Impact of Psychological Screening on Routine Outpatient Care of Hematopoietic Cell Transplantation Survivors

Abstract

Hematopoietic cell transplantation (HCT) recipients are at high risk for psychological distress with reported prevalence rates as high as 40%. Although published guidelines advocate periodic routine screening, it is unclear how screening affects management of psychological symptoms at routine post-HCT outpatient clinic visits. We hypothesized that providers will be more likely to act on patients’ psychological symptoms if a screening survey is completed and reviewed prior to a clinic visit. We used a brief, diagnostically focused Patient Health Questionnaire (PHQ), to assess for depressive disorders, anxiety, substance abuse, and problems in occupational or interpersonal functioning (functional disruption). Adult HCT survivors were randomized to complete the PHQ prior to meeting with their medical provider (n = 50; experimental group) or afterwards (n = 51; control group). Providers used the experimental group PHQ results at their discretion during the visits. Both providers and patients rated their satisfaction with management of psychological concerns after the visit. The prevalence of clinically significant depression (21%), anxiety (14%), or suicidal ideation (8%) did not differ between the two groups. Patients in the experimental group were significantly more likely to have discussion of psychological symptoms than the control group (68% versus 49%, P = 0.05). Medical providers were significantly more satisfied with the management of psychological issues for the experimental group (P < 0.001). Patients with depression or anxiety were significantly more likely to prefer the PHQ be used at future visits (P = 0.02 and P = 0.001 respectively). These findings suggest an informative yet brief self-report psychological screen can be easily integrated into routine care of HCT survivors, stimulates discussion of psychological symptoms, and improves provider satisfaction with psychological symptom management. Future research will evaluate whether serial prospective administration improves patient outcomes.

INTRODUCTION

Many adults cope well following hematopoietic stem cell transplant (HCT). However, HCT elicits strong psychological responses in a sizeable minority of patients, not all of them with a history of pre-HCT psychopathology. The overall prevalence of psychological difficulties during initial hospitalization is high (44%), and includes mood (14%), anxiety (8%), and adjustment (23%) disorders [1]. In the long term, HCT survivors remain at risk for emotional and/or cognitive difficulties as indicated by a psychological distress prevalence rate of 43% in survivors more than 3 years post-HCT [2].

Because diagnosis, prognosis and treatment processes involved in HCT can be expected to elicit a certain amount of normative distress, differentiating between normal and abnormal emotional responses is difficult. Without a systematic approach to identifying psychological distress, many oncologists frequently miss symptoms of depression [3, 4]. Thus a reliable and valid assessment method is required. However, in order to be easily integrated into hectic routine HCT outpatient clinics, assessments need to be time-efficient, to impose a low burden to patients and be easy for providers to interpret.

The aim of the current study was to extend existing Patient Health Questionnaire (PHQ) screening research [5, 6] by evaluating (1) whether using the brief, diagnostically focused, mental-health-oriented PHQ [7] in routine outpatient office visits of HCT patients is feasible and useful in identifying depression, anxiety, and substance abuse, (2) whether it affects how HCT providers address psychological distress, and (3) whether it impacts the satisfaction of patients and physicians with the care provided during these visits. We hypothesized that compared to treatment as usual, patients in the experimental group who were PHQ-screened before their routine outpatient visit would be more likely to discuss emotional issues with their physician, resulting in greater satisfaction of patients and physicians with the care provided.

PATIENTS AND METHODS

Participants

This study was reviewed and approved by the University of Michigan institutional review board. Patients were enrolled during standard scheduled post-HCT clinic visits in the University of Michigan Blood and Marrow Transplant Clinic (August, 2010 – February, 2011). Inclusion criteria included age ≥ 18 years, status ≤ 5 years post autologous or allogeneic HCT, mental competence to provide informed consent and ability to read and speak English.

Procedure

The study was cross-sectional in design, and no patient participated more than once. Upon completion of the informed consent, patients were randomly assigned to complete the PHQ depression, anxiety, and substance abuse modules prior to meeting with their medical provider (n = 50; experimental group) or afterwards (n = 51; control group). All patients completed a demographic background questionnaire before meeting with their physician, and afterwards completed a post-visit questionnaire regarding treatment decisions and satisfaction with the visit. The 10 participating HCT physicians treated patients in both groups. Physicians were given access to the experimental group PHQ scores to use at their discretion during the visits. Physicians provided treatment as usual to patients in the control group without access to PHQ data and without being informed in advance of the visit that control patients were participating in the study. However, after control group visits, research assistants immediately informed providers if patients had checked any response other than “not at all,” in response to the PHQ item which assesses for suicidal ideation. Physicians then responded as needed to that information. After each visit irrespective of group, providers completed a checklist of psychological interventions initiated or continued at that visit, and rated their satisfaction with the visit and with the management of psychological concerns.

Measures

Medical variables were extracted from the medical database. Demographic information and history of prior mental health difficulties and history of mental health treatment (behavioral and/or psychopharmacological) was assessed by patient report in a brief survey consisting of 28 items which was designed for this study. The timing of the diagnoses of mood disorder (major depressive disorder, bipolar disorder, dysthymia) and anxiety (generalized anxiety disorder, phobias, obsessive-compulsive disorder, post-traumatic stress disorder, panic disorder) was individually assessed for by patients choosing one of three possible responses for each diagnosis, (1) prior to HCT, (2) during or after HCT, or (3) never diagnosed with the condition.

Patient Health Questionnaire (PHQ) [7]

The PHQ was the primary questionnaire in this study, which has been validated as a diagnostic screening instrument for mood disorders, anxiety disorders, alcohol abuse disorders and other psychopathology. The PHQ has high sensitivity and specificity for detecting generalized anxiety disorder [8] and substance abuse [7, 9], as well as making criteria-based diagnoses of depressive disorders and indicating depression severity [10].

The PHQ modules used in this study were the 9 item depression screener (PHQ-9; clinically significant depression equates to a score ≥ 10), 2 item anxiety screener (GAD-2; clinically significant anxiety equates to a score ≥ 3), and 6 item alcohol abuse screener (which is only completed in full if the lead question is answered in the affirmative). In addition, we used a one question screen that assessed if any emotional difficulties were impacting occupational and/or interpersonal functioning, scored on a four point scale from “not at all” to “extremely difficult.” Endorsement of anything other than “not at all” is referred to in this study as “functional disruption.” With the permission of the PHQ authors, we created and deployed parallel forms of the alcohol abuse screener to screen for recreational drug use (6 items) and prescription medication abuse (6 items). Thus the PHQ modules consisted of a minimum of 15 questions that every participant answered, and an additional 15 that were only answered if the participant admitted to using alcohol (5 items), recreation drugs (5 items), or overusing prescription medications (5 items). Completing all items on the questionnaire (Supplemental Materials, S1) took approximately two minutes.

Post-visit patient satisfaction questionnaire

Patients completed a post-visit satisfaction questionnaire (Supplemental Materials, S2) which assessed (1) whether the visit included discussion of mental, emotional, or behavioral issues (‘yes’ or ‘no’), (2) if interventions in the form of psychopharmacological medication management or referral to mental health providers occurred (‘yes’ or ‘no’), (3) their attitudes toward using the PHQ in future (‘always,’ ‘sometimes,’ or ‘never’) and (4) their satisfaction with the visit. Patient visit satisfaction was broken down into three categories: (1) overall, (2) the way mental health or emotional issues were addressed and (3) the provider’s concern for the patient as “a whole person” and was rated on a 7-point scale Likert scale from “not at all satisfied” to “very satisfied.”

Post-visit provider report

Providers completed a post-visit survey (Supplemental Materials, S3) that documented (1) whether mental health and emotional issues were discussed during the visit, (2) if interventions in the form of referral to mental health providers or psychopharmacological medication management occurred, (3) whether the information provided by the PHQ influenced intervention for that visit and (4) their satisfaction with the visit overall and with the degree to which psychosocial issues were addressed. The two satisfaction ratings were on a 7-point scale Likert scale from “not at all satisfied” to “very satisfied.”

Statistical analysis

SPSS Version 19 (SPSS Inc. Chicago, IL, USA) was used for statistical analysis. Standard descriptive statistics (frequency, means, standard deviation, and range) were used to summarize the demographic, medical and psychological variables and associations between variables were assessed by Pearson correlation coefficient for continuous variables and Spearman’s ρ for ordinal variables. The two sample t test was used to compare the experimental and control groups on continuous measures of psychopathology and satisfaction, and Pearson’s Chi Square tests to compare groups on the dichotomous principal outcome variables of discussion of mental health issues, referral and psychopharmacological prescription. Multiple linear regression (for continuous outcomes) or logistic regression (for dichotomous outcomes) was used to assess associations among the principal outcome variables and other variables of interest.

RESULTS

Participants

Participants were 101 HCT survivors (83 allogeneic and 18 autologous graft recipients) with a median age of 53 years (range 18 – 70), who were a median of 9.5 months (range 20 days – 4.99 years) post-HCT. The demographic and medical characteristics are listed in Table 1. The experimental and control groups did not differ significantly in self-reported past psychopathology. For the entire cohort, the prevalence of self-reported mood disorder diagnoses was 14% (8% and 6% diagnosed pre- and post-HCT respectively) and self-reported anxiety diagnoses was 40% (22% and 18% diagnosed pre- and post-HCT respectively). The prevalence of self-reported history of substance abuse diagnoses was 6% (Table 2). The proportion of patients who recalled at least some need for mental health services versus no need increased from 30% pre-HCT to 53% post-HCT (OR 8.67; 95% CI 2.62 – 28.6; P < 0.001) (Figure 1).

Table 1.

Characteristics of the 101 Study Participants

	Mean ± SD, or n (%)	Range
Age, years	51.06 ± 13.77	18 – 70
Time from HCT to study, days	481.51 ± 492.70	20 – 1801
Education, years	14.49 ± 2.75	10 – 24
Gendera: Male	52 (51%)
Ethnicitya
White	95 (95%)
Black	4 (4%)
Other	2 (2%)
Married/partnered	69 (68%)
Income, annual
>$100,000	13 (13%)
>$50,000	29 (29%)
>$25,000	25 (25%)
≤$25,000	15 (15%)
Prefer not to say	18 (18%)
Indication for HCT
Acute Leukemia/MDS	49 (48%)
Non-Hodgkin Lymphoma	23 (23%)
Multiple Myeloma	12 (12%)
Other	17 (17%)
Type of HCT
Autologous	18 (18%)
Related Donor	41 (41%)
Unrelated Donor	42 (42%)
Acute GVHD	42 (51%)b
Chronic GVHD	44 (53%)b
Pre-HCT risk of relapse (CIBMTR criteria)
Low	40 (40%)
Intermediate	44 (44%)
High	17 (17%)
Full Intensity Conditioning	52 (51%)
Steroids before study participation	32 (32%)
Steroid dosec at study participation (mg/kg)	.32 ± .31	.01 – 1.08
Relapsed before study participation	6 (6%)
Time from relapse to study, days	310.33 ± 415.5	2 – 112
Karnofsky Performance Scale	81.98 ± 8.83	70 – 100

^an = 100;

^bpercent of the 83 allogeneic graft recipients;

^cMedrol dose equivalent

Table 2.

Self-Reported Incidence of Psychological Disorders Before and After HCT (n(%))

	Before	After	Total
Mood disorders	8 (8%)	6 (6%)	14 (14%)
Mood >1 disordera	7 (7%)	6 (6%)	13 (13%)
Depression	5 (5%)	6 (6%)	11 (11%)
Bipolar	2 (2%)	0 (0%)	2 (2%)
Dysthymia	1 (1%)	0 (0%)	1 (1%)
Anxiety	22 (22%)	18 (18%)	40 (40%)
Anxiety >1 disorderb	19 (19%)	12 (12%)	31 (31%)
Generalized Anxiety Disorder	17 (17%)	10 (10%)	27 (27%)
Phobias	0 (0%)	0 (0%)	0 (0%)
Obsessive Compulsive Disorder	1 (1%)	2 (2%)	3 (3%)
Posttraumatic Stress Disorder	2 (2%)	3 (3%)	5 (5%)
Panic	2 (2%)	3 (3%)	5 (5%)
Substance abuse
Alcohol abuse	3 (3%)	0 (0%)	3 (3%)
Recreational drug abuse	2 (2%)	0 (0%)	2 (2%)
Medication abuse	1 (1%)	0 (0%)	1 (1%)
Axis II Diagnosis	3 (3%)	2 (2%)	5 (5%)
Adjustment and otherc disorders	6 (6%)	14 (14%)	20 (20%)

^aMore than one mood disorder

^bMore than one anxiety disorder

^cOther includes Attention Deficit Disorder, Body Dysmorphic Disorder, interpersonal problems, sexual difficulties, death in family, grief, little patience when in pain, memory and cognitive dysfunction

Figure 1.

Perceived need for mental health assistance increased from pre- to post-HCT (P < 0.001)

Current psychopathology as indicated by the PHQ

There were no statistically significant differences between the experimental and control groups in current psychopathology according to the PHQ modules. However, patients reported a higher incidence of depression (21% versus 8%, P = 0.001), a lower incidence of anxiety (14% versus 22%, P = 0.01) but no statistically significant change in incidence of substance abuse according to the post-HCT PHQ diagnostic criteria when compared to their recall of being professionally diagnosed pre-HCT (Figure 2). Notably, 12% of the study sample met PHQ criteria for more than one form of psychopathology. Education level was unrelated to current psychopathology or mental health diagnosis pre-HCT. However, patients with less education were more likely to be professionally diagnosed with a mental health disorder post-HCT (Spearman’s ρ =−.26, P =.009).

Figure 2.

Percent of patients with clinically significant functional disruption, depression, anxiety and substance abuse, as self-reported pre-HCT and as indicated by PHQ at time of study participation

** P = 0.01; *** P = 0.001

Suicidal ideation was surprisingly common. Eight percent of all participants (4 patients in each arm) reported suicidal thoughts for at least several days in the most recent two weeks. Suicidal ideation was positively correlated with the sum of all other depressive symptoms (Spearman’s ρ = .33, P < 0.001), anxiety (Spearman’s ρ = .23, P = 0.02), and longer time since transplant (Spearman’s ρ = .22, P = 0.03), but negatively correlated with age (Spearman’s ρ = −.21, P = 0.04). Thus suicidal ideation was more likely in younger patients, those with more depressive symptoms, and those farther out from HCT. Pre-HCT risk of relapse was not associated with suicidal ideation. Suicidal ideation was reported by one (out of six, 16%) of the patients who had relapsed post-HCT but prior to study participation. These small numbers preclude any formal statistical analysis.

Psychological intervention

In support of our hypothesis, patients in the experimental arm were significantly more likely to have discussion of psychological symptoms than the control group (68% vs 49%, P = 0.05). Univariate analyses indicated multiple demographic variables (gender, age, and years of education) and medical variables (post-HCT relapse, days post-HCT dichotomized as <180 days or ≥ 181 days) did not impact the likelihood of discussion. When each of these variables was controlled for in a multiple logistic regression, there was still a trend toward more discussion in the experimental group (P = 0.10). However, the two groups did not differ in rates of referral to mental health providers (12% versus 10%, for the experimental and control groups respectively; P = 0.49) or prescription of psychotropic medication (14% in both the experimental and control groups; P = 0.60). As one would expect, patients in either group who discussed “mental, emotional or behavioral problems” during their clinic visit were more likely to be referred to mental health services (OR = 8.16, 95% CI, 1.0 – 66.5; P = 0.05) and prescribed psychotropic drugs (OR = 4.98, 95% CI, 1.1–23.6; P = 0.03). The impact of discussion on intervention (referrals or prescription) was not significantly affected when relevant demographic variables (gender, age, and years of education) and medical variables (post-HCT relapse, and days post-HCT dichotomized as <180 days or >181 days) were controlled for in a multiple logistic regression model.

Provider and patient perspectives

Patient satisfaction was skewed to the high end of the rating scale and did not differ between groups. Patients with clinically significant depression or anxiety based on the PHQ, irrespective of study arm, were significantly more likely to wish that the questionnaire be used at future outpatient visits (depression OR = 3.88, 95% CI 1.27 – 11.9; P = 0.02; anxiety OR = 17.6, 95% CI 2.18 – 142.3; P = 0.001)

Medical providers with access to the completed PHQ prior to the visit were significantly more satisfied with their patient management when compared to the control group. The mean satisfaction of providers for the experimental group was higher both for management of psychological problems (6.11 versus 5.56, P < 0.001) and for the visit as a whole (5.89 versus 4.66 respectively P = 0.005). Providers reported that information provided by the PHQ influenced their psychological management for 48% of the patients in the experimental group.

DISCUSSION

This study’s novel design, comparing effects of screening versus treatment as usual without screening in routine outpatient care of HCT survivors, addresses “the paucity of clinical trials focused on screening and preventive practices” noted by Rizzo [11] regarding the psychological status of HCT survivors. By focusing on the screening process itself in routine outpatient follow-up care, our study demonstrates that when presented with informative psychological screening data, HCT physicians were more likely to discuss patients’ mental health concerns and demonstrated greater satisfaction with the care provided to those patients. In addition, patients who endorsed anxiety or depression wished that the PHQ be used in future visits.

For physicians, the act of acknowledging and reviewing the PHQ results with the patient appeared to function as an intervention, whether or not new referrals or psychopharmaceutical interventions were viewed as needed at the time. The absence of difference in terms of intervention between the two groups could be because the physicians in this study were already quite adept at the detection of psychological issues requiring clinical attention. Alternatively, use of the PHQ in the experimental group may have sensitized physicians to mental health concerns in the control group. From the perspective of patients with depression or anxiety, discussing the PHQ results with their physicians was viewed as beneficial, based on their desire to continue use of the PHQ. Although we did not investigate this directly, we surmise the discussions of psychological issues might serve also to strengthen the therapeutic alliance between physician and patient who often work together for months or years.

Prevalence rates in our sample of clinically significant depression (21%) and anxiety (14%) as indicated by the brief PHQ screening instrument are consistent with rates reported in the literature of patients 1–2 years post-transplant [12, 13].

The prevalence rate of suicidal ideation in our sample, not commonly reported in the HCT literature, was alarming at 8%, although the confidence interval is quite wide and overlaps with the much lower 12 month prevalence rate of 2.8 – 3.3% in a large nationally representative sample in the United States [14]. Furthermore, it is possible that study patients experiencing suicidal ideation may be more likely to choose to participate in psychological screening studies, which would skew our findings. Nonetheless, the brief screening instrument used in this study, the PHQ-9, which has been validated for use with cancer outpatients [15,16] and HCT patients [17], facilitated provider detection of suicidal thoughts, This finding suggests the PHQ could be used regardless of specialty clinic setting, to assess all HCT survivors, both short and long-term. More investigation is warranted to better quantify the degree of psychological distress during the post-HCT recovery period.

Although the majority of transplant centers conduct psychosocial assessments of their patients prior to HCT, our study suggests that ongoing clinical assessment is needed, as not all patients experiencing distress after HCT can be identified on the basis of their distress prior to HCT. In our study, a modified PHQ demonstrated utility for facilitating discussion of emotional wellbeing and resulted in higher physician satisfaction with the care provided during routine outpatient HCT follow-up visits. Furthermore, assessment by the modified PHQ was feasible in the outpatient setting. As a future direction we propose to evaluate patients serially, which should allow for timely identification of emotional distress and promote discussion between patients and providers. We expect serial PHQ assessments would also provide physicians with an objective method to evaluate response to any given intervention. In the meantime, our study results support the use of the modified PHQ as a useful psychological assessment tool for outpatient HCT recipients, particularly when the provider reviews the results prior to meeting with the patient.