Do maladaptive beliefs delay whiplash associated disorders (WAD): A systematic review

The purpose of the study is to try to establish if maladaptive beliefs effect recovery times and poor outcomes in whiplash associated disorders (WAD). In May 2017 the following databases were searched from their inception until June 2017: SPORT Discuss, CINAHL, PsycINFO, MEDLINE, Ovid MEDLINE, Cochrane, AMED, Embase. A combination of sensitive search strategies was used for locating articles on maladaptive beliefs and WAD. Hand-searching of relevant journals and citation tracking were used to maximise the identified study pool. A total of 189 references were retrieved and an additional three studies were identified through different sources, 178 remained after the removal of duplicates. For 43 references, the full text was assessed, and 7 studies were included. The methodological quality was assessed independently by two assessors. Data extraction was carried out using a standardised Erika Gabriela Macias1, Vasileios Georgopoulos2, Alan Taylor3 Affiliations: 1Graduate student, School of Health Sciences, Faculty of Medicine and Health Sciences, The University of Nottingham, University Park, NG7 2RD, Nottingham, UK; 2Doctoral Researcher, School of Medicine, Division of Academic Rheumatology, The University of Nottingham, University Park, NG7 2RD, Nottingham, UK; 3Assistant Professor, School of Health Sciences, Faculty of Medicine and Health Sciences, The University of Nottingham, University Park, NG7 2RD, Nottingham, UK. Corresponding Author: Erika Gabriela Macias, Andador de la Yuca 88, Unidad Ponciano Arriaga, Soledad de Graciano Sánchez, San Luis Potosí, San Luis Potosi, Mexico, 78437; Email: erika.gabriela.macias@gmail.com Received: 27 January 2018 Accepted: 27 April 2018 Published: 17 May 2018 data extraction form. Most articles scored a high overall quality and fourteen percent (14%) of articles (1 out of 7) were rated with moderate overall quality. Meta-analysis was not undertaken due to the heterogeneity of prognostic factors, outcome measures and methods used. Four out of the seven studies presented a correlation between catastrophising and disability in at least one follow-up time point (3, 6 or 12 months) whilst three studies found a correlation between fear-avoidance and disability. Four of the studies showed an association between maladaptive beliefs (catastrophising or fear avoidance) and pain and two found a negative effect. Our findings show that outcomes, such as pain and disability, were found to be associated with maladaptive beliefs (catastrophising and fear avoidance).


INTRODUCTION
Whiplash injury due to a road traffic collision is considered common as it composes 75% of all road www.edoriumjournals.com/ej/drMacias et al. 2 traffic collision injuries [1].The incidence of whiplash associated disorders (WAD) has been increasing over the past 30 years [2].WAD has become an important burden to healthcare systems, insurance organisations and economies of industrialised countries [3].In the UK, whiplash effects the economy with an added cost of approximately £3 billion per annum [4].The Quebec Task Force (QTF) defines the term "whiplash" as an acceleration-deceleration mechanism that results in injury to the neck [5].This may lead to soft tissue damage, hence resulting in clinical manifestations such as neck pain.The term "whiplash associated disorder" is given to the symptoms presented following a whiplash-related injury and to differentiate it from the accelerationdeceleration mechanism of injury [5].Around 83% of individuals, who are affected by a motor vehicle collision (MVC), develop WAD [6].Recovering from WAD can be an extended process that involves significant use of health resources.
Challenges persist in management, prediction of prognosis and functional recovery for patients with WAD.Many patients recover within a normal time frame (less than three months), although a considerable proportion (40-60%) develop chronic and disabling symptoms [7].Additionally, there appear to be no meaningful changes in the recovery of pain and disability beyond three months following the injury [8].A challenge that clinicians often encounter is the difficulty to predict the prognosis of functional recovery for patients with WAD.
If the prognosis appears to be favourable, clinicians have the option of selecting less-intensive interventions, such as advice and education [9].On the contrary, if the prognosis seems to be unfavourable, the clinicians should have the ability to recognise and opt for specific, more targeted interventions [9].
The most consistent risk factors for poor recovery are initial elevated levels of pain and disability [1,10].Contemporary studies have included psychological factors as prognostic factors.These include pain catastrophising, presence and symptoms of depression, anxiety, fear avoidance and posttraumatic stress symptoms [10,11].Low expectations of recovery have also been found to predict outcome in WAD [1,12].Maladaptive beliefs, such as fear avoidance beliefs and pain catastrophising, are described as unhelpful evaluative cognitions related to the self in the domain of interpersonal interactions [13].Screening for maladaptive beliefs incorporates key elements of the biopsychosocial model into clinical practice [14,15].The biopsychosocial model attributes injury outcomes to the complex interaction of biological, psychological and social factors [16].Literature tends to agree that psychological factors such as post-traumatic stress symptoms, pain catastrophising and fear avoidance are common in patients with WAD [17].
Interventions related to cognitive behavioural therapy, propose that beliefs influence pain perception and adjustment to pain.Therefore, it is hypothesised that individuals who present maladaptive beliefs about pain have a higher risk of poor outcomes and impediment to their recovery [18].Poor treatment outcomes include maintained distress, pain and disability [19,20].
The aim of this review was to establish if maladaptive beliefs affect recovery times and poor outcomes in WAD and to discuss whether these results have a clinical application.

Data sources and searches
The identification of relevant studies was performed through a systematic literature search of SPORTDiscuss, CINAHL, PsycINFO, MEDLINE, Cochrane, AMED and Embase, from database inception until June 2017.The synthesis of maladaptive belief-sensitive searches adapted for this study has been used in the past by similar systematic reviews [8,17].However, considering the lack of an accepted and standardised search strategy relevant to WAD, a robust search scheme was needed.The Cochrane Back Group (CBG) [21] proposed a search method for WAD studies.The suggested strategy by the CBG was incorporated by adding recommended keywords to establish the final search plan (Appendix 1).
The search included various terms identified in the literature for maladaptive beliefs (eg, catastophising, catastrophization, maladaptive beliefs, fear of movement) with different combinations and subject headings.Additionally, the reviewers screened bibliographies of all included studies and retrieved review articles by hand to assure the completeness of the literature search and possibly add appropriate references not retrieved by the systematic search.All potential related articles, were included in the full-text review.The inclusion and exclusion criteria also applied to these studies.
A total of 189 references were retrieved, additionally 3 more studies were identified through different sources and 178 remained after the removal of duplicates.For 43 references, the full text was assessed, and 7 studies were included.

Inclusion criteria
Articles were included if they satisfied the following criteria; Adults with WAD, a prospective study design that provided a baseline measure of at least one maladaptive belief variable, such as pain, disability, depression and/ or anxiety and were published as full papers in English or with a translation available.No limitations were applied on the type or predictor variables (history, physical examination, imaging, etc.) nor on the clinical setting and the type of patients with WAD.
All relevant literature collected from various databases were downloaded into EndNote -Web were duplicates were removed.www.edoriumjournals.com/ej/drMacias et al. 3 One independent reviewer was assigned to execute the first and second stage of the screening process.The firststage consisted of the evaluation of titles and abstracts based on the eligibility criteria stated above.All studies allocated for inclusion to the systematic review proceeded to the second-stage along with the studies that were identified through citation tracking or hand searching.

Exclusion criteria
Studies were excluded if they presented the following criteria; Not a primary research study, conference abstracts, PhD thesis, unpublished dissertations and books, did not include at least one maladaptive belief measure, did not present pain and/or disability as an outcome measure and if neck pain was not related with a whiplash injury.Following independent data extraction, the completed forms were compared by two independent reviewers.Disagreement between examiners was resolved by consensus or if needed, by a third party.

Data extraction and quality assessment
Similarly, the quality assessment for the 7 eligible studies was also carried out by 2 reviewers who assessed the studies independently using a quality assessment tool based on clearly defined methodological standards for Randomised control trials (RCTS) and cohort studies.The Quality in Prognosis tool (QUIPS) [22] was used to assess cohort studies whilst the PEDro scale [23] was used to assess the RCT for this review.

Data synthesis
Studies included for this review were expected to manifest elevated levels of heterogeneity, therefore no statistical pooling was attempted for result analysis.This article was composed under the guidelines of the Preferred reporting items of systematic reviews and meta-analysis (PRISMA) [24].

Maladaptive beliefs
Maladaptive beliefs associated with outcomes were extracted from all 6 cohort studies and 1 RCT.A correlation value was considered statistically significant if the reported p-value was less than 0.05, if the article reported that the association was significant, or if the 95% confidence intervals around a rate ratio or similar statistic did not cross 1.Where a maladaptive belief was assessed with respect to the outcome at several timepoints in one cohort, data was extracted for a short-term follow-up (less than 6 months) and for long term followup (the longest follow-up point greater than or equal to six months).

Study selection
An overview of the search results is presented in Figure 1.Overall, 189 articles were identified.Three additional studies were collected and added to the study pool via hand searching of scientific journals and citation tracking.After duplicates were removed, 178 studies were screened through title and abstract assessment.43 records were found eligible to progress to the next screening stage where the full-text copies were retrieved for careful examination.Following the application of eligibility criteria and detailed evaluation, 7 studies were included in this review

Characteristics of included studies
The seven studies involved 1 RCT and 6 cohort studies.Three of these studies originated from Australia [25][26][27], one from Switzerland [28], one from Denmark [29], one from Canada [30] and one from the Netherlands [31].The RCT by Jull [25] aimed to identify whether multidisciplinary treatments for individuals with acute whiplash had an ability to reduce the incidence of chronicity.Although this study did not specifically try to examine an association between maladaptive beliefs and disability, the information was provided.The study by Angst [28] focused on finding factors that predicted pain relief, improved physical function and improved working capacity by examining the association between catastrophising and pain.A limitation of the study is the patients' high drop-out rate at baseline as this may indicate attrition bias.Andersen [29] researched paincatastrophising and fear-avoidance beliefs as mediators to pain as well as post-traumatic stress symptoms by assessing the association between maladaptive beliefs (catastrophising, fear avoidance) and outcome.This study was presented with a short-follow up period and an elevated percentage of individual withdrawal.This should be contemplated when analysing the outcome of the sample size.The cohort study by Buithenhuis [31] inspected the role of catastrophising and causal beliefs with severe persisting neck disorders after a MVC.To assess individual's causal beliefs of post-traumatic neck complaints, the authors used the newly developed Causal Beliefs Questionnaire for Whiplash (CBQ-W).However, the validity and reliability of the CBQ-W has not been formally established.This should be considered when interpreting the results due to the exclusive development of the tool for the purposes of that study.The cohort study by Pedler [27] attempted to examine a model for the maintenance of pain and disability in WAD patients, including PTSD, sensory hypersensitivity, and fear avoidance model factors.The study also assessed the association between catastrophising, kinesiophobia and disability.Bostick [30] studied the predictive capacity of pain beliefs and catastrophising in WAD.Participants were invited to complete maladaptive belief and disability measures at baseline as well as at 3 and 6 months postinjury.Subsequently, using multiple linear regression modelling, baseline belief and catastrophising scores were examined for their association with prospective pain and disability.Findings of the study, must be interpreted with discretion due to the relatively small sample size (n=72).Additionally, Casey [26] intended to determine recovery trajectories based on disability, pain catastrophising and mental health.This study also attempted to investigate linkages between these trajectories, finding a correlation between catastrophising and disability.However, previous history such as trauma, comorbidities or preexisting physical and mental health were not considered for this study.Appendix 2 includes specific study details.

Qualitative appraisal of included studies
Each study was independently examined by two reviewers.The 6 cohort studies were assessed using the Quality in Prognosis Studies (QUIPS) tool [32].The QUIPS tool has an interrater agreement (κ statistic) range from 0.56 to 0.82 (median, 0.75).The RCT in this study was assessed using the PEDro scale [23].The reliability of the PEDro score is described as "fair" to "good", ICC for the total score was 0.56 (95% CI = 0.47 to 0.65).
For this review, there were only four episodes of disagreement in quality assessment, which was resolved by discussion.No disagreements were referred to a third party for settlement.An overall quality score was assigned to each article.Although most articles scored a high overall quality, fourteen percent (14%) of articles (1 out of 7) were rated as a moderate overall quality.The score was due to a lack of adequate description regarding the study participation (place of recruitment not specified), study attrition (differences between participants who completed the study and those who did not) and clear specification of outcome measures.Table 1 displays the quality appraisal scoring for each study.

Pain beliefs
Pain beliefs for catastrophising and fear of movement were measured with various tools.The two most consistent tools used where the pain catastrophising scale (PCS) and the pictorial fear of activities scale -cervical spine (PFActS-C).
The assessment of pain catastrophising for the studies presented in this systematic review were obtained using PCS and the Coping Strategies Questionnaire (CSQ) [33].The PCS assesses three domains specific for catastrophising (helplessness, rumination and magnification), while the CSQ assesses helplessness and pessimism in the context of pain.The PCS may be considered a wider assessment tool of catastrophising than CSQ.The PCS has been previously found to correlate with outcomes such as pain intensity in conditions such as WAD [34,35].Additionally, PCS has been found to have adequate psychometric properties, with acceptable internal consistency (Cronbach's α= 0.66-0.87)[34] and good temporal stability (Pearson's r²=0.92) in both acute and chronic pain populations [36].The PCS was used in 4 of the 7 studies [26,[29][30][31].

Pain and disability
The outcome measures for pain and disability were the Neck Disability Index (NDI), Functional Rating Index (FRI), Whiplash Disability Questionnaire (WDQ), Visual Analogue Scale (VAS), 11-point numerical rating scale, the Short Form-36 (SF-36) and the Posttraumatic Stress Diagnostic Scale (PDS).Three studies measured both pain and disability [25,27,30] whilst two measured only disability [26,31] and two others only assessed pain [28,29].Three out of seven articles measured disability with the NDI.The NDI has been used extensively in patients with WAD and it has been shown to be reliable, valid and with high internal consistency [38].One cohort study [26] used the FRI to measure disability.The FRI has shown to have excellent responsiveness, reliability and validity and internal consistency (Cronbach's α=0.92) [39].The WDQ was also used to assess disability in other studies [21].It has been shown to have psychometric properties that demonstrate high internal consistency (Cronbach's α=0.96), excellent reproducibility, short-term test-retest reliability and receptiveness in a physiotherapy setting for individuals with acute and chronic WAD [40,41].The VAS for pain [95% CI = 0.96 to 0.98] [42] was used by two studies [25,27], the 11-point numerical rating scale [43] was also used by two studies for its high test-retest reliability (r = 0.96 and 0.95) [29,30] and one used the SF-36 (Cronbach's α = 0.59) [44].The PDS was only used by one study [26].The alpha rating for PDS has been shown to be 0.92 with a test-retest reliability of (r=-0.74)[45].

SUMMARY OF EVIDENCE
Based on the evidence retrieved for this review, only two maladaptive beliefs were reported, fear-avoidance beliefs and catastrophising.Various baseline maladaptive belief measurements were assessed with outcome measurements such as pain and disability, at long-term follow-up (6 to 12 months) and one at short term followup (less than 6 months).Results have been summarised in Table 2.
There were only four studies [25,26,30,31] that found an association between fear-avoidance and disability (Table 3).These studies agreed that higher initial scores for beliefs had an impact on initial and follow-up disability scores.The study by Buitenhuis [31] presented a multiple logistic regression model, using the persistence of postwhiplash syndrome at 6 and 12 months as a dependent variable.Findings demonstrated that causal beliefs regarding whiplash (CBQ-W) can contribute to concurrent disability (NDI) at a 6 (β=3.00,p<0.001) and 12-month follow-up (β=2.39,p=0.002).In accordance to these findings, the RCT by Jull [25] introduced a univariate logistic regression analysis of recovery demonstrating that fear of movement (PFActS-C) had asignificant impact (β=2.54,p<0.001) at 6 months (β=2.54,p<0.001) at 12 months.Bostick [30] carried out a multiple linear regression analysis to examine the capacity of baseline beliefs to predict 3 and 6 months pain intensity and selfreported disability.Baseline patient beliefs (PBPI) about pain were demonstrated to be significantly correlated to disability at 3 months (β=0.18,p<0.05) and 6 months (β=0.14, p<0.05).
In the prospective cohort study by Angst [28] findings showed that reduction of catastrophising were associated with improvements in three dependent variables (pain, function, working capacity).Catastrophising was associated to disability and a greater display of pain behaviour.Catastrophising (CSQ) at baseline presented an association with pain (NASS) at a 6-month followup (β=0.348,p=0.039).However, the study uses two outcome measures that are not specific for pain, which challenges the validity of the results, due to inadequate pain evaluation.The SF-36 measures physical function, social function and comprehensive bodily pain whilst the NASS questionnaire assess physical function, neurogenic symptoms and cervical spine-specific pain.The interactive effects of pain with kinesiophobia (TSK) (β=0.03,p=0.259), and pain with fear of activities (PFActS-C) (β=-0.03,p=0.643) was non-significant according to results found by Pedler et al., (2016).Conversely, the study by Casey [26] demonstrated an association between baseline catastrophising (PCS) and pain (FRI) (β=0.04,p≤0.01).

DISCUSSION
Over the last decade there has been a significant increase in physiotherapy studies related to maladaptive beliefs.This movement in healthcare displays an eagerness for advanced clinical decision-making and the acquisition to cost-effective medical management.
Models such as the fear-avoidance model were originally designed for LBP [46].Those models have attempted to analyse precise beliefs that impact LBPrelated outcomes [46].Pain beliefs related to WAD are less well understood [47].
To the author's knowledge, there is no other review that systematically attempted to appraise and establish the effects of maladaptive beliefs on WAD recovery.Up until now, studies addressing maladaptive beliefs have primarily focused on WAD management and the use of cognitive behavioural therapy.However, there is an existing systematic review that has attempted to assess the prognostic importance of catastrophising as a coping strategy in patients with low back pain (LBP) [48].Results found in this systematic review [48] supports the findings of this study.In the review [48] catastrophising was also found to be associated with disability and pain after follow-up for patients presenting LBP in an acute, subacute and chronic stage.Another study [49] published in 2014, reviewed fear avoidance beliefs as a prognostic factor for patients with nonspecific LBP.This systematic review, assessed fear avoidance with the Fear Avoidance Beliefs Questionnaire (FABQ) and the Tampa Scale of Kinesiophobia.The research included in this current analysis have been published after 2006, similar to our studies.The findings found in this systematic review [49], are in line with the findings of our study that suggests that fear avoidance beliefs may be an indicator of a poor prognosis for patients with subacute LBP.
Additionally, another systematic review by Williamson [17] examined the effect of psychological factors and the development of late whiplash syndrome.Although this review considered maladaptive beliefs (catastrophising and fear avoidance) as psychological factors, these variables were not considered primary predictors of outcome.Additionally, maladaptive belief barriers were mediated by pain severity [17].Nonetheless, it is important to also consider additional nonsystematic reviews that focus on the development of chronic pain due to the influence of psychological factors [50,51].
The evidence presented in this review suggests that beliefs relevant to causation, fear-avoidance and selfefficacy correlate with WAD pain and disability.The fear avoidance model leads to the development of maladaptive beliefs (e.g.catastrophising) and such beliefs have the capacity to negatively influence pain and disability.While the hypotheses of these correlations require clarification, there is existing evidence that suggests there is an effect on cognitive factors as well.

STUDY LIMITATIONS
Meta-analysis was impossible due to the heterogeneity of outcome measures and the various maladaptive belief variables.Results from meta-analysis may have demonstrated more precise results.

CONCLUSIONS
This study is the first to systematically identify and evaluate the correlation of maladaptive beliefs to recovery in WAD.The data provided in this systematic review, demonstrates a moderate to strong association between maladaptive beliefs (catastrophising and fear avoidance) and pain and disability.Although, there is a compelling amount of research studying maladaptive beliefs, the approach for evaluating and reporting correlations between maladaptive beliefs and outcome measures is identified as an obstacle for the interpretation of relevant literature.There is a need to compare baseline characteristics of study populations who present maladaptive beliefs in comparison with those who do not and identify plausible reasons why maladaptive beliefs are associated with pain and disability.Even though these findings suggest a plausible association between

A
standardised tool for data extraction was used by two independent reviewers.The data extracted was (a) setting, (b) method of diagnosis, (c) sample selection (d) study population: number of participants and patient characteristics (age, gender, body mass index), (e) study design, (f) duration of study (g) Pain and disability measurements (h) maladaptive belief measure.

Figure 1 :
Figure 1: PRISMA flow diagram for search results and study selection.

Table 2 :
Study Cohorts displaying an association between maladaptive beliefs and treatment outcomes

Table 3 :
Identified association between maladaptive beliefs and disability

Table 5 :
Identified association between maladaptive beliefs and pain www.edoriumjournals.com/ej/drMacias et al. 11 pain, disability, catastrophising and fear-avoidance, we are unable to fully establish if maladaptive beliefs are important in delaying WAD outcome and further research is required.Whiplash Injuries.mp.humans www.edoriumjournals.com/ej/drMacias et al. 14