original articles
Annals of Oncology
Annals of Oncology 26: 231–237, 2015 doi:10.1093/annonc/mdu489 Published online 29 October 2014
Poor patient-reported outcomes reporting according to CONSORT guidelines in randomized clinical trials evaluating systemic cancer therapy O. Bylicki1,2, H. K. Gan3, F. Joly4,5,6, D. Maillet7, B. You7,8,9 & J. Péron7,10,11*
Received 23 June 2014; revised 21 September 2014; accepted 30 September 2014
Background: The Consolidated Standards of Reporting Trials (CONSORT) guidance was extended in 2013 to provide a set of specific recommendations regarding patient-reported outcomes (PROs) reporting in randomized clinical trials (RCTs). There is limited data regarding how well current publications of oncology RCTs report PROs if assessed using these guidelines. Design: All phase III medical oncology RCTs published between 2007 and 2011 were reviewed according to the 2013 PROs CONSORT recommendations and an 11-point PROs reporting quality score (PRORQS) was defined based on the criteria. Results: The majority of trials did not report on PROs at all (201 of 325; 62%). Of the remaining 124 trials, the mean PRORQS score was 5.0 on an 11-point scale. The items related to methods of PROs collection and analysis were poorly reported (Description of the prespecified PRO hypothesis: 26% of RCTs; methods for PRO data collection ( paper, telephone, electronic, other): 16%; statistical approaches for managing missing data: 37%). The only factor significantly associated with improved PROs reporting was where PROs reporting was the subject of a dedicated secondary manuscript, as was the case in 36 of the 124 (29%) of RCTs. Conclusion: Despite their clinical relevance, our findings show that some aspects of PROs reporting may greatly be improved, especially critical methodological aspects of PROs collection and analysis. The exceptions were where PROs were described in PROs-specific secondary publication. Use of the 2013 PROs CONSORT extensions should be encouraged and their effects on PROs reporting subsequently reassessed. Key words: randomized clinical trials, quality of life, patients-reported outcomes, reporting quality
introduction Randomized clinical trials (RCTs) are considered to be the gold standard in assessing medical interventions. Patient-reported outcomes (PROs) are outcomes reported by the patients themselves, without the interpretation of the patient’s responses by a physician or anyone else [1]. PROs measures may vary in complexity, from a single-item question about unique concept, up to multi-item instruments for measuring quality of life, and multiple domains of functional status.
*Correspondence to: Dr Julien Péron, Department of medical oncology, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, F-69310. 165, chemin du grand revoyet, 69495 Pierre-Bénite, France. Tel: +33-4-78-86-43-18; Fax: +33-4-78-86-43-56; E-mail: [email protected]
In the field of medical oncology, the primary end points are frequently some measure of patient survival. PROs are complementary to evaluate both benefits and harm of treatments tested in RCTs. They can arguably be considered as important as patient survival, especially since oncology drugs have lower therapeutic indices compared with drugs in other therapeutic areas [2]. PROs data are increasingly used in modern RCTs [3, 4]. They provide the most direct evidence of whether the prescribed treatments actually improve patients’ general well-being, tumor-related symptoms as well as treatment side-effects [5]. As there are still many methodological challenges on data collection, appropriate timing of assessment, adequate statistical hypothesis and analysis as well as outcomes reporting and interpretation [6–8], data issued from PROs are not totally accepted. Not surprisingly then, evidence suggests that reporting
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Nanyang Technological University on May 25, 2015
1 Department of Pneumology, Desgenettes Hospital; 2Department of Medical Oncology, Centre Léon BERARD, University of Lyon, Lyon, France; 3Joint Austin-Ludwig Oncology Unit, Austin Hospital, Melbourne, Victoria, Australia; 4INSERM, U1086, Caen; 5Clinical Research Unit, François Baclesse Center, Caen; 6Department of Medicine, CHU de Caen, Caen; 7Department of Medical Oncology, Lyon-Sud Hospital Center, Hospices Civils de Lyon, Pierre-Bénite; 8EMR UCBL/HCL 3738, Faculté de Médecine Lyon-Sud, Oullins; 9Department of Medical Oncology, University of Lyon, Lyon; 10Biostatistics Unit, Hospices Civils de Lyon, Lyon; 11Biometry and Evolutionary Biology Laboratory, Health and Biostatistics Team, CNRS UMR 5558, Villeurbanne, France
original articles
Annals of Oncology
methods trial selection We searched Medline via PubMed (http://www.pubmed.gov) to identify all primary report of RCTs assessing systemic therapies for solid tumors and including at least one PROs published between January 2007 and December 2011 in 10 English language journals where oncology RCTs are frequently published: Annals of Oncology; British Journal of Cancer; Breast Cancer Research and Treatment; Cancer; European Journal of Cancer; Journal of Clinical Oncology; Journal of the National Cancer Institute; Lancet; Lancet Oncology; and New England Journal of Medicine. The search was carried out using the terms ‘randomized’ and ‘cancer’ as keywords and ‘English’ plus ‘clinical trials’ or ‘randomized controlled trial’ as limits. Exclusion criteria were: lack of PROs reporting; pediatric studies; hematologic trials, phase I, II,
or IV trials; meta-analyses, overviews, publications using pooled data from two or more trials; and secondary reports on previously published trials [14, 15]. The presence of at least one PROs ( patient-reported HRQL data; patient-reported symptom data; patient-reported satisfaction data) was researched by reviewing of the full manuscripts. We also searched Medline via PubMed for secondary reports of PROs, published secondary or as companion paper to the primary report, using the title and the authors’ names of all the article citing the primary report in their references (Figure 1).
definition of trial characteristics Trials were considered as industry funded if a RCT received any form of industry funding with the exception of those studies where only drug(s) was provided but no funding. Trials were considered intercontinental when patients from more than one continent were included. The journal impact factor (IF) was the mean journals IF between 2007 and 2011. Only the journal IF of the manuscript reporting the main RCT results ( primary manuscript) was used for analyses. The type of investigational therapy was the main drug assessed in trials; it could be either tested in combination with other drugs or as a single-agent therapy.
development of a quantitative scoring system for quality of patient-related outcomes reporting A standardized data extraction form was used by two authors (JP and DM) to capture the data in this review. The two authors examined each article. Where there was a discrepancy in responses to a given item, the authors resolved this by consensus evaluation. The extraction form included a number of guidelines to ensure homogenous data extraction for those recommendations potentially open to interpretation. PROs relevant domains were considered adequately identified (i.e. item P2b) if one or several domains of primary interest were identified
Key-words through pubmed: Cancer; Controlled Clinical trials; Phase III; English; All Adult: 19+years, Publication Date from 2007 to 2011
Primary reports identified (n = 739)
Reports met inclusion and exclusion criteria (n = 124)
Did not met inclusion criteria (n = 414) - pediatric studies - treatment solely with radiotherapy or surgery - phase I, II or phase IV trials - supportive care, palliative care or prevention trials - haematology trials - mets-analyses, overviews or publications - using data from ≥2 trials - secondary reports or interim analyses No PRO data was associated with the report (n = 201)
PRO data reported in a separtate article (n = 36)
PRO data reported in the main article (n = 88)
Figure 1. Selection of randomized clinical trials in the systematic review.
| Bylicki et al.
Volume 26 | No. 1 | January 2015
Downloaded from http://annonc.oxfordjournals.org/ at Nanyang Technological University on May 25, 2015
of PROs remains sub-optimal across both oncology and nononcology RCTs [9–11]. To improve the reporting of PROs in oncology RCTs, a recent extension to the CONSORT statement regarding PROs reporting was published [12]. It included five ‘extension’ statements to the 2010 CONSORT guidance [13] that each addresses a key reporting item for quality reporting from all RCTs using PROs. In addition, components of the existing 2010 CONSORT guidance that were critically relevant to PROs reporting were expanded by six ‘elaboration’ statements. This review was carried out to evaluate the quality of PROs reporting in recent oncology RCT reports, according to the 2013 PROs-specific CONSORT extension, thereby establishing the current adequacy of PROs reporting. In addition, we investigated manuscripts’ characteristics associated with better quality in PROs reporting.
original articles
Annals of Oncology
as those referencing a multidomain concept representing the patient’s general perception of the effect of illness and treatment on physical, psychological, and social aspects of life. Symptom data were defined as any subjective evidence of a disease, health condition, or treatment-related effect that can be noticed and known only by the patient. The main statistical method used to compare PROs between the randomized groups was also collected.
statistical analysis Categorical trials characteristics among RCTs including at least one PROs and RCTs without PRO were compared using χ 2 tests or fisher’s exact tests as appropriate. The PRORQS was the sum of the number of items that were adequately reported (Table 1) and expressed as an integer between 0 and 11. PRORQS scores were summarized using descriptive statistics such as mean, confidence intervals (CIs) and range. Single-item frequencies were compared between categories using χ 2 tests. Univariate and multivariable linear regression analyses were used to identify factors associated with higher PRORQS. The following trial characteristics were investigated: year of primary report publication, tumor site, presence of an industrial funding, primary report journal IF, geographic
Table 1. Quality of PROs reporting, rating using items from the 2013 extensions of the CONSORT statement (N = 124) Descriptor of the 2010 CONSORT criteria
Descriptor of the 2013 PRO-specific extension or elaboration PRO-specific extensions are prefaced by the letter P PRO-specific elaborations are prefaced by the letter E
1b—Structured summary of trial design, methods, results, and conclusions 2a—Scientific background and explanation of rationale 2b—Specific objectives or hypotheses
P1b Identification of the PROs in the abstract as a primary or secondary outcome E2a Background and rationale for PROs assessment P2b Identification of the PROs relevant domains Statement of the PROs hypothesis Statement of the PROs analysis power P6a Evidence of PROs instrument validity Reference of the PROs instrument Statement of the person completing the PROs Methods of data collection (paper, telephone, electronic, other) P12a Statistical approaches for dealing with missing data are explicitly stated E13a Description of the number of PROs outcome data at baseline and at subsequent time points At baseline At subsequent time points E15 Including baseline PROs data when collected
6a—Completely defined prespecified primary and secondary outcome measures
12a—Statistical methods used to compare groups 13a—For each group, the numbers of participants who were randomly assigned, received intended treatment and were analyzed for the primary outcome 15—A table showing baseline demographic and clinical characteristics for each group 16—For each group, number of participants (denominator) included in each analysis 17a—For each primary and secondary outcome, results for each group, the estimated effect size, and its precision 20/21—Trial limitations, addressing sources of potential bias and generalizability of the trial findings 22—Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence
Number of trials in which item was adequately reported, n (%) 47 (28) 53 (43) 80 (65) 32 (26) 12 (10) 45 (36) 104 (84) 47 (38) 20 (16) 46 (37) 64 (52) 76 (61) 87 (70) 49 (40)
E16 Required for PROs results
60 (48)
E17a For multidimensional PROs results from each domain P20/21 PROs-specific limitations and implications for generalizability and clinical practice E22 PROs data should be interpreted in relation to clinical outcomes including survival data
54 (43) 43 (35) 75 (60)
The 11 items of the 2013 PROs-specific CONSORT extension are in bold. PROs-specific extensions are prefaced by the letter P; PROs-specific elaborations are prefaced by the letter E. PROs, patient-reported outcomes.
Volume 26 | No. 1 | January 2015
doi:10.1093/annonc/mdu489 |
Downloaded from http://annonc.oxfordjournals.org/ at Nanyang Technological University on May 25, 2015
by the authors of the reports. The description of the number of PROs outcome data at subsequent time points (i.e. item E13a) was considered adequate if authors provided at least one assessment of the number of PROs data collected in addition to the baseline evaluation. Item E17a was considered correctly reported when the results were reported for all the symptoms or dimensions included in the PROs tool used. A PROs reporting quality score (PRORQS) based on the 2013 CONSORT extension was defined by two of the authors (OB and JP). The score was based on 11 items derived from the 11 recommendations (either extension or elaboration of the 2010 CONSORT statement) (Table 1). Each item was scored 1 if it was adequately reported or 0 if it was not clearly reported or not reported at all; each item was weighted with equal importance. A 1-point difference in mean score between two groups was considered meaningful, as it would suggest the failure of one group to report on one more item compared with the other group. In addition, data were also captured regarding the methods of PROs data collection and PROs results reporting. When PROs results were included in the main article, the space (expressed as the percentage of lines devoted to PROs relative to total section size) devoted to PRO in the Methods and Results sections were collected. The type of PROs used was collected (quality-of-life data and/or symptom data). Quality-of-life data were defined
original articles
results characteristics of selected randomized, controlled trials From the 739 trials initially screened, a total of 325 RCTs reporting phase III trials in the field of medical oncology were identified. Data from this dataset have previously been reported regarding the quality of adverse events reporting [14]. Among them, 124 (38%) included at least one PROs and were included in this analysis (Figure 1). The number of published RCTs with PROs per year was stable (P = 0.29) (Table 2). The presence of at least one PROs was more frequent among RCTs including patients with metastatic disease than among RCTs including patients in adjuvant or neoadjuvant setting (44% versus 27%, P = 0.00015). PROs were also more frequent among RCTs reported in high IF journals (61% for IF >20, 36% for IF between 10 and 20, 27% for IF
Annals of Oncology
Annals of Oncology 26: 231–237, 2015 doi:10.1093/annonc/mdu489 Published online 29 October 2014
Poor patient-reported outcomes reporting according to CONSORT guidelines in randomized clinical trials evaluating systemic cancer therapy O. Bylicki1,2, H. K. Gan3, F. Joly4,5,6, D. Maillet7, B. You7,8,9 & J. Péron7,10,11*
Received 23 June 2014; revised 21 September 2014; accepted 30 September 2014
Background: The Consolidated Standards of Reporting Trials (CONSORT) guidance was extended in 2013 to provide a set of specific recommendations regarding patient-reported outcomes (PROs) reporting in randomized clinical trials (RCTs). There is limited data regarding how well current publications of oncology RCTs report PROs if assessed using these guidelines. Design: All phase III medical oncology RCTs published between 2007 and 2011 were reviewed according to the 2013 PROs CONSORT recommendations and an 11-point PROs reporting quality score (PRORQS) was defined based on the criteria. Results: The majority of trials did not report on PROs at all (201 of 325; 62%). Of the remaining 124 trials, the mean PRORQS score was 5.0 on an 11-point scale. The items related to methods of PROs collection and analysis were poorly reported (Description of the prespecified PRO hypothesis: 26% of RCTs; methods for PRO data collection ( paper, telephone, electronic, other): 16%; statistical approaches for managing missing data: 37%). The only factor significantly associated with improved PROs reporting was where PROs reporting was the subject of a dedicated secondary manuscript, as was the case in 36 of the 124 (29%) of RCTs. Conclusion: Despite their clinical relevance, our findings show that some aspects of PROs reporting may greatly be improved, especially critical methodological aspects of PROs collection and analysis. The exceptions were where PROs were described in PROs-specific secondary publication. Use of the 2013 PROs CONSORT extensions should be encouraged and their effects on PROs reporting subsequently reassessed. Key words: randomized clinical trials, quality of life, patients-reported outcomes, reporting quality
introduction Randomized clinical trials (RCTs) are considered to be the gold standard in assessing medical interventions. Patient-reported outcomes (PROs) are outcomes reported by the patients themselves, without the interpretation of the patient’s responses by a physician or anyone else [1]. PROs measures may vary in complexity, from a single-item question about unique concept, up to multi-item instruments for measuring quality of life, and multiple domains of functional status.
*Correspondence to: Dr Julien Péron, Department of medical oncology, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, F-69310. 165, chemin du grand revoyet, 69495 Pierre-Bénite, France. Tel: +33-4-78-86-43-18; Fax: +33-4-78-86-43-56; E-mail: [email protected]
In the field of medical oncology, the primary end points are frequently some measure of patient survival. PROs are complementary to evaluate both benefits and harm of treatments tested in RCTs. They can arguably be considered as important as patient survival, especially since oncology drugs have lower therapeutic indices compared with drugs in other therapeutic areas [2]. PROs data are increasingly used in modern RCTs [3, 4]. They provide the most direct evidence of whether the prescribed treatments actually improve patients’ general well-being, tumor-related symptoms as well as treatment side-effects [5]. As there are still many methodological challenges on data collection, appropriate timing of assessment, adequate statistical hypothesis and analysis as well as outcomes reporting and interpretation [6–8], data issued from PROs are not totally accepted. Not surprisingly then, evidence suggests that reporting
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Nanyang Technological University on May 25, 2015
1 Department of Pneumology, Desgenettes Hospital; 2Department of Medical Oncology, Centre Léon BERARD, University of Lyon, Lyon, France; 3Joint Austin-Ludwig Oncology Unit, Austin Hospital, Melbourne, Victoria, Australia; 4INSERM, U1086, Caen; 5Clinical Research Unit, François Baclesse Center, Caen; 6Department of Medicine, CHU de Caen, Caen; 7Department of Medical Oncology, Lyon-Sud Hospital Center, Hospices Civils de Lyon, Pierre-Bénite; 8EMR UCBL/HCL 3738, Faculté de Médecine Lyon-Sud, Oullins; 9Department of Medical Oncology, University of Lyon, Lyon; 10Biostatistics Unit, Hospices Civils de Lyon, Lyon; 11Biometry and Evolutionary Biology Laboratory, Health and Biostatistics Team, CNRS UMR 5558, Villeurbanne, France
original articles
Annals of Oncology
methods trial selection We searched Medline via PubMed (http://www.pubmed.gov) to identify all primary report of RCTs assessing systemic therapies for solid tumors and including at least one PROs published between January 2007 and December 2011 in 10 English language journals where oncology RCTs are frequently published: Annals of Oncology; British Journal of Cancer; Breast Cancer Research and Treatment; Cancer; European Journal of Cancer; Journal of Clinical Oncology; Journal of the National Cancer Institute; Lancet; Lancet Oncology; and New England Journal of Medicine. The search was carried out using the terms ‘randomized’ and ‘cancer’ as keywords and ‘English’ plus ‘clinical trials’ or ‘randomized controlled trial’ as limits. Exclusion criteria were: lack of PROs reporting; pediatric studies; hematologic trials, phase I, II,
or IV trials; meta-analyses, overviews, publications using pooled data from two or more trials; and secondary reports on previously published trials [14, 15]. The presence of at least one PROs ( patient-reported HRQL data; patient-reported symptom data; patient-reported satisfaction data) was researched by reviewing of the full manuscripts. We also searched Medline via PubMed for secondary reports of PROs, published secondary or as companion paper to the primary report, using the title and the authors’ names of all the article citing the primary report in their references (Figure 1).
definition of trial characteristics Trials were considered as industry funded if a RCT received any form of industry funding with the exception of those studies where only drug(s) was provided but no funding. Trials were considered intercontinental when patients from more than one continent were included. The journal impact factor (IF) was the mean journals IF between 2007 and 2011. Only the journal IF of the manuscript reporting the main RCT results ( primary manuscript) was used for analyses. The type of investigational therapy was the main drug assessed in trials; it could be either tested in combination with other drugs or as a single-agent therapy.
development of a quantitative scoring system for quality of patient-related outcomes reporting A standardized data extraction form was used by two authors (JP and DM) to capture the data in this review. The two authors examined each article. Where there was a discrepancy in responses to a given item, the authors resolved this by consensus evaluation. The extraction form included a number of guidelines to ensure homogenous data extraction for those recommendations potentially open to interpretation. PROs relevant domains were considered adequately identified (i.e. item P2b) if one or several domains of primary interest were identified
Key-words through pubmed: Cancer; Controlled Clinical trials; Phase III; English; All Adult: 19+years, Publication Date from 2007 to 2011
Primary reports identified (n = 739)
Reports met inclusion and exclusion criteria (n = 124)
Did not met inclusion criteria (n = 414) - pediatric studies - treatment solely with radiotherapy or surgery - phase I, II or phase IV trials - supportive care, palliative care or prevention trials - haematology trials - mets-analyses, overviews or publications - using data from ≥2 trials - secondary reports or interim analyses No PRO data was associated with the report (n = 201)
PRO data reported in a separtate article (n = 36)
PRO data reported in the main article (n = 88)
Figure 1. Selection of randomized clinical trials in the systematic review.
| Bylicki et al.
Volume 26 | No. 1 | January 2015
Downloaded from http://annonc.oxfordjournals.org/ at Nanyang Technological University on May 25, 2015
of PROs remains sub-optimal across both oncology and nononcology RCTs [9–11]. To improve the reporting of PROs in oncology RCTs, a recent extension to the CONSORT statement regarding PROs reporting was published [12]. It included five ‘extension’ statements to the 2010 CONSORT guidance [13] that each addresses a key reporting item for quality reporting from all RCTs using PROs. In addition, components of the existing 2010 CONSORT guidance that were critically relevant to PROs reporting were expanded by six ‘elaboration’ statements. This review was carried out to evaluate the quality of PROs reporting in recent oncology RCT reports, according to the 2013 PROs-specific CONSORT extension, thereby establishing the current adequacy of PROs reporting. In addition, we investigated manuscripts’ characteristics associated with better quality in PROs reporting.
original articles
Annals of Oncology
as those referencing a multidomain concept representing the patient’s general perception of the effect of illness and treatment on physical, psychological, and social aspects of life. Symptom data were defined as any subjective evidence of a disease, health condition, or treatment-related effect that can be noticed and known only by the patient. The main statistical method used to compare PROs between the randomized groups was also collected.
statistical analysis Categorical trials characteristics among RCTs including at least one PROs and RCTs without PRO were compared using χ 2 tests or fisher’s exact tests as appropriate. The PRORQS was the sum of the number of items that were adequately reported (Table 1) and expressed as an integer between 0 and 11. PRORQS scores were summarized using descriptive statistics such as mean, confidence intervals (CIs) and range. Single-item frequencies were compared between categories using χ 2 tests. Univariate and multivariable linear regression analyses were used to identify factors associated with higher PRORQS. The following trial characteristics were investigated: year of primary report publication, tumor site, presence of an industrial funding, primary report journal IF, geographic
Table 1. Quality of PROs reporting, rating using items from the 2013 extensions of the CONSORT statement (N = 124) Descriptor of the 2010 CONSORT criteria
Descriptor of the 2013 PRO-specific extension or elaboration PRO-specific extensions are prefaced by the letter P PRO-specific elaborations are prefaced by the letter E
1b—Structured summary of trial design, methods, results, and conclusions 2a—Scientific background and explanation of rationale 2b—Specific objectives or hypotheses
P1b Identification of the PROs in the abstract as a primary or secondary outcome E2a Background and rationale for PROs assessment P2b Identification of the PROs relevant domains Statement of the PROs hypothesis Statement of the PROs analysis power P6a Evidence of PROs instrument validity Reference of the PROs instrument Statement of the person completing the PROs Methods of data collection (paper, telephone, electronic, other) P12a Statistical approaches for dealing with missing data are explicitly stated E13a Description of the number of PROs outcome data at baseline and at subsequent time points At baseline At subsequent time points E15 Including baseline PROs data when collected
6a—Completely defined prespecified primary and secondary outcome measures
12a—Statistical methods used to compare groups 13a—For each group, the numbers of participants who were randomly assigned, received intended treatment and were analyzed for the primary outcome 15—A table showing baseline demographic and clinical characteristics for each group 16—For each group, number of participants (denominator) included in each analysis 17a—For each primary and secondary outcome, results for each group, the estimated effect size, and its precision 20/21—Trial limitations, addressing sources of potential bias and generalizability of the trial findings 22—Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence
Number of trials in which item was adequately reported, n (%) 47 (28) 53 (43) 80 (65) 32 (26) 12 (10) 45 (36) 104 (84) 47 (38) 20 (16) 46 (37) 64 (52) 76 (61) 87 (70) 49 (40)
E16 Required for PROs results
60 (48)
E17a For multidimensional PROs results from each domain P20/21 PROs-specific limitations and implications for generalizability and clinical practice E22 PROs data should be interpreted in relation to clinical outcomes including survival data
54 (43) 43 (35) 75 (60)
The 11 items of the 2013 PROs-specific CONSORT extension are in bold. PROs-specific extensions are prefaced by the letter P; PROs-specific elaborations are prefaced by the letter E. PROs, patient-reported outcomes.
Volume 26 | No. 1 | January 2015
doi:10.1093/annonc/mdu489 |
Downloaded from http://annonc.oxfordjournals.org/ at Nanyang Technological University on May 25, 2015
by the authors of the reports. The description of the number of PROs outcome data at subsequent time points (i.e. item E13a) was considered adequate if authors provided at least one assessment of the number of PROs data collected in addition to the baseline evaluation. Item E17a was considered correctly reported when the results were reported for all the symptoms or dimensions included in the PROs tool used. A PROs reporting quality score (PRORQS) based on the 2013 CONSORT extension was defined by two of the authors (OB and JP). The score was based on 11 items derived from the 11 recommendations (either extension or elaboration of the 2010 CONSORT statement) (Table 1). Each item was scored 1 if it was adequately reported or 0 if it was not clearly reported or not reported at all; each item was weighted with equal importance. A 1-point difference in mean score between two groups was considered meaningful, as it would suggest the failure of one group to report on one more item compared with the other group. In addition, data were also captured regarding the methods of PROs data collection and PROs results reporting. When PROs results were included in the main article, the space (expressed as the percentage of lines devoted to PROs relative to total section size) devoted to PRO in the Methods and Results sections were collected. The type of PROs used was collected (quality-of-life data and/or symptom data). Quality-of-life data were defined
original articles
results characteristics of selected randomized, controlled trials From the 739 trials initially screened, a total of 325 RCTs reporting phase III trials in the field of medical oncology were identified. Data from this dataset have previously been reported regarding the quality of adverse events reporting [14]. Among them, 124 (38%) included at least one PROs and were included in this analysis (Figure 1). The number of published RCTs with PROs per year was stable (P = 0.29) (Table 2). The presence of at least one PROs was more frequent among RCTs including patients with metastatic disease than among RCTs including patients in adjuvant or neoadjuvant setting (44% versus 27%, P = 0.00015). PROs were also more frequent among RCTs reported in high IF journals (61% for IF >20, 36% for IF between 10 and 20, 27% for IF
