Determinants of dentists' decisions to initiate dental implant treatment:

A judgment analysis
Pieter Koele, PhD, a and Johan Hoogstraten, PhD b
Faculty of Psychology and the Amsterdam Academic Centre for Dentistry, University of Amsterdam,
Amsterdam, The Netherlands
Statement o f p r o b l e m . Do psychosocial patient characteristics influence the decision of a dentist in rec-
ommending implant treatment?
Purpose. This study assessed the importance that dental general practitioners attach to psychosocial patient
characteristics when judging the suitability of these patients for dental implant treatment.
M e t h o d and material. The judgment strategy of 30 Dutch dental practitioners was studied by using 2
tasks. Dentists first judged a series of written scenarios representing fictitious patients, each patient being
characterized by 4 psychosocial aspects to determine the patients' suitability for implant treatment. The sec-
ond task involved rank ordering 14 patient characteristics according to their importance in implant treat-
ment decisions.
Results. Dentists judged the personal appearance of the patient and, to a lesser degree, the patient's
socioeconomic status to influence decisions to treat. However, these characteristics are ranked as quite
unimportant in decision making, whereas motivation, oral hygiene, and level of neuroticism were the most
important patient characteristics.
Conclusions. The results on the 2 tasks were dearly not in agreement. There was a substantial disagree-
ment between what dentists say to be important characteristics (rank order task) and the characteristics they
actually use to judge the suitability for implant treatment (scenario task). Moreover, it appeared that agree-
ment among the dentists is quite low, especially for the scenario task. (J Prosthet Dent 1999;81:476-80.)

Dental implant treatment is seen as a safe and effec-
tive alternative for treatment o f edentulous patients. In
The Netherlands, a large scale nationwide study evaluat-
ed both quantitative (patient and treatment characteris-
tics) and clinical (treatment effects on denture
complaints) factors. 1-4 Other studies have assessed the
edentulous patients' well-being, both before and after
dental implant treatment. 5-7 These studies reported that
for a large proportion o f edentulous patients, dental
implants are an effective treatment alternative and result
in high levels o f patient satisfaction.
Dental implant treatment has been included in the
Dutch National Dental Insurance Schema since 1989,
potentially affecting approximately 1.5 million totally
aAssociate Professor, Department of Methodology, Faculty of Psy-
chology.
bprofessor, Department of Social Dentistry and Dental Health Edu-
cation, Academic Centre for Dentistry; and Professor, Depart-
ment of Methodology, Faculty of Psychology.
edentulous patients. 3 Edentulous patients with func-
tional denture problems who have been helped by con-
ventional dentures are entitled to implant treatment. In
The Netherlands, the treating dentist applies for the
implant treatment for the patient, then the dental con-
sultants o f the health insurance company decide
whether this will be approved for reimbursement. The
judgment strategy o f these dental consultants was thor-
oughly analyzed by Groen, 8 who suggested that gener-
al practitioners " . . . t e n d to be reserved in applying
dental implant treatments. It may, after all, be assumed
that they will not apply for a treatment when they feel
it will not be approved." Because the general practi-
tioner applies for treatment, possibly at the request o f
the patient, but always on behalf o f this patient, the fac-
tors that this dentist takes into account when deciding
whether to seek approval arc investigated. In particular,
what psychosocial characteristics o f the patient may
influence the dentist's decision to recommend implant
treatment? From the literature on the success and fail-

476 THE JOURNAL OF PROSTHETIC DENTISTRY VOLUME 81 NUMBER 4

KOELE A N D HOOGSTRATEN
ures with conventional complete denture prostheses,
these psychosocial aspects are important determinants
of denture success. 9-12
The purpose of this study was to assess both the
importance of psychosocial patient characteristics and
the agreement among practitioners about their impor-
tance. This study was conducted in 2 parts. In the first
part, general practitioners were presented with a series
o f written scenarios representing fictitious patients
described by a number of psychosocial characteristics.
The practitioner judged each patient on his/her suit-
ability for dental implant treatment. In the second part,
general practitioners were asked to indicate the impor-
tance o f a number of patient characteristics in their
decision to consider recommending implant treatment.
METHODS
Fifty-one practicing general dentists were recruited
by telephone for this study. Eighteen dentists did not
meet the criterion for rendering implant treatment and
3 other dentists chose not to participate because o f
time constraints a n d / o r lack of interest. Thirty dentists
were eventually selected to participate in the project.
The 30 dentists were asked to determine the suit-
ability of edentulous patients for dental implant treat-
ment by participating in a series of tasks. The first task
involved a number of written scenarios that described
fictitious dental patients. Dentists were requested to
evaluate the suitability of these patients for implant
treatment on the basis of the information provided in
the scenarios. The second task involved a list of factors
that might influence the dentist's decision to treat
patients with implants. The dentists were asked to rank
order these factors to their importance in the decision-
making process.
F i r s t task: S c e n a r i o s
Four patient factors that may influence the decision
o f a dentist to recommend implant treatment were
selected for consideration. These factors were
(1) patient's socioeconomic status (SES), (2) personal
appearance, (3) comprehension, and (4) risk percep-
tion. In the written scenarios, these factors were opera-
tionalized on 2 levels: h i g h / l o w or g o o d / p o o r . By
systematically combining all levels for all factors, 16
scenarios were developed.
Socioeconomic status was defined by the occupation
and the level of education of the patients, and by their
way of verbalizing dental complaints about the prosthesis
they were wearing. Personal appearance was related to
the value one places on esthetics. Patients scoring "poor"
on this factor are described as looking unkempt, and
patients who scored "good" were perceived as tidy and
well-groomed. Comprehension involved the patient's
ability to understand what the treatment involved.
Patients who scored "low" on this factor showed no
APRIL1999
THE JOURNAL OF PROSTHETIC DENTISTRY
interest in the dentist's explanation of what was going to
happen, and patients scoring "high" listened eagerly and
asked questions. Risk perception was related to the
amount of insight the patient had into the possible neg-
ative consequences of the treatment. Patients scoring
"high" showed that they knew and understood the risks;
patients with "low" scores did not want to be informed
about the consequences or ignored them.
To increase the realism of the patient scenarios, 2
additional factors were introduced, gender and age.
Male and female patients were assigned systematically
to the scenarios in such a way that there was no rela-
tionship between gender and any of the factors. All
patients were between 56 and 59 years of age.
Three decoy scenarios were also created that con-
tained factors other than the ones previously listed, in
an effort not to overemphasize the factors being stud-
ied. The 16 scenarios were administered to the dentists
in a random order. However, the 3 decoy scenarios
were always in the same position (third, eighth, and
thirteenth) in the study document.
Before the scenarios were presented to the dentists,
background information was presented that was valid
for all patients. This information included the general
medical and dental history of the patients and related
conditions:
You may assume that for all patients described
in the scenarios the following hold true:
-They are referred to you by a respectable
colleague who seeks your opinion on their suit-
ability for implant treatment;
-They are edentulous patients and two cylin-
drical implants can be inserted on the site of the
former canines in the anterior mandible;
-Bone quality is good, and the quantity is
sufficient to enable implantation;
-A medical history indicates there are no signs
that might negatively influence the prognosis.
The patients are healthy individuals without con-
traindications for implant surgery.
After reading the text of a scenario the dentist was
asked to answer 2 questions: (1) H o w suitable for den-
tal implant treatment do you consider this patient? The
dentist could answer this question by circling either
"not suitable," "somewhat suitable," "reasonably suit-
able," or "very suitable." (2) Do you think the dental
consultant of the health insurance will approve reim-
bursement for the implant treatment costs o f this
patient? In this case, the dentist would circle "yes,"
"not sure," or "no."
477
THE J O U R N A L OF PROSTHETIC DENTISTRY
Table I. Summary results of the Analysis of Variance on question 1, task 1
Source of variance SS
Within subjects 326.39
Main effects
Socio-economic status (SES) 14.46
Personal appearance 91.79
Comprehension 31.96
Risk perception 4.24
Error 200.28
Between subjects 45.62
Total 408.01
SS = Sums of squares; dr= degreesof freedom; MS = mean squares.
S e c o n d task: R a n k o r d e r i n g
The second task assessed the importance the dentists
assigned to the patient factors in the scenarios by asking
the dentists to rank order these factors. However, to
exclude the possibility that the dentists might recognize
the focus of the second task, 14 patient factors were pre-
sented to the dentists. These factors were suggested by
4 experts in implant therapy as influences on the suit-
ability of patients for implant treatment: occupation,
level of education, complaint formulation, personal
appearance, comprehension, risk perception, coopera-
tiveness, communicative ability, intelligence, oral
hygiene, treatment expectation, motivation, age, and
neuroticism. The first 3 factors defined the socioeco-
nomic status of the patient (factor 1, first task) and the
next 3 factors corresponded with the other 3 factors in
the first task. The 14 patient factors were presented to
the dentists in a list, in alphabetical order. They were
requested to rank order the 14 factors according to their
importance in the implant decision process by writing
down a number between i and 14 next to each factor (1
being the most and 14 the least important factor).
Dependent variables
The first task had 2 dependent variables; namely, the
responses to the 2 questions each dentist had to answer
for all scenarios. In the first question (suitability for
implant treatment), the answer possibilities were "not
suitable," "somewhat suitable, .... reasonably suitable,"
or "very suitable" and scored as 1, 2, 3, and 4, respec-
tively. The answer possibilities for the second question
(approval of dental consultant) "no," "not sure," or
"yes" were scored as 1, 2, and 3, respectively. In the
second task the dependent variable was the rank order-
ing of the 14 patient factors by the dentists.
Data analysis
For the first task of assessing the influence of the
patient factors on the scores of the dependent variables,
two 2 (high vs low SES) x 2 (good vs poor person
appearance) x 2 (high vs low comprehension) • 2 (high
478
KOELE A N D H O O G S T R A T E N
df MS F p q2
434
1 14.46 26.89 .000 0.07
1 91.79 170.24 .000 0.31
1 31.96 59.28 .000 0.14
1 4.24 7.87 .001 0.02
428 0.54
29
472
vs low risk perception) factorial analyses of variance
(ANOVAs) were executed with repeated measures on all
factors. The corresponding F-tests had 1 degree of free-
dom (df) for the numerator and 428 for the denomina-
tor, which gives them sufficient power 13 to detect even
small effect sizes at a significance level of P<.05. Further-
more, agreement among the dentists in their judgment
of the patients was assessed by calculating Kappa's. 14
For the second task, agreement among the dentists
in their rank ordering of the 14 patient factors was
assessed by calculating Kendall's measure of concor-
dance W. is Rank ordering of the factors also used in
the first task were compared with the results of the
ANOVA on the first dependent variable of the first task.
RESULTS
First task: Scenarios
For the first question; " H o w suitable for dental
implant treatment do you consider this patient?", the
factorial ANOVA with repeated measures on all factors
revealed 4 significant main effects, and 4 significant
interaction effects (Table I). Because the sizes of the
interaction effects were small, attention was concentrat-
ed on the main effects. Table I presents the summary
results of the F-tests on the main effects. Eta squared
(112) is the estimated amount of variance in the answers
to the first question explained by a particular factor and
serves to indicate the effect size of that factor. Table I
also presents all main effects that were significant, but it
is also apparent that only personal appearance and com-
prehension had substantial effect sizes, that SES
explained a rather modest amount of variance, and that
risk perception did not really matter much.
The means o f the responses to the first question
confirm this analysis. Scenarios low on SES had a mean
suitability rating of 2.39 versus a mean rating of 2.70
for scenarios high on SES. Scenarios of patients with a
poor personal appearance scored a mean rating of 2.08,
and scenarios with a good personal appearance had a
mean rating of 2.97. Scenarios low on comprehension
had a mean rating of 2.26, and scenarios high on this
VOLUME 81 NUMBER 4
KOELE AND HOOGSTRATEN THE JOURNAL OF PROSTHETIC DENTISTRY

Table II. Summary results of the Analysis of Variance on question 2, task 1

Source of variance SS df MS F P r12

Within subjects 148.81 434

Main effects
Socio-economic status (5ES) 7.28 1 7.28 23.35 .000 0.05
Personal appearance 18.83 1 18.83 61.04 .000 0.13
Comprehension 6.24 1 6.24 20.13 .000 0.05
Risk perception 1.36 1 1.36 4.55 .034 0.01
Error 131.61 428 0.31
Between subjects 97.44 29
Total 270.42 472
SS = Sums of squares; dr= degreesof freedom; MS = mean squares.

factor had a mean rating of 2.77. Finally, scenarios with Table III. Mean rank numbers of the patient factors
a low risk perception had a mean rating of 2.43, and Patient factors Mean SD
scenarios with a high risk perception had a slightly
higher mean suitability rating of 2.60. 1. Motivation 2.5 1.4
For the second question; "Do you think the dental 2. Oral hygiene 3.7 2.7
3. Neuroticism 4.6 3.5
consultant of the health insurance will advise to reim-
4. Treatment expectation 5.5 2.6
burse the implant treatment costs of this patient?", the
5. Risk perception 5.6 2.0
muhifactorial ANOVA with repeated measures on all fac- 6. Comprehension 5.9 2.7
tors also revealed 4 significant main effects and 2 signifi- 7. Cooperativeness 6.5 3.5
cant interaction effects. Again, the sizes of the interaction 8. Complaint formulation 8.3 3.5
effects were minute. Table II summarizes results of the 9. Personal appearance 9.1 3.6
F-tests on the main effects. Personal appearance emerged 10. Communicative ability 9.1 2.6
as the most important factor for explaining variance in 11. Intelligence 9.9 2.9
the ratings on the dependent variable, but this time the 12. Age 11.1 3.2
effect size was not as impressive as it was in the previous 13. Occupation 11.3 2.8
analysis. Risk perception did not really play a role, and 14. Level of education 11.7 2.2

SES and comprehension factors had only limited impor-

tance when amounts of explained variance are concerned.
Mean ratings on this question were as follows: low versus
high SES, 2.11 versus 2.36; poor versus good personal
appearance, 2.03 versus 2.43; low versus high compre-
hension, 2.12 versus 2.35; and low versus high risk
perception, 2.18 versus 2.28.
Agreement among dentists
To assess the degree of agreement among the dentists
in their answers to the 2 questions, a 2-dimensional
cross-table was constructed for each question separately
with the 16 scenarios as 1 entry, and the 3 or
4 answer possibilities of the question as the other entry.
Table cells contained the number o f dentists having
assigned a specific scenario to a specific answer possibil-
ity. For such a table, Fleiss' extension 16 of Cohen's
kappa 14 for the agreement among many raters can be
calculated. It proved to have the value .11 for question
1, and .10 for question 2. Given the fact that kappa
ranges from 0 (no agreement) to 1 (perfect agreement),
the degree of agreement must be labeled as quite poor.
Second task: Rank ordering patient factors
Table III presents the mean rank numbers (and stan-
dard deviations) of the 14 patient factors, calculated for
each factor in the group of 30 dentists. Table III shows
that dentists consider patient motivation to be the most
important factor in recommending implant treatment;
oral hygiene and neuroticism followed. Treatment
expectation, risk perception, comprehension, and
cooperativeness were close together as moderately
important factors, and less important were complaint
formulation, personal appearance, communicative abil-
ity, and intelligence. Factors ranked least important
were age, occupation, and level o f education. The
rather large standard deviations o f the rank numbers o f
the factors suggested that there was disagreement
among the dentists on what they consider to be impor-
tant factors. Kendall's measure of concordance W was
calculated to determine the degree o f agreement
among the number of rankings. The value 0.51 was
found, which is significant above chance level
(Z2=229.5, df=15, P=.000), but not impressive. The
calculation o f Spearman rank correlations among all
possible pairs o f dentists revealed that there were
6 dentists who had no significant correlation with
the other dentists. Omitting the rankings o f these
dentists from the calculation of W leads to a somewhat

APRIL 1999 479

THE JOURNAl- OF PROSTHETIC DENTISTRY
higher value (W=0.65), but the degree of agreement
remains unsatisfactory.
From the analysis of the answers to the first question
in task 1, it appeared that, for their ratings of the suit-
ability fbr implant treatment of the patients in the sce-
narios, personal appearance emerged as the most
important factor and risk perception was the least
important factor (Table I). When asked explicitly to
rank order patient factors, as they did in task 2, the den-
fists considered risk perception to be more important
than personal appearance (Table III).
DISCUSSION
The main finding of this study was that general prac-
titioners were willing to suggest dental implant treatment
for edentulous patients fulfilling the following psychoso-
cial criteria: adequately motivated, showing good oral
hygiene, a low level of neuroticism, and an adequate
understanding of the treatment and the risks implied by
the treatment. Other patient characteristics such as occu-
pation, intelligence, communicative abilities, and age
seemed to be less important. However, these findings are
not as clear because of the following reasons. The results
for the 2 tasks presented to the dentists are not in agree-
ment. When asked to rank order patient characteristics,
dentists report that they consider patient's motivation,
oral hygiene, and level of neuroticism most important for
their decision to suggest implant treatment. It is clear
that rank ordering is a task susceptible to social desirable
responding. The scenario task is less transparent for
researcher's intentions. This might explain why the
results for this task demonstrated that personal appear-
ance was most important and risk perception least impor-
tant, whereas for the rank order task, the results show risk
perception to be far more important than personal
appearance. What was found was a clear discrepancy
between what dentists say are important patient factors
and what they actually used as important patient factors
when they decide to recommend implant treatment.
The degree of disagreement among the dentists
obviously was substantial, both for the scenario task
and the rank ordering task. Other research reports indi-
cated that dentists often have quite a range of opinions
when judging the same dental situation.S, 17 From this
study, it appears that guidelines should be developed to
enable some form of standardization of the application
phase that will lead to a fair procedure, irrespective of
dentist or patient peculiarities.
Another factor that may have influenced the results
was the failure to establish "real life" situations. The
artificial character of the research tasks (scenario, task,
decision making) was well recognized by the dentists,
as illustrated by remarks such as: "I first must see the
patient, .... I want to know how the patient reacts," and
so forth. However, after some further instructions and
explanations, the dentists were able and willing to con-
480
KOEI_E AND HOOGSTRATEN
tinue. Above all, it has been demonstrated that tasks,
like the one used in our study, produce results that can
be safely generalized to real-life decision making. 18
CONCLUSIONS
This study evaluated the importance that general
dental practitioners attach to psychosocial patient char-
acteristics when judging the suitability of patients for
dental implant treatment. From the results of this
study, personal appearance of the patient, and to a less-
er extent the patient's socioeconomic status, influenced
the decision to treat. However, there was a substantial
disagreement between what dentists say to be impor-
tant characteristics and the characteristics they actually
use to judge the suitability for implant treatment.
REFERENCES
1. Cune MS, de Putter C, Hoogstraten J. Characteristics of 5410 edentulous
implant candidates and the treatment they receive. Community Dent Oral
Epidemiol 1995:23:110-3.
2. Cune MS, de Putter C, Hoogstraten J. Treatment outcome with implant
retained overdentures. Part I--clinical findings and predictability of clini-
cal treatment outcome. J Prosthet Dent 1994;72:144-51.
3. Cune M. Overdentures on oral implants. [in Dutch] [PhD thesis.] Utrecht,
The Netherlands: Utrecht State University; 1993.
4. Cune MS, de Putter C, Hoogstraten J. A nationwide evaluative study on
implant-retained overdentures. J Dent 1997;25Suppl:S13-9.
5. Hoogstraten J, Lamers LM. Patient satisfaction after insertion of an
osseointegrated implant bridge. J Oral Rehabil 1987;14:481-7.
6. Grogono AL, Lancaster DM, Finger IM. Dental implants: a survey of
patients' attitudes. J Prosthet Dent 1989;62:573-6.
7. Kiyak HA, Beach DH, Worthington P, Taylor T, Bolender C, Evans J. Psy-
chological impact of osseointegrated dental implants. Int J Oral Maxillo-
fac Implants 1990;5:61-9.
8. Groen HJ. A dental consultant's dilemma: decision making in dentistry.
Amsterdam: Thesis Publishers; 1997.
9. Langer A, Michman J, Seifert I. Factors influencing satisfaction with com-
plete dentures in geriatric patients. J Prosthet Dent 1961 ;11:1019-31.
10. Smith M. Measurement of personality traits and their relation to patient
satisfaction with complete dentures. J Prosthet Dent 1976;35:492-503.
11. Swoope CC. Identification and management of emotional patients. J Pros-
thet Dent 1972;27:434-40.
12. Van Waas MA. The influence of psychological factors on patient satisfac-
tion with complete dentures. J Prosthet Dent 1990;63:545-8.
13. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed.
Hi[]sdale (NJ): Erlbaum; 1988. p. 273-406.
14. Cohen J. A coefficient of agreement for nominal scales. Educ Psycho[
Meas 1960;20:37-46.
15. Kendall MG. Rank correlation methods. London: Griffin; 1948.
16. F[eiss JL. Measuring nominal scale agreement among many raters. Psy-
chol Bull 1971;88:322-8.
1 7. Den Dekker J. Treatment planning in dental practice. [in Dutch] [PhD the-
sis.] Amsterdam: University of Amsterdam; 1990.
18. Brehmer A, Brehmer B. What have we learned about human judgment form
thirty years of policy capturing? In: Brehmer B, Joyce CR, editors. Human
judgment: the SJT view. Amsterdam: North-Holland; 1988. p. 75-114.
Reprint requests to:
DR PIETERKOELE
FACULTYOF PSYCHOLOGY
ROETERSSTRAAT15
1018 WB AMSTERDAM
THE NETHERLANDS
FAX: 31-20-639-0026
Copyright 9 1999 byThe Editorial Council of The Journal of Prosthetic
Dentistry.
0022-3913/99/$8.00 + 0. 10/1/95497
V O L U M E 81 NUMBER 4