Uptake of Team Care Arrangements for adults newly diagnosed with cancer

Keywords: cancer patients, community health care, data linkage, health care utilisation, health services: accessibility, New South Wales, patient-centred care, Team Care Arrangements.

Cancer incidence is increasing as a result of improvements in early detection, diagnostic techniques, and treatments, and an aging population (Miller et al. 2019). In Australia, the 5-year relative survival rate for all cancers combined has increased from 51 to 70% between 1998–92 and 2013–17 (Australian Institute of Health and Welfare 2021). Many cancer survivors live with long-term multiple health conditions, due to the effects of the cancer itself, its treatment, the threat of cancer recurrence, and health conditions that pre-dated the cancer diagnosis (Leach et al. 2015; Vardy et al. 2019). The 2020–25 National Health Reform Agreement (Australian Government, Department of Health and Aged Care 2020) and key recommendations by the Clinical Oncology Society of Australia (Vardy et al. 2019) and Cancer Australia (Cancer Australia, Australian Government 2017) emphasised the need for improved integration and coordination across the healthcare system to improve the health and wellbeing of cancer survivors.

The Chronic Disease Management program was introduced into the Australian Medicare Benefits Schedule (MBS) in 2005 to provide coordinated health care to people with long-term chronic or terminal medical conditions. This program supports GPs to implement shared care via coordinated Team Care Arrangements (TCAs) established between collaborating health professionals, specifically themselves and two or more other healthcare providers. Teams may include specialists, allied health professionals, community service providers, health education providers, and personal care workers (Services Australia, Australian Government 2019).

Previous research has evaluated claims for MBS chronic disease items (Douglas et al. 2011) and TCAs (Harris et al. 2011; Vitry et al. 2012; Welberry et al. 2019a, 2019b), and observed these plans were more frequent among people with chronic and complex care needs, of lower social status (i.e. by income, area of residence or educational attainment), female and of an older age (Douglas et al. 2011; Harris et al. 2011; Vitry et al. 2012; Welberry et al. 2019a, 2019b). However, these studies did not focus on people diagnosed with cancer, a population group with a high prevalence of comorbid health conditions.

This study aimed to quantify the uptake of TCA, the factors independently associated with initiating a TCA, and the subsequent TCA-subsidised allied health services, by people with newly diagnosed cancer in New South Wales (NSW), Australia.

The Sax Institute’s 45 and Up Study is a prospective cohort study of 267 153 people aged ≥45 years residingin NSW, Australia (45 and Up Study Collaborators 2008). Recruitment was performed in 2006–09 using random sampling from the Services Australia (formerly the Department of Human Services) Medicare enrolment database, Australia’s universal health insurance provider. People aged 80+ years and residents of rural and remote areas were oversampled. Participants joined by self-completing a questionnaire about their socio-demographic characteristics, lifestyle behaviours and a range of health, functional and social measures. They also provided written consent to long-term follow up, including linkage with administrative health datasets. Approximately 18% of those invited participated in the study (45 and Up Study Collaborators 2008; Johar et al. 2012).

Data from the 45 and Up Study cohort were probabilistically linked to the NSW Cancer Registry (1972–2016), NSW Admitted Patients Data Collection (July 2001–19), NSW Emergency Department Data Collection (2005–19), and NSW Registry of Births, Deaths and Marriages (2006–19) by the Centre for Health Record Linkage (http://www.cherel.org.au/); and to the MBS (2001–19) and Pharmaceutical Benefits Scheme [(PBS) (2001–19)] by the Sax Institute using deterministic matching and a unique identifier provided by Services Australia.

Cohort participants were eligible for this study if they had:

• an incident-notified cancer diagnosis after 45 and Up enrolment;

• no access to Department of Veterans’ Affairs (DVA) health services, as subsidised health services are not fully captured for these patients;

• no cancer diagnosis in the 3 months immediately after cohort enrolment. This was necessary to minimise the impact of underlying, undiagnosed cancer on self-reported baseline characteristics;

• at least one linked MBS or PBS record; and

• no record of a TCA before the cancer diagnosis.

There were five categories of exposure of interest (Table 1): socio-demographic, health status, and health risk factors, which were measured at 45 and Up Study enrolment, except age at cancer diagnosis; cancer type; and healthcare utilisation prior to cancer diagnosis; that is, hospital admission, visits to emergency departments, GPs, specialists, and other health practitioners (Medicare subsidised services for allied health, multidisciplinary case conferences, optometry, mental health care, allied health group services, Chronic Disease Dental Scheme¹, and others).² For healthcare utilisation, we used a 12-month look-back period. To exclude healthcare utilisation associated with the diagnosis or treatment of the cancer, we disregarded the month of cancer diagnosis and the preceding 3 months, creating a look-back period of 4–15 months. All variables were divided into two subcategories (e.g. age at cancer diagnosis was split into 45–64 and 65 years).

Table 1.  Exposure variables available for analysis.

The outcome measure was the first claim for a TCA (MBS item 723). The secondary outcome was TCA-subsidised eligible allied health services accessed up to 24 months after the initial TCA claim: Aboriginal health workers or Aboriginal and Torres Strait Islander health practitioners (MBS item 10 950), audiologists (10 952), chiropractors (10 964), diabetes educators (10 951), dietitians (10 954), exercise physiologists (10 953), mental health workers (10 956), occupational therapists (10 958), osteopaths (10 966), physiotherapists (10 960), podiatrists (10 962), psychologists (10 968) and speech pathologists (10 970).

All incident-notified primary cancer diagnoses were identified from linked cancer registry records. All patients were followed from the date of cancer diagnosis until TCA uptake, death, or the end of the observation period (December 2019), whichever occurred first (Fig. 1).

Fig. 1.  Visual representation of the project design: Timeline of events for two hypothetical patients. Patient 1 has a cancer diagnosis approximately 2 years after enrolment to the 45 and Up Study. Patient 2 has a cancer diagnosis approximately 6 months after enrolment to the 45 and Up Study.

Characteristics of eligible participants were summarised using descriptive statistics. We used multivariable Cox proportional hazards modelling to identify factors associated with the uptake of a TCA plan after cancer diagnosis. We tested the proportional hazard assumption of the Cox model for each variable of interest independently using the Kaplan–Meier method. We excluded variables whose curves for each subcategory crossed, indicating a violation of the proportional hazards assumption. We also excluded variables for which there was no statistical difference between the subcategories in the uptake of TCA at any time point using the Log Rank test.

We assessed the variance inflation factor (VIF) for the remaining variables. Variables with a VIF >1.5 were considered to exhibit multicollinearity and not included in the multivariable model (Johnston et al. 2018). Univariable and fully adjusted Cox proportional hazards regression models were built to examine the relationship between the exposure variables and a TCA claim. We also performed sensitivity analyses to account for the potential competing risk of death before the uptake of a TCA. We tested the proportional hazards assumption for each variable included in the multivariable model based on the Schoenfeld residuals.

We counted the number and type of eligible TCA-subsidised allied health services in the 24 months after TCAs were initiated. For this 24-month period, we also calculated the inverse of the survival function via the Kaplan–Meier method to measure the cumulative uptake of subsidised services, with death or end of the study as the right censor.

We accessed anonymised unit record data via the Secure Unified Research Environment (SURE). Analyses were performed using SAS (ver. 9.4; SAS Institute Inc., Cary, NC, USA) and R (ver. 3.6.3; R Foundation for Statistical Computing, Vienna, Austria).

This research was approved by the NSW Population and Health Services Ethics Committee (Approval Number: HREC/14/CIPHS/50). The 45 and Up Study was approved by the University of New South Wales Human Research Ethics Committee.

A total of 13 951 people with an incident cancer diagnosis after enrolment in the 45 and Up Study had no prior TCA plan in place (Fig. 2). The mean age of the cohort was 68 years at the time of cancer diagnosis, 8202 were male (58.8%), 10 678 were married or in a de facto relationship (76.5%), and 10 624 were born in Australia (76.2%) (Table 2).

Fig. 2.  Cohort flow chart illustrating the inclusion and exclusion criteria of participants in the study. DVA, Department of Veterans’ Affairs.

Table 2.  Cox proportional hazards models to identify the effect of the selected variables upon the time to the uptake of a new TCA plan after the first cancer diagnosis.

In total, 6630 patients had a TCA plan initiated (47.5%) and the median time from cancer diagnosis to TCA uptake was 24 months (interquartile range (IQR) 9–51 months). All exposure variables were significantly associated with TCA uptake in univariable analyses (Table 2). Residential location could not be included in the models because the Kaplan–Meier curves crossed. In the multivariable model, four variables, sex, age, private health insurance and smoking status, violated the proportional hazards assumption based on the Schoenfeld residuals. We addressed the violation in two ways. We stratified by sex and split the variables age, private health insurance and smoking status into two-time intervals according to their temporal behaviour: short-term (the same year and the year after cancer diagnosis) and long-term (2 years after diagnosis until the end of follow up) (Table 2).

In the final multivariable model, the uptake of TCA was positively associated with being aged >65 years at the time of diagnosis, having lower educational attainment, not having private health insurance, living in areas of greater socio-economic disadvantage, having visited the GP six or more times during the 12-month look-back period, having visited other health practitioners one or more times during the 12-month look-back period, having a lower self-rated overall health status, having one or more comorbidities, and being a past or current smoker (Table 2).

Overall, 4577 patients died during follow up (32.8%). The median time from cancer diagnosis to death was 19 months (IQR 6–50 months). Among those who died, those without a TCA (3057) had a shorter median time to death, 11 months (IQR 3–22 months), compared to those with a TCA (1520), 49 months (IQR 25–71 months). The multivariable models for TCA uptake were unchanged in sensitivity analyses where death was considered a competing risk (data not shown).

The median time from TCA uptake to death was 24 months (IQR 9–49 months). A total of 4084 patients had at least one eligible TCA-subsidised allied health visit within 24 months of TCA initiation (61.6%) (Fig. 3). Of those, 2812 patients had five or fewer visits (68.9%), with five visits being the most common (Table 3). The most consulted TCA-subsidised allied healthcare professionals were physiotherapists and podiatrists, followed by dietitians and exercise physiologists (Table 4). The most common combination of TCA-subsidised allied health visits was physiotherapists and podiatrists followed by physiotherapists and dietitians. The median time between the TCA and the first subsidised allied health service was 1 month (IQR, 0–5 months) (Fig. 3).

Fig. 3.  Cumulative uptake of first TCA-subsided allied health service in the 24 months since TCA initiation.

Table 3.  Number of visits per patient for any allied health service covered by TCA in the 24 months after the TCA.

Table 4.  Number of visits per patient for TCA-subsidised allied health services covered by TCA in the 24 months after the TCA.

Of 13 951 adults aged ≥45 years who were diagnosed with incident cancer in 2006–16, around half had a TCA plan initiated over a median duration of 24 months since diagnosis. TCA uptake was more likely for people with pre-existing health conditions, lower self-rated health, higher healthcare use prior to cancer diagnosis, and those with lower economic means. This patient profile is similar to those observed in prior evaluations of the 45 and Up Study as a whole, unselected for cancer (Douglas et al. 2011; Welberry et al. 2019a). We found that TCA uptake was associated with visits to health practitioners other than GPs and specialists prior to cancer diagnosis, a relationship that no previous study examined. Unlike these prior studies, we did not find an association with prior specialist consultations or language spoken at home (Welberry et al. 2019a). The strongest factors associated with TCA uptake were older age and higher GP attendance prior to cancer diagnosis.

Our results suggest that TCAs for cancer patients are being provided for those with more complex healthcare needs. Individuals newly diagnosed with cancer commonly have concomitant chronic health conditions and factors such as poor diet or inactivity due to the cancer treatment can increase the severity of pre-existing conditions (Jefford et al. 2020). Several studies have demonstrated that cancer patients can be safely and effectively managed by GPs (Lim et al. 2018; Jefford et al. 2020). There is a complex interaction between cancer and other chronic health conditions because cancer and chronic conditions may share risk factors including smoking, poor diet, obesity, physical inactivity, and excessive alcohol intake. Furthermore, cancer may be associated with chronic conditions; for example, uterine cancer is associated with obesity and diabetes, and finally cancer and chronic conditions are both common so they may overlap (Sarfati et al. 2016).

Our results also showed that TCAs were established for those most in need from a socioeconomic perspective. TCAs are specifically indicated for patients who have little or no capacity to access or receive the healthcare services they need via the usual referral process (Hart et al. 2021). The relatively weak association with country of birth and the lack of association with main language spoken at home may indicate a subset of cancer patients in whom TCAs are being underutilised.

Several studies have demonstrated that cancer patients can be safely and effectively managed by GPs and other primary care practitioners like nurses (Lim et al. 2018; Jefford et al. 2020; Chan et al. 2021). Nevertheless, specialist-led care continues to be the norm in Australia and other countries (Chan et al. 2021). GP-led care plans may be more appropriate to deliver follow-up care, especially if they prove to be more sustainable and cost-effective. For this to happen, GP-led teams should consider cancer survivors as a unique cohort based upon their specific additional needs. Holistic, patient-centred cancer survivorship care requires clear communication and clarity regarding the roles of practice nurses, allied health professionals and GPs in the care and follow up of these patients. It is also important to achieve consensus regarding the scope of practice of primary- and specialist-led care, which is another barrier to GPs having a leading role in survivorship care (Fox et al. 2022). A partnership approach is the goal, and a specialist practice nurse could take responsibility for coordinating communication between the general practice team and cancer specialist, with a view to providing a tailored approach for each patient. Evidence-based education of the primary care workforce about cancer survivor care needs and the utility of TCAs within a broader survivorship care plan may help improve TCA uptake and patient outcomes.

Around 40% of cancer patients with a TCA plan did not attend any eligible community-based TCA-subsidised allied health services. There are a number of possible reasons for this finding. Patients may not have understood the need for private allied health services, they may have received this treatment from the public sector, they may have given allied health services a lower priority in order to concentrate on their follow-up cancer care, there may have been restricted availability of appointments for these services, or they may have been unable to pay the fee gap.

The allied health practitioners most utilised under this scheme were podiatrists, physiotherapists, exercise physiologists, and dietitians. Russell et al. (2020) found similar results. Examples of the relevance and benefits of allied health services to cancer patients include physiotherapy for cancer-related fatigue (Pyszora et al. 2017), podiatrists to prevent, diagnose and manage podiatric toxicity of cancer treatments (Lacouture et al. 2018), exercise physiologists and physiotherapists to prescribe and oversee exercise and manual therapies to attenuate cancer treatment-related side-effects (Cormie et al. 2020), and dietitians to help address weight and muscle loss commonly experienced by cancer patients (Sullivan et al. 2021).

We were unable to identify allied health consultations provided by Primary Health Networks, the National Disability Insurance Scheme and cancer centres, as they are not captured in MBS datasets. We were also unable to identify whether some specialist or consultant physicians were part of the TCA plan, as there are no exclusive billing codes for these consultations under the TCA.

We could not examine the association between TCA uptake and household income due to the high proportion of cohort participants not answering this question. And 92.7% of participants spoke English at home limiting conclusions about TCA uptake and non-English speakers. Furthermore, we could not account for the uptake of allied health services with interpreters in the public system. The association with socioeconomic factors may not be generalisable, as the participation rate was low (~18%) and the 45 and Up Study sample is skewed towards higher-income groups. Finally, we could not examine the association between cancer stage at diagnosis and TCA uptake due to the high proportion of patients with unknown cancer stage (>25%) and evidence that this variable was not missing at random.

TCAs are common in older adults after a cancer diagnosis. Our findings suggest that TCAs are being initiated in cancer patients with the greatest health and socio-economic needs. Future research may identify whether GP characteristics are also associated with TCA uptake, and also why some patients do not have any subsequent claims for eligible TCA-subsidised community-based allied health services. It is unclear whether the need for these services was not understood, no longer deemed necessary, were displaced by other care needs (such as management of cancer recurrence), were found in the public sector or could not be accessed. This warrants further investigation of the trajectories of specific patient subgroups, including by cancer type.

Restrictions apply to the availability of these data, which were used under ethics approval for the current study, and so data are not publicly available.

The authors declare that they have no conflicts of interest.

This research received financial support from the Translational Cancer Research Network (a Cancer Institute NSW Translational Cancer Research Centre).