The role of attentional biases in the context of cancer: A systematic review and meta‐analysis

Beyond the modified dot-probe task: A meta-analysis of the efficacy of alternate attention bias modification tasks across domains

Attention biases towards disease-relevant cues have been implicated in numerous disorders and health conditions, such as anxiety, cancer, drug-use disorders, and chronic pain. Attention bias modification (ABM) has shown that changing attention biases can change related emotional processes. ABM most commonly uses a modified dot-probe task, which has received increasing criticism regarding its reliability and inconsistent findings. The purpose of the present review was thus to systematically review and meta-analyse alternative tasks used in ABM research. We sought to examine whether alternative tasks significantly changed attention biases and emotional outcomes, and critically examined whether relevant sample, task and intervention characteristics moderated each of these effect sizes. Seventy-four (completer n = 15,294) study level comparisons were included in the meta-analysis. Overall, alternative ABM designs had a medium effect on changing biases (g = 0.488), and a small, but significant effect on improving clinical outcomes (g = 0.117). We found this effect to be significantly larger for studies which successfully changed biases compared to those that did not. Across all tasks, it appeared that targeting engagement biases results in the largest change to attention biases. Importantly, we found tasks incorporating gaze-contingency - encouraging engagement with non-biased stimuli - show the most promise for improving emotional outcomes.

The role of interpretation biases and symptom burden in fear of cancer recurrence/progression among ovarian cancer survivors

BACKGROUND

Models of fear of cancer recurrence or progression (FCR/P) suggest that the way in which people interpret ambiguous physical symptoms is an important contributor to the development and maintenance of FCR/P, but research has not investigated this claim. The aim of this study is to fill that gap.

METHODS

This was a cross-sectional study. Sixty-two women with ovarian cancer reported completed measures of FCR/P, an interpretation bias task and a symptom checklist. The healthy control group (n = 96) completed the interpretation bias task.

RESULTS

Women with ovarian cancer were more likely to interpret ambiguous words as health-related compared to healthy women (p < 0.001; Cohen's d = 1.28). In women with cancer, FCR/P was associated with overall symptom burden (r = 0.25; p = 0.04) and interpretation bias score (r = 0.41; p = 0.001), but interpretation bias and symptom burden were not related (r = 0.22; p = 0.09). Interpretation bias did not moderate the relationship between symptoms and FCR/P.

CONCLUSIONS

We found that women with ovarian cancer interpreted ambiguous words as health related more often compared to women without cancer, and this bias was greater for women with higher FCR/P. Symptom burden was also associated with FCR/P. However, interpretation bias did not moderate the relationship between physical symptoms and FCR/P. Hence, the central tenet of the Cancer Threat Interpretation model was not supported in women with ovarian cancer.