Profiling cancer-related distress and problems: A latent class analysis approach.

Abstract

e23191 Background: Psychological distresses, emotional troubles, social difficulties as well as both physical and practical issues are all among oppressive problems that oncological patients have to face almost every day. On one hand, these problems seem to have a strong impact on quality of life; and to the other hand, these issues seem to be related to specific cancer-related variables ( e.g.: type and localization of the tumor). Thus, the aim of the study was to use a Latent Class Analysis (LCA) approach to profile a latent structure accounting for the covariance between psychological distress and everyday problems due to cancer. A one-factor model with three classes was hypothesized, which comprised distress and cancer-related problems as indicators and age, type of medical treatment as well as type and localization of tumor as external variables that moderate the latent structure. Methods: Patients ( N = 264, 62.3% female, mean age = 65.3, SD = 12.4) were enrolled at the Oncology Day Hospital, “Presidio Ospedaliero” of Saronno, ASST Valle Olona, Italy. Using the standardized Distress Thermometer and Problem List (TD&PL; NCCN, 2015) patients were tested for: (A) distress; (B) practical problems (Cronbach’s α = .63); (C) social difficulties (Cronbach’s α = .65); (D) emotional issues (Cronbach’s α = .86) and (E) physical problems (Cronbach’s α = .77). Results: First of all, a CFA was performed to test the original factorial structure of the DT&PL. The original five factor solution was supported by adequate fit indices: RMSEA = .063; CFI = .924. Then, the LCA (10000 bootstrap resampling) shows the goodness-of-model fit [χ2 = 10.01; p = .11; LRχ2 = 9.30; p = .09] and the goodness-of-classification quality [Entropy = .80 ( > .7); Average-Latent-Class-Assignment-Probability: .975, .900, and .902 for Class1, Class2, and Class3, respectively]. The LCA identified a latent variable with three classes (VLMR = 193.38; p < .001; Class1 = .49%; Class2 = .23%, and Class3 = .28%). In addition, interactions with the latent variable were found for age ( β = .12; p = .039), type of medical treatment ( β = .29; p = .009), as well as type and localization of tumor ( β = .20; p = .025). Conclusions: These results provided a better understand of psychological distress and cancer-related issues: each class represents a specific “profile” – moderated by age, type of medical treatment, and type and localization of tumor. These results recommend paying more attention to the specific profile expressed by the patient suggesting new ways to improve their quality of life.