Portuguese breast cancer study validates trauma symptom scale, finds mixed growth-scale validity

A Portuguese validation study of 122 women with early-stage breast cancer found strong psychometric support for the PTSD Checklist for DSM-5 (PCL-5), while the Posttraumatic Growth Inventory–Short Form (PTGI-SF) was highly reliable but had limited external validity. Participants were recruited at the Champalimaud Clinical Centre as part of the BOUNCE project and completed both questionnaires at 12 and 18 months after diagnosis. Most analyses of factor structure, internal consistency and construct validity used the 12-month assessment, while test–retest reliability and paired score comparisons used data from both assessments.

The results were stronger for the PCL-5. Its four-factor structure, corresponding to the DSM-5 symptom domains of intrusion, avoidance, negative alterations in cognition and mood, and alterations in arousal and reactivity, showed excellent fit in confirmatory factor analysis. The model produced a comparative fit index (CFI) of 1.00, a Tucker–Lewis index of 1.03, an RMSEA of 0.00 and a standardized root mean square residual (SRMR) of 0.08. Factor loadings ranged from 0.45 to 0.89 and were statistically significant. Because a Tucker–Lewis value can exceed 1.00 when a model fits far better than a poorly fitting baseline, the authors interpreted the result as evidence of a robust model rather than misspecification.

Internal consistency was high across all four PCL-5 domains. Cronbach’s alpha and McDonald’s omega were 0.86 for intrusion, 0.83 and 0.84 for avoidance, 0.85 and 0.75 for negative alterations in cognition and mood, and 0.74 and 0.72 for alterations in arousal and reactivity. Corrected item–total correlations were all above 0.30, and removing any item did not meaningfully improve consistency. Test–retest reliability over the six-month interval was generally acceptable: intraclass correlation coefficients (ICCs) were 0.55 for intrusion, 0.65 for negative alterations in cognition and mood, and 0.69 for alterations in arousal and reactivity. Avoidance was the exception, with an ICC of 0.37 and a 95% confidence interval of 0.19 to 0.53, a level the authors characterized as poor.

The PCL-5 also showed the expected pattern of convergent and divergent validity. Total post-traumatic stress symptom severity was positively correlated with negative affect measured by the Positive and Negative Affect Schedule–Short Form (PANAS-SF), with r = 0.67, and negatively correlated with quality of life measured by the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30), with r = −0.51. It was also negatively associated with positive affect, r = −0.43. Each of the four symptom clusters correlated positively with negative affect, negatively with positive affect and negatively with quality of life. These findings support the PCL-5 as a symptom-severity measure in this group, but the study explicitly states that it was not used to establish a clinical diagnosis of PTSD.

Symptoms were generally low in this cohort. On most PCL-5 items, more than 60% of participants selected one of the two lowest response categories, indicating that symptoms were absent or mild for many women. Several items had strongly skewed, highly peaked distributions. The researchers said this pattern may reflect the sample’s early-stage disease, assessments conducted 12 months after diagnosis, eligibility for systemic treatment and possible protective factors such as social support, optimism and meaning-making. These characteristics also limit the findings’ generalizability: the psychometric performance may not extend to people with advanced cancer or more severe post-traumatic stress symptoms, or to cancer populations beyond breast cancer.

The PTGI-SF showed a different psychometric profile. A unidimensional model provided adequate overall fit, with a CFI of 0.99, a Tucker–Lewis index of 0.98, an RMSEA of 0.07 and an SRMR of 0.09. All item loadings were statistically significant and ranged from 0.60 to 0.83. The result indicates that the 10 items formed a coherent general measure of perceived growth in this sample, even though the instrument was designed around five domains: appreciation of life, new possibilities, personal strength, relating to others and spiritual change. The researchers specified the unidimensional model as the primary solution and did not directly compare it with the competing multidimensional models, partly because stable estimation of five two-item factors would require a substantially larger sample.

Reliability of the PTGI-SF was excellent. Its internal consistency was α = 0.91 and ω = 0.91, and all corrected item–total correlations exceeded 0.30. Removing an item did not improve either coefficient. The measure also showed high test–retest stability, with an ICC of 0.91 and a 95% confidence interval of 0.86 to 0.94 between 12 and 18 months. In paired analyses, PTGI-SF scores did not change significantly over the same interval: the mean was 23.99 at 12 months and 23.03 at 18 months, a small effect of d = 0.13, p = 0.184. This stability does not mean that growth was identical for every patient or that it applies to all cancer trajectories; it describes the average scores in the 102 participants with complete paired data for this measure.

External validity was the PTGI-SF’s main weakness. Its total score was not significantly correlated with quality of life, r = 0.06, p = 0.53, or positive affect, r = 0.08, p = 0.38. It was also unrelated to negative affect, r = −0.01, p = 0.95. The authors offered several possible explanations: the sample had limited variability, the short form may lose sensitivity by reducing five domains to two items each, and a single-factor solution may obscure specific growth experiences. They also cautioned that perceived growth need not translate directly into better current well-being, because oncology-related quality of life also reflects physical functioning, symptom burden and treatment effects. The pattern is compatible with the idea that growth can involve meaning-making or reinterpretation of the cancer experience rather than serving as a simple marker of lower distress. It does not prove that the scale captures illusory growth; rather, its expected external associations were not demonstrated in this sample.

Neither scale changed significantly between 12 and 18 months. For the PCL-5, the mean total score was 13.45 at 12 months and 13.40 at 18 months among 93 participants with complete paired data; the Wilcoxon signed-rank test gave Z = −0.79, p = 0.432, with a small effect size of r = 0.08. The PTGI-SF comparison included 102 paired participants. These descriptive results support relative stability of the average scores in this cohort, but they do not establish a uniform course of symptoms or growth for individual patients. The study did not directly test the relationship between PCL-5 and PTGI-SF scores, and its design cannot determine whether distress and growth co-evolve or influence one another. Its sample was also limited to Portuguese women aged 40 to 70 with non-metastatic stage I–III breast cancer, most of whom were employed and married. Larger, more clinically diverse studies, including patients with higher symptom severity and follow-up beyond 18 months, are needed to clarify the scales’ performance and whether the two constructs change in related ways.

Subject of Research: Psychometric validation of post-traumatic stress and posttraumatic growth measures in women with breast cancer

Article Title: Trauma and posttraumatic growth in cancer: validation of the post-traumatic stress disorder checklist for DSM-5 and the posttraumatic growth inventory-short form in patients with breast cancer

Article References: Castanheira, C., Almeida, S., Manica, I., Abreu, A., Frasquilho, D., Sousa, B., Silva-Nunes, M. V., Oliveira-Maia, A. J., & Lemos, R. (2026). Trauma and posttraumatic growth in cancer: validation of the post-traumatic stress disorder checklist for DSM-5 and the posttraumatic growth inventory-short form in patients with breast cancer. Frontiers in Psychiatry, 17, Article 1855703. https://doi.org/10.3389/fpsyt.2026.1855703

Image Credits: AI Generated

DOI: 10.3389/fpsyt.2026.1855703

Keywords: breast cancer, psychometric validation, PCL-5, PTGI-SF, post-traumatic stress, posttraumatic growth, psycho-oncology, patient-reported outcomes, confirmatory factor analysis, Portuguese women, Trauma, posttraumatic