Illness and treatment beliefs in head and neck cancer

Journal of Psychosomatic Research 63 (2007) 17 – 26

Illness and treatment beliefs in head and neck cancer: Is Leventhal’s common sense model a useful framework for determining changes in outcomes over time? Carrie Diane Llewellyna,4, Mark McGurkb, John Weinmana a

Health Psychology Section, Department of Psychology, Institute of Psychiatry (Guy’s Campus) King’s College London, Guy’s Hospital, London, United Kingdom b Department of Oral Surgery, Guy’s Hospital NHS Trust, London, United Kingdom Received 20 November 2006; received in revised form 16 January 2007; accepted 18 January 2007

Abstract Objective: The main aim of this prospective study was to examine the utility of Leventhal’s common sense model in predicting longitudinal judgement-based outcomes in patients with head and neck cancer (HNC). The study is of potential importance as it focuses on the relations between personality factors, coping styles, informational needs, illness representations, and outcomes using a longitudinal study design. This has particular value as the trend in similar research is to focus on concurrent relations between variables. In addition, the prediction of numerous outcomes from illness perceptions has received relatively scant attention in the field of HNC. Methods: Fifty patients completed the following measures prior to treatment, 1 month and 6–8 months after treatment: IPQ-R, BMQ, Brief COPE, LOT-R, SCIP, EORTC QLQ-C30, SF-12, Patient Gen-

erated Index (PGI), and HADS. Results: Baseline illness and treatment beliefs were not predictive of HR-QoL, individualized QoL, or anxiety 6–8 months after treatment; however, beliefs about the chronicity of the disease (timeline beliefs) were predictive of depression after treatment. Coping strategies employed and levels of satisfaction with information before treatment were significant predictors of several outcomes. Conclusions: Our findings suggest that a common sense model may be a useful framework for eliciting and understanding patients’ beliefs regarding HNC; however, there are concerns regarding the use of a ddynamicT model to predict longitudinal outcomes from baseline factors that may change over the course of an illness. D 2007 Elsevier Inc. All rights reserved.

Keywords: Coping; Head and neck cancer; HR-QoL; Individualized QoL; Common sense model

Introduction There has been minimal improvement in survival rates for head and neck cancer (HNC), with a 5-year relative survival rate at approximately 50% [1]. Consequently, quality of life (QoL) has become an increasingly important outcome measure. Treatment for HNC leads to varying degrees of disability, but it is now apparent that factors such

4 Corresponding author. Brighton and Sussex Medical School, Mayfield House, Falmer, Brighton, BN1 9PH, United Kingdom. Tel.: +44 0 1273 642187; fax: +44 0 1273. E-mail address: [email protected] (C.D. Llewellyn). 0022-3999/07/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jpsychores.2007.01.013

as tumour size or physical functioning after treatment are not wholly responsible for adaptation and QoL [2,3]. There is a paucity of research exploring factors that impact on HR-QoL in HNC [4], but it has been suggested that to understand the process of how QoL judgements are made by patients, a model of QoL needs to go beyond disease- and treatment-related factors, and incorporate the patient’s beliefs about the disease and treatment. These representations may affect the meaning and relative importance of domains involved in making QoL judgments [5]. Explanations for variation in responses to a health threat have been provided by Leventhal’s common sense model [6]. The central proposition of this theory is that the beliefs (or representations) patients have about their illness

18

C.D. Llewellyn et al. / Journal of Psychosomatic Research 63 (2007) 17 – 26

influence coping responses which, in turn, may influence judgement-based outcomes (such as QoL and emotional outcomes). This theory could be useful in explaining variation in patient outcomes that cannot be explained by either sociodemographic, disease-, or treatment-related factors. Based on the common sense model, a response to a health threat is the product of an underlying control system which can be divided into three broad processes. Firstly, the cognitive and emotional representations of the health threat are constructed. These representations reflect the individual’s interpretation of the health threat and can be a result of internal cues, for example, symptoms, and/or external cues such as sources of information. Secondly, an action plan is developed. The coping strategy used is perceived by the individual to be appropriate to the beliefs they hold. The third stage is the process of coping appraisal. This is a process of evaluating the effectiveness of the coping strategy on the outcome or goal. The two key attributes of

the common sense model are that these three stages occur in parallel on both an emotional and a cognitive level, and that the interaction between each level is ddynamicT, meaning that each component is influenced by a process of feedback. Previous cross-sectional research with HNC patients has demonstrated how beliefs such as illness identity and emotional representations are related to variables such as HR-QoL subscales, prior to treatment [7,8]. An extended version of the common sense model has been proposed which stipulates that patients will dnot just have their own ideas about the illness, but also about the treatment being offeredT [9]. This extended model was originally used to provide an explanation for adherence to medication; however, treatment beliefs could equally be important in the perception of outcomes, such as QoL, and thus have been included in this study. The model proposes that individuals form beliefs about their illness based on both abstract and concrete sources of

Table 1 Socio-demographic and clinical characteristics of follow-up samples and statistical analysis comparing patient samples at follow-up with sample at baseline

Characteristic Gender Male Female Age (years) Mean (S.D.) Range Ethnicity White Other Marital status Single/widowed/divorced Married/cohabiting Highest qualification4 None GCSE/O’ level GCE/A’ level Higher education Degree or higher AJCCb Stage of cancer4 Stage 1 Stage 2 Stage 3 Stage 4a–c Stage dichotomized4 Early stage (1 and 2) Advanced stage (3 and 4) Treatment Surgery only Radiotherapy only Surgery and radiotherapy Radiotherapy and chemotherapy Surgery and radiotherapy and chemotherapy a

Baseline sample (T1) (n=82)

Follow-up sample at T2 (n=68)

n (%)

n (%)

Follow-up sample at T3 (n=50) Test statistica m2=0.23; P=.63

Test statistica

n (%)

m2=3.51; P=.06

54 (66) 28 (34)

44 (65) 24 (35)

59.9 (12.5) 23–89

59.6 (13.10) 23–89

75 (92) 7 (8)

62 (91) 6 (9)

Crame´r’s V=0.02; P=.84

47 (94) 3 (6)

Crame´r’s V=0.11; P=.30

32 (39) 50 (61)

28 (41) 40 (59)

Crame´r’s V=0.10; P=.38

22 (44) 28 (56)

Crame´r’s V=0.13; P=.25

27 (33) 15 (18) 11 (13) 10 (12) 16 (20)

21 2 8 9 14

(31) (18) (12) (13) (21)

Crame´r’s V=0.17; P=.88

15 (30) 8 (16) 4 (8) 9 (18) 11 (22)

Crame´r’s V=0.31; P=.30

19 20 12 26

19 16 12 16

(28) (24) (18) (24)

Crame´r’s V=0.48; P=.003

17 (34) 11 (22) 10 (20) 7 (14)

Crame´r’s V=0.53; P=.001

Crame´r’s V=0.21; P=.07

28 (56) 17 (34)

Crame´r’s V=0.27; P=.02

(23) (24) (15) (32)

39 (47) 38 (47)

35 (52) 28 (41)

22 21 26 9

20 17 21 6

(27) (26) (31) (11)

4 (5)

(29) (25) (31) (9)

4 (6)

t(80)=2.74; P=.65

Crame´r’s V=0.26; P=.26

29 (58) 21 (42) 60.94 (12.67)

18 14 12 4

(36) (28) (24) (8)

t(80)=1.17; P=.34

Crame´r’s V=0.36; P=.03

2 (4)

Independent t-tests, m2 tests, and Crame´r’s V tests comparing respondents at each time point with dropouts from original baseline sample. American Joint Committee on Cancer. 4 Data missing/unobtainable.

b


19

Table 2 Means (S.D.), medians, range, and Cronbach’s alpha values for outcomes (individualized QoL, generalized and cancer-specific HR-QoL, anxiety, and depression) at baseline, 1 month, and 6–8 months post-treatment Mean (SD)

Median

Domain

Baseline

1 month

6–8 months

PGIa EORTC QLQC30 Global health status/QoL SF-12v2 Domain MCSa PCSa HADS subscale Anxiety Depression

4.27 (2.3)

3.39 (1.8)

4.34 (2.2)

Range

Cronbach’s alpha

6–8 6–8 Baseline 1 month months Baseline 1 month 6–8 months Baseline 1 month months 4.0

3.3

4.4

0–10

0–8.7

0.6 –10

0–100

0–100

62.2 (23.7) 58.6 (21.2) 66.54 (23.9) 67

58

67

46.0 (11.7) 42.8 (11.0) 47.4 (10.2) 38.9 (10.1)

45 41

50 45

10–69 10–68

14–62 19–56

32– 66 14 – 60

6 5

5 3

0–20 0–11

0–18 0–18

0 –20 0 –19

7.9 (26.9) 6.54 (4.9) 3.9 (3.5) 5.74 (4.1)

49.0 (8.3) 49 42.9 (11.8) 50 5.6 (4.6) 4.7 (4.5)

7 3

0 –100

N/A 0.87

0.82

0.94

N/A N/A 0.89 0.81

0.91 0.85

0.89 0.89

a

Cronbach’s alpha cannot be computed as the scale consists only of one item. 4 Data from one or two participants missing.

information available to them. Leventhal et al. [6,10] suggest that beliefs are potentially formed by three sources of information. The first source being the general pool of dlayT information that the individual has previously assimilated. The second source is derived externally from friends, family, and authoritative resources. The third source of information is the current experience of the illness, such as somatic experiences and symptoms. Many factors are known to influence self-regulatory processes in the health care setting. From earlier research on preparation for stressful investigations and treatments [11], it is apparent that clear information provision can provide patients with a map against which their actual experiences can be evaluated, which in turn can enhance behavioural and informational control. Unmet informational needs and low satisfaction with information provided have been found to be related to unfavourable outcomes, such as lower HRQoL and higher levels of depression and anxiety in HNC [12,13]. Thus, in addition to examining the role of patients’ beliefs and coping, the present study will also examine the contribution of adequate information provision to adjustment and outcome following HNC surgery and treatment. Research has demonstrated that patient beliefs, in particular having a strong illness identity, perceptions of the negative consequences of the illness, and chronic timeline beliefs, are related to worse HR-QoL and emotional distress [14,15]. A recent review highlighted that very few studies have investigated the time-lagged relationships between baseline beliefs, coping, and outcome over time using a longitudinal design, and there is a need to further clarify the relationships between key components of the common sense model at baseline and prospective outcomes [16]. Despite this recommendation, the theoretical reasons why beliefs prior to treatment or intervention should predict longitudinal outcomes based on a dynamic model have rarely been discussed. Indeed, it would be predicted that pretreatment beliefs are associated with coping strategies used after treatment rather than with specific outcomes.

Other theories appropriate to self-regulation and illness [17] have also proposed that personality factors such as dispositional optimism may play an important role in predicting both coping strategy and outcome. The current study is of value as it focuses on the relations between personality factors, coping styles, informational needs, illness representations, and outcomes using a longitudinal study design. In addition, the prediction of outcomes from such factors has received relatively scant attention in the field of HNC which is pertinent given the post-treatment mortality rate and repercussions of treatment to morbidity. The first aim of this study was to ascertain whether patients’ beliefs before treatment were associated with coping over time. The second aim was to assess the extent to which key components of the common sense model (illness and treatment beliefs and coping) were predictive of outcomes. This may clarify the utility of this model for the

Table 3 Means (S.D.), medians, ranges, and Cronbach’s alpha values of illness representations and treatment beliefs at baseline (T1)

IPQ subscale* Illness Identity Timeline Timeline cyclical Consequences Personal control Illness coherence Emotional representations BMQ-Specific subscale* Necessity Concerns

Min–max score

Cronbach’s alpha

Mean (S.D.)

Median

3.95 7.54 7.63 9.28 6.88 10.05 9.83

(3.2) (2.5) (2.7) (2.7) (1.7) (3.0) (3.1)

3 8 7 9 7 10 10

0–12 3–15 3–15 3–15 3–10 3–15 3–15

N/A 0.88 0.74 0.69 0.61 0.78 0.87

21.05 (2.7) 12.92 (3.5)

20 13

14–25 4–20

0.75 0.75

N/A indicates Cronbach’s alpha cannot be computed as the scale consists only of one item. * Data missing.

20


prediction of longitudinal judgement-based outcomes. In addition, the contribution of dispositional optimism and pretreatment satisfaction with information to variation in outcome was investigated. Hypotheses The following hypotheses were tested: Hypothesis 1. Pretreatment illness and treatment perceptions will be associated with coping strategy used after treatment. Hypothesis 2. Pretreatment illness and treatment perceptions will explain a significant amount of variance in longitudinal outcomes of (a) Standardized HR-QoL, (b) Individualized QoL, (c) Depression, (d) Anxiety.

Methods Design Prospective, repeated measures, questionnaire-based study. Procedure In the period July 2003 to July 2004, eligible patients identified from attending four HNC clinics of hospitals in the southeast of England were consecutively recruited following Local Research Ethics Committee approval and informed consent. Inclusion criteria were any newly diagnosed patient with a first and histologically confirmed primary tumour of the head and neck, over 18 years old. Baseline data were obtained in the period between confirmation of diagnosis and prior to treatment through self-completed questionnaires and medical records. Followup assessments 1 month and 6–8 months after all treatment had finished were by self-completed questionnaire. Followup was terminated at 6–8 months, at recurrence of disease, on entering a palliative phase or death. Measures The following measures were assessed at baseline (T1). The Revised Illness Perception Questionnaire (IPQ-R) [18] was used to assess patient’s beliefs and understanding of their illness and has proven validity and reliability across a range of illness groups. It provides a quantitative assessment of the nature and strength of patient beliefs about the following seven components: the nature of the patients illness didentityT, which is the number of symptoms they perceive to be related to their illness; how long the patients think their illness will last (timeline—chronic) and whether symptoms are sustained or cyclical (timeline—cyclical); the perceived consequences of the illness;

how much personal control the patients feel they have over their illness; whether patients have a coherent understanding of their illness (illness coherence); and, finally, the emotional representation the patients hold towards their illness. Scores on subscales of timeline, cyclical timeline, consequences, illness coherence, personal control, and emotional representations range from 3 to 15 with higher scores indicating a stronger belief. The Beliefs about Medicines Questionnaire (BMQ) [19] was designed to assess patients’ beliefs about medicines prescribed for personal use. It is validated for use in a wide range of illness groups [19]. The BMQ-Specific subscale was adapted to assess treatment beliefs in HNC and comprises two subscales assessing beliefs about the dnecessityT of treatment to health and dconcernsT regarding the possible side-effects and the disruptive effects of treatment. Scores on the necessity and concerns subscales used in this study range between 5 and 25, and 4 and 20, respectively. Higher scores reflect stronger beliefs about the necessity of treatment and stronger concerns. The Life Orientation Test—Revised (LOT-R) [20] was used to assess generalized optimism [21]. This short form consists of six items and scores range from 6 to 30, whereby higher values indicate higher levels of optimism. The Satisfaction with Cancer Information Profile (SCIP) [22] is a new measure used to assess the extent to which HNC patients are satisfied with information received about their cancer treatment and the consequences of treatment. Two satisfaction scores are derived: Satisfaction with the amount and content of information (scores can range from 0 to 14) and Satisfaction with the type and timing of the information (scores can range from 7 to 35). Higher scores on both subscales indicate higher levels of satisfaction. The following measures were assessed at all three time-points. The Brief COPE [23] is a validated, multidimensional coping inventory which was used to assess situational coping. This short-form version of the COPE [24] assesses coping strategies on 14 conceptually different subscales. Both adaptive and maladaptive coping strategies are included. Scores for each subscale range from 2 to 8, with higher scores indicating more frequent use of a particular coping strategy. The European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ-C30) (version 3) [25] is a validated self-completion questionnaire with 15 multi- and single-item domains. For this study, only the global QoL/health score was used in analyses, which range from 0 to 100. Higher scores signify better functioning. The MOS Short-Form Health Survey (SF-12v2) is a multi-purpose short-form consisting of 12 items [26]. Similar to the SF-36, this is a generic measure and version 2 results in an eight-domain profile and two dcomponent summary scoresT. In this study, only the standardized Mental and Physical Component Summary scores (MCS and PCS) were used. Scores range from 0 to 100 with higher scores representing better functioning.


21

Table 4 Relationships between pretreatment psychological factors, satisfaction with information, and use of coping strategies over time Baseline (T1) psychological factors

Coping strategies associateda 1 month after treatment (T2)

Coping strategies associateda 6–8 months after treatment (T3)

Identity (IPQ)

SU (0.26)4; V (0.33)44

Timeline (IPQ)

–

Cyclical timeline (IPQ) Consequences (IPQ)

H ( 0.25)4; R (0.26)4 D (0.36)44; V (0.26)4; P (0.33)44

Personal control (IPQ) Coherence (IPQ) Emotional representation (IPQ)

Necessity (BMQ)

– D ( 0.38)44 SD (0.33)44; P (0.44)44; AC (0.31)4; SB (0.39)44; D (0.55)44; SU (0.4)44; BD (0.33)44 SD (0.31)4; V (0.32)44; P (0.29)4; D (0.51)44; SU (0.33)44; IS (0.29)4; BD (0.3)44 AC (0.38)44; SU (0.28)4; PR (0.25)4; P (0.3)44

SD (0.41)44; SU (0.36)44; IS (0.32)4; BD (0.32)4; V (0.54)44; H (0.29)4; SB (0.34)4 SD (0.34)4; AC (0.36)44; V (0.28)4; P (0.42)44; PR (0.29)4 SD (0.29)4; BD (0.43)44; R (0.31)4 SD (0.39)44; D (0.36)44; V (0.43)44; P (0.33)4; PR (0.31)4 BD ( 0.30)4 AC (0.33)4 SD (0.53)44; AC (0.31)4; D (0.46)44; SU (0.4)44; IS (0.31)4; BD (0.28)4; V (0.29)4 D (0.44)44; BD (0.34)4; V (0.37)44; SB (0.34)4

Optimism (LOT-R)

SB ( 0.30)44

SD ( 0.34)4; D ( 0.37)4; BD ( 0.36)44; P ( 0.3)4; SB ( 0.3)4

Satisfaction with information Amount and content Type and timing

– –

ES ( 0.30)4; A ( 0.39)44 –

Concerns (BMQ)

SD (0.32)4

SD indicates self-distraction; AC, active coping; P, planning; V, venting; PR, positive reframing; H, humour; A, acceptance; R, use of religion; SU, substance use; ES, use of emotional support; IS, use of instrumental support; BD, behavioural disengagement; SB, self-blame; D, denial. a Spearman’s correlation coefficient. 4 PV.05. 44 PV.01.

The Patient Generated Index (PGI) is a global measure of individualized QoL and reflects patients’ evaluations of QoL, based on the definition of QoL as, dthe extent to which our hopes and ambitions are matched by experienceT [27]. The PGI aims to provide a measure that represents the effect of a condition on the aspects of patients’ lives that they consider most important. The PGI is completed in three stages and a final score can be computed which ranges from 0 to 10. This score represents the extent to which reality meets or falls short of the patient’s hopes and expectations in the areas of life prioritized and a higher score indicates a greater QoL. The Hospital Anxiety and Depression Scale (HADS) [28] is a validated scale used to provide a brief measure of state anxiety and depression. Higher scores indicate greater anxiety or depression. Scores can range from 0 to 21 with scores of 11 or more indicating probable psychological morbidity. Reliability data for each measure can be found in Tables 2 and 3. Standard socio-demographic information included in this study was age, gender, ethnicity, marital status, and educational attainment. Clinical variables of interest were site of cancer, stage of disease, and type of treatment planned. Statistical analysis Preliminary analysis Independent t-tests, v 2 tests, and Crame´r’s V tests were used to compare characteristics of respondents with nonrespondents. Correlations were conducted between baseline

(T1) psychological factors (LOT-R, IPQ, and BMQ subscales) and coping over time (T2 and T3). Methods of analysis to investigate predictors of long-term QoL To test the associations between explanatory variables measured at baseline (T1) and longitudinal outcomes assessed 6–8 months post-treatment (T3), correlations were conducted. In order to assess a range of outcomes encompassing HR-QoL, individualized QoL, and emotional outcomes, six outcome variables were used throughout these analyses. The EORTC QLQ-C30 global QoL score and SF-12 (PCS scores and MCS scores) were used as markers of HR-QoL, PGI scores for individualized QoL, and depression and anxiety scores for markers of emotional outcomes. Multiple regression analyses using the stepwise method for variable entry were conducted for each outcome variable, entering factors found to be significantly associated with outcome from correlational analyses. Table 5 Predictive factors of Global health status/QoL 6–8 months after treatment (T3) (n=46) Predictive factors assessed at baseline Stage of tumour Acceptance coping Pretreatment Global QoL scores Overall model: R 2=0.58; adj. R 2=0.54; F=17.23; df=3,38**. * Pb.005, ** PV.001.

Std h 0.51** 0.43** 0.33*

22


Table 6 Predictive factors of PCS scores 6–8 months after treatment (T3) (n=49) Predictive factors assessed at baseline Pretreatment PCS scores Gender

Std h 0.75* 0.29*

Overall model: R 2=0.68; adj. R 2=0.66; F=46.63; df=2,45*. * Pb.001.

Results Sample characteristics A sample of 82 newly diagnosed HNC patients were recruited into the study prior to treatment. One month after treatment had finished, 68 patients (83% retention rate) completed the first follow-up questionnaire. Six to 8 months after treatment, 50 patients completed the last follow-up questionnaire (61% retention rate). In total, 6% of patients died during the study, 17% had recurrences, 2% entered palliative care, and one patient had severe complications after surgery. Socio-demographic and clinical characteristics of the follow-up samples are shown in Table 1. There were no significant differences between the sample of patients at first follow-up and those who dropped out, excluding stage of cancer, which was found to be higher in patients who failed to complete follow-up assessments. At longitudinal follow-up, patients were significantly less likely to have combined therapy than those who did not respond. Preliminary analysis Table 2 shows the means (S.D.), medians, ranges, and Cronbach’s alpha values for outcome measures completed at baseline, and both follow-ups.

Table 7 Predictive factors of MCS scores 6–8 months after treatment (T3) (n=49) Anxiety Optimism Satisfaction with information (amount and content) Overall model: R 2 =0.37; adj. R 2=0.33; F=8.27; df=3,42**. * Pb.005, ** Pb.005.

Predictive factors assessed at baseline Pretreatment PGI scores Stage of tumour

Std h 0.58** 0.32*

Overall model: R 2=0.51; adj. R 2=0.48; F=16.31; df=2,31**. * Pb.005, ** Pb.001.

To assess whether coping strategies mediated relationships between illness perceptions and outcomes, procedures suggested by Baron and Kenny [29] were attempted. Previous studies [30 – 32] have reported explained variance in outcome (such as QoL or functional status) using illness representations of approximately 20–35%, giving an effect size of 0.25–0.54. Based on an effect size of 0.54, 49 cases are deemed sufficient to detect an R 2 of 0.35 with 80% power at the .05 level of significance.

Predictive factors assessed at baseline

Table 8 Predictive factors of individualized QoL 6–8 months after treatment (T3) (n=40)

Std h 0.41** 0.28* 0.27*

Relationships between baseline psychological factors and coping strategies used over time Table 3 shows the means (S.D.), medians, ranges, and Cronbach’s alpha’s for beliefs at baseline. Relationships between pretreatment beliefs, optimism, satisfaction with information, and coping strategy employed after treatment are shown in Table 4. Illness and treatment representations Many relationships were found between pretreatment illness and treatment beliefs and post-treatment coping strategies (Table 4). Strongest relationships were found between illness identity and long-term coping through self-distraction and venting; longer timeline beliefs and the long-term use of planning; perceptions of worse consequences and reporting venting; and, stronger emotional representations and reporting substance use, denial, self-distraction, and planning. Stronger treatment concerns were related to higher levels of denial. It is worth noting that more relationships were found between pretreatment beliefs and coping strategies at 6–8 months than at 1 month after treatment. Optimism Dispositional optimism was negatively related to timeline beliefs (r= 0.35; P=.002) and positively related to perceptions of control (r=0.27; P=.02) at baseline. Low levels of optimism were associated with higher levels of self-blame shortly after treatment (r= 0.30; PV.01), and higher levels of self-distraction (r= 0.34; PV.05), denial (r= 0.37; PV.05), behavioural disengagement (r= 0.36; PV.01), self-blame (r= 0.30; PV.05), and planning (r= 0.3; PV.05) longer term (6–8 months after treatment).

Table 9 Predictive factors of depression scores 6–8 months after treatment (T3) (n=49) Predictive factors assessed at baseline Timeline Self-blame Satisfaction with information (amount and content) Acceptance Overall model: R 2=0.70; adj. R 2=0.67; F=24.66; df=4,42*. * Pb.001.

Std h 0.51* 0.42* 0.34* 0.33*


Satisfaction with information Higher levels of satisfaction with the type and timing of information were associated with stronger perceptions of control (r=0.28; P=.015) and of the necessity of treatment (r=0.28; P=.015). Satisfaction with the amount and content of information was associated with perceptions of control (r=0.29; P=.012), coherence (r=0.30; P=.008), and illness identity (r= 0.32; P=.006). These results indicated that greater levels of satisfaction with information were related to stronger beliefs in the necessity of treatment and the controllability of the disease, a greater sense of understanding of the illness, and a weaker illness identity, before treatment. Examining the relationships between pretreatment levels of satisfaction with information and coping after treatment, we found that acceptance coping (r= 0.39; P=.007) and use of emotional support (r= 30; P=.03) were associated at T2 only. The predictive factors of QoL after treatment for HNC1 It was predicted that illness and treatment perceptions assessed prior to treatment would explain significant amounts of variation in outcomes over time. Tables 5-9 show pretreatment factors which were found to significantly contribute to variance in outcomes 6–8 months after treatment. Standardized HR-QoL Global Health status/QoL Stage of tumour, acceptance as a coping strategy, and pretreatment QoL scores were found to explain 54% of the variance in global health status/QoL 6–8 months after the end of treatment (adj. R 2=0.54; F=17.23; df=3,38; PV.001). Tumour stage predicted 24% of the variance, with a further 21% predicted by levels of acceptance and 9% predicted by pretreatment global QoL scores. Standardized beta values indicated that lower levels of global QoL were associated with more advanced tumours, high levels of acceptance coping, and lower levels of global QoL prior to treatment (Table 5). Physical Component Summary scores. Physical Component Summary scores were predicted by two factors: pretreatment PCS scores and gender (adj. R 2=0.66; F=46.63; df=2,54; PV.001). Baseline PCS scores accounted for 58% of the variance and an additional 8% was accounted for by gender. Being female was associated with worse PCS scores over time (Table 6). Mental Component Summary scores. A third of the variance in MCS scores was predicted by three variables (adj. R 2=0.33; F=8.27; df=3,42; Pb.005). Levels of anxiety accounted for 19% of the variance, 8% by levels of optimism, and a further 6% by satisfaction with information. 1 Kolmorogov–Smirnov tests demonstrated that key outcome variables 6–8 months post-treatment were not significantly different from a normal distribution to be problematic for entry into parametric analysis.

23

Higher MCS scores were associated with lower levels of anxiety, higher levels of optimism, and greater satisfaction with information (Table 7). Individualized QoL. Half of the variance in individualized QoL scores was predicted by pretreatment PGI scores and stage of tumour (adj. R 2=0.48; F=16.31; df=2,3; Pb.001). The majority of this variance was accounted for by pretreatment levels of QoL (40%) and a further 8% was accounted for by stage of tumour. Later stages of tumour were associated with lower levels of QoL (Table 8). Emotional outcomes Depression Levels of depression after treatment were predicted by illness beliefs, coping strategies and satisfaction with information (adj. R 2=0.67; F=24.66; df=4,42; Pb.001). Stronger beliefs in the illness lasting a long time, high levels of self-blame for the illness, low levels of satisfaction with information, and high levels of acceptance at baseline were related to high levels of depression over time (Table 9). Timeline beliefs and self-blame accounted for the highest amounts of variance in depression (28% and 21%, respectively). Anxiety The only factor to prove predictive of levels of anxiety 6–8 months after treatment was anxiety at baseline (adj. R 2=0.27; F=19.37; df=1,48; Pb.001) which accounted for 27% of the variance. In conclusion, Hypothesis 1 was accepted as there were significant associations between baseline beliefs and coping over time. Hypotheses 2a, b, and d were rejected, as illness and treatment beliefs were not significant predictors of standardized HR-QoL, individualized QoL, or anxiety. However, Hypothesis 2c was accepted as timeline beliefs at baseline were significantly predictive of levels of depression 6–8 months after treatment.2 No mediational relationships were found between illness perceptions and outcomes. Discussion The main aims of this prospective study were to examine whether baseline beliefs were associated with coping over 2 For each of the six regression models, plots of standardized residuals against standardized predicted values were fairly random and evenly dispersed; therefore, data were probably within the limits for meeting assumptions of homoscedasticity and linearity. Histograms of standardized residuals were normally distributed (excluding individualized QoL), indicating that assumptions had been met for normally distributed errors (individualized QoL displayed non-normally distributed errors). The P–P plots of normally distributed residuals represented normal distributions. The assumption of independent errors was met with Durbin–Watson statistics within the acceptable range. Collinearity statistics of tolerance and variance inflation factors were well within acceptable ranges (N0.2 and b10, respectively) indicating that the assumptions of no multicollinearity were met for each model.

24


time and to examine the utility of components of the common sense model in predicting longitudinal outcomes. This is of particular importance as the trend with previous research has been to focus on concurrent relations between variables. Furthermore, if the model proved constructive in understanding adaptive processes in this patient group, potentially modifiable factors found to be predictive of outcomes may be targeted in future intervention studies to enhance patient outcomes. The application of the common sense model to guide understanding of the processes by which HNC patients understand and react to their illness Univariate analyses demonstrated that various relationships existed between pretreatment illness and treatment beliefs and coping over time. In particular, relationships over time between beliefs about the negative consequences of the illness, emotional representations, and illness identity pretreatment, and dmaladaptiveT coping strategies, such as venting, substance use, and denial posttreatment, were found. More dadaptiveT coping, such as use of planning strategies, was associated with baseline perceptions about the likely length of the illness and emotional representations. Previous studies have highlighted the relationship between beliefs and coping using cross-sectional data [33,30]; however, little was known previously as to the relationships between baseline illness representations and coping over time. These relationships are, however, consistent with the common sense model. Previous research has highlighted the evidence linking components of the common sense model to outcomes such as QoL or depression in HNC [7,8] and other illness groups [30,34,35]. However, the majority of this evidence is based on cross-sectional data and thus the direction of causation cannot be established. Components of the common sense model such as illness perceptions and coping strategies were found to be better explanatory factors of judgement-based outcome variables such as HR-QoL and depression at baseline [7,8] than over time as found in the present study. Depression was an exception to this rule, with the majority of the variance predicted by psychological factors of timeline beliefs, selfblame, acceptance coping, and satisfaction with information. This finding is, however, consistent with the model which maintains that a person’s cognitions and behaviour influence outcomes at that time and cognitions and behaviour at one point of time would not necessarily be expected to influence future outcomes. Coping strategy was found to be significantly associated with two of the outcome measures: Global health status/ QoL and depression. Levels of acceptance coping were negatively related to Global health status/QoL and positively related to depression, indicating that lower levels of acceptance were reported by individuals experiencing high levels of QoL and low levels of depression. This may appear

to be contrary to expectation; however, it may be that acceptance is related to severity of symptoms and impairment resulting from the cancer or, indeed, states of denial. Patients reporting low levels of depression and high QoL may be less accepting of their cancer due to denial or because they are experiencing few signs and symptoms to remind them of their cancer diagnosis. The literature also demonstrates that patient beliefs may be more predictive of outcomes than coping mechanisms [15,36,37]. However, previous studies have shown that coping is related to HRQoL (when beliefs are not assessed) [38] and have suggested that pretreatment coping strategies may be an important focus for intervention [39–41]. The role of personality and cultural context has been acknowledged as potentially influential in the formation of representations [42,43] and could potentially influence outcome directly. Optimism was found to be associated with timeline and personal control beliefs (data not shown), and only MCS scores were predicted by optimism. In support of Scheier’s and Carver’s proposition [17], optimism was also found to be directly related to coping, in particular self-blame, denial, and self-distraction. Previous results have indicated that the common sense model is valuable in providing information about the underlying psychological determinants of a variety of outcomes at baseline [7]. The present results have also demonstrated that patient’s beliefs and coping strategies at the time of diagnosis are related to levels of depression over time. The model proved limited for predicting other longitudinal outcomes; however, it does not actually specify direct relationships between illness beliefs and outcomes or cross-lag relationships (over time). Previous studies have reported the role of coping as a mediator of the relationship between illness perceptions and distress [44,45]. Although mediation analyses were attempted in the present study, no significant relationships were found between illness perceptions, coping strategies, and outcomes. It is probable, given the fairly low sample sizes typical of HNC studies, that there was not sufficient power to find any real effects. However, other studies have also failed to find this effect [46]. Implications for intervention Patients’ beliefs about their illness have successfully been the focus for interventions in a variety of illnesses [47,48]. The only belief to be related to outcome was how long the illness would last. Self-blame and acceptance coping were also predictive of outcome. These particular components of the common sense model could be targeted for intervention in the time period between diagnosis and shortly after treatment in order to maximise longitudinal outcomes. These psychological factors were more successful in explaining outcomes than any demographic or clinical factors, which provides promising data for intervening in the case of vulnerable patients. Nevertheless, findings from the


current research indicate that any positive effects of intervention would be limited if based solely on illness representations and coping strategies. Satisfaction with information has been shown to influence long-term outcomes after treatment. In the present study, satisfaction with information was found to be associated with outcomes of depression and MCS scores. With this in mind, and as recommended by other authors [49,50], it would seem judicious to assess HNC patients’ needs for information on an individual basis. The information-giving process could be developed to encourage realistic (but positive) expectations through more contact with specialist nurses. However, research has suggested that it is the inferences individuals make about the information that determine levels of distress rather than the meanings the information giver intends to convey. It has been documented that people are more likely to experience anxiety if the information is interpreted as threatening [51]. This indicates that people interpret the information they have been given within their own framework of ideas and theories of their illness [11], which is why it is of importance to access patients’ views about their illness and treatments in relation to their satisfaction with information prior to and during treatment. Limitations A problem with QoL data in general is that there is an inherent sampling bias towards those who have less severe problems and indeed those who have survived. Follow-up data tend to be biased towards those with earlier stage tumours and therefore less physical consequences of treatment. A comparison of those who were originally recruited into the study and those who remained at follow-up demonstrated a significant difference in tumour stage, with those remaining in the study being diagnosed with earlier stage tumours at the outset. Relatively large attrition rates for a number of reasons limit the generalisability of the results. It was not possible to fully test the model due to its complexity. In particular, it proved difficult to capture the dynamics of when and how changes to outcome occurred and what factors in particular caused the change. The problem with using formulaic measurement times to assess outcomes, for example, at 1 month and 6 months, is that changes cannot be captured exactly when they occur, and the variability between people in terms of when changes occur cannot be observed. A key component of the common sense model is the appraisal process, and similar to the majority of studies, this part of the model was not tested. The appraisal stage of the model is critical in understanding outcomes as it accounts for the two-way process between the beliefs people have and the coping strategy they employ. This then has repercussions for outcome. The appraisal mechanism is rarely assessed as it is difficult to capture when it occurs and how it manifests. Attempts could be

25

made to assess this process through interviewing, although the full extent of this process may be obscured in patients who are less aware of it occurring. Until we fully understand the mechanism of appraisal and its relationship between coping and beliefs, it is impossible to establish how important this process is in influencing outcome. Acknowledgments CDL was supported by a grant from Guy’s and St Thomas’ Charitable Foundation (R020216). Thanks to all the patients and staff from Guy’s and St Thomas’ Hospitals, The Royal Sussex County Hospital, and The Royal Marsden Hospital who assisted with this project. References [1] Carvalho AL, Nishimoto IN, Califano JA, Kowalski LP. Trends in incidence and prognosis for head and neck cancer in the United States: a site-specific analysis of the SEER database. Int J Cancer 2005;114:806 – 16. [2] de Graeff A, de Leeuw RJ, Ros WJ, Hordijk GJ, Battermann JJ, Blijham GH, Winnubst JA. A prospective study on quality of life of laryngeal cancer patients treated with radiotherapy. Head Neck 1999;21:291 – 6. [3] Hammerlid E, Mercke C, Sullivan M, Westin T. A prospective quality of life study of patients with laryngeal carcinoma by tumor stage and different radiation therapy schedules. Laryngoscope 1998;108: 747 – 59. [4] Llewellyn CD, McGurk M, Weinman J. Are psychosocial and behavioural factors related to health related quality of life in patients with head and neck cancer? A systematic review. Oral Oncol 2005; 41:440 – 54. [5] Leventhal H, Colman S. Quality of life: a process view. Psychol Health 1997;12:753 – 67. [6] Leventhal H, Meyer D, Nerenz D. The common sense representation of illness danger. In: Rachman S, editor. Medical Psychology. New York7 Pergamon Press, 1980. pp. 7 – 30. [7] Llewellyn CD, McGurk M, Weinman J. Head and neck cancer: To what extent can psychological factors explain differences in healthrelated quality of life and individual quality of life? Br J Oral Maxillofac Surg 2006;44:351 – 7. [8] Scharloo M, Baatenburg de Jong RJ, Langeveld TPM, van VelzenVerkaik E, Doom-op den Akker MM, Kaptein AA. Quality of life and illness perceptions in patients with recently diagnosed head and neck cancer. Head Neck 2005;27:857 – 63. [9] Horne R, Weinman J. Self-regulation and self-management in asthma: exploring the role of illness perceptions and treatment beliefs in explaining non-adherence to preventer medication. Psychol Health 2002;17:17 – 32. [10] Leventhal H, Nerenz D, Steele DJ. Illness representations and coping with health threats. In: Baum A, Taylor SE, Singer JE, editors. Handbook of Psychology and Health, Volume 4: Social Psychological Aspects of Health. Hillside (NJ)7 Erlbaum, 1984. pp. 219 – 52. [11] Johnson JE, Leventhal H. Effects of accurate expectations and behavioural instructions on reactions during a noxious medical examination. J Pers Soc Psychol 1974;29:710 – 8. [12] Edwards D. Head and neck cancer services: views of patients, their families and professionals. Br J Oral Maxillofac Surg 1998; 36:99 – 102. [13] Mesters I, van den Borne B, De Boer M, Pruyn J. Measuring information needs among cancer patients. Patient Educ Coun 2001; 43:253 – 62.

26


[14] Heijmans M. Coping and adaptive outcome in chronic fatigue syndrome: Importance of illness cognitions. J Psychosom Res 1998; 45:39 – 51. [15] Scharloo M, Kaptein AA, Weinman J, Hazes JM, Willems LNA, Bergman W, Rooijmans HGM. Illness perceptions, coping and functioning in patients with rheumatoid arthritis, chronic obstructive pulmonary disease and psoriasis. J Psychosom Res 1998;44:573 – 85. [16] Hagger MS, Orbell S. A meta-analytic review of the common-sense model of illness representations. Psychol Health 2003;18:141 – 84. [17] Scheier MF, Carver CS. Optimism, coping, and health: assessment and implications of generalized outcome expectancies. Health Psychol 1985;4:219 – 47. [18] Moss-Morris R, Weinman J, Petrie KJ, Horne R, Cameron LD, Buick D. The Revised Illness Perception Questionnaire (IPQ-R). Psychol Health 2002;17:1 – 16. [19] Horne R, Weinman J, Hankins M. The Beliefs about Medicines Questionnaire: The development and evaluation of a new method for assessing the cognitive representation of medication. Psychol Health 1999;14:1 – 24. [20] Scheier MF, Carver CS. Dispositional optimism and physical well-being: the influence of generalized outcome expectancies on health. J Pers 1987;55:169 – 210. [21] Scheier MF, Carver CS. Effects of optimism on psychological and physical well-being: theoretical overview and empirical update. Cogn Ther Res 1992;16:201 – 28. [22] Llewellyn CD, Horne R, McGurk M, Weinman J. Development and preliminary validation of a new measure to assess satisfaction with information amongst head and neck cancer patients: The Satisfaction with Cancer Information Profile (SCIP). Head Neck 2006;28:540 – 8. [23] Carver CS. You want to measure coping but your protocol’s too long: Consider the Brief COPE. Int J Behav Med 1997;4:92 – 100. [24] Carver S, Scheier F, Weintraub JK. Assessing coping strategies: a theoretically based approach. J Pers Soc Psychol 1989;56:267 – 83. [25] Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, Filiberti A, Flechtner H, Fleishman SB, de Haes JC. The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 1993;85:365 – 76. [26] Ware E, Sherbourne DC. The MOS 36-item Short Form Health Survey (SF-36). Med Care 1992;30:473 – 83. [27] Calman KG. Quality of life in cancer patients—a hypothesis. J Med Ethics 1984;10:124 – 7. [28] Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983;67:361 – 70. [29] Baron RM, Kenny DA. The moderator–mediator variable distinction in social psychological research: conceptual, strategic and statistical considerations. J Pers Soc Psychol 1986;51:1173 – 82. [30] Rutter CL, Rutter DR. Illness representation, coping and outcome in irritable bowel syndrome (IBS). Br J Health Psychol 2002;7:377 – 91. [31] Rutter CL, Durham-Hall A, Weinman J, Fidler HM. The importance of illness beliefs on quality of life and psychological status in patients with ulcerative colitis and Crohn’s disease. The British Psychology Society: Division of Health Psychology Annual Conference, 2003. [32] Scharloo M, Kaptein AA, Weinman J, Bergman W, Vermeer BJ, Rooijmans HGM. Patients’ illness perceptions and coping as predictors of functional status in psoriasis: a 1-year follow-up. Br J Dermatol 2000;142:899 – 907. [33] Helder DI, Kaptein AA, van Kempen GMJ, Weinman J, van Houwelingen HC, Roos RAC. Living with Huntington’s disease:

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

Illness perceptions, coping mechanisms, and patients’ well-being. Br J Health Psychol 2002;7:449 – 62. Moss-Morris R, Chalder T. Illness perceptions and levels of disability in patients with chronic fatigue syndrome and rheumatoid arthritis. J Psychosom Res 2003;55:305 – 8. Vaughan R, Morrison L, Miller E. The illness representations of multiple sclerosis and their relations to outcome. Br J Health Psychol 2003;8:287 – 301. Moss-Morris R, Petrie KJ, Weinman J. Functioning in chronic fatigue syndrome: do illness perceptions play a regulatory role? Br J Health Psychol 1996;1:15 – 25. Orbell S, Johnston M, Rowley D, Espley A, Davey P. Cognitive representations of illness and functional and affective adjustment following surgery for osteoarthritis. Soc Sci Med 1998;47:93 – 102. Hassanein KA, Musgrove BT, Bradbury E. Functional status of patients with oral cancer and its relation to style of coping, social support and psychological status. Br J Oral Maxillofac Surg 2001; 39:340 – 5. Allison PJ, Nicolau B, Edgar L, Archer J, Black M, Hier M. Teaching head and neck cancer patients coping strategies: results of a feasibility study. Oral Oncol 2004;40:538 – 44. Chaturvedi SK, Shenoy A, Prasad KM, Senthilnathan SM, Premlatha BS. Concerns, coping and quality of life in head and neck cancer patients. Supportive Care Cancer 1996;4:186 – 90. List MA, Rutherford JL, Stracks J, Haraf D, Kies MS, Vokes EE. An exploration of the pretreatment coping strategies of patients with carcinoma of the head and neck. Cancer 2002;95:98 – 104. Diefenbach MA, Leventhal H. The common-sense model of illness representation: theoretical and practical considerations. J Soc Distress Homeless 1996;5:11 – 38. Landrine H, Klonoff EA. Culture and health-related schemas: a review and proposal for interdisciplinary integration. Health Psychol 1992;11:267 – 76. Fortune DG, Smith JV, Garvey K. Perceptions of psychosis, coping, appraisals, and psychological distress in the relatives of patients with schizophrenia: an exploration using self-regulation theory. Br J Clin Psychol 2005;44:319 – 31. Hagger MS, Chatzisarantis NLD, Griffin M, Thatcher J. Injury representations, coping, emotions, and functional outcomes in athletes with sports-related injuries: a test of self-regulation theory. J Appl Soc Psychol 2005;35:2345 – 74. Kaptein AA, Helder DI, Scharloo M, van Kempen GMJ, Weinman J, van Houwelingen HJC, Roos RAC. Illness perceptions and coping explain well-being in patients with Huntington’s disease. Psychol Health 2006;21:431 – 46. Candy B, Chalder T, Cleare AJ, Wessely S, Hotopf M. A randomised controlled trial of a psycho-educational intervention to aid recovery in infectious mononucleosis. J Psychosom Res 2004;57:89 – 94. Petrie KJ, Cameron LD, Ellis CJ, Buick D, Weinman J. Changing illness perceptions after myocardial infarction: An early intervention randomized controlled trial. Psychosom Med 2002;64:580 – 6. Newell R, Ziegler L, Stafford N, Lewin RJ. The information needs of head and neck cancer patients prior to surgery. Ann R Coll Surg Engl 2004;86:407 – 10. Semple CJ, McGowan B. Need for appropriate written information for patients, with particular reference to head and neck cancer. J Clin Nurs 2002;11:585 – 93. Teasdale K. Information and anxiety: a critical reappraisal. J Adv Nurs 1993;18:1125 – 32.