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

Part of a larger 

study, selec-

tion unclear

n: 129

Age: 9

Permanent

Mean dmfs

age 9: 9.7

Increment

DMFS age

12: 3.8 

increased 

to 9.3 at 

age 17

3–8

Not 

stated

Möller

1966

BW 

“when

needed”

No of 

exam-

iners 

unclear

Not 

stated

dmfs at age 9

Levels: >8, >14  

and >17

DMFS at

12–17 years

(mean 15.5):

>5, >8, >15

DMFS

Proportion 

high risk: 15, 

25 and 50%

Se, Sp

Best for >14 dmfs

and DMFS >15:

Se about 42%;

Sp about 85%

Low

Large range

of follow-

up years. 

Confounder: 

improvement 

in dental 

health

253

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

Table 5.4 Caries prediction in school children and adolescents.  

Studies with low quality and relevance.

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

Abernathy 

et al

1987 [59]

USA

Representat-

ive of target 

population

Control group 

in prevention 

programme

n: 1 253

Age: 6

n: 1 384

Age: 10

Permanent

Caries 

prevalence 

at start not 

stated

Mean

increment 

not stated

4

52

Analysed

Radike

1968

BW

16 

specially 

trained 

exami-

ners

“Very

consis-

tent with

themsel-

ves”

A total of 11: Age, 

race, sex, socio-

economy, sound 

surfaces, eruption 

status, DMFS, refer-

ral score, fluoride 

in drinking water,

defs molar surfa-

ces, Grainger index 

(general level of

tooth decay scores 

0–5)

DMFS

increment 

(∆DMFS) or

prevalence 

(DMFS) level

not stated

Proportion 

high risk: 

25%

Cor-

relation, 

discri-

minant 

analysis

Log reg-

ression

Se, Sp, 

PPV, NPV

DMFS better than

∆DMFS as outcome.

DMFS

Grade 1

Se: 48%; Sp: 82%

Grade 5

Se: 52%; Sp: 84%

Prediction model 

better for non F 

than for F

Low

Post hoc data 

modelling

Alaluusua 

et al

1987 [70]

Finland

Part of a larger 

study, selec-

tion unclear

n: 129

Age: 9

Permanent

Mean dmfs

age 9: 9.7

Increment

DMFS age

12: 3.8 

increased 

to 9.3 at 

age 17

3–8

Not 

stated

Möller

1966

BW 

“when

needed”

No of 

exam-

iners 

unclear

Not 

stated

dmfs at age 9

Levels: >8, >14  

and >17

DMFS at

12–17 years

(mean 15.5):

>5, >8, >15

DMFS

Proportion 

high risk: 15, 

25 and 50%

Se, Sp

Best for >14 dmfs

and DMFS >15:

Se about 42%;

Sp about 85%

Low

Large range

of follow-

up years. 

Confounder: 

improvement 

in dental 

health

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

254

Table 5.4 continued

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

Alaluusua 

et al

1990 [42]

Finland

Sample not 

specified

n: 122

Age: 12–17

Permanent

Mean DFS:

6.7

Mean DS:

0.8

Increment

Mean new

DFS: 2.6

3

16

Möller

1966

BW

No of 

exam-

iners 

unclear

Not 

stated

Baseline caries 

prevalence salivary 

buffer capacity, 

secretion rate, MS,

LB in saliva

>13 yrs:  

≥10 DFS

13–15 yrs:

≥11 DFS

>15 yrs:  

≥12 DFS

Proportion 

high risk: 

25%

Se, Sp

Baseline DFS best

single predictor

Se: 61%; Sp: 82%;

PPV: 0.54; NPV:

0.86

DFS+MS

Se: 71%; Sp: 79%

DFS+LB

Se: 84%; Sp: 62%

Little value of adding

MS and LB. Salivary

buffer capacity, 

salivary secretion 

rate did not improve 

the model

Low

Unspecified 

sample

Alaluusua 

et al

1993 [71]

Finland

Same as 1990

n: 122

Age: 12–17

Permanent

DFS: 6.7

Mean DS:

0.8

Increment

Mean new

DFS: 2.6

3

16

Möller

1966

BW

No of 

exam-

iners 

unclear

Not 

stated

Baseline DFS, MS,  

LB in saliva

Categorised at 2–3

levels

DFS cut-off

correspond-

ing to MS

and LB

Proportion 

high risk: 15, 

25 and 50%

Se, Sp, 

PPV, NPV

No improvement 

in accuracy compa-

red with previous

results

Same mater-

ial as Ala-

luusua 1990. 

Unspecified 

sample

David et al

2006 [72]

Norway

All children  

at 3 public  

dental clinics  

in Bergen.  

High caries 

prevalence

n: 112

Age: 12

Permanent

Mean

DMFS: 2.7

DMFT: 1.9

Increment:

Mean new

DMFS: 6.0

6

30

Explained, 

analysed

Espelid 

et al 

1990

BW

5 exa-

miners 

at start, 

1 at end 

(not the

same)

Yes

Baseline D

en

FS

appr

 

(enamel and dentine)

>0 of premolars or

2nd molars

>0 to >4

new DFS

of premo-

lars or 2nd 

molars

Proportion 

high risk: 

Premolars 

19%, 2nd

molars 11%

Se, Sp, 

PPV, NPV

>4 new DFS best

Se: 62–85% (pre- 

molar, 2nd molars);

Sp: 90–96%

Low

Retrospect-

ive. Only  

1 predictor

255

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

Table 5.4 continued

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

Alaluusua 

et al

1990 [42]

Finland

Sample not 

specified

n: 122

Age: 12–17

Permanent

Mean DFS:

6.7

Mean DS:

0.8

Increment

Mean new

DFS: 2.6

3

16

Möller

1966

BW

No of 

exam-

iners 

unclear

Not 

stated

Baseline caries 

prevalence salivary 

buffer capacity, 

secretion rate, MS,

LB in saliva

>13 yrs:  

≥10 DFS

13–15 yrs:

≥11 DFS

>15 yrs:  

≥12 DFS

Proportion 

high risk: 

25%

Se, Sp

Baseline DFS best

single predictor

Se: 61%; Sp: 82%;

PPV: 0.54; NPV:

0.86

DFS+MS

Se: 71%; Sp: 79%

DFS+LB

Se: 84%; Sp: 62%

Little value of adding

MS and LB. Salivary

buffer capacity, 

salivary secretion 

rate did not improve 

the model

Low

Unspecified 

sample

Alaluusua 

et al

1993 [71]

Finland

Same as 1990

n: 122

Age: 12–17

Permanent

DFS: 6.7

Mean DS:

0.8

Increment

Mean new

DFS: 2.6

3

16

Möller

1966

BW

No of 

exam-

iners 

unclear

Not 

stated

Baseline DFS, MS,  

LB in saliva

Categorised at 2–3

levels

DFS cut-off

correspond-

ing to MS

and LB

Proportion 

high risk: 15, 

25 and 50%

Se, Sp, 

PPV, NPV

No improvement 

in accuracy compa-

red with previous

results

Same mater-

ial as Ala-

luusua 1990. 

Unspecified 

sample

David et al

2006 [72]

Norway

All children  

at 3 public  

dental clinics  

in Bergen.  

High caries 

prevalence

n: 112

Age: 12

Permanent

Mean

DMFS: 2.7

DMFT: 1.9

Increment:

Mean new

DMFS: 6.0

6

30

Explained, 

analysed

Espelid 

et al 

1990

BW

5 exa-

miners 

at start, 

1 at end 

(not the

same)

Yes

Baseline D

en

FS

appr

 

(enamel and dentine)

>0 of premolars or

2nd molars

>0 to >4

new DFS

of premo-

lars or 2nd 

molars

Proportion 

high risk: 

Premolars 

19%, 2nd

molars 11%

Se, Sp, 

PPV, NPV

>4 new DFS best

Se: 62–85% (pre- 

molar, 2nd molars);

Sp: 90–96%

Low

Retrospect-

ive. Only  

1 predictor

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

256

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 me-

thods

Results*

Sensitivity and 

specificity

Study 

quality and 

relevance

Comments

de Liefde

1989 [73]

New  

Zealand

All children in 

Marlborough

1982, effect of 

programme

n: 921

Age: 5–8,

included  

stepwise

Permanent

5–8

13

Own

criteria

No cal-

ibration

No BW

Not 

clear 

who

perfor-

med  

the dia-

gnosis

Several 

exam-

iners

No

2-step prediction

1. Age 5: Cut-off  

for high risk ≥dmft

2. 2 yrs after erup-

tion of permanent 

1st molars: cutt of 

for high risk DFT

occl 

≥1

DMFT ≥3

DMFS ≥4

Proportion 

high risk: 

Not stated

Se, Sp

For ≥3 DMFT

Se: 78%; Sp: 88%;

PPV: 0.51; NPV:

0.96; accuracy: 0.87

For ≥4 DMFT

Se: 85%; Sp: 85%;

PPV: 0.37; NPV:

0.98; accuracy: 0.85

Low

Heterogen-

eous sample. 

Incom-

plete data 

reporting

Hausen et al

2000 [43]

Finland

Random

sample of 

75% of target

population in 

Vantaa, Finland

n: 583

Age: 13 (mean)

effect of pro-

gramme

Permanent, 

all surfaces

Mean

DMFS in

high risk 

group 3.8;

low risk

group 1.8 

(estimated)

Increment:

Mean new

DMFS:

High risk 

4.4–5.1,

low risk

2.0

3

30

Not ana-

lysed

Möller

1973

BW

4 expe-

rienced 

exam-

iners

Yes, 

except 

for BW 

(read by

an expe-

rienced 

exam-

iner)

Estimated no of new

fillings after 1 yr =

≥2 or caries expe-

rience = >1 DFS or

>0 DFS approximal

or salivary flow ≤0.7 

ml/min and buffer

capacity =1 or

combined MS and LB

in saliva or combined 

DFS and MS, LB in

saliva

≥1 new DFS

≥2 new DFS

Proportion 

high risk: 1/3

Se, Sp, 

PPV,

NPV*

For ≥1 new DFS

Se: 54%; Sp: 70%;

PPV: 0.71; NPV:

0.52

For ≥2 new DFS

Se: 58%; Sp: 67%;

PPV: 0.57; NPV:

0.68*

Low

Bias from 

treatment 

likely. Effect 

of different 

predictors 

cannot be 

evaluated

Table 5.4 continued

257

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 me-

thods

Results*

Sensitivity and 

specificity

Study 

quality and 

relevance

Comments

de Liefde

1989 [73]

New  

Zealand

All children in 

Marlborough

1982, effect of 

programme

n: 921

Age: 5–8,

included  

stepwise

Permanent

5–8

13

Own

criteria

No cal-

ibration

No BW

Not 

clear 

who

perfor-

med  

the dia-

gnosis

Several 

exam-

iners

No

2-step prediction

1. Age 5: Cut-off  

for high risk ≥dmft

2. 2 yrs after erup-

tion of permanent 

1st molars: cutt of 

for high risk DFT

occl 

≥1

DMFT ≥3

DMFS ≥4

Proportion 

high risk: 

Not stated

Se, Sp

For ≥3 DMFT

Se: 78%; Sp: 88%;

PPV: 0.51; NPV:

0.96; accuracy: 0.87

For ≥4 DMFT

Se: 85%; Sp: 85%;

PPV: 0.37; NPV:

0.98; accuracy: 0.85

Low

Heterogen-

eous sample. 

Incom-

plete data 

reporting

Hausen et al

2000 [43]

Finland

Random

sample of 

75% of target

population in 

Vantaa, Finland

n: 583

Age: 13 (mean)

effect of pro-

gramme

Permanent, 

all surfaces

Mean

DMFS in

high risk 

group 3.8;

low risk

group 1.8 

(estimated)

Increment:

Mean new

DMFS:

High risk 

4.4–5.1,

low risk

2.0

3

30

Not ana-

lysed

Möller

1973

BW

4 expe-

rienced 

exam-

iners

Yes, 

except 

for BW 

(read by

an expe-

rienced 

exam-

iner)

Estimated no of new

fillings after 1 yr =

≥2 or caries expe-

rience = >1 DFS or

>0 DFS approximal

or salivary flow ≤0.7 

ml/min and buffer

capacity =1 or

combined MS and LB

in saliva or combined 

DFS and MS, LB in

saliva

≥1 new DFS

≥2 new DFS

Proportion 

high risk: 1/3

Se, Sp, 

PPV,

NPV*

For ≥1 new DFS

Se: 54%; Sp: 70%;

PPV: 0.71; NPV:

0.52

For ≥2 new DFS

Se: 58%; Sp: 67%;

PPV: 0.57; NPV:

0.68*

Low

Bias from 

treatment 

likely. Effect 

of different 

predictors 

cannot be 

evaluated

Table 5.4 continued

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

258

Table 5.4 continued

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