Measuring Mental Health Service Preferences Amongst Illinois Dairy Producers

ABSTRACT

Objectives

To mitigate mental health concerns of farmers, research is needed to investigate strategies that encourage help-seeking behavior in this population. This study attempts to identify those help-seeking strategies. Six mental health service options were examined.

Methods

A survey, implementing a best-worst scaling choice experiment, was disseminated to members of the Illinois Milk Producers Association. Two methods of analysis were conducted. The first, a count-based method, employs a simple count-based approach to measure the relative preferences for the six mental health service options in question. The second is more complex and employs a latent-class logit regression model to measure individual preferences.

Results

The mental-health service options, ranked in order from most preferred to least preferred were: 1) speak to family and friends, 2) keep it to myself, 3) utilize programs offered by agricultural organizations, 4) search online for self-help resources, 5) talk to a mental health professional, and 6) use “tele-health” support services.

Conclusion

This study examined an important gap in the literature concerning help-seeking preferences of dairy farmers. It is the first to utilize a choice experiment approach to assess help-seeking preferences among this understudied population. Results provide important empirical evidence to support distinct categories of farmers who may be weighing options regarding how to best address their mental health concerns.

KEYWORDS:

• Best-worst scaling
• choice experiment
• dairy
• help-seeking
• mental health

Introduction

Males with an occupation in the “farmers, ranchers, and other agricultural managers” sector have a suicide rate over one and a half times that of other occupations.¹ State and federal policymakers have recognized the importance of raising awareness and providing resources for mental health within the agricultural community.^2–8 Agricultural organizations have also recognized the need for increased access to mental health-related resources and programs to assist those working within agricultural industries.^9–12 As such, it is arguably an essential public service to investigate, raise awareness, and provide resources for improving mental health in rural communities. Research has identified numerous internal and external stressors that may contribute to high suicide rates. For example, farmers face internal stressors that include social isolation, labor-intensive tasks, extreme work hours, farm-related injuries or disabilities, fear of losing the farm, and passing the farm down to future generations.^13–22 External factors include weather, diseases, pests, changing government policies and programs, volatile commodity prices, rising input costs, finances, and trade disputes.^13–22 Another well-known external factor is the shortage of agricultural labor,²³ which is exacerbated for dairy producers whose commodity is extremely labor intensive and highly perishable. Nevertheless, to mitigate mental health concerns of farmers, research is needed to investigate strategies that encourage help-seeking behavior in this population. This study attempts to identify those help-seeking strategies most preferred by farmers.

Farmers with mental health concerns can experience both psychological and structural barriers to mental health resources. Stigma related to having a mental health concern and seeking help is often marked by labels suggesting a person is weak, a failure, or inadequate; this may be a salient help-seeking barrier for farmers who may find it important to view themselves as strong, successful, and competent.^24–27 Some farmers may overcome barriers related to stigma but still experience structural barriers. For example, they may be unable to afford mental health care or lack immediate access.^28,29 Of farmers surveyed in a recent national poll, 42% “strongly agree” the cost of treatment would be a barrier if they were to seek help for a mental health-related illness.²⁸ Mental health services are sparse in rural areas, typically requiring people to travel long distances, often to another city or town. Therefore, better access to such services could benefit those in the agricultural sector with mental health concerns.^{21,22,28–33} Mental health concerns that are not addressed can lead to or exacerbate mental health-related illnesses such as anxiety, depression, other stress related illnesses, and even suicidal behaviors.^21,34–43

The present study

This study focuses on Illinois dairy producers and fills a gap within the literature for mental health service preferences among dairy producers, specifically within the Midwestern United States. The dairy industry in Illinois is similar to that of other Midwestern states, where dairy farm bankruptcies have increased in recent years.⁴⁴ The economic downturn in the dairy industry has led to a decline in mental health levels amongst dairy farmers.^45–49 Given the current economic state of the dairy industry and the growing concern over mental health-related illnesses in the agricultural community, a greater understanding of the attitudes dairy producers have towards addressing mental health concerns is needed. Preferences for different mental health services were assessed by completing five objectives: 1) identifying mental health service options preferred by Illinois dairy producers; 2) quantifying trade-offs Illinois dairy producers are willing to make when choosing between mental health service options; 3) identifying factors that affect the tradeoff decisions when choosing between mental health service options; 4) identifying subgroups of Illinois dairy producers, if any, that may influence the trade-off decisions being made; and 5) examining anxiety and depression levels among Illinois dairy producers. Results could help agricultural organizations and policymakers create programs and policies beneficial to dairy producers who may be hesitant to seek help for mental health-related issues. If such policies and programs are created to better align with the preferences of those who may utilize them, it could lead to an increase in the use of mental health services and hopefully a decline in mental health-related illnesses and incidents, such as suicide.

Methods

To address this study’s five objectives, a survey was developed and disseminated to members of the Illinois Milk Producers Association (IMPA) in December 2020 to their 475* members using the Dillman Tailored Design Method.⁵⁰ The Dillman Tailored Design Method is a tested survey design method that has been shown to improve survey-based research and increase response rates.⁵¹ The survey was composed of four sections. The first section asked respondents questions regarding their farm’s dairy production. The second section examined distress via the Kessler Psychological Distress Scale (K6+), a measure that assesses mental distress based on the frequency of experiencing anxious and depressive symptoms.⁵² Previous epidemiological research, validation with clinical populations, and evidence for internal consistency (Cronbach’s $\alpha$ value of 0.85) support the validity of the K6+ scale.^52,53 The third section presented a Best-Worst Scaling (BWS) choice experiment to respondents to assess their preferences for six mental health service options.⁵⁴ This approach allows for the measurement of an individual’s strength of preference across multiple mental health service options.^55,56 The mental health service options examined were 1) keep it to myself, 2) search for online self-help resources, 3) speak to family and friends, 4) talk to a mental health professional, 5) use “tele-health” support services, and 6) utilize programs offered by agricultural organizations. The fourth section asked personal and farm demographic questions. This study was approved by the Illinois State University Institutional Review Board, protocol #2020–367. Informed, written consent was provided by all survey participants.

Choice experiment

The survey was constructed on the basis of a choice experiment. Choice experiments are a stated preference procedure where individuals are asked to hypothetically choose between various goods or services.⁵⁷ Choice experiments are widely applied in various research fields such as marketing, health and transportation economics, and psychology.^58–60 More recently, choice experiments have been used in various agricultural-related studies.^61–65

Lancaster’s (1966) consumer theory and McFadden’s (1974) random utility theory contribute to the conceptual framework of a choice experiment. Consumer theory proposes that individuals will typically derive their utility from a good or service based off the characteristics that comprise it, rather than the good or service itself.⁶⁶ It is assumed that when selecting between goods or services, individuals will select the good or service that provides them the highest level of utility. Yet, when attempting to model such choices, researchers are unable to observe all the factors that influence such a choice. To address these unobservable factors, random utility theory recommends researchers instead model the probability of an individual’s choice. Thus, the choices an individual makes become a function of the probability the utility associated with a particular good or service is higher than alternative options.⁶⁷

The BWS Case 1 approach was utilized in this study.⁶⁸ With this approach, respondents are shown a set of objects and asked to select which is “best” (i.e., most preferred) and “worst” (i.e., least preferred). The “objects” in this case are the six mental health service options previously identified. To reduce the complexity when choosing the best and worst of the six mental health service options, a balanced incomplete block design was implemented. Rather than showing producers all six options at once, they were shown three at a time. This is a convenient way to get a complete ranking from each choice. Under this setup there were 20 possible choice combinations. To reduce respondent fatigue and improve response rates, only 10 choice combinations were presented to respondents. Every service type appeared five times total, twice with each alternative service option to keep the design balanced (Table 1).

Table 1. Best-worst scaling balanced incomplete block design.

Choice Set	Mental Health Service Option
1	A	D	E
2	E	C	A
3	A	B	C
4	F	C	D
5	E	B	F
6	B	C	D
7	D	A	F
8	C	F	E
9	B	E	D
10	F	A	B

Treatment options: A) keep it to myself, B) search online for self-help resources, C) speak to family and friends, D) talk to a mental health professional, E) use “tele-health” support services, and F) utilize programs offered by agricultural organizations.

The first service option, “keep it to myself,” was selected, because research has portrayed farmers as often eschewing mental health care – being self-reliant and aware of the heightened stigma of having mental health concerns that are common in agricultural communities.^{26–28,69–71} “Talk to a mental health professional,” as well as “speak to family and friends,” were chosen because of the preference for these service options found amongst farmers in previous studies and represent formal and informal help-seeking, respectively.^28,69,70 “Utilize programs offered by agricultural organizations” was selected to gain insight into farmers’ perception of the utility of these services that are aimed specifically to address mental health-related concerns. There are a limited, yet growing number of programs offered by agricultural organizations that target mental health.^9–12 With a limited pool of resources available to fund such programs, it is imperative for these organizations to know if their stakeholders are interested in using these services if provided.

The increase in online access to mental health resources led to the inclusion of “search online for self-help resources”. Previous research shows low preference for seeking self-help online as a mental health service option.⁷² The final service option, “use tele-health support services,” was chosen because of the recent expansion of it across the state of Illinois,⁷³ and because of increasing utilization during the COVID-19 pandemic.⁷⁴

Statistical analysis

Two methods of analysis were conducted to examine the BWS choice experiment. The first is a count-based method and employs a simple, count-based approach to measure the relative preferences for the six mental health service options in question. The second is more complex and employs a latent-class logit regression model to measure individual preferences.

Count-based method

The count-based analysis is conducted by counting the number of times respondents chose a mental health service option as either their most preferred (i.e., best) or least preferred (i.e., worst) option while completing the BWS choice experiment. From this, two estimates can be calculated. The first, a count-based estimate (Mathematical Expression 1), allows for the identification of a complete ranking between the six service options. For individual j, their least preferred service option count for option i are subtracted from their most preferred service option count for option i:

(1) $\sum \mathrm{M}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{e}{\mathrm{d}}_{ij}-\sum \mathrm{L}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{t}\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{e}{\mathrm{d}}_{ij}$(1)

leading to individual difference scores ranging from -5 to + 5. The higher (lower) the individual score, the more (least) preferred the service option is. In addition to the individual count-based estimate, a sample level count was also calculated. To examine how consistent service option preferences were amongst respondents, the standard deviation of the count-based estimate for each option was calculated. A low (high) standard deviation indicates more (less) consistency.

The second estimate is a ratio estimate (Mathematical Expression 2) and allows for the calculation of an individual’s preference for one service option relative to another. The ratio estimate divides the most preferred service option count for option i for individual j by the least preferred service option count for option i for individual j, followed by the square root of that value.

(2) $\sqrt{\frac{\sum \mathrm{M}\mathrm{o}\mathrm{s}\mathrm{t}\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{e}{\mathrm{d}}_{ij}}{\sum \mathrm{L}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{t}\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{r}\mathrm{e}{\mathrm{d}}_{ij}}}.$(2)

The ratio estimate can then be standardized on a 1–100-point scale. To determine relative preferences, one compares the standardized ratio of each service option to another. For example, a standardized ratio score of 80 for one service option would be four-times as preferred relative to a service option with a standardized ratio score of 20. A sample level ratio estimate was also calculated.

Latent-class model

A latent-class logit model was employed to identify potential service preference subgroups amongst the respondents.⁷⁵ This model type allows for preference heterogeneity to be examined and places respondents into specific classes based on their shared preferences and demographics.^76,77 A common issue faced by researchers when performing a latent-class logit model is selecting the appropriate number of classes to place their data set into. Although there is no general agreement on the best method to use when selecting the ideal class number, three standards were implemented. The first two were two statistical criteria: the Bayesian information criteria (BIC) and consistent Akaike information criteria (CAIC). For both criteria, the lower the score the better the model fit. The third standard involved a minimum class size of 25 respondents. Following the above criteria, a three-class solution was chosen as the optimal model of choice.

Results

Of the 475 IMPA members surveyed, 170 producers responded, which resulted in a 35.8% response rate. Respondent and farm characteristics are shown in Table 2. Respondents were primarily male (79.4%) with a mean age of 44.5 (Table 2). The average number of years a respondents’ farm has been in business was 47.5 years. Nearly half (48.2%) of respondents reported their farm ownership structure being a co-ownership with family members, and 47.1% reported having a succession plan in place for their farm.

Table 2. Respondent and farm characteristics, N = 170.

Characteristic	n (%)^a
Gender
Male	135 (79.4%)
Female	20 (11.8%)
Age
<40 years	16 (9.4%)
40–59 years	56 (32.9%)
60–79 years	63 (37.1%)
80+ years	5 (2.9%)
Education
High school, G.E.D., or less	66 (38.8%)
Some college/university	30 (17.6%)
Bachelor’s degree or higher	57 (33.5%)
Head of milk cows
<19 milk cows	2 (1.2%)
20–49 milk cows	29 (17.1%)
50–99 milk cows	59 (34.7%)
100–199 milk cows	40 (23.5%)
200–499 milk cows	21 (12.4%)
500+ milk cows	10 (5.9%)
Farmland owned
<199 acres	51 (30.0%)
200–499 acres	61 (35.9%)
500–999 acres	26 (15.3%)
1,000+ acres	15 (8.9%)
Farmland rented
<199 acres	71 (41.8%)
200–499 acres	36 (21.2%)
500–999 acres	19 (11.2%)
1,000+ acres	13 (7.7%)
Farm ownership structure
Sole proprietorship	67 (39.4%)
Co-ownership with family members	82 (48.2%)
Co-ownership with non-family members	2 (1.2%)
Co-ownership with family and non-family members	4 (2.4%)
Other	1 (0.6%)
Farm succession plan
Yes	80 (47.1%)
No	70 (41.2%)
Gross farm income
<$100,000	8 (4.7%)
$100,000–$249,999	29 (17.1%)
$250,000–$499,999	43 (25.3%)
$500,000–$999,999	29 (17.1%)
$1,000,000+	43 (25.3%)
Off-farm income
Yes	77 (45.3%)
No	77 (45.3%)
Fair milk price received
Yes	30 (17.6%)
No	133 (78.2%)
K6+ Distress Level
None	14 (8.2%)
Low	65 (38.2%)
Moderate	70 (41.2%)
Severe	11 (6.5%)

^aPercentages will not equal 100% due to non-responses.

K6+ scale

Results from the K6+ scale are shown in Table 2. The most common distress level reported was “moderate” distress, with 70 (41.2%) respondents being classified in this range. Sixty-five (38.2%) respondents were found to have “low” distress levels. Only 14 (8.2%) respondents showed no signs of distress, while 11 (6.5%) were found to have “severe” levels of distress.

Best-worst count

Of the 170 producers that responded, 107 fully answered all 10 BWS choice experiment scenarios. The service options, ranked in order from most preferred to least preferred were: 1) speak to family and friends, 2) keep it to myself, 3) utilize programs offered by agricultural organizations, 4) search online for self-help resources, 5) talk to a mental health professional, and 6) use “tele-health” support services (Table 3). This ranking is based on the best minus worst score, to which “speak to family and friends” (323) and “keep it to myself” (99) had the only positive scores. “Use ‘tele-health’ support services” was overwhelmingly the least favored service option among respondents with a best minus worst score of -235.

Table 3. Sample-wide best-worst scaling results, N = 107.

Service Option	Best	Worst	B-W	Mean	Standard Deviation	Rank	Ratio Estimate	Standard Ratio
Speak to family and friends	372	49	323	2.99	2.32	1	2.76	100
Keep it to myself	239	140	99	0.92	3.42	2	1.31	47.4
Utilize programs offered by agricultural organizations	136	165	-29	-0.27	2.77	3	0.91	32.9
Search online for self-help resources	126	199	-73	-0.68	2.47	4	0.80	28.9
Talk to a mental health professional	143	228	-85	-0.79	3.37	5	0.79	28.7
Use “tele-health” support services	54	289	-235	-2.18	1.96	6	0.43	15.7

B-W corresponds to the count-based approach for determining rank-ordered preferences. Mean and standard deviation corresponds to the respective mean and standard deviation of the B-W score.

To examine relative preferences amongst the six service options, the standard ratios of each service option are compared. For example, “speak to family and friends” has a standard ratio of 100, while “use ‘tele-health’ support services” has a standard ratio of 15.7. Thus “speak to family and friends” is over six times as preferred (100.0$÷$15.7 = 6.37) to “use ‘tele-health’ support services” when addressing a mental health concern. Interestingly, respondents had relatively equal preferences for “talk to a mental health professional” and “search online for self-help resources” (28.9$÷$28.7 = 1.01).

Examining the standard ratios does not imply that all respondents held identical preferences. The standard deviation of the best minus worst difference estimate scores in Table 3 can provide more insight. “Use ‘tele-health’ support services” had the lowest standard deviation (1.96), implying preferences for this service option were the most consistent across respondents. Conversely, “keep it to myself” (3.42) and “talk to a mental health professional” (3.37) had the highest standard deviations, implying that preferences for these two service options were the least consistent. The heterogeneity of preferences shown amongst the standard deviations suggests the benefit of examining potential subgroups of preferences in the population.

Latent-class model

Results from the latent-class analysis are shown in Table 4. Three distinct classes were identified and assigned names based on their respective treatment preferences: Community Help-Seekers, Prospective Help-Seekers, and Private-Healers. “Use ‘tele-health’ support services” was chosen as the reference, or base, service option for this analysis. A positive (negative) coefficient indicates the specific service type was viewed more (less) favorable than “use ‘tele-health’ support services.” For estimations of the individual-specific covariates to be identified, the Private-Healers class was assigned as the reference, or base, class. The sign of each individual-specific covariate tells the influence it has on the likelihood of a respondent falling into each class relative to the Private-Healers class. A positive (negative) coefficient shows that respondents with the individual-specific covariate are more (less) likely to belong to said class, relative to the Private-Healers class.

Table 4. Latent-class cluster analysis estimates and preference shares, N = 107.

	Community Help-Seekers			Prospective Help-Seekers			Private-Healers
Treatment	Coefficient	Rank	Preference	Coefficient	Rank	Preference	Coefficient	Rank	Preference
FF	2.52***	1	58.7%	2.34***	2	27.8%	2.32***	2	34.1%
MHP	1.55***	2	22.3%	1.33***	3	10.1%	-2.40***	6	0.3%
AG	0.88***	3	11.4%	0.63***	5	5.0%	1.12***	4	10.3%
SHR	-0.01	5	4.5%	0.90***	4	6.6%	1.39***	3	13.5%
KTM	-0.46***	6	2.8%	2.88***	1	47.7%	2.44***	1	38.5%
THS	—	4	4.5%	—	6	2.7%	—	5	3.4%
Covariates		Coefficient		Coefficient
College Degree		-0.88*			-1.74***			—
Acres Owned		0.01**			-0.01			—
Age (Years)		0.01*			-0.01			—
Co-Ownership with Family		0.69			1.42**			—
% Income from Dairy		0.02			-0.01			—
Off-Farm Income		0.82			0.29			—
K6+ Score		-0.02			0.02			—
Milk Head		-0.01			-0.01			—
Fair Price		-0.84			0.76			—
Constant		-0.82			-0.54			—
Class Share		36.8%			28.8%			34.4%

“Tele-health” Services is the base treatment option to which all other treatments are compared. Covariate coefficients for Assured Help-Seekers and Prospective Help-Seekers are relative to Self-Healers.

*p < .10. **p < .05. ***p < .01.

Abbreviations: FF = Family and Friends; MHP = Mental Health Professional; AG = Agricultural Organizations; SHR = Self-Help Resources;THS = tele-health Services; KTM = Keep to Myself; K6+ Score = Assessment of Mental Distress.

Community Help-Seekers comprised 36.8% of the respondents (Table 4). Respondents in this class overwhelmingly preferred to “speak to family and friends” (58.7%) over the other five service options when seeking help for a mental health-related issue. Community Help-Seekers least preferred the service option “keep it to myself” (2.8%). “Speak with family and friends” was nearly 21 (58.7$÷$2.8 = 20.96) times as preferred as “keep it to myself.” Those classified as Community Help-Seekers preferred to “search online for self-help resources” equally (4.5$÷$4.5 = 1.00) as much as to “use ‘tele-health’ support services.” Possessing a college or university degree had a negative effect on the likelihood of respondents belonging to the Community Help-Seekers class, relative to the Private-Healers class.

Prospective Help-Seekers compromised over a quarter (28.8%) of respondents (Table 4). As opposed to Community Help-Seekers, Prospective Help-Seekers ranked “keep it to myself” (47.7%) as their top preferred service option. “keep it to myself” was nearly twice (47.7$÷$27.8 = 1.72) as preferred as “speak to friends and family”. The least preferred service option amongst Prospective Help-Seekers was “use ‘tele-health’ support services” (2.7%). Having a farm structure, being a co-ownership with family members had a positive and significant effect on the likelihood of respondents belonging to the Prospective Help-Seekers class, relative to the Private-Healers class.

Private-Healers accounted for 34.4% of respondents (Table 4). Similar to the Prospective Help-Seekers, Private-Healers also ranked “keep it to myself” (38.5%) as their most preferred service option, with “speak to family and friends” (34.1%) being a close second. The preference share for the service option “talk to a mental health professional” was exceedingly low (0.3%) amongst Private Healers. Thus “keep it to myself” was over 100 times more preferred (38.5$÷$0.3 = 128) than “talk to a mental health professional.”

Summary statistics of respondents across the three distinct classes are shown in Table 5. Community Help-Seekers accounted for the largest class share (36.8%), owned the largest number of acres (416.7 acres) on average, and included the highest percent of respondents who felt they received a fair price for milk with 32.5% of that class answering, “Yes”. The Prospective Help-Seekers were younger on average than the other two classes (53.6 years), owned the least number of average acres (294.4 acres), and included just 12.5% of respondents who indicated receiving a fair price for milk. Additionally, this class experienced the highest K6+ score of each of the three classes. The final class, Private-Healers included the largest number of respondents holding college degrees (60%), while just over half (54.3%) of the respondents in this class co-own the farm with family, while the other two classes included 62.5% co-ownership. The percentage of income from the dairy operation and number of head milked were similar across all three classes.

Table 5. Latent-class cluster analysis class summary statistics, N = 107.

Variable	Sample Average	Community Help-Seekers	Prospective Help-Seekers	Private-Healers
College Degree (%)	45.8	37.5^a	40.6^a	6.0
Acres Owned (Mean)	415.5	461.7	294.4^a	307.2^a
Age (Years) (Mean)	57.0	59.8^a	53.6^a	58.8^a
Co-Ownership with Family (%)	59.8	62.5^a	62.5^a	54.3^a
% Income from Dairy (%)	68.0	68.5^a	67.7^a	67.8^a
Off-Farm Income (%)	5.5	47.5^a,b	43.8^a	60.0^b
K6+ Score (Mean)	5.4	5.1^a	6.1	5.2^a
Milk Head (Mean)	154.0	152.0^a	149.8^a	151.3^a
Fair Price (%)	22.4	32.5	12.5	2.0
Class Share (%)	1.0	36.8	28.8	34.4

Class means with similar superscripts (e.g., ^a,b,c) are not statistically different.

Regarding respondents’ service preferences for specific mental health concerns (Table 6), results were similar to BWS choice experiment results. A surprising result was the low preference for “talk to a mental health professional” and “utilize programs offered by agricultural organizations” when facing various mental health concerns. For mental health concerns relating to relationship problems, stress and anxiety, and health-related issues, respondents ranked “talk to a mental health professional” a distant third. Given these concerns relate to mental or physical problems, it is surprising that seeking help from a mental health professional was not more highly preferred. Respondents ranked “utilize programs offered by agricultural organizations” as third for mental health concerns relating to financial or farm-related issues. With the diverse number of agricultural organizations that exist to aid farmers with any farm-related concerns, it is surprising this service was not ranked higher for these issues.

Table 6. Help-seeking preferences for specific mental health concerns, N = 170^a.

Mental Health Concern	Keep It to Myself	Search for Online Self-Help Resources	Speak to Family and Friends	Talk to A Mental Health Professional	Use “tele-health” Support Services	Utilize ProgramsOffered by Agricultural Organizations
Financial Issue	54 (31.8%)	9 (5.3%)	80 (47.1%)	1 (.6%)	0 (0%)	17 (10%)
Relationship Problem	62 (36.5%)	3 (1.8%)	84 (49.4%)	5 (2.9%)	0 (0%)	2 (1.2%)
Stress and Anxiety	59 (34.7%)	5 (2.9%)	79 (45.6%)	10 (5.9%)	2 (1.2%)	3 (1.8%)
Farm-Related Issue	31 (18.2%)	8 (4.7%)	93 (54.7%)	2(1.2%)	0 (0%)	25 (14.7%)
Health-Related Issue	27 (15.9%)	8 (4.7%)	93 (54.7%)	14(8.2%)	8 (4.7%)	5 (2.9%)

^aStatistics presented: n (%); percentages will not equal 100% due to non-responses.

Discussion

This study examines an important gap in the literature concerning help-seeking preferences of dairy farmers. It is the first to utilize a BWS choice experiment approach to assess help-seeking preferences among this understudied population. Results highlight the benefits of examining help-seeking preferences using BWS procedures and provide important empirical evidence to support distinct categories of farmers who may be weighing options regarding how to best address their mental health concerns.

The BWS choice experiment results shown in Table 3 revealed respondents most preferred to “speak to family and friends” and to “keep it to myself” when seeking help for a mental health-related issue. This was followed by “utilize programs offered by agricultural organizations”, “search online for self-help resources”, and “talk to a mental health professional”. “Use ‘tele-health’ support services” was the least preferred service option. Overall, respondents appeared hesitant to utilize professional services aimed at directly addressing mental health (e.g., therapist, telehealth, etc.).

An examination of differences in the response patterns to the BWS choice experiment presented three different classes of help-seekers: Community Help-Seekers, Prospective Help-Seekers, and Private-Healers. A strong majority of respondents preferred informal sources of help, such as “keep it to myself” and “speak to family and friends.” These service options were ranked first and second respectively for the Prospective Help-Seekers and Private-Healers classes. The heterogeneity surrounding service option preferences suggests that some farmers have significant differences in their openness to engaging in more formal sources of help, such as talking to a mental health professional. It may be helpful for future research to examine factors (environmental, psychological, economic, etc.) that predict greater openness to professional help.

Prospective Help-Seekers prefer informal sources of help such as confiding in family and friends or keeping their issues to themselves. However, they do appear open to the idea of formal help options, having ranked “talk to a mental health professional” third among the six service options. Respondents classified as Prospective Help-Seekers were more likely to be younger adults, which may explain their receptiveness to formal help-seeking services. Studies have found younger adults were more likely to seek help compared to older adults.^78,79 Prospective Help-Seekers could benefit from a program similar to that offered by the Wisconsin Farm Center, which provides vouchers to farmers to meet with mental health professionals.⁷ Since Prospective Help-Seekers are more open to formal sources of help, giving them the opportunity to meet with professionals allows them to test if that service is suitable for them and continue if desired.

As opposed to Prospective Help-Seekers, Private-Healers had “talk to a mental health professional” ranked last. Based on their rank order, Private Healers prefer informal sources of help and are not amenable to seeking out formal help sources. Respondents classified as Private Healers were more likely to have some of their household income come from off-the-farm sources. A potential benefit of off-farm income sources could be the inclusion of health insurance. However, if that is not a benefit from an off-farm job, dairy producers may choose to not seek out formal sources of help due to the cost of such services.^14,28 This is because many farmers do not receive health care benefits if they are self-employed, unless a spouse works off-the-farm where health insurance is included.¹⁴ Private Healers might benefit from programs similar to voluntary stress management training program provided by the “Seeding Rural Resilience Act” that trains individuals to aid farmers in managing stress and suicide prevention.³ Training programs that teach family members the signs and symptoms of mental health-related illnesses, as well as the resources and services available would allow them to be able to help farmers if confided in.

As their name suggests, Community Help-Seekers are confident in utilizing formal mental health services. This class ranked “talk to a mental health professional” second, with “keep it to myself” as the least preferred option. Those classified as Community Help-Seekers would benefit from a program similar to the AgriSafe Network, which connects farmers to mental health professionals and educators during times of crisis, as well as offering a hotline [phone]) that operates 24/7.¹⁰ Since Community Help-Seekers had “utilize programs offered by agricultural organizations” ranked just below “talk to a mental health professional”, it would be fitting to have agricultural organizations that are able to connect farmers to mental health professionals if needed and having a hotline available when other services are not.

Limitations

This study focused solely on mental health service preferences among Illinois dairy producers. Future research should examine service preferences among dairy producers across the U.S. This would allow for more subgroups to be identified that might affect service preferences among dairy producers, as well as develop a list of preferred mental health service types among dairy producers across the country. This could also be done for various mental health concerns to see how service preferences may differ depending on the type of issues that producers are facing. Additionally, a strong limitation of the present study was the omission of religious leaders as a way of addressing mental health concerns. Several respondents in this study wrote comments regarding the importance of their religious beliefs, suggesting some dairy farmers may address mental health concerns via religious leaders; however, this was not assessed. Future research may benefit from including religious leaders as mental health service options for dairy producers. This could help gain insight into dairy farmers’ perceptions of the intersection of religion and mental health treatment.

While the K6+ scale presented in this study asked several anxiety and depression questions, the result was an overall self-reported distress level, rather than a mental health-related illness diagnosis. This study found that nearly 80% of respondents were experiencing either low (38.2%) or moderate (41.2%) distress. Producers are under stress, however whether that stress is contributing to a mental health-related illness is unknown. Evaluating various mental health-related illnesses, such as anxiety and depression, among U.S. dairy producers could lead to insight into the most common mental health-related illnesses that they may be facing. In knowing this, aid and resources could be provided to dairy producers that are struggling.

Conclusion

Results from this study add to previous research findings that show farmers widely prefer to keep their mental health-related problems to themselves. This study found that respondents’ top service preference for mental health-related illnesses was “speak to family and friends”, with “keep it to myself” being a close second. This information could be useful to government agencies and agricultural organizations that have set up programs to provide mental health services to farmers struggling with stress, anxiety, and depression.^80,81 Many of these programs are geared towards connecting farmers with mental health professionals or online resources. However, “talk to a mental health professional” and “search online for self-help resources” are not the most preferred service options among respondents. Instead, government agencies and agricultural organizations might enact programs to teach farmers’ family and friends to identify the signs and symptoms of various mental health-related illnesses. In addition, these programs could give them access to mental health resources and the ability to connect with professionals if needed, since it appears that farmers are less likely to seek help on their own.

Ethics approval

This study was approved by the Illinois State University Institutional Review Board, protocol #2020–367.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.

Notes

* According to the 2017 U.S. Census of Agriculture, 579 Illinois farms were classified under the North American Classification System as being involved in “Dairy Cattle and Milk Production”.