TY - JOUR
T1 - Statistical model estimating the occurrence of otitis media from temporal bone pneumatization
AU - Homøe, Preben
AU - Rasmussen, Niels
AU - Lynnerup, Niels
AU - Skovgaard, Lene Theil
PY - 1994/6
Y1 - 1994/6
N2 - In order to investigate the relationship between the pneumatization of temporal bones and the occurrence of otitis media in Greenlandic Inuit, 36 Greenlandic Inuit were examined by radiography of the temporal bones. The pneumatized cell area was measured planimetrically. All subjects answered a questionnaire on infectious middle ear disease (IMED), and an objective otologic examination was performed. Nine persons of 34 (26%) reported IMED in childhood, and there was IMED reported in all pneumatized areas below 400 mm2. Based on bilateral area measures, a polychotomous logistic regression model was applied. The occurrence of IMED was shown to be associated with smaller areas, and unilateral IMED was associated with pronounced asymmetry. The model has enabled risk assessment, and 8 persons were designated by the model as having had IMED with a sensitivity of 67% (confidence interval 30% to 93%) and a specificity of 92% (confidence interval 74% to 99%). When the model was applied to a historical anthropological Inuit material from 1700 to 1800 AD, 6 of 56 crania were designated as having had IMED. This method has improved the accuracy of estimating the occurrence of IMED in ancient populations.
AB - In order to investigate the relationship between the pneumatization of temporal bones and the occurrence of otitis media in Greenlandic Inuit, 36 Greenlandic Inuit were examined by radiography of the temporal bones. The pneumatized cell area was measured planimetrically. All subjects answered a questionnaire on infectious middle ear disease (IMED), and an objective otologic examination was performed. Nine persons of 34 (26%) reported IMED in childhood, and there was IMED reported in all pneumatized areas below 400 mm2. Based on bilateral area measures, a polychotomous logistic regression model was applied. The occurrence of IMED was shown to be associated with smaller areas, and unilateral IMED was associated with pronounced asymmetry. The model has enabled risk assessment, and 8 persons were designated by the model as having had IMED with a sensitivity of 67% (confidence interval 30% to 93%) and a specificity of 92% (confidence interval 74% to 99%). When the model was applied to a historical anthropological Inuit material from 1700 to 1800 AD, 6 of 56 crania were designated as having had IMED. This method has improved the accuracy of estimating the occurrence of IMED in ancient populations.
KW - Arctic
KW - epidemiology
KW - Eskimo
KW - matched pair data
KW - otopathology
KW - paleopathology
KW - physical anthropology
KW - polychotomous logistic regression
UR - https://www.scopus.com/pages/publications/0028241604
U2 - 10.1177/000348949410300608
DO - 10.1177/000348949410300608
M3 - Article
C2 - 8203813
AN - SCOPUS:0028241604
SN - 0003-4894
VL - 103
SP - 469
EP - 473
JO - Annals of Otology, Rhinology & Laryngology
JF - Annals of Otology, Rhinology & Laryngology
IS - 6
ER -