Sbu • Statens beredning för medicinsk utvärdering The Swedish Council on Technology Assessment in Health Care
Yüklə 4,3 Mb. Pdf görüntüsü
|
|
Obs time (yrs) Drop- out (%) Explained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical methods Results Sensitivity and specificity Study quality and relevance Comments Sullivan et al 1989 [24] Sweden
2 dental clinics in Malmö
n: 89 Age: 5
Primary dmfs: 4.4 2 15
WHO 1987
BW No of
exam- iners
unclear Yes
Oral hygiene (gingival status), salivary secretion and buffer, MS, LB in saliva >0 new caries lesion Proportion high risk: Not stated Se, Sp,
PPV, NPV Best single predictor MS Se: 41%; Sp: 83% Best combination of predictors Gingival status + MS Se: 53%; Sp: 74% Salivary secretion and buffer did not improve the model Low
Small sample not described. Bias from treatment Thibodeau et al 1995 [25] USA Children enrolled in programme; poor areas in Connecticut n: 148 Age: Mean 3.8, range 2.0–5.3 Primary
Low, moderate and high dmfs:
Mean: 0.15, 1.44, 3.36 2 68
1972 No BW
2 exam- iners
Yes Baseline dmfs, MS in saliva >2 dmfs or 3–4 dmfs, ≥5 dmfs
Proportion high risk: 36% Se, Sp,
PPV, NPV Best for ≥5 dmfs increment and high MS group at baseline Se: 72%; Sp: 72%; PPV: 0.34; NPV: 0.93 Low
Large drop- out. Het- erogeneous sample
Tsubouchi et al
1995 [69] Japan
Not clear n: 100
Age: 1.5 Primary
dft not stated
1.5 Not
stated Own
Not stated
“Caries activity”; pH assessment; Cariostat ≥1 new
lesion Proportion high risk: 50%
Se, Sp, PPV, NPV
18–36 months Se: 61%; Sp: 67%; PPV: 81%; NPV: 43% Low
Sampling unclear. Attrition unclear
Twetman et al 1996 [26] Sweden All available in areas with different f-exposure Halland
n: 1 022 Age: 4–5
Primary dfs 0.2 –1.0 2 Not
stated WHO
1987 No BW
Not stated
Caries prevalence at baseline MS in saliva; single or repeated >2 new dfs Proportion high risk: 13–28%
Se, Sp, PPV, NPV
F-exposure influ- enced Se and Sp. Combined baseline caries and MS: Low F-exposure best Se: 65%; Sp: 86% No improvement from repeated MS- tests Low
Incomplete data repor- ted
237 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 (yrs) 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 methods Results Sensitivity and specificity Study quality and relevance Comments Sullivan et al 1989 [24] Sweden
2 dental clinics in Malmö
n: 89 Age: 5
Primary dmfs: 4.4 2 15
WHO 1987
BW No of
exam- iners
unclear Yes
Oral hygiene (gingival status), salivary secretion and buffer, MS, LB in saliva >0 new caries lesion Proportion high risk: Not stated Se, Sp,
PPV, NPV Best single predictor MS Se: 41%; Sp: 83% Best combination of predictors Gingival status + MS Se: 53%; Sp: 74% Salivary secretion and buffer did not improve the model Low
Small sample not described. Bias from treatment Thibodeau et al 1995 [25] USA Children enrolled in programme; poor areas in Connecticut n: 148 Age: Mean 3.8, range 2.0–5.3 Primary
Low, moderate and high dmfs:
Mean: 0.15, 1.44, 3.36 2 68
1972 No BW
2 exam- iners
Yes Baseline dmfs, MS in saliva >2 dmfs or 3–4 dmfs, ≥5 dmfs
Proportion high risk: 36% Se, Sp,
PPV, NPV Best for ≥5 dmfs increment and high MS group at baseline Se: 72%; Sp: 72%; PPV: 0.34; NPV: 0.93 Low
Large drop- out. Het- erogeneous sample
Tsubouchi et al
1995 [69] Japan
Not clear n: 100
Age: 1.5 Primary
dft not stated
1.5 Not
stated Own
Not stated
“Caries activity”; pH assessment; Cariostat ≥1 new
lesion Proportion high risk: 50%
Se, Sp, PPV, NPV
18–36 months Se: 61%; Sp: 67%; PPV: 81%; NPV: 43% Low
Sampling unclear. Attrition unclear
Twetman et al 1996 [26] Sweden All available in areas with different f-exposure Halland
n: 1 022 Age: 4–5
Primary dfs 0.2 –1.0 2 Not
stated WHO
1987 No BW
Not stated
Caries prevalence at baseline MS in saliva; single or repeated >2 new dfs Proportion high risk: 13–28%
Se, Sp, PPV, NPV
F-exposure influ- enced Se and Sp. Combined baseline caries and MS: Low F-exposure best Se: 65%; Sp: 86% No improvement from repeated MS- tests Low
Incomplete data repor- ted
filled teeth (permanent teeth); LB = Lactobacilli; MS = Mutans streptococci; NPV = Negative predictive value; NS = Not significant; OR = Odds ratio; PPV = Positive predictive value; RR = Relative risk; 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 238
Table 5.3 Caries prediction in school children and adolescents. Studies with high or medium quality and relevance. Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Disney et al 1992 [11] USA
Target popu- lation Aiken, South Carolina and Portland, Maine, USA n: 2 185 (Grade 1), 1 932
(Grade 5) Age: 6 and 10 Permanent
Aiken (Port- land) Age 6:
0.3 (0.2) Age 10: 3.0 (1.7)
Age 6:
9.3 (2.9) Age 10: 4.4 (2.4)
Age 6:
1.9 (0.8) Age 10: 3.1 (1.5) 3 19–22 Not ana- lysed
Radike 1968
No BW (fibre
optics) 4 exam-
iners Yes
A total of 22–23 variables Baseline dmfs, DMFS, subjec- tively predicted caries (“gut feeling”), sound permanent sur- faces, white spot lesions, sealants, fissure morpho- logy, MS, LB in saliva (cut-off not stated), plaque score, between meal snacks (sugar, chips or cereal), oral hygiene (brush
fluoride mouth rinse, fluoride tablets, fluoros, age, race, gender, general health, dental visits last year, education of household, combinations of variables New DMFS: Grade 1 ≥2 Grade 5 ≥3 Proportion high risk: 25% Log
regres- sion,
Se, Sp, PPV,
NPV Best model for combined fis- sure morphology, subjectively predicted caries (gut feeling) and DMFS:
Se: 59%; Sp: 84% Grade 5 Se: 62%; Sp: 83%; PPV: 0.49; NPV: 0.88 MS and LB contributed little to the model; oral hygiene and misuse of sugar not significant in the model High
Bias from treatment possible (fissure sealing)
239 K A P I T E L 5 • r I s K b E d ö M n I n g Table 5.3 Caries prediction in school children and adolescents. Studies with high or medium quality and relevance. Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Disney et al 1992 [11] USA
Target popu- lation Aiken, South Carolina and Portland, Maine, USA n: 2 185 (Grade 1), 1 932
(Grade 5) Age: 6 and 10 Permanent
Aiken (Port- land) Age 6:
0.3 (0.2) Age 10: 3.0 (1.7)
Age 6:
9.3 (2.9) Age 10: 4.4 (2.4)
Age 6:
1.9 (0.8) Age 10: 3.1 (1.5) 3 19–22 Not ana- lysed
Radike 1968
No BW (fibre
optics) 4 exam-
iners Yes
A total of 22–23 variables Baseline dmfs, DMFS, subjec- tively predicted caries (“gut feeling”), sound permanent sur- faces, white spot lesions, sealants, fissure morpho- logy, MS, LB in saliva (cut-off not stated), plaque score, between meal snacks (sugar, chips or cereal), oral hygiene (brush
fluoride mouth rinse, fluoride tablets, fluoros, age, race, gender, general health, dental visits last year, education of household, combinations of variables New DMFS: Grade 1 ≥2 Grade 5 ≥3 Proportion high risk: 25% Log
regres- sion,
Se, Sp, PPV,
NPV Best model for combined fis- sure morphology, subjectively predicted caries (gut feeling) and DMFS:
Se: 59%; Sp: 84% Grade 5 Se: 62%; Sp: 83%; PPV: 0.49; NPV: 0.88 MS and LB contributed little to the model; oral hygiene and misuse of sugar not significant in the model High
Bias from treatment possible (fissure sealing)
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 240
Table 5.3 continued Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Beck et al 1992 [10] USA
Target popu- lation Aiken, South Carolina and Portland, Maine, USA n: 2 185 (Grade 1), 1 932
(Grade 5) Age: 6
(Grade 1), 10 (Grade 5) Permanent
Aiken
(Portland) Age 6:
0.3 (0.2) Age 10: 3.0 (1.7)
Age 6:
9.3 (2.9) Age 10: 4.4 (2.4)
Age 6:
1.9 (0.8) Age 10:
3.1 (1.5) 3 19–22 Not ana- lysed
Radike 1968
No BW (fibre
optics) 4 exam-
iners Yes
Included significant predictors 1. Any risk predictor model (grade 1: edu- cation – head of house-hold, use fluoride tablets, LB in saliva, gut feeling) (grade 5: baseline DMFS, white spot lesions, sound permanent sur- faces, LB in saliva, gut feeling)
(including both etiological and other risk fac- tors)
(including only etiological fac- tors; tooth mor- phology, sound permanent tooth surfaces, mean plaque score, MS, LB in saliva)
≥1 new
DMFS 2. High risk Aiken: 4–5 new DMFS Portland: 2–3 new DMFS
Proportion high risk: Not stated Log
regres- sion,
Se, Sp, PPV,
NPV Any risk prediction model Aiken Se: 80–84%; Sp: 54–61%; PPV: 0.73–0.78; NPV: 0.64–0.69 Portland Se: 66–76%; Sp: 71–78%; PPV: 0.57–0.68; NPV: 0.78–0.84
Se: 59–62%; Sp: 81–83%; PPV: 0.48–51; NPV: 0.86–88
Se: 59–62%; Sp: 84%; PPV: 0.47–0.5; NPV: 0.89–0.9
Se: 74–81%; Sp: 50–55%; PPV: 0.68–0.75; NPV: 0.58–0.62 Portland Se: 59–69%; Sp: 65–74%; PPV: 0.51–0.62; NPV: 0.71–0.8 Se+Sp ≤147% in any of the models Any risk prediction model gave highest Se (80–84%) and PPV
High risk prediction model gave the highest Sp (81–84%) and NPV Same
material as Disney 1992 [11] Modelling data
241 K A P I T E L 5 • r I s K b E d ö M n I n g Table 5.3 continued Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Beck et al 1992 [10] USA
Target popu- lation Aiken, South Carolina and Portland, Maine, USA n: 2 185 (Grade 1), 1 932
(Grade 5) Age: 6
(Grade 1), 10 (Grade 5) Permanent
Aiken
(Portland) Age 6:
0.3 (0.2) Age 10: 3.0 (1.7)
Age 6:
9.3 (2.9) Age 10: 4.4 (2.4)
Age 6:
1.9 (0.8) Age 10:
3.1 (1.5) 3 19–22 Not ana- lysed
Radike 1968
No BW (fibre
optics) 4 exam-
iners Yes
Included significant predictors 1. Any risk predictor model (grade 1: edu- cation – head of house-hold, use fluoride tablets, LB in saliva, gut feeling) (grade 5: baseline DMFS, white spot lesions, sound permanent sur- faces, LB in saliva, gut feeling)
(including both etiological and other risk fac- tors)
(including only etiological fac- tors; tooth mor- phology, sound permanent tooth surfaces, mean plaque score, MS, LB in saliva)
≥1 new
DMFS 2. High risk Aiken: 4–5 new DMFS Portland: 2–3 new DMFS
Proportion high risk: Not stated Log
regres- sion,
Se, Sp, PPV,
NPV Any risk prediction model Aiken Se: 80–84%; Sp: 54–61%; PPV: 0.73–0.78; NPV: 0.64–0.69 Portland Se: 66–76%; Sp: 71–78%; PPV: 0.57–0.68; NPV: 0.78–0.84
Se: 59–62%; Sp: 81–83%; PPV: 0.48–51; NPV: 0.86–88
Se: 59–62%; Sp: 84%; PPV: 0.47–0.5; NPV: 0.89–0.9
Se: 74–81%; Sp: 50–55%; PPV: 0.68–0.75; NPV: 0.58–0.62 Portland Se: 59–69%; Sp: 65–74%; PPV: 0.51–0.62; NPV: 0.71–0.8 Se+Sp ≤147% in any of the models Any risk prediction model gave highest Se (80–84%) and PPV
High risk prediction model gave the highest Sp (81–84%) and NPV Same
material as Disney 1992 [11] Modelling data
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 242
Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Stewart et al 1991 [32] USA
Target popu- lation Aiken, South Carolina and Portland, Maine, USA n: 2 185 (Grade 1), 1 932
(Grade 5) Age: 6
(Grade 1) Permanent Baseline mean DMFS Aiken
(Portland) Age 6:
0.3 (0.2) Age 10: 3.0 (1.7)
Age 6:
9.3 (2.9) Age 10: 4.4 (2.4)
Age 6:
1.9 (0.8) Age 10: 3.1 (1.5) 2 19–22 Not ana- lysed
Radike 1968
No BW (fibre
optics) 4 exam-
iners Yes
The model Clas- sification and Regression Tree (CART) decides the appropriate number of vari- ables Portland (low caries preval- ence): dmfs, gut feeling Aiken (high caries prevalence): 9 variables: age, urgency, 4 differ- ent dmfs scores, no of erupted 1st permanent molars, morpho- logy 1st per- manent molars, DMFS
≥2 DMFS Proportion high risk: Not stated Se, Sp
Se: 62%; Sp: 77% Aiken (9 predictors) Se: 64%; Sp: 86% dmfs common predictor for both areas Same material
as Disney 1992 [11] Modelling data
Table 5.3 continued 243 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 (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Stewart et al 1991 [32] USA
Target popu- lation Aiken, South Carolina and Portland, Maine, USA n: 2 185 (Grade 1), 1 932
(Grade 5) Age: 6
(Grade 1) Permanent Baseline mean DMFS Aiken
(Portland) Age 6:
0.3 (0.2) Age 10: 3.0 (1.7)
Age 6:
9.3 (2.9) Age 10: 4.4 (2.4)
Age 6:
1.9 (0.8) Age 10: 3.1 (1.5) 2 19–22 Not ana- lysed
Radike 1968
No BW (fibre
optics) 4 exam-
iners Yes
The model Clas- sification and Regression Tree (CART) decides the appropriate number of vari- ables Portland (low caries preval- ence): dmfs, gut feeling Aiken (high caries prevalence): 9 variables: age, urgency, 4 differ- ent dmfs scores, no of erupted 1st permanent molars, morpho- logy 1st per- manent molars, DMFS
≥2 DMFS Proportion high risk: Not stated Se, Sp
Se: 62%; Sp: 77% Aiken (9 predictors) Se: 64%; Sp: 86% dmfs common predictor for both areas Same material
as Disney 1992 [11] Modelling data
Table 5.3 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 244
Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Russell et al 1991 [36] Scotland
Scottish ado- lescents; from mixed indu- strial and rural areas. Tooth paste trial n: 372 Mean age: 12.6 (11–14) Permanent Baseline mean DMFS: 10 (including radiographic diagnosis) Increment: Mean new DMFS: 4.9 2 23
1978 BW 2 exam- iners Yes
Baseline DMFS, plaque score, MS, LB, Can- dida, Veilonella in saliva, Snyder colorimetric test, salivary buffer (SB), salivary flow rate (SR) >14 DMFS
>2 DS ≥10
6 MS ≥10 5 LB ≥10 2 candida
Proportion high risk: 25–42% Corre-
lation Se, Sp
Best model for combined DMFS, LB and Veilonella Se: 71% (95% CI 61–78); Sp: 74% (95% CI 70–78) (proportion high risk: 39%) Single best predictor DS score Se: 54%; Sp: 72% (pro- portion high risk: 38%)
Se: 40%; Sp: 83% (pro- portion high risk: 25%)
Se: 48%; Sp: 68% (pro- portion high risk: 38%) Salivary flow rate not signi- ficant. Salivary flow rate, salivary buffer did not contribute to the model BW added little to the prediction compared with clinical exam only High Possible bias from fluoride toothpaste use
Table 5.3 continued 245 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 (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Russell et al 1991 [36] Scotland
Scottish ado- lescents; from mixed indu- strial and rural areas. Tooth paste trial n: 372 Mean age: 12.6 (11–14) Permanent Baseline mean DMFS: 10 (including radiographic diagnosis) Increment: Mean new DMFS: 4.9 2 23
1978 BW 2 exam- iners Yes
Baseline DMFS, plaque score, MS, LB, Can- dida, Veilonella in saliva, Snyder colorimetric test, salivary buffer (SB), salivary flow rate (SR) >14 DMFS
>2 DS ≥10
6 MS ≥10 5 LB ≥10 2 candida
Proportion high risk: 25–42% Corre-
lation Se, Sp
Best model for combined DMFS, LB and Veilonella Se: 71% (95% CI 61–78); Sp: 74% (95% CI 70–78) (proportion high risk: 39%) Single best predictor DS score Se: 54%; Sp: 72% (pro- portion high risk: 38%)
Se: 40%; Sp: 83% (pro- portion high risk: 25%)
Se: 48%; Sp: 68% (pro- portion high risk: 38%) Salivary flow rate not signi- ficant. Salivary flow rate, salivary buffer did not contribute to the model BW added little to the prediction compared with clinical exam only High Possible bias from fluoride toothpaste use
Table 5.3 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 246
Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Vanobbergen et al 2001 [34] The Nether- lands
7% of target population n: 3002 Age: 7
Permanent 1st molars Baseline dft/DFT not stated Increment: 68% had 0 new DMFS
3 19 WHO No BW 16 exam-
iners Yes
Baseline dmfs, plaque, dietary habits: Daily use of sugar-containing drinks, between- meal snacks, start tooth brushing, tooth brushing frequency, socio- demographic factors (province, education) >2 new
DMFS in permanent 1st molars Proportion high risk: 10–50%
Log regres-
sion, Se, Sp,
ROC Model (almost all predictive power came from baseline dmfs) For 30% high risk proportion Se: 59% (95% CI 55–63); Sp: 73% (95% CI 71–75)
Se: 51%; Sp: 79% Oral hygiene (brush <1/day) significant in the model mis- use of sugar not significant in the model ROC (A z
dmfs): 0.72 (0.69) High
Possible bias from fluoride toothpaste use
247 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 (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Vanobbergen et al 2001 [34] The Nether- lands
7% of target population n: 3002 Age: 7
Permanent 1st molars Baseline dft/DFT not stated Increment: 68% had 0 new DMFS
3 19 WHO No BW 16 exam-
iners Yes
Baseline dmfs, plaque, dietary habits: Daily use of sugar-containing drinks, between- meal snacks, start tooth brushing, tooth brushing frequency, socio- demographic factors (province, education) >2 new
DMFS in permanent 1st molars Proportion high risk: 10–50%
Log regres-
sion, Se, Sp,
ROC Model (almost all predictive power came from baseline dmfs) For 30% high risk proportion Se: 59% (95% CI 55–63); Sp: 73% (95% CI 71–75)
Se: 51%; Sp: 79% Oral hygiene (brush <1/day) significant in the model mis- use of sugar not significant in the model ROC (A z
dmfs): 0.72 (0.69) High
Possible bias from fluoride toothpaste use
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 248
Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Burt et al 1994 [40] USA
All children in Coldwater, Michigan Low caries prevalence population n: 499 Age: 10–15 Permanent All surfaces, proximal and pit & fissures separately Mean DMFS: boys (girls) Age 10: 0.4 (2.4) Age 13: 4.1 (5.5) Age 15: 4.9 (6.0) Increment Mean new DMFS=2.7 (boys) 3.1 (girls), NS 30% had 0 new DMFS, 28% had 1–2 new DMFS, 20% had ≥1 new approx- imal DMFS 3 33 Explained, analysed
NIDR 1989
No BW 2 experi- enced exam-
iners Yes
Total sugar intake, % sugar of total energy intake, between meal sugar con- sumption (% of total energy intake) age, gender ≥1 new
DMFS all surfaces, proximal and pit & fissures separately Proportion high risk: 25%
RR for highest
versus lowest
quartile of sugar con- sump-
tion Caries incidence poorly rela- ted to sugar intake, whether measured as total daily amount, between meal intake or frequency of consumption Consumption of sugar as % of total energy: No difference between age and gender Average total sugar intake related to caries increment: no pattern Highest quartile Total sugar (≥30% of total energy/day) RR Any DMFS: 1.2 Any DMFS appr
: 1.8 Any DMFS
occl : 1.2 (NS) Between meal sugar (as % of total energy/day) RR Any DMFS: 1.2 (NS) Any DMFS appr
: 1.7 Any DMFS
occl : 1.1 (NS) Medium Important predic- tors not included. Only RR
reported Table 5.3 continued 249 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 (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Burt et al 1994 [40] USA
All children in Coldwater, Michigan Low caries prevalence population n: 499 Age: 10–15 Permanent All surfaces, proximal and pit & fissures separately Mean DMFS: boys (girls) Age 10: 0.4 (2.4) Age 13: 4.1 (5.5) Age 15: 4.9 (6.0) Increment Mean new DMFS=2.7 (boys) 3.1 (girls), NS 30% had 0 new DMFS, 28% had 1–2 new DMFS, 20% had ≥1 new approx- imal DMFS 3 33 Explained, analysed
NIDR 1989
No BW 2 experi- enced exam-
iners Yes
Total sugar intake, % sugar of total energy intake, between meal sugar con- sumption (% of total energy intake) age, gender ≥1 new
DMFS all surfaces, proximal and pit & fissures separately Proportion high risk: 25%
RR for highest
versus lowest
quartile of sugar con- sump-
tion Caries incidence poorly rela- ted to sugar intake, whether measured as total daily amount, between meal intake or frequency of consumption Consumption of sugar as % of total energy: No difference between age and gender Average total sugar intake related to caries increment: no pattern Highest quartile Total sugar (≥30% of total energy/day) RR Any DMFS: 1.2 Any DMFS appr
: 1.8 Any DMFS
occl : 1.2 (NS) Between meal sugar (as % of total energy/day) RR Any DMFS: 1.2 (NS) Any DMFS appr
: 1.7 Any DMFS
occl : 1.1 (NS) Medium Important predic- tors not included. Only RR
reported Table 5.3 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 250
Author Year, reference Country Sample, n Age (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Stenlund et al 2002 [38] Sweden
Target popula- tion of children in a Swedish community with mixes socioeconomy Low caries prevalence n: 432 Age: 12–13 Permanent Approximal surfaces 4 days to 7 months
DMFT: 3.2 DMFS: 3.6 DMF
appr : 0.6
9 19 Analysed Own radio-
graphic criteria
2 exam- iners
Yes Bitewing
Approximal caries at baseline ≥1 new approximal enamel or dentin lesion Proportion high risk: Not applic- able
RR Subject level Number of approximal caries lesions at baseline and corresponding RR of new approximal lesions: 0/1, 2/2.06, 3/3.55, 4–8/3.62, >8/4.85 Medium
Important confoun-
ders not included. Only rel- ative risk reported
BW = Bitewing; 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); DMFT = Decayed, missing, filled teeth (permanent teeth); DS = Decayed surfaces; LB = Lactobacilli; MS = Mutans streptococci; NPV = Negative predictive value; NS = Not significant; PPV = Positive predictive value; ROC = Receiver operating characteristic; RR = Relative risk; Se = Sensitivity; Sp = Specificity; WHO = World Health Organization
251 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 (yrs) at start Teeth tested, caries pre- valence at start, caries increment/ incidence during study period Obs time (yrs) Drop- out (%) Ex- plained Diag- nostic criteria Exam- iner (n) Relia- bility Predictor variables Validating criteria Statist- ical met- hods Results Sensitivity and specificity Study quality and rele- vance Com- ments Stenlund et al 2002 [38] Sweden
Target popula- tion of children in a Swedish community with mixes socioeconomy Low caries prevalence n: 432 Age: 12–13 Permanent Approximal surfaces 4 days to 7 months
DMFT: 3.2 DMFS: 3.6 DMF
appr : 0.6
9 19 Analysed Own radio-
graphic criteria
2 exam- iners
Yes Bitewing
Approximal caries at baseline ≥1 new approximal enamel or dentin lesion Proportion high risk: Not applic- able
RR Subject level Number of approximal caries lesions at baseline and corresponding RR of new approximal lesions: 0/1, 2/2.06, 3/3.55, 4–8/3.62, >8/4.85 Medium
Important confoun-
ders not included. Only rel- ative risk reported
BW = Bitewing; 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); DMFT = Decayed, missing, filled teeth (permanent teeth); DS = Decayed surfaces; LB = Lactobacilli; MS = Mutans streptococci; NPV = Negative predictive value; NS = Not significant; PPV = Positive predictive value; ROC = Receiver operating characteristic; RR = Relative risk; 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 252
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.
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
Yüklə 4,3 Mb. Dostları ilə paylaş:
|