A psychometric evaluation of the Functional assessment of cancer therapy—kidney symptom index (FKSI-19) among renal cell carcinoma patients suggesting an alternative two-factor structure

Kidney cancer has a yearly worldwide incidence of > 400 000 cases, corresponding to 2.2% of all cancers, with about 175 000 deaths occurring each year, equaling 1.8% of all cancer deaths [1]. Renal cell carcinoma (RCC) is the most common type of kidney cancer, accounting for > 90% of all cases [2]. The incidence of RCC is rising, in particular there is an increase in small renal tumors, mostly attributed to the increased use of tomographic imaging. Localized RCC at diagnosis is mainly treated with radical or partial nephrectomy and for smaller tumors also minimal invasive treatments as radiofrequency ablation [3]. During the last decade there has been a trend toward nephron sparing treatments as partial nephrectomy and ablative invasive treatments. The treatment of metastasized RCC at diagnosis is mainly based on systemic therapies including tyrosine kinase inhibitors and immunotherapies, but surgery of primary tumors and metastases should still be considered when immediate medical treatment is not required [4]. With estimated 5-year RCC survival rates ranging from 95% for stage I patients to 20% for stage IV patients [2], studying the health-related quality of life (HRQoL) experienced by the surviving patients is of large interest.

The general awareness and risk for recurrent disease will have an impact on the HRQoL for patients surgically treated for RCC. A number of treatment and surveillance factors will influence HRQoL. These include postoperative renal function, per- and postoperative complications, and positive surgical margins with increased risks for recurrent disease. Clinical management with adequate disease information and follow-up monitoring might be other factors of importance for the perception of well-being after RCC surgery.

Of the instruments available for measuring HRQoL among cancer patients, the Functional Assessment of Cancer Therapy—Kidney Symptom Index (FKSI) is the most kidney cancer-specific instrument, with the 19-item version (FKSI-19) being the most recent version [2, 5, 6]. FKSI-19 includes four subscales or domains/factors measuring different aspects of HRQoL among kidney cancer patients: Disease-Related Symptoms-Physical (DRS-P), Disease-Related Symptoms-Emotional (DRS-E), Treatment Side Effects (TSE), and Function/Well-Being (FWB). A comprehensive independent large-scale evaluation of the psychometric properties of the FKSI-19 instrument is, however, lacking. Specifically, psychometric methods have not previously been used for examining the factor structure of FKSI-19. The present study intends to rectify this shortcoming, using a large cohort of Swedish RCC patients.

Demographic and clinical symptoms characteristics for the 1731 patients with RCC that participated in the present study are given in Table 1, together with total and domain scores for the FKSI-19 instrument, overall and divided by sex. Almost two-thirds (n = 1133, 65.5%) of the patients were males, which was in line with the overall proportion of males (61.5%) in the NSKCR [8]. The participants were at a mean (SD) age of 66.7 (11.0) years at the time of surgery, with the questionnaire being answered a mean (SD) 8.0 (1.2) months after surgery. Most patients (72.7%) were married or cohabiting, while less than a third (29.8%) had a college or university education. Most patients were at TNM stages T1 (66.4%), N0 (90.1%), and M0 (93.5%), while open surgery was performed on almost half (49.1%) of the patients. The mean (SD) total score of 60.8 (10.6) points on FKSI-19 indicated an overall low degree of symptoms. The differences between males and females were in general small for all variables except marital status, where 77.4% of males were married/cohabiting, compared to 63.9% for females.

The main finding of the present study was that the hypothesized four-factor structure of the FKSI-19 instrument provided a barely acceptable model fit when applied to data from the NSKCR, implying that an alternative factor structure may be contemplated. Moreover, for a factor structure to have good psychometric properties, it is generally recommended that each factor should include ≥ 3 items [14], which is not the case for the hypothesized four-factor structure of the FKSI-19 instrument. An alternative 14-item two-factor solution (FKSI-14) with good psychometric properties was thus derived from the available FKSI-19 items. Notably, the five items that were excluded from FKSI-14 (C2, BRM3, L2, RCC2, and C5) were those items describing explicit physical symptoms (weight loss, fever, coughing, blood in urine, and diarrhea). In contrast, items included in FKSI-14 described more vague symptoms and feelings (e.g., lack of energy, pain, fatigue, weakness, and nausea). Moreover, it should be noted that the five excluded items all had low item reliability, with R² values between 0.012 and 0.207. This suggests that FKSI-14 may be a more robust measure of HRQoL among RCC patients. Finally, it should be noted that the four-factor FKSI-19 structure and the alternative two-factor FKSI-14 structure overall seem to measure the same aspects of HRQoL, with the PMS domain of FKSI-14 largely over-lapping with the DRS-P, DRS-E, and TSE domains of FKSI-19, and the DLFWB domain of FKSI-14 over-lapping with the FWB domain of FKSI-19.

The exclusion of items describing explicit physical symptoms while retaining items describing more vague symptoms and feelings is logical considering that the former are objective signs of a disease while the latter are subjective signs, which are better suited for being measured by HRQoL instruments. In general, it may still be important to measure explicit physical symptoms, in order to provide the clinicians working with the patients with sufficient information regarding individual patient's symptom burden, but these questions could then be asked separately and need not be included in the calculation of the HRQoL index.

However, it should be noted that the vast majority of RCC patients undergo surgery as their primary, and in many cases only, treatment. Surgery gives a relatively short postoperative convalescence, after which patients are living a normal daily life. Moreover, the explicit physical symptoms measured by the five excluded items are rare in this particular group of patients. They are more important after radiation therapy or chemotherapy, treatments that seldom are used for RCC patients. Additionally, in the NSKCR, these specific questions are answered by patients after their first line of treatment, with many patients never suffering from relapse. Actually, in a later stage of the disease with several lines of systemic treatment, there might even be a risk of underestimation of symptoms if the clinician relies on the patient’s answers to these specific questions for his or her assessment.

The present study is the first to use psychometric methods for examining the factor structure of the FKSI-19 instrument. The hypothesized four-factor structure of FKSI-19 provided a barely acceptable model fit when applied to data from the National Swedish Kidney Cancer Register. An alternative 14-item two-factor structure (FKSI-14) with good psychometric properties derived from the available FKSI-19 items may be used as an alternative or complement to the four-factor structure when interpreting the FKSI-19 instrument. The usefulness of this two-factor structure should be tested on other datasets.