- Top of page
- MATERIALS AND METHODS
- FUNDING SUPPORT
- CONFLICT OF INTEREST DISCLOSURES
Patients receiving chemotherapy (CTX) experience multiple cooccurring symptoms. On average, these patients report 10 unrelieved symptoms that have a negative impact on their functional status and quality of life (QOL). However, a significant amount of interindividual variability exists, with some patients experiencing a few symptoms whereas others experience every symptom associated with a given CTX regimen. The demographic and clinical characteristics that contribute to this interindividual variability in patients' symptom experiences warrant investigation so that high-risk patients can be identified and preemptive symptom management interventions can be initiated.
Previous work from our research team focused on the identification of these high-risk patients based on an evaluation of their experiences with the 4 most common symptoms associated with cancer and its treatment (ie, pain, fatigue, sleep disturbance, and depression).[2-6] Across 5 separate studies, using either cluster analysis or latent class analysis (LCA), 3 to 5 distinct subgroups of patients were identified. It is interesting to note that across all 5 studies, 1 subgroup of patients was characterized as having low levels of all 4 symptoms and another subgroup was characterized as having high levels of all 4 symptoms. In these studies, compared with patients with low levels of pain, fatigue, sleep disturbance, and depression, patients in the “All High” subgroup were significantly younger and reported lower functional status and decreased QOL.[2-6] In the 2 studies that evaluated for differences in clinical characteristics among the patient subgroups,[2, 6] no differences were identified.
In another group of studies[7, 8] that used symptom occurrence ratings from the Memorial Symptom Assessment Scale (MSAS) to identify high-risk patients, only 2 distinct subgroups were identified, namely patients with low and high symptom occurrence rates. Again, in both of these studies, although clinical characteristics were not associated with subgroup membership, patients in the high symptom subgroup reported decrements in functional status and QOL. The reason for the inconsistent number of subgroups identified across these 7 studies[2-8] may relate to the number of symptoms evaluated, whether symptom occurrence or severity ratings were used to create the patient subgroups, and the statistical procedures used to identify the subgroups, as well as the relatively small sample sizes.
In the era of precision medicine, the specialty of oncology has led efforts to identify distinct tumor subtypes for several cancers (eg, breast cancer[11, 12] and lung cancer) based on tumor-specific characteristics and molecular profiles. The goal of these efforts is to develop mechanistically based cancer treatments. Despite some limitations, the emerging evidence cited above suggests that similar studies need to be performed to identify distinct subgroups of patients who will require more targeted symptom management interventions while undergoing cancer treatment.[2-8] The purposes of the current study, performed in a sample of 582 oncology outpatients who were receiving CTX were to identify subgroups of patients based on their distinct experiences with 25 commonly occurring symptoms and to identify demographic and clinical characteristics associated with subgroup membership. In addition, differences in QOL outcomes were evaluated.
- Top of page
- MATERIALS AND METHODS
- FUNDING SUPPORT
- CONFLICT OF INTEREST DISCLOSURES
To our knowledge, the current study is the first to use LCA to identify 3 distinct subgroups of patients based on their reports of the occurrence of 25 common symptoms before their next cycle of CTX. Consistent with our previous studies,[2, 5, 6] approximately 14% of the patients reported relatively high occurrence rates for all 25 symptoms. The mean number of symptoms reported by patients in the All High class (ie, 20.3) is higher than the mean of 10 symptoms reported in cross-sectional studies that did not use specific analytic techniques to identify interindividual variability in patients' symptom experiences. Equally important, patients in the Low class reported an average of 6 symptoms, which constitutes a fairly high symptom burden. Findings from the current study suggest that rather than simply reporting the mean number of symptoms, future studies should use the types of statistical approaches used in this and other studies[2, 5-8] to be able to identify oncology patients at higher risk of increased symptom burden. The reliance on mean values for the total number of symptoms will overestimate and underestimate symptom burden and not allow for the identification of patients who require more intensive symptom management interventions.
Although none of the clinical characteristics, except KPS and comorbidity, was found to be associated with class membership, several demographic characteristics differentiated among the 3 latent classes. Consistent with previous reports,[5, 34, 35] younger patients were more likely to be in the All High class. Several potential explanations may account for this finding: older patients may receive lower doses of CTX[36, 37]; age-related changes may occur in the hypothalamic-adrenal-pituitary axis that mediate the occurrence of cancer-related symptoms; or older patients may experience a “response shift” in their perception of symptoms.[39, 40] Other characteristics that differentiated among the classes were sex and ethnicity, with a higher percentage of women and nonwhite individuals being in the All High class. Additional research is warranted because findings regarding sex[41-45] and ethnic[46, 47] differences in the occurrence and severity of symptoms in oncology patients are inconsistent.
Equally important, several socioeconomic characteristics were found to be associated with a higher symptom burden. The finding that marital status distinguished among the 3 latent classes may be related to perceived levels of social support. In several studies, oncology patients who reported higher levels of social support reported lower levels of depressive symptoms.[48-50] Although social support was not measured in the current study, this hypothesis is supported by the finding that patients in the Moderate and All High classes reported significantly lower social functioning scores on both the generic and disease specific measures of QOL. Consistent with previous reports,[51-53] patients in the All High class were more likely to report a lower annual household income. The reason for this disparity warrants investigation in future studies.
It is interesting to note that for both the generic and disease-specific measures of QOL, patients in the All High class reported worse QOL outcomes than patients in the Low and Moderate classes. Compared with the Low class, decrements in functional status reported by the Moderate (d = 0.5) and All High (d = 1.1) classes represent not only statistically significant but clinically meaningful differences in KPS scores.[54, 55] The effect size indicator (ie, d) equals the difference between the 2 group means in standard deviation units. Similar effect sizes were found for the various subscales of both the generic and disease-specific QOL measures. In terms of the SF-12 PCS scores, all 3 classes had scores below the US population mean score of 50. The effect size calculations for differences between the Low and Moderate classes, as well as the Low and All High classes, in PCS scores were d = 0.6 and d = 1.0, respectively.
For the SF-12 MCS scores, the Moderate and All High classes had scores below national norms. The effect size calculations for the MCS scores indicated clinically meaningful differences between the Low and Moderate (d = 0.5) as well as the Low and High (d = 1.3) classes. Finally, the effect size calculations for the total score on the disease-specific measure of QOL (ie, QOL-PV[19-21]) identified clinically meaningful differences when the Low class was compared with the Moderate (d = 0.9) and All High (d = 1.6) classes. Taken together and consistent with previous reports,[2, 5-8] these findings emphasize the significant impact that the cooccurrence of multiple symptoms has on patients' ability to function and their QOL.
Although consistent with previous reports,[2-4, 6-8] a surprising finding from the current study is that with the exception of the decrements in functional status and severity of comorbidities, none of the disease and treatment characteristics was found to be associated with class membership. The relatively small and heterogeneous samples in terms of cancer treatment may explain the lack of associations between disease and treatment characteristics observed in previous studies. However, in the current study, the large sample size as well as the relatively even distribution of cancer diagnoses, reasons for CTX, and extent of metastatic disease across the 3 latent classes suggests that alternative explanations are plausible. One potential explanation for the lack of disease and treatment effects is that patients with a higher disease burden and more severe symptoms declined study participation. Another explanation for the lack of disease and treatment effects is that interindividual variability in patients' symptom experiences may be associated with genetic and epigenetic determinants. This hypothesis is supported by work from our research team and others on the association between several candidate genes and individual symptoms (eg, pain,[32, 56, 57] fatigue,[58-63] depression,[30, 64, 65] and sleep disturbance) in oncology patients. Studies currently are underway in the study laboratory to identify specific biomarkers associated with membership in the All High class identified herein.
Several study limitations need to be acknowledged. Patients were recruited at various points in their CTX. In addition, the types of CTX were not homogeneous. Although we cannot rule out the potential contributions of clinical characteristics to patients' symptom experiences, the relatively similar percentages of cancer diagnoses, reasons for current treatment, time since diagnosis, and evidence of metastatic disease suggest that the classes were relatively similar in terms of disease and treatment characteristics. Although it is possible that patients in the Low class were receiving more aggressive symptom management interventions, the occurrence rates for the 5 most common symptoms (ie, lack of energy, difficulty sleeping, pain, feeling drowsy, and difficulty concentrating) were relatively proportional across the 3 classes. The study used symptom occurrence rates in the LCA and did not evaluate for changes over time in latent class membership. It is possible that using ratings of severity or distress to create the latent classes would provide additional information on interindividual differences in the symptom experience of these patients.
In conclusion, the findings from the current study support the clinical observation that some oncology patients experience a differentially higher symptom burden during CTX. Risk factors identified in this study included younger age, being female, being nonwhite, having lower levels of social support, being of lower socioeconomic status, having poorer functional status, and having a higher level of comorbidity. These high-risk patients experience significant decrements in QOL. Future studies will focus on the identification of molecular mechanisms that contribute to this high-risk phenotype, as well as the identification of latent classes using patients' ratings of symptom severity and distress.