Calculating Fluoridation Statistics
Monitoring health status and access to preventive interventions and evaluating quality of population-based health services are among the 10 essential public health services. National- and state-based surveys monitor the oral health status of the U.S. population and the Water Fluoridation Reporting System (WFRS) compiles data useful for monitoring access to and evaluating the quality of water fluoridation programs at the state and water system levels. WFRS data have been summarized every two years since 2000, and these summary reports provide estimates for tracking the Healthy People 2010 objective: Increase the proportion of the U.S. population served by community water systems with optimally fluoridated water.
Reporting of Fluoridated Populations by CDC
The U.S. Public Health Service periodically has reported fluoridation statistics since 1956. The first national Water Fluoridation Census (Census) was published in 1963, and then periodically through 1992. The Fluoridation Census presented the fluoridation status of community water systems and summary statistics for each state and county. It was compiled from the EPA, U.S. Census Bureau, and surveys of the state drinking water and oral health programs. The CDC has been the lead agency in this effort since 1975, and an objective for water fluoridation has been included among the Healthy People national objectives since 1979.
Compiling and publishing the Water Fluoridation Census was a resource-intensive effort; some states lacked the resources to provide accurate information. Biennial fluoridation reports from 2000–2008 reported state and national statistics for water fluoridation, whereas the Water Fluoridation Census also included a comprehensive listing of the fluoridation status of individual community water systems. The 2000 report was the first to summarize water system data submitted by states through the Web-based Water Fluoridation Reporting System (WFRS). Aside from the use of a Web-based reporting system, the data collection methodology and data sources remained quite similar from the earliest reports through the 2004 report.
With the 2006 report, the U.S. Census Bureau data were used differently. Population estimates were used instead of decennial census data; the 2006 vintage of population estimates for 2002 and 2004 was used to revise previously published fluoridation and water system population estimates to improve comparability of estimates published between U.S. censuses. Instead of capping state water system populations at the 2000 census population, for states in which the state water system population exceeded the state population estimate, an adjustment factor was calculated. The factor is a ratio of the U.S. census population estimate and the state water system population. This adjustment factor was applied to reduce the reported number of persons receiving fluoridated water and number of persons served by community water systems. Details of the change in methodology are reported in the 2006 report: Populations receiving optimally fluoridated public drinking water – United States, 1992–2006. MMWR 2008; 57(27):737–741. Further details are provided below.
Estimation of Populations by State Programs
Estimating the precise service population for water systems is difficult. Challenges include estimating the population per service connection, determining whether the connection is at a primary or secondary residence, and harmonizing different estimation methodologies used by different states. When using a service population estimate, it is important to understand how that estimate was derived.
Each state drinking water program derives its own methodology for estimating water system service populations. Typically, water systems track the number of connections or accounts for billing purposes and not the number of people served by that account. Therefore, most service populations are estimates derived from the number of connections, normally based on an assumed number of people per connection. This can introduce errors, however, because even though a state average might be 2.4 people per household based on census figures, a state drinking water program might use an average of 2.6 people per billing account to address the accounts serving multiple dwelling units on a single meter, such as the case with apartments. Although this may provide a satisfactory service population estimate for the state as a whole, it may result in discrepancies for individual water systems, because the applied factor is an assumed value and may be inconsistent between communities.
Double-counting of individuals is a possible limitation. Water systems may report total people served, which could include commuters that would also be counted as a resident in an adjacent water system, or people with primary and secondary residences, such as college students or recreational homes. It is key to identify which location is an individual’s primary residence.
EPA Data Resources on Populations
The Environmental Protection Agency (EPA) uses the Safe Drinking Water Information System (SDWIS) to track approximately 160,000 public water systems. Many public water systems serve schools, gas stations, parks, and other service districts that may have only a few or no residents at all. Of the public water systems tracked by EPA, approximately 54,000 are considered community water systems because they serve a threshold number of residents, in addition to other uses. CDC bases the fluoridation status of a state on the community water systems.
State drinking water programs normally update populations served by water systems during the Sanitary Survey process, which is conducted every three years for most surface water systems and every four years for most groundwater systems. The Sanitary Survey process is a detailed review of a water system and its condition. States may update the population estimate for an individual water system more often if new information is brought to their attention, but they would unlikely update the estimates for all water systems in any single year. The EPA Web siteExternal presents more information on SDWIS and documentation on its data fields, definitions, and uses.
EPA SDWIS Data Fields
State SDWIS administrators generally update SDWIS files by December, and these revisions are compiled by EPA by February of the following year. CDC and EPA have agreed to allow comparison of data concerning water fluoridation and population between SDWIS and WFRS. The two databases can be compared because both are written using the Public Water System (PWS) ID number as the common reference. CDC conducts an annual comparison to identify discrepancies between the databases, which can then be used to improve the records. CDC compares the year-end data in SDWIS to the data entered in WFRS the following April or May, because many states do not complete their fluoridation records update until March 1. Updating the information in both WFRS and SDWIS is a state program responsibility.
SDWIS is capable of identifying up to four service populations for each water system. Of these, the one most useful for the population estimate in the water fluoridation report is the resident population. The resident population in SDWIS-Fed is identified as RetPopSrvd, and is normally in the field known as TINSWYS column in SDWIS-State entitled D_POPULATION_COUNT/AVG_DAILY_CNT. The resident population should represent the service population expressed as the location of residence. Using the resident population systematically should minimize the number of people that might be double-counted because each person is counted only once for location of residence, presuming that the state drinking water program has carefully addressed this issue.
Additional Considerations on Population Estimates
A state program sometimes contacts an individual water system to find out whether that water system has made a concerted effort to derive a more precise service population. Sometimes a system will compile a more precise number for a municipal bond report or other accounting purpose.
In every state, a small portion of the population consumes water from unregulated private wells. As a result, the reported population consuming water from community water systems typically is less than the U.S. Census population for that state. CDC reports percentage of fluoridated populations relative to the populations served by community water systems in that state.
Many state drinking water programs or state oral health programs compare the compiled state totals entered for individual water systems and against the state census totals to determine whether the compiled totals appear reasonable. They will then adjust the populations to be entered into WFRS by a factor so that the compiled totals accurately reflect the actual state population.
For a few states, the total reported service population on community water systems exceeds the estimated census population for that state. If the compiled total in WFRS is greater than the estimated census population for that state, then the service populations on community water systems will need to be adjusted for use in establishing the total population in a state with access to fluoridated water. Because the state will have estimated the service populations in a consistent manner, the proportion of fluoridated population to the total population as reported by WFRS will be accurate. This percentage can be applied to the census total for the state to derive the fluoridated population for the Water Fluoridation Report.
Reference Census Populations
The U.S. Census Bureau conducts the decennial census every 10 years, and estimates the population at mid-year for years after the most recent decennial census, using data from the decennial Census and the annual American Community Survey in a statistical model of population change. The Water Fluoridation Reports use these population estimates to adjust state estimates of the populations receiving fluoridated water and the community water system (CWS) population for states in which the CWS population exceeds the state’s population estimates.
Details of the change in methodology are reported in the 2006 report Populations receiving optimally fluoridated public drinking water – United States, 1992–2006. MMWR 2008; 57(27):737–741.