Throughout the U.S. South, some African-American communities have been systematically excluded from municipalities through exclusionary zoning practices known as "underbounding" (Aiken, 1987; Lichter, Parisi, Grice, & Taquino, 2007). Today, municipalities control land use in these underbounded communities without providing services such as piped water, sewage disposal, and trash collection (Aiken, 1987; Lichter et al., 2007). Underbounded African-American neighborhoods frequently rely on private wells and septic systems, although municipal water and sewer lines encircle or bisect these communities to reach majority White neighborhoods (Heaney et al., 2013; Johnson, Parnell, Joyner, Christman, & Marsh, 2004; MacDonald Gibson, DeFelice, Sebastian, & Leker, 2014). African-American communities excluded from municipal services are disproportionately exposed to water contaminants and face increased health risks (Heaney et al., 2013; Stillo & MacDonald Gibson, 2017).
To minimize waterborne illness risk, households in underbounded areas should routinely test their water and take action when contaminants are detected (Centers for Disease Control and Prevention, 2014). Few private well owners, however, test their water as frequently as public health experts recommend (Schwartz et al., 1998). Although educational programs could promote well testing (Simpson, 2004), we are unaware of any research identifying what information and resources residents of underbounded communities need to improve stewardship of their water quality.
To identify homeowner perceptions, practices, and preferences related to private well and septic system maintenance and operation, we conducted semistructured interviews with residents in underbounded neighborhoods of Wake County, North Carolina. Our interviews followed the "mental models" framework, which involves assessing risk perceptions and behaviors and comparing them with expert recommendations to identify intervention needs (Bruine de Bruin & Bostrom, 2013; Morgan, Fischhoff, Bostrom, & Atman, 2002).
Here, we sought to inform outreach programs targeted at improving drinking water quality in communities without access to municipal water services in North Carolina and elsewhere. Specifically, our research objectives were to 1) assess current well and septic system monitoring and maintenance behaviors in underbounded communities, 2) identify factors influencing these behaviors to guide future risk communication development, and 3) assess community preferences for private wells versus community water systems.
Following approval by the University of North Carolina (UNC) Institutional Review Board, interviewees were recruited from 57 households participating in a previous UNC study of water quality in underbounded Wake County neighborhoods (Stillo & MacDonald Gibson, 2017). Recruitment letters were mailed to all 57 households offering a $50 gift card for participation. The first 20 respondents were enrolled. Two were excluded due to poor interview audio quality.
Interviews began with five open-ended questions about well water, septic systems, and city water (Table 1). Following the mental models approach, the script used neutral wording and avoided leading questions (Morgan et al., 2002). As the interviews progressed, more focused questions were asked. Specifically, we focused on water quality perceptions, water source preferences, well testing, well maintenance, septic maintenance, well and septic system characteristics, and experiences with city and well water. To conclude, participants were invited to discuss any topics not previously covered.
Each interview statement was coded to identify whether it addressed specific topics in expert models of private wells and septic systems. These models are represented as qualitative influence diagrams; they were created through a combination of literature review and expert consultations (see supplemental figures). Nodes in the expert diagrams represented critical factors influencing well and septic system management and performance. For example, private well diagram nodes included contamination sources (e.g., septic systems and groundwater contamination), well system components potentially affecting water quality (e.g., corrosion of plumbing), and specific contaminants that should be routinely monitored. Each node was assigned a code.
If an interview transcript statement referred to a node, it received the corresponding code. When most interviewees vaguely discussed a group of codes rather than mentioning each individually, multiple codes were merged into one new, more general code. For example, septic drain field parts received the same code because most interviewees did not discuss the drain field in detail. Another list of codes was added to represent topics commonly raised by interviewees but absent from expert models. For example, expert diagrams did not include cost, but all participants mentioned cost.
A team of coders was trained to apply codes to statements from three transcripts. Following training, the coding system was adjusted to improve accuracy. Subsequently, two coders independently coded each interview statement. Coders agreed on 55% of statements (Cohen's [kappa] = 0.54, excluding the three training transcripts). In cases of disagreement, a third coder decided between the first two codes. Finally, the number of interviewees mentioning each code was computed.
To assess homeowner practices, perceptions, and preferences related to private well and septic system maintenance in underbounded communities and inform future outreach efforts, we conducted semistructured interviews with 18 homeowners, following the mental models approach (Morgan et al., 2002). We sought to determine whether participants followed recommended monitoring and maintenance practices, to identify key beliefs and factors that might influence adherence to these recommendations, and to ask whether participants preferred private wells or would like to be connected to a nearby, regulated community water supply.
Characteristics of Study Participants
Table 2 compares our 18 participants with the 57 households in UNCs water quality study of underbounded communities and with Wake County. The proportion of African-American participants (55.6%) was slightly higher than for the 57 households in the UNC water quality study (45.6%) and much higher than in Wake County (19.4%). The proportion of participants over age 65 (10.5%) was lower than in UNC's water quality study (34.6%) but higher than in the county population (8.5%). Of interviewees choosing to report an education level (n = 10), 60% had a 4-year degree or higher, which was similar to the UNC water quality study (57.7%) and slightly higher than in the county's adult population (48.9%).
Testing and Maintenance Practices
One of 18 respondents tested their water annually as recommended by the Wake County Department of Health (Table 3, Figure 1, and see supplemental figures). Half of respondents reported testing less than every 5 years (n = 8) or never (n = 3). Additionally, eight respondents reported conducting well maintenance.
The North Carolina Division of Public Health recommends pumping septic systems every 3-5 years. Seven respondents, however, either were unable to recall their last septic system maintenance or reported last pumping more than 5 years ago. One respondent last pumped their system...