Compliance with reporting guidelines by Australian pathologists: an audit of the quality of histopathology reporting in high-grade glioma

Abstract

Background

Diagnostic pathology reports inform management plans for patients with glioma, and there is an increasing clinical need for molecular testing. We assessed the quality of histopathology reports of grade III/IV gliomas.

Methods

Reports were obtained as part of a tumor biobank. From 720 pathology reports, 594 eligible reports were assessed for 28 elements derived from published checklists. A summary quality score incorporated 9 critical parameters for clinical decision making: diagnosis using World Health Organization 2007 criteria; cell type; grade; narrative supporting cell type and grade; absence of equivocal language; conclusion reporting cell type and grade; and conclusion aligned with report narrative.

Results

Of 594 eligible reports, the final conclusion was not supported by the report narrative in 122 (21%). Tumor classification and grade were not supported by the narrative in 105 (18%) and 36 (6%) reports, respectively. Only 145 (24%) reports fulfilled all 9 quality criteria, while 25% contained 6 or fewer key quality indices. Report quality was higher when pathologists had neuropathology subspecialization, when a grade IV tumor was reported, and when the specimen was from an initial resection or grade-progressed tumor rather than recurrent high-grade glioma. Use of molecular testing increased over time, from 29% to 48% over four quartiles of the study. Molecular testing was more frequently done where oligodendroglial elements were reported.

Conclusion

A significant proportion of reports failed to meet key indicators of report quality. Pathology reporting is critical in communicating between pathologists and treating clinicians. Clinicians should be aware of reporting quality and seek clarification when required.

Patients with high-grade glioma, encompassing grade III and IV glial tumors, will receive different management recommendations depending on the final pathology reported. Recent clinical trials have demonstrated improved outcomes with adjuvant chemotherapy, in particular procarbazine, lomustine, and vincristine (PCV) chemotherapy for 1p19q codeleted anaplastic oligodendroglial tumors,^1,2 and chemoradiotherapy and adjuvant temozolomide for grade IV astrocytomas.³ Patients with non-codeleted anaplastic astrocytomas may receive radiotherapy alone or participate in a clinical trial examining the role of adjuvant chemotherapy. Those with an isocitrate dehydrogenase 1 (IDH1) mutation and oligodendroglial elements may particularly benefit from PCV chemotherapy.⁴ The formal pathology report is the critical communication document between the histopathologist and the treating clinician. This exemplifies the need for accurate and interpretable histopathological reporting of primary brain tumors, including molecular testing, in order to guide clinical management.

The quality of anatomic pathology reports is determined by the accuracy, timeliness, and completeness of reporting,⁵ although assessment of these factors can be difficult. For many tumor types, protocols or synoptic reporting structures outline the essential parameters for inclusion in pathology reports. Cancers of the central nervous system (CNS) present a challenge to formalized methods of pathology reporting. CNS tumors are diverse and are not included in the American Joint Committee on Cancer/International Union Against Cancer (AJCC/UIC) TNM classification, from which synoptic pathology reports for other tumor types have been derived.⁶

From 2004, the Commission on Cancer of the American College of Surgeons stipulated the use of checklist components of the College of American Pathologists (CAP) pathology reporting protocols as part of the Cancer Program Standards for Approved Cancer Programs. The protocols were developed by the CAP to assist reporting of clinically essential information from surgical specimens.⁷ The Royal College of Pathologists of Australasia (RCPA) developed a similar protocol to assist in the reporting of tumors of the CNS.⁸ Reporting recommendations include the description of macroscopic features of the specimen including size, number of specimens and appearance; the proportion of the specimen processed; histologic type and grade of tumor as per the World Health Organization (WHO) 2007 classification⁹; and results of immunohistochemical and ancillary studies. Additionally this protocol recommends a diagnostic summary at the conclusion of the report including specimen type; tumor site and laterality; tumor type and WHO grade.⁸

In this study, we assessed the quality of pathology reports for grade III and IV gliomas in an Australian setting using a checklist derived from the RCPA-structured and CAP-structured reporting protocols. Importantly, this study aimed to assess the quality of the report as a communication tool with the treating clinician, not the accuracy of the diagnosis, which would require an alternative study design. Through assessing the adequacy of pathology reporting of high-grade glioma, we aim to provide a basis for further quality improvement, facilitating optimal patient management decisions through pathology reports that contain adequate and unambiguous information for clinical decision making.

Patients and Methods

The Australian Genomics and Clinical Outcomes of Glioma (AGOG) project recruited patients with glial tumors from 6 tertiary centers in Western Australia (1) and New South Wales (5) to a biobank with matched clinical data. AGOG commenced in 2008 and recruited until 2013. Original diagnostic histopathology reports were obtained from each biopsy or resection for every participant. All participants provided written informed consent, and all relevant institutional Human Research Ethics Committees approved the study.

All available original histopathology reports from patients recorded as having a final diagnosis of grade III or IV glioma were accessed from the AGOG database in April 2012. Most reports postdated 2008, although reports from operations performed prior to 2008 were not excluded. Incomplete, duplicate and reports of frozen section or electron microscopy findings only were ineligible for this analysis, as were reports of low-grade glioma that preceded a subsequent diagnosis of high-grade glioma. A total of 720 reports were recalled, of which 594 were eligible.

Assessment of Reports

No prior formal quality assessment criteria have been proposed for neuro-oncology pathology reporting; hence a structured checklist was developed informed by reporting guidelines and recommendations. The structured checklist identified the key elements of quality neuropathology reports based on the RCPA Central Nervous System Tumors Structured Reporting Protocol and the CAP recommendations in conjunction with the expert opinion of a specialist neuropathologist (VF) on our panel and a neuro-oncologist (AN). The checklist had good face validity and incorporated the descriptions and findings necessary for the pathological diagnosis of grade III or IV gliomas and for appropriate clinical interpretation of the report (Table 1). Each report was assessed for the presence or absence of 28 elements derived from the RACP and CAP checklists (Table 1), a technique which has previously been shown to be a valid method of determining quality reporting.¹⁰

Table 1.

Twenty-eight elements of quality reporting assessed for each report

	Elements Assessed
Macroscopic findings	Report date
	Size
	Site/laterality
	Number of specimens
	Proportion processed
Processing	Frozen section
	Smear
	Electron microscopy
Ancillary studies	Immunohistochemistry described
	Molecular studies performed
Microscopic findings	Cell type specified
	Cellularity described
	Cell abnormalities described
	Additional features described
	Mitotic figures presence/absence described
	Necrosis presence/absence described
	Vascularity described
Interpretation	WHO diagnosis given
	Graded
	Narrative in support of cell type classification
	Narrative in support of tumor grade
Reporting style	Absence of equivocal language
	Opinion not given prior to description
Summary	Specimen type
	Specimen site/laterality
	Summary/conclusion reporting tumor type
	Summary/conclusion reporting tumor grade
	Conclusion supported by report narrative

Elements that are bold and italicized represent those considered key components in interpretation of the report, which were included in the 9-element summary quality score.

One investigator reviewed all reports. For reports in which interpretation was difficult or inconclusive, consensus review by a specialist neuropathologist and medical oncologist was applied. A random 10% of reports were blindly reviewed by a second investigator to assess interobserver variability in scoring. Both assessors were medical oncologists, representing the clinician group most likely to alter a treatment recommendation based on pathology reporting. The field of “narrative in support of classification” was determined by the reporting of cell type or the presence of descriptive terms in keeping with a tumor type; for example, a description of fibrillary cells in an astrocytic tumor. Similarly, the field of “narrative in support of grade” was assessed on the basis of reporting the presence of mitoses, necrosis, and vascularity in keeping with the reported grade of the tumor. Narratives in support of classification and grade were determined with reference to the WHO 2007 Classification of Tumors of the Central Nervous System.⁹ Equivocal language was identified as language in the report found to be vague, meaningless, contradictory, or nonspecific.

A summary quality score was derived from 9 critical parameters (bold and italicized in Table 1). These 9 criteria had face validity as the most important for the clinical interpretation of pathology reports for high-grade gliomas. A value of 1 was assigned for the presence of each criterion, with a maximum possible score of 9/9 representing the presence of all quality criteria.

Additional descriptive information was recorded, which was hypothesized to be a potential determinant of the quality score, including the subspecialization of the reporting pathologist, the amount of material available (biopsy or resection), the tumor grade, the year of reporting, and whether the specimen was from a first or subsequent resection (Table 2). The reporting era was considered both as a continuum or divided into 4 quartiles for the purpose of analysis.

Table 2.

Additional descriptive information recorded from each report

Report Parameters	Subgroups
Specimen type	Biopsy
	Resection
Report style	Synoptic
	Nonsynoptic
Pathologist	Specialist neuropathologist
	Surgical neuropathologist
	General anatomical pathologist
	Pathologist specialty unknown
	No listed pathologist
Grade	III
	IV
Resection type	Initial resection
	Subsequent resection
	Resection of tumor progressed from low grade
Date of report

Specialist neuropathologists were defined as those pathologists who had completed specialist fellowship examinations in neuropathology; surgical neuropathologists were defined as those who had completed fellowship examinations in surgical neuropathology with a special interest in neuropathology. Two investigators performed categorization of pathologists with advice from a further specialist neuropathologist and with reference to the Australian and New Zealand Society of Neuropathology membership. We hypothesized that pathologist specialization would be associated with improved quality scores, as would having a larger amount of material available, and more recent reporting.

Statistics

All data extracted from histopathology reports were recorded using Microsoft Excel and analyzed using IBM SPSS for Windows version 20. The two-sample Kolmogorov-Smirnov test for independent variables was used for nonparametrically distributed data, the Mann-Whitney-U test was used to compare mean quality scores between groups, with Pearson's chi-square test used for a comparison of proportions between groups. The intraclass correlation coefficient was used to test agreement between the two assessors. All tests were two-sided.

Results

A total of 720 reports were extracted from the AGOG database, of which 594 met the eligibility criteria. Three hundred sixty-eight (62%) reports were reported by neuropathologists; of which 339 were reported by surgical neuropathologists. One hundred eighty-nine (32%) reports were by general pathologists. The pathologist name was not listed in one report and for 36 (6%) reports the specialty of the pathologist was unknown. All reports were reviewed by assessor 1, with a random selection of 10% of reports also scored blindly by assessor 2. An average measures intraclass correlation coefficient of 0.907 indicated very high agreement (95% CI, 0.80-0.95, P < .001) between assessors.

Figure 1 shows the proportion of reports that met each of the 28 quality parameters and the 9 key quality criteria. In the reports examined, the macroscopic parameters were well reported on although the date of report completion was not recorded in 15%. While the processing of a frozen section was recorded in almost three-quarters of reports, in 68% and 83% of reports there were no records of examination by smear or electron microscopy, respectively. Immunohistochemical studies were commented on in the majority of reports; however, molecular studies were only described as being done in 37% of reports. The proportion of reports recording molecular studies increased over the 4 quartiles of reporting dates (P = .004), with 48% of specimens undergoing molecular testing in reports dating from the final quartile (March 2011 – April 2012) as compared with 29% from the first quartile (Prior to January 2009). Reports in which oligodendroglial features were described were more likely to undergo molecular studies (76% with oligodendroglial features vs 29% without oligodendroglial features, P < .001), and molecular studies were more likely to be ordered by specialist pathologists (P = .03). No report of MGMT methylation status testing was identified.

Fig. 1.

Proportion of eligible reports fulfilling each quality criterion. Dark shaded parameters comprise the key quality criteria and percent of reports fulfilling those criteria are enumerated.

Of the 9 quality score parameters, the cell type was not specifically noted within the body of 60% of the reports and the narrative description of the tumor cells did not support the conclusion tumor cell type classification in 18% of cases. Specialist neuropathologists were more likely to report the cell type (P < .001) and to provide a narrative that supported the cell type (P < .001). The grade was not recorded in 11% of the reports, and the grade was not supported by the narrative of the report in 6% of cases; there was a nonsignificant trend for specialist neuropathologists to report grade more often (P = .08) and they were significantly more likely to report a narrative supporting the grade (P = .01). The diagnosis according to WHO 2007 classification terminology was not given in 24% of reports; this did not change with the date of the report (P = .24) but was more likely to be included in a report by a specialist neuropathologist (P < .001). In report summaries, the tumor type was not given in 4% of reports and the grade was not specified in 15%.

The final report conclusion was not supported by the report narrative in a total of 122 (21%) of the 594 reports. On review of the quality score, only 145 (24%) met all 9 criteria. One hundred ninety-four (33%) reports met 8 criteria, while 104 (17%) met 7 criteria and the remaining 25% of reports met 6 or fewer of the quality score parameters (Fig. 2A).

Fig. 2.

(A) Proportion of eligible reports scoring each category for the summary quality score (out of a maximum possible of 9). (B) Proportion of eligible reports meeting summary quality score criteria for neuropathologists and general anatomical pathologists (out of a maximum possible of 9).

Using only the reports for which pathologist specialty was known, the mean summary quality score for reports by neuropathologists was 7.67 (standard deviation [SD] = 1.33), significantly higher than that for reports by general pathologists (mean = 6.62, SD = 2.12, P < .0001, Fig. 2B). Sensitivity analyses were performed in which the reports by pathologists of unknown specialty were included for analysis with reports by neuropathologists and then with reports by general pathologists. In each analysis, the difference in report quality between those classified as neuropathologists and general pathologists remained statistically significant. Statistically significant differences between reports by neuropathologists compared with general pathologists were found for each of the individual quality score criteria with the exceptions of the reporting of grade and the absence of equivocal language (Fig. 3).

Fig. 3.

Proportion of eligible reports fulfilling key quality criteria comprising the summary quality score for neuropathologists and general anatomical pathologists. *denotes P < .05.

One hundred reports (17%) featured the use of equivocal language, with equivalent use by both specialist and nonspecialist pathologists (P = .35). Representative examples of equivocal language used in the reports are shown in Table 3.

Table 3.

Examples scored as using equivocal language

‘Mitoses are inconspicuous’

‘… an occasional mitotic figure identified a focal suggestion of necrosis and evidence of patchy endothelial proliferation …’

‘There is a suggestion of micovascular proliferation in some areas’

‘… and very focal endothelial proliferation. Only one possible mitosis is seen’

‘Very occasional mitotic figures are identified, and within small vessels associated with the glioma, there is possibly a minimal degree of endothelial proliferation’

‘… and occasional areas are surrounded by tumour cells showing a suggestion of palisading. Microvascular proliferation is inconspicuous’

‘… microscopic foci of possible incipient necrosis …’

‘There are focal areas of palisading around small foci of apparent incipient necrosis …’

‘Focally there is some suggestion of necrosis’

‘… plus a hint of delicate chickenwire-like vascular pattern in the background’

Grade IV tumors accounted for 447 (75%), while 147 (25%) reports were for grade III tumors, categorized according to the final conclusion of the pathology report. A statistically significant difference in report quality was found for grade III compared with grade IV tumor reports. The mean summary quality score for reports for grade III tumors was 7.77 (SD = 1.44) compared with 7.06 (SD = 1.80) for grade IV tumor reports, P < .0001. There was also a borderline statistically significantly higher mean summary quality score for reports in which oligodendroglial elements were documented (mean with oligodendroglial elements = 7.88 (SD = 1.27) vs mean without oligodendroglial elements = 7.14 (SD 1.76), P = .025).

Reports were defined as those from initial resections, recurrent resections, or resections of transformed low-grade tumors. Initial resections accounted for the majority, with 430 (72%) reports; resections of recurrent high-grade tumors accounted for 115 (19%) reports; and the remaining 49 (8%) reports were from transformed low-grade tumors. On comparison of the quality of reports for each of these resection types, statistically significant differences were found with the summary quality score for initial resections (mean = 7.5, SD = 1.54) or for resections of transformed low-grade tumors (mean = 7.53, SD = 1.50) being significantly higher than for resections of recurrent known high-grade disease without reported change in grade (mean = 6.1, SD = 2.09, P < .001).

While very few reports used a synoptic format (28 [5%]), the summary quality score was actually slightly lower for those using synoptic reporting (mean = 7.2, SD = 1.8 vs mean = 6.9, SD = 1.4; P = .041). Nevertheless, use of synoptic reporting may have removed the poorest performing outliers, as no synoptic report scored <5 out of 9. Thirty-five percent of specimens were biopsies and 65% were resections; however, the specimen size did not affect the reporting quality score (P = .67). Reports originating in Western Australia scored significantly higher for quality (mean = 7.7 vs 6.6, P < .001), and more cases from Western Australia were reported by a pathologist with neuropathology expertise (71% vs 59%, P = .004). However the difference between region of reporting remained significant even when neuropathology expertise was included as a covariate. No correlation was found between the quality of the report and how recently it was reported.

Discussion

We have reviewed the histopathology reports of grade III and IV gliomas collated as part of a large tissue collection to gain an appreciation of the quality of neuropathology reports in Australia. Our findings regarding the interpretive aspects of reporting and report conclusions are particularly important as it is the clear communication of report conclusions to clinicians that reduces the risk of misunderstanding.¹¹ A significant proportion of reports lack critical elements that are important in the clinical interpretation of neuropathology reports and therefore most likely to impact upon clinical decisions and prognostication. Some reports omitted key information necessary for decision making, such as use of WHO 2007 diagnostic classification of the tumor or a distinction between grade III or IV disease.

The gold standard for diagnostic accuracy in glial tumors is the findings from histopathology. Unlike other tumor types in which clinical or surgical staging contributes significantly to prognostication or management decisions, in glial tumors the histopathological determination of the grade and type of tumor provides the bulk of prognostic information.^2,4,12,13 In particular, both 1p19q and IDH1 mutation status are of significant prognostic and predictive value, and will determine the use of chemotherapy in addition to radiotherapy in grade III glioma.^1,2,4 In our series, although molecular testing was more common in tumors with oligodendroglial elements (79%), it was not universal. While the reports we examined predated publication of data demonstrating a treatment advantage for patients with co-deleted and IDH1 mutated tumors, clearly molecular testing now needs to approach 100% for optimal patient management.

MGMT promoter methylation is a favorable independent prognostic factor in glioblastoma and is predictive of greater benefit associated with treatment with chemotherapy and radiotherapy.^14–16 We did not identify any reports with comments on MGMT methylation status testing, reflecting lack of reimbursement and availability of the test in Australia outside of the research setting, as well as minimal influence of this result on clinical decision making during the period studied. IDH1 or 2 mutations are both independent prognostic factors, with tumors harboring these mutations associated with better outcomes compared with those without the mutations, regardless of tumor grade.¹⁶ Other factors that can be reported by the pathologist and may have prognostic value include the presence of necrosis in glioblastoma,¹⁷ and an oligodendroglial component.^17–20

An accurate pathologic diagnosis of high-grade glioma also determines clinical trial eligibility. Trial participation is associated with a survival advantage for patients with glioblastoma, regardless of the study arm to which they are enrolled.²¹ Currently the randomized CATNON trial is enrolling patients with non-codeleted anaplastic astrocytoma at many of the study sites to examine the role of adjuvant temozolomide chemotherapy. The pathology report therefore provides key information influencing decisions regarding prognosis, treatment, and clinical trial eligibility.

A key component of a histopathology report is its ability to accurately convey the intended information to its audience. The language used in the practice of anatomical pathologists is distant from that used in clinical practice.¹¹ A study comparing clinician comprehension with pathologist intent demonstrated an overall discordance rate of 30% for surgeons and surgical trainees, although discrepancies were reduced to some extent with seniority.¹¹ While some misinterpretation may relate to lack of understanding of laboratory processes, other discrepancies relate to the language with which pathology findings are reported; hence, language use can affect patient care.

We found that reports by neuropathologists were more complete than those by general pathologists, although we did not aim to assess the veracity of the actual diagnosis. Indeed, disagreement in neuropathology reporting has been demonstrated previously. A review of 457 cases found a concordance rate of 77%.²² There was more discordance between original and review diagnoses for cases initially reported at nonacademic (26%) vs academic (12%) hospitals. Of clinically significant discrepancies, 94% originated from nonacademic hospitals. Discordance between specialist and nonspecialist pathologists has also been described in gynecologic cancers, where in 8% of cases specialist pathologist review led to a change in diagnosis and subsequently to patient management.²³ Aside from concerns regarding the accuracy of the diagnosis, we identified less comprehensive reporting of key findings by nonspecialist pathologists, who may underestimate the clinical relevance of some parameters that are required by treating clinicians in order to guide treatment decisions.

The use of synoptic reporting with checklists or templates may enable more comprehensive reporting and has the potential to overcome some gaps in report quality. Synoptic reports are presented in a tabular rather than descriptive form with clearly defined titles and responses.⁶ Templates include essential features required for a diagnosis, and also those features that provide prognostic information or guide management decisions. Synoptic reporting presents information in a consistent and predictable manner that is easy to read, time-saving in the use of abbreviated phrasing, and also leads to more complete reports and efficient extraction of information.²⁴ In lung, breast, and colon cancers, the implementation of a pathology checklist reporting project by Cancer Care Ontario (CCO) improved the completeness of pathology reports. This was consistent with previous studies that also demonstrated improved physician satisfaction, reduced time to read reports, less pathologist time to produce reports, and fewer phone calls for clarification.²⁵ A study investigating the completeness of synoptic versus nonsynoptic reporting for primary cutaneous melanoma found that all features were more frequently reported in synoptic reports, and in particular synoptic reports more frequently reported features critical to staging and management.²⁴

In our sample, few reports used a synoptic format, with no evidence of improved overall quality scores. However, no synoptic reports had a quality score below 5, and equivocal language was avoided. While there is clear evidence for the benefits associated with synoptic reporting in other tumors, the variety and variability of CNS tumors does present a challenge for the development of synoptic report templates. CNS tumors are not included in the AJCC/UIC TNM staging system, which has formed the basis of synoptic report development for other tumor groups. Checklists to encompass all CNS tumors, such as that from the Neuropathology Task Force for the CAP Cancer Committee, are extensive and lengthy to report and read.⁶

It has been hypothesized that pathologic features that do not require a measurement are more likely to be reported if present, but more often omitted if absent.²⁴ This is particularly pertinent in reporting of gliomas, for which the abundance of mitotic figures, necrosis, or degree of vascular proliferation conveys diagnostic and prognostic information. One suggestion has been that synoptic reporting templates for CNS tumors use headers with answer fields rather than drop-down menus. While menus are used for synoptic reporting in other tumor types, they may not provide the flexibility necessary for CNS tumors.⁶ The Central Nervous System Tumors Structured Reporting Protocol 2011 is an initial collaborative approach to the development of a report template that with further review and refinement will provide a synoptic reporting template to improve the completeness and usability of neuropathology reporting.⁸ More recently, consensus guidelines for CNS tumor classification and grading have been produced, which, if followed, are likely to further address some of the concerns highlighted by our analysis.²⁶

Our finding of variation between Australian states in neuropathology reporting is likely to reflect the centralization of elective neurosurgical services and pathology expertise at two major sites in Western Australia, and relative decentralization in New South Wales. Western Australian specimens were almost exclusively reported at sites with access to neuropathology expertise, even when reported by a general pathologist, and predominantly at a single site. In addition, the major pathology provider in Western Australia, while not providing a true synoptic report template for CNS tumors, included some general structural aspects of the report that were conducive to comprehensive reporting.

A limitation of this study is that we were unable to determine whether the quality of the reporting affected clinical decisions, given the complex nature of decision making in high-grade glioma and the need to consider performance status, comorbidities, and patient wishes in determining management plans. Furthermore, the lack of some scoring criteria may not have always affected the message conveyed to the clinician. The retrospective nature of the study meant that we were unable to assess whether discussion at multidisciplinary team meetings took place, or whether reports lacking quality parameters elicited more requests for clarification. In addition, the use of equivocal language or lack of some quality parameters may on occasion reflect suboptimal tissue quality, a small biopsy sample, or an ambiguous microscopic phenotype and not in all settings a poor quality report. Assessment of glioma specimens following previous treatment, including radiotherapy and chemotherapy, may also impact substantially on histologic appearance and sample quality. Hence our finding of lower quality scores for subsequent resections of grade IV glioma may represent the inherent difficulties of examining these specimens, rather than necessarily representing less diligent reporting in the context of a known previous diagnosis. Nevertheless, we were able to review the information that clinicians will have available when making management recommendations, hence approximating the clinical setting.

In conclusion, in this audit of almost 600 neuropathology reports, we found that a substantial proportion of reports failed to meet defined quality criteria. The diagnostic conclusion of the report was not supported by the narrative description in the report in over one-fifth of cases. Omissions and inconsistencies in the pathology report have the potential to alter patient management. This study supports the importance of centralized and specialist pathology reporting for high-grade glioma, a need for additional training for non-neuropathologist reporting brain neoplasms, and the case for the development and use of synoptic pathology reporting in this setting. Pathologists should consider including pertinent negatives in reports. Clinicians managing patients with high-grade glioma should seek further clarification around neuropathology reports that provide insufficient or conflicting information, which may alter clinical decision making.

Funding

Cancer Council New South Wales Strategic Research Partnership (STREP) grant SRP 08-10.