A prospective study of the relationship between illness perception, depression, anxiety, and quality of life in hematopoietic stem cell transplant patients

Abstract Aim The aim of study was to investigate whether depression and anxiety symptoms and illness perception prior to hematopoietic stem cell transplantation (HSCT) predict health related quality of life (HRQOL) at Day 100 and 1 year following HSCT. Methods A total of 205 patients who underwent HSCT (N = 127 autologous transplants, N = 78 allogeneic transplants) were included in this prospective study. Baseline assessment was assessed prior to transplantation and post HSCT data were collected at Day 100 and 1 year. At baseline we assessed depressive symptoms (Patient Health Questionnaire‐9), anxiety symptoms (Generalized Anxiety Disorder‐7), illness perception (Brief Illness Perception Questionnaire), and HRQOL (Functional Assessment of Cancer Therapy‐BMT). Results Patients who expressed a greater level of concern about the severity, course, and ability to exert control over one's illness (i.e., illness perception) and who reported a greater level of depression and anxiety symptoms prior to HSCT reported lower HRQOL at both Day 100 and 1 year posttransplant, with a similar degree of association observed at the two follow‐up time points. Conclusions Our findings suggest that pretransplant perceptions about their illness and negative mood are significant predictors of HRQOL following HSCT. Illness perception, depression, and anxiety are potentially modifiable risk factors for less than optimal outcome after HCSCT and intervention strategies should be explored.


| INTRODUCTION
Hematopoietic stem cell transplantation (HSCT) is a complex medical intervention used to treat a variety of hematological disorders and other malignant conditions. 1,2While HSCT offers effective disease control or potential cure for hematopoietic malignancies, the experience is demanding for the patient. 3,4Patient demands include 3-4 weeks of hospitalization, a prolonged period of isolation and physical recovery posttransplantation, and significant risk from complications. 1 Complications may include chronic graft versus host disease, cardiopulmonary compromise, endocrinopathies, musculoskeletal disorders, and subsequent malignancies. 5,6ollectively, the uncertainty about success and risk of morbidity and mortality is recognized as a significant source of stress for HSCT patients and their caregivers. 3,4,7[10] The concept of illness perception arose out of Leventhal's self-regulation theory 11 and proposes that cognitive representations of illness influence behavioral responses and coping.These cognitive representations include beliefs about cause, severity, course, and ability to exert control over one's own illness.Illness perception have been found to influence health practices and psychological distress in a wide range of illnesses. 12,13o our knowledge, illness perception in HSCT has been investigated in only one prior study.This study found that HSCT recipients who perceived the consequences of their cancer to be more serious experienced more depression, anxiety, and reduced well-being during the year following transplant. 14Whether illness perception impacts the HRQOL of HSCT patients has not been studied.Information about whether illness perception predicts HRQOL following HSCT has implications both for pretransplant psychosocial assessment and for developing intervention approaches to optimize HSCT outcomes.
7][18] Emotional functioning which included feelings of sadness and anxiety at the time of HSCT was found to predict quality of life 1 year post HSCT. 19Our study adds to the existing literature by investigating not only depression and anxiety symptoms, but also how illness perception may influence HRQOL at both Day 100 and 1 year post HSCT.The Day 100 and 1 year time points are particular interest as prior studies have suggested that mental health factors tend to improve by 100 days post transplant 17,20 while HRQOL tends to improve by 1 year post transplant. 21he aim of study was to investigate whether depression and anxiety symptoms and illness perception prior to HSCT predict HRQOL at Day 100 and 1 year following HSCT.Based on theoretical and prior empirical findings, we hypothesized that a greater level of depression and anxiety symptoms and having a more threatening perception of illness at baseline assessment prior to HSCT will predict diminished HSCT at both Day 100 and 1 year post transplant.

| Patients and data collection
Patients eligible for this study were adults ≥18 years of age undergoing autologous or allogeneic HSCT.Inclusion criteria included the ability to speak and read English.The individuals who participated in this investigation provided written informed consent.All HSCT patients receive a comprehensive multi-specialty assessment prior to transplantation and baseline assessment was completed in person during this period of time.Information was collected regarding baseline characteristics (age at transplant, sex, ethnicity, race, marital status, level of school completed, current employment status, approximate annual household gross income, smoking history, primary disease type, current alcohol consumption, and BMI), psychological factors (Brief Illness Perception Questionnaire [BIPQ], Patient Health Questionnaire-9 [PHQ-9], Generalized Anxiety Disorder-7 [GAD-7]), and outcomes (Functional Assessment of Cancer Therapy-BMT [FACT-BMT] at Days 100 and 1 year following BMT).Results are presented separately for the autologous transplant patients, allogeneic transplant patients, and the overall patient group.Post HSCT data were collected at Day 100 and 1 year primarily via mail, or in person when possible.All study measures were administered using paper versions.This study was approved by the Mayo Clinic institutional review board (IRB study ID: 14-004628).
2.2 | Psychological and health-related quality of life assessments

| Depression symptoms
The PHQ-9 22 is 9-Item reliable and valid self-report measure of depression symptoms that is widely used as a screening tool in health care settings. 23Patients answer items on a Likert scale from 0 (not at all) to 3 (nearly every day).On this measure responses can be categorized according to level of depression including minimal (0-4), mild (5-9), moderate (10-14), moderately severe (15-19), and severe (20-27).The PHQ-9 was administered at baseline, Day 100, and 1 year.

| Anxiety symptoms
The GAD-7 24 is a reliable and valid 7-item self-report measure of anxiety symptoms that is widely used as a screening tool in health care settings. 25Patients answer items on a Likert scale from 0 (not at all) to 3 (nearly every day).On this measure responses can be categorized according to level of anxiety including minimal (0-4), mild (5-9), moderate (10-14), and severe (15-21).The GAD-7 was administered at baseline, Day 100, and 1 year.

| Illness perception
The BIPQ 26 items pertain to illness perception (i.e., consequences, timeline, personal control, treatment control, identity, coherence, and emotional representation).Examples of items include: How much control do you feel you have over your illness?: 0 (absolutely no control) to 10 (extreme amount of control); How much do you think your treatment can help your illness?: 0 (not at all) to 10 (extremely helpful); How well do you feel you understand your illness?: 0 (don't understand at all) to 10 (understand very clearly).Items are answered on a Likert scale from 0 to 10 and response anchors differ for each item depending on the nature of the question asked as described above.Items were summed, with certain items being reverse-scored, to obtain an overall score (range: 0-80), with higher scores her scores on this measure reflecting a more threatening view of the illness.The BIPQ has been found to be both a reliable and valid measure of the cognitive and emotional representations of illness. 26revious research has demonstrated that illness perception is generally stable over longer periods of time (i.e., 2 years) 27,28 therefore, similar to a prior study in patients undergoing HSCT 14 we measured illness perception at baseline.

| Health related quality of life
The FACT-BMT 29 was administered at Day 100 and 1 year post HSCT to assess HRQOL.The FACT-BMT is a reliable and validated measure that is commonly used to assess stem cell transplant specific quality of life in clinical trials. 29The questions are rated on a four-point Likert scale from 0 (not at all) to 4 (very much).Certain items are reverse-scored so that a higher summed score reflects better functioning (range: 0-148).

| Statistical analysis
Missing data for the individual questions used to calculate the BIPQ, PHQ-9, GAD-7, and FACT-BMT total scores was imputed as follows.For patients with missing data for more than 20% of the individual questions for a given measure, the total score was considered to be missing.For patients with missing data for less than 20% of the individual questions for a given measure, missing values for individual questions were imputed using the average value of the patients who answered the given question.Imputation was utilized for individual questions only and was not used for total scores of BIPQ, PHQ-9, GAD-7, and FACT-BMT.
The extent of imputation of missing data was generally minimal.Using the aforementioned missing data imputation strategy, for baseline BIPQ, 202 patients did not require imputation, 1 patient had data imputed (only 1 question was imputed), and 2 patients were considered as missing.Regarding baseline PHQ-9, 200 patients did not need imputation and 5 patients had data imputed (all 5 only had imputation for 1 question).The baseline GAD-7 did not require any imputation for 201 patients, with the remaining 4 patients having data imputed (all 4 only had imputation for 1 question).For the FACT-BMT, at Day 100 a total of 97 patients did not need any imputation, 57 had data imputed (Median number of questions imputed = 1), and 51 patients were considered as missing, while at 1 year 90 patients did not require any imputation, 49 had data imputed (Median number of questions imputed = 1), and 66 patients were considered as missing.Given the greater extent of missing data for the FACT-BMT at follow-up time points compared to baseline BIPQ, PHQ-9, and GAD-7, in Table S1 we have displayed the number of FACT-BMT questions answered (of 37 total) at the Day 100 and 1-year time points.
Continuous variables were summarized with the sample median and range.Categorical variables were summarized with number and percentage of patients.Comparisons of characteristics between patients with and without FACT-BMT information available were made using a Wilcoxon rank sum test (continuous and ordinal variables) or Fisher's exact test (categorical variables).Comparisons of PHQ-9 total score and GAD-7 total score between baseline, Day 100, and 1 year time points were made using a paired Wilcoxon signed rank test; p-values <0.0167 (i.e., 0.05/3) were considered as statistically significant after applying a Bonferroni correction for multiple testing for the three pair-wise comparisons made for each psychological measure.Pair-wise correlations between baseline PHQ-9 total score, baseline GAD-7 total score, and baseline BIPQ total score were assessed using Spearman's test of correlation.
Associations of baseline PHQ-9 total score, baseline GAD-7 total score, and baseline BIPQ total score with FACT-BMT (in the autologous transplant subgroup, the allogeneic transplant subgroup, and in all patients) were evaluated using unadjusted and multivariable mixed effects linear regression models.All models included a random effect for patient and were adjusted for time point (Day 100 or 1 year).Multivariable models were considered as the primary analysis and were additionally adjusted for the pre-defined potential confounding variables of age at transplant, sex, ethnicity, race, smoking history, current drinking, and BMI.Additionally, transplant subgroup was also adjusted for in analysis of all patients in both unadjusted and multivariable analysis.Presence of interactions of time point with PHQ-9 total score, baseline GAD-7 total score, and baseline BIPQ total score were also assessed.Regression coefficients (referred to as β) and 95% confidence intervals (CIs) were estimated and are interpreted as the increase in the mean FACT-BMT total score corresponding to a specified increase in the given psychological factor.We utilized a Bonferroni correction for multiple testing in order to account for the three different psychological measures that were assessed for association with FACT-BMT total score, after which p-values <0.0167 (i.e., 0.05/3) were considered as statistically significant.All statistical tests were two-sided.Statistical analyses were performed using SAS (version 9.4; SAS Institute, Inc., Cary, North Carolina).

| RESULTS
A total of 205 patients who underwent HSCT (N = 127 autologous transplants, N = 78 allogeneic transplants) at the Mayo Clinic in Jacksonville, Florida between December 2014 and October 2020 were included in this prospective study.A total of 83 individuals declined an invitation to participate in this study, resulting in a response rate of 71.2%.Participant characteristics at baseline are summarized in Table 1, while patient scores on psychological and HRQOL measures are displayed in Table 2.The BIPQ, PHQ-9, GAD-7, and FACT-BMT are further illustrated in Figure S1 for the separate autologous and allogeneic transplant groups.A summary of participant responses to the individual items of the BIPQ is presented in Table S2.Of note, when comparing characteristics between the 40 patients without FACT-BMT data available at Day 100 or 1-year to the 165 patients with such data available (Table S3), no differences were noted except for primary disease type (p < 0.001) and type of transplant (p = 0.002).Of note, BIPQ total score (p = 0.78), PHQ-9 total score (p = 0.85), and GAD-7 total score (p = 0.74) were similar between the two groups.
The PHQ-9 and GAD-7 were compared across time points and these results are presented in Table 3.No statistically significant (after multiple testing correction) differences between time points were identified.Of note, there were significant correlations between baseline PHQ-9 total score and both baseline GAD-7 total score (Spearman's r: 0.62, p < 0.001) and baseline BIPQ total score (Spearman's r: 0.45, p < 0.001), as well as between baseline GAD-7 total score and baseline BIPQ total score (Spearman's r: 0.48, p < 0.001).
Associations between psychological factors (i.e., BIPQ, PHQ-9, and GAD-7) with HRQOL at Day 100 and 1 year post HSCT are presented in Table 4. Results are presented separately for autologous and allogeneic transplants as well as in the combined group.In multivariable analysis adjusting for potential confounding variables and after correcting for multiple testing, there were significant associations of a greater BIPQ, PHQ-9, and GAD-7 scores with a lower FACT-BMT total score at the two follow-up time points in all three patient groups (all p < 0.001).There was no evidence of a difference in the degree of these associations between the Day 100 and 1-year time points at which FACT-BMT was measured (all interaction p ≥ 0.062).To illustrate further in the combined group of all patients where sample size is largest, in multivariable analysis FACT-BMT was significantly associated with BIPQ (β: −9.21, p < 0.001), PHQ-9 (β: −11.53, p < 0.001), and GAD-7 (β: −11.08, p < 0.001).
For interpretation, an increase of 10 points in the BIPQ at baseline predicted a 9.21 point mean reduction in the FACT-BMT at follow-up time points across the entire sample.Furthermore, 5-point higher scores for the PHQ-9 and GAD-7 corresponded with 11.53 and 11.08 lower mean FACT-BMT levels at Day 100 and 1 year.For the PHQ-9, GAD-7, and FACT-BMT a 5-point change in score is considered clinically significant. 23,24,29A score that would signify a substantial change in illness perception has not been established for the BIPQ.

| DISCUSSION
This study evaluated the association of illness perception and depression and anxiety symptoms assessed prior to HSCT transplant with HRQOL at Day 100 and 1 year posttransplant.Consistent with our hypothesis, results show that pretransplant illness perception, depression symptoms, and anxiety symptoms were significant predictors of HRQOL post HSCT.Patients who expressed a higher level of concern about the severity, course, and ability to exert control over one's illness (i.e., illness perception) and who reported a higher levels of depression and anxiety symptoms prior to HSCT experienced lower HRQOL at both Day 100 and 1 year posttransplant.These findings were consistent in autologous and allogeneic patients as well as in the combined group that included all patients.

| Clinical Implications
Our study is the first to evaluate whether pretransplant illness perception and depression and anxiety symptoms predict HRQOL at both Day 100 and 1 year following HSCT.This pattern of results for the association of depression and anxiety symptoms with HRQOL is consistent with prior investigations that assessed this relationship at other time points. 16,30Our investigation suggests that patients preparing to undergo HSCT who express feeling of depression and anxiety would benefit from intervention efforts to address these concerns prior to transplant.While HRQOL following HSCT is clearly an important outcome on its own, studies have also found psychological factors including depression and anxiety associated with higher posttransplant hospital readmission, 31 morbidity, 32 and mortality. 33,34Additionally, addressing depression and anxiety prior to and during recovery from HSCT could positively impact patient ability to engage in their recovery, adhere fully to treatment recommendations, and better cope with the inevitable uncertainty and multitude of potential neuropsychological and physical complications. 35ur findings also suggest that patients would benefit from additional efforts to assess and when appropriate enhance illness perception prior to HSCT.Some patients may have inaccurate negative perceptions about their  illness and/or HSCT.For example, a given HSCT patient may express that they have a poor understanding of their illness or have little confidence that HSCT will be of benefit to them.These inaccurate perceptions may be readily addressed via additional education from the health care team.However, as our study was designed to study the association between the construct of illness perception and HRQOL and not the beliefs that influence illness perception future research is needed.Since the median score on the PHQ-9 and GAD-7 was in the minimal symptom range, our findings suggest that HSCT patients with higher scores, particularly those scoring about the minimal symptom range with a scores of 5 or greater on either the PHQ-9 or GAD-7, be closely followed and potentially identified and offered additional mental health support.An additional consideration is that HSCT patients often have fewer medical appointments after Day 100, and thereby potentially access to mental health services are also diminished after this time point.Given our findings that depression and anxiety symptoms remain associated with diminished HRQOL at 1 year, individuals reporting increased depression or anxiety prior to transplant may benefit from continued mental health services beyond Day 100.
Only a few studies been undertaken to develop an intervention to address HRQOL in HSCT patients.One investigation found a positive psychology intervention to be feasible and well accepted by HSCT patients. 36Two other studies found that the addition of inpatient palliative care resulted in improved HRQOL and less depression following HSCT 37 and lower depressive and post-traumatic stress disorder symptoms. 38Thus, as both our findings and prior evidence are consistent in finding illness perception and depression and anxiety symptoms related to HRQOL following HSCT, additional work is clearly needed in this area. 39Research is also needed to examine the optimal time in relation to HSCT to deliver an intervention.If psychological wellness were improved, patients might experience improved HRQOL throughout the HSCT process which also has the potential to improve posttransplant medical outcomes.

| Study limitations
Several limitations of this study are important to consider.First, the sample size is not overly large, and therefore, the possibility of a type II error (i.e., a false-negative finding) is important to consider, particularly in the relatively Abbreviation: β = regression coefficient; CI = confidence interval.
Note: Regression coefficients, 95% CIs, and p-values result from mixed effects linear regression models.All models included a random effect for patient and were adjusted for time point (Day 100 or 1 year).Regression coefficients are interpreted as the change in the mean FACT-BMT total score per each mean outcome level per each 10 unit increase in BIPQ total score and per each 5 unit increase in PHQ-9 total score or GAD-7 total score.Multivariable models were additionally adjusted for age at transplant, sex, ethnicity, race, smoking history, current drinking, and BMI.Additionally, transplant subgroup was also adjusted for in analysis of all patients in both unadjusted and multivariable analysis.p-values <0.0167 were considered as statistically significant after applying a Bonferroni correction for the three different psychological measures that were assessed for association with FACT-BMT total score.There was no evidence of a difference in the degree of the associations of BIPQ, PHQ-9, or GAD-7 with FACT-BMT total score between the Day 100 and 1-year time points (all interaction p ≥ 0.062).
small allogeneic transplant subgroup.Second, we cannot rule out the possibility that our results could have been influenced by variables that we did not adjust for in our multivariable models, or by unmeasured confounding variables.Third, the patients from our single-center study represent a specific population (i.e., most were White with a greater than high school level of education), and therefore our findings may not be generalizable to other patient populations.Fourth, our study did not assess associations of depression, anxiety, and illness perception with clinical outcomes such as survival and relapse, as this was beyond the scope of our study.Nonetheless, this will be an important topic for future study.Finally, information regarding FACT-BMT was unavailable at either follow-up time point for approximately 20% of the study patients, which is not unexpected as this information was primarily collected via mail.However, other than differences regarding primary disease type, type of transplant, and to a lesser extent BMI, no major differences in patient characteristics were noted between patient with and without follow-up FACT-BMT information.Importantly, BIPQ, PHQ-9, and GAD-7 total scores were similar between these two groups.

| CONCLUSIONS
While HSCT continues to remain a promising form of treatment and advances in transplant medicine have led to reduced morbidity and mortality from HSCT, HRQOL remains an important consideration.Our study suggests that addressing possibly modifiable psychological risk factors prior to and/or during recovery from HSCT has the potential to improve this important aspect of patient experience during transplant and also health outcomes.

T A B L E 2 3
Patient scores on psychological and health related quality of life measures.Comparisons of PHQ-9 total score and GAD-7 total score across time points.
Patient characteristics.
T A B L E 1 Note: p-values result from a paired Wilcoxon signed rank test.p-values <0.0167 were considered as statistically significant after applying a Bonferroni correction for multiple testing for the three pair-wise comparisons that were made for each psychological measure.

between the given psychological measure and FACT-BMT total score
Associations of psychological factors at baseline with Day 100 and 1 year FACT-BMT total score.
T A B L E 4