Interval-valued TESS Study
Research type
Research Study
Full title
Exploring Interpretation of Function Using an Interval-Based Toronto Extremity Salvage Score (TESS)
IRAS ID
208043
Contact name
Christian Wagner
Contact email
Sponsor organisation
University of Nottingham
Duration of Study in the UK
1 years, 10 months, 30 days
Research summary
Research Summary
Soft tissue sarcomas are a rare malignancy of connective tissue. The Toronto Extremity Salvage Score (TESS) was developed to monitor the effects of therapeutic interventions (surgery, chemotherapy and/or radiotherapy) on function in patients with soft tissue sarcomas. The TESS consists of a self-completed questionnaire that allows participants to pick a selection of precise responses representing their perceptions about the extent of difficulty to perform daily activities. As such, each of the response levels are not necessarily equidistant and interpretations may vary significantly both amongst and within medical professionals and patients.
Our pilot study demonstrated that there was significant difference in the interpretation of each of the descriptors used in the TESS between surgeons, physiotherapists and patients. It therefore brings into question the use of linguistic descriptors (such as ‘moderately difficult’) in assessment of function. We have therefore developed an alternative questionnaire where participants are asked to draw an ellipse around the appropriate extent of the scale. The position of the ellipse on the scale indicates the difficulty level, while its width indicates the level of uncertainty in their response.
Primary Research Objectives:
1. To explore the use of an interval-valued scale compared to the traditional likert scale used currently in the TESS
2. To develop a tablet based app for completion of the interval-valued scale TESS
Identification and Recruitment:
Patients who have a new diagnosis of an extremity soft tissue sarcoma will be considered for inclusion in this study at Nottingham City Hospital.
Data collection:
Patients currently complete the traditional TESS on a paper copy as part of their routine assessment at the pre-treatment stage of their pathway, 6 and 12 months post treatment while patients are attending their routine clinic appointments. Patients will also be asked to complete the interval-based version of the TESS on a tablet based app.Summary of Results
The research reported in this work showed the potential of using Computational Intelligence methods on human perception representation through intervals and assessment through a series of explorations. In particular, a measure on Fuzzy Sets (FSs) called the Agreement Ratio was proposed allowing to provide a ratio of overall agreement when modelling Linguistic Terms (LTs) from interval valued questionnaires by means of the computational technique called Interval Agreement Approach (IAA). The measure relies on using the Jaccard Similarity Measure for comparing different levels of agreement and an extension of this measure is also presented considering the case of General Type-2 Fuzzy Sets, that is, IAA generated FSs modelling two types of uncertainty (e.g., inter and intra source). This measure is proposed as a means to enable an objective assessment on the extent of consensus of key candidate LTs during questionnaire design such that less ``ambiguous'' LTs can be identified. Furthermore, in order to address our research question, we conducted a series of comparisons of several Computational Intelligence (CI) approaches for IV data aggregation and evaluated as means for assessing data from IV questionnaires. In particular, these comparisons considered the Choquet Fuzzy Integral & data-driven Fuzzy Measures and an extension of standard Likert scale based questionnaires assessment using Interval Arithmetic.Based on the results from comparisons performed, a methodology for construct measurement through FSs using IV responses has been proposed. This method allows a weighted aggregation of responses as commonly done when it is known a priori that the effect of each question on the model construct might vary. Moreover, as a final part of the process of developing a construct measurement tool, we describe how to rate the construct using different outcome spaces related to different groups of stakeholders (accounting for subjectivity) using a FS derived measure. That is, categorise the FS valued construct as similarly done after completion of a standard questionnaire and thus, provide a qualitative understanding of what constitutes different levels of response.
Two specific case studies have been considered in this research in order to demonstrate their potential and applicability within Quality of Life (QoL) assessment and academic abilities. Analysis results on validity and inter-rater reliability in the case of QoL assessment of the presented approach are presented using well-established metrics from Social Sciences, highlighting the potential value of using IV questionnaires and CI tools as psychometric tools to provide reasonable and consistent ways to summarise and represent human perceptions, attitudes, etcetera.
REC name
Wales REC 6
REC reference
16/WA/0391
Date of REC Opinion
5 Dec 2016
REC opinion
Favourable Opinion