key: cord-1034104-1qzp73ww
authors: Chandu, Dr. Viswa Chaitanya; Lingamaneni, Dr. Krishna Prasad; Pachava, Dr. Srinivas; Baddam, Dr. Venkat Ramana Reddy; Marella, Dr. Yamuna
title: The influence of dissonance induction and assessment reactivity in improving adherence to Covid-19 precautionary measures: a cluster randomized controlled trial
date: 2021-03-20
journal: Int Dent J
DOI: 10.1016/j.identj.2021.03.001
sha: 7aab0fde120838369804d00180a6e5ef60aee6cc
doc_id: 1034104
cord_uid: 1qzp73ww

BACKGROUND: In the context of the Coronavirus Disease (COVID-19) pandemic, adherence to suggested precautionary measures has been highlighted to be extremely important in preventing and curtailing the spread of COVID-19. However, strict adherence to precautionary measures could be very demanding. MATERIALS AND METHOD: This cluster randomized controlled trial done among 1517 undergraduate dental students tested the effectiveness of ‘Dissonance Induction’(DI) and ‘Assessment Reactivity’(AR) in improving adherence to WHO suggested COVID-19 precautionary measures as compared to a control group. At baseline, participants in the DI group were tested for their knowledge on COVID-19 precautionary measures, immediately followed by assessment of their adherence to these precautionary measures. This methodology was adapted so as to systematically reveal to the participants the poor adherence to their self-held cognitions, should there be any. The magnitude of dissonance was measured as the proportion of such dissonant cognitions held by an individual. In the AR group, at baseline, participants were inquired about their attitudes alone towards COVID-19 precautionary measures. The control group was neither assessed for knowledge and adherence nor for attitudes towards the precautionary measures at baseline. Two weeks after the administration of the aforementioned interventions in the DI and AR groups, the three study groups were assessed for adherence to COVID-19 precautionary measures. RESULTS: The follow-up adherence scores in the DI group were found to be significantly higher (15.11±4.1) compared to the AR (13.13±2.01) and control (12.87±2.97) groups as analyzed by Kruskal-Wallis ANOVA (H=243.5; P<0.001). Wilcoxon signed-rank test showed that the adherence scores significantly improved in the DI group from baseline to follow-up (z=-8.84; P<0.001). Magnitude of dissonance at baseline was found to be a significant predictor of follow-up adherence scores (R(2)=0.255). CONCLUSION: This study found that ‘dissonance induction’ is an easy to administer intervention in this moment of global crisis to bring an immediate and significant change in the people's adherence to COVID-19 precautionary measures.

The fact that Coronavirus disease (COVID-19) was declared as a public health emergency of international concern reflects the magnitude of global crisis the disease has been responsible for. 1 Since SARS-CoV-2, the virus causing COVID-19, is a novel virus with very recent emergence among humans, we are completely immune-naïve and consequently vulnerable. 2 International experiences suggest that the transmission of the disease can be enormous in rather short periods of time. 3, 4 Though COVID-19 is associated with mortality, the case fatality rate was reported to be lesser for COVID-19 (2%) compared to SARS (10%) and MERS (34%). However, it was reported that COVID-19 has been responsible for more deaths than SARS and MERS combined. These findings reflect on the more contagious nature, though less severe, of COVID-19 compared to SARS and MERS. 5 It is for this reason that the greatest challenge ahead of health care systems now is to curtail the rapidity and magnitude of transmission. As on 24 th December, 2020, official reports from Ministry of Health and Family Welfare, Government of India suggest that 10.1 million individuals have been confirmed positive among suspected cases and contacts of known cases. 6 These numbers illustrate the increasing transmission of the disease. A lot of onus has been placed on the citizens of the nation to prevent themselves from getting affected. The WHO suggested precautionary measures of frequent hand washing, social distancing, avoidance of touching face, nose, mouth, practice of respiratory hygiene, maintaining one meter distance from those coughing and sneezing, and refraining from smoking and activities that weaken the lungs have been widely circulated in various media platforms for people to assimilate and adopt these behaviors. 7 However, literature suggests that though persuasive health messages are often successful in bringing attitudinal changes among people, these changes are short lived. More importantly, it has been strongly established by social psychologists that positive attitudes do not necessarily translate to positive behaviors. Therefore, more thorough and scientifically informed behavioral interventions are warranted in order to improve adherence to precautionary measures and combat the outbreak of infectious diseases.

The benefits of answering questions, termed as 'question benefit effect' (QBE), about a behavior in producing positive changes, though minor in magnitude, in that behavior has been previously studied. 8 Various theories have been proposed to explain the behavior change following responding to questions on that behavior. Sherman postulated that mental stimulation takes place while responding to questions about a behavior which result in formation of cognitive representations or behavioral scripts. 9 These cognitive representations get reactivated while the subject performs that behavior and assist a positive behavioral change. Another explanation is the theory of attitude accessibility which proposes that the questions about behavioral intentions of individuals activate the attitudes intrinsic for that behavior, making them more accessible in memory. 10 This theory was tested with regard to healthy eating behaviors. 11 Cognitive dissonance is another explanation for QBE, which attempts to attribute the benefit to the cognitive conflict felt by reporting behaviors inconsistent with their beliefs. Cognitive dissonance, as proposed by Leon Festinger, is an unpleasant drive state similar to hunger or thirst experienced when two psychologically inconsistent cognitions are simultaneously held by a person which is against the inner drive to maintain harmony between one's attitudes / beliefs and behavior. 12 Aronson proposed that dissonance theory makes its clearest predictions when the self concepts of people are violated by their own actions. Aronson also argued that passing on information to individuals in a persuasive manner would only result in attitudinal changes which are temporary. 13 Dissonance induced persuasion is more effective in bringing a behavioral change as the individuals' inconsistent behavior with their own self concept creates a necessity to attain consistency between their self concepts and behavior. Assessment reactivity, on the other hand, is the influence of behavioral assessment at the present time on later behavior. It was proposed that mere questioning about a behavior influences the individual to change the behavior in question. 14 In the present study, the group where inconsistency of the subjects' behavior with their own concepts is explicitly made evident to them by testing their knowledge and behavior at baseline is identified henceforth as 'Dissonance Induction'

(DI) group. The group where mere questioning of attitudes was done at baseline is identified as 'Assessment Reactivity' (AR) group. There is a control group which was neither tested for knowledge, behavior nor for the attitudes at baseline. Our hypothesis (H1) was inducing dissonance among subjects by systematically making it evident to them the poor adherence to their own cognitions, results in better follow-up adherence to the precautionary measures. If the null hypothesis (H0) was to be true, there would be no difference in the follow-up adherence scores between the three study groups.

This cluster randomized controlled trial was conducted in the months of February and March, 2020 following ethical approval from the Institutional Ethical Committee of SIBAR Institute of Dental Sciences (Pr.69/IEC/SIBAR/2020). The allocation ration was 1:1:1 in the Dissonance Induction (DI), Assessment Reactivity (AR), and control groups. Six teaching dental institutions (clusters) of the sixteen functioning dental institutions in the Southern Indian state of Andhra Pradesh were selected and two each were allocated to DI, AR, and the control groups after randomization. The study details were depicted in Figure 1 . All the study participants were provided with the necessary information on the purpose and process of the study without revealing the specific focus and objectives of the study.

Consent was obtained prior to the subjects' participation in the study, and participation in the study was voluntary. Care was taken to ensure anonymity in the questionnaire leaving no place for coercion. Since all the participants within a cluster received either the same intervention or no intervention, contamination between the study groups was not possible. Furthermore, study participants from each individual dental institution were neither aware of the alternative interventions tested in the study, nor did they know about the parallel conduct of the study among students from other dental institutions.

In the DI group, dissonance was induced at baseline by administering a self-administered, structured questionnaire that reveals the existing cognitive conflict, should there be any. This questionnaire consisted of two sections: knowledge on the COVID-19 precautionary measures; adherence to precautionary measures. To assess knowledge on COVID-19 precautionary measures, a combination of six World Health Organization (WHO) suggested precautionary measures and six distractor options was given for the participants to choose from. The adherence was assessed immediately after the participants had responded to the knowledge questions. This questionnaire serves two purposes: making the subjects explicitly mention their self-concepts in the 'knowledge' section; making it evident for the subjects how consistent/inconsistent their actions are with their own cognitions in the 'adherence section'. Dissonance was considered to be induced if a subject chooses a WHO suggested precautionary measure in the knowledge section, and reports his/her current adherence to that precautionary measure to be 'occasional' or 'never' as the poor adherence of the subjects with their self-held cognitions was systematically revealed to them. On the other hand, consonant cognitions refer to WHO suggested precautionary measures chosen by the subjects in the knowledge section for which they report their adherence to be 'often' or 'almost always'. The knowledge score (ks) of a subject 'i' was calculated by the formula ks(i)= N-Z, as suggested by Kurz, where N is the number of correct responses chosen by the subject 'i' and Z refers to the number of incorrect responses. 15 This ensures that a subject who chooses all the responses randomly will get a 'zero' knowledge score. The baseline adherence to WHO suggested COVID-19 precautionary measures Adherence score of a subject at the follow-up was calculated as the sum of scores obtained for the responses to the six precautions of interest. Therefore, the follow-up adherence scores range from 0-18, with '0' indicating non-adherence to all the six precautionary measures and '18' indicating perfect adherence to all the precautionary measures. A total of sixty students were lost to follow up across both the intervention groups and were excluded from the analysis. The final sample included 1517 subjects with 522 in the DI group, 507 in the AR group and 488 in the control group.

Statistical Analysis: SPSS version 20 software was used and descriptive statistics, Wilcoxon signed rank test to identify the change in adherence scores from baseline to follow-up in the DI group, Spearman's correlation test to check the correlation between attitudinal scores at baseline and follow-up adherence scores in the AR group, Kruskal-Wallis ANOVA to test the differences in follow-up adherence scores between the study groups, and simple linear regression analysis in the DI group with follow-up adherence score as the dependent variable and magnitude of dissonance as explanatory variable were employed in data analysis.

The response rate was 68.4% in the DI group, 66% in the AR group and 63.2% in the control group. There was no significant difference in the gender distribution between the study groups with female students representing 80.07% (418) of the DI group, 77.51% (393) of the AR group, and 78.07% (381) of the control group. Table 1 shows the participants' responses to knowledge questions in the DI group at baseline. The mean knowledge score of COVID-19 precautionary measures among the participants from the DI group was 3.23±2.15 (95% CI 3.04 -3.41), after adjusting for the incorrect responses. The mean score for magnitude of dissonance was observed to be 0.26±0.18, which suggests that the proportion of dissonant cognitions on an average was slightly more than 25%. The mean adherence to precautionary measures score in the DI group at baseline was 13.02±3.74 (95% CI 12.69 -13.34).

The mean attitudinal score towards COVID-19 precautionary measures in the AR group at baseline was 14.02±3.6, where a score of 18 was indicative that the subject is most likely to practice all the six precautionary measures suggested.

An apparent difference was noted between the three study groups with regard to their mean follow-up adherence scores in practicing the COVID-19 precautionary measures, with the DI group demonstrating higher scores compared to AR and control groups (Figure 2 ).

Significantly higher adherence scores in DI group were suggestive of the more thorough adherence to precautionary measures in the DI group compared to AR and control groups at follow-up. No significant differences were noted between the AR and control groups in post hoc tests for multiple pair wise comparisons, while statistically significant differences were noted for DI group with both AR and control groups ( Table 2) .

The adherence scores in the DI group significantly improved from baseline to follow-up (Wilcoxon Z statistic = -8.84; p<0.001) ( Table 3) . It was noted in this group that the follow-up adherence scores exhibited a positive linear relation with the magnitude of dissonance at baseline (Spearman's rho = 0.505; P<0.001). Magnitude of dissonance was found to be a significant predictor of the follow-up adherence scores in linear regression analysis (Table 4) .

In the AR group, there was no correlation between the adherence scores of the study participants at follow-up and their attitudinal scores at baseline (Spearman's rho = -0.006; p=0.88).

Dissonance induction was found to be an effective way in bringing positive changes in adherence to COVID-19 precautionary measures in the present study, demonstrating significant positive differences with the AR and control groups, and consequently, the null hypothesis (H 0 ) can be rejected in support of H1. This study attempted to make subjects in the DI group experience the cognitive conflict, at baseline, by testing their knowledge on and immediately inquiring about their practice of COVID-19 precautionary measures. Such inquiry results in induction of dissonance, a state of cognitive conflict, by making the discrepancy between cognition and behavior explicitly evident to the respondent. This may influence people to adopt behaviors that, they believe, are health promoting in nature. Aronson et al. referred to this method as creation of dissonance by hypocrisy induction. 16 Aronson et al. successfully tested the effectiveness of this method in overcoming denial and improving the intentions to use condoms, where hypocrisy was induced by having the subjects publicly advocate condom use and then systematically making the subjects aware of their own previous failures in condom use. 16 Dickerson CA et al. reported behaviors conducive to water conservation among the hypocrisy induced group. 17 Aronson argued that the change in people's attitudes through informational campaigns is short lived, as such a change is brought about by an external source. They proposed that little investment of the self in formation of the attitude is the reason behind the impermanent nature of these changes and such attitudes are vulnerable for change if there is a stronger counterargument by a different external source in future. Aronson et al. suggested that dissonance-generated persuasion is effective in bringing a long term change as there is a scope for critical reflection and investment of the self in the process of attitudinal or behavioral change. 18 In the present study, magnitude of dissonance explained more than a quarter of variance in the follow-up adherence scores in the DI group adding strength to the argument that induction of dissonance is effective in bringing positive health behavioral changes. Wilding S et al. reported that dissonance-enhanced 'question benefit effect' (QBE) condition was more effective in health behavioral modification compared to a standard QBE intervention. 19 Dissonance was also discussed as the most plausible mechanism explaining QBE in the meta-analyses conducted by Wood C et al. 11 and Spangenberg et al. 20 Besides induction of dissonance, another intervention tested in this study was mere questioning about the attitudes towards COVID-19 precautionary measures at baseline, without assessing the knowledge and practice of these measures. The existing evidence regarding whether mere questioning of the attitudes relating to a behavior may bring a positive behavioral change is equivocal. 14, 18, 21 In the present study there was no significant difference between the AR and control groups in the follow-up adherence scores. Moreover, the attitudes towards practice of the precautionary measures at baseline in the AR group demonstrated no correlation with the follow-up adherence scores. Spangenberg et al. also reported no differences between participants who were asked and not asked to predict their behavior. 22 These findings, however, were inconsistent with those reported by Wood C et al., where participants who were asked to report their intentions demonstrated more accessible attitudes compared to those who were not. 11 Ayres K et al. proposed that QBE alone is insufficient in promoting health behaviors; a combination of motivation and QBE was reported to be effective in significantly increasing behavior in a randomized controlled trial. 23 Thus, the present study adds strength to the existing QBE research and postulates dissonance induction, over mere questioning about attitudes, as an efficient intervention to promote healthy behaviors with regard to COVID-19 precautionary measures.

Possessing concrete insights into positive health behaviors does not warrant practice of these behaviors. 24 Also, It was reported in literature that people believe their behavior to be better than their actual behavior. 25 Therefore, one of the fundamental goals of health behavioral research is to close the gap between knowledge and behaviors. As a cost effective alternative to close this gap, QBE was previously tested in different domains of health care such as health screening 26 , health check-ups 27 , vaccination 28 , adoption of health promoting behaviors and reduction of risk behaviors 19 . However, this is the first time, to our knowledge, that QBE has been tested for behavioral change with respect to limiting the spread of an infectious disease where strict adherence to suggested precautionary measures is regarded as the best way to protect ourselves. It is evident from this study that though the study population possesses good knowledge of the COVID-19 precautionary measures to be followed, they also hold some misconceptions among which eating plenty of garlic was the most common. Other common misconceptions identified were: need to take regular hot water baths; need to use hand dryers after hand wash; need to stay in hot and humid climates. Refraining from smoking was identified by only less than half of the participants in the intervention group as a precautionary measure. The problem with holding erroneous beliefs is that the unwarranted practice of these beliefs may act as a compensatory mechanism for people to ignore the actual precautionary measures to be followed. For instance, a person from India who holds a notion that staying in hot and humid climates is protective against SARS CoV-2, may not feel the necessity to practice social distancing. This study highlights some of the erroneous notions held by the dental students and gives a heads-up on the necessity to more effectively communicate the precautionary measures to reduce the transmission of COVID-19.

The limitations of this study are: randomization at dental institutional level; short follow-up time; increased accessibility to COVID-19 related information through the study period. Randomization was not done at participant level to prevent contamination bias. 29 Students from the same institution randomized to any of the three study groups may share their experiences with colleagues in a different study group than theirs; this is the reason why cluster randomization was preferred. However, cluster randomization carries the risk of reduced statistical power. A follow up time of only two weeks was considered in the present study. It was reported in literature that QBE decays with time. However, in the context of epidemics, even short term changes in behaviors towards the positive are of tremendous importance. In literature, follow-up times after administering the questionnaire varied over a wide range in the previous QBE research with Van Kerckhove et al. 30 measuring the dependent variable immediately after questioning and Murray et al. 31 reported a time interval of five years between administration of questionnaire and measurement of the outcomes. It is important to point here that all the study participants, regardless of the study groups, had increased accessibility to COVID-19 information through the study period which may have an influence on their adherence to the precautionary measures. Nevertheless, access to information is common for participants in all the study groups, and any possible influence could have affected the follow-up adherence scores in all the groups.

Another limitation in this study is the gender imbalance with 78.5% of the study participants being females. However, this imbalance is consistent with the existing gender based imbalance in enrollment into dentistry in India. 32 Dissonance induction was found to be a low-cost, rapid, and effective intervention in this study to improve adherence to COVID-19 precautionary measures. We propose that a more definitive argument in favor of making use of dissonance induction as a cost-effective method can be made by comparing the adherence levels to healthy behaviors between dissonance induction and strict auditing. Such establishment of dissonance induction as a cost-effective alternative goes a long way in the determination of the choice of interventions for behavior change, not just with regard to COVID-19, but in the broader context of various health behaviors. Future research on QBE in health behavioral research must also attempt for randomization at an individual level and optimum follow-up times. 

2019-nCoV outbreak' (Statement on the second meeting of the International Health Regulations (2005) Emergency Committee regarding the outbreak of novel coronavirus (2019-nCoV)

Q&A: The novel coronavirus outbreak causing COVID-19

COVID-19 and Italy: what next?

Estimates of the severity of coronavirus disease 2019: a model-based analysis

Coronavirus covid-19 has killed more people than SARS and MERS combined, despite lower case fatality rate

COVID-19 state-wise status. Ministry of Health and Family Welfare, Government of India

Corona virus prevention

The questionbehavior effect: What we know and where we go from here

On the self-erasing nature of errors of prediction

A critical review of question-behavior effect research

Why does asking questions change health behaviours? The mediating role of attitude accessibility

A theory of cognitive dissonance

Dissonance theory: Progress and problems

Effect of pretesting on intentions and behaviour: A pedometer and walking intervention

A review of scoring algorithms for multiple-choice tests. Paper presented at the annual meeting of the Southwest Educational Research Association

Overcoming denial and increasing the intention to use condoms through the induction of hypocrisy

Using cognitive dissonance to encourage water conservation

Persuasion via self-justification: Large commitments for small rewards

Retrospection on socialpsychology

Using the question-behavior effect to change multiple health behaviors: An exploratory randomized controlled trial

A meta-analytic synthesis of the questionbehavior effect

Asking questions changes behavior: mere measurement effects on frequency of blood donation

Process evidence for the question-behavior effect: Influencing socially normative behaviors

Exploring the question-behaviour effect: randomized controlled trial of motivational and question-behaviour interventions

Why is changing health-related behaviour so difficult?

Most People Think They Are More Pro-Environmental than Others: A Demonstration of the Better-than-Average Effect in Perceived Pro-Environmental Behavioral Engagement

A mere measurement effect for anticipated regret: impacts on cervical screening attendance

The importance of normative beliefs to the self-prophecy effect

Varying cognitive targets and response rates to enhance the question-behaviour effect: An 8-arm Randomized Controlled Trial on influenza vaccination uptake

Cluster-Randomized Studies

A motivational account of the question-behavior effect

The Hawthorne effect in the measurement of adolescent smoking

Harmonizing professional, personal, and social responsibilities: Indian women dentists' perspectives