The Australian-Canadian Adolescent and Young Adult Collaborative Cohort Initiative: Initial Meeting Report

Abstract

Adolescents and young adults (AYAs) with cancer are a vulnerable group whose outcomes have not improved to the extent of those of younger children and older adults.¹ To date, research attempting to identify mechanisms underlying these disparities and aiming at improving AYA outcomes has been limited by selection bias, lack of critical treatment and biologic data, and reliance on extrapolation from pediatric or adult cohorts.^2–7 To address these limitations, two individually established initiatives are currently underway to build detailed population-based AYA cancer databases through a combination of data linkage and chart abstraction: one in Ontario, Canada (the Initiative to Maximize Progress in Adolescent and Young Adult Cancer Therapy—IMPACT) and the other in Australia (AYA Patterns of Cancer Care Cohort—POCC).

Here, we report on a 2-day planning meeting in Melbourne, Australia, which brought together key individuals from both initiatives to: (1) identify research questions that could improve AYA outcomes and that can be answered by collaboration between the two teams; and (2) identify barriers to such collaboration and ways of overcoming these barriers. We describe the consultation with various stakeholders to inform content and discussion for the collaborative meeting, outcomes of the meeting, and the future collaborative efforts agreed upon. Our intent by doing so is to raise awareness of the research opportunities represented by this collaboration, and to encourage other international efforts in joint AYA research.

Pre-Meeting Stakeholder Survey

Before the meeting, an online survey was distributed to Canadian, Australian, and international individuals identified by meeting organizers to be key AYA cancer stakeholders. AYA cancer patients, clinicians, researchers, policymakers, and funders were targeted. Participants were asked to rate the importance of examples of research questions potentially answerable through collaboration between the two research teams, suggest additional questions, and define what a successful meeting would entail. Participants could suggest additional stakeholders to whom the survey should be sent.

One hundred fourteen individuals were contacted, and 50 (44%) returned a completed survey. The majority of respondents were Australian (N = 22, 44%), with an additional 15 from Canada (30%), 10 from the United States (20%), and 3 from Europe (6%). Respondents' involvement in AYA cancer can be seen in Table 1. Survey results are outlined throughout the rest of this article.

Table 1.

Results of Pre-Meeting Survey Characterizing Respondents' Type of Involvement in Adolescent and Young Adult Cancer (Self-Report) (N = 50) ^*

Type of involvement	N (%)
Healthcare professional, pediatric healthcare setting	16 (32)
Healthcare professional, adult healthcare setting	27 (54)
Researcher	19 (38)
Advocacy group member	7 (14)
Current or former AYA patient	5 (10)
Health policymaker	1 (2)
Funding organization	1 (2)

Respondents could report involvement in more than one category.

AYA, adolescent and young adult.

Meeting Format and Participants

The 2-day meeting was held on December 7–8, 2015, in Melbourne, Australia at the main campus of Cancer Council Victoria. The 10 participants represented both study teams and a variety of expertise (clinical, epidemiology, research program management, data privacy). On Day 1, the healthcare system and AYA cancer policy environment were compared between the two jurisdictions, followed by presentations on the methodology of each study. Possible study questions that could be addressed through collaboration were then identified, discussed, and prioritized. On Day 2, the highest priority research questions were discussed further, followed by deliberation on possible barriers to collaboration, ways of overcoming these barriers, and practical next steps.

Comparison of Healthcare Systems and AYA Policy Environments

Readers are referred other articles for general descriptions of the Canadian and Australian healthcare systems.^8,9 Key differences that may be relevant to AYAs with cancer include the provision of healthcare through a mix of public and private providers in Australia, funded through the national public health insurance scheme and supported by private health insurance arrangements. Services in public clinics or hospitals may be covered completely (i.e., bulk billed) or may require a co-payment from patients. In Canada, medically necessary physician services are delivered through mandatory universal public health insurance programs. In both countries, pediatric hospitals have strict upper age limits (generally between 16 and 18 years), with older AYA patients excluded from these hospitals.

Advocacy networks focused on AYAs with cancer exist in both Australia and Canada, but with varying degrees of governmental involvement. In 2007, the Australian federal government awarded the organization CanTeen 15 million AUD to administer the Youth Cancer Network Program. This has led to specialist Youth Cancer Services being developed in five jurisdictions (New South Wales/Australian Capital Territory, Victoria/Tasmania, South Australian/Northern Territory, Western Australia, and Queensland) as well as to the development of a postgraduate certificate in AYA health and well-being.¹⁰ Funding was also provided for two clinical trials, a national AYA cancer dataset, the development of three guidance documents, and a psychosocial assessment tool.¹¹ In 2013, the federal government awarded an additional 18.2 million AUD to continue the Youth Cancer Network Program to 2017. In Canada, the Canadian Partnership Against Cancer established a Task Force on AYAs with cancer in 2008, which, in turn, has established regional committees that together cover the entire country.¹² Despite these differences, AYA cancer survival outcomes are similar in both countries.^13,14

Comparison of Study Methodologies

Both the IMPACT and POCC studies share a common methodology and aim at building population-based AYA databases (IMPACT—Ontario, Canada, POCC—all Australian states) that contain detailed demographic, disease, treatment, and outcome data. Both studies accomplish this by identifying AYAs with cancer, predominantly through established government-funded cancer registries followed by on-site extraction of data from medical charts by trained chart abstractors. Details of the IMPACT study methodology have been previously published.¹⁵ Although the IMPACT and POCC studies differ in the malignancies targeted, study time periods, and age definition of AYAs, substantial overlap exists in all these domains. A detailed comparison is provided in Table 2.

Table 2.

Description of Canadian and Australian Adolescent and Young Adult Cancer Cohorts, Including Population and Study Methodology Details

	Ontario IMPACT study	Australian POCC study
Study population
Diagnostic period	1992–2012	2007–2012
Age at diagnosis	15–21 years	15–24 years
Malignancies	Acute leukemia (ALL, AML), Hodgkin and non-Hodgkin lymphoma, soft tissue and bone sarcomas, and testicular cancer	Acute leukemia (ALL, AML), soft tissue and bone sarcomas, CNS tumors
Identification	Ontario Cancer Registry	State cancer registries (Victoria, Queensland, Tasmania, Western Australia) and hospital health information departments (South Australia, New South Wales)
Study sample size	N = 5349	N = 1041
Jurisdiction population	13.6 million (2014)	23.1 million (2013)
Data sources	1. Ontario Cancer Registry: Identification of cases	1. State cancer registries (identification and minimal additional information)
	2. POGONIS: Detailed data collected by active data managers in each pediatric treatment center	2. Hospital medical records and electronic databases: Detailed data collected by data collectors in each hospital that saw >5 eligible patients over the time period
	3. Chart abstraction: For patients treated in adult/community treatment centers	3. Hospital admission records: Dates, diagnosis and procedure codes
	4. Linkages to population-based health services data, including hospitalization discharge abstracts and physician billing records	4. Clinician surveys: For patients treated at smaller hospitals or where very little data were obtained from records
		5. Linkage to National Death Index

CNS, central nervous system; IMPACT, Initiative to Maximize Progress in Adolescent and Young Adult Cancer Therapy; POCC, patterns of cancer care; POGONIS, Pediatric Oncology Group of Ontario Networked Information System.

A remarkable degree of similarity exists between the data elements captured by the IMPACT and POCC studies. For example, both studies record detailed pathology data, including in many cases scans of the original pathology report. Both studies collect detailed prognostic information (e.g., cytogenetic results in acute leukemia, size of primary tumor in sarcoma) and treatment data (e.g., radiation site, dose and fraction number, chemotherapy with cumulative doses where relevant). Differences, nonetheless, were identified; for example, POCC collects data on the pre-diagnostic pathway (timing and type of symptoms) whereas IMPACT does not.

Substantial differences were also identified in the adjunct methodologies of the two studies. The POCC study methodology also includes surveys of recently diagnosed AYAs, allowing for the collection of data on AYA experiences of care not available through IMPACT. The IMPACT cohort is individually linked with Ontario population-based healthcare databases that are housed at the Institute for Clinical and Evaluative Sciences (ICES), allowing for long-term follow-up of outcomes that are identifiable by administrative healthcare data.¹⁵

Identification and Prioritization of Collaborative Study Questions

Before the meeting, we identified categories of research questions that were potentially amenable to collaboration between the two research teams. These questions are listed in Table 3, along with their ratings of importance by survey respondents. With the exception of comparisons of the incidence of AYA cancers between the two jurisdictions, respondents judged all categories of research questions to be of significant importance. The highest ratings were given to questions linking differences in treatment to differences in outcome, and to studying particularly rare types of AYA cancer by combining the two study populations. Meeting participants brainstormed research questions within each of the a priori identified categories. These questions were combined with those suggested by survey respondents, which included examining clinical trial enrolment across the two jurisdictions, uptake of fertility and psychosocial services, comparisons of vulnerable AYA subpopulations (e.g., rural; indigenous; and lesbian, gay, bisexual, and transgender), and access to novel cancer therapies.

Table 3.

Categories of Possible Collaborative Research Questions and Their Rating by Survey Respondents

Research questions	Rating median ^a	Rating IQR ^a
Comparisons of incidence of specific AYA cancer subtypes between the two jurisdictions (e.g., Is the incidence of specific leukemia subtypes in AYAs different in Australia than in Ontario, Canada?)	2	2–3
Comparisons of where AYAs are treated in each jurisdiction (e.g., Are Canadian AYAs more likely to be treated by pediatric oncologists than Australian AYAs?)	4	3–4
Comparisons of how AYAs are treated in each jurisdiction (e.g., Are Australian AYAs with leukemia more likely to get bone marrow transplants than Canadian AYAs with leukemia?)	4	3–4
Comparisons of AYA survival between the two jurisdictions (e.g., Are Canadian AYAs with bone tumors less likely to die of their disease than Australian AYAs with cancer?)	4	3–5
Linking differences in how AYAs are treated to differences in survival (e.g., Is a greater use of radiation in Australian AYAs with brain tumors associated with better survival as compared with Canadian AYAs with brain tumors?)	4	4–5
Some AYA cancers are so rare that neither study will have enough patients to accurately study them. We may combine both study groups to allow for a greater number of patients so that these rare AYA cancers can be better understood (e.g., What are the prognostic factors that determine survival in AYA rhabdomyosarcoma?)	4	4–5
Comparisons of the care and outcomes of potentially vulnerable subgroups of AYAs with cancer, such as those of particular ethnicities or in rural areas (e.g., Are rural/urban differences in AYA care and survival different between Australia and Ontario, Canada?)	4	3–4

Ratings used a 5-point Likert scale, with 1 representing “not at all important” and 5 representing “very important.”

IQR, interquartile range.

Meeting participants were then asked to prioritize the research questions while taking into account feasibility, potential to improve AYA outcomes, and added value of collaboration beyond what each study group could accomplish on its own. Several questions identified by meeting participants to be of high value to the AYA community were ruled out due to methodologic limitations (e.g., outcomes of lesbian, gay, bisexual, and transgender AYAs; detailed comparisons of fertility and psychosocial services). After significant discussion, three research areas were prioritized:

1. Disparities in cancer survival outcomes for vulnerable subpopulations of AYAs (e.g., rural AYAs)

2. Predictors of clinical trial enrolment among AYAs with cancer

3. Prognostic factors, treatment patterns, and outcomes in rare AYA cancers (e.g., rhabdomyosarcoma, medulloblastoma)

Data Sharing and Privacy Considerations

Data sharing and privacy restrictions may hinder successful collaboration. Ontario's privacy legislation prevents individual or small-cell level data from being exported from ICES, the research institute at which IMPACT data are housed, even in an anonymized form. Although exports of POCC data with appropriate privacy, confidentiality, and ownership safeguards are technically possible, the governance and regulatory burden necessary for such export is currently being explored, with the POCC study team making further enquiries into the feasibility of data sharing. The initial reaction from Australian regulatory bodies has been favorable, with data sharing possible assuming sign off from individual institutions. Should sharing of individual anonymized data prove impossible, collaboration will still be pursued using different proven analytic techniques such as distributed analysis. This would involve the conduct of separate analyses by following the same protocol, with either subsequent meta-analysis of IMPACT and POCC results or the export of aggregated and stratified data for combination and subsequent analyses.¹⁶

Involvement of Additional Stakeholders

Informed by the pre-meeting survey results, additional stakeholders, including policymakers, clinicians, advocacy groups, and AYA representatives, were identified whose input would be crucial in any subsequent collaboration. Examples of identified stakeholders from each jurisdiction can be seen in Table 4, although this list is not meant to be exhaustive. Strong knowledge translation plans were also identified as a necessary component of any future collaboration.

Table 4.

Proposed Stakeholders to Target for Input on Future Research Endeavors and Collaborations, by Jurisdiction

Group	Australia	Canada
AYA Advocacy	CanTeen	Canadian Partnership Against Cancer Taskforce on AYA
AYA representatives	CanTeen Youth advisory group	Young Adult Cancer Canada
Clinicians	Clinical Oncology Society of Australia and New Zealand Children's Haematology/Oncology Group	Canadian Association of Medical Oncologists C17 Children's Cancer and Blood Disorders
Policymakers	Cancer Council Australia	Pediatric Oncology Group of Ontario Cancer Care Ontario Canadian Cancer Society

Future Steps

Several key next steps were identified:

1. Feedback from the wider community of stakeholders

2. Terms of Reference (ToR): Subsequent to the meeting, ToR outlining guidelines for future collaboration were written. The ToR defined the core project team and their roles and responsibilities, the expected structure and frequency of communication between the core collaborative project team and both study teams, a framework by which new projects will be approved or rejected, guidelines for seeking external funding, principles of data ownership, and authorship guidelines. The ToR were endorsed by both the IMPACT and POCC study teams.

3. Pilot collaborative study: Of the identified priority collaborative research questions, one or two will be selected to submit to the project approval mechanism outlined by the ToR, and if approved, conducted. Additional external funding is not anticipated to be necessary for these pilot projects at this stage, though data sharing agreements and contracts will be required.

Once these steps are in place, future goals include joint submissions to external funding agencies to support larger collaborative projects. Other international AYA research teams will also be approached to encourage additional collaboration.

Conclusions

International linkages between national AYA cancer communities have not yet resulted in significant international AYA research collaborations. This article describes the initial efforts to create an international research partnership comprising research teams in Canada and Australia that will create an opportunity to push the AYA research agenda forward in a way not possible without such collaboration. Although challenges exist, such as addressing privacy and regulatory requirements associated with data sharing across jurisdictions, we are optimistic that successful collaboration between the IMPACT and POCC study teams will lead to improvements in AYA cancer care in both jurisdictions, and beyond. We are optimistic that the partnership can serve as a model to other AYA cancer researchers and encourage other international collaborations within the AYA cancer research community.

Footnotes

Acknowledgments

This meeting was supported by Canadian Institutes of Health Research (CIHR—Funding Reference Number 138873). In-kind contributions were provided by the POCC Team and Cancer Council Victoria. The Australian POCC study is supported by a National Health and Medical Research Council of Australia (NHMRC) Project Grant 1012250 and grants from Canteen and The Kids Cancer Project. The Canadian IMPACT study is supported by the C¹⁷ (partially funded by Childhood Cancer Canada Foundation and the Kids With Cancer Society), the Pediatric Oncology Group of Ontario Research Unit, the CIHR (133618), and the Cancer Care Ontario Health Services Research Chair awarded to Dr. Nancy Baxter. The authors would like to acknowledge investigators and members of the study teams who contributed to the preparation and planning of the collaborative meeting, distribution of pre-meeting materials, and preparation of presentations, as well as Dr. Luc te Marvelde for his input. They would also like to thank the survey respondents for their valuable insights.

Author Disclosure Statement

No competing financial interests exist.

References

Barr

, Rogers

, Schacter

. Adolescents and young adults with cancer: towards better outcomes in Canada. Cancer, 2011; 117:2239–354.

Gupta

, Pappo

, Saunders

, et al. Clinical outcome of children and adults with localized Ewing sarcoma: impact of chemotherapy dose and timing of local therapy. Cancer, 2010; 116:3189–94.

Stock

, La

, Sanford

, et al. What determines the outcomes for adolescents and young adults with acute lymphoblastic leukemia treated on cooperative group protocols? A comparison of Children's Cancer Group and Cancer and Leukemia Group B studies. Blood, 2008; 112:1646–54.

Boissel

, Auclerc

, Lheritier

, et al. Should adolescents with acute lymphoblastic leukemia be treated as old children or young adults? Comparison of the French FRALLE-93 and LALA-94 trials. Journal Clin Oncol, 2003; 21:774–80.

Pole

, Alibhai

SMH

, Ethier

, et al. Adolescents with acute lymphoblastic leukemia in Ontario have similar outcomes when treated at pediatric or adult hospitals. Ann Oncol, 2013; 24:801–6.

Furlong

, Rae

, Greenberg

, Barr

. Surveillance and survival among adolescents and young adults with cancer in Ontario, Canada. Int J Cancer, 2012; 131:2660–7.

Albritton

, Wiggins

, Nelson

, Weeks

. Site of oncologic specialty care for older adolescents in Utah. J Clin Oncol, 2007; 25:4616–21.

Steinbrook

. Private health care in Canada. New Engl J Med, 2006; 354:1661–4.

Hall

. Australian health care—the challenge of reform in a fragmented system. New Engl J Med, 2015; 373:493–7.

10.

Osborn

, Little

, Bowering

, Orme

. Youth Cancer Services in Australia: development and implementation. International perspectives on AYAO, Part 3. J Adolesc Young Adult Oncol, 2013; 2:118–24.

11.

Palmer

, Patterson

, Thompson

. A national approach to improving adolescent and you young adult (AYA) oncology psychosocial and care tools. Palliat Support Care, 2014; 12:183–8.

12.

Rogers

, De Pauw

, Schacter

, Barr

. A process for change in the care of adolescents and young adults with cancer in Canada. “Moving to Action”: the Second Canadian International Workshop. International Perspectives on AYAO, Part 1. J Adolesc Young Adult Oncol, 2013; 2:72–6.

13.

Canadian Cancer Society's Advisory Committee on Cancer Statistics. Canadian Cancer Statistics 2013. Toronto, ON: Canadian Cancer Society; 2013.

14.

Pinkerton

, Wills

, Coory

, Fraser

. Survival from haematological malignancy in childhood, adolescence and young adulthood in Australia: is the age-related gap narrowing?. Med J Aust, 2010; 193:217–21.

15.

Baxter

, Daly

, Gupta

, et al. The Initiative to Maximize Progress in Adolescent and Young Adult Cancer Therapy (IMPACT) Cohort Study: a population-based cohort of young Canadians with cancer. BMC Cancer, 2014; 14:805.

16.

Dormuth

, Hemmelgam

, Paterson

, et al. Use of high potency statins and rates of admission for acute kidney injury: multicenter, retrospective observational analysis of administrative databases. BMJ, 2013; 346:f880.