The World Health Organization (WHO) recommends a water fluoride concentration of 1-1.5mg per litre, depending on local climatic conditions.1 This level has been set in order to maximise the caries preventative benefit of Community Water Fluoridation (CWF) while minimising the degree and prevalence of dental fluorosis (which causes stained, mottled teeth). New Zealand was one of the first countries to adopt CWF as a public health intervention to lower rates of dental cavities. Results from early trials found that children born and brought up in fluoridated areas had, on average, 50% fewer cavities compared to children from non-fluoridated areas.2Approximately 56% of New Zealanders have access to fluoridated water.3 The Ministry of Health recommends a water fluoride concentration of 0.7-1.0mg/l to improve oral health; however, responsibility for implementing CWF has remained with local authorities since the 1950s.4The last economic analysis of CWF in New Zealand, by Wright et al (2001), found it was a cost effective intervention in communities of over 1,000 people.5 It has been noted that prevalence of dental caries has decreased over time, both in fluoridated and non-fluoridated communities.2 The aim of the current study was to use more recent data to determine whether community water fluoridation remained a cost effective public health intervention in New Zealand.MethodA cost effectiveness analysis (CEA) was conducted from a societal perspective. CWF at a level of 0.7-1.0mg/l with treatment was evaluated against a treatment only option for the prevention of dental caries. The CEA measured the cost of preventing one newly decayed tooth, rather than a tooth surface,as the effectiveness data for teeth had greater statistical significance than that for tooth surfaces.5In line with Wright et al,5 the CEA examined the relationship between cost effectiveness and community size. Four communities sizes (minor, medium, large 1 and large 2) were identified based on three Register of Drinking Water Suppliers for New Zealand6 derived groupings. The large group was split into two, due to a wide variation in the size of communities included.CEA was conducted separately for children and total population. Fluoride was considered to be effective for all adults that were dentate. Recent evidence suggests the effectiveness of fluoride is both systemic and topical, resulting in a protective effect continuing after teeth have erupted.7Data collectionFluoridation status of water supplies was identified using data from the National Fluoridation Information Service (NFIS), Environmental Scan for 2011-2012.8 Local authorities supplying fluoridated water were contacted to identify the correct person and organisation to direct questions regarding CWF. A questionnaire was emailed to the organisations identified. If replies were not received within two weeks, a follow-up phone call was made and a second email sent. This process was repeated if there was no response after six weeks. After this point, a non-response outcome was assumed and no further contact made.Fixed capital costs, lifespan of plant and machinery, salvage value, and ongoing costs (purchase of chemical, labour from administration and maintenance of CWF and water testing) were requested in the questionnaire. Population which received a fluoridated water supply were cross referenced against the Register of Drinking Water Suppliers for New Zealand.6Cost of adverse side effects (dental fluorosis) was assumed to be negligible and not attributed a value. This was in line with previous studies undertaken.5,9,10 In addition, the most recent New Zealand Oral Health Survey found no statistically significant difference in prevalence of dental fluorosis between 8-30-year-olds living in areas with CWF compared to those in areas without CWF.7Data on mean difference in decayed, missing and filled teeth (dmft: deciduous teeth; DMFT: permanent teeth) between fluoridated and non-fluoridated communities were taken from the New Zealand Oral Health Survey.7 These data represented a mean for the total New Zealand population and were adjusted for age, gender, ethnicity and socioeconomic deprivation. Dental costs averted were identified for total population, and children. The proportion of dentate people within the population was calculated using data from the samesource.7Consistent with previous studies, the following assumptions were made: All carious surfaces would be treated and treatment would comprise of a two surface dental restoration per dmft/DMFT.10-12 Wright et al based their analysis on a dental restoration. Cost was based on the reimbursement rate under the CDA base agreement for the treatment of children and adolescents,13 adjusted to the 2011/2012 financial year using the consumer price index.14This cost may be conservative for total population as adults visiting a private dental practice may be charged at a higher rate. There would be a one hour loss of productivity to attend a dental appointment.9 Indirect costs arising from loss of productivity were based on the average hourly wage rate in 2011.15 The population profile reflects that of the New Zealand population, described in the New Zealand Oral Health Survey,7 from which effectiveness data for CWF was based. Level of effectiveness had already been adjusted for ethnicity, age and gender. Data AnalysisData on costs of CWF received from local authorities supplying fluoridated water was adjusted to the 2011/2012 financial year using the New Zealand Consumer Price Index.14 Average annual equivalised costs were calculated at a base discount rate of 3.5% (as recommended by PHARMAC),16 and a base life span for capital equipment of 15 years based on questionnaire responses and in line with other analyses conducted in Australasia.5,10 The salvage value of plant and machinery (FV) was set at zero in line with questionnaire responses.The Annual Equivalised Cost of capital equipment was calculated and added to the annual costs for the chemical, maintenance and testing. The sum was divided by the population of the community served, to gain an Annual Equivalised Cost per person (AEC per capita).Average AEC per capita was determined for communities with populations of: <5,000 (small); 5,000-10,000 (medium); 10,001-50,000 (large 1) and> 50,000 (large 2).6Present value of cost averted resulting from CWF was calculated for the mean difference in dmft/DMFT between communities with CWF and those without. The cost of treatment was divided by the lifespan of the treatment (12.8 years),17 to determine annual cost averted at a discount rate of 3.5%. Annual cost averted was divided by the percentage of the population who were dentate to give an annual equivalised per person cost averted (AES per capita).Cost effectiveness was calculated using net cost per DMFT averted, by subtracting AES per capita from AEC per capita, (see Appendix 1 for formulae used in the data analysis).Sensitivity AnalysisA univariant sensitivity analysis was conducted. Univariant analysis was used as (with the exception of the discount rate) inputs were prone to parameter uncertainty rather than causing uncertainty to the cost effectiveness ratio as a whole.18Sensitivity analysis was conducted on:Discount rates: 0%, 5% recommended for sensitivity analysis by PHARMAC,16 and 6.4%, the nominal rate recommended by the New Zealand Treasury.19Number of years of life for fluoridation equipment, 8-30 years based on the shortest and longest replacement periods reported for parts in the questionnaire.Number of years of life for a dental restoration, 8-15 years based on protocol from previous studies.5,10ResultsFluoridation Costs QuestionnaireTwenty-eight communities (total population 2,110,778) were identified as receiving fluoridated water from 20 suppliers in 2011/2012. All 20 suppliers were contacted and asked to complete the questionnaire. Sixteen responded, of whom two suppliers refused to answer the questions for commercial reasons, two sets of costs were incomplete and one set of costs was found to be anomalous. The final inclusion rate of 55% covered 11 of the 20 suppliers of fluoridated water, 12 out of 28 fluoridated communities (43%) and a population of 420,616, (20% of those receiving fluoridated water). Responses were evenly distributed geographically and between sizes of community, with two exceptions. Water providers for four out of the five small fluoridated communities responded to the survey. The provider for Auckland, the largest community receiving fluoridated water (population 1,249,858), did not complete the questionnaire.At base rate, the total annual equivalised per capita cost of community water fluoridation ranged from $0.37 to $5.63; mean: $1.95; median: $1.32. Full cost details used in the calculation of AEC per capita can be found in Appendix 2. A strong relationship was evident between community size and cost (Table 1).Table 1: AEC per capita of CWF at base rate (3.5% discount rate, 15 year lifespan of plant and machinery). Size of community served Mean annual equivalised cost (AEC) per capita of CWF Small (< 5000) Range: 744-3,240 $4.38 $5.63-$0.94 Medium (5-10,000) Range: 7,542-9,710 $1.23 $0.71-$1.75 Large 1 (10-50,000) Range: 19,000-34,300 $0.66 $0.68-$0.63 Large 2 (>50,000) Range: 74,953-343,900 $0.53 $0.88-$0.37 Costs Averted from Community Water FluoridationThe difference in mean dmft/DMFT between communities with and without CWF was 1.0 dmft/DMFT (p<0.05) for children (<18 years) and 0.8 dmft/DMFT (p<0.05), for total population, see Table 2.7Table 2: Difference in mean dmft/DMFT between fluoridated and non-fluoridated communities Source: New Zealand Oral Health Survey (2009)7 Mean dmft/ DMFT non-fluoridated area (95% CI) Mean dmft/ DMFT fluoridated area (95% CI) Ratio of means Difference in mean dmft/DMFT Children (2-17 years) 2.4 (2.0-2.6) 1.5 (1.1-1.9) 1.7 1.0 Total (> 2 years) 12.1 (11.5-12.7) 9.6 (9.1-10.1) 1.1 0.8 Costs averted, adjusted to 2011 prices, were $82.33 for a two surface dental restoration and $20.64 for one hours loss of productivity to attend the dental appointment, based on the average wage in 2011.The annual equivalised per capita costs averted resulting from CWF was $4.82 for total population and $5.21 for children.Cost effectiveness of CWFCWF represented a negative net cost (cost saving) at base rate per dmft/DMFT averted for total population and children in all four communities. This indicated that CWF with treatment was a more cost effective oral health intervention than treatment alone for those groups (Table 3).Table 3: Average annual per capita net cost of CWF per dmft/DMFT averted. Population Cost Effectiveness by Community Size ($ per dmft/DMFT averted) Small: (<5000) Medium: (5,000-10,000) Large 1: (10,001-50000) Large 2: (>50,000) Child -$2.23 -$5.38 -$5.95 -$6.09 Total -$0.55 -$4.49 -$5.20 -$5.38 Sensitivity AnalysisA negative net cost (cost saving) per dmft/DMFT averted remained for CWF for total population and children in all communities with populations of over 5,000 when the variables were adjusted for sensitivity analysis, (Table 4).Table 4: Sensitivity Analysis Range for CWF (net cost per dmft/DMFT averted). Population Value Range Cost Effectiveness Range by Community Size ($ per dmft/DMFT averted) Small (<5,000) Medium (5,000-10,000) Large 1 (10,001-50,000) Large 2 (>50,000) Child Low +$2.5 -$1.71 -$2.67 -$2.8 High -$9.57 -$11.98 -$13.71 -$13.81 Total Low +$4.24 -$1.03 -$2.23 -$2.39 High -$7.78 -$10.79 -$11.19 -$11.31 Positive values for communities under 5,000 at the \u2018low end of the value range indicate that CWF with treatment was less cost effective than treatment alone under certain scenarios: when the discount rate was increased above 3.5%, the lifespan of plant was reduced from 15 to 8 years; when the lifespan of a dental restoration was increased from 12.8 to 15 years; and when only direct costs (cost of treatment but not loss of productivity) were included.CWF with treatment was less cost effective than treatment alone, for total population in small communities (<5,000) at all discount rates when lifespan of plant and machinery was <8 years and only direct costs were included. When the base discount rate (3.5%) was used, CWF with treatment was less cost effective than treatment alone when only direct costs were included. When the discount rate was increased to 5% and above, CWF had a positive cost effectiveness ratio when all other variables were at base case.In small communities (<5,000), CWF with treatment was less cost effective than treatment alone, for children at the base discount rate when a dental restoration was assumed to last>15 years. At a discount rate of 6.4% CWF with treatment was less cost effective than treatment alone for children when all other variables were at base case.DiscussionThe baseline results, which show a negative net cost per dmft/DMFT averted for CWF with treatment when compared to treatment alone, were consistent with Wright et al, and a number of other economic analyses undertaken since.5,9-12,20-25 Cost effectiveness remained, despite an overall reduction in dental caries in the population. In other studies, levels of effectiveness for CWF have varied between 50-60%,11 30%,26 25%22 and 15%.23 Even where gains in effectiveness were modest, savings in dental treatment mean that CWF could still provide a net saving to society both in monetary and oral health terms.Community size and cost effectiveness of CWFDue to higher per capita costs for implementation and maintenance, CWF with treatment was less cost effective than treatment alone when lifespan of a filling was reduced to eight years (in line with Wright et al) in small communities. CWF with treatment was also less cost effective than treatment alone for these communities where the lifespan of plant and machinery were reduced and the discount rate increased. These findings were similar to those of previous studies.5,22 Wright et al identified a \u2018break-even community size for CWF of 700-900 people.5 The four communities making up the \u2018small classification in this study ranged from 744 to 3,240 people.Effectiveness of CWFThe levels of effectiveness used in this study were based on adjusted mean differences in dmft/DMFT between fluoridated and non-fluoridated communities, reported by the 2009 New Zealand Oral Health Survey.7 The unadjusted differences represented a 20.7% reduction in dmft/DMFT for the total population and a 47.5% reduction in dmft/DMFT for children. The figure for total population demonstrates a lower level of effectiveness for CWF than found in most previous studies.11,22,26 Despite lower levels of effectiveness, CWF remained cost effective. This finding was supported by Tchouaket et al, who found that cost savings could be achieved even at an effectiveness level of 1%.26In addition to differences in effectiveness between age groups and over time, some populations are more at risk of caries than others.25,5 The analysis undertaken in this study was based on the New Zealand population as a whole. Communities that have a higher risk of dental caries would show greater cost-effectiveness. This might make small rural communities, where the population has a greater proportion of M\u0101ori and/or low-income families, more appropriate candidates for CWF than communities with larger but more affluent populations. Wright et al identified that cost effectiveness increased in areas where the proportion of the population that were M\u0101ori was above the national rate.5 Lopez25 identified higher levels of cost savings from CWF amongst (low income) medicaid recipients in Texas. Differences of this sort were also demonstrated by Birch using a hypothetical model of the UK.27LimitationsThis cost effective analysis has a number of limitations. Costs which were not included, for example adverse effects and \u2018selling CWF to the community, may have led to an over-estimation of the cost effectiveness of CWF. Likewise, cost savings which were not included, such as pain and suffering averted and dental treatments beyond a two surface dental restoration, may have led to an under-estimation of the cost-effectiveness of CWF.Three limitations arose from the data collection.Firstly, not all providers responded to the questionnaire.Secondly, of those providers that did reply, data for five providers was excluded. Data suitable for use in this study equated to around 20% of people who received fluoridated water in New Zealand in 2011/2012.Thirdly, the open nature of the questions gave a wide scope for attributing costs. While this allowed local authorities to attribute the costs of CWF as they perceived them, it did lead to a lack of consistency between local authorities.The data for the effectiveness of CWF was taken from the 2009 New Zealand Oral Health Survey. Fluoridation status was determined by where the respondent lived at the time of the survey. Movement of individuals between communities, resulting in different levels of exposure, and length of time the community had received fluoridated water was not taken into account. The lack of information on lifetime exposure particularly limits the reliability of the data for adults. In addition, confounding factors, such as oral hygiene habits, exposure to other fluoride sources and diet, were not controlled for. It was also assumed that for adults, over 45 years of age, missing teeth had been extracted solely for pathological reasons.9ConclusionThe cost effectiveness analysis supports the findings of an earlier economic analysis of water fluoridation in New Zealand by Wright et al.5 CWF is a cost effective public health intervention in New Zealand for all populations. For smaller communities cost effectiveness is more marginal and dependent on the population profile of the community. Appendix 1: Equations used in the cost effectiveness calculation: Annual equivalent cost of capital for CWF: Where:PMTc = Annual equivalent cost of capital equipment (fluoridation equipment)PV = Present value (purchased cost of the asset)FV = Future value (salvage cost (if any) of the asset when it is replaced)DR = Discount rateT = lifespan of the asset Annual equivalised cost per capita of CWF: Where:PMT = Annual equivalent cost of capital equipmentC = annual cost of chemicalm = cost of maintenance, labour and administrationt = cost of independent testing of water fluoride concentration Costs averted resulting from CWF: Where:PV = present value of dental treatmentdc = direct cost (of a two surface amalgam filling)ic = indirect cost (productivity loss to seek treatment)dif(\u00b5DMFT) = the difference in mean DMFT (for the relevant population) resulting from CWF Annual costs averted resulting from CWF: Where: PMTs = Annual equivalent costs averted PV = Present value (reimbursement rate for treatment)DR = Discount rate (3.5%)T = lifespan of the treatment (12.8 years) Annual equivalised per capita costs averted: Cost effectiveness per dmft/DMFT averted: Where: AEC-AES = net cost of CWF dif(\u00b5DMFT) = difference in mean DMFT resulting from CWF Appendix 2: Cost components of CWF by community # Based on \"Register of Drinking Water Suppliers for New Zealand\" (2013 edition): http://www.esr.cri.nz/SiteCollectionDocuments/ESR/PDF/WaterManagement/RegisterOfSuppliers.pdf * Tank (30 years+), Dosing Pump (8 years), Pipework, Day Tank **10/12 use hydrofluorosicic acid: two use sodium silico powder no significant difference was found in the costs of these two chemicals *** Costs vary from purely external testing ($500 p.a.) to including maintenance and calibration, consumables (parts), safety and administration ($22,000 p.a.)

The World Health Organization (WHO) recommends a water fluoride concentration of 1-1.5mg per litre, depending on local climatic conditions.1 This level has been set in order to maximise the caries preventative benefit of Community Water Fluoridation (CWF) while minimising the degree and prevalence of dental fluorosis (which causes stained, mottled teeth). New Zealand was one of the first countries to adopt CWF as a public health intervention to lower rates of dental cavities. Results from early trials found that children born and brought up in fluoridated areas had, on average, 50% fewer cavities compared to children from non-fluoridated areas.2Approximately 56% of New Zealanders have access to fluoridated water.3 The Ministry of Health recommends a water fluoride concentration of 0.7-1.0mg/l to improve oral health; however, responsibility for implementing CWF has remained with local authorities since the 1950s.4The last economic analysis of CWF in New Zealand, by Wright et al (2001), found it was a cost effective intervention in communities of over 1,000 people.5 It has been noted that prevalence of dental caries has decreased over time, both in fluoridated and non-fluoridated communities.2 The aim of the current study was to use more recent data to determine whether community water fluoridation remained a cost effective public health intervention in New Zealand.MethodA cost effectiveness analysis (CEA) was conducted from a societal perspective. CWF at a level of 0.7-1.0mg/l with treatment was evaluated against a treatment only option for the prevention of dental caries. The CEA measured the cost of preventing one newly decayed tooth, rather than a tooth surface,as the effectiveness data for teeth had greater statistical significance than that for tooth surfaces.5In line with Wright et al,5 the CEA examined the relationship between cost effectiveness and community size. Four communities sizes (minor, medium, large 1 and large 2) were identified based on three Register of Drinking Water Suppliers for New Zealand6 derived groupings. The large group was split into two, due to a wide variation in the size of communities included.CEA was conducted separately for children and total population. Fluoride was considered to be effective for all adults that were dentate. Recent evidence suggests the effectiveness of fluoride is both systemic and topical, resulting in a protective effect continuing after teeth have erupted.7Data collectionFluoridation status of water supplies was identified using data from the National Fluoridation Information Service (NFIS), Environmental Scan for 2011-2012.8 Local authorities supplying fluoridated water were contacted to identify the correct person and organisation to direct questions regarding CWF. A questionnaire was emailed to the organisations identified. If replies were not received within two weeks, a follow-up phone call was made and a second email sent. This process was repeated if there was no response after six weeks. After this point, a non-response outcome was assumed and no further contact made.Fixed capital costs, lifespan of plant and machinery, salvage value, and ongoing costs (purchase of chemical, labour from administration and maintenance of CWF and water testing) were requested in the questionnaire. Population which received a fluoridated water supply were cross referenced against the Register of Drinking Water Suppliers for New Zealand.6Cost of adverse side effects (dental fluorosis) was assumed to be negligible and not attributed a value. This was in line with previous studies undertaken.5,9,10 In addition, the most recent New Zealand Oral Health Survey found no statistically significant difference in prevalence of dental fluorosis between 8-30-year-olds living in areas with CWF compared to those in areas without CWF.7Data on mean difference in decayed, missing and filled teeth (dmft: deciduous teeth; DMFT: permanent teeth) between fluoridated and non-fluoridated communities were taken from the New Zealand Oral Health Survey.7 These data represented a mean for the total New Zealand population and were adjusted for age, gender, ethnicity and socioeconomic deprivation. Dental costs averted were identified for total population, and children. The proportion of dentate people within the population was calculated using data from the samesource.7Consistent with previous studies, the following assumptions were made: All carious surfaces would be treated and treatment would comprise of a two surface dental restoration per dmft/DMFT.10-12 Wright et al based their analysis on a dental restoration. Cost was based on the reimbursement rate under the CDA base agreement for the treatment of children and adolescents,13 adjusted to the 2011/2012 financial year using the consumer price index.14This cost may be conservative for total population as adults visiting a private dental practice may be charged at a higher rate. There would be a one hour loss of productivity to attend a dental appointment.9 Indirect costs arising from loss of productivity were based on the average hourly wage rate in 2011.15 The population profile reflects that of the New Zealand population, described in the New Zealand Oral Health Survey,7 from which effectiveness data for CWF was based. Level of effectiveness had already been adjusted for ethnicity, age and gender. Data AnalysisData on costs of CWF received from local authorities supplying fluoridated water was adjusted to the 2011/2012 financial year using the New Zealand Consumer Price Index.14 Average annual equivalised costs were calculated at a base discount rate of 3.5% (as recommended by PHARMAC),16 and a base life span for capital equipment of 15 years based on questionnaire responses and in line with other analyses conducted in Australasia.5,10 The salvage value of plant and machinery (FV) was set at zero in line with questionnaire responses.The Annual Equivalised Cost of capital equipment was calculated and added to the annual costs for the chemical, maintenance and testing. The sum was divided by the population of the community served, to gain an Annual Equivalised Cost per person (AEC per capita).Average AEC per capita was determined for communities with populations of: <5,000 (small); 5,000-10,000 (medium); 10,001-50,000 (large 1) and> 50,000 (large 2).6Present value of cost averted resulting from CWF was calculated for the mean difference in dmft/DMFT between communities with CWF and those without. The cost of treatment was divided by the lifespan of the treatment (12.8 years),17 to determine annual cost averted at a discount rate of 3.5%. Annual cost averted was divided by the percentage of the population who were dentate to give an annual equivalised per person cost averted (AES per capita).Cost effectiveness was calculated using net cost per DMFT averted, by subtracting AES per capita from AEC per capita, (see Appendix 1 for formulae used in the data analysis).Sensitivity AnalysisA univariant sensitivity analysis was conducted. Univariant analysis was used as (with the exception of the discount rate) inputs were prone to parameter uncertainty rather than causing uncertainty to the cost effectiveness ratio as a whole.18Sensitivity analysis was conducted on:Discount rates: 0%, 5% recommended for sensitivity analysis by PHARMAC,16 and 6.4%, the nominal rate recommended by the New Zealand Treasury.19Number of years of life for fluoridation equipment, 8-30 years based on the shortest and longest replacement periods reported for parts in the questionnaire.Number of years of life for a dental restoration, 8-15 years based on protocol from previous studies.5,10ResultsFluoridation Costs QuestionnaireTwenty-eight communities (total population 2,110,778) were identified as receiving fluoridated water from 20 suppliers in 2011/2012. All 20 suppliers were contacted and asked to complete the questionnaire. Sixteen responded, of whom two suppliers refused to answer the questions for commercial reasons, two sets of costs were incomplete and one set of costs was found to be anomalous. The final inclusion rate of 55% covered 11 of the 20 suppliers of fluoridated water, 12 out of 28 fluoridated communities (43%) and a population of 420,616, (20% of those receiving fluoridated water). Responses were evenly distributed geographically and between sizes of community, with two exceptions. Water providers for four out of the five small fluoridated communities responded to the survey. The provider for Auckland, the largest community receiving fluoridated water (population 1,249,858), did not complete the questionnaire.At base rate, the total annual equivalised per capita cost of community water fluoridation ranged from $0.37 to $5.63; mean: $1.95; median: $1.32. Full cost details used in the calculation of AEC per capita can be found in Appendix 2. A strong relationship was evident between community size and cost (Table 1).Table 1: AEC per capita of CWF at base rate (3.5% discount rate, 15 year lifespan of plant and machinery). Size of community served Mean annual equivalised cost (AEC) per capita of CWF Small (< 5000) Range: 744-3,240 $4.38 $5.63-$0.94 Medium (5-10,000) Range: 7,542-9,710 $1.23 $0.71-$1.75 Large 1 (10-50,000) Range: 19,000-34,300 $0.66 $0.68-$0.63 Large 2 (>50,000) Range: 74,953-343,900 $0.53 $0.88-$0.37 Costs Averted from Community Water FluoridationThe difference in mean dmft/DMFT between communities with and without CWF was 1.0 dmft/DMFT (p<0.05) for children (<18 years) and 0.8 dmft/DMFT (p<0.05), for total population, see Table 2.7Table 2: Difference in mean dmft/DMFT between fluoridated and non-fluoridated communities Source: New Zealand Oral Health Survey (2009)7 Mean dmft/ DMFT non-fluoridated area (95% CI) Mean dmft/ DMFT fluoridated area (95% CI) Ratio of means Difference in mean dmft/DMFT Children (2-17 years) 2.4 (2.0-2.6) 1.5 (1.1-1.9) 1.7 1.0 Total (> 2 years) 12.1 (11.5-12.7) 9.6 (9.1-10.1) 1.1 0.8 Costs averted, adjusted to 2011 prices, were $82.33 for a two surface dental restoration and $20.64 for one hours loss of productivity to attend the dental appointment, based on the average wage in 2011.The annual equivalised per capita costs averted resulting from CWF was $4.82 for total population and $5.21 for children.Cost effectiveness of CWFCWF represented a negative net cost (cost saving) at base rate per dmft/DMFT averted for total population and children in all four communities. This indicated that CWF with treatment was a more cost effective oral health intervention than treatment alone for those groups (Table 3).Table 3: Average annual per capita net cost of CWF per dmft/DMFT averted. Population Cost Effectiveness by Community Size ($ per dmft/DMFT averted) Small: (<5000) Medium: (5,000-10,000) Large 1: (10,001-50000) Large 2: (>50,000) Child -$2.23 -$5.38 -$5.95 -$6.09 Total -$0.55 -$4.49 -$5.20 -$5.38 Sensitivity AnalysisA negative net cost (cost saving) per dmft/DMFT averted remained for CWF for total population and children in all communities with populations of over 5,000 when the variables were adjusted for sensitivity analysis, (Table 4).Table 4: Sensitivity Analysis Range for CWF (net cost per dmft/DMFT averted). Population Value Range Cost Effectiveness Range by Community Size ($ per dmft/DMFT averted) Small (<5,000) Medium (5,000-10,000) Large 1 (10,001-50,000) Large 2 (>50,000) Child Low +$2.5 -$1.71 -$2.67 -$2.8 High -$9.57 -$11.98 -$13.71 -$13.81 Total Low +$4.24 -$1.03 -$2.23 -$2.39 High -$7.78 -$10.79 -$11.19 -$11.31 Positive values for communities under 5,000 at the \u2018low end of the value range indicate that CWF with treatment was less cost effective than treatment alone under certain scenarios: when the discount rate was increased above 3.5%, the lifespan of plant was reduced from 15 to 8 years; when the lifespan of a dental restoration was increased from 12.8 to 15 years; and when only direct costs (cost of treatment but not loss of productivity) were included.CWF with treatment was less cost effective than treatment alone, for total population in small communities (<5,000) at all discount rates when lifespan of plant and machinery was <8 years and only direct costs were included. When the base discount rate (3.5%) was used, CWF with treatment was less cost effective than treatment alone when only direct costs were included. When the discount rate was increased to 5% and above, CWF had a positive cost effectiveness ratio when all other variables were at base case.In small communities (<5,000), CWF with treatment was less cost effective than treatment alone, for children at the base discount rate when a dental restoration was assumed to last>15 years. At a discount rate of 6.4% CWF with treatment was less cost effective than treatment alone for children when all other variables were at base case.DiscussionThe baseline results, which show a negative net cost per dmft/DMFT averted for CWF with treatment when compared to treatment alone, were consistent with Wright et al, and a number of other economic analyses undertaken since.5,9-12,20-25 Cost effectiveness remained, despite an overall reduction in dental caries in the population. In other studies, levels of effectiveness for CWF have varied between 50-60%,11 30%,26 25%22 and 15%.23 Even where gains in effectiveness were modest, savings in dental treatment mean that CWF could still provide a net saving to society both in monetary and oral health terms.Community size and cost effectiveness of CWFDue to higher per capita costs for implementation and maintenance, CWF with treatment was less cost effective than treatment alone when lifespan of a filling was reduced to eight years (in line with Wright et al) in small communities. CWF with treatment was also less cost effective than treatment alone for these communities where the lifespan of plant and machinery were reduced and the discount rate increased. These findings were similar to those of previous studies.5,22 Wright et al identified a \u2018break-even community size for CWF of 700-900 people.5 The four communities making up the \u2018small classification in this study ranged from 744 to 3,240 people.Effectiveness of CWFThe levels of effectiveness used in this study were based on adjusted mean differences in dmft/DMFT between fluoridated and non-fluoridated communities, reported by the 2009 New Zealand Oral Health Survey.7 The unadjusted differences represented a 20.7% reduction in dmft/DMFT for the total population and a 47.5% reduction in dmft/DMFT for children. The figure for total population demonstrates a lower level of effectiveness for CWF than found in most previous studies.11,22,26 Despite lower levels of effectiveness, CWF remained cost effective. This finding was supported by Tchouaket et al, who found that cost savings could be achieved even at an effectiveness level of 1%.26In addition to differences in effectiveness between age groups and over time, some populations are more at risk of caries than others.25,5 The analysis undertaken in this study was based on the New Zealand population as a whole. Communities that have a higher risk of dental caries would show greater cost-effectiveness. This might make small rural communities, where the population has a greater proportion of M\u0101ori and/or low-income families, more appropriate candidates for CWF than communities with larger but more affluent populations. Wright et al identified that cost effectiveness increased in areas where the proportion of the population that were M\u0101ori was above the national rate.5 Lopez25 identified higher levels of cost savings from CWF amongst (low income) medicaid recipients in Texas. Differences of this sort were also demonstrated by Birch using a hypothetical model of the UK.27LimitationsThis cost effective analysis has a number of limitations. Costs which were not included, for example adverse effects and \u2018selling CWF to the community, may have led to an over-estimation of the cost effectiveness of CWF. Likewise, cost savings which were not included, such as pain and suffering averted and dental treatments beyond a two surface dental restoration, may have led to an under-estimation of the cost-effectiveness of CWF.Three limitations arose from the data collection.Firstly, not all providers responded to the questionnaire.Secondly, of those providers that did reply, data for five providers was excluded. Data suitable for use in this study equated to around 20% of people who received fluoridated water in New Zealand in 2011/2012.Thirdly, the open nature of the questions gave a wide scope for attributing costs. While this allowed local authorities to attribute the costs of CWF as they perceived them, it did lead to a lack of consistency between local authorities.The data for the effectiveness of CWF was taken from the 2009 New Zealand Oral Health Survey. Fluoridation status was determined by where the respondent lived at the time of the survey. Movement of individuals between communities, resulting in different levels of exposure, and length of time the community had received fluoridated water was not taken into account. The lack of information on lifetime exposure particularly limits the reliability of the data for adults. In addition, confounding factors, such as oral hygiene habits, exposure to other fluoride sources and diet, were not controlled for. It was also assumed that for adults, over 45 years of age, missing teeth had been extracted solely for pathological reasons.9ConclusionThe cost effectiveness analysis supports the findings of an earlier economic analysis of water fluoridation in New Zealand by Wright et al.5 CWF is a cost effective public health intervention in New Zealand for all populations. For smaller communities cost effectiveness is more marginal and dependent on the population profile of the community. Appendix 1: Equations used in the cost effectiveness calculation: Annual equivalent cost of capital for CWF: Where:PMTc = Annual equivalent cost of capital equipment (fluoridation equipment)PV = Present value (purchased cost of the asset)FV = Future value (salvage cost (if any) of the asset when it is replaced)DR = Discount rateT = lifespan of the asset Annual equivalised cost per capita of CWF: Where:PMT = Annual equivalent cost of capital equipmentC = annual cost of chemicalm = cost of maintenance, labour and administrationt = cost of independent testing of water fluoride concentration Costs averted resulting from CWF: Where:PV = present value of dental treatmentdc = direct cost (of a two surface amalgam filling)ic = indirect cost (productivity loss to seek treatment)dif(\u00b5DMFT) = the difference in mean DMFT (for the relevant population) resulting from CWF Annual costs averted resulting from CWF: Where: PMTs = Annual equivalent costs averted PV = Present value (reimbursement rate for treatment)DR = Discount rate (3.5%)T = lifespan of the treatment (12.8 years) Annual equivalised per capita costs averted: Cost effectiveness per dmft/DMFT averted: Where: AEC-AES = net cost of CWF dif(\u00b5DMFT) = difference in mean DMFT resulting from CWF Appendix 2: Cost components of CWF by community # Based on \"Register of Drinking Water Suppliers for New Zealand\" (2013 edition): http://www.esr.cri.nz/SiteCollectionDocuments/ESR/PDF/WaterManagement/RegisterOfSuppliers.pdf * Tank (30 years+), Dosing Pump (8 years), Pipework, Day Tank **10/12 use hydrofluorosicic acid: two use sodium silico powder no significant difference was found in the costs of these two chemicals *** Costs vary from purely external testing ($500 p.a.) to including maintenance and calibration, consumables (parts), safety and administration ($22,000 p.a.)