Health-related quality of life and predicting survival in cancer: not a simple matter

Machine learning with PROs in breast cancer surgery; caution: Collecting PROs at baseline is crucial

As high breast cancer survival rates are achieved nowadays, irrespective of type of surgery performed, prediction of long‐term physical, sexual, and psychosocial outcomes is very important in treatment decision‐making. Patient‐reported outcomes (PROs) can help facilitate this shared decision‐making. Given the significance of more personalized medicine and the growing trend on the application of machine learning techniques, we are striving to develop an algorithm using machine learning techniques to predict PROs in breast cancer patients treated with breast surgery. This short communication describes the bottlenecks in our attempt to predict PROs.

The prevalence of patient-reported dysphagia and oral complications in cancer patients

PURPOSE

Research investigating swallowing problems (dysphagia) and complications within the oral cavity in non-head and neck cancer patients is limited. The purpose of this study was to determine the prevalence of patient-reported dysphagia and oral complications in all cancer patients and to examine the relationships between cancer types, oral complications and dysphagia.

METHODS

A cross-sectional study was conducted at a specialist cancer centre in Australia. Data on patient-reported dysphagia and oral complications were collected using the Vanderbilt Head and Neck Symptom Survey (version 2.0) which was completed by participants in one of three settings: inpatients, ambulatory patients receiving chemotherapy, or ambulatory patients receiving radiotherapy.

RESULTS

Data were collected on 239 patients, receiving treatment for 14 cancer types. The proportion of patients who reported dysphagic symptoms were as follows: any dysphagia (54%); dysphagia for liquids (20%); and dysphagia for solids (46%). Significantly more head and neck patients and significantly fewer breast patients reported dysphagia, but there were no differences between other tumour types. Oral symptoms across all cancer types were reported at the following rates: taste changes (62%); xerostomia (56%); voice changes (37%); smell changes (35%); thick mucous (33%); difficulty with teeth/dentures (25%); mouth/throat pain (20%); and trismus (19%).

CONCLUSIONS

This is the first time comprehensive data on dysphagia and oral complications across all cancer patients have been collected. We have identified that dysphagic symptoms and oral complications-which have implications for quality of life and function-are common in all cancer patients, not just those with head and neck cancer.

Prediction of long-term mortality by preoperative health-related quality-of-life in elderly onco-surgical patients

Objective

Aim of this study was to evaluate the association between preoperative health-related quality of life (HRQoL) and mortality in a cohort of elderly patients (>65 years) with gastrointestinal, gynecological and genitourinary carcinomas.

Design

Prospective cohort pilot study.

Setting

Tertiary university hospital in Germany.

Patients

Between June 2008 and July 2010 and after ethical committee approval and written informed consent, 126 patients scheduled for onco-surgery were included. Prior to surgery as well as 3 and 12 months postoperatively all participants completed the EORTC-QLQ-C30 questionnaire (measuring self-reported health-related quality of life). Additionally, demographic and clinical data including the Mini Mental State Examination (MMSE) were collected. Surgery and anesthesia were conducted according to the standard operating procedures. Primary endpoint was the cumulative mortality rate over 12 months after one year. Changes in Quality of life were considered as secondary outcome.

Results

Mortality after one year was 28%. In univariable and multivariable logistic regression analysis baseline HRQoL self-reported cognitive function (OR per point: 0.98; CI 95% 0.96–0.99; p = 0.024) and higher symptom burden for appetite loss (per point: OR 1.02; CI 95% 1.00–1.03; p = 0.014) were predictive for long-term mortality. Additionally the MMSE as an objective measure of cognitive impairment (per point: OR 0.69; CI 95% 0.51–0.96; p = 0.026) as well as severity of surgery (OR 0.31; CI 95% 0.11–0.93; p = 0.036) were predictive for long-term mortality. Global health status 12 months after surgery was comparable to the baseline levels in survivors despite moderate impairments in other domains.

Conclusion

This study showed that objective and self-reported cognitive functioning together with appetite loss were prognostic for mortality in elderly cancer patients. In addition, impaired cognitive dysfunction and severity of surgery were predictive for one-year mortality whereas in this selected population scheduled for surgery age, gender, cancer site and metastases were not.

Exercise interventions on health-related quality of life for cancer survivors

BACKGROUND

Cancer survivors experience numerous disease and treatment-related adverse outcomes and poorer health-related quality of life (HRQoL). Exercise interventions are hypothesized to alleviate these adverse outcomes. HRQoL and its domains are important measures for cancer survivorship.

OBJECTIVES

To evaluate the effectiveness of exercise on overall HRQoL and HRQoL domains among adult post-treatment cancer survivors.

SEARCH METHODS

We searched the Cochrane Central Register of Controlled Trials (CENTRAL), PubMed, MEDLINE, EMBASE, CINAHL, PsycINFO, PEDRO, LILACS, SIGLE, SportDiscus, OTSeeker, and Sociological Abstracts from inception to October 2011 with no language or date restrictions. We also searched citations through Web of Science and Scopus, PubMed's related article feature, and several websites. We reviewed reference lists of included trials and other reviews in the field.

SELECTION CRITERIA

We included all randomized controlled trials (RCTs) and controlled clinical trials (CCTs) comparing exercise interventions with usual care or other nonexercise intervention to assess overall HRQoL or at least one HRQoL domain in adults. Included trials tested exercise interventions that were initiated after completion of active cancer treatment. We excluded trials including people who were terminally ill, or receiving hospice care, or both, and where the majority of trial participants were undergoing active treatment for either the primary or recurrent cancer.

DATA COLLECTION AND ANALYSIS

Five paired review authors independently extracted information on characteristics of included trials, data on effects of the intervention, and assessed risk of bias based on predefined criteria. Where possible, meta-analyses results were performed for HRQoL and HRQoL domains for the reported difference between baseline values and follow-up values using standardized mean differences (SMD) and a random-effects model by length of follow-up. We also reported the SMDs between mean follow-up values of exercise and control group. Because investigators used many different HRQoL and HRQoL domain instruments and often more than one for the same domain, we selected the more commonly used instrument to include in the SMD meta-analyses. We also report the mean difference for each type of instrument separately.

MAIN RESULTS

We included 40 trials with 3694 participants randomized to an exercise (n = 1927) or comparison (n = 1764) group. Cancer diagnoses in study participants included breast, colorectal, head and neck, lymphoma, and other. Thirty trials were conducted among participants who had completed active treatment for their primary or recurrent cancer and 10 trials included participants both during and post cancer treatment. Mode of the exercise intervention included strength training, resistance training, walking, cycling, yoga, Qigong, or Tai Chi. HRQoL and its domains were measured using a wide range of measures.The results suggested that exercise compared with control has a positive impact on HRQoL and certain HRQoL domains. Exercise resulted in improvement in: global HRQoL at 12 weeks' (SMD 0.48; 95% confidence interval (CI) 0.16 to 0.81) and 6 months' (0.46; 95% CI 0.09 to 0.84) follow-up, breast cancer concerns between 12 weeks' and 6 months' follow-up (SMD 0.99; 95% CI 0.41 to 1.57), body image/self-esteem when assessed using the Rosenberg Self-Esteem scale at 12 weeks (MD 4.50; 95% CI 3.40 to 5.60) and between 12 weeks' and 6 months' (mean difference (MD) 2.70; 95% CI 0.73 to 4.67) follow-up, emotional well-being at 12 weeks' follow-up (SMD 0.33; 95% CI 0.05 to 0.61), sexuality at 6 months' follow-up (SMD 0.40; 95% CI 0.11 to 0.68), sleep disturbance when comparing follow-up values by comparison group at 12 weeks' follow-up (SMD -0.46; 95% CI -0.72 to -0.20), and social functioning at 12 weeks' (SMD 0.45; 95% CI 0.02 to 0.87) and 6 months' (SMD 0.49; 95% CI 0.11 to 0.87) follow-up. Further, exercise interventions resulted in decreased anxiety at 12 weeks' follow-up (SMD -0.26; 95% CI -0.07 to -0.44), fatigue at 12 weeks' (SMD -0.82; 95% CI -1.50 to -0.14) and between 12 weeks' and 6 months' (SMD -0.42; 95% CI -0.02 to -0.83) follow-up, and pain at 12 weeks' follow-up (SMD -0.29; 95% CI -0.55 to -0.04) when comparing follow-up values by comparison group.Positive trends and impact of exercise intervention existed for depression and body image (when analyzing combined instruments); however, because few studies measured these outcomes the robustness of findings is uncertain.No conclusions can be drawn regarding the effects of exercise interventions on HRQoL domains of cognitive function, physical functioning, general health perspective, role function, and spirituality.Results of the review need to be interpreted cautiously owing to the risk of bias. All the trials reviewed were at high risk for performance bias. In addition, the majority of trials were at high risk for detection, attrition, and selection bias.

AUTHORS' CONCLUSIONS

This systematic review indicates that exercise may have beneficial effects on HRQoL and certain HRQoL domains including cancer-specific concerns (e.g. breast cancer), body image/self-esteem, emotional well-being, sexuality, sleep disturbance, social functioning, anxiety, fatigue, and pain at varying follow-up periods. The positive results must be interpreted cautiously due to the heterogeneity of exercise programs tested and measures used to assess HRQoL and HRQoL domains, and the risk of bias in many trials. Further research is required to investigate how to sustain positive effects of exercise over time and to determine essential attributes of exercise (mode, intensity, frequency, duration, timing) by cancer type and cancer treatment for optimal effects on HRQoL and its domains.

Exercise interventions on health-related quality of life for people with cancer during active treatment

BACKGROUND

People with cancer undergoing active treatment experience numerous disease- and treatment-related adverse outcomes and poorer health-related quality of life (HRQoL). Exercise interventions are hypothesized to alleviate these adverse outcomes. HRQoL and its domains are important measures of cancer survivorship, both during and after the end of active treatment for cancer.

OBJECTIVES

To evaluate the effectiveness of exercise on overall HRQoL outcomes and specific HRQoL domains among adults with cancer during active treatment.

SEARCH METHODS

We searched the Cochrane Central Register of Controlled Trials (CENTRAL), PubMed MEDLINE, EMBASE, CINAHL, PsycINFO, PEDRO, LILACS, SIGLE, SportDiscus, OTSeeker, Sociological Abstracts from inception to November 2011 with no language or date restrictions. We also searched citations through Web of Science and Scopus, PubMed's related article feature, and several websites. We reviewed reference lists of included trials and other reviews in the field.

SELECTION CRITERIA

We included all randomized controlled trials (RCTs) and quasi-randomized controlled clinical trials (CCTs) comparing exercise interventions with usual care or other type of non-exercise comparison intervention to maintain or enhance, or both, overall HRQoL or at least one distinct domain of HRQoL. Included trials tested exercise interventions that were initiated when adults with cancer were undergoing active cancer treatment or were scheduled to initiate treatment.

DATA COLLECTION AND ANALYSIS

Five paired review authors independently extracted information on characteristics of included trials, data on effects of the intervention, and assessed risk of bias based on predefined criteria. Where possible, we performed meta-analyses for HRQoL and HRQoL domains for the reported difference between baseline values and follow-up values using standardized mean differences (SMDs) and a random-effects model by length of follow-up. We also reported the SMD at follow-up between the exercise and control groups. Because investigators used many different HRQoL and HRQoL domain instruments and often more than one for the same domain, we selected the more commonly used instrument to include in the SMD meta-analyses. We also report the mean difference for each type of instrument separately.

MAIN RESULTS

We included 56 trials with 4826 participants randomized to an exercise (n = 2286) or comparison (n = 1985) group. Cancer diagnoses in trial participants included breast, prostate, gynecologic, hematologic, and other. Thirty-six trials were conducted among participants who were currently undergoing active treatment for their cancer, 10 trials were conducted among participants both during and post active cancer treatment, and the remaining 10 trials were conducted among participants scheduled for active cancer treatment. Mode of exercise intervention differed across trials and included walking by itself or in combination with cycling, resistance training, or strength training; resistance training; strength training; cycling; yoga; or Qigong. HRQoL and its domains were assessed using a wide range of measures.The results suggest that exercise interventions compared with control interventions have a positive impact on overall HRQoL and certain HRQoL domains. Exercise interventions resulted in improvements in: HRQoL from baseline to 12 weeks' follow-up (SMD 0.33; 95% CI 0.12 to 0.55) or when comparing difference in follow-up scores at 12 weeks (SMD 0.47; 95% CI 0.16 to 0.79); physical functioning from baseline to 12 weeks' follow-up (SMD 0.69; 95% CI 0.16 to 1.22) or 6 months (SMD 0.28; 95% CI 0.00 to 0.55); or when comparing differences in follow-up scores at 12 weeks (SMD 0.28; 95% CI 0.11 to 0.45) or 6 months (SMD 0.29; 95% CI 0.07 to 0.50); role function from baseline to 12 weeks' follow-up (SMD 0.48; 95% CI 0.07 to 0.90) or when comparing differences in follow-up scores at 12 weeks (SMD 0.17; 95% CI 0.00 to 0.34) or 6 months (SMD 0.32; 95% CI 0.03 to 0.61); and, in social functioning at 12 weeks' follow-up (SMD 0.54; 95% CI 0.03 to 1.05) or when comparing differences in follow-up scores at both 12 weeks (SMD 0.16; 95% CI 0.04 to 0.27) and 6 months (SMD 0.24; 95% CI 0.03 to 0.44). Further, exercise interventions resulted in a decrease in fatigue from baseline to 12 weeks' follow-up (SMD -0.38; 95% CI -0.57 to -0.18) or when comparing difference in follow-up scores at follow-up of 12 weeks (SMD -0.73; 95% CI -1.14 to -0.31). Since there is consistency of findings on both types of measures (change scores and difference in follow-up scores) there is greater confidence in the robustness of these findings.When examining exercise effects by subgroups, exercise interventions had significantly greater reduction in anxiety for survivors with breast cancer than those with other types of cancer. Further, there was greater reduction in depression, fatigue, and sleep disturbances, and improvement in HRQoL, emotional wellbeing (EWB), physical functioning, and role function for cancer survivors diagnosed with cancers other than breast cancer but not for breast cancer. There were also greater improvements in HRQoL and physical functioning, and reduction in anxiety, fatigue, and sleep disturbances when prescribed a moderate or vigorous versus a mild exercise program.Results of the review need to be interpreted cautiously owing to the risk of bias. All the trials reviewed were at high risk for performance bias. In addition, the majority of trials were at high risk for detection, attrition, and selection bias.

AUTHORS' CONCLUSIONS

This systematic review indicates that exercise may have beneficial effects at varying follow-up periods on HRQoL and certain HRQoL domains including physical functioning, role function, social functioning, and fatigue. Positive effects of exercise interventions are more pronounced with moderate- or vigorous-intensity versus mild-intensity exercise programs. The positive results must be interpreted cautiously because of the heterogeneity of exercise programs tested and measures used to assess HRQoL and HRQoL domains, and the risk of bias in many trials. Further research is required to investigate how to sustain positive effects of exercise over time and to determine essential attributes of exercise (mode, intensity, frequency, duration, timing) by cancer type and cancer treatment for optimal effects on HRQoL and its domains.

Health-related quality of life and cancer clinical trials

The measurement of patient-reported outcomes, including health-related quality of life, is a new initiative which has emerged and grown over the past four decades. Following the development of reliable and valid self-report questionnaires, health-related quality of life has been assessed in tens of thousands of patients and a wide variety of cancers. This review is based on a selection of data published in the last decade and is intended primarily for healthcare professionals. The assessments in clinical trials have been particularly useful for elucidating the effects of various cancers and their treatments on patients' lives and have provided additional information that enhances the usual clinical endpoints used for determining the benefits and toxicity of treatment. With growing experience the quality of the health-related quality of studies has improved and, in general, recent studies are more likely to be methodologically robust than those that were performed in earlier decades. Health-related quality of life has become a more accurate predictor of survival than some other clinical parameters, such as performance status. The overall outlook for the routine assessment of patient-reported outcomes in clinical trials is assured and, eventually, it is likely to become a standard part of clinical practice. However, there is still a need for a clear method for determining the clinical meaningfulness of changes in scores. The answer will probably come from the greater use of patient-reported outcomes and the consequent growth of experience that is necessary to make such judgements.