Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=rhpr20

Health Psychology Review

ISSN: 1743-7199 (Print) 1743-7202 (Online) Journal homepage: www.tandfonline.com/journals/rhpr20

The future of health behaviour change interventions: opportunities for open science and personality research

Daryl B. O’Connor

To cite this article: Daryl B. O’Connor (2020) The future of health behaviour change interventions: opportunities for open science and personality research, Health Psychology Review, 14:1, 176-181, DOI: 10.1080/17437199.2019.1707107

To link to this article: https://doi.org/10.1080/17437199.2019.1707107

Published online: 20 Jan 2020.

Submit your article to this journal

Article views: 2433

View related articles

View Crossmark data

Citing articles: 5 View citing articles

COMMENTARY

The future of health behaviour change interventions: opportunities for open science and personality research Daryl B. O’Connor

School of Psychology, University of Leeds, Leeds, UK

ARTICLE HISTORY Received 22 November 2019; Accepted 16 December 2019

Introduction

Understanding self-regulation processes is a central concern for promoting health, wellbeing and longevity. Over many decades scientists have attempted to understand the architecture of self-regu- lation and its associated mechanisms and processes. Moreover, within this context numerous theor- etical models have been developed to facilitate the identification of the key determinants of health behaviours in order to provide suitable targets for intervention. As a result, there has been an explosion of research investigating the effectiveness of health behaviour change interventions. However, what are we to make of these findings? This special issue entitled ‘Understanding and pre- dicting health behaviour change: A contemporary view’ represents an important advance in under- standing the ‘state of the art’ with regards to health behaviour change interventions applied to a range of different behaviours (e.g., chronic disease medications, Wilson et al., 2020; unhealthy risk- taking behaviours, Protogerou, McHugh, & Johnson, 2020), health conditions (e.g., cardiovascular disease, Suls et al., 2020; chronic disease conditions; Hennessy, Johnson, Acabchuk, McCloskey, & Stewart-James, 2020) and contexts (e.g., high versus low social class groups, race/ethnicity, Alcántara et al., 2020; early childhood and adolescence, Miller, Lo, Bauer, & Fredericks, 2020). The special issue ends with a look to the future with examples of how machine learning and natural language proces- sing methods may advance the behaviour change interventions evidence base (Wilson et al., 2020). In this commentary, I outline two critical issues that came to mind when reading this collection of papers. The first relates to the reproducibility of meta-analyses and the principles of Open Science. The second deals with the need to examine the role of personality and individual differences in the context of health behaviour change interventions and self-regulation processes.

The parent meta-review by Hennessy et al. (2020) sets the scene for the special issue overall and for the associated meta-reviews. However, it also raises a number of fundamental issues not only for self-regulation intervention research in the area of chronic disease, but also for health behaviour change research more broadly. Hennessy and colleagues highlight the relatively low quality of reviews in this area (e.g., reviews satisfied less than 50% of the items using the AMSTAR 2 assessment tool) as well as highlighting the inconclusive evidence in terms of which intervention components are consistently important. Similarly, in the sub meta-reviews, Wilson et al. reported variable quality of meta-analyses of interventions targeting self-regulation on adherence to chronic disease medication (AMSTAR 2 item completion – M = 50%; range 31–65%) and Protogerou et al. showed that only four (26.67%) of the meta-analyses included in their meta-review satisfied at least 50% of the AMSTAR 2 quality criteria. Finally, Suls et al. (2020) reported that, on average, the meta-analyses included in their review achieved 56.5% completion of AMSTAR 2 quality items (range 31–88%). Taken together, these findings suggest that there is substantial room for improvement in the conduct and reporting of meta-analytical syntheses of health behaviour change interventions. In addition, they also point to

© 2020 Informa UK Limited, trading as Taylor & Francis Group

CONTACT Daryl B. O’Connor [email protected] Paper accepted by Martin Hagger.

HEALTH PSYCHOLOGY REVIEW 2020, VOL. 14, NO. 1, 176–181 https://doi.org/10.1080/17437199.2019.1707107

concerns about the veracity of the conclusions of meta-analyses in this area. Another important issue relates to the reproducibility of meta-analyses and the extent to which they are conducted in line with the principles of Open Science – an issue I turn to next.

Open Science

Science, and not just psychological science, is undergoing a renaissance. It is an exciting time for psy- chology, and it is great that we, as a discipline, have been leading the way. This renaissance has been prompted by a number of developments: Most notably was the publication of the Open Science Col- laboration (2015) paper estimating the reproducibility of psychological science. This large scale inves- tigation attempted to replicate 100 experimental and correlational studies from three leading journals. The findings were stark; less than 40% of psychology studies were replicated. More recently, another large investigation (the Many Labs 2 project) found that only 14 of 28 classic and contem- porary studies replicated (Klein et al., 2018). In addition, these failures to replicate did not appear to be attributable to sample diversity. Numerous factors have been proposed to explain these low levels of replication including low statistical power, hypothesising after the results are known (HARKING), p-hacking and other questionable research practices (see Munafò et al., 2017 for further discussion). As a consequence, the renaissance in psychology has prompted a new approach to the scientific process known by the umbrella term, Open Science. The aim of Open Science is to increase openness, integrity and reproducibility in scientific research. It is hoped that it will propel psychological researchers forward by improving scientific practice and trigger new ways of working that will ultimately improve the robustness of our evidence base (Norris & O’Connor, 2019).

Health psychology generally, and behaviour change research (including self-regulation interven- tions) specifically, are not immune to issues related to replication and reproducibility. Indeed, Hagger, Peters, Heino, Crutzen, and Johnston (2017) have reported that health psychology has paid relatively limited attention to the issues of replication and reproducibility. However, over the last few years, many health psychology researchers have begun to embrace open science practices, engage in large scale replication efforts and recognise the risks of p-hacking and other questionable research practices. In fact, it is important to note that health psychologists and other behaviour change trialists have been early adopters of a number of important Open Science practices such as pre-registering studies on relevant repositories (e.g., https://clinicaltrials.gov/; https://www.isrctn.com/) as well as pre-registering systematic reviews and meta-analyses (https://www.crd.york.ac.uk/PROSPERO/). Nevertheless, there will be many studies included in the meta-analyses described in this special issue that have not been pre-registered and they may be contaminated by questionable research practices. This is not to undermine the conclusions drawn from the reviews here, but instead it may help to explain some of the inconsistent and variable findings observed in this vast and diver- gent literature base.

Another issue that is worthy of comment relates to the reproducibility of meta-analyses given their central role in meta-reviews. Meta-analyses are the cornerstone of cumulative science. They allow researchers to synthesise evidence across a range of studies of differing sample sizes and outcomes and to draw conclusions about the weight of evidence for the effectiveness of a particular interven- tion or the size of the association between variables of interest while taking into account publication bias. However, relatively recently, concerns have been raised about the reproducibility of meta-ana- lyses (e.g., Gotzsche, Hrobjartsson, Maric, & Kendal, 2007; Lakens, Hilgard, & Staaks, 2016). For example, across 27 meta-analyses, Gotzsche and colleagues attempted to replicate the results of these meta-analyses by independently calculating the standardised mean difference (SMD) from two trials randomly selected from each of the chosen meta-analyses (as well as investigating other data extraction errors). The main findings showed that the authors were unable to replicate at least one of two chosen SMDs in 37% of meta-analyses (N = 10) and other errors were reported in 63% of the meta-analyses (N = 17). Moreover, they concluded that data extraction in meta-analyses

HEALTH PSYCHOLOGY REVIEW 177

is prone to errors that may actually negate the original conclusions or reverse the findings of the study.

What, therefore, can we do to improve the reproducibility of meta-analyses? Prompted by the work led by Gotzsche and colleagues, Lakens et al. (2016) recently published six practical recommen- dations to improve the reproducibility of meta-analyses. In their own words, they argued that there is ‘the need to improve the reproducibility of meta-analyses to facilitate the identification of errors, allow researchers to examine the impact of subjective choices such as inclusion criteria, and update the meta-analysis after several years’ (p. 1). Their six recommendations are summarised briefly below:

(1) Disclose all meta-analytic data (i.e., effect sizes, sample sizes for each condition, test statistics and degrees of freedom etc.),

(2) Facilitate quality control by specifying which effect sizes calculations are used and which assump- tions are made for missing data,

(3) Adhere to established reporting guidelines with the minimum standard being the PRISMA guidelines,

(4) Pre-register the meta-analysis protocol and clearly state confirmatory and exploratory analyses, (5) Facilitate reproducibility by allowing others to re-analyse your data (e.g., provide links to data

files, script, codes etc.), (6) Recruit expertise as required (e.g., consult a librarian about systematic reviewing and/or a statis-

tician before extracting effect size data).

The scientific publishing landscape has changed substantially as a result of the Open Science movement. An important development is the introduction of Registered Reports (https://osf.io/rr/). The aim of this new type of article is to increase the transparency of science, to allow peer review of research studies before the results are known and, crucially, to guarantee acceptance of the paper (irrespective of the findings following review at Stage 1; known as an In Principle Acceptance, IPA). As a consequence, it is hoped this will help reduce questionable research practice while improv- ing the quality of our research protocols, and over time, it is hoped this will ultimately improve the robustness of our evidence base and the reliability and reproducibility of future meta-analyses and meta-reviews. Another area of research that is likely to improve our understanding of the effective- ness of health behaviour change interventions and self-regulation processes relates to personality and individual differences.

What about the role of personality and individual differences in the context of health behaviour change interventions?

Self-regulation processes do not happen in isolation. They happen within individuals who vary in terms of a range of individual differences variables (e.g., personality traits, gender, race/ethnicity) and across different contexts (e.g., high versus low social class, educational settings). Indeed, a myriad of individual differences variables have been identified as important determinants of health behaviours and self-regulation processes. Moreover, these variables are likely to be key mod- erators of behaviour change interventions. Therefore, the inclusion of the Alcántara et al. (2020) syn- thesis on the role of social determinants of health in the context of health behaviour change interventions targeting self-regulation was a welcome addition to this special issue. Interestingly, these authors report that 73.5% of social determinants moderator analyses tested heterogeneity of treatment effects by gender, race/ethnicity, and intervention setting. Of course, there are other important individual differences variables that may directly influence health outcomes and are also likely to influence the effectiveness of health behaviour change interventions. There is a substan- tial body of research that has shown that components of the five-factor model of personality are associated with longevity and health status (e.g., Friedman et al., 1993; Hampson, Goldberg, Vogt,

178 D. B. O’CONNOR

& Dubanoski, 2006, 2013; Hill, Turiano, Hurd, Mroczek, & Roberts, 2011; Jokela et al., in press; Kern & Friedman, 2008; Shipley, Weiss, Der, Taylor, & Deary, 2007; Stephan, Sutin, Luchetti, & Terracciano, 2019). However, I will limit my discussion here to conscientiousness as it has been reliably identified as a determinant of a range of health behaviours and the only ‘Big Five’ factor robustly linked to chronic diseases and mortality across multiple studies (e.g., Bogg & Roberts, 2004; Jokela et al., in press; O’Connor, Conner, Jones, McMillan, & Ferguson, 2009). For example, Friedman et al. (1993), using data from the Terman Life Cycle Study, reported that childhood conscientiousness predicted longevity and the magnitude of this effect (as a risk factor) was comparable to those from elevated serum cholesterol and systolic blood pressure levels in adulthood. In another study spanning forty years, the mechanisms through which childhood personality traits influence health status in adult- hood were assessed (Hampson et al., 2006). Results indicated that conscientiousness influenced health status in adulthood indirectly via educational attainment, healthy eating habits and smoking. Bogg and Roberts (2004) carried out a large meta-analysis of 194 studies, demonstrating that conscientiousness was positively correlated with physical activity and negatively correlated with excessive alcohol use, unhealthy eating, tobacco use, drug use, risky driving, risky sex and suicide. These relationships have also been confirmed in daily diary studies as well as large-scale investigations (Gartland, O’Connor, Lawton, & Ferguson, 2014; Green, O’Connor, Gartland, & Roberts, 2016; Kern & Friedman, 2008; O’Connor et al., 2009). Recent research has also shown positive associations between the facets of conscientiousness and objective markers of health status includ- ing adiposity, blood markers, physical performance and thickness of brain cortical regions (Lewis et al., 2018; Sutin, Stephan, & Terracciano, 2018).

Conscientiousness has been defined as the propensity to follow socially prescribed norms, control impulses and to be goal directed, planful, and able to delay gratification (John & Srivastava, 1999). Each of the latter variables is likely to influence the effectiveness of health behaviour change inter- ventions and self-regulation processes (see Ferguson, 2013). Indeed, Hennessy et al. (2020) identify personalised feedback, goal setting, and self-monitoring as successful intervention components in their meta-review. These key self-regulation interventions align closely with the lower order facets of conscientiousness (industriousness, order, self-control, traditionalism, virtue and responsibility, see Green et al., 2016) and their associated automatic patterns of thoughts, feelings and behaviours. Moreover, one’s ability to control one’s behaviour and to complete tasks is likely to facilitate the per- formance of aversive or difficult health behaviours that individuals may or may not be motivated to perform (O’Connor et al., 2009). Early research by Conner, Rodgers, and Murray (2007) showed con- scientiousness moderated the impact of intentions to exercise on exercise behaviour. Similarly, Rhodes, Courneya, and Jones (2005) reported conscientiousness to significantly moderate the inten- tion–exercise behaviour relationship, with higher levels of conscientiousness associated with stron- ger intention–behaviour relationships. Future research might usefully examine how the facets of conscientiousness (as well as other personality traits) influence self-regulation processes. Do individ- uals high on conscientiousness utilise different strategies to enact their health behaviour intentions? For example, do conscientious individuals formulate clearer plans or simply try harder? In short, are individuals high on conscientiousness better at self-regulation?

Miller et al.’s (2020) ‘Big Picture’ synthesis highlights the importance of considering developmental factors in health-focused self-regulation interventions and reinforces the need to account for devel- opmental stage when delivering behaviour change interventions. For example, they argue that different developmental considerations are required when delivering self-regulation interventions to children and youth compared to adults. Relatedly, it is developmental fact that people tend to become more conscientious as they get older (Roberts, Walton, & Viechtbauer, 2006), therefore, it is likely that the effectiveness of different behaviour change interventions for self-regulation pro- cesses will change over time too. Nevertheless, it is surprising that the moderating effects of key per- sonality factors in health behaviour change intervention research have been relatively underresearched. Indeed, many health-based behaviour change interventions are designed to increase purposeful and planned behaviour (implementation intentions, TPB-based interventions,

HEALTH PSYCHOLOGY REVIEW 179

e.g., O’Connor, Armitage, & Ferguson, 2015) and may be effective by changing trait levels of person- ality, hence the need to assess traits as part of intervention development, delivery and evaluation. In addition, the idea that personality is open to change has led authors such as Roberts, Hill, and Davis (2017) to suggest the intriguing possibility that interventions can be developed to change traits such as conscientiousness that may have important health benefits. Roberts et al. (2017) have recently introduced the Sociogenomic Trait Intervention Model (STIM), an intervention to change conscien- tiousness that is based on behavioural activation theory and is informed by developmental research.

To summarise, understanding, predicting and changing self-regulation processes will continue to be a central concern for health psychology and related disciplines. There is a real opportunity to improve the robustness of the health behaviour change evidence base by continuing to embrace the principles of Open Science together with investigating how individual differences and personality traits interact with these interventions. Future research should adopt more open, transparent and reproducible scientific practices and explore the role of individuals differences variables such as per- sonality in order to provide a fuller understanding of when, where and how self-regulation processes are effective.

Disclosure statement

No potential conflict of interest was reported by the author.

References

Alcántara, C., Diaz, S. V., Giorgio Cosenzo, L., Loucks, E. B., Penedo, F. J., & Williams, N. J. (2020). Social determinants as moderators of health behaviour change interventions: Scientific gaps and opportunities. Health Psychology Review.

Bogg, T., & Roberts, B. W. (2004). Conscientiousness and health-related behaviours: A meta-analysis of the leading behav- ioural contributors to mortality. Psychological Bulletin, 130, 887–919.

Conner, M., Rodgers, W., & Murray, T. (2007). Conscientiousness and the intention–behaviour relationship: Predicting exercise behaviour. Journal of Sport and Exercise Psychology, 29, 518–533.

Ferguson, E. (2013). Personality is of central concern to understand health: Towards a theoretical model for health psy- chology. Health Psychology Review, 7, S32–S70.

Friedman, H. S., Tucker, J. S., Tomlinson-Keasey, C., Schwartz, J. E., Wingard, D. L., & Criqui, M. H. (1993). Does childhood personality predict longevity? Journal of Personality and Social Psychology, 65, 176–185.

Gartland, N., O’Connor, D. B., Lawton, R., & Ferguson, E. (2014). Investigating the effects of conscientiousness on daily stress, affect and physical symptom processes: A daily diary study. British Journal of Health Psychology, 19, 311–328.

Gotzsche, P. C., Hrobjartsson, A., Maric, K., & Kendal, B. (2007). Data extraction errors in meta-analyses that use standar- dised mean differences. JAMA, 298, 430–437.

Green, J., O’Connor, D. B., Gartland, N., & Roberts, B. W. (2016). The Chernyshenko conscientiousness scales: A new facet measure of conscientiousness. Assessment, 23, 374–385.

Hagger, M., Peters, G.-J. Y., Heino, M. T., Crutzen, R., & Johnston, M. (2017). The replication crisis in (health) psychology: Reflections and solutions. The European Health Psychologist, 19(Supp.).

Hampson, S. E., Edmonds, G. W., Goldberg, L. R., Dubanoski, J. P., & Hillier, T. A. (2013). Childhood conscientiousness relates to objectively measured adult physical health four decades later. Health Psychology, 32, 925–928.

Hampson, S. E., Goldberg, L. R., Vogt, T. M., & Dubanoski, J. P. (2006). Forty years on: Teachers’ assessment of children’s per- sonality traits predict traits predict self-reported health behaviours and outcomes at mid life. Health Psychology, 25, 57–64.

Hennessy, E. A., Johnson, B. T., Acabchuk, R. L., McCloskey, K., & Stewart-James, J. (2020). Self-regulation mechanisms in health behaviour change: A systematic meta-review of meta-analyses, 2006-2017. Health Psychology Review.

Hill, P. L., Turiano, N. A., Hurd, M. D., Mroczek, D. K., & Roberts, B. W. (2011). Conscientiousness and longevity: An exam- ination of possible mediators. Health Psychology, 30, 536–541.

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 102–138). New York, NY: Guilford Press.

Jokela, M., Airaksinen, J., Virtanen, M., Batty, G. D., Kivimäki, M., & Hakulinen, C. (in press). Personality, disability-free life years, and life expectancy: Individual participant meta-analysis of 131,195 individuals from 10 cohort studies. Journal of Personality. doi:10.1111/jopy.12513

Kern, M. L., & Friedman, H. S. (2008). Do conscientious individuals live longer? A quantitative review. Health Psychology, 27, 505–512.

180 D. B. O’CONNOR

Klein, R. A., Vianello, M., Hasselman, F., Adams, B. G., Adams, R. B., Alper, S.,… Nosek, B. A. (2018). Many labs 2: Investigating variation in replicability across sample and setting. Advances in Methods and Practice in Psychological Science, 1, 443–490.

Lakens, D., Hilgard, J., & Staaks, J. (2016). On the reproducibility of meta-analyses: Six practical recommendations. BMC Psychology, 4, 24.

Lewis, G. J., Dickie, D. A., Cox, S. R., Karama, S., Evans, A. C., Starr, J. M.,… Deary, I. J. (2018). Widespread associations between trait conscientiousness and thickness of brain cortical regions. Neuroimage, 176, 22–28.

Miller, A. L., Lo, S., Bauer, K. W., & Fredericks, E. M. (2020). Considering developmental factors in health-focused self-regu- lation interventions. Health Psychology Review.

Munafò, M. R., Nosek, B. A., Bishop, D. V., Button, K. S., Chambers, C. D., Du Sert, N. P.,… Ioannidis, J. P. (2017). A manifesto for reproducible science. Nature Human Behaviour, 1(1), 0021.

Norris, E., & O’Connor, D. B. (2019). Science as behaviour: Using a behaviour change approach to increase uptake of Open Science. Psychology and Health, 34, 1397–1406.

O’Connor, D. B., Armitage, C. J., & Ferguson, E. (2015). Randomized test of an implementation intention-based tool to reduce stress-induced eating. Annals of Behavioral Medicine, 49, 331–343.

O’Connor, D. B., Conner, M., Jones, F., McMillan, B., & Ferguson, E. (2009). Exploring the benefits of conscientiousness: An investigation of daily stressors and health behaviours. Annals of Behavioral Medicine, 37, 184–196.

Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716. Protogerou, C., McHugh, R. K., & Johnson, B. T. (2020). Effectiveness of self-regulation interventions to reduce unhealthy

risk-taking behaviours: A meta-review of evidence syntheses. Health Psychology Review. Rhodes, R. E., Courneya, K.

Related Tags