Setting
The setting was Klamath County in rural southeast Oregon USA [8, 9]. The population in 2008 was 66,425 and growth 2000 to 2008 was 4.2%. The county in 2008 was 82.7% white but not of Hispanic or Latino origin, 9.2% Hispanic, 4.1% American Indian or Alaska Native, 1.0% Asian, 0.8% black, 0.2% Pacific Islander and 2% other or mixed races. There was no artificial fluoridation and little naturally occurring fluoride. Over half (432, 51.6%) of the 836 births in Klamath County in 2007 were to low-income women covered by the Oregon Health Plan (OHP), the Medicaid program.
Participants
Two groups of children were examined. The first group consisted of a sample of offspring of mothers in Klamath County who were eligible for the counseling program and received dental care during pregnancy or within two months of delivery. The children of 235 mothers were invited for examinations in random order and 113 (48%) were recalled and examined. The primary reason for non-participation was that the program was unable to contact the mother after multiple attempts. The comparison group consisted of 56 children in Deschutes, Crook and Jefferson counties. The children were selected from a list of all children between 24 and 35 months of age enrolled with OHP and eligible for care by the dental care organization Advantage Dental Services, LLC. The universe of eligible children was identified and the parents contacted by telephone. In Deschutes County there were 70 children identified: 30 (42.8%) were examined. In Jefferson County there were 37 children eligible: 16 (43.2%) participated. In Crook County there were 22 eligible children: 10 (45.4%) participated. All three counties are in rural Oregon and have similar levels of births to women served by the Oregon Health Plan. The Institutional Review Board of the University of Washington approved the study and the informed consent of the parents was obtained.
Examinations
The examinations were conducted by one of the dental managed care organizations in cooperation with the Klamath County Department of Health as part of the ongoing evaluation of the program. Visual examinations were conducted using artificial light and a front plane mirror by two examiners (Shirtcliff and Woll). World Health Organization criteria [10] were used: only frank cavitation was recorded as tooth decay. Dental exams were performed in dental offices in Bend, (Deschutes County) Madras (Jefferson County) and Prinville (Crook County). The detection of caries was done using visual/tactile examination, using a #23 explorer without significant axial force. The interrater reliability of the examiners was assessed on seven children seen by both examiners: the intraclass correlation (ICC) for the number of teeth with decay was .95. Examiners were not blind to which sample was being examined. The outcome measures in the study were the number of children with any decayed deciduous tooth and the number of such teeth.
Descriptive information on the children and their mothers
The child's age, gender and race, and whether this was the mother's first child were collected by interview or from the WIC database. For the comparison group, whether the mother received any dental care during her pregnancy or immediately post partum was obtained from the dental managed care organizations.
Analysis plan
The de-identified data were recorded in an Excel spreadsheet by program personnel, edited, and then imported into SPSS (version 16 for Mac). We tested the hypothesis that children of mothers in the Klamath County program would be less likely to have any tooth decay and have fewer decayed or filled teeth than the children of mothers in the comparison counties. Binomial regression using a log link was used to calculate the relative risk of being caries free, adjusted for child age and Hispanic race. The binomial regression was implemented using generalized estimating equations with a robust variance estimator to estimate valid standard errors and perform statistical inference [11].