 576
experimental subjects and 153 controls.
In an attgmpt t o separate out
a control group of untreated patients, experirnentals were those who had
continued a f t e r f i r s t contact with the PMC.
had visited the centre only
Controls were those who
Data Set A
This mailout yielded, a f t e r 11 weeks, 197 experimental subject
returns (34%of experimental subject pool), and 33 controls (21.5% of
control subject pool).
A telephone campaign aimed a t increasing return
rate resulted i n five more returns out of 162 recontacts.
point, efforts t o increase response rate ceased.
A t this
Overall, 202
experimental returns and 33 control returhs were received.
experimental subjects were deleted because only half of the questions
had been answered. The final sample comprised 34.7% o f the experimental
subject pool and 21.5% of the control pool. This was a somenhat
disappointing amount of analyzable data.
However, many o f the
non-returns had moved, leaving no forwarding. address.
For those who.
were contacted more than once, i t was decided t h
y e t another
reminder may be i n t e r p r e t e d as i n t r u s i v e and would r e f l e c t badly on t h e
Time and money were a l s o important considerations i n t h e decision
t o proceed w i t h a t o t a l o f 233 subjects.
Each subject was t e s t e d twice,
f o r a t o t a l o f 466 completed C M I 's.
Since one subsection o f t h e questionnaire d i f f e r e d f o r males and
females (H
genitourinary), there were 126 separate questions f o r
females and 124
0 '
f o r males.
used f o r data analysis.
HTS raw data
f i l e s were analyzed using analysis of variance f o r repeated measures
( B W P ~ V )Total
nunber of "yes" responses on the C M I over sex, group
(experimental o r c o n t r o l ) and-iime (intake-Time 1; r e t e s t = T i m 2)
From t h e 496 p a t i e n t s who d i d n o t r e t u r n a useable CMI, 236 were
randomly selected to roughly coincide w i t h the 233 member responder
The 236 subjects were l a b e l l e d nnonresponders".
Two responders
were deleted who had questionable patterns of response on t h e CMI, and
t h e r e s u l t i n g sample of 467 c l i e n t s was used t o t e s t r e l i a b i l i t y o f t h e
SPSS ~obptw~ram
Pearson C o r r e l a t i o n between odd and even items
was used, along w i t h . t h e Spearman-Brown Correction f o r basic odd-even
r e l i a b i 1i t i e s .
Year of b i r t h , sex and education were recorded for a l l 236
Age data were recorded for 203 responders and sex and
education levels for 205 responders.
The resulting combined sample of
responders and non-responders was analyzed using a stepwise logistic
regression (BMDPLR) for "group membership".
From an original pool of
229 responders and 236 non-responders, those subjects w i t h missing data
or CMI scores of 85 or more were deleted for this analysis.
While this
would not have affected the large sample, analyses using the smaller
subsample (Data Set B) would be strongly influenced by outliers. Since
the number of subjects w i t h scores above 85 waj small for b w p l e s
.(See Appendix B)
they were deleted.
This, along w i t h deletions for
missing data, resulted i n a 436 subject group being used for the
logistic regression.
An SPSS Partial Correlation between group membership and CMI Time1 score was done to examine the relationship of these two variables. A
chi-square was also calculated for number of males and females i n each
Data Set B.
The Krantz Health Opinion Survey ( W S ) and the Wallston
Health Locus of Control (HLC) were mailed t o a 126 member subsample o f
t h e 229 subject responder group.
A stamped, self-addressed envelope
was mclosed to f a c i l i t a t e returns, along w i t h a covering l e t t e r
requesting subject cooperation and written permission t o use the data
for research purposes.
- - -
Only 52 of these packets were originally
Telephone contact w i t h 50 o f thi remaining patients yielded
eight more returns, f o r a total of 60 useable subjects (21 male and
39 female).
These 60 subjects composed Data Set B of the study. (See
Appendix A).
Test scores, year'of b i r t h , sex, years of education, smoking
habits (increase, decrease, no change) and alcohol consumption (increase,
Some behaviour
decrease, no change) were recorded for each subject.
change measures were included i n an attempt to monitor compliance w i t h
suggested 1ifestyle changes.
I t was expected that actual behaviour
change would correlate w i t h either ckange i n symp€oma€o?o~or wtth
.locus of control for health.
Smoking habits and alcohol consumption
were chosen because pre-test data were available for these variables
from a subgroup of the original data pool.
Age, sex and education
were chosen as demographic variables because' they tend t o covary w i t h
number of symptoms of i l l health on the CMI (.Abramson e t a1 ., 1953) and
w i t h Locus of Control
con otter, 1966). Originally,
m i les driven per year
and seat be1 t use were alsd to be included i n the analyses.
since no change i n these behaviours over ,the two year period was reported
for .the 60 avai 1able subjects, these variables were d'ropped.
Outliers were deleted, and the resulting subject pool of 58 was
analyzed using a stepwise mu1 tip1 e regression (BPIDP2R). Change score
over a two year period was the dependent variable and CMI Time 1 score,
- -
- - - --
- -
iCtC score, HOS score, age, sex, education level, smoking behaviour, and
aTcoho7 [email protected] as independent variables.
This was donekin order to
d tennine the best predictor(s) of decrease i n symptoms.
For further
cormberation, the data were standardized and another regression was run
-- - --
using CMI Time 2 score as dependent variable and retaining the same
independent variables. In order t o determine the magnitude of the effect, SPSS
Subprogram Regression analyses were done on these data i n standard and
non-standardized uform, u s i n g CMI change score as dependent variable
and two different s e t s o f independent variables.
he f i r s t regression
used CMI rime 1 score, HLC score, HOS score, age,
ex, education level,
smoking behaviour and a1coho1 consumption as independent variables.
seccmd regression used cm'iy CHI Time 1 score as an independenk uarktde,
retaining the same dependent variable.
The difference between these two
R-square values was then taken as a measure of the magnitude of the
" effect of the group of independent variables, excluding CMI Time 1
These data were also.'submitted to t-tests between males and females
w i t h i n the three t e s t s and the demographic variables.
A t - t e s t between
CMI Time 1 and CMI Time 2 was done on the 58 subject sample, as well as
a t - t e s t between the CMI Time 1 and CMI Time 2 scores for t h i s sample and
the 229 member responder group from Data Set A. Titests between groups
were used to detennine whether the 58 subjects were representative. A
chi-square for sex between this group and the 205 responders w i t h this
Data Set A
The f i r s t step i n the i n v e s t i g a t i o n was analysis o f variance
(BMDP2V) between t h e uW) experimental (open f i l e ) and 33 c o n t r o l
(drop-out) subjects gained f r o m t h e CMI mailout.
S i g n i f i c a n t effects
from t h i s analysis were f o r Time o f Administration i n c l u d i n g section H
(genitourinary) and Time of Administration excluding s e c t i o n H (see
Table 3).
fable 3
Analysis o f Variance Between Drop-outs and Open F i l e s
CMI i n c l u d i n g section H
.002 1
&MI excluding s e c t i o ~H
*. 0024
There was no s i g n i f i c a n t d i f f e r e n c e found between males and females
i n t h e number o f "yes" responses, n o r were there s i g n i f i c a n t differences
between those who had v i s i t e d t h e PMC o n l y once (drop-outs) and those
w f t h continued m t a c t (open fi-te).
Further m v e r t i g a t m r r meate&tkt
p r e f e r r e d h e a l t h h a b i t s - a t f i r s t contact.
Since t h i s comprises a l a r g e
p a r t o f t&e prevention aspect o f the Ce~tre'sa c t i v i t i e s , ns real
d t f f e r e n c e i n preventive treatment couTd be j u r t i i a % ? y assumed. 6 i n
t h a t these c l i e n t s a r e a self-selected group'from t h e outset, a l l o f
whom receive a t l e a s t p a r t of the same "treatment",jand
given t h a t the
analysis of variance between groups was n o t s f g n i f i c a n t , these samples
were combined f o r f u r t h e r analyses.
Since both c o n t r o l and experimental subjects were self-selected,
cannot be assumed from these data t h a t t h i s sample i s representative o f
t h e population i n general.
Additionally',' the PMC p a t i e n t sample was
narrowed f u r t h e r by the, f a i 1ure o f many c l i e n t s t o r e t u r n a compl eted
CMI questionnaire.
Therefore, t h e 233 responders also cannot be assumed
t o represent t h e t o t a l c l i e n t population a t the PMC.
However, i f group membership i s n o t r e l a t e d t o o r i g i n a l i n t a k e
scores on the CMI, then i t can be suggested t h a t the people f o r whom ah
improvement over t i h e has been shown (responders) are from t h e same
population as t h e e n t i r e p a t i e n t p o l a t t h e Centre.
A logisttc
regression (BMDPLR) on 436 of these subjects who had scores o f 85 or l e s s
and were n o t missing r e l e v a n t demographics was calculated,
A score o f
85 was chosen as a cut-off p o i n t because i t was w e l l above the highest'
score reported' i n the 1i t e r a t u r e (Desroches & Larson, 1963) as a
representative mean t o t a l f o r - a p a t i e n t population (69.7).
It was
decided t h a t 85-plus could be considered extreme i n t h i s case without
excluding any r e l e v a n t population.
Frequency d i s t r i b u t i o n s o f scores on
t h e CMI, HOS and HLC a r e provided i n Appendix B.
This regression showed t h a t group membership could be predicted by
education, age, sex and the age-by-education i n t e r a c t i o n (see Table 4) ,
but not by CMI Time 1 score: In order t o examine t h e relationship of
group membership and CMI Time 1 scores, alone, an SPSS Partial
Correlation' of these 'two variables was calculated.
education were he1 d constant.
Sex, age and
No significant correlation was found
between original intake CMI score and group membership (r=0.03,
p 0.50).
Both the l o g i s t i c regression and the partial correlation showed that
these groups do not d i f f e r relative t o original scores on the sjmptm
They can, therefore, be considered t o be from t h e s a m
population with respect t o the v y i a b l e i n question.
Table 4
Logistic Regression for Group Membership
Term Entered
Log like1 i b d
Improvement: x2
- 24.756
step #
0 .OOO
Age x Education 1
-.'.: ;
The distribution o f sex . w i t h i n groups was examined,, as well, for
Data Set A,
A chi-square f o r frequency of~malesand .females .showed that
there was a significant relationship between sex and group membership
within t h i s sample (see Table 5 ) .
Given that the responder and
non-responder groups differed harkedlp on number of males and females,
the responder group cannot be assumed t o be t o t a l l y representative of
the type of client who attends the PMC.
The 'logi s t i c 'regression also
indicates that these groups d i f f e r on a group of demographic variables.
However, since group membership cannot be predicted by CMI Time 1
score, responders can be cautiously assumed to be representative on
that dimension.
Descriptive data are provided for responder and
- non-responder grpups i n Table 6.
Table 5
Relationship Between Sex and Group ~embershipb '
Umber o f Females
on- responders
Number of Males
Tota 1
The CMI was tested for r e l i a b i l i t y within the 467 member group who
had been administered the t e s t during the intake interview.
separated by group and by sex.
As can be seen from
These were
7, for a l l
o f the combinations and' groupings tested, re1 i abil i t i e s were above .85.
Data Set 6
from a combination of attitudes and other variables, a Stepwise Multiple
Regressicm (BWDP2R) nas run,
CHI Change score as depertbt'variable.
- - ---
Table 6
Descriptive S t a t i s t i c s f o r Data Set A
Vari ab1 e
Data S e t A :
CHI Time 1
Year of B i r t h
CMI Time 1
CMI Time 2
Year of B i r t h
* ,7
Nonresponders :
CMI Time 1
Year of B i r t h
Educa t i on
Mean age = 41 y e a r s
b ~ e a nage = 43 y e a r s
%an age = 39 y e a r s
Table 7
R e l i a b i l i t i e s f o r CMI: Data Set A
CHI Time (1,2)
Adjusted r
Responder Group:
Fema 1e
Ma1 e
Nonresponder Group:
Ma1 e
Fema l e
a ~ o m p u t e rwas instructed t o i d e n t i f y sex based on section H (genitourinary)
o f the CMI.
Therefore, a l l 231 responder subjects were included i n the
Independent variables were CMI Time 1 score, HOS score, HLC score, sex;
year of bi r t h , years of education, smoking habits (increase, decrease,
no change) and a1 coho1 consumption (increase, decrease, no change).
Best predictor of change in CMI score f r o m intake t o readministration
two or more years l a t e r was found t o be original number of symptoms
(CMI T i m e 1 score).
No other variable included i n the analysis
contributed enough v a r i a b i l i t y t o be entered into the equation.
BMDPZR using standardized input data revealed t h a t the best predictor
of CMI Time 2 score was also CMI Time 1 score (see Table 8 ) .
Table 8
Stepwi se Mu1 t i ple egression to Predj c t CMI Improvement
~ e r mEntered R~
CMI Change CMI Time 1
Non-standard data
St. Error F t o enter
Standard data
CMI Time 2 CMI Time 1
Independent variables were CMI Time 1, WS, HLC, sex, age, education,
smoking and alcohol consumption.
Aside frm the strorig correlation between CMI Time 1 and CMI Time 2
scores (r=0.76), few other interesting correlations were t o be found.
There was a noticeable correlation between age and education within the
samp'le (r=0.32), and between smoking increase and a ' l c o h ~ l n t l r e a s e
(r=0.38), a1 though these are not particularly strong.
For t h i s sample, mean Time 1* score was 23.14 and mean Time 2 score
was 16.3.
It should be noted t h a t these scores are w e l l below thos
c i t e d i n the 1i t e r a t u r e as i n d i c a t i v e of psychoneurosi s (Brodman e t
1952; Ryle 8 Hamil ton, 1962) unless accompanied by a congruent
diagnosi s by a physician (Pond e t a1
., 1963).
D e s c r i p t i v e data f o r 'a1 1
measures w i t h i n t h i s sample a r e shown i n Table 10.
The magnitude o f e f f e a t f o r the independent variables on CMI Change
score was examined.
Two SPSS Subprogram Regression analyses were used,
w i t h CMI Time 2 score as dependent variable.
The f i r s t used CMI Time 1
score p l u s a l l previously mentioned a t t i t u d e and demographic variables
as. independent variables.
A second regression used o n l y CMI Time 1
score as a p r e d i c t o r . (see Table g ).
Subtracting these two R~ val w s
revealed t h a t whqn t h e e f f e c t o f o r i g i n a l CMI score i s taken i n t o
consideration, only approximately 10% of. t h e variance i n f i n a l CMI score
combined (3% w i t h adjusted r2).
i s contributed by a l l other variables:
Table 9
Magnitude o f E-ffect f o r A t t i t u d e and ~emographics*
CMI Time 1 + a l l
Adjusted 'R
C M I Time 2
Q. !xi3
S t . Error
CM1Time lalone
Magni tude :
SPSS Subprogram ~e~ression,i/
Table 10
Descriptive i t a d s t i c s f& Datq S,et B
Krantz HOS
Wallston HLC
'Year o i Birth
CMI Time 1
CMI Time 2
CMI Change
A1 coho1 Increpse
Alcohol Decrease
Smoki ng Increase
Smoking Decrease
- -
%ean age = 44 years
A t - t e s t t o ensure t h a t a significant change had occurred over
time for t h i s sample was significant.
However, no signficant difference
was found 'between this iample and the 229 member respdnder group on CMI
scores, age, or education level. (see Table 11).
here fore,
t h i s small
sample can be assumed t o represent the larger group from uhieh i t was
drawn with respect t o these dimensions,
Table 11
Testing Means f a r Data Set 0: CMI Scores, A*
a d Ecksthn
CMI Time 2 (same group)
0 .OOO
CMI Time 1 (229 responders)
CMI Time 2 with:
CMI Change Score w i t h :,
CMI Chanye (229 responders)
Age w i t h :
Age (229 responders)
Education with:
Education (229 responders)
2- tai 1 probabil i ty
CMI Time 1 with:
CMI Time 2 (229 responders)
Table 12
T-tests Between M a l f s and Femaled for All Measures: Data Set B
HOS Score
HLC Score
CMI Time 1
CHI T i m e 2
CMI Change
Smoking Increase
Sm k ing Decrease
A1 cohol Increase
A1 cohol Decrease
2 - t a i 1 probabi 1 i t y
This sample was a l s o tested f o r dffferences betwee'n males and
females on a l l p a s u r e s .
No s i g n i f i c a n t d i f f e r e n c e was found between
males and females on any o f the variables measured (see Table 12).
A d d i t i o n a l l y , when t h e p r o p o r t i o n of males and females w i t h i n t h i s group
and the 229 member respunder group was compared, no r e l a t i o n s h i p was found
m e r n b e ~ s h i(see
Table 13).
between sex and
Table 13
Re1a t i o n s h i p Between Sex land Graup ~ e m b e r s h i ~ ~
Number o f Females Number o f Males
Data S e t B
Responders ( ~ a t Set
Both the HOS and t h e HLC were tested f o r r e l i a b i l i t y w i t h i n the 58
subject group comprising Data Set B.
SPSS Subprogram Re1i a b i 1it y produced
alpha l e v e l s o f 0.80 f o r the Krantz HOS, and 0.78 Tor t h e Wal i s t o n HLC.
Overall, i t i s apparent t h a t the best p r e d i c t o r o f change f o r t h i s
sel f-sel ected c l i e n t population i s number o f presenting symptoms.
o f the otheP variables monitored i n t h i s . study contributed s i g n i f i c a n t l y
t o t h a t change.
Although equivalent on some measures, the Data Se$ B .
f i n d i n g s cannot be generalized t o the l a r g e r (responder) o w e v e r y v a r i a b l e
measured i n the study.
small group.
The HLC and IXlS findings should be confined t o t h i s
General Considerations
Two o f the o b j e c t i v e s f o r t h i s research were t o provide information
about t h e outcome o f a preventive medicine program, and t o answer a more
t h e o r e t i c a l question concerning the r e l a t i o n s h i p o f h e a l t h care
a t t i t u d e s t o symptom reduction.
Perhaps an even mare important i'ssue i n .
t h i s study was the attempt t o a c c ~ pi sl h these objectives w i t h i n an
ongoing program, using non-intrusive k t h o d s .
Had t h i s method been
successful i n answering t h e questions posed, i t would have l e n t support
t o t h e suggestion o f Xackler (1979) and others t h a t evaluators and o t h e r
researchers must rever'se the research process.
That i s , they must f i t
t h e i r questions t o t h e data and s i t u a t i o n at-hand when dealing w i U l
ongoing programs.
A s i m i l a r p o i n t o i view i s put fo'rward strongTy by
program managers, l i n e workers and government funders.
This argumen? can be very seductive.
I t i s e s p e c i a l l y convincing
f o r those areas o f research where thewprimary i n t e n t i.s t o provide
i n f q n n a t i o n t h a t w i l l be used by p r a c t i c i n g c l i n i c i a n s i n designing
i n t e r v e n t i o n s o r by program planners i n r e s o l v i n g funding issues.
these are t h e main reasons f o r doing research r a t h e r than more s c h o l a r l y
motives, i t can be d i f f i c u l t t o maintain an o b j e c t i v e view.
Somg basic
methodological p r i o r i t i e s must be maintained, however, l e s t the research
process be rendered useless a t t h e program l e v e l .
I n t h e present study, some useful' information was obtained from t h e
rather long and arduous process that evolved.
However, the questions
that were ultimately answerable i n this situation are of very- limited
interest .
The process of doing research under the restrictions imposed by
program and the problems encountered along the way, provide a
convincing argunent for the traditional order of doing research.
i s , ask your questions, t a i l o r a research design to directly address
those questions and do not lose sight of your objectives.
was found that symptoms do decrease over a f a i r l y lengthy time period
for those clients who self-refer t o the Vancouver Preventive Medicine
However, these results cannot be general i zed to the population
nor can i t be assumed that the decrease i i due to the
treatment received. No relationship was found between symptom decrease
and a t t i tudes toward health care, locus of control for health or a
number of demographic variables.
For t h i s client group, number o f
presenting symptoms was the best predictor of change.
Given the limited
nature of these results, the discussion will centre on problems
encountered, rather than the findings, themselves .
The Process
As t h i s study progressed, both the sample and the possibilities
A number o f t h e problems encountered related to t h e setting
imposed treatment situation, decisions points arose where choices had t o
be made between e x p e r k n t a l r i g o u r and o t h e r considerations.
As can
be expected i n an actual program s e t t i n g , these decisions were n o t
always l e f t up t o t h e researcher.
I t was considered o f primary
importance i n t h i s s i t u a t i o n t o d i s t u r b t h e s e t t i n g and i t s c l i e n t s as
l i t t l e as possible.
The choice was t o use o n l y those data already
a v a i l a b l e a t the Centre, o r some e a s i l y c o l l e c t a b l e additions.
also decided t h a t the research be kept t o t a l l y out o f t h e way of t h e
physicians and t h e i r c l i e n t s .
The advantages o f f e r e d by t h i s s e t t i n g
which made t h i s prospect seem worth pursuing, were a l a r g e data base,
long term t e s t - r e t e s t opportunity and the p o s s i b i l i t y o f c o l l e c t i n g
useable data under these conditions.
Non-instrusive data c o l l e c t i o n methods were chosen both t o t e s t t h e
methodology and t o comply w i t h the program physicians' requirement t h a t
p a t i e n t s not be d i s t u r b e d - o r inconvenienced.
consequences which affected t h e r e s u l t s .
This had a nunber o f
The measures used i n the f i r s t
phase o f t h e i n v e s t i g a t i o n were predetermined. From an o r i g i n a l
expectation of 4 o r 5 useable measures o f d i f f e r e n t aspects o f t h e
program, o n l y one materialized.
.research purposes.
This measure was n o t a good one f o r
The a t t i t u d e measures chosen f o r t h e second aspect
o f t h e research were short, and the behavioural measures o f change
(smoking and d r i n k i n g ) were n o t as s e n s i t i v e as .one would have liked.'
Also, the sample became very narmw due t o a high rate af nmresponseI t would have been too i n t r u s i v e and c o s t l y t o expend more resources i n
o b t a i n i n g a more representative sample, a t the Centre.
S t a f f turnover
and differences among physicians i n record keeping h a b i t s a l s o had an
This r e s u l t e d i n missing t e s t s and demographic i n f o r m a t i o n f o r
some c l i e n t s .
Another d i f f i c u l t y was the Jack o f c o n t r o l o r
No such information had been c b l l e c t e d a t an e a r l y
comparison data.
stage i n t h e p r w a m t o provide long-term comparison o p p o r t u n i t i e s .
attempt t o make up f o r t h i s lack by using program drop-outs as c o n t r o l s
was n o t f r u i t f u l .
Proceeding without a c o n t r o l group and w i t h a d i s a p p o i n t i n g r e t u r n
rate, steps were taken t o sajvagq s m degree o f general i z a t i o n f o r t h e
eventual r e s u l t s - o f t h e analyses.
This involved
a l o t o f data c o l l e c t i o n
and manipulation on t h e outcome phase o f the study, w i t h r e l a t i v e l y
l i t t l e return.
I n t h e end, l i t t l e could be said about t h e program.
More problems were encountered when the second research questioi!
was addressed.
Again, no c o n t r o l group was available.
necessary f o r o b t a i n i n g an even marginal l y comparable4 c o n t r o l group were
considered e i t h e r too i n t r u s i v e , too costly, o r i n need o f
time c&i
long a
ttment t o be feasible under e x i s t i n g c o n s t r a i n t s ;,-; The sample
narrowed d r a s t i c a l l y , again.
This was,
i n part, due t o poorGresponse
but t h e primary f a c t o r i n t h e shrinkage was that,few p a t i e n t s a t the
Pt4C had had behavioural masures taken a t f i r s t contact.
l i m i t i n g f a c t o r a t t h i s stage was t h a t a r e l a t i v e l y important p o s t - t e s t
measure (Health ~ a z a r d~ ~ p rsal
a )i was abandandoned because o f cost.
Results o f t h i s phase were eveh l e s s generalizable than those o f . the
outcome phase.
Overall, the process was f r u s t r a t i n g from a.research p o i n t o f view.
Attempting t o evaluate a program o r .answer a. research question while
having almost no c o n t r o l over t h e data base was, i n t h i s case, a
wasteful process.
Ruch e f f o r t was expended i n d a s c o l l e c t i o n ,
analysis and changes o f plan as obstacles arose.
were o n l y p a r t i a l l y answered.
The o r i g i n a l questions
A d d i t i o n a l l y , the n o n g e n e r a l i z a b i l i t y o f
r e s u l t s renders them r a t h e r t r i v i a l .
On a more p o s i t i v e note, t h i s process d i d o f f e r some useful
There comes a p o i n t where non-intrusive methods l o s e t h e i r
usefulness 'as research tools.
This study made t h a t f a c t abundantly clear.
Here, t h i s method o f maintaining cooperation and t r u s t between program
managers and the researcher n e i t h e r made the process f l o w more smoothly,
nor helped provide more meaningful r e s u l t s .
Perhaps t h i s method can be
used t o advantage i f t h e researcher has been involved i n the o r i g i n a l
program planning stages.
A t t h a t time, instruments could be b u i l t i n
which would e s t a b l i s h a sound data base f o r l a t e r use.
F a i l i n g that,
t h e reseacher must have c o n t r o l over the measures introduced l a t e r and
9 1
over the data c o l l e c t i o n process.
I n doing research w i t h i n a ongoing program, a f a l l b a c k p l a n o f
a c t i o n should be established from t h e o u t s e t -
The researcher should be
f r e e t o use t h i s strategy when necessary, as we1 1 , if meaningful r e s u l t s
are t o be gained.
process, i t
F l e x i b i l i t y i s important i n working w i t h r e a l world
However, when changes of plan are made i n the middle o f thg
easy t o lose s i g h t of the o r i g i n a l objectives.
A well
.thought out a l t e r n a t i v e plan can help avoid t h i s situation'.
Another important p o i n t t o come out o f t h i s study i s t h a t while
non-intrusive methods may be used t o advantage, this can become
obstructive i f carried t o the point of non-involvement.
During this
project, the doctors a t the Centre were minimally involved w i t h the
Aside from t h i s , > t h e researcher was, essentially,
c u t off from the program being psearched.
Carried t o t h i s degree,
bun-intrusion becomes non-involvement, and good results are h i g h l y
unlikely under these circumstances.
The Outcome Phase
This stage of the investigation
i n symptomatology for those who become involved with the PMC,
these results are n o t generalizable nor attributable t o an identified
treatment plan.
I t must not be assumed t h a t the symptom reduction i s
actually due t o the treatment.
A placebo effect may besoperating such
t h a t symptom reduct ion may simply be due to an expectation of
This i s not likely over such a long period as two years,
although further research i s needed before this possibility can be
The time span between testing periods on the CMI also makes
i t unlikely that the symptom ddrease i s due to a "savings" on the
Second t e s t o r a general tendency t o improve the second tim a t e s t i s
Since this i s not
&a& one's
a learning situation, b u t rather questions
heal tb, there are no correct answers to be learned.
&tit w e
teeat& mmmti-rrg otfrer
i n stmftar s-r'tuatTons the
re1 w a n t compar-isons were not made.
I t i s more likely that t h e decrease is due t o a genera? improvement
i n the population's health and increased warenesf of heam ifstles e w r
the p a s t several years.
I t i s also quite likely t h a t part of the
decrease i s due t o regression toward the mean, since most norms for the
CMf in the general population are slightly lower t h a n the:mean CMI Time
1 score for this client group.
In addition t o the lack of control group and missing behaviour change
measures, the actual measures available posed s t i l l further problems.
The PMC had planned t o collect a number of measures on a l l clients who
came for consultation.
These measures were also t o be readministered a t
regular intervals throughout; contact with the Centre.
Very few of these
measures were actually taken in the f i r s t instance, and almost none were
This posed one o f the most important obstructions to
assessing the effectiveness of this four-part program of lifestyle
The measures originally selected had covered a number of areas
of living (e.g., eating habits, syrrtptmato'kgy, physiological measures
such as blood constituents, external stressors, smoking habits, seat
belt use, e t c . ) .
Upon inspection of client f i l e s , i t became apparent t h a t only the CMI
could be used to assess changes. Other measures promised to offer useful'
information (e.g., seat belt use, smoking and drinking, Health Hazard
Appraisal ) , however none of these were useful in the end.
The CMI , a1 though available for most patients, posed problems as well.
I t was subjxtively designed as a symptom checklist before factor analysis
and other sophisticated t e s t design techniques were widely
has not yet been objectively shown that this t e s t actually measures those
asDects o f health that i t s designers expected, nor i s i t s *reliability
However, i t does have a certain measure of face validity.
I t can, a t l e a s t , be considered a quantitative measure of "number of
complaints", and was shown to reliably measure something f o r t h i s
sampl e.
Another problem for the CHI i s that i t i s cumbersome.
symptom chec'klists such as the Cmulative Illness Rating Scale (Linn,
L i n n , & Gurel , 1968) and the Pbysical Sympto~sInventory (gahler, 1968) a r e
shorter and have demonstrated content validity.
These would be better
measures 'to use i f more work i s to be done in the area.
The two year t e s t - r e t e s t opportunity was an important aspect o f the
Had the opportunity to introduce a different measure presented
i t s e l f the CMI would s t i l l have been retained in t h i s study, because i t
was assumed that introducing a nw measure would have answered an
entirely different question ( a short term e f f e c t ) . I t would 'now be
useful to begin using other measures a t intake and to compare these with
the available CMI d a t a , both a t intake and for change differences over
two years.
This would not only provide a clearer picture of the
a t t i t u d e versus symptom reduction issue, b u t would also investigate the
validity of the C M I .
T h e validity issue i s an important one, here.
had been adhinistered by program s t a f f since the
4 less t h a n
Because the measures
PMC opened i t s doors with
clear plan for future evaluation, a control or comparison
group was not s e t up.
Nor were the measures chosen researched for t h e i r
Both of these f a c t o r s affected the
methodological a p p l i c a b i l i t y .
eventual g e n e r a l i z a b i l i t y of r e s u l t s .
Even i f a1 1 of the original
measures had been a v a i l a b l e , i t i s npt c l e a r t h a t these would pipvide
useable research d a t a .
Further, in the absence of a non-treatment
control group of any s o r t , conciusions based on t h e study r e s u l t s are
confined t o self-referred us-ers of the PMC.
Attitudes and Symptm Reduction
Results of numerous analyses revealed l i t t l e about t h e relationship
of symptom decrease over time t o a t t i t u d e change.
No relationship was
found between a t t i t u d e toward health care or health locus of control and
a decrement i n 'symptmatology f o r t h i s sample: T h e o r i g i n a l supposition
of t h i s section of the research was t h a t symptom reduction m u l d be
s i g n i f i c a n t l y more e v i d e n t f o r those who wanted a c t i ve,.&rticipation
their own care, and/or f e l t t h a t they have control oGr t h e i r own health.
This was not the case.
Rather, i f patients actualiy come t o the Centre
they are l i k e l y to improve t o the degree that they e x h i b i t t h e type of
symptoms sampled by the C M I .
F u r t h e n c r e , f o r t h i s sgb-sample,
iriprovenent i s l i k e l y t o occur regardless o f a t t i t u d e toward personal
cesponsibil it y f o r health.
However, these r e s u l t s cannot be generalized t o the overall
I t i s a l s o questionable whether they can be assumed t o apply
t o the t o t a l c l i e n t group from which t h e 58 subject sub-sample was drawn.
. -, e t h e sub-sample can be considered representative of the Centre's
~ o p t l i a t i o non several variables
(CMI score, sex, age, education), i t i s
not clear whether attitudes are generalizable.
Therefore, canclusions
must be confinedeto the very small 58 subject group.
I n order to render these results more widely applicable, a randunly
selected sample of the PMC's c l i e n t s could be administered t h e a t t i tude
measures for comparison with these data.
These measures., a l o n g w i t h
demographic d a t a would . i n d i c a t e whether the present small sample was
I t 5 s interesting t o note t h a t the mean score on the MIS
( 4 . 5 7 ) i s well below n o m s established i n a college population (see
T a b f ~1 ) . This g r o u p also scored more externally !50.45) than d s d
those subjects used to e s t a b l i ' s h norms for t h e HLC. ( s e e Table 2 ) .
these subjects demonstrated t h e i r cooperativeness by returning the
two sets of questionnaires, and since they also seem less self-directed
than the norfi i n health matters, they may represent a specjal
("cooperativeii?j p~pulationsubgroup.
Given thdat l i t t l e otner information o f interest has been obtained
f r o m t h i s study,
and PMC s t a f f was not prepared to a s s i s t w i t h d a t a
col iection or t o have the n o w I rout?ne
s r u p t e d , f u r t n e r avenues were
In the process of doing research i n such a situation, there
comes a point where further attempts would yield l i t t l e useful information,
relative t o the' costs involved.
I n t h i s case, the relationship of
attitudes to symptom reduction could be investigated elsewhere in a much
more rigorous manner, without: intruding further on t h i s s e t t i n g .
Sugqestions for Further Research
Since most normative data for the HOS and the HLC were gathered from
so7 fege samples, i t would be i n t e r e s t i n g to i n v e s t i g a t e t h e relationship
between these measures and symptom reduction in s i m i l a r samples.
Information i s a l s o needed on the relationship between personality type
and spptom decrease, a n d on personal i ty , a t t i t u d e s and decrement i n
s p ~ p t m t d o g y . Intuitively, i t
seems t h a t some relationship would
e x i s t , but i t is not c l e a r what t h a t relationship m'ight be.
I n identifying t h e a t t i tude/symptorn.. relationship, preferences for
a particular type o f treatment might' be measured by comparing the
subsections of the Krantz H e a l t h Opinion Survey w i t h symptom decrease.
These subsections discriminate between those c l i e n t s who prefer active
involvement and those who prefer infomation.
However, i n t h i s study
a i I o f the atti'tade, control and demographic variables together
accounted f o r i.3: or l e s s o f the variance in C M I scores. Thus, i t was
.icr wssi3:e
tc- :rtvestisate t h i s retationship in tt7.i~study, but. i t does
sugsest ar area where furtner research may be useful.
I t should be noted t h a t the a t t i t u d e and personality measures were
taken a t t h e second CVI adminis~ratiofi,z f t e r t h e treztnent had idng
been underway.
I t could, in f a c t , be considered complete in many cases.
h d i f f e r e n t relationship might be found i f a t t i t u d e measures were
administered a t i n t a k e .
Further information of value may be
found i f changes i n - b o t r a t t i t u d e s and symptom number were compared over
Another area of inteFest would be whether the P M and other such
programs of.prevention a r e more e f f e c t i v e f o r p a r t i c u l a r c l i e n t
subgroups such as depende& p e r s o n a l i t i e s , various psychological
disturbances, smokers, the aged and others.
Private preventive programs
could then be monitored t o determine i f some aspects are more effective
t h a n others ( i e . education, improved nutrition, decreased s t r e s s o r s ) ,
and f o r nhom these separate components are most effective.
i t wou1 d be most useful t o know whether the people who receive t h i s
type of counselling actually p u t the recornended l i f e s t y l e changes i n t o
Furthermore, i f they do, does t h i s result in symptom reduction?
I t m y be that information alone would provide a reduction i n s t r e s s or
affect the c l i e n t in some other way t h a t results i n decreased symptms,
whether or not the physician's advice i s actually implemented.
a t t e m p t was made during t h i s study t o answer t h i s question.,However,
the available behaviour change measures were neither sensitive nor
broadly based enough to o b t a i n answers.
I n sumnary, there were numerous problems encountered i n this
research endeavour.
Some interesting and useful information was gained.
h e w r , the most useful knowledge acquired during t h i s study was not
r e l a t e d to the actual research questions.
T h a t aspect of the research
was judged t o have contributed less really conclusive evidence thav the
amount o f e f f o r t warranted.
The primary gain from t h i s study was the
support i t lends t o the need, to maintain ri'gorous research methodology i n
doing program related research.
I t is no use trying to force a match
bet;.;een r i g i d exper',wntal rigour and sloppy r e a l i t y .
However, i t is
important t o adhere to s o w basic methodological guidelines.
suggestion o f Hackler (1969j and others t h a t programers and researchers
must allow f o r one another's needs i s we?? taken.
However, i t must riot
be forgotten t h a t this a1 lowance must work both ways.
I f metbdology
i s stretched too far i n an e f f o r t t o accomnodate c l i n i c a l reality,
n e i t h e r the c1 i n i c i a n nor the research comnuni t y will benefit.
Appendix A
Figure 1 : D e t a i l s of Sample Shrinkage f o r Data Set A and Data S e t 0
: Mailout o f :
r . ~3 = ~ ~ ~
:Experimental s: :Controls:
: Iq=m . . . .:. :- .R.=.3.5 . . .:.
.: : I . . ..... . E . .
:Used i n later:
:Data Set 8:
. * t *
. t t * . **.-**-
. . . . . . .1. . . .
: recorded .
:. . . N=203
. . . . . . . .f . . . . .
:. R=60
* * * = * - * *
* * . * * * * * * *
:Used i n Analysis
:Data Set A .
: w i t h sex &
:Hz44 1
. . . . . . . . . . . . a
:Data Set A :
' w i t h age
;. . . N=236
... .:::r:::::,;
:Data Set A
:with outliers:
:and m i s s i n g :
:data Ss
: D a t a Set B
'Sent A t t i t u d e
Randlatly selected:
. . . . . . I . . . . . .
:Sex & Education:
:. ~=465...........
(a1 1 Ss)
:. N=229
. " . .]" . . . . . .
. . . . . .1. . . . . . . . . .4. .
:d e l e t e d
:. . N=436
Table A
Proportion o f Sample ~ S a s e Data
Set A
Condi t ion
- r e c e i v i n g CMI a t i n t a k e
Dropped out a f t e r f i r s t
Return Rate ( X )
32.0 (233
N/ A
N/ A
21.5 ( o f 153)
"Open-fi l e s " returns (experimentals)
34.7 ( o f 576)
T o t a l non-responders t o 2nd CHI
2 o r more v i s i t s
Drop-out returns ( c o n t r o l s-phase 1)
Non-responder group (subsampl e )
-47.6 ( o f 496)
Responder g r o u p with d e l e t i o n s
31 - 4 (of 729)
Data.. Set A
6 3 . 8 ( o f 729)
Table 8
Proportion o f Sample Decrease: Data Set
% of
% o f 233
% o f 126
Unprompted returns
. 7.2
U s a b l e records
Table C
Frequencies for CMf Scores: Data Set A
Hunber o f Ss
CMI Tim 2
Nunber o f Ss
70- 79
100- 109
110- 119
120- 129
120- 129
Table D
Frequencies.for C M I Scores.
Data Set B
CMI Time 2
CHI Time 1
Number of Ss
6- 10
5- 09
10- 14
115- 19
26- 30
30- 34
35- 39
50- 54
55- 59
Tota 1
70- 74
Nllmber of Ss
. CMI ~ h d n s e
itumber o f Ss
Frequencies for H t C and HOS Scores: Data Set B
HLC Scores
HOS Scores
Number o f Ss
Nunber o f Ss
Abramson, J . The Cprnell Medical Index as an epidemiological tool.
American Journal of Public Health, 1966, 56, 287-298.
Abramson, J . H., Terespolsky, I . , Brook, J . , & Kark, S. The Cornell
W i c a l Index a s a health measure i n epidemiological s t u d i e s . B r i t i s h
Journal o f P r e v e ~ t i v eand Social Medicine, 1965. 19, 103-110,
Aharan, C. Clinical indications of m t i v a t i o n i n alcoholic patients.
Quarter1y Journal of Studies on Alcohol, 1967, 28, 486-492.
Ajzen, I . , 8 Fishbein, M. Atti tude-behavior relations: a theoretical
analysis and review of empi ri'cal research. Psycholoqical Bul J e t i n ,
1911, 84, &@-918Arnoff, F . , Strough, L . , & Seymour, R . The Cornell Medical Index in a
psychiatric outpatient c l i n i c . Journal of Clinical Psychology, 1956,
12, 263-268.
Auerbach, S. . Trai t - s t a t e anxiety and adjustment t o surgery.
Consulting and Clinical Psychology, 1973, 40, 264-271.
Journal of
Bagozzi, R . Attitudes, intentions and behavior: A t e s t of some key
hypotheses. Journal of Personality and Social Psychology, 1981,
4 1 ( 4 ) , 607-627.
Becker, M . , Kaback, M . , Rosenstock, I . , 8 R u t h , M. Some influences on
pub1 i c participation i n a genetic screening program. Journal of
Cornunity Health, 1975, 1, 3-14.
Bemij, K . Current approaches t o the eiioiogi a n d treatment of anorexi a
nervosa. Psychological B u l l e t i n , 1978, 85(3), 593-617.
Bloom, F., Segal,'D., 6 Guillemin, R. Endorphins: Profound behavioral
e f f e c t s in r a t s suggest new e t i o l o g i c a l factors i n menta 7 i l l n e s s .
Science, 1976, 194, 630-632.
b-ro u d i n, H.
Contingency contracting: A therapeutic t o o l i n t h e
deceleration of amphetamine use. Behavior Therapy, 1972', 3, 604-608.
Breslow, L .
R i s k factor, fntervention f o r health maintenance.
Breslow, L . A positive s t r a t e g y f o r 'the nation's health. Journal of
the American k d i c a t Association, 1979, 242(19), 2043-2095.
Brodman , K. , Erdmann , J. , Lorge, I . , & Wolff, H. The Cornel 1 -Medical
Index. Journal of t h e American Medical Association, 1949, 140,
Brodman, K. , ErQnann, J . , Lorge, I . , & Wol f f , H. The CMI Health
Questionnaire as a diagnostic instrument. Journal o f the American
145(3), 152-157.
Medical Association, 1951, Brodisan, K . , Erdmann, J . , Lorge, I . , Gershenson, P., & Wolff, H. The
CMI Health Questionnaire: The evaluation of emotional disturbance.
Journal of Clinical Psychology, 1952, 8, 119-124.
Brodman, K . , Erdmann, J . , Lorge, I . , & Wolff, H. The CMI Health
Questionnaire: The r e l a t i o n of p a t i e n t s ' complaints t o age, sex,
race, and education. Journal of Gerontology, 1953, 8, 339-342.
Brodman, K,, Erdmntl, J . , t o w , I . , k u t s c h b e r g e r , 3., & Wolff, H.
CMI Health Questionnaire: The prediction of ~svchosomaticand
psychiatric -disabi 1i t i e s in arky training. ~ & r can
i Journal of
Psychiatry, 1954, 3, 37-40.
Brodman, K . , van Woerkem, A . , Erdmann, J . , & Goldstein, t .
Interpretation of symptoms with a data processing machine.
of Internal Medicine, 1959, 103, 116-122.
Brown A . , & Fry, J . The CMI Questionnaire in the i d e n t i f i c a t i o n of
neurotic patients i n general practice. Journal of Psychosomatic
6 , 185-190.
Research, 1962, Buck, R.
Human motivation and emotion.
New York: J . Wiley & Sons, 1973.
Caldbeck-Meenan, J . Screening university students with the CMI.
of Psychosomatic Research, 1 x 6 , 9,.331-337.
Caplan, G .
P r i n c i p l e s of preventive psychiatry.
New York: Basic Books,
Clum, G . , Kole, D., Flag, J . , & Holberg, A. The s t a b i l i t y of CMI scores
and syrnpto,ms in a m i l i t a r y popqlation. Journal of Clinical
1970, 5 ( 2 6 ) , 188-191.
Corrado, R . R . Using experiments in evaluating delinquency prevention
program, In R . Raesch d R . R. Corrado (Eds.), Evaluation and
Criminal J u s t i c e Policy. Beverly H i l l s : Sage, 1981.
Cowen, E . Some problems in community program evaluation research.
Journal of Consulting and Clinical Psychology, 19%, 4 6 ( 4 ) 792-805.
Crisp, A. Some aspects of the relationship between body weight and
sexual behaviour with p a r t i c u l a r reference t o massive obesity and
2, 17-32.
anorexia. International Journ?l of Obesity, 1978, -
Crisp, A . , & Stonehi 11 , E . Relation between 'aspects of n u t r i t i o n a l
disturbance and menstrual a c t i v i t y i n anorexia nervosa. British
Medical Journal, 1971, 3, 149-151.
Cumnings, M., J e t t e , A . , Brock, B . , & Hoefner, D. Psychological
determinants of imnunization behavior i n a swine influenza campaign.
Medical Care, 1979, 17, 639-649.
Davidson, A . , & Jaccard, J . Population psychology: A new look a t an
old problem. Journal of Personality and Social Psychology, 1975, 31,
Davis, F., & Ditman, K. The e f f e c t of court r e f e r r a l and disulfuram on
motivation of alcoholics. Quarterly Journal of Studies on Alcohol,
1963, 24, 276-279.
Desroches, H . , & Larson, I . The CMI a s a screening device i n a VA
hospital . Journal of Cl inical Psychology, 1963, 19, 416-425.
Dudley, D. The CMI as an adjunct t o paraprofessional evaluation of
a1 coho1 addiction. Journal of Studies on A1 coho1 , 1976, 97-99.
Dunham, R . , & Mass, A . Reluctant r e f e r r a l s : The effectiveness of legal
coercion i n outpatient treatment f o r problem drinkers. Journal of
Drug Issues, WinterL, 1982, 6-19.
E l l i o t t R . , & Tighe, T. Breaking t h e c i g a r e t t e habit: Effects of a
technique involving threatened l o s s of money. Psychological Record,
18, 503-513.
1968, Fishbein, M . , & Azjen, I . Attitudes toward objects a s predictors of
s i n g l e and mu1 ti pl e behavior c r i t e r i a . Psychological Review, 1974,
81, 59-74.
Geyman, J . Preventive medicine in family practice: A reassessment.
8 ( 1 ) , 35-36.
Journal of Family Practice, 1979, -
Gottsman, - D . , b Lewis, M. Differences in c r i s i s reactions amona cancer
and s k g e & p t i e n t s . Journal o f Consulting and Clinical ~ ; y c h o l o ~ ~ ,
1982, 50(3) , 381-388.
Hackler, J . 1nvi t a t i o n t o e r r o r : The dangers o f evaluation and some
a1 t e r n a t i v e s . Canadian Journal o f Criminology, 1979, 21, 39-51.
Hall, J . , Robbins, L;, & Gesner, N. The Health Hazard Appraisal Program:
I d e n t i f i c a t i o n of r i s k f a c t o r s . Pwtgraduate Medicine, 1972, 51,
114- 123.
Holmes, T . , & Rahe, R . The Social Readjustment Rating Scale. Journal of
Psychosomatic Reseaich, 1967, 11,213-218.
Home, L . , Csanyi, A . , Gonzales, M . , & Rechs, J . How t o use continqency
contracting in t h e classroom. I1 1 i n o i s : Research Press, 1969.
Jacobsom, M. Problem solving and contingency contracting i n t h e
treatment of marital discord. Journal of Consulting and C l i n i c a l
45, 92- 100.
Psych01 ogy, 1977, J e n k i n s , C . Psychosbcial modifiers of response t o s t r e s s .
Human S t r e s s , December 1979, 3-15.
Journal of
Johns, M. Syrnptoms_of neurotic i 1l n e s s ' i n general hospital p a t i e n t s .
The Medical Journal of A u s t r a l i a , J u l y 1, 1972, 41-44.
J u r j e v i c h , R . Short i nterval t e s t - r e t e s t s t a b i 1i t y of MIPI, CPI , CMI
and a symptom c h e c k l i s t . The Journal of Psychology, 1966, 74,
Kanfer, F. Self management methods. In F . Kanfer & A . Goldstein ( E d s . ) ,
Helping people change. New York: Pergammon Press, 1980.
Keane, T . , Martin, J . , Berler, E . , Wooten, L., Fleece, E . , & Williams, 3 .
Arei hypertensj ves l e s s a s s e r t i ve: A control1 ed evaluation. Journal
50(3) ,. 499-508.
of t o n s u l t i n g and C l i n i c a l Psychology, 1982, Krantz, D., Baum, A . , & vanWideman, M. Assessment of preferences f o r
s e l f-treatment and information i n health care. Journal of
Personal i t y and Social Psycho1 ogy , 1980, 3 9 ( 5 ) , 977-990.
Lawton, M . The screening value of t h e CMI.
Psycho1 oqy, 1959, 2 3 ( 4 ) , 352-356.
Journal of Consulting
l e v i t t , E , The Cornell M i c a 1 Index, In 0 . L Buros ( E d - ) , The
seventh mental rneasureinents yearbook (Vol. 7 ) . New J e r s e y , The
Gryphon Press, 1974.
L i n n , B . S , Linn, M. N., & Gurel, L . Cumulative I l l n e s s Rating'Scale.
Journal o f . t h e Amri can G e r i a t r i c Society, 1968, 16,622-626.
Lykken, D. The ~ o r n e l iMedical Index H e a l t h Q u e s t i o n n a i r e . , In 0.
Buros (Ed. ) , The seventh mental measurement yearbook (Vol 7 ) .
Jersey: The Gryphon Press, 1974.
McCamy J., & Presley, J. Human l i f e s t y l i n g : A program f o r keeping
whole i n t h e 20th Century. New York: Harper & Row, 1978.
McDonald, C. Measures o f n e u r o t i c i s m i n t h e e l d e r l y .
1, 44-57.
Zealand J o u r n a l o f P s y c h i a t r y , 1967, -
~ u s t ' r a l i aNew
Marks, I. Components and c o r r e l a t e s o f p s y c h i a t r i c q u e s t i o n n a i r e s :
o b s e r v a t i o n s on t h e CHI, EPI, and T a v i s t o c k Self-Assessment
I n v e n t o r y i n phobic p a t i e n t s . B r i t i s h Journal o f Medical ~ s y c h o f l o ~ ~ ,
1967, 9,
M i l l e r , P. The use of b e h a v i o r a l c o n t r a c t i n g i n t h e t r e a t m e i t o f
a l c o h o l i s m : A case r e p o r t . Behavior Therapy, 1972, 3, 593-596.
M i l l e r , P., Herson, M., & E i s l e r , R. R e l a t i v e e f f e c t i v e n e s s o f
i n s t r u c t i o n s , agreements and r e i n f o r c e m e n t s i d b e h a v i o r a l c o n t r a c t s
w i t h ' a1 coho1 ics. J o u r n a l ~f Abnormal Psycho1 ogy , 1974, 83, 548-553.
M i 1sum, J
'The Health, Hazard A p p r a i s a l : A p l a n n i n g t o o l f o r good
h e a l t h . C a t h o l i c H o s p i t a l , March, 1978, 2, 40-47,
P e t t y , R., Cacioppo, J . , & Goldman, R. ..Personal involvement as a
d e t e n i nant o f argument-based persuasion.
41(5), 847-855.
S o c i a l Psychology, 1981 , -
Journal o f Personal it y and
Pomerleau, 0. & Brady, J . Behavioral medicine: Theory and p r a c t i c e .
B a l t i m o r e : W i l l i a m s & W i l k i n s , 1979.
Pond, D., Ryle, A., & Hamilton, M. S o c i a l f a c t o r s and n e u r o s i s i n a
working c l a s s p o p u l a t i o n . B r i t i s h Journal
o f P s y c h i-a t a , 1963, ---587-594.
Richman, A. , Slade, H. , & Gordon, G. Symptom q u e s t i o n n a i r e v a l i d i t y i n
assessing t h e need f o r p s y c h i a t r i s t ' s care. B r i t i s h J o u r n a l o f
P s y c h i a t r y , 1966, 112, 549-555.
Rosenbgrg , M., & L i f t i k, J . Use of c o e r c i o n i n t h e o u t p a t i e n t t r e a t m e n t
of a l c o h o l i s m . J o u r n a l o f Studies on A l c o h o l , 1976, 37, 58-65.
Rosenstock, I , H i s t o r i c a l o r i g i n s o f t h e H e a l t h Be1i e f Mode1
Education Monographs, 1924, 1, 328-335.
R o t t e r , J. Generalized expectancies f o r i n t e r n a l versus e x t e r n a l c o n t r o l
o f r e i n f o r c e m e n t . Psycho1 o g i c a l Monographs, 1965. 80(1, Whole No.
Ryle, -A., f i a m i 1ton, ti. Neurosis i n f i f t y married couples.
108, 265-272.
o f Mental Science, 1962, .
Salber, E., & Abelin, T. Smoking behavior o f Newton school c h i l d r e n
five-year f o l low-up. P e d i a t r i c s , 1967, 40, 363-372.
Schuman, S. Prevention: The v i t a l and unique . r o l e o f t h e f a m i l y
physician. Journal o f Family P r a c t i c e , 1979, 9 ( 1 ) , 97-102.
Selye, H.
The s t r e s s o f l i f e . New York: McGraw H i l l , 1956.
Seymour,- G. The s t r u c t 6 r e and p r e d i c t i v e a b i l i t y o f t h e ChI f o r a
normal sample. Journal of Psychosomatic Research, 1976, 20, 469-478.
Sloane, P. # p r e v e n t i o n o r i e n t e d m e d h l record.
P r a c t i c e , 1979, ?( 1) , 89-96.
Mi t k , R. Changes i n locus o f c o n t r o l as a func
r e s o l u t i o n . Journal o f Abnormal Psychology,
Journal o f Family
n of l i f e crisis
75, 328-332.
Spielberger, C . (Ed. ) . , Anxiety: Current trends i itheory and research.
( 0 1 1 ) New York: Academic Press, 1972.
Stout, C . , Wight, M., & Bruhn, 3 . The C M I i n d i s a b i l i t y evaluation.
B r i t i s h ~ o u r n a lo f Preventive and s o c i a l Medicine, 1969, 251-254.
S t u a r t , R. Operant-i nterpersonal treatment f o r mari ta,! discord.
o f Consul t i ng and C1 in i ca3 Psycholdqy , 1969, 33, 675-682.
V a n n i c e l l i , M. Treatment c o n t r a c t s i n an i n p a t i e n t alcoholism
treatment s e t t i n g . Journal of Studies on Alcoho1,,1979, 40, 457-471.
Yerghese, A. Relationships between t h e E P I , t h e C M I and t h e
psychogal vanic s k i n r e f l e x . B r i t i s h Journal o f Psychiatry, 1970,
116, 27-32.
Vigersky, R. (Ed. )
Anorexia nervosa.
New York: Academic Press, 1977.
Voegle.r, R. Alcohol consumption under v o l u n t a r y and i n v o l u n t a r y
treatment. Journal o f Studies on A1 coho1 , 1976, 37, 890-897.
Wakeling, A., de Souza, V., & Beardwood, C. Assessment bf t h e negatiCe
anct.positive feedback e f f e c t s o f administered estrogen on
gonadotropin r e 1ease i n p a t i 4 n t s w i t h anorexia nervosa. Psychological
7, 397-405.
M i c i n e , 1977, -
Wahl e r , H. J . The Physical Symptoms Inventory:' Measuring 1eve1 s of
somatic complaint b e ~ a v i o r . Journal of C l i n i c a l Psychology, 1968,
- 207-211.
Wallston, K., Kaplan, G . , & Maides, S. , Development arqd v a l i d a t i o n of
t h e Heal th'Locus of Control (HLC) Scale. Journal o f ,Consul t i n q and
C l i n i c a l Psychology, B76, 4 4 ( 4 ) , 580-585.
Ueigel, R . , & Newman, L . Increasing attitude-behavior correspondence
by broadening t h e scope of t h e behavioral measure. Journal of
33, 793-802.
Personafi t y and Social Psychology, 1976, Weiss, S. Psychosomatic aspects o f symptom p a t t e r n s among major surgery
13, 109-114.
p a t i e n t s . Journal of Psychosomatic Research, 1969, Wexberg, L.
Alcohol consumption and s o c i a l improvement: motivation f o r
Q u a r t e r l y Journal of Studies on Alcohol, 1953, 14,300-314.
W h i t e , ' < , Reznikoff, M . , & E v e l l , J . Usefulness of t h e CMI in a c o l l e g e
42, 94- 105.
health department. Mental Hygei ne, 1958, ,Williams, C . , Arnold, C . , 8 Wynd; E . Measurement of demographics and
rehsymptom c l u s t e r s . Journal of Consulting Psychology, 1959,
231 3 ) , 250-255.
Winnett, R .
Parameters of deposit c o n t r a c t s in t h e modification of
2 3 , 49-60.
Psychological Record, 1973, B
Wurtele, S . , Roberts, M . , & Leeper, J . Health b e l i e f s and i n t e n t i o n s :
P r e d i c t o r s of return .cornpl iance i n a TB detection d r i v e . Journal of
Appl ied Social Psychology, 1982, g ( 2 ) , 128-136.