Fluoride Action Network

Armfield & Spencer (2004): Consumption of Nonpublic water: Implications for Children’s Caries Experience

Fluoride Action Network | August 2012 | By Michael Connett - UPDATED October 31, 2016

In 2004, a large-scale Australian study tested the hypothesis that increasing consumption of bottled water and other forms of “non-public” water (e.g., rainwater tanks) in Australia would increase the rate of tooth decay in children due to the reduced exposure to fluoridated tap water. To test the hypothesis, the author, Jason Armfield and John Spencer, studied the dental health records of over 13,000 children in South Australia, and ascertained each child’s lifetime exposure to fluoridated water and non-public (non fluoridated) water.

Although the authors make their pro-fluoridation favorable views very clear in the paper, the results of the study provide further evidence that water fluoridation is unnecessary and ineffective. As shown in the following table, boys who had consumed fluoridated water their entire life had the same rate of tooth decay as boys who had never consumed fluoridated water, while girls who consumed fluoridated water their entire life only had 0.15 fewer decayed surfaces than girls who had not consumed fluoridated water:

Decayed, Missing or Filled Surfaces (DMFS) in Children
Based on Percent of Life Drinking Fluoridated/Non-Fluoridated Drinking Water
 % of Life Drinking Non-Fluoridated Water  Males  Females
 0%  1.38  1.33
1-50%  1.32 1.39
51-99% 1.37 1.63
100% 1.38 1.48

Despite this finding, the authors conclude that the study further proves the need for more water fluoridation. They reach this conclusion based on their finding that lifetime exposure to fluoridated water was associated with a reduction in tooth decay in the deciduous teeth (i.e., baby teeth). However, as a cursory inspection of this study shows, the contribution of fluoridation to the reduction in tooth decay was extraordinarily weak. This is evident when looking at the multiple regression analyses where the authors assessed the size of the fluoridation effect by simultaneously controlling for other variables that are known to affect tooth decay rates, such as socioeconomic status and parental education, etc. (Amazingly, prior to 2000, not a single fluoridation/tooth study had ever controlled for 3 or more confounding variables.)

After doing what so few pro-fluoridation researchers had done before (i.e., controlling for income, etc), the authors found that that fluoridation’s effect on deciduous teeth was the statistical equivalent of a drop in the bucket in terms of predicting the decay of the deciduous teeth. In other words, the presence or absence of fluoridated water explained very little of the variation in deciduous tooth decay across the population. Had the authors not been such passionate advocates of fluoridating water supplies, this data would have likely caused them to reassess the necessity of fluoridating water.

Excerpts from Study:

  • SOURCE: Armfield JM, Spencer AJ. 2004. Consumption of nonpublic water: implications for children’s caries experience. Community Dentistry & Oral Epidemiology 32:283-296.


Recently, concerns have been raised over the consumption of both bottled water and water from rainwater tanks. The nature of these concerns reflects varying perspectives. . . . From a dental public health perspective, however, concerns have been voiced for some time about the substitution in the diet of tank or bottled water for fluoridated tap water. More recently, this sentiment has been reiterated by the Centers for Disease Control and Prevention in the US, which stated that the full benefits of community water fluoridation might not be received by persons substituting unfluoridated bottled water for fluoridated mains water. However, while fluoridated bottled water is becoming more common in the US, with more than 20 companies offering the product in 2002, companies in Australia have been reluctant to produce bottled water with what is considered to be appropriate fluoride

. . . .

In Australia the use of domestic rainwater tanks has had a long history. Australia has a generally hot, dry climate and water is therefore widely perceived as a valuable commodity. Certainly, in rural areas, historical problems with connections to mains water and inadequacies with bore water have led to rainwater tanks becoming widely adopted. However, in states such as South Australia, which receives less rainfall than any other state or territory in the country, the prevalence of rainwater tanks in towns and cities is also high. A survey conducted in South Australia in 1996 showed that while 82% of the rural population used rainwater as their primary source of water for drinking, 28% of the population in metropolitan Adelaide, the capital city, also used water from rainwater tanks (29). . . .

Consumption of nonpublic water is increasingly coming to replace mostly fluoridated tap water as the drinking water of choice. This is reflected in the continuing decline in the percentage of adults using mains/town water as their main source of drinking water and the increase in the percentage of adults using bottled water. Although it appears intuitively likely that a decrease in consumption of fluoridated water accompanying the use of bottled or tank water may lead to an increase in caries, literature searches fail to identify previous studies that have investigated this matter. The current study, therefore, sought to describe the relationship between consumption of nonpublic water and dental caries experience among South Australian children with varying levels of lifetime availability of optimally fluoridated water. It was hypothesized that the relationship between nonpublic water consumption and caries would only be significant for children living continuously in fluoridated areas. Additionally, the influence of socioeconomic status (SES) on bottled and tank water consumption was investigated. Although 95.4% of South Australians had access to mains water in 1994 and therefore almost all families have the choice of using public water, it was predicted that children from lower SES families would have higher use of nonpublic water than children from higher SES families.


This study reports on cross-sectional results from the baseline component of a longitudinal study of caries experience in South Australian children conducted between 1991 and 1995. Despite the age of the data set, the data still represent a rare opportunity to investigate this increasingly important oral health issue due to the paucity of scientific data related to the relationship between the consumption of nonpublic water and caries experience.


This study sampled children from fluoridated and nonfluoridated metropolitan and rural areas of South Australia. Children attending the South Australian School Dental Service for a periodic dental examination between June 1991 and May 1992 were randomly sampled on the basis of birth date, the sampling frame varying by the fluoridation status of the region in which a patient lived.


The dmfs (deciduous dentition) and DMFS (permanent dentition) indices, which represent the number of decayed, missing or filled tooth surfaces per child, were the principal outcome variables in this investigation. The mean dmfs index was analysed only for children aged between 4 and 9 years old while the mean DMFS index was analysed only for children aged between the ages of 10 and 15 years old. . . .


A total of 13,911 children were sampled with data obtained on 9988 children aged 3–18 years old. . . .

Mean dmfs and DMFS scores by demographic and socioeconomic variables are shown in Table 1. For both the deciduous and permanent dentition caries experience scores decreased with increasing parental education. Mean dmfs scores were 81.8% and 48.6% higher in the deciduous and permanent dentition, respectively, for children whose parents had not completed secondary school compared with children whose parents had completed university or college. A similar effect was born out for household income, with children from the families with the lowest household income having 52.9% and 68.3% higher mean caries scores in their deciduous and permanent dentitions, respectively, than children from families with incomes over $50 000 per year. The socioeconomic differentials were even more marked with relation to occupational prestige. Children whose parents had no usual occupation had 137.9% and 115.6% higher caries experience in the deciduous and permanent dentitions, respectively, than did children with at least one parent with the highest occupational prestige scores, between 12 and 27 on the Daniel Score. The differences in caries experience by single-parent status were not as appreciable as for the other socioeconomic variables, and for children aged 4–9 was not statistically significant. However, mean DMFS among 10–15-year-old children was 24.4% higher for those children living within a single-parent family compared with children not in a single-parent family and this was statistically significant. Caries experience also varied according to geographical location although this was only statistically significant among 4–9-year-olds in the deciduous dentition, where children living in rural or remote areas had 68.9% and 111.4% higher caries experience, respectively, than children living in metropolitan Adelaide.

Differences in consumption of nonpublic tank and bottled waters by SES are shown in Table 2, with results again presented for both 4–9- and 10–15-year-old children. Using the highest level of education completed by either parent, the relationship with nonpublic water consumption was most clear for the older children, where lower educational attainment was related to increased exposure to nonpublic water. The relationship between SES and water consumption was more apparent for taxable family income, where there was a significant decline in lifetime consumption of nonpublic water across income groups for both the younger and older children. . . .

Table 3. Deciduous (4–9-year-old) and permanent (10–15-year-old) caries experience by consumption of nonpublic water

There was a strong and consistent relationship between consumption of bottled or tank water and caries experience in the deciduous dentition (Table 3). Male and female children who had consumed nonpublic water as their main source of drinking water for their entire life had a mean dmfs score 47.9% and 57.1% higher, respectively, than children who had never consumed tank or bottled water as their main source of drinking water. However, the pattern is less clear in the permanent dentition, where differences across groups for males were minimal and the overall means appeared to be driven by results for females who had a mean DMFS score 11.3% higher for children with 100% lifetime consumption of nonpublic water than for children with 0% consumption of nonpublic water.

To test the significance of the apparent relationship between caries experience and consumption of either tank or bottled water, a series of general linear models were computed using deciduous dmfs and permanent DMFS scores as the dependent variables. To control for variations in the availability of fluoridated tap water, the models were run under three conditions: no lifetime availability of fluoridated tap water, some lifetime availability of fluoridated tap water (between 1% and 99%), and 100% lifetime availability of fluoridated tap water. The distributions of socioeconomic variables and percent lifetime consumption of nonpublic water across these conditions are shown in Table 4. The relationships were similar in both the deciduous and permanent dentitions.

All SES [socioeconomic status] variables were also significantly related to per cent lifetime availability of fluoridated tap water, with lower SES being related to lower lifetime availability and higher SES being characterized by increased availability of fluoridated tap water. . . .

Two models were evaluated under each category of access to fluoridated water. The first looked at the relationship between lifetime consumption of nonpublic water and caries experience after controlling for the demographic variables of age and sex. The second model again examined the relationship between consumption of nonpublic water and caries experience, this time controlling for the demographic variables of age and sex, the sociodemographic variables of family income, parental education, occupational prestige, family type (assessed by single-parent status), residential location, and the potentially confounding fluoride exposure sources of fluoridated toothpaste (assessed by frequency of tooth brushing with fluoridated toothpaste) and fluoride tablets (assessed by whether or not children had at any time consumed fluoride tablets or drops).

Looking at the deciduous dentition, in the first model, per cent lifetime consumption of nonpublic water was entered with age and sex as covariates for children who had had no access to fluoridated water across their lifetime (see Table 5). The relationship between consumption of nonpublic water and dmfs scores was not significant, P > 0.05. In the second model per cent lifetime consumption of nonpublic water was entered with age, sex, family income, parental education, occupational prestige, family type, residential location, frequency of brushing with fluoridated toothpaste and fluoride tablet use. Again, the relationship between consumption of nonpublic water and caries experience was not significant. Similar results were apparent for both models under the second condition, mixed lifetime access to fluoridated water. Again, in both models the relationship between consumption of nonpublic water and dmfs was not significant. However, under the third condition, children who had spent their entire life in localities with fluoridated water available, the relationship between consumption of nonpublic water and dmfs scores was significant after controlling for age and sex. Indeed, after controlling for age, sex, income, education, occupational prestige, family type, residential location, frequency of brushing with fluoridated toothpaste and fluoride tablet use the association between caries experience and per cent lifetime consumption of nonpublic water remained significant, although the effect size (measured using partial eta-squared) of this relationship was small (n2 = 0.01).

Running the same series of models, this time using the permanent DMFS scores as the dependent variable (see Table 6), it can be seen that the effect of consumption of nonpublic water on caries experience was not significant for either of the models under any of the three conditions of access to fluoridated tap water. Although the multivariate models control for residential location, due to the possible confounding of the relationship between SES and consumption of nonpublic water with residential location, the models were also run using only those children living in metropolitan areas. No differences were found in the outcome of the results and these results are therefore not presented here.

. . . .


A less unexpected result of this study, given the findings for the deciduous dentition, was the lack of a significant relationship between consumption of nonpublic water and caries experience in the permanent dentition across any of the differing conditions of access to fluoridated tap water. Earlier research using these data on South Australian children had noted the small absolute mean number of permanent tooth surfaces upon which caries was prevented by exposure to fluoridated water supplies. The benefit was considerably less than noted for the deciduous dentition, and was smaller in South Australia than another comparison state, Queensland. Slade et al. suggested that these results reflected the lower caries experience of children in their permanent dentition, the possible impact of fissure sealants and the possible operation of a halo effect in South Australia. Although the halo effect would be expected to pertain to deciduous as well as to permanent teeth, the lower caries experience in permanent teeth may make the halo effect more of a problem in obtaining significant results in the permanent dentition than in the deciduous dentition. Against this background it was not unexpected that the association for exposure to nonpublic water was not strong, or statistically significant, for the permanent dentition.

Another possible explanation for this is that children with higher lifetime consumption of tank or bottled water are substituting the drinking of nonpublic water for the drinking of soft drinks, effectively counter-balancing the reduced caries prevention of nonpublic water with a reduced consumption of caries promoting soft drinks. In the Australian population there is a noticeable increase between the ages of 4 and 15 in both the total consumption of soft drinks (128.9 average g/person increasing to 525.4 average g/person) and consumption of soft drinks as a percentage of total nonalcoholic beverages (15.0% increasing to 34.4%). It may be that those children consuming bottled water at older ages are substituting water for the generally high consumption of soft drinks of their peers. However, because dietary intake was not assessed in the current study, this possibility could not be investigated. Further explanation of how nonpublic water is consumed and its potential substitution for either fluoridated tap water or sugary, acidic soft drinks and other beverages by age would be highly desirable. The differential effect between the deciduous and permanent dentitions might be explained in a number of ways. Caries activity is greater in the deciduous dentition. For instance at approximately the same interval after the first deciduous and

It is also time that bottled water manufacturers in Australia began marketing fluoridated water. In the US more than 20 companies produce water with optimum fluoride concentrations. This is becoming increasingly important as consumption of tap water decreases and bottled water’s share of all consumed drinks increases. Between 1998 and 2001 there was a further increase in the percentage of Australians consuming bottled water as their main source of drinking water, from 5.1% to 6.9%. However, consumers currently have little choice in Australia and the imminent introduction of fluoride-containing bottled water does not look likely. The Australian Chapter of the International Bottled Water Association proposes that one reason why people are drinking bottled water is actually to avoid chemicals such as fluoride used in the treatment of public water supplies. Bottled water is promoted as a healthy, chemical-free alternative. There is a need for bottled water manufacturers to take a stand on the issue of the benefits of appropriately fluoridated water and provide consumers with choice.