Abstract
Accurate age estimation is an integral part of the identification process. Although used infrequently when compared to more established methods, the Haavikko method can be used in cases where other dental age estimation methods have proven ineffective.The aim of this study was to adapt the Haavikko method as a means of improving age estimation on a representative sample of Croatian children and to establish an applicable standard for the Haavikko developmental stages. To achieve this objective, digital, standardized orthopantomograms of children aged 5 to 16 were collected in four Croatian cities in 1997. Drawing upon a previously published study of the Croatian population, a simple adaptation named ‘shift’ was introduced to the Haavikko method by adding the average difference between chronological and dental age to the estimated dental age. Square deviations were used to compare the results of the original Haavikko method with the ‘shift’ adaptation. Accuracy of age estimation was presented as the percentage of correct estimations within intervals of ±0.5 years, ±1 year, ±1.5 years and ±2 years. The ‘shift’ adaptation was tested through simulation to assure population applicability. The average age for every stage of each permanent tooth was then calculated to provide Croatian specific tables for the Haavikko method. The ‘shift’ adaptation significantly improved age estimation accuracy among boys and girls in all age groups. Simulation confirmed the representativity of the sample and its population applicability. The Croatian specific tables comprise a standard when estimating age using the Haavikko method among Croatian children.
Key words: Age estimation; Haavikko method; “Shift” adaptation; Forensic odontology, Croatia
Keywords: MeSH terms: Age Determination by Teeth, Methods, Research Design, Reference Standards
Introduction
Age estimation is a crucial part of every identification process. More frequent use of x-rays has enabled greater insight into bone and teeth mineralization during growth and development. Chronological age can be estimated based on the recorded pattern of bone and teeth mineralization. Skeletal and dental age, which represent an estimated age based on bone and teeth development, have shown the highest data correlation with chronological age (1-3). While bone development can easily become affected by environmental factors and various diseases, the development of teeth is controlled mainly by genes. Accordingly, skeletal age may be said to represent the measure of development while dental age estimation is more frequently employed as a forensic tool (4).
Conversant with the survey of Moorrees, Fanning and Hunt (5), Karina Haavikko introduced a dental development scale comprising twelve different stages (6). Utilizing this scale, she noted that age was related to specific stages of development in Finnish children. Haavikko evaluated all maxillary and mandibular teeth and constructed gender specific tables with age medians for each developmental stage of every tooth. Consequently, dental age is estimated by dividing the sum of age medians for the developmental stage of specific teeth with the number of teeth assessed. In a study published in 1974, Haavikko introduced a simplified method that utilizes only four teeth for dental age estimation (7).
In practice, the Haavikko method has been tested in significantly fewer studies and its adaptations are relatively rare compared to the Demirjian (8) and Willems (9) methods for dental age estimation. The Demirjian and Willems methods utilize seven left mandibular permanent teeth; their greatest disadvantage is that these methods cannot be used in cases with mandibular hypodontia. The Haavikko method uses different teeth in its selection and is therefore applicable in special cases in which the Demirjian method fails to estimate age. Based on the relatively few published studies testing the Haavikko method in Croatia, the results of which occasionally differ, the purpose of this study was to adapt or ‘shift’ the Haavikko method as a means of increasing accurate age estimation within the Croatian population.
Materials and Methods
The sample included 1, 997 digital, standardized orthopantomograms (OPGs) of children aged between 5 and 16 collected in four Croatian cities: Zagreb, Split, Osijek and Varaždin (Table 1). The OPGs were taken by the Cranex device (Soredex, Finland). Individuals whose OPGs were used for this investigation were referred by their dentist for diagnostic purposes. Their parents or legal guardians signed a letter of informed consent that enabled the OPGs to be used in scientific investigations. This study was approved by the Ethics Committee of the School of Dental Medicine, University of Zagreb.
Table 1. Sample structure.
| Age group | Total | Boys | Girls |
|---|---|---|---|
| 5 – 5.99 | 9 | 4 | 5 |
| 6 – 6.99 | 43 | 31 | 12 |
| 7 – 7.99 | 135 | 82 | 53 |
| 8 – 8.99 | 206 | 121 | 85 |
| 9 – 9.99 | 226 | 121 | 105 |
| 10 – 10.99 | 228 | 135 | 93 |
| 11 – 11.99 | 253 | 126 | 127 |
| 12 – 12.99 | 212 | 112 | 100 |
| 13 – 13.99 | 236 | 129 | 107 |
| 14 – 14.99 | 229 | 141 | 88 |
| 15 – 15.99 | 200 | 107 | 93 |
| Total | 1977 | 1109 | 868 |
The OPGs were coded and stored on a server without name, sex, date of birth or date of record. One investigator (IB) accessed the OPGs from a different computer using the TeamViewer 8.0 programme (TeamViewer GmbH, Germany). The developmental stages of all the teeth were determined using the developmental scale introduced by Haavikko (7), which consists of twelve different stages. Microsoft Excel (Microsoft, USA) was used to store the obtained data. On determining the developmental stages of the teeth in the OPGs, information pertaining to sex, date of birth and date of record was appended.
After a period of two months, 100 randomly chosen OPGs were reassessed; through this process, the developmental stages of all the teeth were redetermined to assess intra-examiner reliability.
Drawing upon and expanding results from a study previously published by Bedek et al. (10), in which the Haavikko method underestimated age by 0.61 years for boys and 0.82 years for girls (Table 2, Table 3), two adaptations of the method were devised.
Table 2. The Haavikko method. Comparison between chronological and dental age for boys [10].
| Boys | Chronological age | Dental age | Deviation | p* |
|---|---|---|---|---|
| Age group | ||||
| 5 – 5.99 | 5.66 | 4.94 | -0.72 | 0.250 |
| 6 – 6.99 | 6.65 | 6.09 | -0.56 | <0.001 |
| 7 – 7.99 | 7.57 | 6.97 | -0.60 | <0.001 |
| 8 – 8.99 | 8.51 | 7.85 | -0.66 | <0.001 |
| 9 – 9.99 | 9.51 | 8.87 | -0.65 | <0.001 |
| 10 – 10.99 | 10.46 | 10.14 | -0.32 | 0.015 |
| 11 – 11.99 | 11.42 | 11.25 | -0.17 | 0.231 |
| 12 – 12.99 | 12.48 | 12.34 | -0.14 | 0.743 |
| 13 – 13.99 | 13.49 | 13.04 | -0.45 | <0.001 |
| 14 – 14.99 | 14.47 | 13.63 | -0.84 | <0.001 |
| 15 – 15.99 | 15.43 | 13.80 | -1.63 | <0.001 |
| Total | 10.51 | 9.90 | -0.61 |
*Wilcoxon signed-rank test
Table 3. The Haavikko method. Comparison between chronological and dental age for girls [10].
| Girls | Chronological age | Dental age | Deviation | p* |
|---|---|---|---|---|
| Age group | ||||
| 5 – 5.99 | 5.33 | 4.50 | -0.84 | 0.063 |
| 6 – 6.99 | 6.62 | 6.32 | -0.30 | 0.519 |
| 7 – 7.99 | 7.60 | 7.07 | -0.52 | <0.001 |
| 8 – 8.99 | 8.52 | 8.10 | -0.42 | <0.001 |
| 9 – 9.99 | 9.52 | 9.27 | -0.25 | <0.001 |
| 10 – 10.99 | 10.54 | 10.31 | -0.23 | 0.120 |
| 11 – 11.99 | 11.50 | 11.51 | 0.00 | 0.388 |
| 12 – 12.99 | 12.48 | 11.96 | -0.52 | <0.001 |
| 13 – 13.99 | 13.50 | 12.35 | -1.15 | <0.001 |
| 14 – 14.99 | 14.42 | 12.47 | -1.94 | <0.001 |
| 15 – 15.99 | 15.48 | 12.66 | -2.82 | <0.001 |
| Total | 10.50 | 9.68 | -0.82 |
*Wilcoxon signed-rank test
The first adaptation of the Haavikko method for the Croatian population was named ‘shift’ and it consists of adding the average difference between chronological and dental age to the estimated dental age. The following equations support this adaptation:
| dental age=HDA+0.61 |
for boys, and
| dental age=HDA+0.82 |
for girls
where HDA represents dental age estimation using the Haavikko method and the numbers 0.61 and 0.82 denote the average difference between chronological and dental age for boys and girls obtained from the earlier Bedek et al. study (10).
Square deviations were used to compare the original Haavikko method with its ‘shift’ adaptation for the Croatian population. The accuracy of the age estimation is presented as the percentage of correct estimations within the intervals of ±0.5 years, ±1 year, ±1.5 years and ±2 years.
The ‘shift’ adaptation of the Haavikko method was tested for applicability in the Croatian population using a simulation in which the sample was randomly split in half: the first half was used to determine the average difference between chronological and dental age using the Haavikko method; the second half was used to test the ‘shift’ adaptation of that method. The average value of ten simulations was calculated and validated for consistency.
As the developmental stages of all the teeth were determined, in the second adaptation the average age for Haavikko’s developmental stages was calculated for every tooth, with the notable exception of the third molar, by providing specific tables pertaining to the Croatian population.
Results
Based on 100 randomly reassessed OPGs, the Kappa value for intra-examiner agreement was 0.83, which indicates a level of high reliability in the developmental stages’ determination.
Square deviations were used to compare the original Haavikko method with the ‘shift’ adaptation for the Croatian population (Table 4). The accuracy of both methods is presented in parallel in Table 5 as a percentage of correct estimations within a specific interval.
Table 4. The Haavikko method and its ‘shift’ adaptation for the Croatian population. Square deviations between chronological and dental age for boys and girls are stipulated. The lower value indicates the more accurate age estimation.
| Boys | Haavikko | Shift | Girls | Haavikko | Shift | |
|---|---|---|---|---|---|---|
| Age group | Age group | |||||
| 5 – 5.99 | 1.33 | 0.83 | 5 – 5.99 | 1.19 | 0.49 | |
| 6 – 6.99 | 0.90 | 0.60 | 6 – 6.99 | 0.80 | 0.98 | |
| 7 – 7.99 | 1.08 | 0.73 | 7 – 7.99 | 0.76 | 0.57 | |
| 8 – 8.99 | 1.16 | 0.72 | 8 – 8.99 | 0.74 | 0.72 | |
| 9 – 9.99 | 1.37 | 0.95 | 9 – 9.99 | 0.39 | 0.64 | |
| 10 – 10.99 | 2.14 | 2.12 | 10 – 10.99 | 1.19 | 1.48 | |
| 11 – 11.99 | 1.99 | 2.15 | 11 – 11.99 | 0.58 | 1.26 | |
| 12 – 12.99 | 1.04 | 1.24 | 12 – 12.99 | 0.57 | 0.39 | |
| 13 – 13.99 | 0.79 | 0.62 | 13 – 13.99 | 1.52 | 0.30 | |
| 14 – 14.99 | 0.89 | 0.24 | 14 – 14.99 | 3.96 | 1.45 | |
| 15 – 15.99 | 2.81 | 1.19 | 15 – 15.99 | 8.08 | 4.14 | |
| Total | 15.50 | 11.37 | Total | 19.79 | 12.42 |
Table 5. Accuracy of age estimation of the Haavikko method and its ‘shift’ adaptation for the Croatian population. Percentage of correct estimation within interval.
| Interval | Haavikko | Shift |
|---|---|---|
| Boys | ||
| ± 0.5 years | 28.4 | 41.3 |
| ± 1 year | 58.2 | 69.2 |
| ± 1.5 years | 78.7 | 86.0 |
| ± 2 years | 91.9 | 94.6 |
| Girls | ||
| ± 0.5 years | 33.6 | 32.1 |
| ± 1 year | 57.6 | 62.8 |
| ± 1.5 years | 73.2 | 81.3 |
| ± 2 years | 82.6 | 93.3 |
The previously described simulation was used to evaluate the applicability of the adaptation of the Haavikko method in the Croatian population (i.e., not only in the sample). Tables 6 and 7 present the average values from ten simulations.
Table 6. Square deviations between chronological and dental age of the Haavikko method and its ‘shift’ adaptation for the Croatian population (average values of ten simulations).
| Boys | Haavikko | Shift | Girls | Haavikko | Shift | |
|---|---|---|---|---|---|---|
| Age group | Age group | |||||
| 5 – 5.99 | 1.24 | 0.56 | 5 – 5.99 | 1.37 | 0.63 | |
| 6 – 6.99 | 0.89 | 0.58 | 6 – 6.99 | 0.88 | 0.97 | |
| 7 – 7.99 | 1.10 | 0.71 | 7 – 7.99 | 0.70 | 0.57 | |
| 8 – 8.99 | 1.14 | 0.68 | 8 – 8.99 | 0.70 | 0.69 | |
| 9 – 9.99 | 1.34 | 0.89 | 9 – 9.99 | 0.40 | 0.64 | |
| 10 – 10.99 | 2.30 | 2.33 | 10 – 10.99 | 1.26 | 1.58 | |
| 11 – 11.99 | 1.86 | 2.14 | 11 – 11.99 | 0.60 | 1.34 | |
| 12 – 12.99 | 1.12 | 1.27 | 12 – 12.99 | 0.54 | 0.40 | |
| 13 – 13.99 | 0.72 | 0.55 | 13 – 13.99 | 1.49 | 0.29 | |
| 14 – 14.99 | 0.91 | 0.25 | 14 – 14.99 | 3.98 | 1.44 | |
| 15 – 15.99 | 2.75 | 1.15 | 15 – 15.99 | 8.10 | 4.11 | |
| Total | 15.37 | 11.10 | Total | 20.02 | 12.65 |
Table 7. Average accuracy of ten simulations of the Haavikko method and its ‘shift’ adaptation for the Croatian population. Percentage of correct estimations within interval.
| Interval | Haavikko | Shift |
|---|---|---|
| Boys | ||
| ± 0.5 years | 28.6 | 41.4 |
| ± 1 year | 57.7 | 69.3 |
| ± 1.5 years | 79.1 | 85.9 |
| ± 2 years | 91.8 | 94.5 |
| Girls | ||
| ± 0.5 years | 33.0 | 32.3 |
| ± 1 year | 57.4 | 62.0 |
| ± 1.5 years | 73.3 | 80.8 |
| ± 2 years | 82.3 | 93.0 |
A complete adaptation of the Haavikko method for the Croatian population was made by calculating the average age for each developmental stage for every tooth with the exception of the third molar (Table 8 and Table 9). This was achieved by adding together the age of all participants who had reached the developmental stage of the specific tooth and then dividing the sum by the number of participants.
Table 8. Average age for the Haavikko developmental stages for the Croatian population (boys).
| Stage | 11 | 12 | 13 | 14 | 15 | 16 | 17 |
|---|---|---|---|---|---|---|---|
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | . | . | 5.4 |
| C1/2 | . | . | . | . | 7.1 | . | 7.4 |
| C3/4 | . | 6.4 | 6.7 | 6.8 | 7.3 | . | 8.1 |
| Crc | . | 7.1 | 6.9 | 7.5 | 8.2 | . | 8.3 |
| Ri | 6.9 | 7.5 | 7.3 | 8.3 | 8.6 | . | 8.9 |
| R1/4 | 7.2 | 7.6 | 8.0 | 8.9 | 9.6 | 5.9 | 10.2 |
| R1/2 | 7.7 | 8.0 | 9.2 | 9.9 | 10.3 | 7.2 | 10.9 |
| R3/4 | 8.1 | 8.8 | 10.7 | 10.5 | 11.0 | 8.1 | 11.6 |
| Rc | 9.4 | 10.0 | 12.3 | 11.6 | 12.1 | 8.9 | 12.8 |
| Ac | 12.6 | 13.1 | 14.2 | 13.8 | 14.1 | 12.1 | 14.5 |
| Stage | 21 | 22 | 23 | 24 | 25 | 26 | 27 |
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | 5.4 | . | 5.4 |
| C1/2 | . | . | . | . | 7.2 | . | 7.4 |
| C3/4 | . | 6.4 | 6.7 | 6.8 | 7.3 | . | 7.9 |
| Crc | . | 7.0 | 7.0 | 7.5 | 8.2 | . | 8.2 |
| Ri | 6.9 | 7.5 | 7.5 | 8.3 | 8.7 | . | 9.0 |
| R1/4 | 7.2 | 7.7 | 8.1 | 8.9 | 9.5 | 5.9 | 10.1 |
| R1/2 | 7.7 | 8.1 | 9.1 | 10.0 | 10.4 | 7.2 | 10.8 |
| R3/4 | 8.1 | 8.8 | 10.8 | 10.5 | 11.0 | 8.1 | 11.4 |
| Rc | 9.3 | 9.9 | 12.2 | 11.5 | 12.2 | 8.8 | 12.7 |
| Ac | 12.6 | 13.0 | 14.2 | 13.8 | 14.1 | 12.1 | 14.5 |
| Stage | 31 | 32 | 33 | 34 | 35 | 36 | 37 |
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | 5.4 | . | 5.4 |
| C1/2 | . | . | . | . | 6.9 | . | 7.3 |
| C3/4 | . | . | 6.8 | 6.9 | 7.4 | . | 7.3 |
| Crc | . | . | 7.1 | 6.9 | 7.8 | . | 8.1 |
| Ri | . | 6.1 | 7.2 | 7.8 | 8.4 | . | 8.7 |
| R1/4 | 6.6 | 6.1 | 7.9 | 8.6 | 9.0 | 5.7 | 9.7 |
| R1/2 | 7.4 | 7.1 | 8.8 | 9.5 | 10.0 | 6.8 | 10.6 |
| R3/4 | 7.8 | 7.8 | 10.5 | 10.6 | 11.2 | 8.1 | 11.3 |
| Rc | 7.9 | 8.5 | 12.2 | 11.6 | 12.6 | 8.5 | 12.9 |
| Ac | 11.9 | 12.4 | 14.2 | 13.9 | 14.3 | 12.4 | 14.6 |
| Stage | 41 | 42 | 43 | 44 | 45 | 46 | 47 |
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | 5.4 | . | 5.4 |
| C1/2 | . | . | . | . | 6.9 | . | 7.3 |
| C3/4 | . | . | . | 7.0 | 7.2 | . | 7.4 |
| Crc | . | 6.0 | 7.2 | 7.0 | 7.9 | . | 8.1 |
| Ri | . | 6.1 | 7.2 | 7.7 | 8.4 | . | 8.6 |
| R1/4 | 6.6 | 6.1 | 7.8 | 8.6 | 9.2 | 5.7 | 9.6 |
| R1/2 | 7.3 | 7.0 | 8.9 | 9.6 | 10.1 | 6.8 | 10.6 |
| R3/4 | 7.7 | 7.8 | 10.5 | 10.6 | 11.3 | 8.0 | 11.3 |
| Rc | 7.9 | 8.5 | 12.1 | 11.5 | 12.6 | 8.4 | 13.0 |
| Ac | 11.9 | 12.4 | 14.3 | 13.9 | 14.3 | 12.4 | 14.6 |
Table 9. Average age for the Haavikko developmental stages for the Croatian population (girls).
| Stage | 11 | 12 | 13 | 14 | 15 | 16 | 17 |
|---|---|---|---|---|---|---|---|
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | 5.9 | . | . |
| C1/2 | . | . | . | 6.0 | 6.2 | . | 6.3 |
| C3/4 | 5.1 | 5.2 | 5.3 | 6.2 | 7.7 | . | 7.5 |
| Crc | 5.3 | 6.3 | 6.1 | 7.3 | 8.0 | . | 8.0 |
| Ri | 6.5 | 6.7 | 6.9 | 7.9 | 8,7 | 5.0 | 8.9 |
| R1/4 | 6.5 | 7.6 | 7.8 | 8.8 | 9.4 | 5.2 | 10.0 |
| R1/2 | 7.5 | 7.9 | 8.8 | 9.5 | 10.1 | 6.1 | 10.5 |
| R3/4 | 7.8 | 8.7 | 9.9 | 10.3 | 11.0 | 6.2 | 11.2 |
| Rc | 8.8 | 9.3 | 11.3 | 10.9 | 11.5 | 7.9 | 12.6 |
| Ac | 12.4 | 12.7 | 13.6 | 13.4 | 13.7 | 12.1 | 14.2 |
| Stage | 21 | 22 | 23 | 24 | 25 | 26 | 27 |
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | 5.4 | 5.3 | . | . |
| C1/2 | . | . | . | 5.9 | 6.4 | . | 5.9 |
| C3/4 | 5.1 | 5.2 | 5.9 | 6.0 | 7.4 | . | 7.3 |
| Crc | 5.3 | 5.9 | 6.1 | 7.4 | 7.9 | . | 7.9 |
| Ri | 6.5 | 6.3 | 7.1 | 8.0 | 8.9 | . | 9.0 |
| R1/4 | 6.5 | 7.5 | 7.8 | 8.8 | 9.6 | 5.2 | 10.1 |
| R1/2 | 7.5 | 7.9 | 8.7 | 9.5 | 10.0 | 5.9 | 10.5 |
| R3/4 | 7.8 | 8.6 | 10.0 | 10.2 | 11.1 | 6.4 | 11.4 |
| Rc | 8.8 | 9.2 | 11.3 | 10.9 | 11.6 | 7.9 | 12.5 |
| Ac | 12.4 | 12.7 | 13.6 | 13.4 | 13.8 | 12.0 | 14.3 |
| Stage | 31 | 32 | 33 | 34 | 35 | 36 | 37 |
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | . | . | . |
| C1/2 | . | . | . | . | 5.4 | . | 6.0 |
| C3/4 | . | . | 5.4 | 5.4 | 7.2 | . | 7.0 |
| Crc | . | . | 5.5 | 7.0 | 7.7 | . | 7.9 |
| Ri | . | 5.3 | 6.7 | 7.4 | 8.1 | . | 8.8 |
| R1/4 | 5.4 | 6.2 | 7.6 | 8.3 | 8.6 | 5.3 | 9.8 |
| R1/2 | 5.8 | 6.6 | 8.4 | 9.2 | 9.8 | 5.6 | 10.1 |
| R3/4 | 7.0 | 7.4 | 9.6 | 10.2 | 11.2 | 6.3 | 11.2 |
| Rc | 7.6 | 8.3 | 11.0 | 11.0 | 12.0 | 8.2 | 12.7 |
| Ac | 11.9 | 12.2 | 13.5 | 13.6 | 14.0 | 12.2 | 14.5 |
| Stage | 41 | 42 | 43 | 44 | 45 | 46 | 47 |
| 0 | . | . | . | . | . | . | . |
| Ci | . | . | . | . | . | . | . |
| Cco | . | . | . | . | . | . | . |
| C1/2 | . | . | . | . | 5.4 | . | 6.6 |
| C3/4 | . | . | 5.4 | 5.4 | 6.3 | . | 7.0 |
| Crc | . | . | 5.8 | 6.6 | 7.8 | . | 7.8 |
| Ri | . | 5.3 | 6.2 | 7.5 | 8.1 | . | 8.8 |
| R1/4 | 5.4 | 6.2 | 7.7 | 8.3 | 8.7 | 5.3 | 9.7 |
| R1/2 | 5.8 | 6.6 | 8.3 | 9.2 | 9.7 | 5.5 | 10.2 |
| R3/4 | 6.9 | 7.4 | 9.5 | 10.2 | 11.1 | 6.3 | 11.2 |
| Rc | 7.7 | 8.3 | 10.9 | 11.3 | 12.2 | 8.1 | 12.7 |
| Ac | 11.9 | 12.2 | 13.5 | 13.6 | 14.0 | 12.2 | 14.5 |
Discussion
The aim of this study was to adapt the Haavikko method for the Croatian population and to provide average ages for the Haavikko developmental stages based on a sample of Croatian children aged 5 to 16. The size of the sample, its geographical origin and the digital, standardized orthopantomograms taken on the same device ensure an accurate representation of the Croatian population. While the original Haavikko method significantly underestimated age in most of the groups in the sample (10), the simple ‘shift’ adaptation significantly improved the results of age estimation. Due to the large and representative sample, the authors were able to calculate the average age for each developmental stage of every tooth according to Haavikko’s developmental scale. The results shown in Tables 8 and 9 represent the complete adaptation of the Haavikko method for the Croatian population.
To remain within the age parameters of this study, the results for the third molars were deliberately omitted to prevent the possibility of error, particularly with regard to the higher developmental stages. An accurate calculation of the average age on reaching the developmental stages of the third molars demands that a representative sample includes OPGs of children older than 16 years of age, which is not the case in this survey.
It must be observed that due to the limited number of OPGs recorded in the two youngest age groups, a degree of caution is required in estimating age accurately among the youngest individuals. Ethical considerations ensured that OPGs were taken for diagnostic purposes only; this affected the authors’ ability to collect additional OPGs in these two age groups. As a result, the statistical methods used in this survey were selected with discretion to eliminate the influence of the two youngest age groups on the results of age estimation for the other age groups. The low number of OPGs in the youngest age groups is noted in the cited literature and may be qualified by the above stated reasons (11-13).
The Haavikko method has been tested previously in Croatia. Many of the methods and results contained in this study build upon and reiterate an earlier study on the Haavikko method in the Croatian population (10). In addition, the sample sizes of previous studies have also provided a store of data on which to build. On a sample of 324 children aged 6 to 16, Bagić et al. determined that the Haavikko method underestimated age by 0.5 years in boys and 1 year in girls (14). Although the age of participants differed slightly, Borčić et al. stipulated that the Haavikko method underestimated age by 0.5 years in both boys and girls on a small sample of 115 orthopantomograms of children aged 4 to 17 (15). Based on a larger sample of 1, 492 Croatian children aged between 6 and 14, Brkić et al. posited an average underestimation of 0.17 years for both genders when using the Haavikko method; but of significance, the mean absolute error is 1.01 years in boys and 1.18 years in girls (16). Accordingly, the authors of this study found the Haavikko method to be more precise than the Demirjian and Willems methods in their testing on the Croatian population, although it should be observed that the sample of children between 6 and 14 narrows the scope of the data (16). The results of Bedek et al. (10) drawn from the large sample of 1, 997 children from the Croatian population were found to be similar to those previously published, albeit with reference to a significantly smaller sample (14, 15).
The Haavikko method has also been tested worldwide. In the Chinese population, the method was found to be highly accurate with a mean difference between dental and chronological age of 0.14 years (17). In the Italian population, an average underestimation of 0.35 years was observed in boys and 0.41 years in girls (18). Using the Haavikko method to determine age estimation in the Turkish population, an underestimation average of 0.49 years was established (19); in a comparable study, an underestimation of 0.60 in boys and 0.56 years in girls is present (20). An underestimation of 0.94 years in boys and 1.59 in girls was found in the Malaysian population (21), whereas in the Indian population a much lower underestimation of 0.17 years in boys and 0.29 in girls was recorded (22). A study in southern India by Mohammed et al. (23) found that the Haavikko method underestimated age by 2.84 years in boys and 2.96 years in girls. A likely reason for this substantial underestimation is the sample selection, in which one third of the sample was taken from children older than 14 years of age; moreover, there is no information provided with regard to how many of those children were in the significant 15 to 16 age group. As the apex of the second molar, the last tooth to finish its development (apart from the third molar, which was omitted from the study), closes at the age of 15 in girls and before 16 in boys, it must be reasoned that a large number of participants close to 16 would inevitably increase the Haavikko’s method underestimation of age.
The ‘shift’ adaptation of the Haavikko method, together with the related data tables with an average age for the developmental stages for boys and girls, may be considered the most significant contribution of the present study. The comparatively large sample used by Bedek et al. (10) in their data analysis is also of appreciable significance. To date, this is the only study conducted among the Croatian population that provides an average age for the Haavikko stages of development using a representative sample. It is important to note that the average values of the ten simulations increase the merit of the study as this method ensured that the sample was truly representative and that the results of the estimations were consistent regardless of the manner in which the sample was divided (i.e., they were not affected by which part of the sample was used as a reference for the adaptation).
Although little more than a simple tweak on the Haavikko method, the authors’ ‘shift’ adaptation significantly increases the accuracy of the results in this study of age estimation. There have been other attempts at modification; however, the most sensible objective is to provide an average age for the different stages of development based on a representative sample. This has not always proven to be the case. An attempt at adapting the Haavikko method in the Italian population using models based on smoothing splines decreased the difference between dental and chronological age, albeit according to the new formulæ (24). Two adaptations practiced in the Indian population, one using seven and the other using four teeth, yielded more accurate results with the seven teeth model (25). Nonetheless, it would be mistaken to regard these studies as true adaptations as only the selection of the teeth is changed while the mean age for each stage is taken from the original Haavikko study.
It should also be observed that age representation and sample size affect the estimation results. This differentiation in age, the number of participants and geographical location make accurate comparison a challenge. In a study drawn from a sample of 1, 974 Bosnian-Herzegovinian children aged 6 to 13, Galić et al. posited that the Haavikko method underestimated age by 0.09 years in boys and 0.29 years in girls (26). The results determined by Bedek et al. (10) in the Croatian population therefore differ markedly from those obtained by Galić et al. in the neighbouring Bosnian-Herzegovinian population. Although other conclusions may be drawn, the authors of the present study contend that age is the determining factor in the results. Galić et al. based their survey on a sample of children aged 6 to 13 (27), which would account for the discrepancies between this study and the conclusions drawn from Bedek et al. (10) and Brkić et al. (16). It may be argued that the accuracy of the Haavikko method or lack thereof derives from these notable differences in the sample. While Bedek et al. analysed the OPGs of children up to 16 years of age, Galić and Brkić limited the scope of their study to children up to 13 and 14. It must be reiterated that the most significant difference between chronological and dental age was observed in the groups between 14 and 16 – the latter stages of development (Table 1 and 2). The ‘shift’ adaptation of the Haavikko method was designed not merely to increase the accuracy of sample results but to remedy and in part explain the reason for the sizeable discrepancies among the various studies on the subject.
Conclusion
It may be concluded that the simple ‘shift’ adaptation of the Haavikko method increased the accuracy of dental age estimation among Croatian children. The simulation provided appropriate sample representativity and confirmed its application on the population. Of particular significance, the comparatively large sample size broadened the scope of the analysis and facilitated a keener insight into previous studies in terms of their strengths and possible limitations. This representative analysis for the Haavikko developmental stages in Croatian specific tables provides a valuable reference for dental development in Croatian children and may be used as a tool for more accurate age estimation elsewhere. These tables may therefore be regarded as the current standard, the future utilisation of which should not be limited to a single population.
Acknowledgements
The authors whish to thank Dr. Rupert Thorough for proofreading and valuable advice regarding the English language.
Compliance with Ethical Standards
This study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments. Ethical approval was granted by the Ethics Committee of the School od Dental Medicine, University of Zagreb.
Informed consent was obtained by parents or legal guardians of all participants in the study enabling the orthopantomograms to be used in scientific investigations.
The authors have no relevant financial or non-financial interests to disclose.
No funding was received for conducting this study.
The authors have no competing interests to disclose.
Footnotes
Conflict of interest
The authors declare that they have no conflict of interest.
References
- 1.Gupta M, Divyashree R, Abhilash P. A Bijle MN, Murali K. Correlation between Chronological Age, Dental Age and Skeletal Age among Monozygoyic and Dizygotic Twins. J Int Oral Health. 2013;5(1):16–22. [PMC free article] [PubMed] [Google Scholar]
- 2.Demirjian A, Buschang PH, Tanguay R, Patterson DK. Interrelationships among measures of somatic, skeletal, dental, and sexual maturity. Am J Orthod. 1985. November;88(5):433–8. 10.1016/0002-9416(85)90070-3 [DOI] [PubMed] [Google Scholar]
- 3.Rai V, Saha S, Yadav G, Tripathi AM, Grover K. Dental and skeletal maturity- A biological indicator of chronologic age. JCDR. 2014;8(9):ZC60–4. 10.7860/JCDR/2014/10079.4862 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Manzoor Mughal A, Hassan N, Ahmed A. Bone age assessment methods: a critical review. Pak J Med Sci. 2014. January;30(1):211–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Moorrees CF, Fanning EA, Hunt EEJ. Age variation of formation stages for ten permanent teeth. J Dent Res. 1963;42:1490–502. 10.1177/00220345630420062701 [DOI] [PubMed] [Google Scholar]
- 6.Haavikko K. The formation and the alveolar and clinical eruption of the permanent teeth. An orthopantomographic study. Suom Hammaslaak Toim. 1970;66(3):103–70. [PubMed] [Google Scholar]
- 7.Haavikko K. Tooth formation age estimated on a few selected teeth. A simple method for clinical use. Proc Finn Dent Soc. 1974. February;70(1):15–9. [PubMed] [Google Scholar]
- 8.Demirjian A, Goldstein H, Tanner JM. A new system of dental age assessment. Hum Biol. 1973. May;45(2):211–27. [PubMed] [Google Scholar]
- 9.Willems G, Van Olmen A, Spiessens B, Carels C. Dental age estimation in Belgian children: Demirjian’s technique revisited. J Forensic Sci. 2001. July;46(4):893–5. 10.1520/JFS15064J [DOI] [PubMed] [Google Scholar]
- 10.Bedek I, Dumančić J, Lauc T, Marušić M, Čuković-Bagić I. Applicability of the Demirjian, Willems and Haavikko methods in Croatian children. J Forensic Odontostomatol. 2022;40(2):21–30. [PMC free article] [PubMed] [Google Scholar]
- 11.Djukic K, Zelic K, Milenkovic P, Nedeljkovic N, Djuric M. Dental age assessment validity of radiographic methods on Serbian children population. Forensic Sci Int. 2013. September;231(1–3):398.e1–5. 10.1016/j.forsciint.2013.05.036 [DOI] [PubMed] [Google Scholar]
- 12.Willems G, Lee SS, Uys A, Bernitz H, Cadenas de Llano-Perula M, Fieuws S, et al. Age estimation based on Willems method versus new country-specific method in South African black children. Int J Legal Med. 2018. March;132(2):599–607. 10.1007/s00414-017-1686-3 [DOI] [PubMed] [Google Scholar]
- 13.Wang J, Ji F, Zhai Y, Park H, Tao J. Is Willems method universal for age estimation: A systematic review and meta-analysis. J Forensic Leg Med. 2017. November;52:130–6. 10.1016/j.jflm.2017.09.003 [DOI] [PubMed] [Google Scholar]
- 14.Čuković Bagić I, Sever N, Brkić H, Kern J. Dental Age Estimation in Children Using Orthopantomograms. Acta Stomatol Croat. 2008;42(1):11–8. [Google Scholar]
- 15.Borcic I, Petrovecki M, Brkic H. Studies of Two Different Methods for Dental Age Estimation in Croatian Children. Acta Stomatol Croat. 2006;40(2):135–41. [Google Scholar]
- 16.Brkić H, Galić I, Vodanović M, Dumančić J, Mehdi F, Anić Milošević S. The Cameriere, Haavikko, Demirjian, and Willems methods for the assessment of dental age in Croatian children. Int J Legal Med. 2022. November;136(6):1685–96. 10.1007/s00414-022-02891-1 [DOI] [PubMed] [Google Scholar]
- 17.Wang Y, Huang S, Liu H. Use of Haavikko’s method to assess dental age in Chinese children. Community Dent Health. 2011. June;28(2):160–4. [PubMed] [Google Scholar]
- 18.Butti AC, Clivio A, Ferraroni M, Spada E, Testa A, Salvato A. Haavikko’s method to assess dental age in Italian children. Eur J Orthod. 2009. April;31(2):150–5. 10.1093/ejo/cjn081 [DOI] [PubMed] [Google Scholar]
- 19.Sezer B, Çarıkçıoğlu B. Accuracy of the London Atlas, Haavikko’s Method and Cameriere’s European Formula of dental age estimation in Turkish children. Leg Med (Tokyo). 2022. February;54:101991. 10.1016/j.legalmed.2021.101991 [DOI] [PubMed] [Google Scholar]
- 20.Kırzıoğlu Z, Ceyhan D. Accuracy of different dental age estimation methods on Turkish children. Forensic Sci Int. 2012. March;216(1–3):61–7. 10.1016/j.forsciint.2011.08.018 [DOI] [PubMed] [Google Scholar]
- 21.Kumaresan R, Cugati N, Chandrasekaran B, Karthikeyan P. Reliability and validity of five radiographic dental-age estimation methods in a population of Malaysian children. J Investig Clin Dent. 2016;7(1):102–9. 10.1111/jicd.12116 [DOI] [PubMed] [Google Scholar]
- 22.Hegde S, Patodia A, Dixit U. A comparison of the validity of the Demirjian, Willems, Nolla and Häävikko methods in determination of chronological age of 5–15 year-old Indian children. J Forensic Leg Med. 2017;50:49–57. 10.1016/j.jflm.2017.07.007 [DOI] [PubMed] [Google Scholar]
- 23.Mohammed RB, Sanghvi P, Perumalla KK, Srinivasaraju D, Srinivas J, Siva Kalyan U, et al. Accuracy of four dental age estimation methods in Southern Indian children. JCDR. 2015;9(1):HC01–08. 10.7860/JCDR/2015/10141.5495 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Spada E, Butti AC, Ferraroni M, Milani S. Adapting Häävikko’s dental age for the assessment of Italian children: use of LMS and other models based on smoothing splines. Stat Med. 2009. December;28(28):3554–61. 10.1002/sim.3717 [DOI] [PubMed] [Google Scholar]
- 25.Hedge S, Shah K, Dixit U. A comparative evaluation of the applicability of two adapted Häävikko Methods for age estimation of 5-15 year-old Indian children. J Forensic Odontostomatol. 2016;34(2):21–34. [PMC free article] [PubMed] [Google Scholar]
- 26.Galić I, Vodanović M, Cameriere R, Nakaš E, Galić E, Selimović E, et al. Accuracy of Cameriere, Haavikko, and Willems radiographic methods on age estimation on Bosnian-Herzegovian children age groups 6-13. Int J Legal Med. 2011;125(2):315–21. 10.1007/s00414-010-0515-8 [DOI] [PubMed] [Google Scholar]
- 27.Galić I, Vodanović M, Janković S, Mihanović F, Nakaš E, Prohić S, et al. Dental age estimation on Bosnian-Herzegovinian children aged 6-14 years: Evaluation of Chaillet’s international maturity standards. J Forensic Leg Med. 2013;20(1):40–5. 10.1016/j.jflm.2012.04.037 [DOI] [PubMed] [Google Scholar]