Evaluating the effectiveness of community-informed resource dissemination at increasing knowledge and testing rates among private well water owners in a statewide population

Abstract

Background

Threats to groundwater quality pose health risks to private well owners. Knowledge gaps are the main reason for low testing rates. Yet, few studies have examined the extent to which community-informed resource distribution increases knowledge and promotes private well testing.

Objectives

Evaluate the effectiveness of the resource dissemination in promoting increased knowledge and private well testing.

Methods

Community-informed resource that included 6 domains: regulation, testing recommendations, local resources, rationale for testing, state and federal resources, and mitigation actions. They were disseminated to n=1423 Survey of the Health of Wisconsin participants. Participants completed evaluations at the time of dissemination and one year later. Logistic regression models examined knowledge and behavior changes because of the resource sheet by demographics.

Results

About 50% of respondents reported increased knowledge on most domains of the resource sheet; 80% reported increased knowledge on at least one domain. About 13% tested their well in the last year because of the resource sheet. Because of the resource sheet, seeking out information in the last year was 2.47 (95% CI: 1.04–5.87) times higher among those ≥65 years of age after adjustment, and private well testing was 5.46 (95% CI: 2.15–13.9) times higher among households with ≥$100,000 annual income.

Discussion

Direct information distribution to private well owners can benefit many rural residents and promote well testing. Findings from this study highlight outreach disparities to private well owners. Future work should identify unique barriers and motivators to testing, and preferred communication media, among low-income and younger private well owners.

1. Introduction

Changes in agricultural practices and the effects of climate change are threats to groundwater quality.^1,2 These threats pose an increased health risk to rural residents relying on groundwater from private wells as their primary source of drinking water. About 15% of the United States (US) population, rely on groundwater from private wells for drinking and bathing.^3–5 Various chemical and microbial contaminants from naturally occurring and anthropogenic sources have been measured in private wells including arsenic, radium, manganese, bacteria, microbes, perfluoroalkyl substances, nitrates, hormones, pesticides, viruses, and pharmaceuticals.^6–12 Several adverse health effects are associated with exposures to these contaminants including endocrine disruption, cancer, liver and kidney problems, gastrointestinal illness, reproductive issues, and neurological disorders.^{6–8,13–18} Therefore, mitigating and managing pollution of private well water due to agricultural and other sources is key to advancing public health in rural areas. Unlike municipal water supplies, private wells serving fewer than 25 people are not regulated under the US Environmental Protection Agency’s Safe Drinking Water Act (Public Law 93–523) which ensures drinking water meets health-based standards.¹⁹ Rather, it is the responsibility of homeowners to test, treat, and manage their private wells to ensure their water is safe for consumption. Yet, several surveillance programs and observational research studies have discovered water testing behavior remains low among private well owners, as low as 4–10% in the prior 12 months.^{20–24, 28–30} State-level surveys and private well sampling found 40–58% of private wells exceed at least one Safe Drinking Water Act health-based standard,^25–27 demonstrating the significant at-risk population for this preventable public health problem in the US.

Regular testing of private well water is a necessary first step for identifying groundwater contaminants and protecting human health. While research on private well stewardship is limited, lack of relevant knowledge has been identified across studies as one of the main barriers to testing.^{20–24,31–34} Despite this understanding, few efforts have been reported which disseminate information about private well testing.³⁵ Even fewer evaluations have been conducted to measure program effectiveness at increasing knowledge and testing among private well owners.³⁵ Reported studies have relied on resource intensive methods (e.g., costly media campaigns, free testing kits and sample analysis, statewide testing requirements, and data reporting and notification programs).^36–40 These methods may not be feasible for under-resourced local and state public health departments, the primary institutions responsible for distributing public health information and resources in the US.^24,41,42 More research is warranted to better understand what strategies can realistically increase private well testing, and promote the health of more than 43 million, mostly rural, Americans served by private wells.^3–5

The current study was conducted in response to survey findings among private well owners in the Survey of the Health of Wisconsin (SHOW) cohort, where <10% of respondents reported testing their private well in the last year.²³ The majority of the respondents reported they did not have the information they needed to make informed decisions about their well, and requested local information.

The present study examined the effectiveness of using statewide cohort data on facilitators and barriers to testing private wells²³ and an advisory group consisting of local and state agencies. This study will inform the development of a sustainable strategy to improve private well stewardship in Wisconsin. However, methods and findings are applicable to other US states and countries where private well resources (e.g. personnel contacts, information, labs) are available at local or regional levels, not just at the state or country agency headquarters. In the US, many states include rural residents reliant on private wells and have county or local governments of health in addition to state health department resources. It follows a Dissemination and Implementation (D&I) framework, which posits that programs which respond to audience needs and concerns will be the most effective at increasing acceptance of the message, reducing barriers of uptake of an intervention, and promoting positive behavior change.^{24,31–33,41,42} The study aimed to (1) Develop community-informed resources, tailored to private well owners’ needs, (2) Disseminate the resources to private well owners, and (2) Evaluate the effectiveness of the resource sheet in promoting increased knowledge and private well testing.

2. Methods

2.1. Population

Participants were identified from SHOW, the only statewide representative survey modeled after the National Health and Nutrition Examination Survey (NHANES). Details on the SHOW program, the baseline survey, and sampling methods are described elsewhere.⁴⁶ In brief, an address-based area probability sampling without replacement (PPSWOR) was used to generate statewide representative samples among a total of n=5742 adults. SHOW cohort members who indicated a private well as their primary drinking water source were invited to participate, which equated to n=1423 adults (age 18+ years), from n=951 households. Supplementary Figure 1 depicts a map of the eligible sample.

2.2. Private Well Resource Sheet

Stakeholders from the Wisconsin Departments of Health Services (DHS) and Natural Resources (DNR), the University of Wisconsin-Madison Extension (UW-Extension), Wisconsin Public Health Research Network (WPHRN), and local health departments (LHD) made up an advisory board who collectively developed a one-page informational sheet with weblinks to additional information. The DNR list of labs included both commercial and public use, and its format was identified as a potential barrier for public uptake. Project staff contacted all 85 LHDs for information for local testing kits, labs, mitigation, county-specific contaminants of concern, and local contact information. Project staff tailored the state agency’s list of labs to the public and made it searchable by county.

We followed the PRACTical planning for Implementation and Scale-up (PRACTIS) guide to ensure best practices were followed when developing the private-well resource sheet and surveys (Table S1).⁴⁷ The advisory board determined the essential information on private well stewardship included 6 domains: regulation, testing recommendations, local resources, rationale for testing, other state and federal resources, mitigation actions available (Table S2). A county-specific private well resource sheet was developed for all 72 counties. Domains 1–2 & 4–6 (Table S2) were identical on all county resource sheets, and Domain 3 was tailored by county, with county-specific resources. See Supplemental materials for an example of the resource sheet. All LHDs were provided with a template so they could update their location information in the future for continued use.

2.3. Evaluation Survey

The evaluation surveys were developed in consultation with the University of Wisconsin Institute for Clinical and Translational Research (ICTR) program. The initial evaluation survey was designed to assess whether the resource sheet increased knowledge and awareness and promoted positive behavior change (8 questions). It also asked for residents’ preferred modality for similar resources in the future, what aspects of the resource sheet were most useful, and to report any unanswered questions (3 questions). Additionally, private well testing behaviors and demographic questions were captured.

The initial dissemination and evaluation effort took place from June to September 2022. Eligible adults were mailed a letter, a county-specific resource sheet for the county they resided in, an initial evaluation form, and a pre-paid return envelope. The letter included findings from the Private-well Stewardship Survey²³ and explained why they were being contacted. A website link was provided for online evaluation option via Qualtrics, a secure, web-based software platform for research data capture, hosted at the University of Wisconsin-Madison. Only one invited adult per household was asked to complete survey. The evaluations were returned anonymously with no way to identify the participant. All eligible adults were mailed a follow-up with identical materials two weeks after the initial mailing. A final reminder postcard was sent 5 weeks after the initial mailing.

A second evaluation survey was designed to assess whether the resource sheet promoted behavior change that resulted in seeking out additional information or testing their well. The second evaluation was conducted one-year later and asked whether participants remember receiving the initial resource sheet, if they had sought out additional information about testing or had tested their private-well in the last year, and if those actions were because of the resource sheet (5 questions). It included the same screener from the initial evaluation (2 questions), and demographic items (4 questions). See Supplementary Table 3 for evaluation items.

The one-year follow-up evaluation took place from June to October 2023. Eligible adults were mailed a letter, along with a county-specific private-well resource sheet for the county they resided in, the one-year evaluation form, and a pre-paid return envelope. Thirty participants from 16 distinct households became ineligible from the initial evaluation to the one-year evaluation due to being deceased, having no valid contact information, or requesting not to be contacted again.

The study was exempt from full review by the University of Wisconsin Health Sciences Institutional Review Board as it was determined to be program evaluation rather than research. All SHOW participants consented to be contacted for future evaluation under University of Wisconsin Health Sciences Institutional Review Board (IRB# 2013–0251 & 2007–0261).

2.4. Statistical Analysis

All analyses were conducted using SAS v.9.4 and ESRI ArcGIS version 10.8.2. Descriptive statistics and frequencies were used to report participants’ responses on both evaluations. To examine selection bias, cross tabulations and Chi-square test statistic were used to evaluate differences (p<0.05 considered statistically significant) in demographics among (1) those who were invited versus those who participated and (2) those who completed the initial evaluation versus those who completed the one-year evaluation. Summary scores were created based on the number of domains, out of six, of which respondents reported increased knowledge or awareness from the resource sheet. Multivariate logistic regression was used to determine what demographics and behaviors were more likely to result in higher odds of reporting (1) increased knowledge on 3–6 domains compared to 0–2 domains, and (2) the resource sheet impacted their decision to test and/or seek out information in the last year.

Full models were adjusted for gender, age (less than 65 years of age vs. greater than or equal to 65 years of age), annual household income (above or below $100,000), and educational attainment (above or below a bachelor’s degree). Models using data from the initial survey were additionally adjusted for whether their private well had been tested in the past year (also examined “testing in past 10 years”), whether they had lived at their home for 10 years or more, and whether they had sought out information on private well water testing prior to this project. Models examining results from the initial evaluation also considered knowledge and perceptions as a predictor. Those who reported not knowing how to test, what to test for, or that it was their responsibility to test as reasons for not testing their private well in the last year were considered to have a prior lack of knowledge compared to those who did not select these. Those who reported “not being concerned about my/our well water” and/or “have been drinking the well water for years without any problems” as reasons they did not test well in last year were considered to have low-risk perceptions compared to those who did not select these. To reduce selection bias from missing data, parsimonious models excluded variables which were not statistically significant at <0.05 level and which had >10% missing data.

3. Results

3.1. Participants

Among the n=951 invited households, 37% (n=366) responded to the initial invite; 52 were ineligible, resulting in a 33% participation rate (n=314). The one-year evaluation had a lower response rate (29%) and a participation rate of 26% (n=241) (Table S4). Compared to those who were invited, those who participated were statistically significantly more likely to be 65 years of age or older, male, and have higher educational attainment (Table S5). For both surveys, over 60% of participants were 65 years of age or older and just under 50% had a 4-year college degree or higher. Just under 30% of participants were from a household with more than $100,000 annual income (Table 1). Participant demographics did not vary significantly between evaluations.

Table 1.

Demographics of participants who completed the initial and one-year evaluation.

	Initial Evaluation (n=314)	One-year Evaluation (n=241)
Demographics	N (%)	N (%)	p-value
Current age, years			0.18
<45	32 (10.2)	14 (5.8)
45 to 64	94 (29.9)	65 (27)
65 to 74	96 (30.6)	86 (35.7)
75+	92 (29)	76 (31.5)
Gender			0.57
Male	149 (47.4)	112 (46.5)
Female	152 (48)	126 (52.3)
Missing (‘other’ n=1)	13 (4.1)	3 (1.2)
Educational attainment			0.12
High school/GED or less	71 (22.6)	51 (21.2)
Some college/2-year degree	93 (29.6)	77 (32)
4-year degree or more	150 (47.8)	109 (45.2)
Missing		4 (1.7)
Household Income			0.85
<$50,000	121 (38.5)	95 (39.4)
$50,000-$99,999	110 (35)	79 (32.8)
≥ $100,000	83 (26.4)	67 (27.8)
Sought out private well informationa			0.01
Yes	96 (32.5)	54 (22.4)
No	199 (67.5)	184 (75.9)
Missing	19 (3.9)	3 (1.7)
Length of current residence
<10 years	59 (18.8)	-
10+ years	253 (80.6)	-
Missing	2 (0.6)	-
Tested private well in the last 1 year			0.05
Yes	50 (15.9)	54 (22.4)
No	260 (82.8)	183 (75.9)
Missing	4 (1.3)	4 (1.7)
Tested private well in last 10 years
Yes	190 (60.5)	-
No	118 (37.6)	-
Missing	6 (1.9)	-

Prior to initial evaluation (for the initial evaluation), since July 2022 or when participant received the initial evaluation (for -year evaluation); p-value: chi-square test for significance.

3.2. Initial Evaluation

Sixteen percent (n=50) tested their well in the last year and 60% (n=190) tested their well in the last 10 years (Table 1). The top two reasons participants did not test their wells in the last year were because they were not concerned and/or because they have been drinking the well water for years without any problems; 16.8% of those who had not tested in the last year reported they did not know how to test their well (Supplementary Figure 2a). Nearly 60% of the those who tested their private well in the last 10 years did so because they wanted to know if their well water was safe to drink (Supplementary Figure 2b).

Roughly half of participants (48–54%) reported increased awareness and knowledge on four of the six domains on the private well resource sheet: (1) recommendation to test annually, (2) how to find resources on well testing, (3) how to test their private well, and (4) how to address well problems (Figure 1a). Sixty-two percent were already aware of the importance of testing, and 82% already knew testing was their responsibility. Over 80% reported their knowledge increased on at least one informational domain because of the resource sheet (figure 1b). Knowledge increased on 3 or more domains for 50% of participants, whereas 19% of participants reported no knowledge increase. Recommended tests, reasons to test, and local resources were the most useful information on the resource sheet (Figure 1c). A mailed paper resource sheet was the preferred modality for receiving future information and resources about private well water testing (81%), with email being the second most preferred (28%) (Figure 1d).

Fig. 1.

(a) Self-reported effectiveness of the resource sheet increasing knowledge or awareness of six key domains of private well water stewardship (n = 314), (b) the percent of participants (n = 314) that reported knowledge increase on at least 1, at least 3, and none of the six domains, (c) the domain or topic respondents (n = 279) reported as being most useful, and (d) preferred modalities (select all that apply) among respondents (n = 293) for receiving future information and resources about private well water testing.

Participants who reported knowledge increase on 3–6 domains were more likely to have not previously sought out well testing information and more likely to have not tested their well in the last year (or last 10 years), compared to participants who reported knowledge increase on 0–2 domains from the sheet (Tables S6 & S7). These associations held for comparisons between those with knowledge increase on at least one domain versus those with no knowledge increase. Increased knowledge from the resource sheet did not differ by gender, age, income, or education.

In adjusted models, those who reported increased knowledge because of the resource sheet on three or more of the domains were 3.22 (95%CI: 1.52–6.85) times more likely to not have tested their private well in the last year and had 1.80 (95%CI: 1.04–3.10) times the odds of not seeking out testing information prior to receiving the resource sheet, compared to those who reported increased knowledge on <3 domains (Table 3). Those who reported any knowledge increase because of the sheet (on one or more domains vs. no domains) had nearly 4 times the odds of not having tested their private well in the last five years (Table 2). Similar findings arose in the model examining well testing in the last 10 years, but with smaller effect sizes. In multivariate models, lack of prior testing, seeking of information, testing knowledge, and a perception of low risk of contamination and health problems from drinking water, were primary predictors of increased knowledge. To increase sample size used in models, gender and length of residency were not retained in final models. Table S8 & S9 depict full and parsimonious models.

Table 3.

Adjusted Odds Ratios (OR) with 95% Confidence Intervals (CI) of participants reporting one year later that they sought out testing information in the last 12 months because of the resource sheet (n=41) compared to those who reported the resource sheet did not impact their decision (n=187); and they tested their private well in the last 12 months because of the resource sheet (n=30) compared to those who reported the resource sheet did not impact their decision (n=196). Adjusted odds ratios are reported by demographic comparisons.

	Adjusted ORs of seeking out information in last 12 months (n=228)	Adjusted ORs of testing private well in last 12 months (n=226)
	OR (95%CI)	OR (95%CI)
Female vs. Male	1.03 (0.51,2.08)	1.07 (0.47,2.41)
≥ 65 years vs. <65 years	2.47 (1.04,5.87)	3.58 (1.27,10.1)
< Bachelor’s degree vs. Bachelor’s or Higher	1.71 (0.81,3.61)	0.80 (0.33,1.92)
Household Income ≥ $100,000 vs. < $100,000	2.13 (0.95,4.77)	5.46 (2.15,13.9)

Table 2.

Adjusted Odds Ratios (OR) with 95% Confidence Intervals (CI) of participants reporting (a) knowledge increase on 3–6 domains (n=152) vs. knowledge increase on 0–2 domains (n=162) of private well information because of the resource sheet, and knowledge increase on 1 or more domains (n=228) vs. knowledge increase on none of the domains (n=86) of private well information because of the resource sheet. Adjusted odds ratios are reported by demographic and behavior comparisons for four different models of behaviors with the same demographics.

	Adjusted ORs of knowledge increase on 3–6 domains vs. 0–2	Adjusted ORs of knowledge increase on 1+ domains vs. 0
Model with testing in the last year:	OR (95%CI)	OR (95%CI)
<65 years vs. ≥65 years	1.01 (0.62,1.67)	1.16 (0.66,2.05)
Bachelor’s or Higher vs. < Bachelor’s	1.12 (0.66,1.99)	1.52 (0.83,2.81)
Household Income <$100,000 vs. ≥ $100,000	1.31 (0.73,2.34)	1.55 (0.79,3.02)
Have not tested well in the last year vs. have tested	3.22 (1.52,6.85)	3.95 (1.93,8.09)
Did not seek information prior to this vs. did seek info	1.80 (1.04,3.10)	0.78 (0.41,1.48)
Model with testing in the last 10 years:
<65 years vs. >=65 years	1.00 (0.61,1.64)	1.13 (0.64,2.00)
Bachelor’s or Higher vs. < Bachelor’s	1.19 (0.70,2.00)	1.64 (0.90,3.02)
Household Income <$100,000 vs. ≥ $100,000	1.27 (0.71,2.27)	1.47 (0.76,2.85)
Have not tested well in last 10 years vs. have tested	1.88 (1.14,3.11)	2.26 (1.23,4.17)
Did not seek information prior to this vs. did seek info	1.88 (1.10,3.24)	0.93 (0.50,1.71)
Model with lack of knowledge as a predictor:
<65 years vs. ≥ 65 years	1.02 (0.64,1.64)	0.92 (0.55,1.56)
Bachelor’s or Higher vs. < Bachelor’s	1.22 (0.75,1.98)	1.78 (1.02,3.09)
Household Income <$100,000 vs. ≥ $100,000	1.23 (0.70,2.15)	1.12 (0.60,2.10)
Did not know what to test/how to test vs. did knowa	2.83 (1.57,5.10)	4.70 (1.93,11.41)
Model with low-risk perception as a predictor:
<65 years vs. ≥ 65 years	1.06 (0.67,1.69)	0.97 (0.57,1.63)
Bachelor’s or Higher vs. < Bachelor’s	1.16 (0.71,1.88)	1.69 (0.97,2.95)
Household Income <$100,000 vs. ≥ $100,000	1.19 (0.68,2.07)	1.09 (0.58,2.06)
Low risk perception vs. not low risk perceptionb	1.72 (1.09,2.71)	2.08 (1.25,3.47)

^aReported not knowing how to test, what to test for, or that it was there responsibility as reasons for not testing their well in the last year vs. did not select these as reasons for not testing their well in the last year or did test their well in the last year

^bReported “not being concerned about my/our well water” and/or “have been drinking the well water for years without any problems” as reasons they did not test well in last year vs. did not select these as reason for not testing either well or did test their well in the last year.

Thirty-three participants listed additional concerns or questions they had about private well water testing that were still unclear after receiving the resource sheet (Table S10). Costs of tests, free or reduced cost testing options, and mitigation options were the main topics about which participants still had questions (n=9). Several participants also reported having specific questions and concerns about nitrates, per- and polyfluoroalkylsubstances (PFAS), and other contaminants from nearby concentrated animal feeding operations (CAFOs) and other industrial and agricultural sources (n=8).

3.3. One-Year Evaluation

Figure 2 displays the self-reported effectiveness of the resource sheet impacting participants’ behavior one year after receiving the initial study materials. Among the n=229 who completed the one-year evaluation, 80% remembered receiving the initial sheet one-year prior, 23% reported seeking information on testing since then, and 30% said the resource sheet impacted their decision to seek out information; 18% (n=41) of respondents sought out information in the last 12 months because of the resource sheet. Additionally, since July 2022 when the initial resource sheet was sent, 32% said the resource sheet impacted their decision to test or not to test their private well. While 23% (n=54) reported testing their private well since July 2022, only 13% (n=30) indicated it was because of the resource sheet (Table S11).

Fig. 2.

Self-reported effectiveness of the resource sheet impacting participants’ behavior one year after receiving the initial resource sheet and evaluation (n = 229).

Those who tested their private well in the 12 months since receiving the resource sheet were more likely to have household income >$100k (46.3% vs. 22.4%) and more likely to report the resource sheet impacted their decision (56.6% vs. 25.3%) compared to those who did not test their private well in the last 12 months (Table S11). Similarly, those who sought out testing information in the last 12 months had higher household incomes (40.7% vs. 23.9%) and were more likely to report the resource sheet impacted their decision (75.9% vs. 16.6%) compared to those who did not seek out information in the last 12 months. Additionally, those who sought out information in the last 12 months were much more likely to have remembered receiving the resource sheet one year prior (92.6% vs. 76.1%) and indicate the resource sheet impacted their decision (75.9% vs. 16.6%) compared to those who did not seek out any additional information in the last 12 months (Table S11).

The adjusted odds of seeking out information on private well water testing in the last 12 months because of the resource sheet was 2.47 (95% CI: 1.04–5.87) higher among those 65 years and older compared with those younger than 65 years of age (Table 3). Statistically significant differences in seeking out testing information were not seen by sex, education, or household income in the adjusted model. The adjusted odds of testing one’s private well water in the last 12 months because of the resource sheet was 3.58 (95% CI: 1.27–10.1) times higher among those 65 years and older compared with those younger than 65 years, and 5.46 (95% CI: 2.15–13.9) times higher among those with household incomes greater than or equal to $100,000 compared with those from households with incomes less than $100,000 (Table 3). Statistically significant differences in testing because of the resource sheet were not seen by sex or education in adjusted models.

4. Discussion

This study demonstrates that the direct distribution of informational sheets to private well owners can promote well testing. Dissemination of information is an affordable and scalable outreach medium, and findings suggest it is important for increasing testing. Very few studies have conducted dissemination, implementation, or intervention on private well testing behaviors. Findings from the current study suggest direct distribution of private well information, if tailored to the community, can have huge impacts on decreasing the risk of exposure to contaminants and promoting the health of rural residents. Future efforts would benefit from learning and implementing outreach which yields effective engagement with younger and lower income private well owners.

4.1. Effectiveness of increasing knowledge

Most respondents reported the resource sheet increased their knowledge. Most importantly, it reached those who needed it most: those who had not tested their private well in the last year, 10 years, and those who have not previously sought testing information. Interestingly, the impact of the resource sheet on increasing knowledge was consistent across demographics of age, gender, income, and education. Similar findings were observed in studies conducted elsewhere.^{22–24,30,48,49} Findings suggest this medium of dissemination has the potential to benefit all private well owners’ awareness. It is important to note that those 65 years and older were 2.5 times more likely to seek out additional information compared to their younger counterparts. So, while the resource sheet effectively increased knowledge across demographics, it was more likely to impact behaviors among older adults. Private well resources are found on governmental websites or contacting governmental personnel via phone or email; methods of retrieving information that are more common among older age groups.^50,51 Future work in this area should consider promoting private well stewardship resources via social media and understanding preferred modes of information sharing among younger adults.

4.2. Effectiveness of increasing testing

About 13% of respondents tested their private well in the last 12 months and reported it was because of the resource sheet. This was an 8% increase from 2014 when only 5% of Wisconsin private well owners in SHOW (n=434) reported testing their well in the last 12 months.²³ In this study alone, testing rates in the last 12 months from the initial evaluation to the one-year evaluation also saw a similar increase of 6.5%. If the 6.5–8% increase in testing were scaled to the 800,000 private-well owners in Wisconsin, it would result in an additional 52,000–64,000 private well owners having tested their private well per the annual guidelines from the Wisconsin Department of Health Services, enabling more Wisconsin residents to identify, mitigate, and manage issues related to well water quality.

Few studies have evaluated the effectiveness of outreach interventions on testing behavior. However, when a study in New Jersey sent a free testing kit and risk communication about arsenic levels in neighboring wells to private well owners, it resulted in the same testing rate: 13% (n=224) of n=274 invited.³⁹ Those who received a high-risk letter were more likely to test their wells compared to those who received a medium-risk or low-risk letter (14.9% vs. 10.5%, a 4.4% difference in testing). The fact that outreach alone resulted in a similar testing rate suggests more resource intensive interventions such as subsidized testing kits may not produce better testing outcomes. However, a study in Quebec, Canada found a community-based arsenic-specific outreach intervention implemented after a mass media campaign resulted in an increase in arsenic testing from 4% to 16%. This finding suggests outreach focused on specific contaminants may result in greater testing and participation. While the interventions evaluated to-date have led to increased testing, none have resulted in ideal testing rates. A multi-prong approach of tailored resource dissemination, targeted campaigns around specific contaminants of concern, information via multiple mediums and annual reminders may be needed to reach higher testing rates.

4.3. Demographics and health equity

Over 60% of the participants were 65 years or older. This age group was also 2.5 times more likely to seek out information between the initial and one-year survey, and 3.5 times more likely to test their well because of the resource sheet, when compared with those <65 years of age. While not statistically significant, the young age group were more likely to report knowledge increase about private well stewardship, but it was not as likely to result in a behavior change the following year. Other studies have observed mixed findings with age as a predictor or modifier of testing.^22,32,36,39 This may be due to the difference in the underlying sample populations. In the United States, most private well owners reside in rural areas; 17.5% of rural populations are aged 65+ compared with 13.8% of urban populations.⁵² Even so, participants were significantly older than the eligible non-participants, suggesting selection bias was introduced. The current study may have suffered from cohort attrition or an inability to reach young-to-midlife adults about private well stewardship. Retirees may have more time and be more likely to have the financial security to pay for testing and mitigation. Also, older adults are more likely to reside in their home for longer, warranting concerns that new contaminants have arisen.

While only 30% of participants were in households of $100,000 annual income or higher, they were 5 times more likely to report testing their private well in the last 12 months because of the resource sheet. Cost may remain a substantial barrier to testing, a factor consistent with what other surveys have identified as main barrier to testing.^24,32,33,41 Yet, even when free and reduced cost well testing initiatives have been offered as interventions, higher income and higher educated adults were more likely to participate.^22,36 These collective findings indicate outreach programs tend to benefit wealthier populations and are not reaching less advantaged populations. More research is needed into the potential underlying health and social inequities that result in lower income residents being less responsive to outreach programs and less likely to test their wells after acquiring increased awareness of potential drinking water risks.

4.4. Strengths and limitations

This study responded to needs identified by private well owners and followed the Health Belief Model and D&I frameworks, effective methods of risk communication for pro-health behavior change identified by evidence in the literature.^1,2 The study leveraged community-academic partnerships and collaborated with stakeholders to develop county-specific resource sheets that convey state agency testing recommendations and offer local contacts, test kits, and lab options. It built capacity and sustainability by disseminating resources to state and LHDs for continued adaptation and use. The model does not rely on subsidized or free testing kits, large media campaigns, or resource-intensive delivery of data linkages with neighboring well results as prior studies have done.^{22,36,39,40,54} This study relies on a cost-effective and generalizable approach to providing reproducible educational resources for rural populations. The longitudinal nature of the study allowed adequate time for residents to test their private well. To our knowledge, this is the first outreach dissemination that evaluated knowledge and behavior change at two timepoints, spread one year apart. We were also able to leverage a statewide representative cohort providing a general population-based response to testing, different than past studies which have focused interventions on high-risk communities.

This study has several limitations. The response rate for the initial evaluation in this study was 38%, nearly half the response rate seen in the 2014 survey among private well owners from the same cohort. This could be due to typical attrition seen in longitudinal cohorts, or additional adversities faced by participants in the years following COVID-19. As mentioned earlier, this resulted in respondents being older, male, and having higher education compared to non-responders. Another limitation of the study was the inability to effectively engage younger participants. While knowledge change did not differ by age-groups, testing was not prevalent among younger ages in response to the resource sheet. Because this project was set up as evaluation and not research, it limited our ability to look at change over time among the same individuals who participated in both evaluations, and to capture other demographics, behaviors, and beliefs about testing that would have been informative. Finally, due to the lack of racial diversity in the rural participants of the SHOW cohort, we were not able to examine disparities in communities of color that may be serviced by a private well. SHOW includes racially diverse participants, but the majority reside in urban areas, serviced by municipal facilities. There is a need to evaluate well user stewardship, including testing, among rural communities of color, including tribal communities.

5. Conclusions

This study provided evidence of the potential pro-health behavioral impact outreach and resource dissemination can have across a state population in promoting private well stewardship. It highlights the importance of tailoring communication to the community and learning what their needs and interests are before designing outreach communication. Findings also highlight the disparities that exist in outreach to private well owners. Future work should focus on identifying the unique barriers and motivators to testing, and the preferred communication mediums among low income and younger private well owners, and private well owners of color.

Data Sharing

Data is available upon request for researchers. Please reach out to the corresponding author or visit the University of Wisconsin website for details on accessing the Survey of the Health of Wisconsin (SHOW) data and data from ancillary studies from the cohort