Andy P. Field

A systematic review of psychological factors as predictors of chronicity/disability in prospective cohorts of low back pain.

Study Design. A systematic review of prospective cohort studies in low back pain. Objectives. To evaluate the evidence implicating psychological factors in the development of chronicity in low back pain. Summary of Background Data. The biopsychosocial model is gaining acceptance in low back pain, and has provided a basis for screening measurements, guidelines and interventions; however, to date, the unique contribution of psychological factors in the transition from an acute presentation to chronicity has not been rigorously assessed. Methods. A systematic literature search was followed by the application of three sets of criteria to each study: methodologic quality, quality of measurement of psychological factors, and quality of statistical analysis. Two reviewers blindly coded each study, followed by independent assessment by a statistician. Studies were divided into three environments: primary care settings, pain clinics, and workplace. Results. Twenty-five publications (18 cohorts) included psychological factors at baseline. Six of these met acceptability criteria for methodology, psychological measurement, and statistical analysis. Increased risk of chronicity (persisting symptoms and/or disability) from psychological distress/depressive mood and, to a lesser extent, somatization emerged as the main findings. Acceptable evidence generally was not found for other psychological factors, although weak support emerged for the role of catastrophizing as a coping strategy. Conclusion. Psychological factors (notably distress, depressive mood, and somatization) are implicated in the transition to chronic low back pain. The development and testing of clinical interventions specifically targeting these factors is indicated. In view of the importance attributed to other psychological factors (particularly coping strategies and fear avoidance) there is a need to clarify their role in back-related disability through rigorous prospective studies.

How to do a meta-analysis.

Meta-analysis is a statistical tool for estimating the mean and variance of underlying population effects from a collection of empirical studies addressing ostensibly the same research question. Meta-analysis has become an increasing popular and valuable tool in psychological research, and major review articles typically employ these methods. This article describes the process of conducting meta-analysis: selecting articles, developing inclusion criteria, calculating effect sizes, conducting the actual analysis (including information on how to do the analysis on popular computer packages such as IBM SPSS and R) and estimating the effects of publication bias. Guidance is also given on how to write up a meta-analysis.

A meta-analysis of risk factors for post-traumatic stress disorder in children and adolescents

Post-traumatic stress disorder (PTSD) is a complex and chronic disorder that causes substantial distress and interferes with social and educational functioning. Consequently, identifying the risk factors that make a child more likely to experience traumatic distress is of academic, clinical and social importance. This meta-analysis estimated the population effect sizes of 25 potential risk factors for PTSD in children and adolescents aged 6-18 years across 64 studies (N=32,238). Medium to large effect sizes were shown for many factors relating to subjective experience of the event and post-trauma variables (low social support, peri-trauma fear, perceived life threat, social withdrawal, comorbid psychological problem, poor family functioning, distraction, PTSD at time 1, and thought suppression); whereas pre-trauma variables and more objective measures of the assumed severity of the event generated small to medium effect sizes. This indicates that subjective peri-trauma factors and post-event factors are likely to have a major role in determining whether a child develops PTSD following exposure to a traumatic event. Such factors could potentially be assessed following a potentially traumatic event in order to screen for those most vulnerable to developing PTSD and target treatment efforts accordingly. The findings support the cognitive model of PTSD as a way of understanding its development and guiding interventions to reduce symptoms.

Memory for emotionally neutral information in posttraumatic stress disorder : A meta-analytic investigation

Studies have come to conflicting conclusions about whether posttraumatic stress disorder (PTSD) is associated with poorer memory for emotionally neutral information. The authors report a meta-analysis of 27 studies that investigated verbal and/or visual memory in samples with PTSD and healthy controls. The results indicated that the association between PTSD and memory impairment appears to be robust, small to moderate in size, and stronger for verbal than for visual memory. Effect sizes did not vary according to whether recall was immediate or delayed. The association is found in both civilian and military samples and cannot be readily explained as being due to the use of nontraumatized healthy control groups or concurrent head injury. The findings are placed in the context of recent neurobiological and experimental cognitive research.

Fear information and the development of fears during childhood: effects on implicit fear responses and behavioural avoidance

Field, Argyris and Knowles (Behav Res Ther 39 (2001) 1259), and Field, Hamilton, Knowles and Plews (Behav Res Thera 41 (2003) 113) have developed a prospective paradigm for testing Rachmans (Behav Res Ther 15 (1977) 375) proposition that fear information is important in the development of fears and phobias in children. Despite this paradigm being an advance on retrospective reports, the research so far has been restricted to self-reported fear beliefs measured after the information is given. This gives rise to two possible shortcomings: (1) the effects could simply reflect demand characteristics resulting from children conforming to the experimental demands, and (2) although fear information changes beliefs, this might not translate into the behavioural change that would be expected if this information has a powerful effect relevant to the development of pathological fear. This paper describes an experiment that attempts to address these concerns by improving Field et al.s (2001, 2003) basic paradigm but with the addition of two measures: (1) a behavioural measure of avoidance, and (2) an implicit attitude task that should not be susceptible to deliberate attempts to conform to experimental demands. The result showed that negative and positive information have dramatic, and opposite, effects on self-reported fear beliefs, behavioural avoidance and implicit attitudes. There were no effects of gender on any of these results. This study fully supports Rachmans model and suggests that past work does not merely reflect demand characteristics and that fear information increases behavioural avoidance as well as fear beliefs.

The Problems in Using Fixed-Effects Models of Meta-Analysis on Real-World Data

There are 2 approaches to meta-analysis: One assumes that studies in a meta-analysis are sampled from populations with the same effect size (the fixed-effects case), the other assumes that studies are taken from populations that have varying effect sizes (the random-effects case). As such, 2 corresponding meta-analytic frameworks have been developed: fixed- and random-effects methods. Recent evidence suggests that the assumption of fixed population effect sizes is not tenable for virtually all real-world data (e.g., Hunter & Schmidt, 2000), and yet fixed-effects methods of meta-analysis are routinely applied to real-world data (see National Research Council, 1992). This article describes some of the problems in using fixed-effects models on random-effects data by presenting 2 Monte Carlo simulations. In keeping with statistical theory (e.g., Hunter & Schmidt, 2000) results show a radical inflation of the significance tests of the mean effect sizes (above and beyond theoretical predictions). These results ...

Who's afraid of the big bad wolf: a prospective paradigm to test Rachman's indirect pathways in children

Rachmans theory [The conditioning theory of fear insition: a critical examination. Behav. Res. Ther. 15 (1977) 375-387] of fear acquisition suggests that fears and phobias can be acquired through three pathways: direct conditioning, vicarious learning and information/instruction. Although retrospective studies have provided some evidence for these pathways in the development of phobias during childhood [see King, Gullone, & Ollendick, Etiology of childhood phobias: current status of Rachmans three pathways theory. Behav. Res. Ther. 36 (1998) 297-309 for a review], these studies have relied on long-term past memories of adult phobics or their parents. The current study was aimed towards developing a paradigm in which the plausibility of Rachmans indirect pathways could be investigated prospectively. In Experiment 1, children aged between 7 and 9 were presented with two types of information about novel stimuli (two monsters): video information and verbal information in the form of a story. Fear-related beliefs about the monsters changed significantly as a result of verbal information but not video information. Having established an operational paradigm, Experiment 2 looked at whether the source of verbal information had an effect on changes in fear-beliefs. Using the same paradigm, information about the monsters was provided by either a teacher, an adult stranger or a peer, or no information was given. Again, verbal information significantly changed fear-beliefs, but only when the information came from an adult. The role of information in the acquisition of fear and maintenance of avoidant behaviour is discussed with reference to modern conditioning theories of fear acquisition.

Distorted cognition and pathological anxiety in children and adolescents

The present article provides a review on the role of distorted cognition in the pathogenesis of childhood anxiety problems. A comprehensive model of information processing that can be employed for discussing various types of anxiety-related cognitive distortions is presented. Evidence for the occurrence of these cognitive distortions in anxious children and adolescents is summarised. Then, the origins of cognitive distortions in anxious children and adolescents are addressed with reference to genetic-based vulnerability, environmental influence, and gene–environment interactions. Finally, the article provides a critical discussion of the developmental aspects of cognitive distortions, their precise role in the aetiology and maintenance of childhood anxiety disorders, and their relevance for the treatment of this type of psychopathology. Throughout the article many leads are given that may guide future research in this area.

Is the meta-analysis of correlation coefficients accurate when population correlations vary?

One conceptualization of meta-analysis is that studies within the meta-analysis are sampled from populations with mean effect sizes that vary (random-effects models). The consequences of not applying such models and the comparison of different methods have been hotly debated. A Monte Carlo study compared the efficacy of Hedges and Veveas random-effects methods of meta-analysis with Hunter and Schmidts, over a wide range of conditions, as the variability in population correlations increases. (a) The Hunter-Schmidt method produced estimates of the average correlation with the least error, although estimates from both methods were very accurate; (b) confidence intervals from Hunter and Schmidts method were always slightly too narrow but became more accurate than those from Hedges and Veveas method as the number of studies included in the meta-analysis, the size of the true correlation, and the variability of correlations increased; and (c) the study weights did not explain the differences between the methods.

Redefine Statistical Significance

We propose to change the default P-value threshold for statistical significance from 0.05 to 0.005 for claims of new discoveries.