Trait reactance and trust in doctors as predictors of vaccination behavior, vaccine attitudes, and use of complementary and alternative medicine in parents of young children.

OBJECTIVE: The aim of the present study was to investigate whether anti-vaccination attitudes and behavior, and positive attitudes to complementary and alternative medicine (CAM), are driven by trait reactance and a distrust in medical doctors.
METHODS: The sample consisted of 770 Finnish parents who filled out an online survey. Structural equation modeling (SEM) was used to examine if trait reactance plays a role in vaccination decisions, vaccine attitudes, and in the use of CAM, and whether that relationship is mediated by trust in medical doctors.
RESULTS: Parents with higher trait reactance had lower trust in doctors, more negative attitudes to vaccines, a higher likelihood of not accepting vaccines for their children and themselves, and a higher likelihood to use CAM treatments that are not included in evidence-based medicine. Our analyses also revealed associations between vaccination behavior and CAM use and vaccine attitudes and CAM use, but there was no support for the previous notion that these associations would be explained by trait reactance and trust in doctors.
CONCLUSIONS: Taken together, higher trait reactance seems to be relevant for attitudes and behaviors that go against conventional medicine, because trait reactance is connected to a distrust in medical doctors. Our findings also suggest that high trait reactance and low trust in doctors function differently for different people: For some individuals they might be associated with anti-vaccination attitudes and behavior, while for others they might be related to CAM use. We speculate that this is because people differ in what is important to them, leading them to react against different aspects of conventional medicine.

Introduction

Vaccination is widely regarded as one of the most important public health achievements. Thanks to successful immunization programs, many serious and highly contagious diseases have become rare and, in some cases, eliminated and even eradicated [1]. Despite the unquestionable benefits of vaccination, previous research has shown that many individuals have concerns about accepting vaccines for themselves or for their children, and some individuals choose to delay or reject vaccinations altogether [2-6]. This phenomenon, labeled vaccine hesitancy [7], poses a threat to global health, as it undermines vaccination coverage and can lead to outbreaks of vaccine-preventable diseases (see [8], for figures on measles outbreaks). Why then do some individuals hesitate in their decision to get vaccinated, or refuse vaccinations altogether, despite the medical consensus about the safety and benefits of vaccines, and the risks of not getting vaccinated?

The results from a number of studies show that the decision to get vaccinated is a complex process that can be influenced by a wide range of factors (for reviews, see e.g., [5,9-15]). Studies that aim at identifying key determinants of vaccination decision-making, suggest that vaccine acceptance is more likely among individuals who perceive vaccines as available, affordable, beneficial, safe, and effective, and who trust the actors involved in vaccinations [7,15,16]. However, even though the relationship between people’s vaccine attitudes and their vaccination behavior has received a lot of research interest, the questions of why some, but not all, individuals perceive vaccines negatively, has not been studied as extensively and systematically. Hence, an awareness of which factors influence people’s vaccine attitudes is important, for example, when designing interventions that address negative vaccination attitudes. This is because it is the “underlying fears, identity issues and worldviews that motivate people to embrace the surface attitudes” (17; p. 308). Therefore, attempts to increase vaccination uptake may be inefficient if these underlying factors are not properly considered and addressed [17].

One way to think about vaccine attitudes is that they reflect an individual’s tendency to agree with the medical consensus that approved vaccines are safe and beneficial. Hence, considering vaccines to be unnecessary or unsafe, means having opinions or beliefs that go against a medical consensus. To investigate this idea further, one line of research has focused on exploring if individuals embrace negative attitudes to vaccines because they have a general unwillingness to accept scientific evidence. One of the first studies on this topic [18] examined the association between conspiratorial thinking, political worldviews, and attitudes to vaccines in a sample of 1001 adults in the U.S.. The study showed that negative attitudes to vaccination were related to a higher tendency for conspiratorial thinking. According to the authors of that study, conspiratorial ideation stands in direct opposition to scientific reasoning, which may explain why individuals with a tendency for conspiratorial thinking would be motivated to reject scientific evidence that challenges their beliefs [18]. Conspiratorial thinking was also found to be an important predictor of parental vaccination decisions in a recent survey of 4010 U.S. adults [19]. In that study, parents with higher levels of conspiratorial thinking were more likely to have delayed vaccines for their children. A higher belief in conspiracy theories has recently been shown to be related to negative vaccine attitudes also among 518 U.S. adults [20] as well as in a sample of adults from 24 countries [17]. In the latter study, a higher belief in conspiracy theories was shown to be related to more negative attitudes to vaccines in all countries.

Another suggested predictor of opposition to vaccinations is reactance. Reactance refers to the motivational state that arises when people feel that their behavioral freedom has been threatened or taken away [21]. When this occurs, individuals may act contrary to the prescribed action in order to protect or restore their feeling of freedom and control. In a recent study, Hornsey et al. [17] found that individuals with higher trait reactance were more likely to reject vaccinations. Trait reactance refers to an individual’s predisposition to perceive situations as threats to his/her freedom and to act with reactance (for an overview of trait reactance, see e.g., [22]). The role of reactance in the vaccination context is unsurprising because national immunization programs, the medical consensus around the benefits and safety of vaccines, and the fact that accepting vaccines is considered the norm, may be perceived as threats to people’s freedom of choice. Reactance may manifest itself as negative attitudes towards vaccines and medical authorities, and in some individuals, even in a behavior that favors the option that they feel has been taken away from them, that is, to postpone vaccinations or to not get vaccinated altogether. Therefore, if reactance is the motive behind the negative perceptions of vaccines, educational interventions by health authorities, which represent one of the most widely used methods to counter negative attitudes to vaccines, may prove inefficient (for reviews on interventions, see e.g., [9,23,24]). Attempts at improving vaccine-related knowledge and correcting misperceptions about vaccinations by presenting scientific evidence, may in fact backfire and result in even stronger negative attitudes to vaccines [25,26]. Reactance may thus undermine the efficiency of educational interventions.

Other studies have looked at the relationship between vaccine attitudes and the unwillingness to agree with the medical consensus from a different angle; namely, from the perspective of complementary and alternative medicine (CAM). In an Australian study with adults, Browne et al. [27] showed that negative attitudes to vaccines were associated with a tendency to prefer CAM over conventional medicine. On the basis of those results, and the fact that CAM refers to treatments and substances that are not included in evidence-based medicine, Browne et al. [27] speculated that negative attitudes to vaccines might be related to a reluctance to accept conventional medicine, and to a distrust in authorities providing that kind of evidence. This speculation received support in a later qualitative study [28] with 29 Australian parents who had rejected or postponed vaccines. For many of the parents, CAM was considered a natural way to strengthen the immune system, whereas vaccines were considered toxic and harmful. Many of the parents who reported using CAM also mentioned the importance of trusting one’s own expertise in knowing what is best for his/her own children. Finally, for many of the parents, CAM also represented an expert system that is free from the influence of “Big Pharma” and that stands in opposition to conventional medical epistemology. The relationship between vaccine hesitancy, use of CAM, trust in CAM, and trust in conventional treatments was recently investigated in 5,200 Spanish adults [29]. Even though the results showed that more CAM use was associated with greater vaccine hesitancy, a distrust in conventional treatments played a more important role in explaining vaccine hesitancy than did trust in CAM. Based on this, the authors speculated that people do not become vaccine hesitant because they trust CAM, but rather because they distrust conventional medicine. The connection between positive attitudes to CAM and negative attitudes to vaccines has recently been found also among parents in 18 European countries [30], and adults living in America [31]. Finally, the results from an Australian study with 2758 adults [32], indicated that the negative association between CAM and vaccine attitudes could largely be explained by magical beliefs about health, which lends support to the idea that negative attitudes to vaccinations, as well as CAM, may be due to an underlying view on health that is not evidence-based.

In the present study, we wanted to shed more light on the role of trait reactance and trust in medical doctors in the vaccination context in parents of young children. This population is highly relevant when studying vaccination acceptance, because decisions about vaccinations are of immediate importance for this group. However, instead of looking only at attitudes, which was the focus of the study by Hornsey et al. [17], we explored vaccination behavior as well, that is, whether the parents had accepted childhood vaccines for their children and influenza vaccines for themselves. The second aim of the present study was to investigate the role of trait reactance and trust in medical doctors in predicting parents’ use of CAM. Previous studies have suggested that both negative attitudes to vaccines and positive attitudes to CAM may be due to an underlying unscientific view on health and a reluctance to adhere to evidence-based medicine [27,28,32]. The present study tests these speculations in the following two ways: 1) by exploring to what degree trait reactance and trust in medical doctors predict anti-vaccination attitudes and behavior, and CAM use, and 2) by investigating if the association between anti-vaccination attitudes and behavior and CAM use, can be explained by trait reactance and trust in medical doctors. The assumption that trait reactance plays a role also in the decision to use CAM, is based on the idea that CAM represents nonconventional treatments that fall outside the prevailing medical recommendations. Reactance may thus manifest itself in use of CAM in individuals who experience conventional medicine as a threat to their freedom of choice. To the best of our knowledge, this is the first study to look at actual vaccination behavior in this context.

We used structural equation modeling (SEM) to examine if trait reactance predicted vaccination behavior (accepting influenza vaccines for oneself and childhood vaccines for one’s children), vaccine attitudes, and CAM use, and whether these relationships were fully or partially mediated through trust in doctors. Because previous studies have suggested that negative attitudes to vaccines and positive attitudes to CAM are related to reluctance in accepting conventional medicine [27-29,32], we hypothesized that higher trait reactance would predict lower trust in doctors, more negative attitudes to vaccines, a lower likelihood of accepting vaccines and a higher likelihood of using CAM. Finally, we tested the hypothesis that trait reactance and trust in doctors would explain some of the association between the vaccine-related outcomes and CAM use.

Materials and methods

Study context

In Finland, childhood vaccinations are administered free of charge at child health clinics in accordance with the national vaccination program [33]. The influenza vaccines are included in the national vaccination program free of charge for all risk groups, including children under the age of 7 years. All vaccinations are voluntary.

Participants and procedure

An invitation to participate in a 20-minute electronic survey was sent out per mail to 3401 Finnish parents participating in the FinnBrain Birth Cohort Study (hereafter called Finnbrain), which is an ongoing longitudinal project investigating child development [34]. All parents who received the invitation were caregivers to at least one child younger than 4.5 years. In all, 833 parents responded to the survey, but for 50 of them, informed consent was missing, and 13 indicated that they did not allow their responses to be connected to previously gathered data. These individuals were excluded, resulting in a sample of 770 parents (response rate 22.6%; Table 1). Their mean age was 36.43 years (SD = 4.87, range = 22–61). In 155 cases, both parents of the same child had answered the survey.

Table 1

Descriptive information about the participants.

Variable	n	%
Sex
Female	500	64.94
Male	270	35.06
Language
Finnish	648	84.16
Swedish	122	15.84
Educationa
Basic/Upper secondary	180	23.38
University of applied sciences	216	28.05
University	288	37.40

an = 684.

Ethics statement

The study received ethical permission by the Ethics Committee of the Hospital District of Southwest Finland. In the invitation letter, the parents received information about the study and that they could terminate their participation at any time. All parents were asked to give their informed consent to participate and to indicate whether they allowed their responses to be connected to their personal data previously collected in the project.

Measures

The survey was administered in either Finnish or Swedish, depending on the preference of the participant. The measures included in the current study are described below. See S1–S3 Questionnaires for the questionnaires in English, Swedish, and Finnish.

Childhood vaccination behavior

The following three questions queried parents’ past vaccination behavior concerning their children’s childhood vaccinations: 1) Have you ever hesitated in letting your child(ren) receive any of the childhood vaccines?, 2) Have you ever postponed a childhood vaccination for your child(ren)?, and 3) Have you ever decided not to let your child(ren) receive any of the childhood vaccines? The parents could answer either “yes” or “no” to each question. These questions were combined into a single measure of childhood vaccination behavior as follows: 0 = had never hesitated in a childhood vaccination decision, or postponed or rejected a childhood vaccine, 1 = had hesitated or postponed, but not rejected, a childhood vaccine, 2 = had rejected a childhood vaccine. The response was coded as 0 if the child had medical contraindications for vaccination.

The parents were informed that the term “childhood vaccines” referred to the vaccines included in the national vaccination program for children up to the age of six: the rotavirus vaccine, the chickenpox vaccine, the pneumococcal conjugate vaccine (PCV), the DTaP-IPV-Hib (”5-in-1”) vaccine, the MMR vaccine, and the DtaP-IPV (”4-in-1”) vaccine.

Influenza vaccination behavior

To get a measure of influenza vaccination behavior, the parents were asked whether they had taken the influenza vaccine for themselves during the preceding influenza season. The response alternatives were coded as: 0 = had received the vaccine against influenza, and 1 = had not received the vaccine against influenza. The response was coded as 0 if the parent had a medical contraindication for vaccination.

Use of CAM

To measure the use of CAM, the parents were presented with a list of 39 CAM items, from which they were asked to select the ones they had used during the past 12 months to treat an illness or to maintain good health. For the purpose of the present study, we included those CAM items that are not in the Finnish Current Care Guidelines [35], which are national evidence-based guidelines for the treatment and prevention of diseases in medical practice. The final list included the following 18 items: colloidal silver, turmeric, ginger, health powders, natural products for flu, aloe vera, kombucha, cupping, healing, laying on of hands, reiki, the Rosen method, zone therapy, salt therapy, chakra therapy, homeopathy, oil-pulling, and Ayurveda. The CAM variable was coded according to the number of CAM items used (0, 1, 2, 3, or 4 or more items).

Reactance

Trait reactance was measured with the 14-item version of the Hong Psychological Reactance Scale (HPRS; [36]). For each of the statements, the parents were asked to indicate their agreement on a scale from 1 (completely disagree) to 5 (completely agree). Only nine of those items were used in the analyses, based on a study that investigated the factor structure of the 14-item HPRS using the Finnish-speaking respondents of the present sample [37]. A higher HPRS score indicates higher trait reactance.

Trust in doctors

Six statements for measuring trust in doctors were created for the study (S1 Table; e.g., “I let doctors make the decisions concerning my health”, “I trust doctors' ability to make correct diagnoses”). The statements were of varying polarity, and the participants were asked to indicate their agreement with each statement on a scale from 1 (completely disagree) to 4 (completely agree). Reverse-scored items were recoded so that a higher score indicated more trust.

Vaccine attitudes

Attitudes towards the benefit and safety of vaccines were measured with 15 statements created by the authors after literature review and discussions (S1 Table). The statements concerned childhood vaccines and vaccines in general (e.g., “The risk of side effects outweighs the protective benefits of childhood vaccines”, “Vaccinating healthy children helps to protect others by stopping the spread of disease”), and influenza vaccines (e.g., “The risk of side effects outweighs the protective benefits of influenza vaccines”, “It is not worth getting the influenza vaccine, as the influenza symptoms are not serious”). The participants were asked to indicate their agreement with each statement on a scale from 1 (completely disagree) to 4 (completely agree). The polarity of the statements varied, but the items were recoded so that higher scores indicated more positive attitudes.

Statistical analyses

A preregistration of the statistical analyses can be found at [https://osf.io/wda4k?view_only=a3406aee4dbc45d2a094b56ec9a29525]. See S1 Preregistration for changes to the preregistered analyses. The analyses were conducted using structural equation modeling (SEM) in Mplus 8.4 [38]. SEM models can be used for modeling relationships between both latent and observed variables. As the present data collection was cross-sectional, the analyses cannot establish causality between the variables but allows us to test whether our data are consistent with a putative causal model. Trait reactance (Reactance; nine indicators), trust in doctors (Trust; six indicators), and vaccine attitudes (VaccAtt; 15 indicators) were represented by latent factors in the analyses. We first conducted confirmatory factor analyses (CFA) to test the fit of the factors. Second, we assessed the zero-order correlations between all measures. Third, in an attempt to replicate the results of Hornsey et al. [17], who showed that higher trait reactance is related to more negative attitudes to vaccines, we examined whether reactance and trust in doctors predicted vaccine attitudes, by specifying a structural regression (SR) model with the vaccine attitudes factor as the outcome measure (Model 1). The vaccine attitudes factor was regressed on reactance and trust in doctors. Trust was also regressed on reactance to investigate whether trust mediated the associations between reactance and vaccine attitudes. Fourth, to examine our main research questions, we specified a similar SR model with vaccine behavior and CAM use as the outcome variables (Model 2). The outcome variables were again regressed on reactance and trust and trust was regressed on reactance.

As a fifth step, we investigated whether reactance and trust in doctors explained the possible associations between vaccination attitudes and CAM use, and vaccination behavior and CAM use. This was done by assessing whether the disturbance correlations between the outcome measures were weaker than the zero-order correlations between the outcomes. Disturbance correlations constitute the correlations between the proportions of the variances that are not explained by the model. If the disturbance correlations are weaker than the zero-order correlations, it means that the model explains variance that is shared between the outcome measures. To obtain the disturbance correlation between vaccine attitudes and CAM use, the CAM use measure was included as an outcome in Model 1.

Robust WLS (WLSMV) estimation was applied in the SR and CFA analyses, as the indicators and outcome variables were ordinal and responses were non-normally distributed. The relationships between the measures are represented by probit regression coefficients. This coefficient indicates the change in the outcome variable’s standard normal distribution (z-score), given a one-unit increase in the predictor. As the data partly consisted of parents from the same family, responses can be considered clustered. Because of this, non-independence between observations was accounted for when computing standard errors and χ2 statistics in all analyses. Missing data were handled with pair-wise deletion.

Results

The parents’ responses on the outcome variables are presented in Table 2. A majority of the parents had never hesitated in a childhood vaccination decision or postponed or rejected a childhood vaccine. Half of the parents had taken the influenza vaccine the preceding season. Most parents reported that they had not used any of the CAM items during the past 12 months. The parents’ responses to the statements of the four factors can be seen in S2 and S3 Tables.

Table 2

Parents’ responses concerning vaccination behavior and CAM use.

Variable	n	%
Childhood vaccination
No hesitation/postponing/rejectiona	559	73.46
Hesitated	187	24.57
Postponed	98	12.88
Rejected	55	7.22
Influenza vaccination
Yes	391	51.72
No	365	48.28
CAM use
No	483	62.73
One item	161	20.91
Two items	66	8.57
Three items	34	4.42
Four or more items	26	3.38

The responses to Hesitated, Postponed, and Rejected are not mutually exclusive, as a parent may have answered yes to all three questions.

aIncludes nine individuals who reported that their child had medical contraindications for vaccination.

Latent factor modeling

The factors Reactance, χ2(26) = 155.18, CFI = .951, TLI  =   .932, RMSEA =   .081; 90% CI[.069, .093], SRMR = .040, and Trust, χ2(8) = 36.35, CFI = .993, TLI  =   .987, RMSEA =   .068; 90% CI[.047, .092], SRMR = .022, showed appropriate fit to the data with one correlated error term in each model. However, the fit of the factor VaccAtt was unsatisfactory, χ2(90) = 816.57, CFI = .858, TLI  =   .835, RMSEA =   .103; 90% CI[.097, .110], SRMR = .076. The residual covariance matrix indicated that the model underestimated the relationships among the indicators concerning influenza vaccine attitudes, whereas the relationships between these indicators and the indicators measuring attitudes to childhood vaccines or vaccines in general, were overestimated. Modification indices also suggested the inclusion of several correlated error terms between the indicators measuring influenza vaccine attitudes. Therefore, we decided to split the VaccAtt factor into two factors: one with the indicators for attitudes towards childhood vaccines or vaccines in general (VaccAttGeneral), and one with the indicators for attitudes towards influenza vaccines (VaccAttFlu). Both VaccAttGeneral, χ2(34) = 112.68, CFI = .964, TLI  =   .953, RMSEA =   .055; 90% CI[.044, .067], SRMR = .038, and VaccAttFlu, χ2(4) = 31.60, CFI = .992, TLI  =   .981, RMSEA =   .095; 90% CI[.066, .127], SRMR = .019, fitted the data well after the inclusion of one correlated error term in each model. All residual correlations specified in the one-factor models were retained in the subsequent analyses.

The factor loadings and variances can be seen in S4 Table. Zero-order correlations between all measures are shown in Table 3. The relationship between reactance and trust was negative and statistically significant, indicating that individuals with higher trait reactance tended to have lower trust in doctors.

Table 3

Zero-order correlations between measures.

Measure	1	2	3	4	5	6	7
1. Reactance	-
2. Trust	-.33	-
3. Childhood vaccination behavior	.16	-.44	-
4. Influenza vaccination behavior	.18	-.21	.33	-
5. CAM use	.08	-.24	.19	.12	-
6. VaccAttGen	-.27	.52	-.66	-.41	-.24	-
7. VaccAttFlu	-.25	.48	-.58	-.78	-.22	.76	-

All other correlations statistically significant at p < .001, except for the correlations between reactance and childhood vaccination behavior (p = .001), and reactance and CAM use (p = .087). Reactance = trait reactance; Trust = trust in doctors; CAM = complementary and alternative medicine; VaccAttGen = attitudes towards childhood vaccines or vaccines in general; VaccAttFlu = attitudes towards influenza vaccines.

Association between reactance and attitudes to vaccines

Due to the split of the vaccine attitudes factor, Model 1, investigating the relationship between reactance, trust, and vaccine attitudes, was re-specified to include two outcome measures: VaccAttGeneral and VaccAttFlu. The model showed good fit to the data, χ(395) = 862.25, CFI = .955, TLI  =   .950, RMSEA =   .039; 90% CI[.036, .043], SRMR = .052. The results revealed that reactance was directly and statistically significantly related to both VaccAttGeneral and VaccAttFlu (Table 4), indicating that parents with higher trait reactance had more negative attitudes to vaccines. Also, the indirect effects of reactance on both vaccine attitude measures, mediated by trust in doctors, were statistically significant. The total effect (the sum of direct and indirect effects) of reactance on VaccAttGeneral was β = .27, SE = .04, t = 6.55, p < .001, whereas the total effect on VaccAttFlu was β = .25, SE = .04, t = 6.19, p < .001.

Table 4

Direct and indirect effects in the SR models.

	Path		Unstandardized				Standardized
			b	SE	t	p	β	SE	t	p
Model 1
Direct effects
Reactance		→ Trust	-0.21	0.03	6.15	< .001	-.33	.04	7.50	< .001
Reactance		→ VaccAttGeneral	-0.13	0.06	2.42	.016	-.12	.05	2.39	.017
Reactance		→ VaccAttFlu	-0.15	0.06	2.30	.021	-.11	.05	2.30	.021
Trust		→ VaccAttGeneral	0.85	0.12	6.97	< .001	.48	.04	11.30	< .001
Trust		→ VaccAttFlu	0.95	0.12	7.70	< .001	.45	.04	10.52	< .001
Indirect effects
Reactance	→ Trust	→ VaccAttGeneral	-0.18	0.04	4.99	< .001	-.16	.03	5.69	< .001
Reactance	→ Trust	→ VaccAttFlu	-0.20	0.04	5.34	< .001	-.15	.03	5.46	< .001
Model 2
Direct effects
Reactance		→ Trust	-0.20	0.03	6.18	< .001	-.33	.04	7.53	< .001
Reactance		→ Influenza vaccine	0.18	0.09	2.14	.033	.12	.06	2.12	.034
Trust		→ Childhood vaccine	-1.13	0.15	7.75	< .001	-.45	.05	9.83	< .001
Trust		→ Influenza vaccine	-0.44	0.15	2.98	.003	-.17	.06	3.10	.002
Trust		→ CAM use	-0.62	0.13	4.68	< .001	-.24	.05	5.23	< .001
Indirect effects
Reactance	→ Trust	→ Childhood vaccine	0.23	0.04	5.49	< .001	.15	.03	5.55	< .001
Reactance	→ Trust	→ Influenza vaccine	0.09	0.03	2.73	.006	.06	.02	2.72	.006
Reactance	→ Trust	→ CAM use	0.12	0.03	4.04	< .001	.08	.02	4.00	< .001

Reactance = trait reactance; Trust = trust in doctors; Influenza vaccine = influenza vaccination behavior; Childhood vaccine = childhood vaccination behavior; CAM = complementary and alternative medicine; VaccAttGen = attitudes towards childhood vaccines or vaccines in general; VaccAttFlu = attitudes towards influenza vaccines.

Association between reactance, vaccination behavior, and CAM use

The SR model including vaccination behavior and CAM use (Model 1) fitted the data well, χ2(126) = 288.78, CFI = .973, TLI  =   .967, RMSEA =   .041; 90% CI[.035, .047], SRMR = .040. The model showed that reactance did not have a statistically significant direct effect on childhood vaccination behavior and CAM use (β = .02, SE = .06, t = 0.35, p = .726, and β = .00, SE = .05, t = 0.02, p = .983, respectively). We therefore compared a more parsimonious model, where these coefficients were constrained to zero, to the unconstrained model. The constrained model did not result in a statistically significant loss of fit, Δχ(2) = 0.16, p = .926. Fig 1 displays the final model, χ(128) = 276.12, CFI = .975, TLI  =   .970, RMSEA =   .039; 90% CI[.032, .045], SRMR = .040.

Fig 1

Standardized estimates (standard errors) from model 2.

Factor indicators, loadings, and variances, as well as disturbances and their covariances are not shown in the figure. The paths from Reactance to Childhood vaccine and CAM use are set to zero. * p < .05; ** p < .01; *** p < .001.

Reactance had a small and statistically significant direct effect on influenza vaccination behavior (Table 4), indicating that parents with higher trait reactance were less likely to have taken the influenza vaccine during the previous influenza season. Furthermore, reactance had small and statistically significant indirect effects on all outcome measures that were mediated by trust in doctors. Hence, the results were consistent with a model where individuals with higher trait reactance are more likely to have lower trust in doctors, and as a consequence, are more likely to have rejected a childhood vaccine for their children and the influenza vaccine for themselves, and to use more CAM. The total effect of reactance on influenza vaccination was β = .17, SE = .05, t = 3.44, p = .001.

Zero-order and disturbance correlations

Tables 5 and 6 show the zero-order correlations between the measures for vaccination behavior, vaccine attitudes, and CAM use, as well as their disturbance correlations from Model 1 with CAM use included, χ(421) = 887.86, CFI = .956, TLI  =   .951, RMSEA =   .038; 90% CI[.034, .041], SRMR = .051), and from the un-constrained Model 2. When it comes to the association between vaccination behavior and CAM use, the disturbance correlations between the outcome variables, after controlling for Trust and Reactance, were lower than the zero-order correlations. However, the confidence intervals of the disturbance correlations were wide and overlapped with the zero-order correlations, suggesting that trait reactance and trust in doctors do not explain the association between vaccination behavior and CAM use. The difference between zero-order correlations and disturbance correlations was larger for the association between vaccine attitudes and CAM than for vaccination behavior and CAM, and the confidence intervals for the disturbance correlations showed minimal overlap with the zero-order correlations. This suggests that reactance and trust may explain a small part of the association between vaccine attitudes and CAM use. It is, however, important to note that the zero-order correlations between the vaccination-related variables and the CAM use variable were small (r range: .12-.24).

Table 5

Zero-order and disturbance correlations [95% CI] between childhood vaccination behavior, influenza vaccination behavior, and CAM use.

Outcome variable	Zero-order			Disturbance
	1	2	3	1	2	3
1. Childhood vaccination behavior	-			-
2. Influenza vaccination behavior	.33*** [.21, .44]	-		.26*** [.14, .39]	-
3. CAM use	.19*** [.09, .29]	.12* [.01, .22]	-	.10 [-.01, .21]	.07 [-.04, .18]	-

Table 6

Zero-order and disturbance correlations [95% CI] between VaccAttGen, VaccAttFlu, and CAM use.

Outcome variable	Zero-order			Disturbance
	1	2	3	1	2	3
1. VaccAttGen	-			-
2. VaccAttFlu	.76 [.72, .81]	-		.68 [.62, .74]	-
3. CAM use	-.24 [-.33, -.16]	-.22 [-.31, -.13]	-	-.14 [-.25, -.04]	-.12 [-.22, -.02]	-

Discussion

Previous studies have suggested that an unwillingness to agree with the medical consensus may lie behind both negative attitudes to vaccines and positive attitudes to CAM [27,28,32]. The present study investigated this idea further by exploring if trait reactance plays a part in vaccine attitudes, vaccination decisions, and in the use of CAM, and whether these relationships are mediated by trust in doctors. To the best of our knowledge, this is the first study that examined the association between trait reactance and actual vaccination behavior, and that jointly investigated the role of trait reactance in predicting vaccine attitudes and CAM use, and vaccination behavior and CAM use.

The results from the present study, conducted in a relatively large sample (N = 770) of Finnish parents of young children, showed that trait reactance had a statistically significant direct effects on the parents’ attitudes to influenza vaccines and to vaccines in general, indicating that higher trait reactance was related to more negative attitudes to vaccines. These results are in line with the study by Hornsey et al. [17]. Our findings, however, shed more light on the relationship between trait reactance and vaccine attitudes by specifying that trust in doctors plays an important role in the association.

Concerning actual vaccination behavior, trait reactance had a small direct effect on the parents’ decision to take the influenza vaccine, but there was no direct effect of trait reactance on the parents’ decisions to accept childhood vaccinations for their children. However, as was the case with vaccine attitudes, all indirect paths between trait reactance and the vaccination behavior variables were statistically significant, meaning that parents with higher trait reactance had less trust in doctors and a smaller likelihood of having accepted vaccines for their children and for themselves. Our results thus extend previous research by showing that trait reactance not only affects attitudes to vaccines, but it has small effects on the actual vaccination decisions as well. The finding that parents with more trust in doctors were more likely to have accepted vaccinations, and more likely to have positive attitudes to vaccines, was also in line with previous studies (for reviews, see e.g., [5,9-15]).

Based on the results, it seems that trait reactance and trust in doctors explain somewhat more of the variance in vaccine attitudes than in actual vaccination behavior. Also, the relationship between vaccination attitudes and CAM use is slightly stronger than the one between vaccination behavior and CAM use. One possible explanation for this discrepancy is that embracing anti-vaccination attitudes may be a way of expressing one’s personal identity and of communicating that to others [39]. However, when it comes down to the actual vaccination decision, it is possible that also people who express anti-vaccination attitudes choose vaccinations after all.

Trait reactance did not have a direct effect on parents’ CAM use, but in the same way as for vaccination attitudes and behavior, there was a statistically significant indirect effect of trait reactance that went via trust in doctors. As expected based on previous studies [27,28,32], the effects of the predictors on CAM use were in the opposite direction, compared to their effects on vaccination behavior, as higher trait reactance was associated with less trust in doctors, which in turn was associated with more use of CAM treatments and substances that are not included in evidence-based medicine.

Taken together, the results of the present study are thus consistent with a model that suggests that one of the reasons why some individuals high in trait reactance have negative attitudes to vaccines, do not accept vaccines for their children and for themselves, and use CAM, is that they have low trust in doctors.

We also tested the speculations put forth in previous studies [27,28,32] that negative attitudes to vaccines and positive attitudes to CAM may be driven by a shared underlying reluctance to agree with the medical consensus. Our analyses indeed revealed weak associations between vaccine attitudes and CAM use (r = -.22 –-.24), which were roughly in line with previous research [31,32], and between vaccination behavior and CAM use (r = .12 –.19), but there was no clear support for the hypothesis that the associations would be explained by trait reactance and trust in doctors. These findings suggest that high trait reactance and low trust in doctors has different consequences for different people. In some individuals, high trait reactance and a distrust in doctors might result in anti-vaccination attitudes and behavior, while for others, they might lead to CAM use. One possible explanation for this is that people vary in what is important to them, leading them to react against different aspects of conventional medicine.

Limitations

As the present study employs a cross-sectional design, all causal interpretations are speculative. However, trait reactance refers to the predisposition to act with reactance in situations that are perceived as threats to the freedom of choice [22]. Individuals who tend to be reactant may embrace attitudes or engage in behavior that go against the option that has been imposed on them. Therefore, the present study assumes that trait reactance results in attitudes and behavior (i.e., distrust in medical doctors, anti-vaccination attitudes and behavior, and use of CAM), and not the other way around.

Another limitation that may affect the validity of the results, is the fact that the present study is based on self-reported attitudes and behavior. The responses may thus have been influenced by factors such as social desirability bias or memory issues. Also, the questionnaires regarding vaccine attitudes, trust in doctors, and CAM use, have not been validated in other samples. However, during the process of developing the questionnaires for the present study, experts in the field assessed the face validity of the questions. Also, when it comes to the questionnaires probing vaccine attitudes and trust in doctors, factor analysis was used to assess the factor loadings of the questions on the constructs and to handle measurement error.

Concerning possible limitations to generalizability, the parents in the present study are part of a birth cohort study that includes health-related measurements during multiple time points over several years [34]. It is therefore possible that these parents have higher trust in doctors and are less reactant than the general population. Finally, the response-rate was rather low, which possibly resulted in selection bias.

Conclusions

The results from the present study involving Finnish parents of young children show that parents with higher trait reactance are more likely to distrust doctors, and because of that, have more negative attitudes to vaccines, and have a higher likelihood of not accepting vaccines for their children and themselves. Parents with higher trait reactance and a distrust in doctors are also more likely to turn to CAM treatments and substances that are not included in evidence-based medicine. Furthermore, high trait reactance and low trust in doctors have different consequences for different people. In some individuals, high trait reactance and a distrust in doctors might result in anti-vaccination attitudes and behavior, while in others, they might lead to CAM use. One possible explanation for this is that people vary in what is important to them, leading them to react against different aspects of conventional medicine.

However, even though reactance is important to keep in mind when addressing parents’ concerns about vaccines, it is important to note that the parents’ use of CAM, their attitudes towards vaccines, and their decisions to accept or reject vaccines, are mainly due to other factors than trait reactance. Also, parents with high trait reactance constitute a clear minority and it would therefore seem plausible to assume that the main focus when trying to increase immunization rates should still be on the non-reactant parents. This is also supported by the results from a recent study showing that mandatory vaccinations are associated with higher vaccination coverage [40].

Survey questions measuring trust in doctors and vaccine attitudes.

(DOCX)

Click here for additional data file.

Parents’ responses to the included items of the HPRS.

(DOCX)

Click here for additional data file.

Parents’ responses to statements measuring trust in doctors and attitudes to vaccines.

(DOCX)

Click here for additional data file.

Factor loadings and variances from confirmatory factor analyses.

(DOCX)

Click here for additional data file.

Transparent changes.

(DOCX)

Click here for additional data file.

Questionnaire translated into English.

(DOCX)

Click here for additional data file.

Questionnaire in Swedish.

(DOCX)

Click here for additional data file.

Questionnaire in Finnish.

(DOCX)

Click here for additional data file.

7 May 2020

PONE-D-20-10006

Trait Reactance as a Predictor of Vaccination Behavior and Use of Complementary and Alternative Medicine in Parents of Young Children

PLOS ONE

Dear Mrs Soveri,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Jun 21 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Peter Karl Jonason

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information.

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: .

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I enjoyed reading this article and think it would be a good fit for PLoSONE. The writing is clear, the lit review thorough, and the data set is impressive. Analyses seem appropriate, and limitations are laid out in a clear and non-defensive fashion. I know that novelty is not necessarily a primary criterion for acceptance in PLoSONE, but I note that this is the first study that I am aware of that has explored relationships between reactance and vaccination behavior (as opposed to attitudes).

Some minor suggestions for improvement:

• I’d prefer if you expanded the correlation table to incorporate the outcome variables as well. I always find correlations reassuring, particularly when there’s the hint of suppression

• I also like to see authors provide a little insight into measures in the main manuscript, rather than just referring people to supplementary files. I’m thinking of the attitudes in particular: 2 or 3 example items would be fine.

• I will send you directly some recent papers of mine that might be relevant to the current analysis: one in JESP and one in SS&M (both are currently in press but can be found online I think). The paper in SS&M is particularly relevant, as it examines the relationship between CAM use, vaxx attitudes, and trust in various medical interventions. The message of the paper is that anti-vaxx attitudes are associated with mistrust of conventional medicine … not so much trust in CAM. So it’s a suspicion issue primarily: people are turning away from vaccinations because they don’t trust conventional medicine, and they are turning to CAM for the same reason. This parallels your data: most of the effect between trait reactance and CAM use is mediated by trust in doctors. It suggest that this mistrust issue is the proximal cause of both antivaxx attitudes and CAM use (and that the link between CAM use and anti-vaxx attitudes might be partly an artifact of having this one predictor in common)

• I found it interesting that reactance predicted attitudes (as in Hornsey et al., 2018) but not behaviours. I think you could make more of that in the Discussion. It suggests to me that for people high in reactance, anti-vaxx views serve a performative, identity-expressive function (communicating something about who you are). But when it comes down to it, these people also go for the safety of the vaccination. In other words they engage in cheap antivaxx talk but they take the vaccinations anyway. Interesting! I talk a bit about this in the Hornsey & Fielding American Psych paper on personal identity expression, and your data provide a good circumstantial case for this process.

That’s it! I thank you for the chance to read your paper, and wish you all the best with your ongoing program of research.

Matthew Hornsey

Reviewer #2: Summary

This is an interesting article that furthers the investigation of the psychology of anti-vaccination attitudes and scepticism towards conventional medicine more generally.

The writing is generally good, and the literature reviewed is comprehensive and pertinent. However, as discussed below, the MS still requires more proof reading to correct a number of minor expression issues throughout.

My version appeared to be missing Figure labels, which contributed to ambiguity in understanding the results. For example, I am assuming the reported effects are standardised, but I'm not sure.

Although the analysis and conclusions are reasonable, I strongly advise that the authors consider a simpler analysis technique. The CFA is sound, but the causal effects can be more transparently reported using conventional regressions. This is related to my concerns regarding the causal structure, and also my feeling that investigating the mediating role of attitudes on behaviour is not very interesting, and tangential to the main focus of the paper.

The overly-complex analysis leads to a lack of confidence on the part of the reader, which is a shame given the hypotheses are simple, and the variables are few. To illustrate, one result the reader will be interested in is comparing the relative influence of Trust and Reactance on CAM and VAX. It appears that Trust might be more important, but even this elementary result is somewhat obscured.

Overall, I believe the article is fundamentally sound, and should ultimately be published. With simplification of the analyses so that the results are more transparent, and a basic edit, I believe it could be.

Details

One question I have in relation to the rationale for the research question provided in the introduction. The reasoning for trait reactance being an explanatory factor in vaccine scepticism is set out well. However, I don't see a similar argument for CAM? Perhaps the idea is that CAM adherents are motivated by reactance against conventional medical advice more generally - but this seems a little more tenuous as compared to vaccination attitudes. Can you please address this issue?

A related point is that the rationale for trust in doctors mediating the effect of reactance on CAM and anti-vac. It's not clear to me that it should be a mediating effect. Indeed, it seems at least (or more) plausible that trust in doctors would add to, or exacerbate (captured by an interaction) the direct effect of reactance. When adopting a path analytic or SEM approach, it's very important to have a very strong rationale for the proposed model being much more plausible than alternative formulations. Can this be provided? Alternatively, the plausible models might be fit and compared.

278 - small correction to terminology. Since you're using SEM, your three constructs are not technically measured, they're latent

Figures - the labelling needs to be improved, and formatted so they are contained within the boxes. 'CV' and 'IV' for example, would benefit from more informative labels. I think the structure was created using automatic software. Manual formatting using software (I can recommend OmniGraffle) is necessary. These diagrams also usually include * and/r standard errors.

Frankly, after seeing the SEM structure, in which (almost) everything is related to everything else, I am more uncomfortable about this analytic approach. SEM or PA models are essentially defined by the causal links that are *not* in the model. We usually are motivated to apply SEM/PA when we hypothesis a much simpler structure than the correlation matrix. Further, especially in an exploratory context, SEM/PA is focused on comparing alternative plausible models.

I think the CFA approach used to refine the constructs (e.g. vaccination attitudes) makes sense. However, the subsequent analyses become very complicated, for an analysis that involves just a few variables. I'm also struggling to relate the beta coefficients mentioned in the text and diagram to those reported in Table 4. None of the standardised effects in Table 4 exceed .21, yet mention of direct, indirect and total effects in the text are often greater.

Overall, the large number of 'effects' the reader has to wade through tends to obscure the results. The issue is compounded by the issue mentioned earlier, whereby specifying that reactances causes (decrease in) trust.

Both those issues could be resolved by putting aside that causal assumption, and specifying simply:

1. Reactance and Trust causes (with potential interaction) vaccination attitudes

2. Trust and reactance cause CV, IV, CAM

There's no real benefit to including attitudes as mediating variables in the main model, since it's trivially true that attitudes drive behaviour. It greatly complicates the results, without providing any real benefit.

This could be done with some regressions. They will allow you, for instance, to provide a straight-forward comparison of the relative influence on trait reactance on anti-vax and CAM.

The reader will have much more confidence in the results, and they will be much more transparent, if the analyses can be simplified. Less is definitely more, when it comes to statistical analyses.

It is beyond the scope of the review process to provide detailed close editing when there are a great many required edits. I have identified expression issues in the first page. However, the authors will need to take steps to improve expression throughout the manuscript. The issues are generally quite minor, but polished expression is needed for journal publication.

64 - widely regarded

70 - Salmon et al

71 delete great

72 can lead to the

76 delete great

76 the decision to vaccinate

80 vaccination decision-making

81 delete for instance

83 "actors in the vaccine chain" odd wording

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

23 Jun 2020

Response to reviewers

Editor’s comments

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Response: The manuscript has now been formatted according to PLOS ONE’s style requirements.

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information.

Response: The parts of the questionnaire that have been developed for this study are now included as supporting information in English, Swedish, and Finnish (S1-S3 Questionnaires).

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

Response: Due to Finnish federal legislation on personal data protection in medical research, the original research data cannot be made available online, but data can potentially be shared with Material Transfer Agreement. Requests and collaboration initiatives can be directed to the Board of the FinnBrain Birth Cohort Study. Please contact data manager Teemu Kemppainen (teekem@utu.fi).

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Response: Captions and in-text citations have been formatted according to PLOS ONE’s requirements.

Reviewer #1

I enjoyed reading this article and think it would be a good fit for PLoSONE. The writing is clear, the lit review thorough, and the data set is impressive. Analyses seem appropriate, and limitations are laid out in a clear and non-defensive fashion. I know that novelty is not necessarily a primary criterion for acceptance in PLoSONE, but I note that this is the first study that I am aware of that has explored relationships between reactance and vaccination behavior (as opposed to attitudes).

Response: We thank the reviewer for the positive feedback. We now highlight the novelty of the study both in the Introduction and the Discussion:

“To the best of our knowledge, this is the first study to look at actual vaccination behavior in this context.” (p. 7)

“To the best of our knowledge, this is the first study that examined the association between trait reactance and actual vaccination behavior, and that jointly investigated the role of trait reactance in predicting vaccine attitudes and CAM use, and vaccination behavior and CAM use.”(p. 20)

Some minor suggestions for improvement:

1. I’d prefer if you expanded the correlation table to incorporate the outcome variables as well. I always find correlations reassuring, particularly when there’s the hint of suppression

Response: The table (Table 3) now includes the zero-order correlations between all measures.

2. I also like to see authors provide a little insight into measures in the main manuscript, rather than just referring people to supplementary files. I’m thinking of the attitudes in particular: 2 or 3 example items would be fine.

Response: Based on the Reviewer’s suggestion, we have added example statements for the Trust and Attitudes measures in the Method section.

3. I will send you directly some recent papers of mine that might be relevant to the current analysis: one in JESP and one in SS&M (both are currently in press but can be found online I think). The paper in SS&M is particularly relevant, as it examines the relationship between CAM use, vaxx attitudes, and trust in various medical interventions. The message of the paper is that anti-vaxx attitudes are associated with mistrust of conventional medicine … not so much trust in CAM. So it’s a suspicion issue primarily: people are turning away from vaccinations because they don’t trust conventional medicine, and they are turning to CAM for the same reason. This parallels your data: most of the effect between trait reactance and CAM use is mediated by trust in doctors. It suggest that this mistrust issue is the proximal cause of both antivaxx attitudes and CAM use (and that the link between CAM use and anti-vaxx attitudes might be partly an artifact of having this one predictor in common)

Response: We thank the reviewer for sending us the relevant recent papers. We have added them to the manuscript.

4. I found it interesting that reactance predicted attitudes (as in Hornsey et al., 2018) but not behaviours. I think you could make more of that in the Discussion. It suggests to me that for people high in reactance, anti-vaxx views serve a performative, identity-expressive function (communicating something about who you are). But when it comes down to it, these people also go for the safety of the vaccination. In other words they engage in cheap antivaxx talk but they take the vaccinations anyway. Interesting! I talk a bit about this in the Hornsey & Fielding American Psych paper on personal identity expression, and your data provide a good circumstantial case for this process.

Response: We wish to thank the Reviewer for the interesting idea. We have added the following section to the Discussion:

“Based on the results, it seems that trait reactance and trust in doctors explain somewhat more of the variance in vaccine attitudes than in actual vaccination behavior. Also, the relationship between vaccination attitudes and CAM use is slightly stronger than the one between vaccination behavior and CAM use. One possible explanation for this discrepancy is that embracing anti-vaccination attitudes may be a way of expressing one’s personal identity and of communicating that to others (39). However, when it comes down to the actual vaccination decision, it is possible that also people who express anti-vaccination attitudes choose vaccinations after all.” (pp. 21-22)

That’s it! I thank you for the chance to read your paper, and wish you all the best with your ongoing program of research.

Matthew Hornsey

Reviewer #2

Summary

This is an interesting article that furthers the investigation of the psychology of anti-vaccination attitudes and scepticism towards conventional medicine more generally. The writing is generally good, and the literature reviewed is comprehensive and pertinent. However, as discussed below, the MS still requires more proof reading to correct a number of minor expression issues throughout.

Response: We thank the Reviewer for the constructive feedback. Please see the responses to the specific comments below.

1. My version appeared to be missing Figure labels, which contributed to ambiguity in understanding the results. For example, I am assuming the reported effects are standardised, but I'm not sure.

Response: In line with PLOS ONE instructions for figures, the Figure captions and notes can be found in the manuscript immediately following the paragraph where the figure is first cited (e.g., Figure 1 on page 18), and not together with the figure.

The effects reported in Fig 1 are indeed standardized. This information has now been added to the figure legend: “Fig 1. Standardized Estimates (standard errors) from Model 2”(p.18)

2. Although the analysis and conclusions are reasonable, I strongly advise that the authors consider a simpler analysis technique. The CFA is sound, but the causal effects can be more transparently reported using conventional regressions. This is related to my concerns regarding the causal structure, and also my feeling that investigating the mediating role of attitudes on behaviour is not very interesting, and tangential to the main focus of the paper.

The overly-complex analysis leads to a lack of confidence on the part of the reader, which is a shame given the hypotheses are simple, and the variables are few. To illustrate, one result the reader will be interested in is comparing the relative influence of Trust and Reactance on CAM and VAX. It appears that Trust might be more important, but even this elementary result is somewhat obscured.

Overall, I believe the article is fundamentally sound, and should ultimately be published. With simplification of the analyses so that the results are more transparent, and a basic edit, I believe it could be.

Response: We have simplified the analyses (described in more detail in our response to Reviewer 2, comment 7). We have kept the analyses within a SEM environment to be able to include the latent constructs in the analyses. Both the relative influence of Trust and Reactance on the outcome measures, as well as the theorized mediating effect of Trust, are more easily interpreted from the new models.

Details

3. One question I have in relation to the rationale for the research question provided in the introduction. The reasoning for trait reactance being an explanatory factor in vaccine scepticism is set out well. However, I don't see a similar argument for CAM? Perhaps the idea is that CAM adherents are motivated by reactance against conventional medical advice more generally - but this seems a little more tenuous as compared to vaccination attitudes. Can you please address this issue?

Response: We wish to thank the reviewer for pointing out that the justification for the hypothesis that trait reactance would predict CAM, was missing in the manuscript. We apologize for this and have added the following paragraph to the Introduction: “The assumption that trait reactance plays a role also in the decision to use CAM, is based on the idea that CAM represents nonconventional treatments that fall outside the prevailing medical recommendations. Reactance may thus manifest itself in use of CAM in individuals who experience conventional medicine as a threat to their freedom of choice.” (p.7)

4. A related point is that the rationale for trust in doctors mediating the effect of reactance on CAM and anti-vac. It's not clear to me that it should be a mediating effect. Indeed, it seems at least (or more) plausible that trust in doctors would add to, or exacerbate (captured by an interaction) the direct effect of reactance. When adopting a path analytic or SEM approach, it's very important to have a very strong rationale for the proposed model being much more plausible than alternative formulations. Can this be provided? Alternatively, the plausible models might be fit and compared.

Response: The rationale behind the mediation model is that we measure trait reactance and not state reactance. Our theory thus suggests that trait reactant individuals will, as a consequence of their reactance, also distrust medical authorities such as doctors, and that this increased distrust manifests itself also as negative attitudes to vaccines. We discuss this in the revised Introduction and Limitation sections:

“Reactance may manifest itself as negative attitudes towards vaccines and medical authorities, and in some individuals, even in a behavior that favors the option that they feel has been taken away from them, that is, to postpone vaccinations or to not get vaccinated altogether.” (p. 5)

and

“As the present study employs a cross-sectional design, all causal interpretations are speculative. However, trait reactance refers to the predisposition to act with reactance in situations that are perceived as threats to the freedom of choice (22). Individuals who tend to be reactant may embrace attitudes or engage in behavior that go against the option that has been imposed on them. Therefore, the present study assumes that trait reactance results in attitudes and behavior (i.e., distrust in medical doctors, anti-vaccination attitudes and behavior, and use of CAM), and not the other way around.”(p.23)

5. 278 - small correction to terminology. Since you're using SEM, your three constructs are not technically measured, they're latent

Response: Based on the Reviewer’s comment, we have checked the manuscript for inaccuracies in terminology. The text on row 278 in the first version of the manuscript referred to the observed outcome measures, which are not latent.

6. Figures - the labelling needs to be improved, and formatted so they are contained within the boxes. 'CV' and 'IV' for example, would benefit from more informative labels. I think the structure was created using automatic software. Manual formatting using software (I can recommend OmniGraffle) is necessary. These diagrams also usually include * and/r standard errors.

Response: We apologize for using unclear labels in the figure and have corrected this in the revised manuscript. We have formatted the diagram and include standard errors and significance levels.

7. Frankly, after seeing the SEM structure, in which (almost) everything is related to everything else, I am more uncomfortable about this analytic approach. SEM or PA models are essentially defined by the causal links that are *not* in the model. We usually are motivated to apply SEM/PA when we hypothesis a much simpler structure than the correlation matrix. Further, especially in an exploratory context, SEM/PA is focused on comparing alternative plausible models.

I think the CFA approach used to refine the constructs (e.g. vaccination attitudes) makes sense. However, the subsequent analyses become very complicated, for an analysis that involves just a few variables. I'm also struggling to relate the beta coefficients mentioned in the text and diagram to those reported in Table 4. None of the standardised effects in Table 4 exceed .21, yet mention of direct, indirect and total effects in the text are often greater.

Overall, the large number of 'effects' the reader has to wade through tends to obscure the results. The issue is compounded by the issue mentioned earlier, whereby specifying that reactances causes (decrease in) trust.

Both those issues could be resolved by putting aside that causal assumption, and specifying simply:

1. Reactance and Trust causes (with potential interaction) vaccination attitudes

2. Trust and reactance cause CV, IV, CAM

There's no real benefit to including attitudes as mediating variables in the main model, since it's trivially true that attitudes drive behaviour. It greatly complicates the results, without providing any real benefit.

This could be done with some regressions. They will allow you, for instance, to provide a straight-forward comparison of the relative influence on trait reactance on anti-vax and CAM.

The reader will have much more confidence in the results, and they will be much more transparent, if the analyses can be simplified. Less is definitely more, when it comes to statistical analyses.

Response: We have simplified the analyses according to the Reviewer’s suggestions, with one exception. Instead of the suggested interaction between Reactance and Trust, we still include Trust as a mediator, as this is central for our hypotheses (please see our response to Reviewer 2, comment 4). We conduct two structural regression models: 1) one where Reactance and Trust predict vaccination behavior and CAM use, and 2) another one where Reactance and Trust predict vaccine attitudes. In both models, Trust is also regressed on Reactance to investigate the mediation hypothesis. This change addresses the Reviewer’s concern on whether the SEM approach is motivated, as all paths in the models now specifically test our hypotheses. The mediating role of attitudes on behavior, for which we did not specify a hypothesis, has been omitted. The new analysis scheme also considerably simplifies the interpretation of the results. To further utilize the SEM approach, we re-specify paths that show very weak and non-significant associations to be fixed to zero, and test whether the fit is significantly reduced.

We thank the Reviewer for the suggestions and agree that the simplified analysis scheme is easier to interpret.

8. It is beyond the scope of the review process to provide detailed close editing when there are a great many required edits. I have identified expression issues in the first page. However, the authors will need to take steps to improve expression throughout the manuscript. The issues are generally quite minor, but polished expression is needed for journal publication.

64 - widely regarded

70 - Salmon et al

71 delete great

72 can lead to the

76 delete great

76 the decision to vaccinate

80 vaccination decision-making

81 delete for instance

83 "actors in the vaccine chain" odd wording

Response: We have corrected all language issues pointed out by the Reviewer. We have also polished the language throughout the manuscript and changed the reference style to the Vancouver system.

Submitted filename: Response-to-Reviewers.docx

Click here for additional data file.

4 Jul 2020

PONE-D-20-10006R1

Trait reactance and trust in doctors as predictors of vaccination behavior, vaccine attitudes, and use of complementary and alternative medicine in parents of young children

PLOS ONE

Dear Dr. Soveri,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Aug 18 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Peter Karl Jonason

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have taken care in the revision and have addressed all my original concerns. My only remaining comment: in revising the abstract, the conclusions section contains quite a bit of causal language, which should be avoided given that it's a correlational study. Apart from this I have no further comments and I don't need to see a revision. I look forward to seeing the paper in print at some stage soon.

All the best, Matthew

Reviewer #2: Thanks for addressing my comments seriously. As well as being more polished overall, the revised analyses and results provide a much clearer and less ambiguous interpretation, which well supports the conclusions. An interesting result, and good work.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Matthew Hornsey

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

5 Jul 2020

PONE-D-20-10006R1

Trait reactance and trust in doctors as predictors of vaccination behavior, vaccine attitudes, and use of complementary and alternative medicine in parents of young children

PLOS ONE

Dear Dr. Soveri,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Aug 18 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Peter Karl Jonason

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

________________________________________

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have taken care in the revision and have addressed all my original concerns. My only remaining comment: in revising the abstract, the conclusions section contains quite a bit of causal language, which should be avoided given that it's a correlational study. Apart from this I have no further comments and I don't need to see a revision. I look forward to seeing the paper in print at some stage soon.

All the best, Matthew

Response: We have now revised the Conclusions section in the Abstract so that there is no causal language in the sentences that relate to our results. We decided to keep the causal language in the last sentence, because that sentence is based on speculation about the meaning of the results. The text now reads as follows:

”Taken together, higher trait reactance seems to be relevant for attitudes and behaviors that go against conventional medicine, because trait reactance is connected to a distrust in medical doctors. Our findings also suggest that high trait reactance and low trust in doctors function differently for different people: For some individuals they might be associated with anti-vaccination attitudes and behavior, while for others they might be related to CAM use. We speculate that this is because people differ in what is important to them, leading them to react against different aspects of conventional medicine.”

Reviewer #2: Thanks for addressing my comments seriously. As well as being more polished overall, the revised analyses and results provide a much clearer and less ambiguous interpretation, which well supports the conclusions. An interesting result, and good work.

________________________________________

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Matthew Hornsey

Reviewer #2: No

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

9 Jul 2020

Trait reactance and trust in doctors as predictors of vaccination behavior, vaccine attitudes, and use of complementary and alternative medicine in parents of young children

PONE-D-20-10006R2

Dear Dr. Soveri,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Peter Karl Jonason

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

13 Jul 2020

PONE-D-20-10006R2

Trait reactance and trust in doctors as predictors of vaccination behavior, vaccine attitudes, and use of complementary and alternative medicine in parents of young children

Dear Dr. Soveri:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Peter Karl Jonason

Academic Editor

PLOS ONE