Table 2 Descriptives of included studies
| ID | References | Author, (date of publication) | Study population: no. of participants, age (mean (S.D.) and sex (n (%)) | Tumour type, country of study | PROMs used | Study aims | Statistical analysis, covariates | Summary of findings |
|---|---|---|---|---|---|---|---|---|
| 1 | [22] | de Rooij et al. (2018) | N = 2457; Age: 68.1 (10.1); M: 1457 (59); F: 1000 (41) |
Colorectal, Urological, Gynaecological The Netherlands | EORTC QLQ-C30, Hospital Anxiety and Depression Scale (HADS) |
Identification of subgroups of cancer survivors with realistic, pessimistic or optimistic illness perceptions (IPs) relative to prognosis at time of questionnaire Assessment of HRQoL and survival associated with these subgroups |
Multivariate Cox proportional Hazards regression models Clinical/demographic characteristics |
Functioning and global QoL were higher and symptom burden lower in those with optimistic illness perceptions compared with realistic IPs (all P < 0.01) Functioning was lower and symptoms were higher in those with pessimistic IPs compared to those with realistic IP (all P < 0.01) All-cause mortality higher in survivors with pessimistic IPs compared with those with realistic IPs (HR 1.52, 95% CI (1.27–1.84) |
| 2 | [23] | Graham et al. (2018) | N = 68; Age: 64.15; M: 53 (77.9); F: 15 (22.1) |
Renal United States | Edmonton Symptom Assessment System (ESAS) | Assessment of whether ESAS measured at baseline can provide prognostic information for patients receiving standard first-line sunitinib for metastatic renal cell carcinoma (mRCC) | Multivariate Cox proportional Hazards regression models, Known prognostic indicators of mRCC |
In multivariate analysis, higher baseline symptom burden was associated with inferior overall survival (HR 1.21, 95% CI (1.01–1.44) for each 10-unit increase in the ESAS total score when controlling for the mskcc risk group and 1.240 (P = 0.019) for each 10-unit increase when controlling for the imdc risk group Baseline symptom burden, as measured by ESAS, provides modest degree of prognostic information independently of other widely used prognostic factors |
| 3 | [24] | Baekelandt et al. (2016) | N = 66; Age: 68 (34–83); M: 20 (30.3); F: 44 (69.7) |
Pancreatic Norway | EORTC QLQ-C30, EORTC QLQ-PAN26, Edmonton Symptom Assessment System (ESAS) | Assess the prognostic significance of pretreatment HRQoL and symptom scores on survival in patients with resectable pancreatic ductal adenocarcinoma |
Multivariate Cox proportional Hazards regression models Clinical/demographic characteristics |
Six reported variables with P < 0.20 eligible for multivariate model. After stepwise backward selection only cognitive function remained in model Based on results of prognostic impact of PROs on survival, patients divided into 2 groups: those with high versus those with low cognitive function Hazard ratio for death (HR 3.5, 95% CI (1.7–7.3) higher in patients with low cognitive function compared to higher cognitive function |
| 4 | [25] | Moningi et al. (2015) | N = 110; Age: NI; M: NI; F:NI |
Pancreatic United States | EORTC QLQ-PAN26 | Examine associations between QoL as they relate to self-reported symptoms, clinical characteristics and performance status | Pearson's Chi-squared Test | Patients with lower performance status measured by both ECOG and KPS had worse QoL scores: pain, digestive symptoms, cachexia, ascites (P < 0.05). More aggressive symptom management may result in improved PFS and better outcomes for pancreatic patients |
| 5 | [14] | Bingener et al. (2015) | N = 431; Age: 69.0 (11.2); M: 219(50.8); F: 212 (49.2) |
Colorectal United States | Symptom Distress Scale, 5-item Quality of Life Index |
Investigate whether deficits in preoperative QoL scores are associated with surgical outcomes such as 30-day morbidity Determine whether change in QoL postoperatively associated with morbidity |
Stepwise logistic and linear models Demographics |
Changes from baseline to day 2 QoL indicators, including concentration (OR1.27 (1.00–1.61) P = 0.049), appearance (OR 1.38 (1.02–1.87) P = 0.037), breathing (OR 1.50 (1.02–2.21), P = 0.038) significantly associated with early complications Changes from baseline to day 14 in ‘activities’, ‘daily living’ and ‘total QLI’ were also associated with early complications. Using stepwise logistic model, the variables significantly associated with having any early complications (yes/no) were age, ASA III and change in ‘activity’ from baseline to day 14 Significant predictors for being readmitted to the hospital within 2 months were baseline pain distress severity and changes from baseline to day 2 in fatigue. Also associated with readmission were changes from baseline to postoperative day 14 in ‘daily living’ and outlook |
| 6 | [21] | Quinten et al. (2013) | N = 2603; Age: NI; M: NI, F: NI |
Colorectal, Urological, Gynaecological, Pancreatic Belgium | EORTC QLQ-C30 | Investigate the relative contribution of different HRQoL domains as prognostic for separate cancer types |
Multivariate Cox proportional Hazards regression models Established clinical/socioeconomic prognostic indicators |
Results demonstrated that, for each cancer site, at least 1 HRQOL domain provided prognostic information that was additive over and above clinical and sociodemographic variables. However, the HRQOL parameters of greatest prognostic value differed across the cancer groups; and the effect size of each HRQOL parameter, indicated by the HR, depended on the tumor site Physical functioning linked to survival in colorectal cancer (HR 0.93, 95% CI (0.96–0.99)) Nausea and vomiting significant assocation in colorectal and ovarian cancer; (HR 1.06, 95% CI (1.01–1.07) and (HR 1.16, 95% CI (1.07–1.25) respectively |
| 7 | [26] | Robinson et al. (2011) | N = 723; Age: NI; M: 0 (0), F: 723 (100) |
Gynaecological Denmark | EORTC QLQ-C30 | Hypothesised that diagnostic delay could be a indicator of poorer health care system performance, which may affect survival and patient satisfaction. Investigated associations between QoL, patient satisfaction and survival | Poisson regression modelling and multivariate Cox proportional Hazards models, Demographic/clinical factors |
Association between QoL and survival differed between ovarian and endometrial cancer. In ovarian cancer significant association between increased fatigue and reduced survival (HR 1.84, 95% CI (1.09–3.10) In endometrial cancer significant association with survival remained after adjustment for physical functioning (HR 3.30, 95% CI (1.27–8.57), role functioning (HR 5.40, 95% CI (1.57–18.58), emotional functioning (HR 3.41, 95% CI (1.33–8.74), nausea (HR 2.7, 95% CI (1.13–6.45), appetite loss (HR 3.78, 95% CI (1.71–8.34) |
| 8 | [32] | Lehto et al. (2018) | N = 104; Age: 66.5 (51–82); M: 104 (100), F: 0 (0) |
Prostate Finland | Ways of Coping Questionnaire, Anger Expression Scale, Life Experience Scale, Rotterdam Symptom Checklist, EORTC QLQ-C30, LENT-SOMA outcome measure | Investigated the baseline and early predictors of disease-free and overall survival times in prostate cancer patients of all ages and treated with external beam radiotherapy | Multivariate Cox proportional Hazard models, Demographic/clinical factors | Different QOL measures exhibited either a favorable or an unfavorable impact, i.e., an increased level of pain (HR 0.05; 95% CI 0.01–0.32) predicted longer survival, whereas prostate-area symptoms (HR 1.18; 95% CI 1.03–1.36), increased fatigue (HR 7.08; 95% CI 1.77–28.32), and reports of no or few physical symptoms (HR 9.90; 95% CI 1.48–66.30) were significant predictors of shorter survival time. However, when the overall quality-of-life index (total scale 1–7) was tested instead of the prostate-area symptom scale, it predicted a longer survival (HR 0.51, 95% CI 0.27–0.95, P = 0.033); when both scales were included, the effect of the overall QOL was weaker (HR 0.56; 95% CI 0.27–1.15, P = 0.113) |
| 9 | [27] | Gupta et al. (2015) | N = 917; Age: 63 (40.8–89.3); M: 917 (100), F: 0 (0) |
Prostate United States | PS Questionnaire | Investigate whether self-rated health is a potential confounder of the relationship between patient satisfaction with service quality and survival in patients with prostate cancer undergoing treatment | Multivariate Cox proportional Hazard models, Demographic/clinical factors | Self-rated health found to be independent predictor of survival in multivariate analysis after controlling for patient satisfaction (HR 0.30, 95% CI (0.11–0.86). Finding coupled with observation that performance status and self-rated health correlated suggest that self-rated health is potential confounder in relationship between patient satisfaction and survival in prostate cancer |
| 10 | [28] | Jayadevappa et al. (2009) | N = 318; Age: 57.25 (4.75); M: 318 (100), F: 0 (0) |
Prostate United States | SF-36, UCLA Prostate Cancer Index | Analysis of association between race/ethnicity, risk of biochemical recurrence and recovery pattern of patient reported outcomes and cost in younger men with newly diagnosed prostate cancer | Linear mixed effect models, Demographic/clinical factors |
Ethnicity is not a predictor of generic and prostate-specific HRQoL after adjustment for demographic and clinical variables Low risk of biochemical recurrence was associated with better physical function, vitality, mental health general health, urinary function, bowel function, bowel bother and sexual bother |
| 11 | [34] | Lis et al. (2008) | N = 230; Age: 61.8 (40–87); M: 230 (100), F: 0 (0) |
Prostate United States | The Ferrans and Powers Quality of Life Index | Aimed to determine whether patient satisfaction with HRQoL might predict length of survival in patients with prostate cancer undergoing routine oncology treatment | Multivariate Cox proportional Hazard models, Clinical factors |
Patient satisfaction with health and physical subscale was significantly associated with survival (RR:0.82, 95% CI (0.67–0.98) Association independent of previous treatment history and Gleason score |
| 12 | [33] | Lis et al. (2006) | N = 55; Age: 56.2 (33–74); M: 31 (56.4), F: 24 (43.6) |
Pancreatic United States | The Ferrans and Powers Quality of Life Index (QLI) | Determine whether patient satisfaction with QoL predicts length of survival in patients with pancreatic cancer undergoing care in a nonclinical trial setting | Multivariate Cox proportional Hazard models, Disease Stage | Four variables (health and physical subscale, family subscale, global QoL and stage at diagnosis) found to be significant upon univariate analysis. No QoL subscale found to be significant after adjustment for stage at disease although health and physical subscale marginally significant (RR: 0.94, 95% CI (0.89–1.00) (P = 0.053) |
| 13 | [15] | Bernhard et al. (2010) | N = 295, Age: 63 (26–83); M: 138(46.8), F: 157 (53.2) |
Pancreatic Switzerland | Global linear-analogue self-assessment (LASA) | Investigated the prognostic value of QoL relative to CA-19, and the role of CA-19 in estimating palliation in patients with advanced pancreatic cancer receiving chemotherapy within RCT | Multivariate Cox proportional Hazard models and linear mixed effect models |
At baseline, less pain and tiredness (i.e. less symptom burden) predicted better survival (HR 0.65, 95% CI (0.45–0.96)) and (HR 0.63, 95% CI (0.44–0.92)) respectively baseline CA 19–9 did not predict QOL or time on study treatment, besides a marginal effect on pain. Neither CA 19–9 nor QOL predicted tumour response to chemotherapy. Survival is influenced by different factors than response to chemotherapy, although response impacts on survival. Thus, CA 19–9 and QOL at baseline provide limited information for estimating palliation by chemotherapy |
| 14 | [16] | Gourgou-Bourgade et al. (2013) | N = 342; Age: 61 (25–76); M: 211(62.0), F: 131 (38.0) |
Pancreatic France | EORTC QLQ-C30 | To compare the quality of life (QoL) of patients receiving oxaliplatin, irinotecan, fluorouracil, and leucovorin (FOLFIRINOX) or gemcitabine as first-line chemotherapy and to assess whether pretreatment QoL predicts survival in patients with metastatic pancreatic cancer | Multivariate Cox proportional Hazard models, QoL domains/clinical factors | Performance status, constipation (HR 1.06, 95% CI (1.01–1.11)) and dyspnea (HR 1.06, 95% CI (1.00–1.14)) remained significant after backward and forward selection procedures |
| 15 | [29] | Braun et al. (2013) | N = 186, Age: 55.1 (24–85); M: 121(65.1), F: 65 (34.9) |
Pancreatic United States | EORTC QLQ-C30 | Investigate whether pretreatment QoL parameters as well as changes in QoL scores from baseline until 3 months after treatment could predict survival in patients with stage IV pancreatic cancer | Multivariate Cox proportional Hazard models, Demographic/clinical factors |
Found that every 10-point increase in baseline global QoL score was associated with a 12% decreased risk of death (HR, 0.88; 95% CI, 0.81Y0.95; P = 0.001) Improvement in cognitive function at 3 months is an indicator of improved patients’ survival after adjusting for other covariates (HR 0.89, 95% CI (0.77–0.98)) |
| 16 | [19] | Roychowdhury et al. (2003) | N = 364; Age: 63.5; M: 288(79.1), F: 76 (20.9) |
Bladder United States | EORTC QLQ-C30 | Analysis performed to determine the prognostic significance of HRQoL parameters on time-to-event end points in patients with locally advanced or metastatic bladder cancer who participated in a phase III randomized study comparing gemcitabine and cisplatin (GC) with methotrexate, vinblastine, doxorubicin, and cisplatin (MVAC) | Multivariate Cox proportional Hazard models, Demographic/clinical factors |
Identified HRQoL parameters (physical functioning, anorexia, and fatigue) to be significant and independent prognostic factors for time-to-event end points Physical functioning: (HR:1.56, 95% CI (1.15–2.10)), Anorexia: (HR 1.84, 95% CI (1.36–2.49)), Fatigue: (HR 1.46, 95% CI (1.11–1.94) |
| 17 | [30] | Maisey et al. (2002) | N = 501; Age: NI; M: NI, F: NI |
Colorectal United Kingdom | EORTC QLQ-C30 | Examine the prognostic value of baseline QoL measurements in patients with locally advanced and metastatic colorectal cancer treated with systemic chemotherapy within the context of clinical trials |
Multivariate Cox proportional Hazard models Clinical/demographic factors |
Majority of the QoL domains remained significant independent predictors of overall survival in the final multivariate model: Physical functioning: (HR 1.35, 95% CI (1.07–1.69)) Role functioning: (HR 1.34, 95% CI (1.08–1.66)) Social fuctioning: (HR 1.43, 95% CI (1.13–1.81)) Emotional functioning: (HR 1.28, 95% CI (1.04–1.57)) |
| 18 | [20] | Collette et al. (2004) | N = 391; Age: 70.7 (34.3–89.3); M: 391 (100), F: 0 (0) |
Prostate Europe | EORTC QLQ-C30 | Utilise HRQOL data from three RCTs to assess clinical and biochemical parameters to identify independent prognostic factors for overall survival | Stratified multivariate Cox proportional Hazard models |
Symptom items of appetite loss and insomnia from the EORTC QLQ-C30 were retained as independent prognostic factors of overall survival: Appetite loss: (HR:1.47, 95% CI (1.16–1.86)) Insomnia: (HR:1.45, 95% CI (1.15–1.84) |
| 19 | [17] | Stucky et al. (2011) | N = 449; Age: 68.8 (11.25); M: 223 (50), F: 226 (50) |
Colorectal United States | Symptom Distress Scale, 5-item Quality of Life Index | Evaluate the effect of baseline QOL on subsequent QOL and survival | Intention to treat analysis and stepwise logistic/linear regressions | Baseline outlook [hazard ratio (HR) = 0.58, 95% confidence interval (CI) 0.38–0.88, P = 0.01] and support (HR 2.85, 95% CI 1.52–5.35, P = 0.001) were significantly associated with overall survival |
| 20 | [18] | Chau et al. (2004) | N = 1080; Age: 62 (28–84); M: 842 (78), F: 238 (22) |
Gastric United Kingdom | EORTC QLQ-C30 | Identify baseline patient- or tumor-related prognostic factors and assess whether pretreatment QoL predicts survival in patients with locally advanced or metastatic oesophagus-gastric cancer | Multivariate Cox proportional Hazard models, Clinical/demographic factors | When pretreatment QoL data were tested against the baseline prognostic models, physical functioning (HR 0.76, 95% CI (0.60–0.97)) (P = .003), role functioning (HR 0.69, 95% CI (0.54–0.88) (P < .001), and global QoL (HR 0.57, 95% CI (0.45–0.72)) (P < .001) had significant prognostic impact |
| 21 | [31] | Coates et al. (1997), 47 | N = 47; Age: 57.7; M: NI, F: NI |
Gastric, Colorectal Australia, Germany, Canada | EORTC QLQ-C30 | Evaluate the prognostic association of QL scores among patients with advanced malignancies in routine practice |
Multivariate Cox proportional Hazard models, Clinical/demographic factors |
Single-item QL scores for overall physical condition, overall quality of life, and the global and social functioning scales remained independently prognostic: Global QoL: (HR 0.99, 95% CI (0.98–0.99)) Social: (HR 0.93, 95% CI (0.98–0.99)) |