1 Introduction

Hematopoietic stem cell transplant (HSCT) is a well-established and increasingly utilized procedure to address malignant and non-malignant conditions [1]. Despite being a potentially life-saving procedure, the intense nature of HSCT-related procedures challenges patients' ability to adhere to medical interventions and healthcare provider recommendations prior to and well after HSCT [2-4]. Intensive conditioning regimes, high chemotherapy doses that may cause drug toxicities, infections, and immunodeficiency complications as well as the long recovery process post-HSCT, remain common experiences for many patients [5, 6].

Depending on the clinical presentation, individuals may have several treatment options to consider, while others may be debating between pursuing HSCT and palliative measures. Given the treatment decisions individuals with cancer must make, regret has been an area of interest for researchers across varying cancer populations [7]. Regret, as it relates to cancer and medical decision making, has been defined several ways; however, Connolly and Reb have noted that most definitions include four parts: (1) regret will be avoided when possible, (2) it includes emotional and cognitive components, (3) it is unique from concepts of negative affect including disappointment, and (4) there is a comparison between the decision and another option [8].

Although much of the regret literature has looked at non-hematological cancer populations [7], a small body of research has examined regret in patients undergoing HSCT [9, 10]. Overall, these studies suggest that regret occurs in a small percentage of cases. In a study exploring regret at 100 days, 5 months, and 12 months post-HSCT, 6%–8% of patients endorsed regret at each timepoint [9]. In a sample of HSCT survivors (mean years since transplant = 5.7 years, SD = 3.4, range = 1–14 years), 9.2% of patients who underwent allogenic transplantation reported regret in the past 7 days [10]. Findings suggest that individuals with lower quality of life and household income, greater depressive symptoms and number of medical care visits, and those who experienced graft-versus-host disease (GVHD) or recurrence following transplant were more likely to endorse regret [9, 10]. In a qualitative study of 25 patients exploring the impact of experimental stem cell transplant, no patients reported regret related to undergoing transplant because HSCT was viewed as an opportunity to lengthen their lives or as their only option [11]. However, themes of regret did emerge relative to post-transplant physical and cognitive changes [11]. These findings underscore the complex decision-making process to pursue HSCT as well as the evolving emotional and cognitive appraisals of the decision following the procedure.

Given the small number of studies that have examined HSCT regret, and the varying methodologies used, the prevalence and predictors of transplant regret are unknown. Therefore, the first aim of the study was to document the prevalence of regret and explore demographic, disease, and psychological factors at pre-transplant that may be predictive of regret 1-year post-HSCT. Our second aim was to determine if post-transplant variables were associated with regret 1-year post-HSCT. Additionally, we sought to address two gaps in the current literature by including individuals who received either allogeneic or autologous transplant and by utilizing multiple sources of information including medical record abstraction, self-report, and mental health data obtained by diagnostic interview with a licensed mental health provider. We hypothesized that the individuals who had greater psychological burden (e.g., mental health diagnosis), more stressors, lower Psychosocial Assessment of Candidates for Transplantation (PACT) score, and those with greater disease severity at pre-transplant would be more likely to endorse future regret. We also hypothesized that individuals with complications post-transplant (e.g., GVHD, unplanned hospitalizations) would be significantly more likely to endorse regret.

2 Methods

The parent study was a prospective cohort study of 1000 individuals preparing to undergo HSCT at a large academic medical center in the Midwest that examined health behaviors prior to and 1 year following HSCT. Given the significant resources required to abstract psychological and social work records, we conducted a case–control analysis nested within the cohort. Nested case–control studies are advantageous in situations such as this while minimally impacting statistical performance [12]. Age, sex, and transplant type were selected as matching variables through discussions with our multidisciplinary team with the goal of reducing the chance of biased comparisons in the nested sample. The advantage of matching is to optimize group equivalency when studying a rare outcome. All participants provided written informed consent. The protocol was approved by the Mayo Clinic Institutional Review Board.

Of the 1000 people who enrolled in the parent study, 870 (87%) completed study surveys prior to HSCT (defined as baseline). Among the people who completed baseline questionnaires, 638 (73%) completed and returned the 1-year follow-up survey. Forty-nine of these participants endorsed regret (cases). The 49 cases were matched 1:2 with participants who did not endorse regret at the 1-year follow-up [13, 14] which resulted in a final sample size of 147. We successfully matched all participants on age (±2 years), sex, and transplant type (allogeneic vs. autologous) to account for potential confounding variables.

Following matching, psychosocial clinical data, including psychological and social work records in the year prior to transplantation, were reviewed and abstracted. Psychological evaluations included psychiatric diagnostic interviews inclusive of mental health diagnosis using the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) [15]. Two authors (J.L.K., H.C.E.) independently rated each participant on the PACT [16, 17] measure based on the available clinical records. Rating disagreements were discussed with the senior author (S.L.E.), and consensus was achieved.

3 Results

Approximately 8% of the individuals (n = 49) in the original cohort who underwent HSCT transplant reported regret at 1 year. Individuals who endorsed transplant regret tended to be middle aged (mean = 57 years ±10 years), White (n = 44, 90%), married or partnered (n = 44, 90%), diagnosed with multiple myeloma (n = 22, 45%) and underwent autologous transplant (n = 40, 82%). Forty-five percent of people who endorsed regret were still completing treatment 1-year post-transplant, and 61% had experienced at least one hospitalization.

Cases and controls were well matched on age (56.6 vs. 57.2 years), sex (34.7% female for both groups), and transplant type (81.6% autologous for both groups). Forty percent rated their health at 1-year post-HSCT as “excellent” or “very good” and over half of patients reported no hospitalizations in the year since transplant. Disease and demographic data by group are provided in Table 1.

Prior to HSCT, 73% of patients completed a psychological evaluation with a diagnostic interview to detect the presence of mental health diagnoses and other psychosocial stressors that may impact the transplant trajectory. The overall prevalence at the pre-HSCT evaluation of at least one mental health diagnosis was similar between the regret and no regret groups (39% and 42%, respectively). Although the cell sizes were small, the group of patients who went on to develop regret had a higher prevalence of co-occurring mental health diagnoses at baseline (16% vs. 8%). The median PACT score for patients who endorsed regret was 3 (mean: 2.59, range: 0–4). For those who did not endorse regret, the median PACT score was also 3 (mean: 2.74, range: 1–4).

Conditional logistic regressions were conducted to examine whether pre- and post-HSCT disease, demographic, and psychosocial variables were associated with post-HSCT regret at 1-year follow-up. There was no association for the number of mental health diagnoses, number of psychosocial stressors (e.g., caregiving/family stress, financial worry, and employment concerns), or PACT score with transplant regret (Table 2). Post hoc analyses did not reveal significant group differences based on neither mental health diagnostic groups (i.e., depressive, anxious, adjustment, and substance use disorders) nor specific psychosocial stressors. Additionally, there were no significant group differences based on disease group, pre-HSCT ECOG score, or conditioning regimen (Table 2).

In contrast, post-HSCT factors showed a significant pattern of association with transplant regret. Participants endorsing regret 1-year post-HSCT were 3.7 times more likely to have been hospitalized at least once in the year following HSCT compared to individuals who did not endorse regret (Table 2). Similarly, patients endorsing regret were 5.8 times more likely to not be in complete remission compared to controls. Patients endorsing regret were three times more likely to report current active treatment compared to controls (who did not endorse regret) (Table 2). There were no significant group differences based on the occurrence of acute or chronic GVHD.

Multivariate conditional logistic regression analyses found that only remission status 1-year post transplant remained significant in the model (Table 3). After controlling for the number of hospitalizations and active treatment status, patients who endorsed transplant regret were 3.7 times (95% CI = 1.37–9.69, p = 0.008) more likely not to be in remission compared to controls at 1-year post-HSCT.

Post hoc analyses were conducted to determine if there were any differences in findings based on allogeneic or autologous transplant. Given the small sample size for allogeneic transplant recipients, only univariate models were performed to reduce the risk of overfitting the model. There were no significant findings for any of the variables of interest, which suggests the model was underpowered. For the autologous patients, the same pattern of results was observed in multivariate analyses as was seen in the combined autologous and allogeneic model. Patients who endorsed regret were 3.9 times (95% CI = 1.37–11.30, p = 0.011) more likely to not be in remission compared to controls at 1-year post-HSCT. Cross-sectional post hoc analyses were conducted separately from prospective analyses to explore whether regret was associated with emotional wellbeing at the shared timepoint of 1-year post-transplant. Regret was associated with lower emotional wellbeing (odds ratio: 0.79, 95% CI = 0.71–0.88, p < 0.001). Given that we were interested in examining variables that may be predictive of regret, cross-sectional variables were not entered into the prospective multivariate analyses.

4 Discussion

Approximately 8% of patients in the current study endorsed transplant regret 1-year post-HSCT, which is consistent with past research [9, 10]. The prevalence of mental health diagnoses prior to HSCT in cases (39%) and controls (42%) was slightly higher than past research, which found that 32.8% of people in oncology and hematology settings experienced comorbid psychological distress and/or disorder (e.g., depression, anxiety), with the specific rates varying by assessment timing and methodology [20]. Contrary to hypothesis, our nested case–control analysis revealed that mental health diagnosis and pre-transplant stressors were not significantly associated with regret 1-year after HSCT. This finding is surprising, as psychological distress has been linked to negative outcomes including higher risk of acute GVHD, greater transplant-related symptoms, avoidant coping, and shorter survival [21-27].

However, as expected, standardized measures indicating a lack of improvement in health status after HSCT (i.e., at least one hospitalization, the absence of complete remission) were consistently associated with regret. Concurrent treatment at 1-year post-HSCT was also significantly associated with regret, whereas the presence of GVHD was not significant in the models. However, the non-significant GVHD finding may be due to the small percentage of patients who experienced GVHD in the sample. Overall, these findings suggest that the post-HSCT clinical course may be the most impactful factor contributing to patients' positive or negative feelings regarding their choice to undergo transplant. It can be theorized that the disappointment or negative emotions associated with a more challenging post-HSCT clinical course could lead those who are experiencing regret to question whether they had made the correct choice to undergo transplant.

In this sample, regret was most strongly influenced by remission status at 1-year post-HSCT. In the multivariate model, only remission status remained significant, with the odds of endorsing regret being almost four times greater for individuals who were not in remission. This is consistent with past research which found that post-transplant complications and disease relapse were significant predictors of regret [9, 10].

Author Contributions

Janae L. Kirsch: conceptualization (equal), formal analysis (equal), methodology (equal), writing – original draft (equal), writing – review and editing (equal). James R. Cerhan: conceptualization (equal), methodology (equal), supervision (equal), writing – review and editing (equal). William J. Hogan: writing – review and editing (supporting). Holly C. Edwards: data curation (supporting), writing – review and editing (supporting). Christi A. Patten: resources (equal), writing – review and editing (equal). Tabetha Brockman: data curation (supporting), writing – review and editing (supporting). Christine Hughes: data curation (supporting), writing – review and editing (equal). Angela Dispenzieri: writing – review and editing (supporting). Stephen M. Ansell: writing – review and editing (supporting). Dennis A. Gastineau: data curation (equal), writing – review and editing (supporting). Shawna L. Ehlers: conceptualization (equal), funding acquisition (lead), investigation (lead), methodology (equal), project administration (lead), supervision (equal), writing – review and editing (equal).

Ethics Statement

The protocol was approved by the Mayo Clinic Institutional Review Board.

Consent

All participants provided written informed consent.

Conflicts of Interest

The authors declare no conflicts of interest.