The development of the Chemotherapy Symptom Assessment Scale (C-SAS): a scale for the routine clinical assessment of the symptom experiences of patients receiving cytotoxic chemotherapy

https://doi.org/10.1016/S0020-7489(00)00106-1Get rights and content

Abstract

The management of symptoms experienced by patients receiving cytotoxic chemotherapy influences quality of life during treatment. Symptom management may be improved through a structured approach to symptom assessment. This paper describes the development of the Chemotherapy Symptom Assessment Scale (C-SAS), a 24-item scale designed for the routine assessment of symptoms experienced by patients receiving cytotoxic chemotherapy. The scale development process focused both upon the psychometric properties and the clinical usefulness of the scale. Patients and health professionals played a significant role in item selection and scale design in order to maximise the clinical utility of the C-SAS.

Introduction

The management of symptoms experienced by patients receiving cytotoxic chemotherapy influences their quality of life during treatment (Richardson et al., 1988). A wealth of empirical evidence indicates that detailed and specific knowledge of patients’ experiences is vital to effective side-effects management (Richardson, 1995; Burish and Tope, 1992; Hill, 1992; Given, 1991; Dodd, 1988). Different approaches to symptom assessment may be adopted, from unstructured communication between patients and health professionals to the use of documentation such as checklists or diaries. Nursing staff who administer chemotherapy may argue that they ‘assess’ patients in the communication that occurs during treatment. However, research indicates that nurses do not accurately perceive the extent of chemotherapy-related problems suffered by the individual patient, particularly less ‘visible’ problems such as nausea (Holmes and Eburn, 1989). In part this may be explained by evidence which suggests that communication in oncology nursing is characterised by avoidance of patients’ expression of their concerns and poor appreciation of patients’ needs (Wilkinson, 1991). Nurses may not routinely detect and monitor the concerns of individuals who have cancer, and both a poor level of facilitative communication and distancing and blocking tactics are common (Dennison, 1995). Therefore, poor clinical practice in part may be due to the lack of a routine and systematic approach to the assessment of patients’ symptoms.

Research suggests that the systematic assessment of symptoms is associated with reduced symptom distress over time (Sarna, 1998). Structured symptom assessment encourages patients to report in a systematic way the symptoms they are experiencing at each treatment cycle. This information allows clinical staff to follow patients’ experiences throughout their course of treatment and to evaluate the efficacy of treatments used in symptom control (Osoba, 1993). Structured symptom assessment may be achieved through the use of a structured interview or checklist. Whilst providing a prompt for discussions about symptom experience, the usefulness of ‘home-produced’ checklists is limited. They may lack comprehensiveness and validity, and often fail to incorporate a scoring method, and so are unsuitable for quantifying change in symptom experience over time. The use of a structured assessment tool provides a standard scoring system allowing both the documentation of change over time and the evaluation of clinical interventions and nursing care (Conner and Eggert, 1994).

Assessment tools for routine clinical use must combine clinical usefulness and psychometric robustness. If an assessment tool is to be clinically useful, it must be simple, appropriate for the intended use in clinical practice, acceptable to patients, and should include a clear and interpretable scoring system (Corcoran and Fischer, 1987). An assessment tool must also demonstrate the psychometric properties of reliability and validity.

Reliability refers to whether the tool is measuring what is intended in a reproducible manner. Reliability may be measured in several ways. Measures of internal consistency explore whether the items in the tool are homogeneous and capable of measuring a single concept. If a tool intends to measure several concepts, the internal consistency of each subscale is measured independently. Measures of stability explore the reproducibility of scores on different administrations of the measure. Stability may be measured by examining the degree of agreement between two observers (inter-observer reliability), through agreement between observations made by the same observer on two occasions (intra-observer reliability), or through assessment on two occasions separated by time (test–retest reliability). Where assessments are in a self-report form, it is possible only to assess stability using a test–retest approach. Symptom experience is subjective and may fluctuate over time, therefore tools designed to assess such experiences may not demonstrate high levels of stability; however, the tool should demonstrate internal consistency exceeding α=0.8 (Streiner and Norman, 1995).

Validity refers to whether the tool measures what it is intended to measure (Messick, 1980). Validity may be measured in a variety of ways.

Content validity examines whether the tool measures all the relevant domains or content. Content validity may be inferred by the process of item generation used in the tool's development. As a minimum, the item generation process for a tool should be systematic and include a review of the relevance of items to be included by an expert panel.

Criterion validity refers to the degree of agreement between the tool being evaluated and other tools designed to measure the same concept. In many cases a new tool is developed because no other exists; in this case attention needs to be paid to demonstrating construct validity. Where comparable tools exist, both tools should be administered, and a reasonable correlation (r=0.4–0.8) should be demonstrated between related items.

Construct validity is demonstrated by linking the attribute that is being measured with another attribute via a hypothesis or construct. Such hypotheses often explore the differences between two or more populations who would be expected to have different amounts of the property which is being assessed by the tool. The hypothetical construct may be tested by administering the tool to the two samples and examining whether the observed relationship supports the hypothesis. A tool should demonstrate some evidence of construct validity, particularly in the absence of comparable scales allowing criterion validity testing. Construct validation is an ongoing process requiring the testing of a series of hypothetical constructs (Cronbach and Meehl, 1955).

A variety of instruments are employed for chemotherapy symptom assessment. Clinicians traditionally have used toxicity criteria such as the ‘National Cancer Institute Common Toxicity Criteria’ or ‘World Health Organisation Toxicity Grading’ to grade patients’ symptoms or side-effects. In terms of symptom management, toxicity criteria used alone have limited utility. Toxicity criteria are rated by the clinician rather than by patients themselves; however, research indicates that health professionals do not accurately perceive the extent of patients’ symptom experiences (Grossman et al., 1991; Slevin et al., 1988), suggesting that symptom assessment should be broader than toxicity assessment alone and should include the use of a tool which records patients’ self-reports of symptom experience.

Many symptom assessment scales have been developed for research purposes (Portenoy et al., 1994), however their suitability for routine clinical practice has not been established. Instruments incorporating symptom assessment, such as the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire for Cancer (QLQ-C30) (Aaronson et al., 1993), have been developed to assess cancer-related quality of life rather than chemotherapy-related symptom experience specifically. Other scales have been developed as outcome measures to evaluate the efficacy of different types of chemotherapy (Schipper et al., 1984). At a practical level, symptom assessment scales developed for research purposes are often lengthy and too time consuming to use routinely in clinical practice.

If symptom assessment scales are to be effective in patient care their development must be informed by a greater understanding of both patient experience and the practicalities of symptom assessment in clinical practice. Tools must be short, simple and relevant, with a well-defined and standardised scoring system. The development of the tool must ensure that its content and design are acceptable to both the patients and the health professionals that will use it. There is a clear need for a psychometrically robust instrument designed specifically for use in the routine clinical assessment of patients’ symptom experience during cytotoxic chemotherapy treatment.

This paper describes the development of the Chemotherapy Symptom Assessment Scale (C-SAS), a scale designed for the routine assessment of symptoms experienced by patients receiving cytotoxic chemotherapy for common cancers in outpatient facilities. It is intended that the scale will be completed by patients before each treatment, with the responses providing a focus for discussion and treatment planning with clinical staff. The C-SAS project places a strong emphasis on the involvement of patients and clinical staff in scale development, to ensure that the scale meets the needs of those who will use it in practice. It is hoped that the routine use of the C-SAS in clinical practice will result in improvements in patient experience and in the quality of care provided.

A previous descriptive study of chemotherapy side-effects generated a large pool of potential items for inclusion in the C-SAS (Sitzia et al., 1997). This paper describes a six-stage item generation and reduction process to produce a prototype C-SAS for psychometric testing. A first aim was to establish the content validity of the items included in the item pool. However, for the C-SAS to be clinically useful it is important that it is short and simple whilst being comprehensive enough to maintain content validity; therefore, a second aim was to reduce the number of items in a systematic and meaningful way through the involvement of both patients and clinicians, the potential users of the C-SAS.

Section snippets

Method and results

Item selection for the C-SAS involved six stages (Table 1). Due to the sequential nature of the scale development process, method and results are presented together for each stage.

Item generation and selection process

A robust process of item generation and selection is fundamental to the validity of a new clinical assessment instrument. We attempted to use a diverse range of approaches in this process, and while the process was lengthy and somewhat complex, we feel this diversity was invaluable in developing a scale suitable for routine clinical use. The six-stage item generation and selection process allowed the original 64-item pool generated for the WCQ to be reduced to 24 items in a meaningful and

Conclusion

Assessment scales for use in clinical practice must be both psychometrically robust and clinically useful. The clinical utility of a scale may be improved by involving the potential users of the instrument throughout the scale development process. The comprehensive six-stage approach to item generation described in this paper has allowed the development of a clinically focused symptom assessment scale and established the content validity of the C-SAS. Further psychometric evaluation is now

Acknowledgements

This scale has been developed by Worthing and Southlands Hospitals NHS Trust in collaboration with Bromley Hospitals NHS Trust. The work reported here forms the first phase of a project to develop and psychometrically evaluate the Chemotherapy Symptom Assessment Scale. This study was funded by NHS Executive South East R&D Directorate.

We are most grateful to all the patients and hospital staff who have given up their time to support this project, and the Project Steering Group who have provided

References (27)

  • R.A Darton

    Rotation in factor analysis

    The Statistician

    (1980)
  • S Dennison

    An exploration of the communication that takes place between nurses and patients whilst cancer chemotherapy is being administered

    Journal of Clinical Nursing

    (1995)
  • P.A Field et al.

    Nursing ResearchThe Application of Qualitative Approaches

    (1985)
  • Cited by (51)

    • The effect of telephone triage on symptom management in patients with cancer undergoing systemic chemotherapy: A randomized controlled trial

      2022, European Journal of Oncology Nursing
      Citation Excerpt :

      An increase or decrease in the median value refers to an increase or decrease in symptom severity or degree of discomfort. In the test-retest reliability analysis of the scale, “Weighted Kappa” was found as 0.61 ± 0.17 for the frequency of symptoms, 0.60 ± 0.23 for the severity of symptoms, and 0.56 ± 0.23 for the discomfort of symptoms (Brown et al., 2001). The Turkish validity and reliability study of the scale was conducted by Aslan et al. (2006).

    • Remote real-time monitoring for chemotherapy side-effects in patients with blood cancers

      2017, Collegian
      Citation Excerpt :

      Where a side-effect/impairment was reported, patients rated the severity and the bother caused by the issue. Questions were derived from the Chemotherapy Symptom Assessment Scale (Brown, Sitzia, Richardson, Hughes, & Hannon, 2001), with severity descriptors adapted from the Common Criteria for Chemotherapy Adverse Events (US Department of Health and Human Services, 2009) (Fig. 2). Patient data was transmitted to a secure server in real-time via Global Packet Radio Services.

    • Effects of inhaled ginger aromatherapy on chemotherapy-induced nausea and vomiting and health-related quality of life in women with breast cancer

      2015, Complementary Therapies in Medicine
      Citation Excerpt :

      As many as 80% of patients who received anthracycline-based chemotherapy and cyclophosphamide; a commonly prescribed adjuvant regimens for breast cancer, did experience some degree of nausea and vomiting.3,4 Poorly controlled CINV symptoms may result in multiple physiologic consequences, pose impact on patient's quality of life (QoL) and alter patient's compliance with treatment.5–7 In view of the gaps in the current practices, attention given to the use of complementary and alternative medicine (CAM) as an adjuvant treatment has recently been escalating.

    • Effect on symptom control of structured information given to patients receiving chemotherapy

      2014, European Journal of Oncology Nursing
      Citation Excerpt :

      Patient information form (PIF): this is a 19-question form requesting information such as age, sex, educational status, marital status, diagnosis, duration of disease, number of courses, chemotherapy regimen, home town, and social support of the patient. C-SAS: the validity and reliability analysis of the study which was developed by Brown et al. (2001) was performed by Aslan and Vural (2006). The C-SAS includes the 24 chemotherapy symptoms observed in cancer patients receiving chemotherapy.

    View all citing articles on Scopus
    View full text