Attitude toward livestock farming does not influence the earlier observed association between proximity to goat farms and self-reported pneumonia

Abstract

Attitudes toward environmental risks may be a source of bias in environmental health studies because concerns about environmental hazards may influence self-reported outcomes.

Objective:
The main aim was to assess whether earlier observed associations between proximity to goat farms and self-reported pneumonia were biased by participants' attitude toward farming.

Methods:
We developed an attitude-score for 2,457 participants of the Dutch Livestock Farming and Neighbouring Residents' Health Study (veehouderij en gezondheid omwonenden) by factor analysis of 13 questionnaire items related to attitude toward livestock farming. Linear regression analysis was used to assess associations between attitude and potential determinants. The effect of attitude on the association between goat farm proximity and pneumonia was analyzed by evaluating (1) misclassification of the outcome, (2) effect modification by attitude, and (3) exclusion of participants reporting health problems due to farms in their environment.

Results:
In general, the study population had a positive attitude toward farming, especially if participants were more familiar with farming. Older participants, females, ex-smokers, and higher-educated individuals had a more negative attitude. Both self-reported respiratory symptoms and exposure to livestock farms were associated with a more negative attitude. Misclassification of self-reported pneumonia was nondifferential with regard to participants' attitude. Furthermore, no indication was found that the association between proximity to goat farms and pneumonia was modified by attitude. Excluding subjects who attributed their health symptoms to livestock farms did also not change the association.

Conclusions:
The association between goat farm proximity and pneumonia was not substantially biased by study participants' attitude toward livestock farming.

Full text

Downloaded from https://journals.lww.com/environepidem by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3z484KQalTVEsmeDH9sZBIaQkaYYv4hpuUm3LlrqiHiBvs9y+xOMuEw== on 02/26/2019
Downloadedfromhttps://journals.lww.com/environepidem by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3z484KQalTVEsmeDH9sZBIaQkaYYv4hpuUm3LlrqiHiBvs9y+xOMuEw== on 02/26/2019
Original Research
ISEE
E
NVIRONMENTAL
EPIDEMIOLOGY
1
Introduction
Environmental hazards—such as air or drinking water pollu-
tion—may be a source of concern in exposed individuals.1,2
In epidemiologic studies, information on health outcomes is
often self-reported, which has well-documented limitations
such as recall bias and social desirability bias.3 Study partici-
pants’ attitudes toward environmental risks may be a source of
information bias as well because concerns about environmental
hazards may inuence self-reported outcomes. Moffatt et al1
describes such “awareness bias” as the propensity to report
more illness and symptoms as a result of proximity to a po-
tential hazard, in the absence of a biologic effect. Perception of
exposure, causal beliefs and concerns, and media coverage play
an important role in symptom reporting.4–8
Actual or perceived exposure to a hazard, and cultural and
social factors may inuence someone’s risk perception, which
results in a variation of attitudes toward a potential environ-
mental risk among individuals.9 Marcon et al2 found that
determinants of environmental risk perception mainly com-
prise demographic, socioeconomic, and exposure indicators.
However, the authors did not investigate whether risk percep-
tion affected epidemiologic associations between environmental
pollution and self-reported health outcomes.2
There is an ongoing debate about intensive livestock farming
and potential health risks for surrounding populations.10–14 The
Netherlands is a small country with one of the highest popu-
lation densities in the world in combination with one of the
highest livestock densities.15 A small survey (n = 1,090) on the
public’s view on intensive livestock farming showed disagree-
ment among the Dutch general population about large-scale
intensive farming.16 Most arguments against intensive livestock
farming were focused on animal welfare, and potential risks for
public health.
The veehouderij en gezondheid omwonenden (VGO) study
(Dutch acronym for Livestock Farming and Neighbouring
Residents’ Health) investigated a wide range of health risks
(respiratory health, zoonotic infections, and antimicrobial
aInstitute for Risk Assessment Sciences, Utrecht University, Utrecht, The
Netherlands; bNetherlands Institute for Health Services Research, NIVEL, Utrecht,
The Netherlands; and cDepartment of General Practice & Elderly Care Medicine,
Amsterdam Public Health Research Institute, VU University Medical Center,
Amsterdam, The Netherlands.
Sponsorships or competing interests that may be relevant to content are
disclosed at the end of the article.
Supplemental digital content is available through direct URL citations in
the HTML and PDF versions of this article (www.epidem.com).
*Corresponding author. Address: Institute for Risk Assessment Sciences, Yalelaan
2, 3584 CM, Utrecht, The Netherlands. Tel.: +3130 253 8696. E-mail address:
l.a.smit@uu.nl (L. A. M. Smit).
Attitude toward livestock farming does not
influence the earlier observed association
between proximity to goat farms and self-reported
pneumonia
Floor Borléea,b, C. Joris Yzermansb, Floor S. M. Oostwegela, François Schellevisb,c, Dick Heederika,
Lidwien A. M. Smita*, VGO Consortium
Environmental Epidemiology (2019) 3:e041
Received: 14 August 2018; Accepted 13 January 2019
Published online 14 February 2019
DOI: 10.1097/EE9.0000000000000041
Background: Attitudes toward environmental risks may be a source of bias in environmental health studies because concerns
about environmental hazards may influence self-reported outcomes.
Objective: The main aim was to assess whether earlier observed associations between proximity to goat farms and self-reported
pneumonia were biased by participants’ attitude toward farming.
Methods: We developed an attitude-score for 2,457 participants of the Dutch Livestock Farming and Neighbouring Residents’
Health Study (veehouderij en gezondheid omwonenden) by factor analysis of 13 questionnaire items related to attitude toward live-
stock farming. Linear regression analysis was used to assess associations between attitude and potential determinants. The effect of
attitude on the association between goat farm proximity and pneumonia was analyzed by evaluating (1) misclassification of the out-
come, (2) effect modification by attitude, and (3) exclusion of participants reporting health problems due to farms in their environment.
Results: In general, the study population had a positive attitude toward farming, especially if participants were more familiar with
farming. Older participants, females, ex-smokers, and higher-educated individuals had a more negative attitude. Both self-reported
respiratory symptoms and exposure to livestock farms were associated with a more negative attitude. Misclassification of self-re-
ported pneumonia was nondifferential with regard to participants’ attitude. Furthermore, no indication was found that the association
between proximity to goat farms and pneumonia was modified by attitude. Excluding subjects who attributed their health symptoms
to livestock farms did also not change the association.
Conclusions: The association between goat farm proximity and pneumonia was not substantially biased by study participants’ at-
titude toward livestock farming.
Copyright © 2019 The Authors. Published by Wolters Kluwer Health, Inc. on
behalf of Environmental Epidemiology. All rights reserved. This is an open access
article distributed under the Creative Commons Attribution License 4.0 (CCBY),
which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly cited.

Borlée et al. • Environmental Epidemiology (2019) 3:e041 Environmental Epidemiology
2
resistance) among residents living in close proximity of livestock
farms in the Netherlands.17–24 One of the main ndings was a
higher risk of pneumonia for residents living in close proximity
to goat farms.22,24 Pneumonia was dened based on question-
naire data22 and/or a diagnosis of pneumonia by the general
practitioner (GP), recorded in the Electronic Medical Record
(EMR).22,24 As a direct policy implication, ve Dutch provinces
have stopped issuing building permits for goat farms, awaiting
further evidence. However, one can raise the criticism that po-
tential awareness bias—overreporting of pneumonia by exposed
individuals—may have resulted in a biased association.
In the present analysis, we constructed an “attitude toward
farming” score as a proxy for awareness of farming as an envi-
ronmental hazard. The main aim of the current study is to assess
whether the earlier observed association22 between proximity to
goat farms and pneumonia was biased by participants’ attitude.
Methods
Study design and population
The VGO study population originates from participants of a
cross-sectional questionnaire survey (n = 14,163) among ran-
domly selected GP patients (18–70 years old) living in small
towns or villages in a livestock-dense area in the south of the
Netherlands.18 Respondents who were willing to participate in a
follow-up study and who were not working or living on a farm
were eligible for a medical examination (n = 8,714). Based on
their home addresses, 12 temporary research centers were es-
tablished. Between March 2014 and February 2015, all respon-
dents living within 10 km of one of these temporary research
centers (n = 7,180) were invited to the nearest center for med-
ical examination and 2,494 participated (response, 34.7%). The
medical examination consisted among others of a second and
more extended questionnaire and spirometry.17,25 The study pro-
tocol (13/533) was approved by the Medical Ethical Committee
of the University Medical Centre Utrecht. All 2,494 subjects
signed informed consent. In total, data from 37 subjects were
excluded from the analyses because of missing data, resulting in
a study population of 2,457 subjects.
Medical examination
The questionnaire comprised among others items on education,
profession, residential history, smoking habits, and respiratory
health. Moreover, the questionnaire also contained 15 state-
ments on attitude toward farming in their residential environ-
ment (statements are shown in Table1). Statements were mostly
adopted from a survey among the general Dutch population
which was focused on the public’s view on intensive livestock
farming.16 To assess lung function, pre- and postbronchodila-
tor spirometry was conducted among 2,037 participants.25
We had two sources of information on pneumonia: (1) self-re-
ported, physician-diagnosed pneumonia over the past 3 years,
or (2) having had at least one pneumonia episode recorded in
the EMR during the 3 years preceding the medical examination.
Although our original nding was based on a combination of
both sources,22 or EMR data alone,24 in the current analysis,
we focused on the effect of attitude on associations with self-re-
ported pneumonia because the impact of attitude was expected
to be most pronounced for a self-reported outcome.
Construction of a score for attitude toward livestock
farming in the residential environment
Based on the 15 statements on attitude toward farming, we
developed an “attitude-score” using factor analysis. Response
options of the 15 statements were coded based on a ve-point
Likert scale (Table1). Principal factor analysis was used to iden-
tify one or more latent factors that can be interpreted as an
attitude toward farming. Standardized factor scores (z-scores,
hereafter named attitude-score) were computed as linear com-
binations of scoring coefcients and standardized questionnaire
Table 1
Statements regarding attitude toward farming in the residential environment and the distribution of 2,457 participants’ responses
to the 15 statements
Question
Reverse
scored?
Included in
final factor?
Factor
loading
More negative
attitude, %
Neutral
attitude, %
More positive
attitude, %
Missing,
%
S1. Livestock farms are a heavy burden for my living environment Yes Yes 0.70 15.0 28.4 56.2 0.3
S2. Farmers do their best to prevent heavy disturbances in my living
environment
No Yes 0.54 11.4 43.0 45.0 0.6
S3. Livestock farms are important for the Dutch economy No Yes 0.60 5.4 20.9 73.3 0.4
S4. I am happy with the livestock farmers in my neighbourhood No Yes 0.71 21.1 50.2 28.6 0.1
S5. There is too much discussion about the disadvantages of livestock
farming
No Yes 0.60 23.4 38.2 38.0 0.5
S6. The odor of manure disturbs me every time Yes Yes 0.58 30.1 26.3 43.4 0.2
S7. Livestock farming is a threat for my health Yes Yes 0.80 15.5 41.6 42.6 0.2
S8. I think the threat for my health due to livestock farming increased
in the last decade
Yes Yes 0.74 25.2 32.4 42.2 0.2
S9. If farmers monitor the health of their animals well, livestock
farming is not a threat for my own health
No Yes 0.49 18.2 35.1 46.6 0.2
S10. I am concerned about the impact of antibiotic use in livestock
farming for my own health
Yes No - 65.8 23.4 10.7 0.2
S11. I am concerned about new diseases that can be transmitted from
animals to humans
Yes No - 63.4 24.7 11.9 0.1
S12. I have health problems that are caused by livestock farms in my
living environment
Yes Yes 0.56 4.3 31.6 63.5 0.6
S13. A livestock farmer loves his animals and takes good care of them No Yes 0.57 7.2 31.9 60.6 0.3
S14. If there is no disturbance for me or my family, livestock farming
may increase
No Yes 0.65 33.0 29.9 36.8 0.2
S15. Construction of bigger stables disturbs the landscape Yes Yes 0.53 50.5 27.4 22.0 0.1
For comparability purposes, responses of negatively-keyed statements were reverse scored. Response options are coded based on a five-point Likert scale (“Strongly disagree,” “Disagree,” “Neutral,”
“Agree,” “Strongly agree”) but are represented in the table as a three-point scale. The answers Strongly disagree and Disagree were merged and translated as “More negative attitude,” the answers Agree
and Strongly agree were also merged and translated to a “More positive attitude.”

Borlée et al. • Environmental Epidemiology (2019) 3:e041 www.environmentalepidemiology.com
3
responses for each participant, where a higher score indicates a
more positive attitude toward farming.
Livestock farm exposure variables
Distances between home addresses and livestock farms were
computed using a geographic information system (ArcGis 10.1;
Esri, Redlands, CA) as described previously.18,25,26 The following
livestock farm exposure proxies were studied for each subject:
(1) number of farms within 500 and 1,000 m, and (2) presence
of a farm (pig, poultry, cattle, goat, sheep, horse) within 1,000
m (Yes/No).
Data analysis
First, we assessed the association between the attitude-score and
potential determinants using linear regression analysis. Results
were expressed as regression coefcients (β) and 95% CIs repre-
senting the mean change in the attitude-score given a change in
the determinant (one unit or otherwise stated in the Tables). The
potential determinants of attitude studied were as follows: (1)
personal characteristics, (2) respiratory health, and (3) exposure
to livestock farms. Two adjusted models were assessed: model
A, adjusted for age and gender, and model B, adjusted for age,
gender, born in study area, childhood on a farm, BMI ≥ 30, vis-
ited a farm last 12 months, and high education.
Second, to study the impact of attitude on information bias
(i.e., differential misclassication of self-reported pneumonia),
we compared self-reported and EMR-based pneumonia, and
computed sensitivity and specicity in a group with a more neg-
ative (< median attitude-score) and a more positive attitude (>
median attitude-score). To study effect modication by attitude,
the association between proximity to goat farms and pneu-
monia was also analyzed in the “more negative” and “more pos-
itive” group, and we tested interaction between farm proximity
and attitude-score.
Third, sensitivity analyses were conducted after excluding
subjects who attributed their symptoms to presence of livestock
farms in their environment. The association between pneumonia
and goat farm proximity (within 1,000 m as in Ref. 22) was
analyzed with logistic regression, and expressed as odds ratios
(ORs) and 95% CI. Data were analyzed using SAS 9.4 (SAS
Institute Inc., Cary, NC).
More details on the study methodology are provided in the
online supplement; http://links.lww.com/EE/A34.
Results
Study population
Participants were on average 56.4 ± 11.1 years old, and 54.6%
of the study population consisted of women (Table2). In total,
76.1% was born in the study area and one third (33.8%) had
grown up on a farm. The number of missing answers to the
15 statements was low for all items (<0.6%) (Table 1). The
majority of participants answered neutral or positive to all
statements, with the exception of three statements regarding
concerns about antibiotic usage in livestock farming, zoonotic
diseases, and disturbance of the landscape due to construction
of bigger sheds.
Construction of attitude-score
After rst exploratory factor analyses, statements 10 and 11
were removed because their residual correlation coefcients
were >0.1. The nal factor analysis was performed on the re-
maining 13 statements, and one latent factor was identied (ei-
genvalue = 5.14) and explained 97.6% of the total variance.
Cronbach’s alpha was 0.89, suggesting a good internal con-
sistency. Factor loadings (i.e., the correlations of the individual
questionnaire items with the factor) ranged from 0.49 to 0.80
(Table1). Including one of the initially removed statements (10
or 11) resulted in a very similar factor solution (correlation be-
tween factor scores based on 13 or 14 statements was 0.998).
Determinants of attitude
Older participants, women, ex-smokers (vs. never smokers), and
individuals with a higher education (vs. low and middle educa-
tion) had a more negative attitude toward farming (Table 2).
As expected, determinants related to familiarity with a farming
environment—such as childhood on a farm, born in the study
Table 2
Characteristics of the study population of 2,457 adults from a general, nonfarming population, and association between potential
determinants and the attitude-score
Personal characteristics
Mean
(SD) or %
Unadjusted
β (95% CI)
Model Aa Adjusted
β (95% CI)
Model Ba Adjusted
β (95% CI)
Age (per 10 years), mean (SD) 56.4 (11.1) −0.17 (−0.21, −0.14) −0.18 (−0.21, −0.14) −0.21 (−0.24, −0.17)
Female (%) 54.6 −0.02 (−0.10, 0.05) −0.08 (−0.16, −0.01) −0.09 (−0.17, −0.02)
Born in the study area (%) 75.6 0.29 (0.20, 0.38) 0.22 (0.13, 0.31) 0.23 (0.14, 0.31)
Childhood on a farm (%) 33.8 0.22 (0.15, 0.30) 0.30 (0.22, 0.38) 0.11 (0.02, 0.20)
Ex-smoker (%) 44.6 −0.18 (−0.26, −0.11) −0.08 (−0.16, −0.01) −0.09 (−0.17, −0.01)
Current smoker (%) 10.2 0.18 (0.05, 0.30) 0.13 (0.01, 0.26) 0.09 (−0.03, 0.21)
BMI ≥ 30b (%) 20.6 0.22 (0.13, 0.31) 0.25 (0.16, 0.34) 0.24 (0.15, 0.33)
Higher education (%) 30.2 −0.19 (−0.27, −0.11) −0.29 (−0.37, −0.21) −0.24 (−0.32, −0.16)
Paid work (%) 57.5 0.20 (0.12, 0.28) −0.05 (−0.15, 0.04) −0.08 (−0.17, 0.02)
Retired (%) 28.3 −0.25 (−0.33, −0.17) 0.02 (−0.09, 0.12) 0.08 (−0.02, 0.19)
Having pets, last 5 years (%) 52.4 0.17 (0.09, 0.25) 0.07 (−0.01, 0.15) 0.04 (−0.03, 0.12)
Having farm animals as a hobby,
last 5 years (%)
18.2 0.12 (0.02, 0.21) 0.09 (0.00, 0.19) 0.02 (−0.08, 0.11)
During current work/study contact
with animals (%)
6.1 0.18 (0.03, 0.34) 0.21 (0.05, 0.36) 0.10 (−0.05, 0.26)
Visited a farm last 12 months (%) 62.6 0.22 (0.14, 0.29) 0.16 (0.09, 0.24) 0.12 (0.05, 0.20)
Potential determinants of the “attitude-score” (z-score obtained from factor analysis) were analyzed with linear regression analysis. Regression coefficients display a change in the attitude-score for a
difference in determinants as indicated in the table (e.g., for 10 years increase in age, or for being female vs. male). A negative association means that the determinant is associated with a more negative
attitude toward farming and a positive association means that the determinant is associated with a more positive attitude toward farming.
aModel A was adjusted for age and gender, and model B was adjusted for age, gender, born in study area, childhood on a farm, BMI ≥ 30, visited a farm last 12 months, and high education.
bBMI = mass (kg)/height (m)2.
BMI indicates body mass index.

Borlée et al. • Environmental Epidemiology (2019) 3:e041 Environmental Epidemiology
4
area, or a recent farm visit—were associated with a more posi-
tive attitude toward farming.
All self-reported respiratory health outcomes were associated
with a lower attitude-score, whereas objectively measured res-
piratory health such as lung function and chronic obstructive
pulmonary disease (COPD; based on lung function) was not as-
sociated with attitude (Table3).
The following proxy measures of livestock farm exposure
were associated with a more negative attitude-score: a larger
number of farms within 500 and 1,000 m of the home, presence
of a pig farm (β, −0.13 [95% CI = −0.22, −0.04]), or a goat farm
(β, −0.19 [95% CI = −0.31, −0.08]) within 1,000 m (supplemen-
tary Table S1; http://links.lww.com/EE/A34).
As expected, subjects who attributed their health complaints
to livestock farming had a more negative attitude toward farm-
ing (Table 3). Excluding subjects who attributed their health
symptoms to livestock farms in their environment (n = 191,
7.8%) did not change associations between attitude and per-
sonal characteristics and associations with farm exposures (data
not shown). However, associations between the attitude-score
and self-reported respiratory health symptoms were attenuated
in the sensitivity analyses (data not shown).
Impact of attitude on the association between proximity to
goat farms and pneumonia
Sensitivity and specicity of self-reported pneumonia (com-
pared with an EMR-based diagnosis) did hardly differ between
those with a more negative attitude (sensitivity, 52%; specicity,
97%) and those with a more positive attitude (sensitivity, 56%;
specicity, 98%). Residents living within 1,000 m of a goat farm
had a higher risk of self-reported pneumonia (OR, 1.78 [95%
CI = 1.07, 2.95]) (Figure1), which differed slightly from the pre-
viously reported OR (2.0 [95% CI = 1.3, 3.1]) that was based
on both EMR and self-reported pneumonia,22 and from the OR
based on EMR only (2.3 [95% CI = 1.4, 3.9]).
No signicant interaction was observed between attitude
and living within 1,000 m of a goat farm (P value for interac-
tion 0.63), suggesting that the association between goat farms
and pneumonia was not modulated by attitude. In addition,
dividing the population in a group with a more negative and
a more positive attitude did not substantially change the asso-
ciation, but CIs were wider (< median attitude-score: OR, 1.65
[95% CI = 0.85, 3.23]; > median attitude-score: OR, 2.06 [95%
CI = 0.94, 4.52]).
Excluding subjects who attributed their health symptoms to
livestock farms in their environment did not change the associ-
ation between self-reported pneumonia and living within 1,000
m of a goat farm (OR, 1.75 [95% CI = 1.02, 3.01]).
Discussion
Our present study shows that the earlier observed associa-
tion22 between proximity to goat farms and pneumonia in the
Livestock Farming and Neighbouring Residents’ Health Study
was not substantially biased by participants’ attitude toward
farming.
Misclassication of self-reported pneumonia resulted in at-
tenuated risk estimates when compared with EMR-based di-
agnosis, but misclassication was nondifferential with regard
to participants’ attitude. Furthermore, the association between
goat farm proximity and pneumonia was similar in groups
with a more positive or more negative attitude (i.e., no effect
modication), and excluding participants who attributed their
health problems to livestock farming (7.8% of the population)
did not meaningfully change the association. The attitude-score
as dened in this article was used as a measure of information
quality (quality of self-reported physical health). Because atti-
tude is not a causal ancestor of physical health, it does not meet
the causal structure required of a confounder or a causal inter-
mediate. Adding the attitude-score as if it were a confounder
hardly changed the association between goat farm proximity
and self-reported pneumonia (OR, 1.73 [95% CI = 1.03, 2.93]).
We found several determinants that are associated with atti-
tude toward farming in residential environments. In general, the
study population had a relatively positive attitude toward farm-
ing. Most questions were answered with a neutral to positive
tendency. Familiarity with farming could possibly explain the
predominantly positive attitude. One third of the study popula-
tion had grown up on a farm. The study area, in which 75.6%
Table 3
Associations between the attitude-score and self-reported and objectively measured respiratory health outcomes
Health status %
Unadjusted
β (95% CI)
Model Aa Adjusted
β (95% CI)
Model Ba Adjusted
β (95% CI)
Self-reported respiratory health
Self-reported ever asthma 6.3 −0.16 (−0.31, 0.00) −0.23 (−0.38, −0.08) −0.20 (−0.36, −0.05)
Self-reported current asthma 4.9 −0.18 (−0.36, −0.01) −0.26 (−0.43, −0.09) −0.24 (−0.41, −0.07)
Self-reported COPD 5.1 −0.25 (−0.43, −0.08) −0.17 (−0.34, −0.00) −0.16 (−0.33, 0.02)
Self-reported pneumonia confirmed by GP
or specialist
5.3 −0.21 (−0.38, −0.04) −0.19 (−0.35, −0.02) −0.24 (−0.41, −0.08)
Attribution health complaints by livestock
farming
7.8 −1.25 (−1.38, −1.11) −1.20 (−1.33, −1.08) −1.19 (−1.32, −1.06)
Objectively measured respiratory health mean (SD) (lung function parameters expressed as IQR increase)b
COPD based on lung function (%)c,d 9.0 −0.09 (−0.22, 0.04) −0.12 (−0.26, 0.02) −0.10 (−0.24, 0.03)
Lung function parameters (mean [SD]), per IQRd
FEV1 % predicted 99.4 (15.0) −0.06 (−0.10, −0.01) −0.05 (−0.09, 0.00) −0.02 (−0.07, 0.03)
FVC % predicted 103.1 (12.8) −0.10 (−0.15, −0.05) −0.10 (−0.15, −0.05) −0.04 (−0.09, 0.01)
FEV1/FVC % predicted 95.8 (8.5) 0.03 (−0.01, 0.07) 0.04 (0.00, 0.09) 0.01 (−0.03, 0.05)
MMEF % predicted 94.0 (32.2) 0.00 (−0.05, 0.04) 0.01 (−0.04, 0.06) 0.01 (−0.04, 0.06)
Associations between the “attitude-score” (z-score obtained from factor analysis) and self-reported respiratory health and objective measured respiratory health were analyzed with linear regression
analysis. Regression coefficients display a change in the attitude-score for a difference in health determinants as indicated in the table. A negative association means that the determinant is associated with
a more negative attitude toward farming and a positive association means that the determinant is associated with a more positive attitude toward farming.
aModel A was adjusted for age and gender, and model B was adjusted for age, gender, born in study area, childhood on a farm, BMI ≥ 30 (BMI = mass [kg]/height [m]2), visited a farm last 12 months, and
high education.
bIn total, 2,059 subjects had lung function measurements of good quality (C or better).17
cCOPD based on lung function: a post-BD measurement of FEV1/FVC below the lower limit of normal AND/OR a post-BD measurement of FEV1/FVC < 0.70.17,25
dAdjusted models (A + B) with self-reported respiratory health, COPD, and lung function parameters were also adjusted for current smoking.
BD indicates bronchodilator; BMI, body mass index; IQR, interquartile range.
FEV1 indicates forced expiratory volume in 1 second; FVC, forced vital capacity; MMEF, maximal mid-expiratory flow.

Borlée et al. • Environmental Epidemiology (2019) 3:e041 www.environmentalepidemiology.com
5
of the study population was born, is characterized by the highest
farm density of the Netherlands. Previous studies on risk per-
ception show that common risks are judged more acceptable
than uncommon and unknown risks.27 Agricultural activities
are familiar and common among the majority of the study pop-
ulation and therefore probably more acceptable. Attitude was
indeed positively associated with determinants related to famil-
iarity with a farming environment such as childhood on a farm,
being born in the study area, or a recent farm visit.
We found that self-reported health symptoms were associated
with a more negative attitude. Subjects who reported to attribute
their health complaints to livestock farming had a much lower
average attitude-score than other participants. This is in line
with previous studies that showed positive associations between
concern and reporting factors related to illness.1,6 Awareness
bias1 might have played a role since we only observed an as-
sociation between attitude and self-reported respiratory health
and not with objectively measured respiratory health. Several
indicators of livestock farm exposure were associated with a
more negative attitude. Subjects who live in areas with a high
number of livestock farms, especially in close proximity of pig
and goat farms, had a more negative attitude toward farming
than subjects living in areas with less livestock farms. The asso-
ciation with goat farms might be explained by an unprecedented
outbreak of Q-fever, a zoonosis caused by Coxiella burnetii, that
occurred in the study area between 2007 and 2010.28 Dairy goat
farms with C. burnetii–induced abortions were implicated as
the major source of infection in the neighboring human popula-
tion. More than 3,500 acute Q-fever patients, mostly presenting
as pneumonia, were ofcially registered, and it was estimated
that 95 patients died. A study focused on regional differences in
public perceptions regarding Q-fever found that this epidemic
caused increased perceived anxiety and preventive behavior
among subjects living in regions with high Q-fever incidence.29
The observed association with pig farms could possibly be
explained by odor annoyance. Pig farms emit more offensive
odor in comparison with cattle and poultry farms.30 Odor
annoyance is common in populations living in the proximity of
livestock farms and is a main source of annoyance.31,32 A Dutch
study showed that the number of pigs, but also the number of
poultry and cattle, around homes of residents was associated
with odor annoyance.33
In 2011, a survey on the general public’s view of the Dutch
population on intensive livestock farming was conducted.16 This
survey consisted of two parts: a qualitative part that explored
arguments that play a role in the discussion on intensive live-
stock farming in the Netherlands, and a second part that con-
sisted of an online survey among 1,090 subjects from the Dutch
general population. The 15 statements in our questionnaire
were adopted from or inspired by this survey. Results of the
online survey showed a lot of similarities with the answers to
the statements given by our study population, even though our
study population is living in a rural area with high livestock
farm density. This might explain why our study population
considers the benets for the local (and Dutch) economy more
important than the general Dutch population from the previous
survey (73.3% vs. 52%). In the online survey, one of the most
important arguments against intensive livestock farming was
focused on potential risks for public health, and especially on
antibiotic resistant bacteria and zoonotic diseases.16 The ma-
jority of our study population mentioned to be concerned about
antibiotic usage in livestock farming and zoonotic diseases. The
use of antibiotics in livestock production can lead to increased
occurrence of antimicrobial resistance in bacteria which may
transmit to humans.34 Previous studies show increased risks of
livestock-related antimicrobial resistance among farmers with
direct animal contact.35,36 This may have contributed to concerns
about antimicrobial resistance in the study population, despite
the large reduction of antimicrobial use of more than 60% in
livestock farming since 2009 in the Netherlands.37 In the current
VGO study, no increased risk was observed between farm prox-
imity and carriage of extended-spectrum β-lactamase (ESBL)-
and plasmid-encoded AmpC-producing Enterobacteriaceae.19
However, a slightly increased risk was observed between living
near farms and carriage of livestock-associated methicillin-re-
sistant Staphylococcus aureus (LA-MRSA), although the preva-
lence was low (0.4%), and there is a high likelihood of a chance
nding.20 The Q-fever outbreak in the study area between 2007
and 2010 is likely to have contributed to our study population’s
concerns on emerging zoonotic infections.28,29,38
Strengths of our study are our large, population-based sample
and the low amount of missing data on the attitude statements.
Both self-reported and objectively assessed data on respiratory
health were available; this enabled us to compare associations
with attitude and to explore awareness bias. Nevertheless, a
number of limitations should be considered. First, the cross-sec-
tional design makes it difcult to infer causality. Second, atti-
tude toward farming may have contributed to the decision
whether or not to participate to the medical examination and
to the questionnaire survey where the study population origi-
nates from. Our previous studies showed that participants of
the medical examination17 and responders to the questionnaire
survey18 lived in closer proximity to farms compared with sub-
jects who did not participate and with nonresponders, respec-
tively. We have no information on attitude toward farming from
the source population; therefore, it was not possible to analyze
the effect of participation bias on the average reported attitude.
In conclusion, we developed an attitude-score to measure at-
titude toward farming in the residential environment. In general,
the study population had a positive attitude toward farming,
in particular if participants were more familiar with farming.
Older participants, females, ex-smokers, and individuals with
a higher education had a more negative attitude. Self-reported
symptoms were also associated with a more negative attitude.
However, we did not nd any indication that the previously re-
ported association between proximity to goat farms and self-re-
ported pneumonia was biased by attitude. Overall, results of
the current study indicate that attitude might play a role when
Figure. Effect of attitude on the previously observed association between self-reported pneumonia and living within 1,000 m of a goat farm (“main model”22).

Borlée et al. • Environmental Epidemiology (2019) 3:e041 Environmental Epidemiology
6
using self-reported data in environmental health studies. When
relying on self-reported data, we recommend to estimate atti-
tude toward a potential hazard to assess the potential inuence
of awareness bias on epidemiologic associations.
Conflicts of interest statement
The authors declare that they have no conicts of interest with
regard to the content of this report.
This work was supported by grant from the Lung Foundation
Netherlands (grant number: 3.2.11.022) and funded by the
Ministry of Health, Welfare and Sports and the Ministry of
Economic Affairs of the Netherlands.
Acknowledgments
The VGO study is conducted by a consortium (VGO consortium)
of different research institutes (in alphabetic order): Institute
for Risk Assessment Sciences of the Utrecht University (IRAS),
Netherlands Institute for Health Services Research (NIVEL),
National Institute for Public Health and the Environment
(RIVM), Wageningen Bioveterinary Research of Wageningen
University and Research (WUR), and Wageningen Livestock
Research (WUR).
References
1. Moffatt S, Mulloli TP, Bhopal R, Foy C, Phillimore P. An explora-
tion of awareness bias in two environmental epidemiology studies.
Epidemiology. 2000;11:199–208.
2. Marcon A, Nguyen G, Rava M, Braggion M, Grassi M, Zanolin ME. A
score for measuring health risk perception in environmental surveys. Sci
Total Environ. 2015;527–528:270–278.
3. Coughlin SS. Recall bias in epidemiologic studies. J Clin Epidemiol.
1990;43:87–91.
4. Baliatsas C, Van Kamp I, Bolte J, Schipper M, Yzermans J, Lebret E.
Non-specic physical symptoms and electromagnetic eld exposure in
the general population: can we get more specic? A systematic review.
Environ Int. 2012;41:15–28.
5. Porsius JT, Claassen L, Smid T, Woudenberg F, Petrie KJ, Timmermans
DR. Symptom reporting after the introduction of a new high-voltage
power line: a prospective eld study. Environ Res. 2015;138:112–117.
6. Shusterman D, Lipscomb J, Neutra R, Satin K. Symptom prevalence
and odor-worry interaction near hazardous waste sites. Environ Health
Perspect. 1991;94:25–30.
7. Baliatsas C, Bolte J, Yzermans J, et al. Actual and perceived exposure to
electromagnetic elds and non-specic physical symptoms: an epidemio-
logical study based on self-reported data and electronic medical records.
Int J Hyg Environ Health. 2015;218:331–344.
8. Witthöft M, Rubin GJ. Are media warnings about the adverse health
effects of modern life self-fullling? An experimental study on idio-
pathic environmental intolerance attributed to electromagnetic elds
(IEI-EMF). J Psychosom Res. 2013;74:206–212.
9. Jacobs J, Taylor M, Agho K, Stevens G, Barr M, Raphael B. Factors asso-
ciated with increased risk perception of pandemic inuenza in Australia.
Inuenza Res Treat. 2010;2010:947906.
10. van Cleef BA, Verkade EJ, Wulf MW, et al. Prevalence of livestock-asso-
ciated MRSA in communities with high pig-densities in The Netherlands.
PLoS One. 2010;5:e9385.
11. Smit LA, van der Sman-de Beer F, Opstal-van Winden AW, et al. Q fever
and pneumonia in an area with a high livestock density: a large popula-
tion-based study. PLoS One. 2012;7:e38843.
12. Poulsen MN, Pollak J, Sills DL, Casey JA, Nachman KE, Cosgrove SE,
et al. High-density poultry operations and community-acquired pneu-
monia in Pennsylvania. Environ Epidemiol. 2018;2:e013.
13. Huijbers PM, de Kraker M, Graat EA, et al. Prevalence of extend-
ed-spectrum β-lactamase-producing Enterobacteriaceae in humans liv-
ing in municipalities with high and low broiler density. Clin Microbiol
Infect. 2013;19:E256–E259.
14. O’Connor AM, Auvermann BW, Dzikamunhenga RS, et al. Updated
systematic review: associations between proximity to animal feeding
operations and health of individuals in nearby communities. Syst Rev.
2017;6:86.
15. Statistics Netherlands. Agriculture; crops, livestock and land use by
general farm type, region. Available at: http://statline.cbs.nl/Statweb.
Accessed 14 August 2018.
16. Verhue D, Vieira V, Koenen B, van Kalmthout R. Opvattingen over
megastallen [Views on intensive livestock farming]. Available at: http://
edepot.wur.nl/168111. Accessed 14 August 2018.
17. Borlée F, Yzermans CJ, Aalders B, et al. Air pollution from livestock
farms is associated with airway obstruction in neighboring residents.
Am J Respir Crit Care Med. 2017;196:1152–1161.
18. Borlée F, Yzermans CJ, van Dijk CE, Heederik D, Smit LA. Increased res-
piratory symptoms in COPD patients living in the vicinity of livestock
farms. Eur Respir J. 2015;46:1605–1614.
19. Wielders CCH, van Hoek AHAM, Hengeveld PD, et al. Extended-
spectrum β-lactamase- and pAmpC-producing Enterobacteriaceae
among the general population in a livestock-dense area. Clin Microbiol
Infect. 2017;23:120.e1–120.e8.
20. Zomer TP, Wielders CC, Veenman C, et al. MRSA in persons not living
or working on a farm in a livestock-dense area: prevalence and risk
factors. J Antimicrob Chemother. 2017;72:893–899.
21. van Gageldonk-Lafeber AB, van der Hoek W, Borlée F, et al. Hepatitis E
virus seroprevalence among the general population in a livestock-dense
area in the Netherlands: a cross-sectional population-based serological
survey. BMC Infect Dis. 2017;17:21.
22. Freidl GS, Spruijt IT, Borlée F, et al. Livestock-associated risk factors for
pneumonia in an area of intensive animal farming in the Netherlands.
PLoS One. 2017;12:e0174796.
23. van Dijk CE, Garcia-Aymerich J, Carsin AE, et al. Risk of exacerbations
in COPD and asthma patients living in the neighbourhood of livestock
farms: observational study using longitudinal data. Int J Hyg Environ
Health. 2016;219:278–287.
24. Kalkowska DA, Boender GJ, Smit LAM, et al. Associations between
pneumonia and residential distance to livestock farms over a ve-year
period in a large population-based study. PLoS One. 2018;13:e0200813.
25. Borlée F, Yzermans CJ, Krop E, et al. Spirometry, questionnaire and elec-
tronic medical record based COPD in a population survey: comparing
prevalence, level of agreement and associations with potential risk fac-
tors. PLoS One. 2017;12:e0171494.
26. Smit LA, Hooiveld M, van der Sman-de Beer F, et al. Air pollution from
livestock farms, and asthma, allergic rhinitis and COPD among neigh-
bouring residents. Occup Environ Med. 2014;71:134–140.
27. Slovic P. Perception of risk. Science. 1987;236:280–285.
28. van der Hoek W, Morroy G, Renders NH, et al. Epidemic Q fever in
humans in the Netherlands. Adv Exp Med Biol. 2012;984:329–364.
29. Bults M, Beaujean D, Wijkmans C, Richardus JH, Voeten H. Q fever
in the Netherlands: public perceptions and behavioral responses in
three different epidemiological regions: a follow-up study. BMC Public
Health. 2014;14:263.
30. Ni JQ, Robarge WP, Xiao C, Heber AJ. Volatile organic compounds at
swine facilities: a critical review. Chemosphere. 2012;89:769–788.
31. Wing S, Horton RA, Marshall SW, et al. Air pollution and odor in com-
munities near industrial swine operations. Environ Health Perspect.
2008;116:1362–1368.
32. Radon K, Schulze A, Ehrenstein V, van Strien RT, Praml G, Nowak
D. Environmental exposure to conned animal feeding operations
and respiratory health of neighboring residents. Epidemiology.
2007;18:300–308.
33. Hooiveld M, van Dijk C, van der Sman-de Beer F, et al. Odour annoy-
ance in the neighbourhood of livestock farming - perceived health and
health care seeking behaviour. Ann Agric Environ Med. 2015;22:55–61.
34. Marshall BM, Levy SB. Food animals and antimicrobials: impacts on
human health. Clin Microbiol Rev. 2011;24:718–733.
35. Huijbers PM, Graat EA, Haenen AP, et al. Extended-spectrum and
AmpC β-lactamase-producing Escherichia coli in broilers and
people living and/or working on broiler farms: prevalence, risk
factors and molecular characteristics. J Antimicrob Chemother.
2014;69:2669–2675.
36. Van den Broek IV, Van Cleef BA, Haenen A, et al. Methicillin-resistant
Staphylococcus aureus in people living and working in pig farms.
Epidemiol Infect. 2009;137:700–708.
37. Veldman KT, Mevius DJ, Wit B, van Pelt W, Heederik D. MARAN
2018 Monitoring of Antimicrobial Resistance and Antibiotic Usage
in Animals in the Netherlands in 2017. Lelystad, The Netherlands:
Wageningen Bioveterinary Research (WBVR); 2018.
38. Pijnacker R, Reimerink J, Smit LAM, et al. Remarkable spatial variation
in the seroprevalence of Coxiella burnetii after a large Q fever epidemic.
BMC Infect Dis. 2017;17:725.