Princess Alexandra Hospital model of comprehensive geriatric assessment of cancer patients: methodological and practical aspects



  1. Division of Cancer Services, Princess Alexandra Hospital, Brisbane, Queensland
  2. School of Nursing and Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Queensland
  3. School of Social Sciences, University of Tasmania


There is increasing momentum in oncology to implement a two stage assessment process that accurately determines the ability of older patients to cope with, and benefit from, chemotherapy for cancer. The two-step approach aims to ensure that patients clearly fit for chemotherapy can be accurately identified and referred for treatment without undergoing a time and resource intensive Comprehensive Geriatric Assessment. Ideally, a two-step process removes the uncertainty of how to classify and then appropriately treat the older cancer patient. After trialing a two-stage screen and Comprehensive Geriatric Assessment process in the Division of Cancer Services at Princess Alexandra Hospital in 2011-12, a model of oncogeriatric care was implemented based on the findings. In this paper, the methodological and practical aspects of implementing the Princess Alexandra Hospital model are explored and further work needed to refine the process is outlined.

A scan of the oncogeriatric research literature reveals a wide variety of tools used for screening. Common approaches range from gross functional assessments such as the Karnofsky Performance Index, to more detailed functional assessment by way of the Barthel Index or the Lawton Activities of Daily Living Scale. Geriatric-specific assessment using the Geriatric-8 and Vulnerable Elders Survey-13 (VES-13) is also common, designed to capture the wider range of variables that affect function and the subjective variables likely to influence outcomes in older patients. There are also cancer-specific screening tools, such as the abbreviated Comprehensive Geriatric Assessment, that elicit a range of variables thought specific to gerioncology.1

Irrespective of the method chosen, the screen should be sufficiently sensitive to ensure that older patients deemed suitable for systemic treatment are actually fit to proceed to treatment. The screen should also have sufficient specificity to ensure that the time consuming process of the second stage – a Comprehensive Geriatric Assessment (CGA) for those judged more vulnerable on screening – is actually necessary.2 Given the diversity of tools recommended in the literature however, the composition of the screening tool is clearly open to interpretation. In the authors experience, it is not possible to standardise screening across sites, as the patient variables that must be assessed are highly specific to the setting in which they occur. The treatment context, which is a major metropolitan referral centre servicing one third of Queensland, has sociodemographic challenges in terms of health literacy, language differences, social deprivation and degree of rurality. These greatly influence the treatment outcomes, but might not have the same impact in other settings. A robust screen is one that provides a careful appraisal of variables, with the potential to affect treatment outcomes in a given situation.

VES-13 is perhaps the most popular screening tool, and the one trialled in this study after appraising the evidence.3 In non-cancer settings, VES-13 is commonly used to quickly evaluate patients’ functional abilities and deficits, their self-reported health status and their chronological age. The VES-13 then grossly categorises older people into the categories of ‘fit’ for treatment, or ‘vulnerable’ to treatment and in need of further comprehensive assessment. While not cancer-specific, VES-13 is recommended in oncology settings to quickly differentiate individuals fit for standard chemotherapy from those who may benefit from a more thorough assessment and possible therapy modification.4

The data from the study (under review) revealed that while it had potential, VES-13 alone was not sufficient to meet clinical needs.5 The conclusions were supported by a paper published just after the preliminary data analysis was completed. Hamaker et al’s systematic review of the seven most common geriatric screening tools – including VES-13 – concluded that irrespective of their value in other geriatric settings, none of these tools had sufficient discriminative power in cancer settings to recommend their use.2 Pooled data in this meta-analysis indicated that VES-13 had a sensitivity of 68% and specificity of 78% in predicting patient outcomes.2 Despite their limitations, Hamaker et al concluded that screening tools did have the potential for modification and enhanced rigour in cancer care.2 If this can be achieved, the value of screening in terms of rationalising resources and referring only those patients potentially vulnerable to treatment for a comprehensive assessment is significant.4

As a result of the study, a composite screen has been implemented. It includes VES-13, which is supplemented with tools that elicit other important contextual patient variables commonly encountered in service, such as malnutrition risk, body mass index, level of distress and health-related quality of life (table 1). The validity and reliability of this composite screener in future studies will be rigorously evaluated.

Which CGA tools?

A CGA comprising a suite of validated tools is designed to determine fitness for treatment in older people after potential vulnerabilities are identified through screening.4 CGA was developed in community and rehabilitation aged care settings to thoroughly assess health problems amenable to correction. In these settings it has demonstrated robustness in identifying older people’s health risks, and for improving their health outcomes by linking them to appropriate interventions.6 More recently, it has translated well to selected acute care settings.7

CGA generally comprises a full medical and social history, followed by a battery of scales that assess the physical, psychological, cognitive and functional domains of health in older patients. While there is consensus in the literature that these domains should be evaluated, as with screening tools there is less agreement as to precisely how this should be undertaken. Irrespective of its composition, a CGA in cancer settings should ensure that: those individuals who are amenable to intensive chemotherapy (after their deficits are identified and remedied) are appropriately treated; that vulnerable patients more suited to modified or supportive regimens are determined; and that frail individuals who would benefit most from palliative regimens, or no treatment at all, are also identified and offered the appropriate level of care.8,9

Hence, it is vital that comorbidities are assessed because pre-existing illnesses affect recovery from cancer treatment and are also correlated with cancer treatment efficacy.10 In the study that informed this model, the gold standard Cumulative Index Rating Scale-Geriatrics (CIRS-G) to elicit comorbidities was used.11 In the course of the study however it was found that while the CIRS-G was indeed methodologically robust, it was logistically cumbersome to build into the model of oncogeriatric assessment that the context demanded. For this reason, the current version of the model uses the Charlson Comorbidity Index – an equally rigorous tool in the cancer context, but one hopefully more useable in practice.

In the psychosocial domain, a CGA should elicit the cognitive and affective status of older patients, which are also believed to contribute to cancer treatment outcomes.8,10 These variables were assessed in this study with the Standardised Mini-Mental State Examination (SMMSE) and the Geriatric Depression Scale (GDS), both of which proved problematic. The GDS, for example, while a widely-accepted component of CGA in non-cancer settings, is not well-validated in oncology-specific contexts.12 It was also not clinically acceptable, in that its fixed yes/no response set was difficult to deal with in this socially-sensitive situation, when patients had been newly diagnosed. It was also realised that the equal weighting of items in the GDS might be misleading in patients, who were often reasonably and reactively distressed by their new diagnosis, as opposed to endogenously depressed over the longer term.13 In addition, it is likely that the SMMSE and GDS do not reliably measure the fitness potential of patients with low levels of education, those from different ethnic backgrounds, and those whose first language is not English. Tests such as the SMMSE and the GDS are culture-specific, reflecting Western cultural values and expectations about cognitive capacity and emotional function. These values do not necessarily translate to all patients in the culturally-diverse treatment context,14 which serves large immigrant Chinese, Burmese, Pacific Islander and refugee African communities. Hence, the assessment could produce false results for those patients.15 So even though the SMMSE is used for the present, it is recognised that it is a rather blunt instrument in certain situations. The evidence has also been further appraised and the CES Depression Scale in the model of oncogeriatric care has been tested and evaluated.

A range of factors such as fiscal pressures, staff expertise and availability, patient demographics, and organisational resources and commitment influences the choice of tools, how they are used and how the data they produce are interpreted in a given context. In Australia, for example, local factors resulted in the development of quite different CGA approaches in South Australia and New South Wales to the model tailored to the Queensland situation.16,17 Table 1 outlines the full suite of tools in the current Princess Alexandra Hospital model.

In addition to these, the study  data  indicated  that  the  assessment  should  summarise  past  and  present  medical history, including all medications to elicit potential problems related to polypharmacy, investigate continence and gather information about material and social support structures in the patient’s home environment.


Logistical issues

In addition to the methodological tensions described above, logistical issues must be considered when implementing an oncogeriatric model of care in a fiscally-challenging climate. Good assessment of elderly cancer patients is not only complex, it is time-consuming. A CGA can take up to two hours to complete, not including the time taken to then review the data, call in a geriatrician to consult on problematic areas and to formulate a plan for further management.8,26 The experiences during the trial emphasise that oncologists lack the time and resources to undertake these complex assessments. Furthermore, it is difficult to meet the needs of our rural patients, some of whom live up to 10 hours drive away. Using the available resources in a way that would not upset the equilibrium of existing people and processes within the service, a creative solution needed to be determined which would still meet the needs of clinicians and patients.

Issues that were flagged by the service as problematic were: the need to avoid replication of existing procedures; to clearly define the roles of personnel within the model so that professional boundaries were clear; to establish efficient procedures for assessment and referral; and to ensure the assessment outcomes were meaningful for all the patients and health professionals concerned i.e. there had to be buy-in from all clinicians from all disciplines. The mechanism of delivery also needed to be scoped. Some CGA models for example, entail mailing surveys to the patient in a reply-paid envelope prior to their first clinic visit; others involve the patient entering their own data into a computer system while waiting for their oncologist appointment.27 These delivery modes are driven by contextual needs and are successful for that reason. At Princess Alexandra Hospital however, the language and computer literacy of a significant proportion of patients required consideration. In addition, patients and clinicians in our study had clearly expressed a preference for face-to-face assessment.

To resolve these tensions, a CGA model wherein the CGA data are collected by a skilled nurse trained in oncogeriatric assessment, was developed. The CGA nurse personally interviews each patient prior to the first oncologist visit where possible, to administer the survey tools. Where a patient cannot attend personally, the screener plus any CGA tools that don’t need to be administered face-to-face are administered by phone, with the rest of the data collected personally at the first clinic appointment. The CGA nurse then enters the data into the computerised oncology patient management system, along with a summary of the objective and subjective data. The summary is sent to the treating doctor and also made available to the tumour stream co-ordinators and relevant allied health personnel for multidisciplinary review to enable a collective treatment decision to be made. This model overcomes the logistic obstacle of overstretched health professional resources and is financially viable, given that a dedicated nurse trained in this procedure is clearly more cost-effective than a two hour assessment undertaken by a time poor oncologist and a further geriatrician consult. It also ensures that treatment is truly multidisciplinary and is convenient to patient needs.

Next step: a computerised decision support system

The model will be subject to ongoing evaluation and refinement. The process must be made more efficient, and fiscal pressures permitting, this will be done by developing telehealth procedures for more remote patients, satellite clinics in outer urban referral services, and a computerised decision support system linked to all patients’ assessments. Algorithms incorporated into computerised decision support systems are an effective way to simplify access to the often disparate data necessary to make decisions, reduce costs, integrate workflow and to alert clinicians when new or significant patterns in patient data arise, while maintaining a patient-centred focus on the individual who is the point of care.28 As with any clinical practice however, it is vital that the computations underpinning the system can be tempered by judgments derived from clinical experience in specific populations.29 To ensure clinical acceptability, it is also imperative that the methods used to develop the computerised algorithm, and the way the items within it are weighted, are open to external scrutiny so that their rigour can be verified, they can be improved if necessary and adapted to specific situations.29 With respect to CGA, the InterRai Consortium has developed computerised decision-making algorithms for various settings, excluding cancer care. While the way InterRai developed the weightings underpinning the algorithms is not clear, the group reports that the algorithms achieve good outcomes in community, residential and other acute care settings.30 In cancer care, Massa et al developed an algorithm underpinning their CGA data but similarly, the method used to weight the components of the CGA and how they contribute to their algorithm was not explained. The work from here entails testing the screening and CGA data to determine the most accurate algorithm to guide care.

This work to date indicates the potential of the two stage assessment process to improve the care of older cancer patients, but not its actual ability to do so. It is clear that the process can identify problems within the older patient’s domains of health that require further consideration and referral, and it certainly contributes to a holistic and multidisciplinary model of care. However, the predictive properties of screening and CGA are not truly understood at this stage.2,31,32 While it is an excellent mechanism for referral and for correction of deficits that can affect treatment, it is not known whether such a process can accurately identify older cancer patients’ fitness or otherwise for chemotherapy, or guide appropriate therapeutic choices. It is also not known whether the data from CGA can be weighted to reliably inform a computerised algorithm to support clinical decision-making. However, given the personal, social and clinical consequences for older cancer patients of inappropriate or under-treatment, there is clearly a need to continue to robustly test context-specific oncogeriatric assessment processes like this in the Australian setting.


  1. Overcash JA, Beckstead J, Moody L, Extermann M, Cobb S. The abbreviated comprehensive geriatric assessment (aCGA) for use in the older cancer patient as a prescreen: Scoring and interpretation. Critical Reviews in Oncology/Hematology. 2006;59(3):205-210.
  2. Hamaker ME, Jonker MM, de Rooij SE, Vos AG, Smorenburg CH, van Munster BC. Frailty screening methods for predicting outcome of a comprehensive geriatric assessment in elderly patients with cancer: A systematic review. Lancet Oncology. 2012;13:437-444.
  3. National Comprehensive Cancer Network (NCCN). Senior Adult Oncology v.2.2007. NCCN Clinical Practice Guidelines in Oncology. 2007; Accessed 23rd May, 2012.
  4. National Comprehensive Cancer Network (NCCN). Senior Adult Oncology v.2.2007. NCCN Clinical Practice Guidelines in Oncology 2007; Accessed 4th August, 2008.
  5. Thomson D, McCarthy A, Yates P, Skerman H, Walpole E. Comparison of ratings of fitness for chemotherapy using Vulnerable Elders Survey-13 versus physician’s judgments Annual meeting of the American Society of Clinical Oncology: Building Bridges to Conquer Cancer. McCormick Place, Chicago, Illinois, United States. : American Society of Clinical Oncology; 2013.
  6. Gray L, Wootton R. Comprehensive geriatric assessment ‘online’. Australasian Journal on Ageing. 2008;27(4):205-208.
  7. Gray L, Bernabei R, Berg K, Finne-Soveri H, Fries BE, Hirdes JP., Arino-Blasco, S. Standardising assessment of elderly people in acute care: The interRAI Acute Care Instrument. Journal of the American Geriatrics Society. 2008;56:536-541.
  8. Singhal N, Rao AV. Tools for assessing elderly cancer patients. Cancer Forum. 2008;32(1):6-7.
  9. Massa E, Madeddu C, Astara G, Pisano M, Spiga C, Tanca FM. An attempt to correlate a “Multidimensional Geriatric Assessment” (MGA), treatment assignment and clinical outcome in elderly cancer patients: Results of a phase II open study. Critical Reviews in Oncology Haematology. 2008;66:75-83.
  10. Hurria A, Lachs MS, Cohen H, Muss H, Kornblith AB. Geriatric assessment for oncologists: Rationale and future directions. Critical Reviews in Oncology and Haematology. 2006;59:211-217.
  11. de Groot V, Beckerman H, Lankhorst GJ, Bouter LM. How to measure comorbidity: a critical review of available methods. Journal of Clinical Epidemiology. 2003;56(3):221-229.
  12. Nelson CJ, Cho C, Berk AR, Holland J, Roth AJ. Are Gold Standard Depression Measures Appropriate for Use in Geriatric Cancer Patients? A Systematic Evaluation of Self-Report Depression Instruments Used With Geriatric, Cancer, and Geriatric Cancer Samples. Journal of Clinical Oncology. January 10, 2010;28(2):348-356.
  13. Adams KB. Depression symptoms, depletion, or developmental change? Withdrawal, apathy, and lack of vigor in the Geriatric Depression Scale. The Gerontologist. 2001;41(6):768-777.
  14. Struhkamp R, Mol A, Swierstra T. Dealing with independence: Doctoring in physical rehabilitation practice. Science, Technology and Human Values. 2009;34(1):55-76.
  15. Brijnath B. Screening for dementia: Fluidity and the Mini Mental State Examination in India. Transcultural Psychiatry. 2011;48(5604-623).
  16. Steer CB, Chia PL, O’Connor J, Underhill C, Donnelly J,Myers R et al. Utilising existing community-based supportive care and aged care resources for older patients with cancer: Updated results of the Care Coordination in the Older Adult with Cancer (CCOAC) project. Journal of Geriatric Oncology. 2012;3(S1):S31.
  17. To THM, Okera M, Prouse J, Prowse RJ, Singhal N. Infancy of an Australian geriatric oncology program: Characteristics of the first 200 participants. Journal of Geriatric Oncology. 2010;1(2):81-86.
  18. Koppie TM, Serio AM, Vickers AJ, Vora, K, Dalbagni G, Donat SM. Age-adjusted Charlson comorbidity score is associated with treatment decisions and clinical outcomes for patients undergoing radical cystectomy for bladder cancer. Cancer, 2008;112(11):384-2392.
  19. Sherbourne CD, Stewart AL. The MOS Social Support Survey. Soc Sci Med. 1991;32(6):705-714.
  20. Kai H, Hoi SF, Wong E, Lo R, Woo J. Do pain and disability differ in depressed cancer patients? Palliative Medicine. 2000;14(1):11-17.
  21. Hopko DR, Bell J L, Armento MEA, Robertson SMC, Hunt MK, Wolf NJ, et al. The Phenomenology and Screening of Clinical Depression in Cancer Patients. Journal of Psychosocial Oncology, 2007;26(1):31-51. Retrieved from doi:10.1300/J077v26n01_03
  22. Cohen ME, Marino RJ. The tools of disability outcomes research functional status measures. Archives of Physical and Medical Rehabilitation. 2000;81(Suppl. 2):S21-S29.
  23. Graf C. The Lawton Instrumental Activities of Daily Living (IADL) Scale. Annals of Long-Term Care. 2007;15(7):21-22.
  24. Rydwik E, Bergland A, Forsen L, Frandin K. Psychometric properties of Timed Up and Go in elderly people: a systematic review. Physical and Occupational Therapy in Geriatrics. 2011;29(2):102-125.
  25. Thornton M, Travis SS. Analysis of the reliability of the Modified Caregiver Strain Index. Journal of Gerontology. 2003;58B(2):S127-132.
  26. Hurria A, Gupta S, Zauderer M, Zuckerman EL, Cohen HJ, Muss H. Developing a cancer-specific geriatric assessment. Cancer, 2005;104:1998-2005.
  27. McCleary NJ, Wigler D, Berry D, Sato K, Abrams T, Chan J, Meyerhardt JA. Feasibility of computer-based self-administered cancer-specific geriatric assessment in older patients with gastrointestinal malignancy. The Oncologist, 2013:18, 64-72.
  28. Berner ES. Clinical Decision Support Systems: State of the Art: Agency for Healthcare Research and Quality (AHRQ);2009.
  29. Trivedi M, Daly E, Kern J, Grannemann B, Sunderajan P, Claassen C. Barriers to implementation of a computerized decision support system for depression: an observational report on lessons learned in “real world” clinical settings. BMC Medical Informatics and Decision Making. 2009;9(1):6.
  30. Gray LC, Berg K, Fries BE, Henrard JC, Hirdes JP, Steel K, et al. Sharing clinical information across care settings: The birth of an integrated assessment system. BMC Health Services Research, 2009. doi: 10.1186/1472-6963-9-71.
  31. Ellis G, Whitehead MA, O’Neill D, Langhorne P, Robinson D. Comprehensive geriatric assessment for older adults admitted to hospital. The Cochrane Database of Systematic Reviews. 2011;2011(7).
  32. Puts MTE, Hardt J, Monette J, Girre V, Springall E, Alibhai SMH. Use of geriatric assessment for older adults in the oncology setting: A systematic review. Journal of the National Cancer Institute. 2012;104:1133-1163.

Be the first to know when a new issue is online. Subscribe today.