Professional-patient discrepancies in assessing lung cancer radiotherapy symptoms: An international multicentre study

We investigate discrepancies in the assessment of treatment-related symptoms in lung cancer between healthcare professionals and patients, and factors contributing to these discrepancies.

Data from 515 participants in the REQUITE study were analysed. Five symptoms (cough, dyspnoea, bronchopulmonary haemorrhage, chest wall pain, dysphagia) were evaluated both before and after radiotherapy. Agreement between healthcare professionals and people with lung cancer was quantified using Gwet's-AC₂ coefficient. The influence of clinical variables, comorbidities, and quality-of-life outcomes on agreement was examined through stratified analyses.

We found varying levels of agreement between healthcare professionals and people with lung cancer. Bronchopulmonary haemorrhage and dysphagia exhibited very good agreement (meanAC₂ > 0.81), while cough and chest wall pain showed substantial agreement (meanAC₂ = 0.64 and 0.76, respectively). Dyspnoea had the lowest agreement (meanAC₂ = 0.59), with prior chemotherapy significantly reducing agreement levels. Chronic obstructive pulmonary disease (COPD) and early cancer stages also contributed to discrepancies in dyspnoea assessments. Regarding quality-of-life, the most relevant factor was fatigue, which reduced agreement in the assessment of dyspnoea (AC₂ = 0.55 vs 0.70), dysphagia (AC₂ = 0.48 vs 0.69), cough (AC₂ = 0.58 vs 0.82), and chest wall pain (AC₂ = 0.77 vs 0.91).

Our findings indicate strong alignment between healthcare professionals' and people with lung cancer evaluations of observable treatment-related symptoms, but less consistency for subjective symptoms such as dyspnoea. Factors such as prior chemotherapy, COPD, and cancer stage should be considered when interpreting symptom assessments. Furthermore, our study underscores the importance of integrating quality-of-life considerations, particularly fatigue, into symptom evaluations to mitigate potential biases in symptom perception.