Telephone monitoring of distress in patients aged 65 years or older with advanced stage cancer

A cancer and leukemia group B study

Abstract

BACKGROUND.

Significant barriers to cancer patients receiving mental health treatment for distress have been reported in the literature. The objective of the current study was to determine whether distress in older patients (aged 65 years and older) would be reduced with educational materials (EM) supplemented by monthly telephone monitoring (TM) (TM + EM) compared with the use of EM alone because of more timely referrals to appropriate health professionals.

METHODS.

One hundred ninety-two older patients with breast, prostate, and colorectal cancers who had advanced disease and currently were receiving treatment were randomized to receive either TM + EM or EM alone. One hundred thirty-one patients were evaluated by telephone interview for psychologic and physical distress and for social support at baseline and at 6 months using the Hospital Anxiety and Depression Scale (HADS), the European Organization for Research and Treatment of Cancer (EORTC)-QLQ-C30 quality-of-life questionnaire, and the Medical Outcomes Study (MOS) Social Support Survey. Patients who in the TM + EM group were called monthly for 6 months to monitor their distress using the HADS and EORTC physical symptom items and the MOS Social Support Survey items, with cutoff levels were established to indicate which patients were in greater distress. Those patients who scored above the cutoff levels were referred to their oncology nurse for referral to the appropriate professional. Patients in the EM group received written materials regarding cancer-related psychosocial issues and available resources.

RESULTS.

At 6 months, patients in the TM + EM group reported significantly less anxiety (HADS; P < .0001), depression (HADS; P = .0004), and overall distress (HADS; P < .0001) compared with patients in the EM group.

CONCLUSIONS.

Monthly monitoring of older patients' distress with TM and EM along with referral for appropriate help was found to be an efficient means of reducing patients' anxiety and depression compared with patients who received only EM. Cancer 2006. Published 2006 American Cancer Society.

Although, in the aggregate, older cancer patients may have less psychologic distress than younger patients,1–4 it is likely that there is a subset of patients who are in need of mental health care. For this distressed subset, the stress of cancer can be exacerbated by other illnesses or disabilities5–7 at a time in life when social support has diminished8–11 and financial resources have decreased.

The medical literature indicates that, for a sizable number of patients, psychologic distress remains underreported and undertreated.12, 13 Barriers to identifying older patients who are in need of mental health services include a lack of reporting emotional problems because of the stigma of having a psychiatric disorder,14 fear of the oncologist's reaction to the expression of a patient's emotional problems, oncologists not knowing how to assess or treat patients with psychiatric conditions,15 oncologists' discomfort in discussing psychologic problems with their patients, a lack of insurance coverage for mental health services, and a lack of physical mobility or transportation needed to obtain mental health services.

Brief screening measures to assess older patients' pain, depression, and fatigue have been suggested as a means for overcoming multiple barriers to their identification and treatment.16 The primary objective of the current randomized trial was to test whether monthly telephone monitoring (TM) of older patients' psychologic, physical, and social support distress, in conjunction with educational materials (EM) (TM + EM), could reduce their psychologic distress significantly more than EM alone because of more timely referrals to the appropriate professional.

MATERIALS AND METHODS

Research Design and Sample

This study was conducted by The Cancer and Leukemia Group B (Study 119801). Patients who were eligible for the study included those aged 65 years or older who currently had stage III or IV breast cancer (T-any,N2-N3,M0; T3,N1,M0; T4,N0-N1,M0; or T-any,N-any,M1), Duke stage C or D colon cancer (T1-T4,N1-N2,M0 or T1-T4,N0-N2,M1) or stage C or D prostate cancer (T3-T4,N1,M0 or T-any,N-any,M1), who were actively receiving treatment that was initiated ≤2 months prior to recruitment to the study, had a life expectancy ≥12 months, and were able to provide informed consent. Eligible patients were randomized to receive either TM + EM or EM alone. The rationale for including the first 2 months of active treatment as part of the criteria was that patients receiving active treatment most likely would be in greater distress than patients after the completion of treatment and therefore would be more likely to have unmet needs detected. All patients were evaluated at the time of study entry, at 6 months, at the completion of TM + EM, and at 9 months to test whether any gains in the intervention group were retained 3 months after the completion of the monthly TM.

Research Procedures

All participants provided informed consent in accordance with the approval of each institution's Institutional Review Board, and an institutional assurance was filed with and approved by the Department of Health and Human Services. Upon consent and after patients had met the criteria for cognitive status as determined by the Blessed Orientation-Memory–Concentration (BOMC) test,17 eligible patients were randomized either to the TM + EM group or to the EM group.

The research interviewer conducted the evaluation interviews of patients by telephone at study entry, at 6 months, and at 9 months. The interview took approximately 45 to 60 minutes using standardized psychosocial instruments. The monthly monitoring sessions conducted by the telephone monitor, using standardized measures, lasted approximately 10 to 15 minutes. TM and research evaluation interviews were conducted by different individuals, so that the intervention and evaluation would not be contaminated. The qualifications for both the research interviewer and the monitors were that they had to be at least college graduates who had received rigorous training by the principal investigator (A.B.K.) using role playing.

Intervention

TM + EM

Patients who were randomized to the TM + EM group received 1 telephone call each month for 6 months from centralized, trained telephone monitors in addition to receiving EM. The EM included the following: Taking Time: Support for People with Cancer and the People who Care about Them,18Eating Hints for Cancer Patients,19 and Helping Hand.20 In addition, EM was included that was specific to each disease site regarding available resources.21–23 Patients were identified as in distress if they scored above established or specified cutoff levels for the following problems: psychologic distress (Hospital Anxiety and Depression Scale [HADS] total score ≥15; or HADS Anxiety and Depression subscale scores ≥8)24; physical problem items on the European Organization for Research and Treatment of Cancer (EORTC) quality-of life questionnaire (QLQ-C30)25 with ratings of ‘quite a bit’ or ‘very much’ that were untreated or not treated sufficiently in patients who had no scheduled appointment with the oncologist in the next few days; and social support distress on the Medical Outcomes Study (MOS) Social Support Survey,26 consisting of no close family or friends or 1 problem in each of the tangible support, emotional support, and social participation subscales. The oncology nurse at the treating institution was called by the telephone monitor within 24 hours concerning those patients who scored above these cutoff levels. The oncology nurse then called the patient to discuss these findings and, if warranted, made a treatment recommendation. Patients also were referred to the oncology nurse if the patient reported levels of distress above cutoff scores at study entry and at the 6-month and 9-month evaluation interviews.

EM alone

Patients in the EM arm of the study received only EM, as specified above. Those in the EM group also were referred to the oncology nurse when they were evaluated at study entry, at 6 months, and at 9 months if they were distressed significantly and scored above the same cutoff levels as patients in the TM + EM group.

Measures

The following measures were included in the study: the EORTC QLQ-C30 questionnaire,25 which assessed general physical symptoms, physical functioning, fatigue/malaise, social functioning, and psychologic distress; the Geriatric Depression Scale (short form)27–29; the HADS30; the MOS Social Support Survey26; the Physical Health subscale of the Older American Resources and Services Questionnaire,31, 32 which measured the presence of any of 22 comorbid conditions and the degree to which they interfered with functioning by using a 3-point Likert scale that ranged from ‘not at all’ to ‘a great deal’; the Utilization of Mental Health and Psychosocial Services instrument,33 which assessed past and current use of mental health and counseling services; the Geriatric Schedule of Recent Experience (GSRE) instrument,34 which assessed stressful life events for individuals aged 70 years or older; the Patient Satisfaction with the Research Program BOMC test,17 which screened for gross cognitive impairment, with scores ≥11 indicating gross cognitive impairment; and medical and sociodemographic characteristics.35

Statistical Analysis

Sixty-five patients from both groups (total, 130 patients) were required at the end of the 6-month study period to detect a median difference between groups with regard to psychologic distress and mental health care use at an α level of .025 (2-tailed) with 80% power. To account for an attrition rate of 33% and a refusal/ineligibility rate prerandomization of 15%, the study's registration and randomization objectives were set at 240 patients and 196 patients, respectively. An intent-to-treat analysis that would include all patients was not feasible in this study because no follow-up data were available for patients who had only baseline data, largely because of illness, death, or inability to contact the patient. Without follow-up data, changes between the 2 groups could not be computed. Imputation of missing follow-up data would have entailed assuming that the correlation between available characteristics of patients with and without follow-up were the same, which was not found to be the case. The analyses reported below illustrated that there were differences identified between patients with and without follow-up that limited our conclusions to only those patients who had follow-up data available. Because the imputation of missing data did not appear appropriate, analyses were conducted with the available data. An analysis of covariance was used to test for physical and psychosocial differences between the 2 treatment arms using the baseline assessment as the covariate. Exact chi-square tests, including Fisher exact tests, were used to examine correlations between categoric outcome variables and patient descriptors. Student t tests were used to compare groups at baseline relative to psychologic measures. Univariate regression analyses were used to test for significant predictors of the HADS total score and the HADS Anxiety and Depression subscale scores.

RESULTS

Sample

Two hundred thirty-one patients initially were registered to the study from September 1998 to January 2003. However 39 patients (16.9%) were not randomized because of patient refusal (9.5% [n = 22 patients]), illness (2.6% [n = 6 patients]), death (7.7% [n = 4 patients]), ineligibility (2.6% [n = 6 patients]), and loss to follow-up (<1% [n = 1 patient]). One hundred ninety-two patients were randomized to the 2 treatment arms, however only 189 patients (96 patients in the TM + EM group and 93 patients in the EM group) were evaluated at baseline (3 patients cancelled, 1 patient was ineligible, 1 patient was too ill, and 1 patient was not evaluable because of an interviewer error). One hundred thirty-one patients (69 patients in the TM + EM group and 66 patients in the EM group) completed both the baseline (study entry) assessment and the 6-month assessment. Reasons for attrition of 58 patients from study entry to the 6-month assessment were illness (41.4%), death (22.4%), refusal (15.5%), inability to reach the patient (13.8%), interviewer error (5.2%), and anxiety (1.7%). There were only 73 patients (38.6%) who completed the baseline, 6-month, and 9-month assessments (39 patients in the TM + EM group and 34 patients in the EM group). Of the patients who were not assessed at 9 months (n = 58 patients), the reasons for attrition were illness (34%), death (7%), and inability to reach the patient (34%). Consequently, the analyses reported below involve only the 131 patients who were evaluated at study entry and at 6 months, when the TM sessions had been completed. The 131 patients who completed the study entry and 6-month evaluations were compared with the 58 patients who were not evaluated at 6 months. No significant differences with regard to sociodemographic or disease characteristics were observed. However, patients who were evaluated both at study entry and at 6 months had significantly lower HADS Depression subscale scores (P = .0329), higher EORTC Emotional Functioning subscale scores (P = .0069), and higher EORTC Role Functioning subscale scores (P = .0249) at baseline compared with the patients who were not assessed at both time points.

The sample's characteristics are presented in Table 1. There were no significant differences in sociodemographic characteristics, disease site characteristics, or psychosocial factors between treatment arms with the exception of pain. Pain scores tended to be higher in the TM + EM arm than in the EM arm (26.06 vs. 17.44, respectively; P = .0502).

Table 1. Subset: Patients' Medical and Sociodemographic Characteristics Assessed at Baseline and at 6 Months

Characteristic	Percent (No. of patients)
	TM and EM	EM
TM indicates telephone monitoring; EM, educational materials; SD, standard deviation. * Patients checked all items that applied.
Total no. of patients	100 (69)	100 (62)
Disease site
Breast	36 (25)	35 (22)
Colon/rectum	29 (20)	32 (20)
Prostate	35 (24)	32 (20)
Sex
Male	52 (36)	53 (33)
Female	48 (33)	47 (29)
Race
White	83 (57)	90 (56)
Black	13 (9)	8 (5)
Other	4 (3)	2 (1)
Age
65–69 y	32 (22)	29 (18)
70–79 y	55 (38)	56 (35)
≥80 y	13 (9)	15 (9)
Mean ± SD (y)	73 ± 5.7	74 ± 6.8
Marital status
Married	62 (43)	56 (35)
Separated/divorced	4 (3)	15 (9)
Widowed	30 (21)	24 (15)
Single, never married	3 (2)	5 (3)
Current employment*
Part-time or full-time	16 (11)	16 (10)
Homemaker	16 (11)	8 (5)
Retired	75 (52)	77 (48)
Disabled	3 (2)	0 (0)
Other	0 (0)	3 (2)
Karnofsky performance status
90–100%	58 (40)	66 (41)
70–89%	32 (22)	16 (10)
50–69%	1 (1)	6 (4)
Unknown	9 (6)	11 (7)

Impact of TM and EM on Patient Distress

Patients in the TM + EM arm had significantly lower anxiety and depression than patients in the EM arm (HADS total score, P < .0001; HADS Depression subscale, P = .0004; HADS Anxiety subscale, P < .0001) at 6 months when controlling for study entry levels (Table 2). The mean HADS total and subscale scores at baseline and at 6 months for the 2 study arms went in opposite directions over time (i.e., as distress decreased in the TM + EM arm, distress increased in the EM arm) (Fig. 1). The other measures of psychologic distress (Geriatric Depression Scale, EORTC Emotional Functioning subscale) did not reveal differences between the 2 study arms.

Figure 1. Total scores on the Hospital Anxiety and Depression Scale (HADS) are illustrated by treatment arm. EM indicates educational materials; TM, telephone monitoring; pre, before treatment; post, after treatment.

Table 2. Physical Functioning and Psychological Distress: The Telephone Monitoring and Educational Materials Group versus the Educational Materials Alone Groups

Measure	No. of patients	Mean (SD)		P (ANCOVA)
		Baseline	6 months
SD indicates standard deviation; ANCOVA, analysis of covariance; HADS, Hospital Anxiety and Depression Scale; TM, telephone monitoring; EM, educational materials; EORTC, European Organization for Research and Treatment of Cancer; QoL, quality of life; OARS, Older American Resources and Services Questionnaire. * High numbers indicate a worse QoL. † High numbers indicate a better QoL.
HADS total score*
TM and EM	69	7.49 (6.38)	6.01 (4.95)	<.0001
EM	61	6.41 (5.57)	8.20 (5.59)
HADS Depression subscale*
TM and EM	69	3.65 (3.41)	3.20 (2.92)	.0004
EM	61	2.95 (2.65)	4.08 (2.85)
HADS Anxiety subscale*
TM and EM	69	3.84 (3.79)	2.81 (2.65)	<.0001
EM	61	3.46 (3.40)	3.25 (3.39)
Geriatric Depression*
TM and EM	68	3.35 (3.74)	2.98 (3.19)	.24
EM	60	2.83 (3.08)	3.15 (3.18)
EORTC overall QoL†
TM and EM	67	61.45 (21.18)	64.79 (20.71)	.24
EM	61	68.99 (18.17)	65.55 (20.40)
EORTC Emotional Functioning†
TM and EM	68	83.07 (17.85)	84.44 (15.30)	.15
EM	61	85.90 (14.63)	82.91 (16.18)
EORTC Physical Symptoms*
TM and EM	68	23.28 (13.12)	19.46 (12.53)	.25
EM	61	19.83 (12.78)	19.58 (12.29)
EORTC Physical Functioning†
TM and EM	68	71.18 (21.69)	65.15 (22.09)	.28
EM	61	72.79 (24.77)	69.67 (23.50)
OARS Comorbid Condition Interference with Functioning*
TM and EM	69	0.013 (0.022)	0.011 (0.021)	.99
EM	62	0.013 (0.025)	0.011 (0.023)
Current use of mental health services
TM and EM	68	10.3%	8.8%	.34
EM	59	16.95%	11.86%

Twenty-eight percent of patients in the TM + EM group (n = 19 of 69 patients) scored above cutoff levels on the HADS total score (≥15) and on the HADS Depression or Anxiety subscales (≥8) both at study entry and at the 6-month evaluation. However, when all TM sessions and evaluation interviews were considered, the percent of patients in the TM + EM group that scored above HADS cutoff levels on the HADS total score or on the Anxiety or Depression subscales increased to 42% (n = 29 of 69 patients). This was significantly different from the EM group, in which 24% of patients (15 of 62 patients) scored above the HADS cutoff levels when considering both the baseline assessment and the 6-month assessment (P = .0414).

There were numerous referrals (‘alerts’) to the oncology nurse for physical symptoms over the course of the study period, primarily in the TM + EM arm, due to the TM sessions. There were 50 alerts caused by physical problems (45 patients in the TM + EM arm and 5 patients in the EM arm), representing 32 patients (27 patients in the TM + EM arm and 5 patients in the EM arm). The alerts for physical symptoms mostly were for pain (26 patients in the TM + EM arm and 1 patient in the EM arm), weakness (17 patients in the TM + EM arm and 2 patients in the EM arm), insomnia (14 patients in the TM + EM arm and 2 patients in the EM arm), and appetite loss (9 patients in the TM + EM arm and 2 patients in the EM arm). There were 7 alerts for social support problems in 7 patients (4 patients in the TM + EM arm and 3 patients in the EM arm). There were no significant differences noted between the 2 groups with regard to the percent of physical symptom alerts and, because the number of alerts for social support problems was so low, statistical analyses could not be performed.

There were 53 instances of some referral by the oncology nurse to either a psychiatrist/psychologist, social worker/social services, oncologist, or oncology nurse (51 patients in the TM + EM arm and 2 patients in the EM arm) for physical, psychologic, and social support problems, representing 29 patients (27 patients in the TM + EM arm and 2 patients in the EM arm).

When we examined for significant correlations between the HADS total score and other variables, the only significant univariate correlations with higher HADS total scores were EM (P = .001), greater GSRE scores (P = .0187), and higher Karnofsky performance status (≥70 vs. <70; P = .0047 [it is worth noting that only 5 patients had Karnofsky performance status scores <70]). There were no other significant univariate correlations noted between HADS total scores and other variables. Because of the interaction between treatment arm and baseline measures of the HADS total score, multivariate regression analyses could be conducted only separately for each arm of the study.

Patient Satisfaction with TM + EM and EM Conditions

Eighty-eight percent of patients in the TM + EM arm reported ‘good’ to ‘excellent’ overall satisfaction with the program, and at least 60% reported satisfaction with individual aspects of the program (Table 3). However, only 14% of patients in the TM + EM arm and 37% of patients in the EM arm rated the program ‘very helpful’ to ‘extremely helpful’ (P = .0064) in coping with an important problem. When patients were asked about what was most helpful in coping with an important problem, they reported talking about problems over the phone (TM + EM arm: 41% [n = 28 of 68 patients]; EM arm: 26% [n = 16 of 62 patients]; P = .0956), feeling cared for between office visits (TM + EM arm: 60% [n = 41 of 68 patients]; EM arm: 61% [n = 38 of 62 patients]), the telephone monitor's personality (TM + EM arm: 51% [n = 35 of 68 patients]), and getting help for problems (TM + EM arm: 22% [n = 15 of 68 patients]).

Table 3. Satisfaction with the Telephone Monitoring and Educational Materials Program

Response	Percentage of patients (No.)		P
	TM and EM	EM
TM indicates telephone monitoring; EM, educational materials; NA, not applicable. * Interviewer error: the patient was randomized into the EM arm but was followed as a member of the TM and EM arm.
Overall satisfaction with the program			.228
Good/excellent	88 (59)	74 (46)
Satisfactory	10 (7)	21 (13)
Fair/poor	1 (1)	5 (3)
How much the program helped with important problems			.018
Very/extremely	14 (9)	37 (22)
Somewhat	41 (26)	28 (17)
A little/not at all	44 (28)	35 (21)
Talking about problems over the telephone			.771
Good/excellent	77 (51)	74 (46)
Satisfactory	20 (13)	23 (14)
Fair/poor	3 (2)	3 (2)
Topics discussed			.864
Good/excellent	79 (53)	79 (49)
Satisfactory	21 (14)	19 (12)
Fair/poor	0 (0)	2 (1)
Quality of educational materials			.241
Good/excellent	63 (41)	68 (40)
Satisfactory	37 (24)	25 (15)
Fair/poor	0 (0)	7 (4)
Evaluation interviews			.258
Good/excellent	76 (51)	84 (52)
Satisfactory	19 (13)	15 (9)
Fair/poor	1 (1)	2 (1)
Interviewer error*	3 (2)	0 (0)
Skill of monitor			—
Good/excellent	88 (60)	NA
Satisfactory	9 (6)
Fair/poor	0 (0)
Interviewer error*	3 (2)
Monitor caring about the patient			—
Good/excellent	90 (60)	NA
Satisfactory	7 (5)
Fair/poor	0 (0)
Interviewer error*	3 (2)
No. of times monitor called			—
Good/excellent	84 (56)	NA
Satisfactory	13 (9)
Fair/poor	0 (0)
Interviewer error*	3 (2)
Length of monitoring call			—
Good/excellent	81 (54)	NA
Satisfactory	15 (10)
Fair/poor	1 (1)
Interviewer error*	3 (2)
Skill of referral nurse			—
Good/excellent	64 (24)	NA
Satisfactory	3 (1)
Fair/poor	0 (0)
RN did not evaluate	11 (4)
Missing	24 (9)

Among all patients who rated the program from ‘very helpful’ to ‘extremely helpful,’ the TM + EM group had significantly less overall distress (HADS, P = .0029), depression (HADS Depression subscale, P = .0176), and anxiety (HADS Anxiety subscale, P = .0120) at 6 months from baseline levels compared with patients in the EM arm (Table 4). Therefore, although patients the EM group reported that the program was ‘very helpful’ to ‘extremely helpful’ significantly more often than patients in the TM + EM group (P = .0064), the HADS scores reflected that patients in the TM + EM group had a significantly greater psychologic benefit from the intervention than patients in the EM group.

Table 4. Psychologic Distress of Patients Who Rated the Program Very Helpful to Extremely Helpful by Treatment Arm

Psychological distress	Mean (SD)				P (ANOCVA)
	TM and EM (N = 9)		EM (N = 22)
	Baseline	6 months	Baseline	6 months
SD indicates standard deviation; TM, telephone monitoring; EM, educational materials; ANCOVA, analysis of covariance; HADS, Hospital Anxiety and Depression Scale. * Higher scores indicate greater distress.
HADS total*	5.44 (5.27)	3.89 (3.95)	6.27 (4.33)	8.64 (4.77)	.0029
HADS Depression*	3.44 (2.96)	2.89 (2.37)	2.91 (1.90)	4.41 (2.38)	.0176
HADS Anxiety*	2.00 (2.87)	1.00 (2.00)	3.36 (2.75)	4.23 (3.10)	.0120

DISCUSSION

The highly significant differences in psychologic distress between the TM + EM arm and the EM arm from baseline to 6 months, as measured by the HADS, indicate the importance of scheduled monitoring of patients' psychologic distress, even when that monitoring is conducted by nonhealth professionals using standardized questionnaires. The current findings suggest that there was a tendency for patients in the TM + EM group to be comforted by just talking to someone about their problems compared with patients in the EM group. Given the various barriers that prevent patients from discussing psychosocial problems with their oncologists, providing patients with the opportunity to talk with someone about their problems may be an under-appreciated aspect of the success of the program.

Our findings are in contrast to those of Maunsell et al.,36 who observed no significant differences in psychologic distress between a TM group and a nonmonitored group of patients with breast cancer. Maunsell et al.36 theorized that a minimal intervention, such as TM, may not have provided a significant improvement over usual care, because usual care included a social worker's assessment of patients' need for mental health and social services at the time they were first seen in the clinic. It is uncommon for institutions to routinely provide patients with a social worker's assessment at the beginning of their treatment. Therefore, the patients in the TM + EM arm of our study most likely represented a sample with undetected psychosocial needs at the beginning of their treatment compared with the study reported by Maunsell et al.36 and therefore were in greater need of our intervention.

There were substantial numbers of patients in the TM + EM group who scored above cutoff levels of psychologic distress over the course of the 6-month monitoring period (42%), nearly 33% more than those who scored above cutoff levels at study entry and at the 6-month assessment. Perhaps this is a reflection of the differences in the ‘density’ of the intervention. By monitoring patients monthly over 6 months, in addition to the study entry and 6-month assessments, there was a greater opportunity to identify psychologically distressed patients in the TM + EM group. The implication of these findings is that, to identify those in psychologic distress, patients need to be assessed multiply over time.

In terms of the study's limitations, attrition was a major problem throughout the current study, which was not unexpected because the study involved patients with advanced disease. However, attrition was similar for both treatment arms. Testing for sample bias largely revealed that the patients who had both the baseline assessment and the 6-month assessment had higher scores on quality-of-life measures than patients who were not assessed at 6 months. Therefore, the findings of the current study may not apply to all older cancer patients with advanced-stage disease.

Second, results across the 3 measures of psychologic distress were inconsistent. Only the HADS demonstrated markedly significant differences between the TM + EM arm and the EM arm, whereas the Geriatric Depression Scale and the EORTC Emotional Functioning subscale did not. It is possible that the latter 2 measures were not sufficiently sensitive to psychologic state to reveal more subtle differences between the TM + EM arm and the EM arm compared with the HADS. The EORTC Emotional Functioning subscale consists of 4 items, whereas the HADS consists of 14 items. Although the Geriatric Depression Scale consists of 15 items, all the response categories were dichotomous (yes or no), which decreased the likelihood of identifying more subtle differences between the treatment arms.

In conclusion, the findings of the current study suggest that, with multiple assessments, there are substantial numbers of older patients who report distress that is suggestive of a possible psychiatric disorder. These findings support the use of multiple telephone monitoring sessions in older patients to assess their distress as a mechanism for referral to an oncology nurse for further referral to treatment to reduce their distress. This intervention also would be done at minimal cost, because the alternative would be for the oncology nurse to conduct these monitoring calls, which would cost much more compared with the use of college graduates as telephone monitors. The intervention's simplicity suggests that it could be disseminated widely to other cancer centers.

Acknowledgements

We thank all the patients who participated in this study by sharing their thoughts and feelings and Ms. Florence Bernstein, who generously donated her time to serve as one of the telephone monitors in this study. We also acknowledge the contribution of the following investigators to this project: Nancy Bartlett, MD (Washington School of Medicine); Thomas Reid, MD (Walter Reed Army School of Medicine); Hyman Muss, MD (Vermont Cancer Center); Jean Donnelly, PhD (Memorial Sloan-Kettering Cancer Center); and Enid Zuckerman, MSW (Memorial Sloan-Kettering Cancer Center) for their contribution to the study.