Background: Current clinical care may not address behavioural and psychosocial elements which can influence quality of life (QoL) and recurrence risk of people living with and beyond cancer (PLWBC). There is a lack of validated tools to assess diet, lifestyle and mental health in PLWBC. We have developed a screener to identify individuals who may need further clinical support beyond cancer recurrence.
Objectives: To validate the new screener in PLWBC.
Methods: PLWBC were recruited at a local cancer patient organisation (AECC-Balears) and at the University Hospital Son Espases. Face validity (n=12, 75% female, average age 51.9 y, 50% breast cancer) assessed construct interpretation and acquiescence of the screener (closed- and open-ended questions) and time to complete (min). For convergent validity, participants (n=105, 63.5% female, average age 57.8 y, 39% breast cancer) completed the screener together with validated diet, lifestyle, and mental health questionnaires, as well as body composition (bioimpedance), physical activity (PA) and sleep (accelerometer), strength (dynamometer, 30-s chair test) and cortisol levels (ELISA) assessment. All participants answered the screener 7-10 days later for reliability analysis (Pearson correlation, ongoing). Agreement between answers obtained from the screener and validated tools is being analysed using a domain-specific approach: item- and score-level associations between the screener and validated tools, and distributional and cut-off-based agreement/mismatch assessments when variables are directly comparable. Classification models (ordinal regression, SVM, C5.0 and Random Forest) were used to predict validated risk categories from individual items and/or domain scores to evaluate predictive performance and identify the most important variables in the models.
Results: The screener is divided into 6 domains: (1) Body composition & sarcopenia; (2) PA, pain and fatigue; (3) Diet and alcohol; (4) Smoking; (5) Sleep; and (6) Mental health. Face validity analysis indicated average time to complete the screener was 10.6 min, was easy to comprehend, and participants did not require help to answer. Suggestions were made regarding the nº of answers per question, ambiguity of certain questions (e.g. what is considered PA), and the inclusion of other factors (e.g. income, employment status) into the questionnaire. 83.3% of participants considered that the content of the questions was relevant to PLWBC. Preliminary agreement analysis between the screener and validated questionnaires indicated the screener has a strong ability to determine PA levels, sleeping behaviour and smoking patterns; moderate ability to identify malnutrition risk and overall dietary behaviours; and low ability to detect sarcopenia risk. Ability to assess intake of individual food groups varies (e.g. high for alcohol and sugary drinks, low for legumes, wholegrains and red meat). Analysis is still ongoing regarding body weight history, pain, fatigue and mental health (including depression, stress and social support). Objective measurements (bioimpedance, accelerometry, strength assessment and cortisol levels) will be integrated into the final analysis.
Conclusions/implications: The validation process will allow the identification of under-performing domains and the refinement of the screener, with the aim to become a validated tool which detects PLWBC's needs for its use in follow-up clinical care plans.