Sbu • Statens beredning för medicinsk utvärdering The Swedish Council on Technology Assessment in Health Care

Author

Year,  

reference

Country

Sample, n

Age (years) 

at start

Teeth 

tested,

Caries 

preval-

ence at 

start

Obs 

time 

(yrs)

Drop-out

Explained

Diag-

nostic 

criteria

Exam-

iner (n) 

Relia-

bility

Predictor  

variables

Validating 

criteria

Statist-

ical met-

hods

Results*

Sensitivity and 

specificity

Study 

quality and 

relevance

Comments

Wilson et al

1989 [46]

England

Sample not 

described

n: 83

Age: 11–12

Permanent

Mean

DMFS 3.8

Increment:

Median

new DFS

5.3

3

17

Own

criteria 

unclear

No BW 

(fibre

optics)

No of 

exam-

iners 

unclear

Unclear

Caries experience 

at baseline, salivary 

buffer, M

1–2

S, LB in

saliva, sugar con-

sumption

2 yr:  

≥5 new DFS

3 yr:  

≥8 new DFS

Proportion 

high risk: 

25%

Se, Sp, 

PPV, NPV

No model;

predictors 

tested 

separately

Caries experience 

best

2 yr

Se: 62%; Sp: 71%;

accuracy: 0.69

3 yr

Se: 58%; Sp: 69%;

accuracy: 0.66

MS 2 yr

Se: 71%; Sp: 52%

MS 3 yr

Se: 79%; Sp: 53%

LB 2 yr

Se: 58%; Sp: 63%

LB 3 yr

Se: 79%; Sp: 53%

SB 3 yr

Se: 84%; Sp: 50%

Low

Small sample 

not well

described. 

Incom-

plete data 

reporting

Zhang et al

2006 [35]

China

All children 

from 7 primary 

schools in 

Wuhan city

n: 433

Age: 6–7

Permanent

dmft: 4.1

DMFT

occl

 

0.06

Increment:

Not stated;

28%

developed 

caries 

during the 

observa-

tion period

2

24

Explained

Own

criteria 

No BW 

(fibre

optics)

2 exam-

iners

Yes

>0 d

1–2

mft primary 

molars or >0 d

1–2

mft 

all primary teeth or 

>0 d

1–2

mft including 

initial lesions and 

D

1–2

MFT including

initial lesions (pit  

and fissures)

>0 new

DMFT

Proportion 

high risk: 

27–72%

Se, Sp, 

PPV, NPV,

relative 

risk, A

z

dmft molars >4

Se: 42%; Sp: 78%;

PPV: 0.43; NPV:

0.78

dmft molars >0

Se: 87%; Sp: 34%;

PPV: 0.34; NPV:

0.87

Best for d

1–2

mft  

+ D

1–2

MFT

occl

 

(including enamel).

Not presented in Se, 

Sp though (A

z

=74)

Low

Obvious bias 

from treat-

ment

Table 5.4 continued

Accuracy = Proportion correctly identified; A

z

= Area under a ROC curve; BW = Bitewing; 

CFU = Colony forming units; DFS = Decayed, filled surfaces (permanent teeth); 

dft = Decayed, filled teeth (primary teeth); DFT = Decayed, filled teeth (permanent teeth); 

dmfs = Decayed, missing, filled surfaces (primary teeth); DMFS = Decayed, missing,  

filled surfaces (permanent teeth); DS = Decayed surfaces; LB = Lactobacilli; 

MS = Mutans streptococci; NPV = Negative predictive value; PPV = Positive predictive  

value; OR = Odd ratio; ROC = Receiver operating characteristic; Se = Sensitivity; 

Sp = Specificity; WHO = World Health Organization

269

K A P I T E L   5   •   r I s K b E d ö M n I n g

Author

Year,  

reference

Country

Sample, n

Age (years) 

at start

Teeth 

tested,

Caries 

preval-

ence at 

start

Obs 

time 

(yrs)

Drop-out

Explained

Diag-

nostic 

criteria

Exam-

iner (n) 

Relia-

bility

Predictor  

variables

Validating 

criteria

Statist-

ical met-

hods

Results*

Sensitivity and 

specificity

Study 

quality and 

relevance

Comments

Wilson et al

1989 [46]

England

Sample not 

described

n: 83

Age: 11–12

Permanent

Mean

DMFS 3.8

Increment:

Median

new DFS

5.3

3

17

Own

criteria 

unclear

No BW 

(fibre

optics)

No of 

exam-

iners 

unclear

Unclear

Caries experience 

at baseline, salivary 

buffer, M

1–2

S, LB in

saliva, sugar con-

sumption

2 yr:  

≥5 new DFS

3 yr:  

≥8 new DFS

Proportion 

high risk: 

25%

Se, Sp, 

PPV, NPV

No model;

predictors 

tested 

separately

Caries experience 

best

2 yr

Se: 62%; Sp: 71%;

accuracy: 0.69

3 yr

Se: 58%; Sp: 69%;

accuracy: 0.66

MS 2 yr

Se: 71%; Sp: 52%

MS 3 yr

Se: 79%; Sp: 53%

LB 2 yr

Se: 58%; Sp: 63%

LB 3 yr

Se: 79%; Sp: 53%

SB 3 yr

Se: 84%; Sp: 50%

Low

Small sample 

not well

described. 

Incom-

plete data 

reporting

Zhang et al

2006 [35]

China

All children 

from 7 primary 

schools in 

Wuhan city

n: 433

Age: 6–7

Permanent

dmft: 4.1

DMFT

occl

 

0.06

Increment:

Not stated;

28%

developed 

caries 

during the 

observa-

tion period

2

24

Explained

Own

criteria 

No BW 

(fibre

optics)

2 exam-

iners

Yes

>0 d

1–2

mft primary 

molars or >0 d

1–2

mft 

all primary teeth or 

>0 d

1–2

mft including 

initial lesions and 

D

1–2

MFT including

initial lesions (pit  

and fissures)

>0 new

DMFT

Proportion 

high risk: 

27–72%

Se, Sp, 

PPV, NPV,

relative 

risk, A

z

dmft molars >4

Se: 42%; Sp: 78%;

PPV: 0.43; NPV:

0.78

dmft molars >0

Se: 87%; Sp: 34%;

PPV: 0.34; NPV:

0.87

Best for d

1–2

mft  

+ D

1–2

MFT

occl

 

(including enamel).

Not presented in Se, 

Sp though (A

z

=74)

Low

Obvious bias 

from treat-

ment

Table 5.4 continued

Accuracy = Proportion correctly identified; A

z

= Area under a ROC curve; BW = Bitewing; 

CFU = Colony forming units; DFS = Decayed, filled surfaces (permanent teeth); 

dft = Decayed, filled teeth (primary teeth); DFT = Decayed, filled teeth (permanent teeth); 

dmfs = Decayed, missing, filled surfaces (primary teeth); DMFS = Decayed, missing,  

filled surfaces (permanent teeth); DS = Decayed surfaces; LB = Lactobacilli; 

MS = Mutans streptococci; NPV = Negative predictive value; PPV = Positive predictive  

value; OR = Odd ratio; ROC = Receiver operating characteristic; Se = Sensitivity; 

Sp = Specificity; WHO = World Health Organization

K A R I E S   –   d I A G N O S T I K ,   R I S K B E d ö M N I N G   O C H   I C K E - I N vA S I v   B E H A N d L I N G

270

Table 5.5 Caries prediction in adults. Studies with medium quality and relevance.

Author

Year,  

reference

Country

Sample, n

Age (yrs)  

at start

Teeth, 

surfaces

Obs 

time 

(yrs)

Drop-out Diag-

nostic 

criteria

Exam-

iner (n)

Relia-

bility

Predictor  

variables

Validating 

criteria

Inciden-

ce/incre-

ment

Statis-

tical met-

hods

Results

Sensitivity 

and speci-

ficity

Study 

quality and 

relevance

Comments

Beck et al

1988 [52]

USA

19% of target

population 

Iowa

n: 525

Age: 65+

All root 

surfaces

1.5

15%

Root

caries 

criteria

unclear

Radike

1968?

No BW

4 exam-

iners

Calib- 

rated, 

inter-

examiner 

only

General health 

physical function, 

age, water F,

gingival reces-

sion, periodontal 

pockets, calculus, 

baseline DFS,

DFT, number of

teeth

≥1 vs 0 new

root DFS

or 3 vs 0–2

new root

DFS

Proportion 

high risk: 

Not stated

(After 18

months)

Mean

root DS

incidence: 

men 1.1, 

women

0.95

43% men,

46%

women

developed 

≥1 new

root DS

Discrim- 

inant 

analysis, 

Se, Sp

Model (≥1 vs 

0 new DFS)

All vari-

ables men 

(women)

Se: 75%

(79%); Sp:

79% (83%)

Dental 

variables

Se: 73%

(77%); Sp:

77% (82%);

accuracy: 

0.77–0.82

Most

important 

variables: 

previous 

root caries, 

periodontal 

pockets  

>3 mm and

number of 

teeth (>23

negative 

relation-

ship)

Medium

Incom-

plete data 

reporting

271

K A P I T E L   5   •   r I s K b E d ö M n I n g

Table 5.5 Caries prediction in adults. Studies with medium quality and relevance.

Author

Year,  

reference

Country

Sample, n

Age (yrs)  

at start

Teeth, 

surfaces

Obs 

time 

(yrs)

Drop-out Diag-

nostic 

criteria

Exam-

iner (n)

Relia-

bility

Predictor  

variables

Validating 

criteria

Inciden-

ce/incre-

ment

Statis-

tical met-

hods

Results

Sensitivity 

and speci-

ficity

Study 

quality and 

relevance

Comments

Beck et al

1988 [52]

USA

19% of target

population 

Iowa

n: 525

Age: 65+

All root 

surfaces

1.5

15%

Root

caries 

criteria

unclear

Radike

1968?

No BW

4 exam-

iners

Calib- 

rated, 

inter-

examiner 

only

General health 

physical function, 

age, water F,

gingival reces-

sion, periodontal 

pockets, calculus, 

baseline DFS,

DFT, number of

teeth

≥1 vs 0 new

root DFS

or 3 vs 0–2

new root

DFS

Proportion 

high risk: 

Not stated

(After 18

months)

Mean

root DS

incidence: 

men 1.1, 

women

0.95

43% men,

46%

women

developed 

≥1 new

root DS

Discrim- 

inant 

analysis, 

Se, Sp

Model (≥1 vs 

0 new DFS)

All vari-

ables men 

(women)

Se: 75%

(79%); Sp:

79% (83%)

Dental 

variables

Se: 73%

(77%); Sp:

77% (82%);

accuracy: 

0.77–0.82

Most

important 

variables: 

previous 

root caries, 

periodontal 

pockets  

>3 mm and

number of 

teeth (>23

negative 

relation-

ship)

Medium

Incom-

plete data 

reporting

The table continues on the next page

K A R I E S   –   d I A G N O S T I K ,   R I S K B E d ö M N I N G   O C H   I C K E - I N vA S I v   B E H A N d L I N G

272

Table 5.5 continued

Author

Year,  

reference

Country

Sample, n

Age (yrs)  

at start

Teeth, 

surfaces

Obs 

time 

(yrs)

Drop-out Diag-

nostic 

criteria

Exam-

iner (n)

Relia-

bility

Predictor  

variables

Validating 

criteria

Inciden-

ce/incre-

ment

Statist 

ical met-

hods

Results

Sensitivity 

and speci-

ficity

Study 

quality and 

relevance

Comments

Gilbert et al

2001 [53]

USA

Florida Dental

Care study

n: 723

Age: 45+

All root 

surfaces

2

17%

DFS

RCI

(Katz

1996)

No BW

5 exam-

iners

Yes

Root DFS at

baseline, root 

fillings, regular 

dental attendee, 

number of teeth, 

attachment loss, 

flossing fre-

quency, ability  

to pay, 65 yrs  

or older

≥1 new root

DS or ≥1 

new filling

(F) or both

Proportion 

high risk: 

Not stated

(After  

2 yrs)

Mean inci-

dence:

root 

DS=0.7;

root 

DFS=1.0

36% deve-

loped ≥1 

new DFS

≥1 new

root D:

17%

≥1 new

root F: 

14%

≥1 root 

D+F: 5%

Log reg-

ression, 

OR

Significant

OR for ≥1

new root DS

Active root 

decay at 

baseline: 

3.6

9–16 teeth

at baseline: 

3.0

17–24 teeth

at baseline: 

1.9

≥34% of

teeth with

attach-

ment loss 

+4 mm:

2.9–3.1

Flossing at 

least daily: 

0.3–0.5

Not able  

to pay: 2.5

Medium

Se, Sp not 

reported

273

K A P I T E L   5   •   r I s K b E d ö M n I n g

Table 5.5 continued

Author

Year,  

reference

Country

Sample, n

Age (yrs)  

at start

Teeth, 

surfaces

Obs 

time 

(yrs)

Drop-out Diag-

nostic 

criteria

Exam-

iner (n)

Relia-

bility

Predictor  

variables

Validating 

criteria

Inciden-

ce/incre-

ment

Statist 

ical met-

hods

Results

Sensitivity 

and speci-

ficity

Study 

quality and 

relevance

Comments

Gilbert et al

2001 [53]

USA

Florida Dental

Care study

n: 723

Age: 45+

All root 

surfaces

2

17%

DFS

RCI

(Katz

1996)

No BW

5 exam-

iners

Yes

Root DFS at

baseline, root 

fillings, regular 

dental attendee, 

number of teeth, 

attachment loss, 

flossing fre-

quency, ability  

to pay, 65 yrs  

or older

≥1 new root

DS or ≥1 

new filling

(F) or both

Proportion 

high risk: 

Not stated

(After  

2 yrs)

Mean inci-

dence:

root 

DS=0.7;

root 

DFS=1.0

36% deve-

loped ≥1 

new DFS

≥1 new

root D:

17%

≥1 new

root F: 

14%

≥1 root 

D+F: 5%

Log reg-

ression, 

OR

Significant

OR for ≥1

new root DS

Active root 

decay at 

baseline: 

3.6

9–16 teeth

at baseline: 

3.0

17–24 teeth

at baseline: 

1.9

≥34% of

teeth with

attach-

ment loss 

+4 mm:

2.9–3.1

Flossing at 

least daily: 

0.3–0.5

Not able  

to pay: 2.5

Medium

Se, Sp not 

reported

The table continues on the next page

K A R I E S   –   d I A G N O S T I K ,   R I S K B E d ö M N I N G   O C H   I C K E - I N vA S I v   B E H A N d L I N G

274

Author

Year,  

reference

Country

Sample, n

Age (yrs)  

at start

Teeth, 

surfaces

Obs 

time 

(yrs)

Drop-out Diag-

nostic 

criteria

Exam-

iner (n)

Relia-

bility

Predictor  

variables

Validating 

criteria

Inciden-

ce/incre-

ment

Statist-

ical met-

hods

Results

Sensitivity 

and speci-

ficity

Study 

quality and 

relevance

Takano et al

2003 [54]

Japan

Non-institu-

tionalised  

from Nigata

n: 373

Age: 70+

All root 

surfaces

2

2%

WHO 

1997

No BW

4 exam-

iners

Yes

Baseline DFS,

oral hygiene 

(use of dental

floss or inter-

dental brush),

MS, LB, perio-

dontal pocket 

depth, loss of 

attachment, 

socioeconomics, 

sociodemo-

graphic general 

health, BMI

≥1, 2 or 3 

new root

DS

Proportion 

high risk: 

Not stated

(After  

2 yrs)

mean root 

DS incre-

ment=0.9

36% deve-

loped ≥1 

new root

DFS

Log reg-

ression, 

OR

Model 1: ≥1

new root DS,

sign OR

≥1 baseline 

root DS:

3.7;

≥2 prosthe-

tic crowns:

2.3;

poor oral 

hygiene: 2.1

mean loss 

of attach-

ment >3,6

mm: 2.3

Model 2: ≥2 

new root

DS, sign OR

≥1 baseline 

root DS:

3.4

≥2 prosthe-

tic crowns:

3.1

LB: 2.1

Poor oral 

hygiene: 2.8

Mean loss

of attach-

ment >3,6

mm: 3.0

BMI ≥20: 

2.2

Medium

Se, Sp not 

reported

Table 5.5 continued 

BMI = Body mass index; BW = Bitewing; DFS = Decayed, filled surfaces (permanent teeth);  

DFT = Decayed, filled teeth (permanent teeth); DS = Decayed surfaces; LB = Lactobacilli; 

MS = Mutans streptococci; OR = Odds ratio; Se = Sensitivity; Sp = Specificity

275

K A P I T E L   5   •   r I s K b E d ö M n I n g