Examining the Factor Structure of an Adapted Posttraumatic Growth Inventory in a Sample of Childhood Cancer Survivors: A Brief Report

Abstract

Posttraumatic growth (PTG) represents positive changes following a trauma, crisis, and/or psychologically distressing event. Experiencing cancer can serve as a traumatic event for patients, resulting in life changes among survivors. Various PTG measures have been used to assess post-cancer change among childhood cancer survivors (CCS), but few have been evaluated for use in this population. This study examined the factor structure of an adapted, 11-item version of the Posttraumatic Growth Inventory (PTGI) among CCS. A randomly selected subgroup of participants (N = 332) was selected from the Los Angeles Cancer Surveillance Program (mean age of 26.5 years at time of survey, mean age at diagnosis of 12 years, primarily male [53.6%], and Hispanic [51.5%]). Participants indicated the degree to which they experienced positive, negative, or no change in their life because of their cancer experience. An exploratory factor analysis (EFA) identified two factors: Appreciation of New Possibilities and Spiritual Change. The adapted, 11-item PTGI was deemed appropriate for use among CCS. Additional research is needed to confirm the use of the two-factor model with confirmatory factor analysis in an independent sample. Future research on PTG among CCS can consider spiritual change as a potential independent factor.

Keywords

post-traumatic growth childhood cancer survivor psychometrics validation exploratory factor analysis

Introduction

Posttraumatic growth (PTG) includes positive life changes following a trauma, major crisis, and/or psychologically distressing event (Tedeschi & Calhoun, 2004). Traumatic events are associated with changes in self, changes in relationships with others, and/or changes in life philosophies (Tedeschi & Calhoun, 1996). For example, the diagnosis and treatment of cancer can serve as a traumatic event for patients (Cordova et al., 2017), which may result in positive or negative changes in quality of life (Arpawong et al., 2013). Cancer survivorship includes unique challenges such as follow-up care, financial burden, changes in functional abilities, as well as further physical and psychosocial obstacles (Arpawong et al., 2013). This is particularly true among childhood cancer survivors (CCS), for which few evidence-based interventions to address disparities and barriers to survivorship care exist (Mobley et al., 2021).

Past research, both qualitative and quantative, has sought to characterize PTG among CCS. A review characterizing correlates of PTG among childhood and adolescent cancer survivors found that PTG was negatively related to time from diagnosis and time from treatment completion, and was positively related to age at diagnosis, age at survey, post-traumatic stress symptoms, and social support (Turner et al., 2018). Optimism was also found to be positively related to PTG (Turner et al., 2018). Similarly, in another study, older age, shorter time from diagnosis, younger age at diagnosis, and female gender were found to be correlates of PTG among CCS (Tremolada et al., 2016). Given its multidimensional nature, characterizing PTG after cancer is important to ensure that appropriate psychological intervention is offered to cancer survivors.

The Posttraumatic Growth Inventory (PTGI) is a 21-item questionnaire designed to measure change after trauma and has been used to assess post-cancer change among CCS; however, it has not been validated for use in this population. The PTGI has been shown to have normally distributed scores, good internal consistency (α = 0.90), and acceptable test-retest reliability. Prior psychometric validation among university students revealed five meaningful factors: Relating to Others (α = 0.85), Spiritual Change (α = 0.85), New Possibilities (α = 0.84), Personal Strength (α = 0.72), and Appreciation of Life (α = 0.67) (Tedeschi & Calhoun, 1996). To decrease participant response burden in the present study, approximately two high-loading items per factor (Tedeschi & Calhoun, 1996) were selected for use, resulting in an adapted PTGI of 11 items. These 11 items have been successfully used in prior research among both clinical and non-clinical populations in the targeted age range (Milam, 2004; Milam et al., 2004). Notably, the original PTGI primarily measures positive change (Tedeschi & Calhoun, 1996). However, given the multidimensional nature of PTG, not allowing respondents to report change on a scale ranging from negative to positive change hinders the full characterization of PTG after cancer. As such, the 11-item PTGI was modified to measure negative, none, or positive change after experiencing cancer, allowing for a more comprehensive examination of CCS’ cancer experiences.

Present Study

Understanding the psychometric properties of the adapted PTGI in a sample of CCS is necessary to reliably measure and understand cancer-related PTG to ensure that optimal supportive care is delivered to survivors. The purpose of the present study is to utilize exploratory factor analysis (EFA) to examine the factor structure of an adapted, 11-item version of the PTGI stemming from a cancer experience among CCS (Yi et al., 2015). It is hypothesized that multiple latent factor constructs will emerge as predictors for the PTG items.

Methods

Procedures

The Project Forward Cohort is a population-based, cancer-registry derived, study of risk and protective factors associated with cancer-related follow-up care in young adult CCS (Milam et al., 2021). Data on all cases were obtained from the Los Angeles Cancer Surveillance Program (LA CSP), the cancer registry for LA County (part of the Surveillance, Epidemiology, and End Results [SEER] program). Eligible participants included CCS who were diagnosed up to 19 years of age between 1996–2010 in LA County with any cancer type (stage 2 or greater, except for brain and melanoma, which included stage 1 or greater), who were at least 5 years from diagnosis, and 18–39 years of age at time of study launch in 2015. Patients who were treated less than 2 years prior to the study were excluded, except for those with chronic myeloid leukemia due to the routine use of protracted maintenance therapy with tyrosine kinase inhibitors.

Introductory postcards and self-report surveys were mailed in English or Spanish (upon request), with the option to complete the survey online, over the phone, or in person in either language. Reminder mailings and calls occurred for those who did not respond. Participants received $20 cash and a lottery entry ($300) as incentives to participate. Procedures were approved by the California State Committee for the Protection of Human Subjects, the Institutional Review Board (IRB) at the University of Southern California (USC), and the California Cancer Registry.

Measures

Demographics: Age at diagnosis, race/ethnicity, age at survey, cancer type, and sex were obtained from the LA CSP. Current health insurance status was self-reported. Posttraumatic Growth: Items for use in the present study were selected based on previously determined high factor loadings (Tedeschi & Calhoun, 1996). Participants were asked to indicate the degree to which they have experienced changes because of their cancer experience. Minor changes in phrasing were made to allow for the endorsement of negative change, no change, or positive change for each item: (1) “Priorities about what is important in my life;” (2) “Appreciation for the value of my own life;” (3) “Involvement in things that interest me;” (4) “Direction for my life;” (5) “My compassion for other;” (6) “Willingness to express my emotions;” (7) “Handling my difficulties;” (8) “My religious faith;” (9) “My own inner strength;” (10) “My understanding of spiritual matters;” and (11) “My sense of closeness with others.” Response options were a 5-point Likert-type scale including: (1) “Highly Negative Change;” (2) “Somewhat Negative Change;” (3) “No Change;” (4) “Somewhat Positive Change;” and (5) “Highly Positive Change.”

Statistical Analysis

Demographic and clinic variables from the cancer registry were examined to determine differences between responders and non-responders. No differences between responders and non-responders in age at diagnosis, years since diagnosis, age, cancer diagnosis, or stage of disease were found. However, survivors who responded were more likely to be female (vs. male), non-Hispanic White, and have higher (vs. lower) socioeconomic status (SES). Additional details are described elsewhere (Milam et al., 2021). Sample: The Project Forward Cohort analytical sample included 1106 CCS. From these, 30% (N = 332) were randomly selected for EFA (Hoe, 2008). The remaining 70% (N = 774) were reserved for future confirmatory factor analysis. All analyses were conducted using SAS statistical software (Version 9.4; SAS Institute Inc). Descriptives: Descriptive statistics were utilized to examine sample demographics and frequencies of item responses. Exploratory factor analysis: The principal factor method was used for initial extraction to calculate linear combinations of the original variables to explain as much of the total variance as possible. An initial oblique rotation with principal factors was performed to allow each factor to account for the maximum variance that had not been accounted for and would not be correlated with the prior factors (Osborne, 2015). The initial analysis found that the two factors were relatively correlated (0.458). Therefore, an oblique rotation was performed as it allows correlated factors, while an orthogonal rotation assumes factors are independent from one another (Osborne, 2015).

Meaningful factors: Evaluation of a scree plot confirmed the use of two factors (Yong & Pearce, 2013). Eigenvalues were used to understand the amount of variance explained by each factor. In accordance with the eigenvalue-one criterion (also referred to as the Guttman-Kaiser criterion), an item with an eigenvalue greater than or equal to one was retained (Yeomans & Golder, 1982). The eigenvalues confirmed the use of two factors, as two items had an eigenvalue of greater than one (retained: items 1 [eigenvalue: 5.29] and 2 [1.22] did not retain items: 3 [0.84], 4 [0.71], 5 [0.63], 6 [0.55], 7 [0.47], 8 [0.38], 9 [0.35], 10 [0.30], and 11 [0.28]). Therefore, we determined to proceed with the two selected factors.

Results

Demographics: Participants (N = 332) had a mean age at time of survey of 26.45 years (SD ± 5.03), a mean age at diagnosis of 11.99 years (SD ± 5.12), were primarily male (53.61%), Hispanic (51.51%), and privately insured (61.73%). The most prominent cancer type was acute lymphoblastic leukemia (36.35%). Adapted Posttraumatic Growth Inventory item responses: The most positively endorsed item was: “Appreciation for the value of my own life” representing ‘somewhat positive change’ (M = 4.37, SD ± 0.86). The most negatively endorsed item was: “My religious faith” representing ‘no change’ to ‘somewhat positive change’ (M = 3.52, SD ± 1.15). Item response means ranged from 3.52 to 4.37, indicating that, on average, participants endorsed positive changes attributed to their cancer experience. Exploratory factor analysis: Results of the 11-item adapted PTGI questionnaire underwent EFA using the principal factor method to extract factors.

Proportion of variance accounted for: The analysis was constrained to have the cumulative proportion of common variance equal to 1.0 after final factor extraction. Evaluation of the proportion of all factors extracted revealed no negative values, therefore, mathematically, the communality estimates can be considered relatively accurate estimates of the actual communality possessed by the variables. In further exploration of the two previously identified factors, we examined the proportion of common variance accounted for by each factor. Factors accounting for 10% or greater of the common variance were retained. As such, Factor 1, accounting for 48% of the common variance, and Factor 2, accounting for 11% of the common variance, were retained.

Interpretability: Acceptable factor loading was defined as 0.4 or greater (Velicer et al., 1982). Factor 1 had eight items with significant loadings and Factor 2 had three items with significant loadings (Raubenheimer, 2004). The variables that load on each factor were determined to share a conceptual meaning, as identified by a content expert (JEM). Each item displayed high factor loadings on one factor (defined as greater than 0.4) and low factor loadings on the other factor (defined as less than 0.4; i.e., no cross-loading) (Velicer et al., 1982).

Oblique rotation: The final two-factor rotated pattern with standardized regression coefficients is displayed in Table 1. The items represent linear combinations of each factor and regression “weights” are associated with each of the factors comprised of the PTG items. This rotated factor pattern provides the simple structure and was used to contextually determine the meaning of each factor. Only meaningful loadings (defined as 0.4 or greater) (Velicer et al., 1982) were considered and cross-loading was not present in this factor pattern, indicating that the items are clean measures of the construct and all items can be retained. No items had high loadings on more than one factor, therefore, no items were dropped.

Table 1.

Final Two-Factor Model With Oblique Rotation and Equations.

Item		Standardized Regression Coefficients		Two-factor Equations
		Factor 1	Factor 2
1	Priorities about what is important in my life	0.88		$YA = 0.87506 F 1 - 0.12154 F 2 + E$
2	Appreciation for the value of my own life	0.76		$YB = 0.76158 F 1 + 0.01813 F 2 + E$
3	Involvement in things that interest me	0.81		$YC = 0.80968 F 1 - 0.01597 F 2 + E$
4	Direction for my life	0.79		$YD = 0.79452 F 1 + 0.02344 F 2 + E$
5	My compassion for others	0.43		$YE = 0.43319 F 1 + 0.29108 F 2 + E$
6	Willingness to express my emotions		0.47	$YF = 0.35571 F 1 + 0.47498 F 2 + E$
7	Handling my difficulties	0.56		$YG = 0.56488 F 1 + 0.26227 F 2 + E$
8	My religious faith		0.92	$YH = - 0.10505 F 1 + 0.91815 F 2 + E$
9	My own inner strength	0.69		$YI = 0.69248 F 1 + 0.08669 F 2 + E$
10	My understanding of spiritual matters		0.87	$YJ = 0.00551 F 1 + 0.86506 F 2 + E$
11	My sense of closeness with others	0.41		$YK = 0.41118 F 1 + 0.38156 F 2 + E$

Note. The correlation between the two factors is 0.458.

Factors: Of the 11 items in the adapted PTGI measure, eight items loaded on Factor 1 and three items loaded on Factor 2. Table 1 displays the 11-item, two-factor model with oblique rotation (factors correlating at 0.458) and the final two-factor equations. The final communality estimates produced a value of 6.51 and the 11 items displayed good internal consistency (α=0.89) with acceptable values of alpha (α=0.87 [Factor 1] and α=0.74 [Factor 2]) (Nunnally & Bernstein, 1994). After consultation with a context expert (JEM), the two factors were labeled: Factor 1, Appreciation of New Possibilities, and Factor 2, Spiritual Change.

Discussion

The present study sought to utilize EFA to examine the factor structure of an adapted, 11-item version of the PTGI stemming from a cancer experience among CCS. These data suggest the adapted PTG assessment is appropriate for use in a CCS sample, representing two PTG factors: Appreciation of New Possibilities and Spiritual Change.

The factors identified in the present study are partially comparable to those identified in the original measure validation among general, non-cancer affected university students (Tedeschi & Calhoun, 1996). Two of the five previously identified factors, Appreciation of Life and New Possibilities, have item-overlap with Factor 1 in the present study. Further, the previously identified Spiritual Change factor was also found in the present study, representing Factor 2. Therefore, despite a decrease in items and subsequently a decrease in factors, these findings suggest similarities between the original 21-item measure and the adapted 11-item measure.

The identification of two distinct factors indicates independent value should be attributed to each factor. Appreciation of New Possibilities reflects a core PTG construct. As such, future research concerning PTG should consider separating the spiritual change items to be unique enough to examine independently, in relation to their topic of interest, among CCS. Factor structure similarities between the original 21-item PTGI and the adapted 11-item measure, in addition to the internal consistency among the 11-items, suggest that the adapted PTGI is applicable for use in this population.

These findings further contribute to those of prior research supporting the option to report none or negative change as a result of ones’ cancer experience (Milam, 2004; Milam et al., 2004). It is possible that prior research using the PTGI may conflate low change with negative change due to lacking item response options. As such, the response format used in the present study reflects the broader, multidimensional nature of PTG and likely yields a more comprehensive characterization of PTG after cancer. However, given limited research modifying the PTGI, the addition of reporting negative change is worthy of further exploration. For example, Taku et al., 2021 has assessed negative change as posttraumatic depreciation (Taku et al., 2021).

The present study has several strengths and some limitations. These findings contribute to the limited research surrounding PTG among CCS and provide a novel and brief approach to measuring PTG by including the response option of negative change. Confirmatory factor analysis is to be conducted which will enable future researchers to verify the use of the two-factor model in an independent sample. Further, prior research indicates PTG differs by gender, cancer type, and age (Boyle et al., 2017; Liu et al., 2021), and therefore should be considered in future examinations of factor structure. These findings may inform future qualitative studies aimed to better understand individual cancer survivorship experiences (e.g., researchers using a narrative approach to characterize life-histories that reflect PTG elements, such as appreciation of new possibilities and spiritual changes stemming from the cancer experience) (Tremolada et al., 2018).

Conclusion

Results suggest that an adapted, 11-item PTGI is well-suited for use among CCS. Providing a response option that includes negative change may facilitate clinicians’ abilities to identify patients who require additional support. Thus, this adapted PTGI yields a rapid screening of PTG that can be used to detect both positive and negative change attributed to the cancer experience among CCS. These findings also suggest the potential for spiritual change to be considered as an independent factor.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grant 1R01MD007801 from the National Institute on Minority Health and Health Disparities of the National Institutes of Health and P30CA014089 from the National Cancer Institute of the National Institutes of Health. Author JT was supported by the VA Office of Academic Affiliations through the Advanced Fellowship Program in Health Services Research and Development as well as the T32CA009492 from the National Cancer Institute of the National Institutes of Health.

Disclaimer

The contents do not represent the views of the U.S. Department of Veterans Affairs or the United States Government.

ORCID iD

Julia Stal

References

Arpawong

T. E.

Richeimer

S. H.

Weinstein

Elghamrawy

Milam

J. E.

(2013). Posttraumatic growth, quality of life, and treatment symptoms among cancer chemotherapy outpatients. Journal of Health Psychology, 32(4), 397–408. https://doi.org/10.1037/a0028223

Boyle

C. C.

Stanton

A. L.

Ganz

P. A.

Bower

J. E.

(2017). Posttraumatic growth in breast cancer survivors: Does age matter? Psychooncology, 26(6), 800–807. https://doi.org/10.1002/pon.4091

Cordova

M. J.

Riba

M. B.

Spiegel

(2017). Post-traumatic stress disorder and cancer. Lancet Psychiatry, 4(4), 330–338. https://doi.org/10.1016/S2215-0366(17)30014-7

Hoe

S. L.

(2008). Issues and procedures in adopting structural equation modeling technique. Journal of Applied Quantitative Methods, 3(1), 76–83. https://jaqm.ro/issues/volume-3,issue-1/pdfs/hoe.pdf

Liu

Thong

M. S. Y.

Doege

Koch-Gallenkamp

Bertram

Eberle

Holleczek

Waldmann

Zeissig

S. R.

Pritzkuleit

Brenner

Arndt

(2021). Prevalence of benefit finding and posttraumatic growth in long-term cancer survivors: Results from a multi-regional population-based survey in Germany. British Journal of Cancer, 125(6), 877–883. https://doi.org/10.1038/s41416-021-01473-z

Milam

(2004). Posttraumatic growth among HIV/AIDS patients. Journal of Applied Social Psychology, 34(11), 2353–2376. https://doi.org/10.1111/j.1559-1816.2004.tb01981.x

Milam

Freyer

D. R.

Miller

K. A.

Tobin

Wojcik

K. Y.

Ramirez

C. N.

Ritt-Olson

Thomas

S. M.

Baezconde-Garbanati

Cousineau

Modjeski

Gupta

Hamilton

A. S.

(2021). Project forward: A population-based cohort among young adult survivors of childhood cancers. JNCI Cancer Spectr, 5(5), pkab068. https://doi.org/10.1093/jncics/pkab068

Milam

Ritt-Olson

Unger

(2004). Postraumatic growth among adolescents. Journal of Adolescent Research, 19(2), 192–204. https://doi.org/10.1177/0743558403258273

Mobley

E. M.

Moke

D. J.

Milam

Ochoa

C. Y.

Stal

Osazuwa

Bolshakova

Kemp

Dinalo

J. E.

Motala

Baluyot

Hempel

(2021). Disparities and barriers to pediatric cancer survivorship care. https://www.ncbi.nlm.nih.gov/pubmed/33724748

10.

Nunnally

Bernstein

(1994). Psychometric theory. McGraw-Hill Higher, INC.

11.

Osborne

J. W.

(2015). What is rotating in exploratory factor analysis? Practical Assessment, Research, and Evaluation, 20(2). https://doi.org/10.7275/hb2g-m060.

12.

Raubenheimer

(2004). An item selection procedure to maximise scale reliability and validity. SA Journal of Industrial Psychology, 30(4), 59–64. https://doi.org/10.4102/sajip.v30i4.168

13.

Taku

Tedeschi

R. G.

Shakespeare-Finch

Krosch

David

Kehl

Romeo

Di Tella

Kamibeppu

Soejima

Hiraki

Volgin

Dhakal

Zieba

Ramos

Nunes

Leal

Gouveia

Calhoun

L. G.

(2021). Posttraumatic growth (PTG) and posttraumatic depreciation (PTD) across ten countries: Global validation of the PTG-PTD theoretical model. Personality and Individual Differences, 169, 110222. https://doi.org/10.1016/j.paid.2020.110222.

14.

Tedeschi

R. G.

Calhoun

L. G.

(1996). The posttraumatic growth inventory: Measuring the positive legacy of trauma. Journal of Traumatic Stress, 9(3), 455–471. https://doi.org/10.1007/BF02103658

15.

Tedeschi

R. G.

Calhoun

L. G.

(2004). Posttraumatic growth: Conceptual foundations and empirical evidence. Psychological Inquiry, 15(1), 1–18. https://doi.org/10.1207/s15327965pli1501_01

16.

Tremolada

Bonichini

Basso

Pillon

(2016). Post-traumatic stress symptoms and post-traumatic growth in 223 childhood cancer survivors: Predictive risk factors. Front Psychol, 7, 287. https://doi.org/10.3389/fpsyg.2016.00287.

17.

Tremolada

Bonichini

Basso

Pillon

(2018). Adolescent and young adult cancer survivors narrate their stories: Predictive model of their personal growth and their follow-up acceptance. European Journal of Oncology Nursing: the Official Journal of European Oncology Nursing Society, 36, 119–128. https://doi.org/10.1016/j.ejon.2018.09.001.

18.

Turner

J. K.

Hutchinson

Wilson

(2018). Correlates of post-traumatic growth following childhood and adolescent cancer: A systematic review and meta-analysis. Psychooncology, 27(4), 1100–1109. https://doi.org/10.1002/pon.4577

19.

Velicer

W. F.

Peacock

A. C.

Jackson

D. N.

(1982). A comparison of component and factor patterns: A Monte Carlo approach. Multivariate Behavioral Research, 17(3), 371–388. https://doi.org/10.1207/s15327906mbr1703_5

20.

Yeomans

K. A.

Golder

P. A.

(1982). The Guttman-Kaiser criterion as a predictor of the number of common factors. The Statistician, 31(3), 221–229. https://doi.org/10.2307/2987988

21.

Zebrack

Kim

M. A.

Cousino

(2015). Posttraumatic growth outcomes and their correlates among young adult survivors of childhood cancer. Journal of Pediatric Psychology, 40(9), 981–991. https://doi.org/10.1093/jpepsy/jsv075

22.

Yong

A. G.

Pearce

(2013). A beginner’s guide to factor analysis: Focusing on exploratory factor analysis. Tutorials in Quantitative Methods for Psychology, 9(2), 79–94. https://doi.org/10.20982/tqmp.09.2.p079