Elise J Devlin1, Hayley S Whitford2, Linley A Denson3, Andrew E Potter4. 1. School of Psychology, Faculty of Health and Medical Sciences, The University of Adelaide, South Australia, Australia. Electronic address: elise.devlin@adelaide.edu.au. 2. School of Psychology, Faculty of Health and Medical Sciences, The University of Adelaide, South Australia, Australia. Electronic address: hayley.whitford@adelaide.edu.au. 3. School of Psychology, Faculty of Health and Medical Sciences, The University of Adelaide, South Australia, Australia. Electronic address: linley.denson@adelaide.edu.au. 4. School of Psychology, Faculty of Health and Medical Sciences, The University of Adelaide, South Australia, Australia; GenesisCare Radiation Oncology, Adelaide, South Australia, Australia. Electronic address: Andrew.Potter@genesiscare.com.
Abstract
OBJECTIVE: Response expectancies of cancer treatment toxicities are often, but not always, associated with subsequent experiences. A recent meta-analysis indicated that response expectancies, measured using different assessment formats, reveal different effect sizes, potentially explaining mixed outcomes. Utilizing a clinical sample, we compared 5-point assessments and visual analogue scales, as measures of response expectancies for the incidence and severity of subsequent toxicities. METHODS: Four weeks pre-radiotherapy, 45 men with prostate cancer rated their response expectancies of the same 18 toxicities on 5-point assessments and visual analogue scales, presented in random order. Descriptors anchored each end of visual analogue scales and every point of 5-point assessments was labelled, including an 'unsure' midpoint. Toxicities were subsequently assessed 2-weeks into radiotherapy on 100-point visual analogue scales. RESULTS: Across all toxicities, 17.5-62.8% of patients selected 'unsure' on 5-point assessments. No response expectancies were reported on 5-point assessments for 'blood in stools' or 'rectal urgency' yet 54.8%-64.3% of patients indicated response expectancies for these toxicities on visual analogue scales. Visual analogue scales and 5-point scales demonstrated small-to-moderate associations (r = 0.30-0.58) as measures of response expectancy incidence, but mostly large associations when visual analogue scales captured severity (r = 0.43-0.76). Response expectancies measured with visual analogue scales predicted more toxicities to a moderate degree or greater (68.8%) than 5-point assessments (37.5%). CONCLUSION: This novel investigation demonstrated an 'unsure' midpoint is often selected, potentially reducing the sensitivity of 5-point assessments. Based on their associations, and outcomes, these assessment formats should be considered independent in response expectancy research of cancer treatment toxicities.
OBJECTIVE: Response expectancies of cancer treatment toxicities are often, but not always, associated with subsequent experiences. A recent meta-analysis indicated that response expectancies, measured using different assessment formats, reveal different effect sizes, potentially explaining mixed outcomes. Utilizing a clinical sample, we compared 5-point assessments and visual analogue scales, as measures of response expectancies for the incidence and severity of subsequent toxicities. METHODS: Four weeks pre-radiotherapy, 45 men with prostate cancer rated their response expectancies of the same 18 toxicities on 5-point assessments and visual analogue scales, presented in random order. Descriptors anchored each end of visual analogue scales and every point of 5-point assessments was labelled, including an 'unsure' midpoint. Toxicities were subsequently assessed 2-weeks into radiotherapy on 100-point visual analogue scales. RESULTS: Across all toxicities, 17.5-62.8% of patients selected 'unsure' on 5-point assessments. No response expectancies were reported on 5-point assessments for 'blood in stools' or 'rectal urgency' yet 54.8%-64.3% of patients indicated response expectancies for these toxicities on visual analogue scales. Visual analogue scales and 5-point scales demonstrated small-to-moderate associations (r = 0.30-0.58) as measures of response expectancy incidence, but mostly large associations when visual analogue scales captured severity (r = 0.43-0.76). Response expectancies measured with visual analogue scales predicted more toxicities to a moderate degree or greater (68.8%) than 5-point assessments (37.5%). CONCLUSION: This novel investigation demonstrated an 'unsure' midpoint is often selected, potentially reducing the sensitivity of 5-point assessments. Based on their associations, and outcomes, these assessment formats should be considered independent in response expectancy research of cancer treatment toxicities.