BRFSS Drinking Water Results Colorado Department of Public Health and Environment Drinking Water Supply and Consumption Trends: Results from the Behavioral Risk Factor Surveillance System Water Quality Control Division (WQCD), Center for Health and Environmental Information and Statistics (CHEIS) and the Environmental Information Management Unit (EIMU), Colorado Department of Public Health and Environment Introduction The Behavioral Risk Factor Surveillance System (BRFSS) is the longest running and largest ongoing collection of public health behavior data in the United States. The BRFSS is administered by the Centers for Disease Control and Prevention (CDC) and supported by grants to individual states. In Colorado the BRFSS is administered by the Center for Health and Environmental Information and Statistics (CHEIS) a division of the Colorado Department of Public Health and Environment (CDPHE). Household drinking water is typically supplied through one of two sources: a public water system or a private well. Public drinking water supplies are regulated federally by the Safe Drinking Water Act (SDWA) and overseen by the Colorado Department of Public Health and Environment Water Quality Control Division (WQCD). Levels of contaminants are monitored in water sources and in treated water supplies to ensure the safety of drinking water. Multiple barriers of removal are used to help ensure that contaminants are not present at levels exceeding regulatory limits. Disinfection ensures that bacteria and viruses pose no unacceptable risk in the public water supply. In 2009 and 2011, CDPHE added questions to the Colorado Behavioral Risk Factor Surveillance System (BRFSS). Included in this set of environmental health indicators was a series of questions on household water supply and drinking water consumption among survey participants. The survey was designed to determine the source of Colorado residents’ home water supply: a city, county, or town water system, small water system operated by a homeowners association or a private well. Survey participants were also asked what type of water they most often drink at home: unfiltered tap water, filtered tap water or bottled water. For the purpose of this report the results indicating a small water system as the water source can be grouped with responses indicating a community water system, to indicate service by a public water system, since both are regulated for public health and safety. If a participant’s water came from a private well they were asked how often they tested the well, if specific contaminants were tested for, and was the test positive for that contaminant. All participants were asked if they drank unfiltered tap water, filtered tap water, bottled or vended water while at home regardless of the source of their home water supply. Participants that answered that they drank filtered, bottled or vended water were asked why they filter their tap water, or why they choose bottled or vended water over unfiltered tap water. In 2009 85% of residents in Colorado received water from either a municipal public water system, or a small water system operated by a homeowners association, for example. According to the 2011 BRFSS that number has remained constant at 85%. Further, approximately 9% of residents get water from a private well. Private wells are not regulated and the water quality varies depending on surrounding geology, land use and well depth. It is recommended that home owners with a private well test their water quality regularly for the presence of contaminants that are otherwise regulated by the SDWA in public water. Bottled drinking water is regulated by the Food and Drug Administration (FDA) which sets bottled water standards based on the EPA standards for safe drinking water. Both bottled and tap waters are considered safe to drink if they meet the standards. The Centers for Disease Control and Prevention (CDC) also provides information about commercially bottled water and public health.
1
Because WQCD is responsible for the oversight of public water system compliance with drinking water regulations, the results of this survey can help the WQCD understand trends in drinking water supply and consumption when developing public policy and prioritizing resources. Of particular interest to CDPHE are the results for bottled water consumption. Choosing bottled water over tap water when tap water treated by a public water system meets health standards for quality highlights an opportunity to improve program outreach. By educating consumers about the benefits of choosing drinking water from a public water system, CDPHE can begin to address disparities observed in some of these populations. Not only does water treatment reduce or eliminate certain harmful chemicals but treated public drinking water supplies often contain levels of fluoride that promote oral health. Further, in some cases economic burden has been associated with the cost of bottled water while environmental burden has been associated with the packaging. Methodology This report is based on the Colorado added questions to the BRFSS which is an ongoing health survey of adults ages 18 years and older. The survey is given in two versions, A and B, that include a core set of national questions developed by the CDC and state added questions developed by CDPHE split between versions A and B. Both versions are available in English or Spanish and are delivered as appropriate to the participant. The two versions allow an appropriate amount of data to be gathered on the topics while maximizing survey sampling. A total of 5,520 adults were asked the drinking water questions in 2011 on Version B of the BRFSS. All households with telephones had an equal chance of being selected to participate in the survey, with individual survey participants being randomly selected from each successfully contacted household. The participants were asked; “what is the main source of your home water supply” and “which of the following best describes the water that you drink at home most often?” If a participant answered “a private well” in response to “what is the main source of your home water supply” they were asked a series of additional questions relating to well water testing that will be discussed later in this report. New in 2011 participants that indicated they drink filtered or bottled water at home were asked if the aesthetics, or perceived poor water quality were a factor in their choice to not drink unfiltered tap water. Survey results were analyzed to determine if statistically significant differences existed between responses from different regions of the state or between demographic groups. Weighted population (estimate), percent (response rate), and ninety‐five percent confidence intervals (CIs) were calculated. Statistical significance was determined by examining overlap of the CIs of each response rate. Overlap of regional or demographic CIs with the statewide CI indicated no significant difference between that behavior for a demographic group or region and the entire state. No overlap of a CI for a demographic group or region with the statewide CI indicated a statistical difference from the statewide rate. When the CI for a group or region was below the statewide CI that group or region is considered statistically less likely to behave a certain way. When the CI for a demographic group or region was above the statewide CI that group or region is considered statistically more likely to exhibit a certain behavior. The same comparison can be done within demographic groups as well, comparing for example the response levels between Hispanics and Whites, or those in a higher versus lower income bracket. The weighted population estimates, percent of the population and CIs were calculated from the raw BRFSS survey results data using the statistical software package SAS (v 9.3). Data are grouped by two categories; demographic and regional. The demographic characteristics are those used by the CDC for the BRFSS. The counties are aggregated into “health statistics regions,” (Regions) that were developed by the Health Statistics Section of CDPHE. Aggregating populations across the state into these regions allows for consistent calculations of descriptive statistics in large areas of the state with low population density, for example rural counties.
2
Results What is the Main Source of Your Home Water Supply? The first drinking water question in the survey was; What is the main source of your home water supply? Participants selected from a set of predetermined answers including: a city, county or town water system, a small water system operated by a homeowners association, a private well, other source or don’t know. Survey responses indicated that ~ 85% of Coloradans received water from some type of public water system including small systems operated by a homeowners association, for example. Both City, county, or town water systems and Small water systems are types of public water systems regulated by WQCD. Only about one percent of participants reported that they received their water from the other category while four percent reported they did not know where their drinking water came from in 2011. An estimated nine percent of Coloradans reported that they drank water from an unregulated private well on the 2011 BRFSS (Figure 1). These figures remain largely unchanged from the 2009 results. While it is assumed that these ground water sources are generally safe CDPHE Laboratory Services Division recommends that homeowners test the water in their well on a regular basis as water quality may change and is subject to geology and surrounding land use, both on the surface and below ground. Figure 1: Source of Colorado Adults’ Household Water Supply Reported in the 2011 Colorado BRFSS. The survey results were evaluated by demographics based on preliminary information gathered at the beginning of the survey. That is, survey participants were asked to self identify race, age, sex, education, income and employment status as part of the survey’s information gathering. General response rates for all demographic groups and responses are illustrated in Figure 2.
3
Figure 2: Demographic Characteristics of Water Supply Sources Reported in the 2011 BRFSS.
Looking in greater detail at the demographic distribution the following observations were made about the source of the participants’ drinking water: Participants that self‐identified as Black or Hispanic were statistically less likely to have a private well as their primary source of home water supply, and Hispanics are more likely to get there water from a public water system. White participants were more likely to drink from a private well than other racial or ethnic groups (Figure 3).
4
Figure 3: Statistical Differences Between Racial/Ethnic Groups and Home Water Supply.
It is worth noting that additional trends may be observed in other demographic categories. These numbers however were small enough to be considered insignificant, or did not reflect a statistically significant disparate behavior based on race, age, sex, income or education. For the purpose of this report they have not been displayed. Statistically significant disparities in employment and water source that were seen in the 2009 BRFSS results were not observed in the 2011 data. Numbers for Black‐Non‐Hispanic participants became too small, or insufficient, to evaluate for some of the variables and so appear to be missing. Information was also gathered on where participants resided in Colorado. Regional differences were observed in the survey results for home water supplies and included: Participants in Regions 1, 5, 8, 9, 11, 13 and 17 were statistically more likely to get their drinking water from a private well. These counties comprise approximately 130,000 adults, or about 39% of the Colorado adults that get drinking water from a private well. Participants in Regions 7, 14, 15 and 19 were statistically less likely to get their household water from a well (Figure 4). Participants in 7, 15, 19 and 20 were more likely in 2011 to get there water from a public water system. These counties comprise approximately 920,000 adults or about 31% of the Colorado adults getting their water from a public water system (Figure 5).
5
Figure 4: Counties (Health Statistics Regions) with Statistically Different Levels of Private Well Use for the Main Source of Home Drinking Water Compared to the State Average. Region 1: Logan, Morgan, Phillips, Sedgwick, Washington, Yuma Region 2: Larimer Region 3: Douglas Region 4: El Paso Region 5: Cheyenne, Elbert, Kit Carson, Lincoln Region 6: Baca, Bent, Crowley, Huerfano, Kiowa, Las Animas, Otero, Prowers Region 7: Pueblo Region 8: Alamosa, Conejos, Costilla, Mineral, Rio Grande, Saguache Region 9: Archuleta, Dolores, La Plata, Montezuma, and San Juan Region 10: Delta, Gunnison, Hinsdale, Montrose, Ouray, San Miguel Region 11: Jackson, Moffat, Rio Blanco, Routt Region 12: Eagle, Garfield, Grand, Pitkin, Summit Region 13: Chaffee, Custer, Fremont, Lake Region 14: Adams Region 15: Arapahoe Region 16: Boulder, Broomfield Region 17: Clear Creek, Gilpin, Park, Teller Region 18: Weld Region 19: Mesa Region 20: Denver Region 21: Jefferson
Figure 5: Counties (Health Statistics Regions) with Statistically Different Levels of Public Water System Use for the Main Source of Home Drinking Water Compared to the State Average.
6
Participants were then asked “which of the following best describes the water that you drink at home most often?” Responses to the question indicated that 82% of participants drink tap water (down from 84% in 2009), either filtered or unfiltered, at home most often. Tap water, whether filtered or unfiltered, may be from a public water system or a private well. Sixteen percent of participants drink bottled/vended water most often when they are at home (up from 15% in 2009), avoiding tap water. 41% of participants use a water filter for their tap water. This figure remains constant wheather sourced from a public water system or a private well. The remaining two percent drink water from some other or an unknown source (Figure 6). When the results for drinking bottled water, and using a water filter were evaluated against the source of participants’ household drinking water there was no significant difference observed in drinking bottled water or using a water filter based on the water source. Figure 6: Total Reported Responses to Which of the Following Best Describes the Water that You Drink at Home Most Often?
General demographic trends for home water consumption are shown in Figure 7. When the demographic data is evaluated further however, a number of disparate behaviors are observed: Participants that did not finish high school were more likely (32.6%) to drink bottled water at home than all groups with greater academic achievment . This group was also less likely (18.8%) to use a water filter at home (Figure 8). Participants in the lowest income group, less than $15,000 annually, were more likely to drink bottled/vended water (27.3%) than those in the highest income group (12.1%), more than $50,000 annually. Statistically significant differences were observed between low and high income earners as well. Lower income earners are less likely to use a water filter, and are more likely to drink unfiltered tap water than high income earners (Figure 9). Higher rates of bottled/vended water consumption were observed in Black and Hispanic participants (30.0% and 30.6%, respectively) than in Whites (11.9%) (Figure 10).
7
Figure 7: Demographic responses to: Which of the following best describes the water that you drink at home most often?
8
Figure 8: Statistical Differences in Water Consumption Choices at Home Between Different Education Levels.
9
Figure 9: Statistical Differences in Water Consumption Choices at Home Between Different Income Levels.
10
Figure 10: Statistically Significant Differences in Overall Bottled and Vended Water Consumption Between Racial/Ethnic Groups.
The BRFSS is delivered to participants in one of two language versions, Spanish or English, based on the preference of the participant at the start of the survey. The results of these separate language surveys allow us to evaluate the results in greater detail. Seperating the results of the English and Spanish version of the survey and performing the same statistical evaluation identifies behaviors of those that choose to have the survey delivered in Spanish. Evaluating the results we make the assumption that these results represent the behaviors of the population segment that are primary Spanish speakers. When we see what appear to be disparate behaviors among Hispanics in the total set of survey results we can evaluate the Spanish version results independently to see if there is truly a disparate behavior in the population represented by primary Spanish speakers. After identifying higher rates of bottled/vended water consumption among Hispanic participants in the summary responses the survey results for the individual language versions were analyzed. Two observations were clear in the results.
Spanish version participants were significantly more likely to drink bottled or vended water (55.7%) than either filtered (23.1%) or unfiltered (19.6%) tap water. Spanish version participants (55.7%) were also significantly more likely to drink bottled or vended water than English version participants (14.1%) (Figure 11).
11
Figure 11: Results from the Spanish and English surveys for the question: Which of the Following Best Describes the Water that you Drink at Home Most Often?
Following the source and type of water you drink at home questions participants were asked why they drink bottled or vended, or filtered water. The responses to the question show that the look, smell, or taste are the statistically more likely reason for drinking bottled or filtered water (Figures 12 and 14). The results for English speaking participants mirrored this observation by showing they were statistically more likely to drink bottled water and filter water for this reason. Spanish version responses showed that this group was more likely to drink bottled water because they are concerned tap water is unsafe than English speaking participants (Figures 13 and 15).
12
Figure 12: Statistical Relationships in Results from Reason for Drinking Bottled Water Between Racial/Ethnic Groups.
13
Figure 13: Statistical Relationships in Results from Reason for Drinking Bottled Water Between Spanish and English Results.
14
Figure 14: Statistical Relationships in Results from Reason for Drinking Filtered Water Between Racial/Ethnic Groups.
15
Figure 15: Statistical Relationships in Results from Reason for Drinking Filtered Water Between Spanish and English Participants.
In 2009 when these questions were last asked in Colorado a number of regions in the state were observed to have significantly higher or lower levels of bottled or vended water consumption at home. In 2011 those differences were not observed except for one region. Health Statistics Region 18, Weld County, had a slightly less significant level of bottled or vended water consumption at home. At this time it is not understood what factors may influence this or why the bottled or vended water consumption levels reported in the 2009 BRFSS may have changed. When Was the Last Time Your Well Water Was Tested by a Laboratory? Each of the participants that stated the source of their household water was a private well was asked when the last time was that they had the well tested. Following this they were then asked if they tested for specific chemicals and if the test revealed presence of the chemical in the well water. Again the responses were evaluated against preliminary location and demographic data collected at the onset of the survey. The following summary observations were made:
16
An estimated 325,958 Coloradans get their household drinking water from a well. That figure is up from the 2009 estimate of 314,114 (Table 1). 15% have had their well tested within the last year. That figure is down from the 2009 estimate of 21%. Approximately 41% of Coloradans either do not know when the last time their well was tested or their well has never been tested (Figure 16). That figure is up from the 2009 figure of 32 %.
Table 1: Population Estimates for Source of Household Water Supply in Colorado from 2011 BRFSS.
From the results indicating that a well test was conducted for a specific contaminant the survey shows that:
6.7 % of wells tested positive for bacteria. Up from 5% in 2009 7.0 % tested positive for nitrates. Down from 8% in 2009. 3.5 % tested positive for volatile organic compounds (VOCs). Up from 2% in 2009 1 % tested positive for pesticides (Table 2). Unchanged from 1% in 2009 11.5 % tested positive for radionuclides (uranium and/or radium). Radionuclides were not asked about in the 2009 BRFSS.
Figure 16: Percent Responses to: When Was the Last Time Your Well Water Was Tested by a Laboratory?
17
Table 2: Responses to well water testing question and results of reported tests:
Conclusion Eighty five percent (85%) of Coloradans received safe drinking water from public water systems, including small water systems, regulated by CDPHE‐WQCD in 2011. Sixteen percent (16%) of Coloradans, however, drink bottled or vended water at home primarily, and 41% filter their tap water at home before drinking it. Further, when we evaluate demographics we see that there are some disparate behaviors between a number of groups including income and education, and race and ethnicity. These behaviors can be further observed when we separate the English and Spanish responses and evaluate the responses and confidence intervals in those sub‐groups. When we look at the Spanish language results we see increases in responses, and widening between confidence intervals, for bottled water consumption. This signifies much more bottled and filtered water consumption in this group of participants. Reasons for this are unknown at this time but are possibly due to culture or language, lack of public information, awareness, or even product marketing. Education and income groups also exhibited some disparate behaviors in bottled and filtered water consumption. Participants that did not graduate high school were statistically more likely to drink bottled water than other education groups. Participants earning less than $15,000 annually were more likely to drink bottled water than those making over $50,000 annually.
18
Approximately nine percent (9%) of Coloradans get their drinking water from a private well. Typically private drinking water wells are not monitored by any state environmental or public health agency. The water quality in a private well can vary widely and it is the responsibility of the homeowner to regularly test the water for contaminants including bacteria, nitrates, radionuclides, volatile organic compounds (VOCs) and pesticides. While general wellwater surveillance is useful it may be more helpful to evaluate well water quality and trends at a finer resolution across the state. Regional analysis of well results would help to further identify areas of the state that are more likely to be impacted by certain types of well contaminants, and where well owners should be aware of specific hazards to their wells. Understanding the surrounding land use and geology is helpful in determining what chemicals should be sampled for in a particular well. Approximately 41% of Coloradans that use a private well for drinking water have either never had it tested or cannot remember when the last time was their well was tested. This figure was only about 31% when this survey was conducted in 2009. This may indicate that well users and owners are testing wells less as a result of public non‐awareness of the importance of well water quality and testing. An additional 23% have not had it tested in the last 5 years. When results of those with private wells that drink bottled or filtered water are evaluated no significant relationship is observed. That is, the results do not show that those with a private well are more likely to drink bottled water or filter the tap water in their house. When the survey results are evaluated for well owners that have not tested in over five years, or have never tested their well with consumption patterns for bottled or filtered tap water the numbers become too small and provide unstable results. A greater understanding of these trends in drinking water consumption can inform program policy decisions and prioritization of resources. Of particular interest from the results of the BRFSS is the consumption of bottled or vended water, or filtering household tap water. Future efforts should continue to improve the understanding of the notable differences in rates of bottled or vended water consumption in different demographics, particularly race and ethnicity, education and income, and in the general population. The information collected by the BRFSS could be evaluated across other known drinking water data. For example, public water quality data in regions with high rates of filtered and bottled water consumption, or fluoridation and health benefit in regions with unfluoridated water supplies. In a continued effort to understand this information the questions asked on the 2011 BRFSS were largely the same as the 2009 water questions. This allows comparison of results across multiple years. In 2011 the well water testing question was changed to specifically ask if a certain chemical was tested for. Prior to this the well water testing question only assumed that these chemicals were tested for. In 2011 the well water testing question included for the first time radionuclides. It has long been known that some areas in Colorado have challenges with radionuclides, like radium and uranium that are naturally occurring in groundwater. Interestingly, the survey results for radionuclide tests showed the largest number of positive tests of any chemical tested. Efforts to gather data for groundwater quality in private wells are also under way. Regional analysis of wellwater quality would be useful for identifying areas of the state with specific types of wellwater contaminants. This would help with environmental public health tracking and other local and county public health efforts focused on well water health. Information on private wellwater testing is available at the CDPHE Laboratory Services Division website, www.coloradostatelab.us.
19
