Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Nov 1.
Published in final edited form as: Br J Health Psychol. 2009 Jan 24;14(Pt 4):701–715. doi: 10.1348/135910708X400462

Trait anxiety as an independent predictor of poor health-related quality of life and post-traumatic stress symptoms in rectal cancer

Stephen L Ristvedt 1,*, Kathryn M Trinkaus 2
PMCID: PMC2756319  NIHMSID: NIHMS108915  PMID: 19171084

Abstract

Objectives

To determine the influence of trait anxiety on patient reports of health-related quality of life (HRQoL) and post-traumatic stress symptoms (PTSS) in a sample of rectal cancer survivors.

Design

Eighty patients who had been diagnosed with rectal cancer were assessed at two points in time in a longitudinal study.

Methods

At Time 1, soon after initial treatment, participants completed the State-Trait Anxiety Inventory and the Temperament and Character Inventory Harm Avoidance scale, which were combined into a composite measure of trait anxiety. At Time 2, 2-5 years following Time 1, participants were assessed for HRQoL using the Functional Assessment of Cancer Therapy-Colorectal scale (FACT-C) and for PTSS using the Impact of Event Scale-Revised (IES-R).

Results

HRQoL and PTSS were generally favourable on average, although many of the patients reported faring poorly. Higher levels of trait anxiety were predictive of poorer scores on all of the FACT-C and the IES-R total and subscale measures. More severe faecal incontinence was associated with poorer scores on the FACT Emotional well-being subscale, the FACT-Colorectal Cancer Scale, and all of the IES-R scales. Males were more likely than females to have poorer scores on the FACT Social well-being subscale, and those patients who were further out from active treatment had more favourable scores on the FACT-Colorectal Cancer Scale. The presence of a colostomy did not impact HRQoL or PTSS.

Conclusion

Trait anxiety had a significant influence on HRQoL and PTSS several years following diagnosis and treatment of rectal cancer.

The number of people who can count themselves as cancer survivors has increased dramatically, due to advances in early detection and successful treatment, the ageing of the population, and broader definitions of ‘survivorship’ (Rowland, Hewitt, & Ganz, 2006). With these changes has come increased interest in the ways in which cancer and its treatment impacts survivors, in terms of both objective and subjective outcomes. Objective outcomes include side-effects of treatment, long-term and late effects of treatment, and economic burden. Subjective (i.e. patient-reported) outcomes include treatment satisfaction, health-related quality of life (HRQoL) and, in some, symptoms of post-traumatic stress (Aziz, 2007; Gotay, Lipscomb, & Snyder, 2005; Kangas, Henry, & Bryant, 2002). Interestingly, objective and subjective outcomes often do not coincide. A common finding is that HRQoL among cancer survivors is more favourable than would be expected, often comparable to healthy controls, even when deleterious biomedical effects persist (Gotay et al., 2005; Schwartz & Sprangers, 2002; Rapkin & Schwartz, 2004). However, most studies report only mean values, which do not reflect the fact that HRQoL data are typically skewed. Although there may be a clustering of patients towards the favourable end of the distribution, the longer tail represents a great number of patients who report relatively poor HRQoL (Curran et al., 2000; Cella et al., 2003). Also, those reporting poor HRQoL might not differ appreciably from their higher HRQoL counterparts in terms of objective health status (Wilson & Cleary, 1995). The question we address in this study is this: why do some cancer survivors report poor HRQoL when their apparent medical status is not different from their peers?

One explanation for the mismatch between subjective and objective outcomes could be that subjective measures are inherently open to the influence of patients’ individual perspectives and predispositions, while objective biomedical outcomes are not (Gotay et al., 2005; Schwartz & Sprangers, 2002; Rapkin & Schwartz, 2004; Cella et al., 1993). For example, one possible influence on HRQoL is patients’ affective disposition, which would not be expected to significantly influence biomedical outcomes related to cancer. There is a broad base of evidence demonstrating that negative affective traits (i.e. tendencies towards depression, anxiety, and pessimism) have a potent influence on self-reported physical symptoms and well-being, often outweighing the influence of objective markers of functioning (Pennebaker, 2000; Watson & Pennebaker, 1989). In the present study, we hypothesized that a commonly measured form of negative affect (trait anxiety) would predict self reports of well-being (HRQoL and post-traumatic stress symptoms, PTSS), while controlling for objective medical outcomes (stage of disease, presence of colostomy, and severity of faecal incontinence).

Only a few studies have investigated relationships between various measures of negative affect and HRQoL. In the original validation study of the Functional Assessment of Cancer Therapy measure (FACT; Cella et al., 1993), the FACT-G summary score was inversely related to scores on the Brief Profile of Mood States (BPOMS; Cella et al., 1987) as well as to scores on the Taylor Manifest Anxiety Scale (Taylor, 1953). In a study of the FACT-Colorectal measure (FACT-C; Ward et al., 1999), all of the summary and subscale scores were inversely related to the BPOMS (Cella et al., 1987) and the Profile of Mood States - Short Form (McNair, Lorr, & Droppleman, 1992). In a study of patients with colorectal cancer (Yoo, Kim, Eremenco, & Han, 2005), all subscale and total scores on the FACT-C (Ward et al., 1999) were inversely related to scores on both the BPOMS(Cella et al., 1987) and the Neuroticism scale of the Eysenck Personality Questionnaire (Eysenck & Eysenck, 1976). And lastly, in a study of 60 cancer patients conducted at the time of surgical follow-up, patients’ global HRQoL as measured by the Quality of Life Questionnaire (Aaronson et al., 1993) was inversely related to their scores on the Negative Affect scale of the Positive and Negative Affect Schedule (Watson, Clark, & Tellegen, 1988), but was not correlated with physicians’ ratings of their HRQoL or overall physical condition (Koller et al., 1996).

Another patient-reported outcome that has received increased interest is post-traumatic stress disorder (PTSD). With the introduction of the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1994), the criteria for PTSD were altered to allow the inclusion of life-threatening illnesses as traumatic events. Since the publication of DSM-IV in 1994, scores of articles have been published on the relation between various serious illnesses and either PTSD or PTSS. Of the studies that have been done with regard to cancer, most have focused on either breast cancer or on childhood cancer survivors and their parents, while relatively few studies have looked at these issues in colon and rectal cancers (cf. Salsman, Segerstrom, Brechting, Carlson, & Andrykowski, 2009). While most cancer patients never meet diagnostic criteria for PTSD, a high proportion of patients experience subclinical PTSS following diagnosis (Kangas et al., 2002; Deimling, Kahana, Bowman, & Schaefer, 2002). Although PTSS will decline considerably within the first few months, such symptoms can trouble some patients for years (Kangas et al., 2002). Similar to the discrepancies regarding HRQoL, neither the incidence of PTSD nor the magnitude of PTSS is clearly related to biomedical aspects of the disease, but they have been linked to various measures of negative affect (Kangas et al., 2002; Kangas, Henry, & Bryant, 2005; Palmer, Kagee, Coyne, & DeMichele, 2004). In a longitudinal study of cancer patients, scores on the Beck Depression Inventory (Beck, Steer, & Brown, 1996) and the Trait scale of the State-Trait Anxiety Inventory (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1970), administered within 1 month of diagnosis, were correlated with both PTSD caseness and PTSS severity, which were assessed at 6 months post-diagnosis (Kangas et al., 2005).

All of the studies just described showed that greater levels of negative affect were related to either poorer HRQoL or more problems with PTSS. However, most of those studies were cross-sectional and most used measures of negative mood state rather than trait negative affect, which has made the direction of causality difficult to infer. The purpose of the present longitudinal study was to examine the influence of stable negative affective disposition on both HRQoL and PTSS in a sample of rectal cancer survivors while controlling for objective medical outcomes. With regard to objective outcomes, the treatment of rectal cancer often entails permanent biomedical ramifications, including chronic bowel, urinary, and sexual dysfunction (Vironen, Kairaluoma, Aalto, & Kellokumpu, 2006). In addition, appearance and function are both significantly altered for those patients who receive a permanent colostomy (Bleday & Garcia-Aguilar, 2007). Nevertheless, several reports indicate that mean HRQoL scores for rectal cancer survivors improve within 3-6 months post-surgery to the point that they are comparable to those reported in community studies (Camilleri-Brennan & Steele, 2001; Neuman et al., 2007; Rauch, Miny, Conroy, Neyton, & Guillemin, 2004). Also, there is no clear evidence that individuals with a colostomy have worse HRQoL than those without (Cornish et al., 2007; Pachler & Wille-Jorgensen, 2005). On the other hand, faecal incontinence and chronic diarrhoea have been shown to have a negative effect on certain domains of HRQoL, particularly social functioning (Vironen et al., 2006).

But again, even though mean HRQoL among rectal cancer survivors is comparable to population controls, these distributions are usually skewed, with the long tail representing a great number of survivors reporting poor HRQoL (Camilleri-Brennan & Steele, 2002; Engel et al., 2003; Schmidt, Bestmann, Kuchler, Longo, & Kremer, 2005a, b). No previous studies that we could find have looked at the role of PTSD or PTSS in rectal cancer. In the present study, we wanted to investigate whether stable negative affective disposition is related to poor HRQoL or problems with PTSS. With the recent attention given to the specific role of anxiety in issues related to perceptions of health and illness (Cameron, 2003), we investigated longitudinally the relative influence of trait anxiety, demographic and medical variables on HRQoL and PTSS in a group of rectal cancer survivors. It was hypothesized that trait anxiety would predict self-reported HRQoL and PTSS, independent of other characteristics of the person or their disease.

Method

Participants and procedure

A longitudinal study was conducted in which rectal cancer patients were assessed at two points in time. The primary predictor (trait anxiety) was assessed at Time 1, while the primary outcomes (HRQoL and PTSS) were assessed at Time 2. Participants were patients who received treatment for primary rectal tumours at Washington University School of Medicine. At Time 1, between July 1998 and July 2001, 123 patients were recruited at the time of their follow-up visit after initial surgical intervention by one of three clinic nurses. Participants were given a set of paper-and-pencil questionnaires that could be completed at home and returned by mail. Assessment involved gathering information regarding sex, age, race/ethnicity, and education. Also, two measures of trait anxiety were included (see descriptions in the Measures section). Any additional medical treatment - whether it included chemotherapy, radiation therapy, or both - occurred subsequent to the assessment, with some of the patients eventually receiving a permanent colostomy.

The participants were contacted again at Time 2, between July 2003 and February 2004, and invited to participate in a telephone interview to assess HRQoL and PTSS. The time that elapsed between Time 1 and 2 assessments ranged from about 2 to 5 years. Potential participants were sent a letter from their treating surgeon to introduce the study and to inform them that they would be contacted by telephone. A computer-assisted telephone interview system was used by trained interviewers to administer three standardized measures: the FACT-C (Ward et al., 1999), as a measure of HRQoL, the Impact of Event Scale - Revised (IES-R; Weiss & Marmar, 1997), as a measure of PTSS, and the Faecal Incontinence Severity Index (FISI; Rockwood et al., 1999). Chart reviews were conducted to determine the grade and stage of patients’ tumours as well as the types of treatments received. A total of 80 patients participated in the assessment at Time 2, out of the original 123 patients who had participated at Time 1. All aspects of this study had approval from the Institutional Review Board at Washington University.

Measures

Trait anxiety

Trait anxiety was measured at Time 1 with two questionnaires. The first was the Trait scale of the State-Trait Anxiety Inventory (Spielberger et al., 1970), which requires subjects to estimate how much of the time they typically experience various symptoms of anxiety. The Trait scale consists of 20 statements that subjects use to describe how they generally feel, using a 4-point frequency scale ranging from almost never to almost always, resulting in total scores that can range from 20 to 80. This scale has been found to be internally consistent as well as stable over time, with coefficient alphas ranging from .72 to .96 and test-retest correlations ranging from .82 to .94 (Barnes, Harp, & Jung, 2002).

Patients also completed the short form (144 items) of the Temperament and Character Inventory (TCI; Cloninger, Przybeck, Svrakic, & Wetzel, 1994). The TCI Harm Avoidance scale was the scale of interest, since it measures stable individual differences in sensitivity to signals of possible threat (Heath, Cloninger, & Martin, 1994). The Harm Avoidance scale consists of 20 true-false statements in which subjects are asked to determine ‘which choice best describes you’. The Harm Avoidance scale has also demonstrated reasonable stability, with test-retest correlations ranging from .76 to .84 over the course of two years (Heath et al., 1994). The Harm Avoidance scale has been found to correlate with several other measures of negative affective traits including the Trait scale of the State-Trait Anxiety Inventory (Carver & White, 1994; Costa & McCrae, 1992; Eysenck & Eysenck, 1976; Spielberger et al., 1970). Six patients missed two or fewer items on one or the other of the two scales, in which case scale scores were calculated as the mean of the nonmissing items multiplied by the number of items in the scale.

Since both the Trait scale of the State-Trait Anxiety Inventory and the TCI - Harm Avoidance scale are quite stable over time, they are considered to be measures of dispositional tendencies towards anxiety. Even though the two measures were highly correlated in our sample (rank correlation = 0.65), they can be seen as complementary measures of trait anxiety, in that the Trait scale of the State-Trait Anxiety Inventory leans towards the somatic aspects of anxiety, while the TCI - Harm Avoidance scale leans towards the cognitive aspect (i.e. worry). The two measures were thus combined into one composite measure by summing their standardized scores. A composite measure such as this is appropriate when the research question involves a broad theoretical construct, i.e. trait anxiety, rather than the idiosyncrasies of the individual scales (Zelenski & Larsen, 2002).

Health-related quality of life

HRQoL was assessed at Time 2 with the FACT-C (Ward et al., 1999), which is a two-part measure of HRQoL. The first component (FACT-G) is a measure of overall HRQoL and is made up of four subscales that cover physical (7 items), social (7 items), emotional (6 items), and functional (7 items) well-being. The second component is the Colorectal Cancer Scale (7 items), which assesses the presence of symptoms that are often experienced by colorectal cancer survivors (e.g. cramps, diarrhoea). Each item requires the patient to indicate ‘how true each statement has been for you during the past 7 days’, with five options ranging from ‘Not at all’ to ‘Very much’. Scores on each subscale range from 0 to 28 (0-24 on the Emotional subscale), with higher scores indicating more favourable functioning. One patient missed 1 item on the Social well-being subscale, so the score was calculated as the mean of the nonmissing items multiplied by the number of items in the subscale.

Post-traumatic stress symptoms

The assessment of PTSS at Time 2 was done with the IES-R (Weiss & Marmar, 1997), which measures the residual psychological impact of having experienced a traumatic or life-threatening situation. The instructions were tailored for this study, so that participants were asked to ‘indicate how distressing each difficulty has been for you during the past 7 days with respect to your rectal cancer and the treatment you had’. The avoidance subscale (8 items) asks questions regarding the avoidance of situations related to the experience; the Hyperarousal subscale (8 items) asks questions regarding episodes of physiological arousal related to thoughts about the event; the Intrusiveness subscale (7 items) asks questions regarding intrusive recollections of the experience. Scores for each subscale are calculated as the mean of the non-missing individual items (i.e. 0-4), with higher scores indicating more severe difficulties. The IES (Horowitz, Wilner, & Alvarez, 1979), which is the predecessor of the IES-R, has been used most often in studies of PTSS in cancer (Kangas et al., 2002). However, we decided to use the IES-R since it includes a scale measuring hyperarousal, which is one of the diagnostic criteria for PTSD (American Psychiatric Association, 1994) but which was not included in the original IES.

Faecal incontinence

The FISI (Rockwood et al., 1999), also administered at Time 2, measures both type (gas, mucus, liquid stool and solid stool) and frequency of incontinence. For each type of incontinence, the patient is asked to indicate how often in the past month they experienced accidental bowel leakage. An overall severity rating is calculated by weighting the component scores according to patients’ estimations of severity derived from the original validation study (Rockwood et al., 1999). Scores can range from 0 to 61, with higher scores indicating greater severity.

Statistical analysis

The primary aim of the statistical analysis was to determine the influence of trait anxiety on both HRQoL and PTSS, while accounting for demographic and medical variables. As expected, scores on the outcome measures were skewed, with the majority of patients clustered at the more favourable end of the ranges of all of the FACT and IES-R subscales. A normal distribution could not be approximated by transformation for any of the scales, so each scale was divided into two groups by median split. Logistic regression was used as the primary analytic tool. Logistic models were constructed to predict which patients had poorer scores on the FACT-C or IES-R total and subscale scores. In addition to the trait anxiety composite score, the covariates of interest included demographic information (age at Time 2, sex, and years of education) and medical information (stage of disease at diagnosis, presence of colostomy, and current severity of faecal incontinence). Because there is evidence to suggest that the passage of time following colon resection brings amelioration of HRQoL (Camilleri-Brennan & Steele, 2002), we also included the time elapsed between the Time 1 and Time 2 assessments.

Statistical analyses were conducted in two stages, as has been recommended for HRQoL research (Curran et al., 2000), by first conducting analyses with summary scores as the outcomes, followed by analyses with the individual subscale scores as the outcomes. So, we first included all of the independent variables simultaneously in analyses of the main summary scores: the FACT-G, the FACT-Colorectal Cancer Scale, and the IES-R total score. The second set of analyses was conducted with each of the individual subscale measures as well as the summary scores as the outcome variables. In these analyses, best-fit models were constructed in which covariates were included based on measures of goodness-of-fit (e.g. the Hosmer-Lemeshow test; Hosmer & Lemeshow, 2000) and of information (e.g. Akaike’s Information Criterion; Collett, 2003).

Results

Participants

A total of 150 patients were originally approached to participate in this study at Time 1. Twenty-seven of those patients declined, either because of time constraints or simple lack of interest. Of the 123 patients who participated at Time 1, 16 were deceased prior to the assessment at Time 2, 14 could not be located, and 13 declined to participate. We compared these three groups with the 80 patients who participated in both Time 1 and 2 assessments on age, educational level, trait anxiety, and stage of disease at diagnosis. The 16 patients who died prior to the Time 2 interview tended to have a more severe stage of disease than the participants (Mann-Whitney z = 1.779, two-tailed p = .075), as did the 14 patients who could not be located (Mann-Whitney z = 2.840, two-tailed p = .005), so many of them may well have died prior to our attempts to contact them. Those patients who declined to participate were significantly older than the participants at the time of contact (74.7 vs. 67.5, t = 2.084, p = .040). The span of time that elapsed from Time 1 to Time 2 ranged from 112.1 weeks to 260.6 weeks (approximately 2-5 years), with a mean of 203.2 weeks (SD = 39.5). Table 1 shows the socio-demographic and clinical characteristics of the 80 patients who participated in both assessments.

Table 1.

Sociodemographic and clinical characteristics

Characteristic
Sex, n (%)
 Female 35 (43.8)
 Male 45 (56.2)
Age, mean years ± SD (range) 67.5 ± 12.0 (29.1-88.3)
Race/ethnicity, n (%)
 Caucasian 73 (91.2)
 African-American 6 (7.5)
 Hispanic 1 (1.2)
Educational level, mean years ± SDa 12.8 ± 3.5
 < High school, n (%) 15 (18.8)
 High school graduate, n (%) 29 (36.2)
 Some college or college graduate, n (%) 22 (27.5)
 Post-graduate, n (%) 10 (12.5)
Stage of disease, n (%)b
 Stage I 37 (46.2)
 Stage II 24 (30.0)
 Stage III 15 (18.8)
 Stage IV 1 (1.2)
Permanent colostomy, n (%) 19 (23.8)
Open in a new tab
a

Education unavailable on 4 patients

b

Stage unavailable on 3 patients.

Measures of HRQoL and PTSS

Table 2 shows the descriptive statistics for the FACT-C and IES-R subscale and total scores, with the means and medians for all participants as well as the means and ranges for the groups above and below the medians. As can be seen, all of these measures were skewed with a clustering of scores on the more favourable ends (i.e. higher scores for FACT scales and lower scores for IES-R scales). The skewness is reflected in two ways. First, the mean is less than the median for all of the FACT scales, while the mean is greater than the median for all of the IES-R scales. Second, the range of scores is much greater on the skewed side of the median for all of the scales (e.g. a range of 10-24 below the median on the Physical well-being subscale, but a range of 25-28 at or above the median).

Table 2.

Descriptive statistics for FACT and IES-R, total and after median split

All participants
 Median split [mean (range)]
Measure Mean Median Below median Above median
FACT-G subscales
 Physical 23.6 25 19.5 (10-24) 26.8 (25-28)
 Social 22.9 24 18.6 (5-23) 26.3 (24-28)
 Emotional 20.5 22 16.6 (5-21) 23.5 (22-24)
 Functional 20.5 21 15.2 (4-20) 25.0 (21-27)
FACT-G total 87.4 92 73.4 (38-90) 100.4 (92-107)
Colorectal Cancer Scale 20.7 22 16.7 (9-21) 24.5 (22-28)
IES-R subscales
 Avoidance 0.5 0 0.00 (0.0-0.0) 1.10 (0.1-3.3)
 Hyperarousal 0.35 0 0.00 (0.0-0.0) 0.90 (0.1-3.1)
 Intrusiveness 0.37 0.1 0.02 (0.0-0.1) 0.86 (0.3-2.7)
IES-R total 1.22 0.4 0.08 (0.0-0.4) 2.61 (0.5-8.6)
Open in a new tab

Less favourable scores are below the median for the FACT and above the median for the IES-R.

Predictors of HRQoL and PTSS

As noted in Table 1, 7 patients were missing either education or tumour stage information. Inspection revealed that the missing values did not affect the joint distribution of covariates used in modelling HRQoL, and there was no evidence that education or stage was missing for any reason related to HRQoL. As a result, the values were considered to be missing at random. The logistic models were thus developed using the 73 patients who had complete data.

The results from the first stage of data analysis are shown in Table 3, which includes the full multivariate logistic models for the FACT-G summary score, the FACT-Colorectal Cancer Scale, and the IES-R total score. These models include all of the covariates in predictions of those patients who fell on the poorer side of the medians of the outcome measures. Trait anxiety and faecal incontinence were the strongest predictors in all of these models, although faecal incontinence did not reach the customary .05 level in the FACT-G model. In addition, the time elapsed from Time 1 to 2 was a strong predictor in the model of the FACT-Colorectal Cancer Scale. None of the other predictors were particularly strong in any of the models.

Table 3.

Full multivariate logistic models of FACT-G, FACT Colorectal Cancer Scale, and IES-R total

Predictor Odds ratio 95% CI Wald χ2 P-value
FACT-G
Trait anxiety (per 0.1 unit increase) 1.8 (1.28, 2.54) 11.4 .0007
Sex
 Female 1
 Male 2.52 (0.70, 9.03) 2.01 .16
Age (per year increase) 1.02 (0.96, 1.07) 0.32 .57
Education (per year increase) 1.06 (0.89, 1.27) 0.41 .52
Tumour stage
 I 1
 II 0.74 (0.16, 3.30) 0.16 .69
 III/IV 0.52 (0.11, 2.53) 0.65 .42
Ostomy
 None 1
 Permanent 1.54 (0.38, 6.32) 0.36 .55
FISI score (per unit increase) 1.04 (0.996, 1.09) 3.28 .07
Time elapsed (per year increase) 0.99 (0.98, 1.01) 1.06 .3
FACT Colorectal Cancer Scale
Trait anxiety (per 0.1 unit increase) 1.61 (1.16, 2.24) 8.1 .0044
Sex
 Female 1
 Male 1.63 (0.46, 5.70) 0.58 .45
Age (per year increase) 1.02 (0.97, 1.07) 0.46 .5
Education (per year increase) 1.1 (0.92, 1.32) 1.03 .31
Tumour stage
 I 1
 II 1.31 (0.30, 5.75) 0.13 .72
 III/IV 1.06 (0.23, 4.88) 0.01 .94
Ostomy
 None 1
 Permanent 0.77 (0.18, 3.18) 0.13 .71
FISI score (per unit increase) 1.08 (1.03, 1.14) 8.48 .0036
Time elapsed (per year increase) 0.98 (0.969, 0.999) 4.14 .042
IES-R total
Trait anxiety (per 0.1 unit increase) 1.35 (1.01, 1.80) 4.21 .04
Sex
 Female 1
 Male 0.92 (0.28, 3.02) 0.02 .89
Age (per year increase) 1.01 (0.96, 1.06) 0.13 .72
Education (per year increase) 1.1 (0.92, 1.32) 1.19 .28
Tumour stage
 I 1
 II 2.29 (0.58, 9.05) 1.39 .24
 III/IV 0.57 (0.12, 2.65) 0.51 .48
Ostomy
 None 1
 Permanent 1.16 (0.30, 4.40) 0.05 .83
FISI score (per unit increase) 1.07 (1.02, 1.12) 7.94 .0048
Time elapsed (per year increase) 0.99 (0.98, 1.004) 1.87 .17
Open in a new tab

All odds ratios correspond to participants in the poor versus better score groups, i.e. below (vs. above) median for FACT, above (vs. below) median for IES-R.

Table 4 shows the best-fitting models for the FACT scales from the second stage of data analysis. As can be seen, higher levels of trait anxiety were closely and consistently related to poorer HRQoL. Each increase of 0.1 units in the trait anxiety score was associated with a 28-56% increase in the probability of having a FACT score below the median (physical well-being, 56%; social well-being, 47%; emotional well-being, 45%; functional well-being, 28%; FACT-G 56%; and FACT-Colorectal Cancer Scale, 42%). Faecal incontinence was more varied in incidence and severity, as well as being measured on a more finely graded scale than trait anxiety. As a result, there is a consistent but smaller effect of FISI scores on the HRQoL measures. Each increase of one unit in the FISI was associated with an increased probability of scores falling below the median on the FACT Emotional Well-Being scale (6% increase in probability) and the FACT-Colorectal Cancer Scale (7% increase in probability). Males were about 3.8 times more likely than females to fall on the poor side of the median on FACT Social Well-Being scores. Time elapsed between Time 1 and 2 was marginally associated with scores on the FACT-Colorectal Cancer Scale, indicating that those patients who were further out from active treatment tended to report fewer gastrointestinal symptoms.

Table 4.

Best-fit multivariate logistic models of all FACT scales

Predictor Odds ratio 95% CI Wald χ2 P-value
Physical well-being
Trait anxiety (per 0.1 unit increase) 1.56 (1.21, 2.02) 11.75 .0006
Social well-being
Trait anxiety (per 0.1 unit increase) 1.47 (1.14, 1.88) 9.05 .0026
Sex
 Female 1
 Male 3.81 (1.21, 11.99) 5.25 .022
Emotional well-being
Trait anxiety (per 0.1 unit increase) 1.45 (1.12, 1.88) 8.02 .0046
FISI score (per unit increase) 1.06 (1.01, 1.10) 6.95 .0084
Functional well-being
Trait anxiety (per 0.1 unit increase) 1.28 (1.04, 1.57) 5.24 .022
FACT-G
Trait anxiety (per 0.1 unit increase) 1.56 (1.20, 2.02) 11.28 .0008
Colorectal Cancer Scale
Trait anxiety (per 0.1 unit increase) 1.42 (1.08, 1.87) 6.26 .012
FISI score (per unit increase) 1.07 (1.02, 1.12) 8.21 .0042
Time elapsed (per year increase) 0.987 (0.973, 1.001) 3.55 .06
Open in a new tab

Table 5 shows the best-fitting models for the IES-R subscales and total score. Trait anxiety and faecal incontinence entered the models for all 3 of the subscales and the total IES-R score, although trait anxiety was just marginally significant in the total score model. Each increase of 0.1 units in the trait anxiety score was associated with a greater probability of IES-R scores falling above the median, meaning more difficulties with symptoms of post-traumatic stress (IES-R: avoidance, 41%; hyperarousal, 30%; intrusiveness, 45%; and total score, 24%). Severity of faecal incontinence also had a deleterious impact. Each unit increase in FISI scores was associated with a 5-7% increase in the probability of IES-R scores falling above the median (IES-R: avoidance, 7%; hyperarousal, 6%; intrusiveness, 6%; and total score, 5%).

Table 5.

Best-fit multivariate logistic models of all IES-R scales

Predictor Odds ratio 95% CI Wald χ2 P-value
Avoidance
Trait anxiety (per 0.1 unit increase) 1.41 (1.09, 1.82) 6.7 .0096
FISI score (per unit increase) 1.07 (1.03, 1.12) 9.83 .0017
Hyperarousal
Trait anxiety (per 0.1 unit increase) 1.3 (1.03, 1.64) 4.91 .027
FISI score (per unit increase) 1.06 (1.02, 1.10) 7.48 .0062
Intrusiveness
Trait anxiety (per 0.1 unit increase) 1.45 (1.12, 1.88) 8 .0047
FISI score (per unit increase) 1.06 (1.01, 1.10) 6.82 .009
Total score
Trait anxiety (per 0.1 unit increase) 1.24 (0.99, 1.55) 3.57 .059
FISI score (per unit increase) 1.05 (1.01, 1.09) 6.38 .012
Open in a new tab

Discussion

The present study examined the influence of trait anxiety on HRQoL and PTSS in a sample of rectal cancer survivors, while simultaneously considering the effects of the medical ramifications of illness (stage of disease, presence of permanent colostomy, presence/severity of faecal incontinence) and socio-demographic characteristics. The mean FACT subscale scores were comparable to those reported by 304 participants in a representative sample of U.S. citizens with no history of specified illness (Cella et al., 2003; Holzner et al., 2004). Mean scores from the present study versus the population study, respectively, were: physical well-being (23.6 vs. 24.8); Social well-being (22.9 vs. 20.3); Emotional well-being (20.5 vs. 19.8); Functional well-being (20.5 vs. 21.5). Thus, long-term HRQoL among this sample of rectal cancer survivors was quite favourable 2-5 years following post-operative assessment. About half of these patients had scores of zero on the IES-R, suggesting no discernible problems with PTSS.

On the other hand, there were many individuals who were faring rather poorly. Several variables were investigated for their associations with the poorest levels of HRQoL and PTSS, and two consistent findings emerged. First, higher levels of trait anxiety were associated with the poorest scores on every one of the outcome measures. This finding was consistent with the primary hypothesis as well as with a broad array of findings that indicate that dispositional negative affect has a potent influence on self-assessments of health and well-being (Cameron, 2003; Pennebaker, 2000; Watson & Pennebaker, 1989). Watson and Pennebaker (1989) posited that the correlations between negative affect and self-reported functioning ‘reflect a common, underlying disposition of somatopsychic distress’ (p. 235). Their ‘symptom perception hypothesis’ states that subjects who are high in negative affect are also more likely to notice and complain about negative changes in physical functioning. Similarly, it has been suggested that reports of PTSS in cancer survivors might reflect generalized distress (Deimling et al., 2002; Palmer et al., 2004). In follow-up analyses, we looked at the relationship between FACT-G scores and the IES-R total score using Fisher’s exact test, and found that the two measures were significantly related (p < .0001). Thus, those patients who reported poorer HRQoL were also more likely to have problems with PTSS, and both of those measures were significantly related to trait anxiety.

Second, the one biomedical predictor of poorer HRQoL was severity of faecal incontinence, showing a relationship with the Emotional well-being subscale and the FACT-C Colorectal Cancer Scale. This finding is similar to other studies showing that lingering problems with bowel control can have a negative impact on certain domains of HRQoL (Vironen et al., 2006). Faecal incontinence also had a negative impact on PTSS, which fits with reports that PTSS in cancer survivors is related to current physical symptoms, but not to severity of the original illness or the complexity of treatment (Deimling et al., 2002). On the other hand, the presence of a colostomy did not enter any of the models. It may be that incontinence represents an uncontrollable and chronic stressor in one’s life, while a colostomy can be routinely managed with experience. Other patient characteristics were less consistent predictors. For example, male patients were more likely to report poorer social functioning, and the passage of time following active treatment brought decreased problems related to gastrointestinal functioning.

There are several weaknesses of this study. The first weakness is in the timing of the assessments of trait anxiety, which took place following initial treatment. One might argue that patients with worse disease consequently had greater levels of psychological distress, causing higher scores on the anxiety measures as well as poorer HRQoL, which would leave open the question of causality. However, further analyses showed that neither the stage of disease nor the grade of the tumour was related to trait anxiety or to any of the HRQoL measures. Also, both measures of trait anxiety used here have been shown to be quite stable over time. In that regard, the anxiety measures were strongly related to the HRQoL measures taken 2-5 years later, so there was apparently something captured by the anxiety measures that persisted long enough to be related to the outcome measures. Clearly, a longitudinal study that included baseline measures of HRQoL and PTSS would provide better answers. Second, we did not measure all variables that could have an impact on HRQoL and PTSS, such as the nature and extent of adjuvant treatments. However, it is difficult to know how one would characterize the various permutations in adjuvant treatment in a way that they could be included as valid predictors in regression analyses, although this is not to say that some metric could not be devised. In the current study, the authors chose to include stage of disease as an imperfect proxy for intensity and duration of treatment. While it is expected that the inclusion of information regarding adjuvant treatment would have accounted for more variance in the outcome measures, it would most likely not supplant variance accounted for by the trait anxiety measures, since trait anxiety had been measured prior to any adjuvant treatment. A third weakness is in the attrition of participants from Time 1 to 2. Compared to the Time 2 participants, the deceased and lost patients had worse disease, and those who were unwilling to participate were older. These losses of participants may thus have contributed to the skewness of the HRQoL measures. Lastly, the relatively modest sample size, as well as the small proportions of certain subgroups of patients (e.g. African-American, patients with colostomy), did not allow for more rigorous tests of associations with the outcome measures.

Still, this study could inform research designed to understand the complex nature of HRQoL. The self-assessment of HRQoL is primarily a process of cognitive appraisal, and is thus open to the influence of psychological factors (Rapkin & Schwartz, 2004). The present study suggests that more attention should be given to the role of affective disposition. It must be emphasized, however, that such findings would not relieve health care providers of the responsibility to help patients improve their HRQoL. From the surgical standpoint, long-term HRQoL will likely benefit from new techniques that will reduce the likelihood and severity of faecal incontinence (Wise & Kodner, 2006). And, while personality traits might influence patient-reported outcomes, providers still need to work to alleviate survivor distress. Such work in medicine and psychology will add to our growing understanding of HRQoL such that future efforts towards improving the quality of patients’ lives will be individually tailored and thus more effective.

Acknowledgements

This research was supported by the Alvin J. Siteman Cancer Center, National Cancer Institute Grant #1R03 CA84845 01, and the American Society of Colon and Rectal Surgeons (LPG 073).

References

  1. Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: A quality-of-life instrument for use in international clinical trials in oncology. Journal of the National Cancer Institute. 1993;85:365–376. doi: 10.1093/jnci/85.5.365. [DOI] [PubMed] [Google Scholar]
  2. American Psychiatric Association . Diagnostic and statistical manual of mental disorders: DSM-IV. 4th ed. American Psychiatric Association; Washington, DC: 1994. [Google Scholar]
  3. Aziz NM. Cancer survivorship research: State of knowledge, challenges and opportunities. Acta Oncologica. 2007;46:417–432. doi: 10.1080/02841860701367878. [DOI] [PubMed] [Google Scholar]
  4. Barnes LLB, Harp D, Jung WS. Reliability generalization of scores on the Spielberger State-Trait Anxiety Inventory. Educational and Psychological Measurement. 2002;62:603–618. [Google Scholar]
  5. Beck AT, Steer RA, Brown GK. Manual for the beck depression inventory. 2nd ed. Psychological Corporation; San Antonio, TX: 1996. [Google Scholar]
  6. Bleday R, Garcia-Aguilar J. Surgical treatment of rectal cancer. In: Wolff BG, Fleshman JW, Beck DE, Pemberton JH, Wexner SD, editors. The ASCRS textbook of colon and rectal surgery. 1st ed. Springer; New York, NY: 2007. pp. 413–436. [Google Scholar]
  7. Cameron LD. Anxiety, cognition, and responses to health threats. In: Cameron LD, Leventhal H, editors. The self-regulation of health and illness behaviour. Routledge; New York, NY: 2003. pp. 157–183. [Google Scholar]
  8. Camilleri-Brennan J, Steele RJ. Prospective analysis of quality of life and survival following mesorectal excision for rectal cancer. British Journal of Surgery. 2001;88:1617–1622. doi: 10.1046/j.0007-1323.2001.01933.x. [DOI] [PubMed] [Google Scholar]
  9. Camilleri-Brennan J, Steele RJ. Objective assessment of morbidity and quality of life after surgery for low rectal cancer. Colorectal Disease. 2002;4:61–66. doi: 10.1046/j.1463-1318.2002.00300.x. [DOI] [PubMed] [Google Scholar]
  10. Carver CS, White TL. Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS BAS scales. Journal of Personality and Social Psychology. 1994;67:319–333. [Google Scholar]
  11. Cella DF, Jacobsen PB, Orav EJ, Holland JC, Silberfarb PM, Rafla S. A brief POMS measure of distress for cancer patients. Journal of Chronic Diseases. 1987;40:939–942. doi: 10.1016/0021-9681(87)90143-3. [DOI] [PubMed] [Google Scholar]
  12. Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, et al. The functional assessment of cancer therapy scale: Development and validation of the general measure. Journal of Clinical Oncology. 1993;11:570–579. doi: 10.1200/JCO.1993.11.3.570. [DOI] [PubMed] [Google Scholar]
  13. Cella D, Zagari MJ, Vandoros C, Gagnon DD, Hurtz HJ, Nortier JW. Epoetin alfa treatment results in clinically significant improvements in quality of life in anemic cancer patients when referenced to the general population. Journal of Clinical Oncology. 2003;21:366–373. doi: 10.1200/JCO.2003.02.136. [DOI] [PubMed] [Google Scholar]
  14. Cloninger CR, Przybeck TR, Svrakic DM, Wetzel RD. Temperament and character inventory (TCI): A guide to its development and use. Center for Psychobiology of Personality; St. Louis, MO: 1994. [Google Scholar]
  15. Collett D. Modelling binary data. 2nd ed. Chapman & Hall/CRC.; Boca Raton, FL: 2003. [Google Scholar]
  16. Cornish JA, Tilney HS, Heriot AG, Lavery IC, Fazio VW, Tekkis PP. A meta-analysis of quality of life for abdominoperineal excision of rectum versus anterior resection for rectal cancer. Annals of Surgical Oncology. 2007;14:2056–2068. doi: 10.1245/s10434-007-9402-z. [DOI] [PubMed] [Google Scholar]
  17. Costa PT, McCrae RR. Normal personality assessment in clinical practice. Psychological Assessment. 1992;4:5–13. [Google Scholar]
  18. Curran D, Aaronson N, Standaert B, Molenberghs G, Therasse P, Ramirez A, et al. Summary measures and statistics in the analysis of quality of life data: An example from an EORTC-NCIC-SAKK locally advanced breast cancer study. European Journal of Cancer. 2000;36:834–844. doi: 10.1016/s0959-8049(00)00056-3. [DOI] [PubMed] [Google Scholar]
  19. Deimling GT, Kahana B, Bowman KF, Schaefer ML. Cancer survivorship and psychological distress in later life. Psychooncology. 2002;11:479–494. doi: 10.1002/pon.614. [DOI] [PubMed] [Google Scholar]
  20. Engel J, Kerr J, Schlesinger-Raab A, Eckel R, Sauer H, Holzel D. Quality of life in rectal cancer patients: A four-year prospective study. Annals of Surgery. 2003;238:203–213. doi: 10.1097/01.sla.0000080823.38569.b0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Eysenck HJ, Eysenck SB. Manual of the EPQ (Eysenck Personality Questionnaire) Educational and Industrial Testing Service; San Diego, CA: 1976. [Google Scholar]
  22. Gotay CC, Lipscomb J, Snyder CF. Reflections on findings of the cancer outcomes measurement working group: Moving to the next phase. Journal of the National Cancer Institute. 2005;97:1568–1574. doi: 10.1093/jnci/dji337. [DOI] [PubMed] [Google Scholar]
  23. Heath AC, Cloninger CR, Martin NG. Testing a model for the genetic structure of personality: A comparison of the personality systems of Cloninger and Eysenck. Journal of Personality and Social Psychology. 1994;66:762–775. doi: 10.1037//0022-3514.66.4.762. [DOI] [PubMed] [Google Scholar]
  24. Holzner B, Kemmler G, Cella D, De PC, Meraner V, Kopp M, et al. Normative data for functional assessment of cancer therapy: General scale and its use for the interpretation of quality of life scores in cancer survivors. Acta Oncologica. 2004;43:153–160. doi: 10.1080/02841860310023453. [DOI] [PubMed] [Google Scholar]
  25. Horowitz M, Wilner N, Alvarez W. Impact of event scale: A measure of subjective stress. Psychosomatic Medicine. 1979;41:209–218. doi: 10.1097/00006842-197905000-00004. [DOI] [PubMed] [Google Scholar]
  26. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. Wiley; New York, NY: 2000. [Google Scholar]
  27. Kangas M, Henry JL, Bryant RA. Posttraumatic stress disorder following cancer: A conceptual and empirical review. Clinical Psychology Review. 2002;22:499–524. doi: 10.1016/s0272-7358(01)00118-0. [DOI] [PubMed] [Google Scholar]
  28. Kangas M, Henry JL, Bryant RA. Predictors of posttraumatic stress disorder following cancer. Health Psychology. 2005;24:579–585. doi: 10.1037/0278-6133.24.6.579. [DOI] [PubMed] [Google Scholar]
  29. Koller M, Kussman J, Lorenz W, Jenkins M, Voss M, Arens E, et al. Symptom reporting in cancer patients: The role of negative affect and experienced social stigma. Cancer. 1996;77:983–995. doi: 10.1002/(sici)1097-0142(19960301)77:5<983::aid-cncr27>3.0.co;2-z. [DOI] [PubMed] [Google Scholar]
  30. McNair DM, Lorr M, Droppleman LF. EdITS manual for the profile of mood states. Educational and Industrial Testing Service; San Diego, CA: 1992. [Google Scholar]
  31. Neuman HB, Schrag D, Cabral C, Weiser MR, Paty PB, Guillem JG, et al. Can differences in bowel function after surgery for rectal cancer be identified by the European Organization for Research and Treatment of Cancer quality of life instrument? Annals of Surgical Oncology. 2007;14:1727–1734. doi: 10.1245/s10434-006-9283-6. [DOI] [PubMed] [Google Scholar]
  32. Pachler J, Wille-Jorgensen P. Quality of life after rectal resection for cancer, with or without permanent colostomy. Cochrane Database of Systematic Reviews. 2005:1–23. doi: 10.1002/14651858.CD004323.pub3. [DOI] [PubMed] [Google Scholar]
  33. Palmer SC, Kagee A, Coyne JC, DeMichele A. Experience of trauma, distress, and posttraumatic stress disorder among breast cancer patients. Psychosomatic Medicine. 2004;66:258–264. doi: 10.1097/01.psy.0000116755.71033.10. [DOI] [PubMed] [Google Scholar]
  34. Pennebaker JW. Psychological factors influencing the reporting of physical symptoms. In: Stone AA, Turkkan JS, Bachrach CA, Jobe JB, Kurtzman HS, Cain VS, editors. The science of self-report: Implications for research and practice. Erlbaum; Mahwah, NJ: 2000. pp. 299–315. [Google Scholar]
  35. Rapkin BD, Schwartz CE. Toward a theoretical model of quality-of-life appraisal: Implications of findings from studies of response shift. Health and Quality of Life Outcomes. 2004;2:14. doi: 10.1186/1477-7525-2-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rauch P, Miny J, Conroy T, Neyton L, Guillemin F. Quality of life among disease-free survivors of rectal cancer. Journal of Clinical Oncology. 2004;22:354–360. doi: 10.1200/JCO.2004.03.137. [DOI] [PubMed] [Google Scholar]
  37. Rockwood TH, Church JM, Fleshman JW, Kane RL, Mavrantonis C, Thorson AG, et al. Patient and surgeon ranking of the severity of symptoms associated with fecal incontinence: The Fecal Incontinence Severity Index. Diseases of the Colon and Rectum. 1999;42:1525–1532. doi: 10.1007/BF02236199. [DOI] [PubMed] [Google Scholar]
  38. Rowland JH, Hewitt M, Ganz PA. Cancer survivorship: A new challenge in delivering quality cancer care. Journal of Clinical Oncology. 2006;24:5101–5104. doi: 10.1200/JCO.2006.09.2700. [DOI] [PubMed] [Google Scholar]
  39. Salsman JM, Segerstrom SC, Brechting EH, Carlson CR, Andrykowski MA. Posttraumatic growth and PTSD symptomatology among colorectal cancer survivors: A 3-month longitudinal examination of cognitive processing. Psycho-Oncology. 2009;18:30–41. doi: 10.1002/pon.1367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Schmidt CE, Bestmann B, Kuchler T, Longo WE, Kremer B. Impact of age on quality of life in patients with rectal cancer. World Journal of Surgery. 2005a;29:190–197. doi: 10.1007/s00268-004-7556-4. [DOI] [PubMed] [Google Scholar]
  41. Schmidt CE, Bestmann B, Kuchler T, Longo WE, Kremer B. Prospective evaluation of quality of life of patients receiving either abdominoperineal resection or sphincter-preserving procedure for rectal cancer. Annals of Surgical Oncology. 2005b;12:117–123. doi: 10.1245/ASO.2005.12.036. [DOI] [PubMed] [Google Scholar]
  42. Schwartz CE, Sprangers MA. An introduction to quality of life assessment in oncology: The value of measuring patient-reported outcomes. American Journal of Managed Care. 2002;8:S550–S559. [PubMed] [Google Scholar]
  43. Spielberger CD, Gorsuch RL, Lushene RE, Vagg PR, Jacobs GA. STAI manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press; Palo Alto, CA: 1970. [Google Scholar]
  44. Taylor JA. A personality scale of manifest anxiety. Journal of Abnormal and Social Psychology. 1953;48:285–290. doi: 10.1037/h0056264. [DOI] [PubMed] [Google Scholar]
  45. Vironen JH, Kairaluoma M, Aalto AM, Kellokumpu IH. Impact of functional results on quality of life after rectal cancer surgery. Diseases of the Colon and Rectum. 2006;49:568–578. doi: 10.1007/s10350-006-0513-6. [DOI] [PubMed] [Google Scholar]
  46. Ward WL, Hahn EA, Mo F, Hernandez L, Tulsky DS, Cella D. Reliability and validity of the functional assessment of cancer therapy-colorectal (FACT-C) quality of life instrument. Quality of Life Research. 1999;8:181–195. doi: 10.1023/a:1008821826499. [DOI] [PubMed] [Google Scholar]
  47. Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology. 1988;54:1063–1070. doi: 10.1037//0022-3514.54.6.1063. [DOI] [PubMed] [Google Scholar]
  48. Watson D, Pennebaker JW. Health complaints, stress, and distress: Exploring the central role of negative affectivity. Psychological Review. 1989;96:234–254. doi: 10.1037/0033-295x.96.2.234. [DOI] [PubMed] [Google Scholar]
  49. Weiss DS, Marmar CR. The impact of event scale - revised. In: Wilson JP, Keane TM, editors. Assessing psychological trauma and PTSD. The Guilford Press; New York, NY: 1997. pp. 399–411. [Google Scholar]
  50. Wilson IB, Cleary PD. Linking clinical variables with health-related quality of life: A conceptual model of patient outcomes. Journal of American Medical Association. 1995;273:59–65. [PubMed] [Google Scholar]
  51. Wise PE, Kodner IJ. Treatment of fecal incontinence. In: Bieber EJ, Sanfilippo JS, Horowitz GM, editors. Clinical gynecology. Elsevier Science; Philadelphia, PA: 2006. pp. 355–374. [Google Scholar]
  52. Yoo HJ, Kim JC, Eremenco S, Han OS. Quality of life in colorectal cancer patients with colectomy and the validation of the functional assessment of cancer therapy-colorectal (FACT-C), Version 4. Journal of Pain and Symptom Management. 2005;30:24–32. doi: 10.1016/j.jpainsymman.2004.12.009. [DOI] [PubMed] [Google Scholar]
  53. Zelenski JM, Larsen RJ. Predicting the future: How affect-related personality traits influence likelihood judgments of future events. Personality and Social Psychology Bulletin. 2002;28:1000–1010. [Google Scholar]

RESOURCES