The expectations of cancer treatment questionnaire and the experiences of cancer treatment questionnaire: Development and validation

Abstract Patients’ expectations of cancer treatment could impact on their treatment experiences. Research in this area tends to focus on a particular aspect of expectations, and no suitable measure exists to explore patients’ expectations and experiences of treatment. The current study developed and validated two new, matched, measures: The Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) and the Experiences of Cancer Treatment (EXPER-CTQ). Items were generated using the previous expectations literature, alongside findings from qualitative interviews, and refined with assistance from cancer patients and a wider pool of researchers. A sample of 200 cancer patients completed the measures, and factor analysis was performed to validate the EXPECT-CTQ, resulting in a 39-item measure with subscales assessing treatment efficacy, physical side-effects, psychological side-effects, and the impact of treatment on daily life, social life and self-care. The EXPER-CTQ was matched to the EXPECT-CTQ, excluding the treatment efficacy subscale, resulting in a 36-item measure. Reliability analyses were conducted on both measures with good results (α > 0.6). Analyses of the expectations-experiences relationship encompassing a broad range of expectations showed that expectations are positively associated with experiences of cancer treatment. These two new measures are suitable for use in future research exploring both expectations and experiences of treatment for cancer and the implications of gaps between these constructs.


Background
Research indicates that patients' expectations of cancer treatment have the potential to influence patients' experiences of cancer treatment and that there may also be implications when patients' experiences do not match their expectations.One perspective on the expectations-experiences relationship comes from response expectancy theory, which suggests that expecting a certain outcome makes that outcome more likely to occur Kirsch (1985;).This has been explored within the cancer literature in relation to treatment efficacy expectations and treatment side-effect expectations.Treatment efficacy expectations research has focused on the placebo effect and although there is support for the presence of placebo effects in other conditions, in the cancer literature a systematic review concluded that there is little to no evidence of the placebo effect present in cancer treatment Chvetzoff and Tannock (2003;) .Treatment efficacy expectations are also important, however, because they have the potential to influence psychological outcomes.Koller et al.Koller et al. (2000;) studied cancer patients undergoing radiotherapy and found that more positive treatment efficacy expectations were related to higher quality of life (QOL), at least in the short-term.This is consistent with findings from the literature on hope in cancer care, where hope has been found to related to a range of positive psychological outcomes Wnuk et al. (2021;), Rustoen et al. (2010); Research from this perspective has also explored expectations of treatment-related side-effects and three separate meta-analyses have concluded that expecting a particular side-effect makes its occurrence much more likely Devlin et al. (2017;), Colagiuri and Zachaerie (2010;), and Sohl et al. (2009;).For example, Devlin et al.'s Devlin et al. (2017;) meta-analysis was the most recent and comprehensive, including a wider range of cancer and side-effect types but was still restricted in that most research in this area has been conducted with patients undergoing chemotherapy, and concentrated on nausea and vomiting as side-effects.This research therefore suggests that holding negative pre-treatment expectations about side-effects can lead to more negative treatment experiences of the related side-effect.Response expectancy theory and the evidence associated with it would therefore suggest that, at least in terms of treatment efficacy and side-effects, fostering positive expectations may be preferable, to encourage more positive experiences of cancer treatment.
At times, however, research indicates a gap between expectations and experiences and expectancy violations theory maintains that when an individual's expectations are not met this can lead to increased arousal and the evaluation the violation of the expectations as either positive or negative Burgoon and Hale (1988;).This theory has its roots in understanding social interactions and suggests that if a patient's treatment experiences are more negative than previously expected, patients can have worse psychological outcomes than if these expectations had been more realistic.There is little evidence exploring this within the cancer literature but that which is available tends to support this perspective by demonstrating that QOL can be negatively impacted if expectations are higher than experiences Koller et al. (2000;), Maguire et al. (2017), Winterling et al. (2008;), Wan et al. (1997;).Whilst research exploring the gap between expectations and experiences is limited this emphasis highlighted by expectancy violations theory provides a challenge to the conclusions drawn from response expectancy theory.In particular, whilst response expectancy theory promotes positive expectations prior to treatment, expectancy violations theory suggest that more realistic expectations may be of more benefit to the patient.
Further support for providing realistic expectations of cancer treatment is found in the preparation for treatment literature.The theory suggested that providing patients with accurate information, about both the physical medical procedures and the sensations likely to be felt, would allow patients to psychologically prepare for coping Janis (1958).There is a large body of research across different health conditions to support this position, where surgery is the treatment of interest Johnston (1993;;), Suls and Wan (1989;), Johnson et al. (1973;), and Egbert et al. (1964;).This research has also been extended into treatments other than surgery, including radiotherapy and chemotherapy treatments for cancer, with similar findings Aranda et al. (2012;), Burish et al. (1991;), Johnson et al. (1989;), and Johnson et al. (1988;).Similar to an expectancy violations perspective, these studies suggest setting realistic pre-treatment expectations but they do, however have problems with measurement of expectations so this would benefit from further exploration.
Some previous research has therefore explored the association between expectations and subsequent experiences although to date evidence is limited.To provide a more in-depth analysis the current authors conducted a qualitative study of patients' expectations and experiences of cancer treatment involving interviews with sixteen men and women diagnosed with cancer who had undergone one or more forms of cancer treatment Cockle and Ogden (2021;).The results from this study indicated that patients have a much wider range of pretreatment expectations than those often explored in the previous research, including expectations relating to psychosocial side-effects and the impact treatment will have on their daily life.Further, analysis showed that cancer treatment was experienced as a challenge that could be understood in terms of two themes: i) "investing in treatment" reflecting trust in treatment and a sense that treatment was a necessary evil and ii) "encountering tensions or disconnects" illustrating the mismatch between expectations and experiences with patient reporting feeling unprepared, pleasantly surprised or shock.In addition, a transcending theme "searching for stability and certainty" was also described reflecting a desire for normality and a shift in focus between treatment process and outcome.Overall, the results from this qualitative study indicated that in line with an expectancy violations perspective, disconnects were apparent between participants' expectations and experiences and that setting more realistic expectations prior to treatment may be of benefit.However, due to their trust in treatment, participants were also able to shift their focus between the treatment itself and future outcomes as a means to cope with the challenges they face.
Further research is therefore needed to address the broader range of expectations and to clarify this relationship between expectations and experiences quantitatively.How they are measured, however, is problematic and different measures focus on different aspects of expectations, making synthesising findings difficult.There are a number of existing measures of expectations that have been used in the context of cancer such as the Side-effect Expectancy Questionnaire (SEEQ), the Stanford Expectations of Treatment Scale (SETS), the Cancer Therapy Satisfaction Questionnaire (CTSQ), the recovery-related expectations questionnaire (RRE), The Cancer Treatment Survey (CaTS) Devlin et al. (2017;), Winterling et al. (2008;), Younger et al. (2012;), Schofield et al. (2012;), and Abetz et al. (2005;).Some of these measures focus on a particular aspect of expectations, such as treatment efficacy or side-effects, and as such do not facilitate a broader exploration of expectations.Where measures are a little broader, they are either retrospective, or focus on expectations in the post-treatment survivorship period.Several other measures exist that relate expectations of information or care provision, rather than expectations of treatment or impact Schofield et al. (2012;), Peck et al. (2000;), Kravitz et al. (1997;), and Williams et al. (1995;).There are also methodological limitations with some existing measures, such as in item generation being based on existing literature of experiences, rather than encompassing an exploration of patient expectations.
In summary, research indicates that patients' expectations of cancer treatment may relate to their subsequent experiences.To date, however, whilst some measures focus on a narrow range of experiences and address treatment efficacy or side effects, no measures have addressed the wider experiences of cancer treatment, nor have they addressed the relationship between expectations and experiences.The current study therefore aimed to develop and validate two new, matched, measures of patients' expectations of cancer treatment, and their experiences of cancer treatment to address a broader range of expectations and to assess the relationship between these two constructs.

Design
This study used a quantitative cross-sectional design whereby participants were asked to complete newly devised measures of their treatment expectations and treatment experiences.

Developing the new measures
The questionnaire development process involved item generation, face validity and initial piloting for both new measures.

Item generation:
Data from a qualitative study reported elsewhere were used to inform item generation Cockle and Ogden (2021;).The sample for this qualitative study were twelve women and four men who had received, or who were currently receiving, cancer treatment.Participants were encouraged to reflect on their expectations and experiences of cancer treatment freely but within a semi-structured interview schedule.For the purposes of item generation, transcripts were examined for any reference to expectations and extracts taken.A process of pooling similar extracts together then began, with two researchers in consultation over the similarity of statements and agreeing the final pooled groups.Once this was decided, the phrasing deemed most suitable and representative for an item was chosen from amongst the statements in each group, generating an initial item list.
Next existing literatures and measures were consulted to ensure that the items generated reflected the previous research in this area, and that no key concepts were omitted.This also ensured the initial items were over-inclusive to allow for later deletion.Existing measures that informed this process include measures of expectations, such as the Stanford Expectations of Treatment Scale (SETS), Younger et al. (2012;); the Side-Effect Expectancy Questionnaire (SEEQ) Devlin et al. (2017;); and the recovery-related expectations questionnaire (RRE) Winterling et al. (2008;).Validated measures of patients' QOL and functioning were also consulted in order to provide a perspective on what patients might realistically expect Aaronson et al. (1993), Bergner et al. (1981;), Bucks et al. (1996;).
Face validity: An initial version of the Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) was created using the items generated.To assess the face validity of the items, feedback was sought regarding each item's relevance in the scale, whether there was anything missing, whether the scale would measure what it intends, its length, and if there were any other general comments.This was sent to three health psychology researchers and one psychology researcher.For one of these researchers, English was not their first language, so they provided comment on wording from this perspective also.Based on feedback, adjustments were made to item wording and order, and another version of the scale generated.This new version was then sent to ten cancer patients, past and present, inviting their feedback, one of whom was an experienced psychology researcher in addition to having been a cancer patient.The scale was also sent to a Macmillan healthcare professional and psycho-oncology researcher, who provided further feedback.Responses were received from six of the cancer patients.Feedback was synthesised and reviewed item by item, informing several changes to the measure, including wording adjustments, and item addition and removal.The resultant version of the scale was carried forward for further validation, consisting of 64 items, categorised into four areas of expectations: efficacy, physical effects, psychological effects, and impact on daily life.Items chosen were over-inclusive to allow for item removal during validation.This version of the questionnaire was then used to design the Experiences of Cancer Treatment Questionnaire (EXPER-CTQ), a scale that was matched to the EXPECT-CTQ, to allow the relationship between expectations and experiences to be directly analysed.The EXPER-CTQ consisted of 51 items, which resembled all items from the EXPECT-CTQ, except for those from the "efficacy" subscale.

Scale validation
The next step was to validate the measures in a sample of cancer patients.There are many divergent views on the number of participants required to carry out scale validity testing Mundfrom et al. (2005;), Comrey and Lee (1992), Everitt (1975;),Cattell (1978).
Due to this evidence, and for pragmatic and ethical reasons, it was decided that the validation study would aim to recruit three participants per item.Therefore, validation of the EXPECT-CTQ would require a minimum of 192 participants and for the EXPER-CTQ, it would be 153.The remainder of this paper details this validation process and subsequent analyses conducted of patients' expectations and experiences of cancer treatment.

Sample
Two hundred participants completed the expectations measure, of whom 183 also completed the experiences measure.The sample was mostly female (n = 177) with a mean age of 51.6 (ranging 29-77).Most participants were White British, did not have dependent children and were employed at the time of treatment.Participants had either received treatment in the past (n = 135) or were currently receiving treatment (n = 65), for a range of cancer types with a mean time since diagnosis ranging from 2 weeks to 19 years (mean = 3 years).Participants had received a range of cancer treatments including surgery, chemotherapy and radiotherapy, either alone or in conjunction with other modalities, with curative (n = 167) or palliative (n = 33) intent.Participants' demographics are shown in Table 1.

Measures
Participants completed demographic, cancer and treatment information.The newly developed Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) was administered to gain retrospective data about what participants expected before they began treatment for cancer.The matched Experiences of Cancer Treatment Questionnaire (EXPER-CTQ) was delivered in one of two formats, depending on whether participants were currently receiving treatment or had done in the past, to gain data about participants' experiences of cancer treatment.

Procedure
Ethical approval was gained from the University Ethics Committee (UEC/2015/107/FHMS), following which advertisement began on social media.Facebook groups with a connection to cancer were also approached for permission to advertise to their group.200 groups and pages were contacted, 63 of which agreed to advertise the study.Following reading the participant information sheet and giving written consent participants completed the demographic information and expectations and experiences of cancer treatment questionnaires.Data were collected between May and July 2017.

Statistical methods
Data were downloaded and analysed in SPSS.Factor analysis and Cronbach's alphas were used for scale validation and reliability analyses for both the patient expectations (EXPECT-CTQ) and the patient experiences (EXPER-CTQ) questionnaires.Correlational analyses (Pearson's R and Spearman's where appropriate) assessed relationships between expectations and experiences.

Expectations of Cancer Treatment Measure (EXPECT-CTQ): scale validation using factor analysis and reliability analysis
Factor analysis was conducted, initially including the 68 items of the expectations measure, with relevant items reversed, using oblique rotation as it could not be assumed that the factors would be uncorrelated.All analyses were interpreted using a threshold of 0.4 for factor loading Field (2013).
The first factor analysis output produced a 15-factor model, with most factors having good face validity.Some initial approximate labels were given to the first 12 factors, which included factors relating to physical side-effects, psychological effects, impact on daily life and treatment efficacy.Despite this, some items double-loaded, others did not load on any of the factors and the total number of items needed to be reduced so that the final measure would not create undue response burden with future participants.Therefore, the next factor analyses aimed to reduce the number of items that better fit within the factors of the model.This process involved removing items that only loaded on factors lacking face validity, those that double-loaded, and some items that proved problematic.Items were removed systematically and with detailed consideration, and at times returned if their removal had caused undue influence on the model produced.During this process, it became apparent that the items related to weight were problematic.First, items relating to both weight gain and weight loss would load together indicating that individuals could expect to both gain and lose weight.Second, although weight change is a common cancer side-effect, weight loss or weight gain cannot be assumed to be always viewed as negative.The decision was therefore taken to remove the two problematic weight items, leaving only one relating to weight: "my weight will be out of control".Reliability analyses using Cronbach's alpha statistics were conducted at various stages to assist with the identification of items to be deleted, and to assess the reliability of factors produced leading to the removal of two items relating to trust.

Age
Mean 51.6 years range 29-77 Ethnicity White British = 187 (93.5%) White Irish = 3 (1.5%)White Other = 2 (1%) White American = 1 (0.5%) White Ashkenazi Jewish = 1 (0.5%) White and Asian = 1 (0.5%) Mixed Irish and Native American = 1 (0.5%) Mixed White European Indian Caribbean = 1 (0.5%) Indian = 1 (0.5%) Pakistani = 1 (0.5%) Chinese = 1 (0.The final factor analysis produced a 10-factor solution with 38 items.One final assessment of the face validity was conducted and although overall this was deemed to be very acceptable, there was one item that had not been included in this analysis that it was necessary to re-introduce.This item pertained to the physical side-effect of nausea.Given the high prevalence of this side-effect in cancer treatment, it is not justified to have a scale that includes patients' expectations of physical side-effects without having an item measuring nausea.The decision was therefore made to include this in the gastrointestinal side-effects subscale.Cronbach's Alpha reliability analyses conducted on each of the factors.Alpha values are all acceptable and range between 0.66 and 0.87.The final solution for the patient expectations questionnaire (EXPECT-CTQ) is shown in Table 2. Details of all iterations of the Factor Analysis with factor loading and eigen values can be found online Cockle (2019).A copy of the final Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) is available in Appendix A.

Experiences of cancer treatment measure (EXPER-CTQ): factor analysis and reliability analysis
For the purposes of the utility of the experiences measure, it was necessary that the items matched items in the EXPECT-CTQ, with the exception of those relating to treatment efficacy.The included items and factors were analysed using Cronbach's Alpha, the results from which are shown in Table 3. Alpha values were acceptable and ranged between 0.60 and 0.91.The EXPER-CTQ is therefore presented as a 9-factor scale with 36 items, detailed within Table 3.A copy of the final EXPER-CTQ is available in Appendix A. Therefore, at the end of the process of scale development two measures were finalised: The Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) and the matched Experiences of Cancer Treatment Questionnaire (EXPER-CTQ).

Analysing the relationship between expectations and experiences
The data were analysed to assess relationships between expectations and experiences using the subscales of Physical Side-Effects, Psychological Side-Effects, and Impact, which were also summated and analysed as Treatment Consequences.A higher score for expectations indicates more positive expectations, which, for physical side-effects means that a higher score equates to a lower expectation.For this analysis the sample was split according to whether treatment was in the past or was currently occurring.The results showed significant positive correlations between cancer treatment expectations and treatment experiences across all subscales and for both those whose treatment was currently occurring and those whose treatment was in the past, with medium to large effect sizes.Details of the effect sizes and significance values are shown in Table 4.This indicates that the more positive participants' expectations of cancer treatment were, the more positive their experiences were.

Discussion
Previous research addressing patient expectations of cancer treatment has been narrow in its focus and explored patient expectations largely in terms of treatment efficacy and treatment sideeffects.Previous findings from a qualitative study, however, indicated that patients also have expectations about a broader range of factors, including the impact cancer treatment will have socially and on daily life, and the psychological effects of treatment Cockle and Ogden (2021;).The two measures developed in the present study have successfully captured this broader notion of patient expectations and experiences of cancer treatment.The Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) is a 39-item self-report measure that can be used alone or in conjunction with its matched counterpart, the 36-item Experiences of Cancer Treatment Questionnaire (EXPER-CTQ).It reflects how positive or negative patients' expectations are overall towards their cancer treatment based on their expectations of physical side-effects (appearance, pain, fatigue, and gastrointestinal), psychological side-effects (negative emotion and positive outlook), and impact (self-care, impact on daily life, and social impact); and how positive or negative their expectations related to treatment efficacy are.The experiences measure reflects how positive or negative patients' experiences of cancer treatment are based on physical and psychological sideeffects, and impact.The current exploratory validation of the EXPECT-CTQ suggests that it successfully represents the included expectations, with well-defined subscales.Reliability analyses suggest that all subscales on both the EXPECT-CTQ and the EXPER-CTQ have good internal consistency and are therefore reliable.
The current study also explored the relationship between expectations and experiences of cancer treatment, in a broader sense than has been included in the previous literature, using these newly developed measures.The results indicate that there were significant positive correlations between expectations and experiences on all subscales, including when these were combined to study Treatment Consequences, indicating that an overall more positive expectation is related to an overall more positive experience.Effect sizes for all relationships were medium or large and this held across those who had received treatment in the past and those currently receiving treatment.This could indicate support for a response expectancy perspective by suggesting that expectations predict experiences and therefore it is better to have more positive expectations, in order to have more positive experiences Kirsch (1985;).Whilst data was cross sectional which limits conclusions about prediction, this is the first attempt at using a broader measure of expectations and experiences to examine the relationship between the two variables shows that this warrants further attention with prospective longitudinal methods.

Study limitations
There are some limitations, however, that need to be considered.Primarily, all participants provided retrospective data on their expectations of cancer treatment, as they were measured during or after treatment.The initial intension of the measure was to be used in a prospective manor by measuring patients' expectations prior to their cancer treatment but recruiting a sample of this nature was problematic, in the quantities required for validation.The time post-diagnosis and prior to treatment is a very difficult and sensitive time for patients.It may also be a time when altruism is understandably low, as patients are attempting to cope with the shock and fear of a cancer diagnosis.In addition, pragmatically, there is often a small window of opportunity in terms of recruiting patients at the correct timepoint, as there may well be very little time between diagnosis and initiation of treatment.The decision was therefore made to gain retrospective data that would inform item deletion and reduce the scales to a more manageable length.Second, and relatedly, both expectations and experiences data were collected at the same time which limits explorations of the expectations-experiences relationship.Measuring the factors so close together could have led to more contamination of responses between the two factors; differentiating between expectations and experiences may be more challenging when the questions are presented concurrently.Finally, the use of social media to recruit participants may have led to a younger sample of cancer patients than may have been obtained in a clinical environment.These limitations do not mean that findings should be dismissed but they should be treated with appropriate caution with these in mind.

Clinical implications
The results from the present study indicate that these broader expectations and experiences of cancer treatment can be measured using the new tools and are related to each other.If expectations predict and change subsequent experiences, then health care professionals working with patients pre-treatment could use these tools to assess patient expectations and manage them in ways to promote better subsequent experiences.Therefore more positive expectations may promote a more subsequent experience which could be encouraged as patients are being supported at the start of their cancer journey.
To conclude, whilst patients' expectations of cancer treatment could impact on their treatment experiences no research to date has explored the relationship between these two factors.The present study therefore developed two new, matched, measures which showed good reliability and validity: The Expectations of Cancer Treatment Questionnaire (EXPECT-CTQ) and the Experiences of Cancer Treatment (EXPER-CTQ).These measures encompass broader expectations than previously assessed in terms of treatment efficacy, physical side-effects, psychological sideeffects, and the impact of treatment (on daily life, social life and self-care) and analysis indicated that expectations are associated with experiences of cancer treatment within these same domains.Further longitudinal research is needed to explore the effects of expectations on experience and the impact of the gap between these two components which in turn would have clear implications for clinical practice and supporting patient with cancer to have the best experience of cancer treatment possible.You are free to: Share -copy and redistribute the material in any medium or format.Adapt -remix, transform, and build upon the material for any purpose, even commercially.The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms: Attribution -You must give appropriate credit, provide a link to the license, and indicate if changes were made.You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.

No additional restrictions
You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.

©
2023 The Author(s).This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license.
and dialog with, expert editors and editorial boards • Retention of full copyright of your article • Guaranteed legacy preservation of your article • Discounts and waivers for authors in developing regions Submit your manuscript to a Cogent OA journal at www.CogentOA.com