Understanding Unwarranted Variation in Care for Victorian Oesophagogastric Cancer Patients Using Linked Data: A Comparative Analysis Between 2012–2016 and 2017–2021

ABSTRACT

Background

The Victorian Integrated Cancer Services convened the second Oesophagogastric (OG) Cancer Optimal Care Summit to determine unwarranted variation in care between 2017 and 2021 and compare trends with the first summit (reporting 2012–2016). Statewide administrative datasets were assessed retrospectively at the population level in alignment with the optimal care pathway steps for OG.

Methods

Victorians with a primary diagnosis of OG cancer were identified via the Victorian Cancer Registry dataset and linked to various administrative datasets to identify unwarranted variations.

Results

Of the 9868 Victorians diagnosed with OG cancer between 2012 and 2021, 39.4% were metastatic at diagnosis. Improved outcomes between time periods included reduced mortality for OG cancer and post‐surgical mortality 1 year post gastrectomy. Patient multidisciplinary meeting presentation increased from 74% to 86%. Unwarranted variations included time from diagnosis to any treatment within 6 weeks for non‐metastatic gastric cancer (64%, 2012–2016 and 60%, 2017–2021) and for oesophageal cancer patients (58%, 2012–2016 and 60%, 2017–2021). One regional Integrated Cancer Service (ICS) demonstrated statistically lower survival compared to the statewide average for gastric cancer in 2017–2021. There was greater variation in survival for gastric cancer for 2017–2021 compared to 2012–2016. Rates of dietitian and/or physiotherapist involvement within 3 months of diagnosis for OG cancer surgical and/or chemotherapy patients were low across all ICS and both time periods.

Conclusions

The Victorian cancer system demonstrates improved survival outcomes for OG cancer patients between the two time periods. Time between diagnosis and treatment, lower survival rates in some ICS, and access to supportive care (e.g., dietetics and physiotherapy) remain areas of deficiency.

1. Introduction

Oesophagogastric (OG) cancer is a key contributor to the global disease burden [1]. The five‐year relative survival for OG is low, due largely to patients often being diagnosed at a late stage when treatment options are limited [1]. These factors may be compounded by a lack of access to potentially curative therapies and services offered in mostly major metropolitan centres [2]. Regional variations in quality care may mean that patients living outside large metropolitan centres do not receive optimal care and therefore have poorer cancer outcomes [3].

The Optimal Care Summits programme is a leading initiative of the Victorian Integrated Cancer Services (VICS) delivered in collaboration with the Victorian Government Department of Health (DH). The purpose of the VICS Optimal Care Summits programme is to identify unwarranted variation in cancer practice and outcomes and provide recommendations for the VICS to address at the state, network, and/or health service level. The programme defines an unwarranted variation as a difference not explained by underlying illness or preferences, that is, a chance to improve the quality and equity of clinical care [4]. The VICS Optimal Care Summits programme also supports a broader programme of work of implementing nationally endorsed Optimal Care Pathways (OCPs) [5] for people with cancer, including one specific to OG cancer [6]. The pathway outlines seven steps in providing optimal care (Figure S1).

The first Victorian OG cancer summit was held in 2016. In March 2024, the VICS delivered the second Victorian OG cancer summit in collaboration with the DH. A multidisciplinary OG cancer expert advisory group representing all Integrated Cancer Services (ICS) (Figure 2) oversaw the analysis of population‐level and health record audit data and statewide stakeholder consultation. The present study aims to describe Victorian unwarranted variations in access, timeliness, and outcomes of OG cancer care across the eight‐adult ICS against five of the seven steps of the OG OCP (steps 2–5 and 7). Data were compared across two time periods, 2012–2016 (the period used for the first OG cancer summit) and 2017–2021 (the most recent cohort at the time of analysis). The focus of this study is on evaluating system‐level performance using linked, population‐level data to support evidence‐informed discussions and drive service improvement initiatives.

2. Methods

Data collection included access, analysis of a range of linked administrative cancer datasets for de‐identifiable OG cancer information. This included the Victorian Cancer Registry (VCR) dataset, Victorian Admitted Episode Dataset (VAED), Victorian Emergency Management Dataset (VEMD), and the Victorian Radiotherapy Minimum Dataset (VRMDS). Other data included Victorian Cancer Service Performance Indicators (CSPI).

2.1. Data Sources

The VCR is a population‐based cancer registry that collects demographic and tumour details for all Victorian residents who are diagnosed with cancer. The Department's Centre for Victorian Data Linkage performed an annual data linkage between the VCR and administrative datasets including the VAED, VEMD, VRMDS, and the Victorian and National Death Index. This linkage added information captured within the inpatient setting in all Victorian public and private hospitals (from the VAED), as well as data from admitted and non‐admitted radical and palliative radiotherapy delivered in Victorian public and private radiotherapy centres (from the VRMDS). Deaths were identified through linkage to the National and Victorian Death Index. VEMD provided data for presentations to the emergency department.

Unlinked data sources included the Victorian CSPI health record audits from 2017 [7] and 2022 [8]. These audits collected data such as multidisciplinary meetings (MDM) information from the health records of a sample of oesophageal and gastric cancer patients treated across 47 Victorian hospitals. For this analysis, the presence or absence of MDM recommendation in a patient's health record was used as a measure of whether or not a MDM discussion occurred. Cancer Council Victoria's Human Research Ethics Committee provided ethics approval 1312 and 1412 for this study.

2.2. Patient Selection

The VCR was used to identify Victorian residents ≥ 18 years or older with a primary diagnosis of OG cancer (International Classification of Diseases—ICD‐10 C15, C16) between 2012 and 2021 [9]. Second incident tumours were excluded (N = 20), as were patients diagnosed via death certificate only (N = 171).

Comorbidities were extracted from diagnosis codes of admitted episodes in the year prior up to 1 month after a patient's cancer diagnosis date and classified using the Charlson comorbidity index (CCI) [10]. Comorbidity scores were grouped into three categories of increasing severity (0, 1, and 2+).

Morphology codes from the VCR were used to group oesophageal cancer patients into adenocarcinoma, squamous cell carcinoma, and other. Gastric cancer patients were not separated by morphology. As a proxy for cancer stage, patients were considered “metastatic at diagnosis” if pathology reports received by VCR indicated metastatic disease, or if any hospital admission within 4 months of diagnosis included ICD‐10‐AM metastatic disease or palliative care codes. No other staging data were available from the VCR.

Socio‐economic status (SES) was defined using the Index of Relative Socioeconomic Disadvantage (IRSD). IRSD scores by Statistical Area 1 (SA1; an area with average population size of 400) were obtained from the Australian Bureau of Statistics and assigned using the patient's residential address at the time of cancer diagnosis. Scores were grouped into quintiles (1‐most disadvantaged, 5‐least disadvantaged) [11].

2.3. Statistical Methods

Statistical methods were used to systematically identify factors indicating potential unwarranted variation. Patient demographics, tumour, and treatment characteristics were summarised using descriptive statistics, including counts and proportions for categorical variables and mean, standard deviation, median, and interquartile range (IQR) for continuous variables.

Age‐standardised incidence was calculated for each ICS region of residence for oesophageal and gastric cancer (2012–2021) by sex and year using the direct method [12]. Rates were standardised to the World Standard Population and expressed as diagnoses/100 000 population.

Survival at 1‐year and 5‐year post‐diagnosis was estimated for oesophageal and gastric patients using the Kaplan–Meier method [13]. A Cox proportional hazard model was used to estimate hazard ratios for the association between risk of death and ICS of residence, adjusted for age, sex, comorbidity score, and SES. Survival time was calculated from the date of diagnosis to the date of death from all causes.

All analyses were performed in R (version 4.3.2).

3. Results

3.1. Patient Profile

Table S1 presents the characteristics of the 9868 Victorians included in the dataset between 2012 and 2021 (oesophageal, N = 3696; gastric, N = 6172). For both periods, the median diagnosis age was around 70, with males more commonly affected than females in both cancer types. A quarter of Victorians with OG cancer were in the most disadvantaged SES quintile for both time periods. The proportion of Victorians who were current smokers or had a history of smoking for both cancers was similar for both time periods.

3.2. Cancer Profile

The age‐standardised incidence rate for oesophageal cancer remained the same between the two time periods, 3.5/100 000 cases (95% CI 3.4–3.7, 2012–2016; 95% CI 3.3–3.6, 2017–2021) but decreased from 6.1/100 000 cases (95% CI 5.9–6.4, 2012–2016) to 5.9/100 000 cases (95% CI 5.7–6.2, 2017–2021) for gastric cancer. Age‐standardised rates for mortality for oesophageal cancer also remained the same between the two time periods, 2.1/100 000 cases (95% CI 2.0–2.2, 2012–2016) and 2.0/100 000 cases (95% CI 2.0–2.2, 2017–2021) but decreased from 3.2/100 000 cases (95% CI 3.1–3.4, 2012–2016) to 3.0/100 000 cases (95% CI 2.9–3.2, 2017–2021) for gastric cancer.

For both cancer types, there was a decrease in the number of patients who had no comorbidities (CCI of zero) between the two time periods. In Victoria, 39.4% of OG patients (2012–2021) were metastatic at diagnosis (39.8%, oesophageal; 37.5%, gastric). The distribution of morphology groups was similar between time periods. For the most recent time period, 52.9% were diagnosed with adenocarcinoma compared to 38.4% with squamous cell carcinoma (Table S1).

For those with oesophageal or gastric cancer, 28% and 31%, respectively visited an emergency department within 28 days prior to diagnosis between 2017 and 2021.

3.3. Survival

Overall survival for oesophageal patients showed no statistical difference between ICS across the two time periods (p = 0.39, 2012–2016 and p = 0.13, 2017–2021, Figure S3). This was contrary to the overall survival for gastric cancer patients, which showed significant variation between the ICS across both time periods (p < 0.05, 2012–2016 and p < 0.01, 2017–2021, Figure S3). The statewide 5‐year survival (2017–2021) for oesophageal cancer was 23% (95% CI: 21–26) ranging from 18% in LMICS to 27% in NEMICS. For gastric cancer, five‐year survival (2017–2021) was higher at 32% (95% CI 30–34), ranging from 25% in LMICS to 35% in SMICS and HRICS. It should be noted, though, that any interpretation of HRICS data should be applied with caution, as there is missing data from the Albury Wodonga Health—Albury hospital campus, which is outside of the Victorian state border.

The hazard ratios obtained from the Cox proportional hazards for oesophageal and gastric cancer are presented in Figure 1. After adjusting for age, sex, comorbidities, metastatic disease, and SES, there was no significant variation between ICS compared to the Victorian average for both time periods for oesophageal cancer. However, for gastric cancer, two metropolitan ICS (SMICS and WCMICS) had statistically better survival than the Victorian average 2017–2021, and one regional ICS (LMICS) had statistically poorer survival compared to average (Figure 1).

FIGURE 1.

Overall survival for oesophageal and gastric cancer by ICS of residence, 2012–2016 and 2017–2021. Source: VCR (2012–2021), VAED (2012–2022).

3.4. Treatment for Non‐Metastatic and Metastatic Patients

Treatment modality (within 1 year of diagnosis) for Victorians with non‐metastatic oesophageal and gastric cancer, within 1 year of diagnosis, is depicted in Figure S4. After adjusting for age, sex, and comorbidities, when compared to the Victorian average, oesophageal cancer patients who resided in GRICS were more likely to receive chemotherapy and/or radiotherapy between 2017 and 2021. This was contrary to patients residing in NEMICS and GICS, who were less likely to receive radiotherapy for both time periods. Gastric cancer patients living in GICS were less likely to receive surgery or radiotherapy compared to the Victorian average between 2017 and 2021.

Of the patients diagnosed as non‐metastatic who had an oesophagectomy, 74.8% underwent neoadjuvant chemotherapy between 2012 and 2016, increasing to 78.5% in 2017–2021. The proportion of patients who underwent a gastrectomy and were treated with neoadjuvant chemotherapy increased slightly between the two time periods from 38% (2012–2016) to 41% (2017–2021). There was a significant increase in the proportion of patients who had an oesophagectomy and were treated with neoadjuvant chemoradiation from 28% (2012–2016) to 57% (2017–2021).

There was significant variation between ICS in the proportion of non‐metastatic oesophageal and gastric patients and time from diagnosis to any treatment (surgery, IV chemo, or radiotherapy) within 4 and 6 weeks (Figure 2). Overall, 29% (2012–2016) and 31% (2017–2021) of non‐metastatic oesophageal patients had treatment within 4 weeks and 58% (2012–2016) and 60% (2017–2021) within 6 weeks. One regional ICS, BSWRICS, was below the Victorian average for both time points (4 and 6 weeks) in both time periods (2012–2016 and 2017–2021). There were similarities for non‐metastatic gastric patients, which showed that overall, 38% (2012–2016) and 36% (2017–2021) of patients had treatment within 4 weeks and 64% (2012–2016) and 60% (2017–2021) within 6 weeks (Figure 2). Two regional ICS, BSWRICS and GRICS, were below the Victorian average for both time points (4 and 6 weeks) compared to metropolitan ICS such as NEMICS, which was above the Victorian average (Figure 2).

FIGURE 2.

Time from diagnosis to any treatment for non‐metastatic (a) oesophageal cancer (N = 1747) and (b) gastric cancer (N = 2739). Source: VCR (2012–2021), VAED (2012–2022).

There was variation between the ICS for the median time from commencing neoadjuvant chemotherapy or chemoradiation to surgery for non‐metastatic OG cancer patients. One metropolitan ICS (NEMICS) showed a decrease in the median time from commencing neoadjuvant chemotherapy between the two time periods ranging from 51 days (2012–2016) to 40 days (2017–2021). This was contrary to a regional ICS (LMICS) which showed an increase in the median time from commencing neoadjuvant chemotherapy between the two time periods, increasing from 49 days (2012–2016) to 61 days (2017–2021). Additionally, one metropolitan ICS (WCMICS) showed an increase for time from neoadjuvant chemoradiation to surgery between the two time periods, ranging from a median time of 55 days (2012–2016) increasing to 60 days (2017–2021). Overall, there was a shift from neoadjuvant chemotherapy to neoadjuvant chemoradiation (Table S2).

Approximately 43% of metastatic oesophageal patients received intravenous (IV) chemotherapy or radiotherapy within 1 year of diagnosis in 2012–2016. This was similar for IV chemotherapy utilisation in 2017–2021 (46.4%) however, increases were seen in radiotherapy usage (54.9%, p < 0.001). There was a lower proportion who received surgery, (2.8%, 2012–2016 and 1.3%, 2017–2021). The remaining 35.8% in 2012–2016 and 28.8% in 2017–2021 had no surgery, chemotherapy, or radiotherapy (p < 0.01).

For metastatic gastric patients, approximately 51% received IV chemotherapy within 1 year of diagnosis in both time periods. Radiotherapy usage showed a slight increase between the two time periods (26.8%, 2012–2016 and 31.2%, 2017–2021, p < 0.05). As with metastatic oesophageal cancer, there was also a lower proportion who received surgery, 11% (2012–2016) and 7.6% (2017–2021, p < 0.01). The remaining 34.5% in 2012–2016 and 32.7% in 2017–2021 had no surgery, chemotherapy, or radiotherapy.

From 2012 to 2021, 2337 OG cancer patients received no treatment within 1 year of diagnosis. Of these OG patients, a high proportion were aged over 70 years (70.9% non‐metastatic, 75.2% metastatic), had more than one comorbidity (44.7% non‐metastatic, 59.1% metastatic) or died within 1 year of diagnosis (44.5% non‐metastatic, 92.4% metastatic) (Figure S5).

3.5. Perioperative Outcomes

Median length of stay (LOS) following oesophagectomy was 15 days (IQR = 12–20 days) and 14 days (IQR = 11–20 days) respectively between 2012–2016 and 2017–2021, as shown in Figure 3. Median LOS following gastrectomy (total and partial) decreased from 12 days (IQR = 9–17 days for 2012–2016) to 10 days (IQR = 8–16 days for 2017–2021). There was statistically significant variation between the ICS in 2017–2021 for LOS following gastrectomy, compared to the statewide average (p < 0.001) (Figure 3).

FIGURE 3.

Length of stay (LOS) following oesophagectomy for patients diagnosed in 2012–2016 and 2017–2021 (a) and LOS following gastrectomy for patients diagnosed in 2012–2016 and 2017–2021 (b). Source: VCR (2012–2021), VAED (2012–2022).

Postoperative 30‐ and 90‐day mortality was similar between the two time periods for both oesophagectomy (0.71% (< 30 day) and 2.48% (< 90 day), 2012–2016; 0.81% (< 30 day) and 2.74% (< 90 day), 2017–2021) and gastrectomy (0.95% [< 30 day] and 1.91% [< 90 day], 2012–2016; and 0.74% [< 30 day] and 2.23% [< 90 day], 2017–2021). There was a significant decrease in postoperative one‐year mortality following gastrectomy from 13% (2012–2016) to 8.9% (2017–2021), p < 0.01. Postoperative one‐year mortality following oesophagectomy remained stable, (16%, 2012–2016; 15%, 2017–2021). After adjusting for comorbidity and age, there was a significant variation in postoperative one‐year mortality following gastrectomy between metropolitan and regional hospitals in 2017–2021.

3.6. Multi‐Disciplinary Meeting Presentation and Supportive Care Screening

The statewide average for documented OG MDM discussion in a central health record was 74% (N = 175) in 2017, increasing to 86% (N = 107) in 2022.

There was significant variation between the ICS in both time periods. Only one of the eight ICS (NEMICS) achieved the Victorian Department of Health's target of 85% in both time periods.

The CSPI audits indicated that documented evidence of supportive care screening for OG patients in health records stayed consistent at 53% between 2017 (N = 135) and 2022 (N = 107). This was well below the benchmark of 80% set by the DH.

3.7. Dietitian and Physiotherapy Contact

The statewide average proportions of surgical OG cancer patients who were seen by a dietitian within 3 months of diagnosis were approximately 60% for both time periods. For the proportion of OG chemotherapy patients who were seen by a dietitian within 3 months of diagnosis, there were no significant changes between the two time periods (55% in 2012–2016 and 57% in 2017–2021). Rates of dietitian involvement were low across all ICS and in both time periods.

The statewide average for surgical OG cancer patients who were seen by a physiotherapist within 3 months of diagnosis decreased from 45% (2012–2016) to 41% (2017–2021). The statewide average was even lower for OG chemotherapy patients who were seen by a physiotherapist within 3 months of diagnosis (approximately 28% across both time periods).

3.8. Palliative Care

For 2012–2016, the proportion of oesophageal cancer patients who were referred to inpatient palliative care at least three months prior to death was 16.1% (N = 528) compared to 15.4% (N = 1094) for gastric cancer patients. This observation was similar for 2017–2021 (16.3% [N = 824]; 13.3% [N = 1704]) for oesophageal and gastric cancer respectively. This was well below the DH benchmark of 80% [14].

Seventy‐two per cent of deaths from OG cancer occurred while in Victorian hospitals in 2012–2016 but decreased to 66% across 2017–2021. In Victoria, about 7.5% of OG cancer patients received chemotherapy in the last 30 days of life for both time periods (Figure S6).

4. Discussion

The effects of regionalisation in OG cancer care have been well‐documented internationally [15, 16]. While this study is set in Victoria, Australia, the implications extend beyond this jurisdiction. The findings reveal statewide variations in Victoria's adherence to key components of the OCP for OG cancer, including low survival rates, delays between diagnosis and treatment, and limited access to dietetics and physiotherapy. This study demonstrates the value of data linkage for identifying unwarranted variations but also highlights limitations in measuring detailed care aspects. It is the first population‐level study in Australia to examine adherence to the full continuum of the OCP for OG cancer, including public and private services, and multiple domains of care rather than focusing specifically on surgical outcomes at high‐volume centres [2]. By leveraging linked administrative and clinical data, this study provides a real‐world assessment of statewide variation in care delivery, including diagnostic delays, access to allied health, and survival outcomes. These elements are relevant to international audiences because they speak to broader challenges in implementing complex care pathways at a population level, regardless of geographic context.

Despite advancements in OG care, gastric cancer patients in regional areas face poor prognoses, consistent with studies showing lower oncology consultations, reduced chemotherapy access, and worse survival for those far from cancer centres [17]. Timely treatment is crucial, as delays are linked to poorer outcomes [18]. Regional variations in care and delays have been reported in Victoria, with differences in treatment timelines between ICS for non‐metastatic oesophageal and gastric patients [3, 19]. Nearly a third of patients visited an ED within 28 days before diagnosis, possibly due to unrecognised symptoms or delays in seeking primary care [20]. Variations may stem from differences in patient populations or regional clinical practices.

Other factors which may explain delays in OG cancer diagnosis and treatment include the impact of the COVID‐19 pandemic [21]. This was evident in an Australian study which demonstrated that the impact of COVID‐19 resulted in oesophageal cancer patients having the lowest compliance rate with the OCP of 4% (p < 0.001) [22]. Furthermore, oesophageal cancer patients commenced treatment in more than double the OCP recommended timeframe (< 28 days) [22]. The required infrastructure and systems are needed to facilitate compliance with the OCP and improve delays in diagnosis and time to treatment, which may also positively affect survival outcomes.

The shift from neoadjuvant chemotherapy to neoadjuvant chemoradiation is consistent with previous research [2]. The increase in the statewide average for non‐metastatic patients who had an oesophagectomy treated with neoadjuvant chemoradiation may be in response to results from the TOPGEAR (Trial Of Preoperative therapy for Gastric and Esophagogastric junction AdenocaRcinoma) randomised trial [23] which has demonstrated significant survival benefit. This is contrary to a more recent study that highlighted the superiority of neoadjuvant chemotherapy among patients with resectable oesophageal adenocarcinoma as compared with chemoradiotherapy [24].

To ensure consistency of decision making, it is widely accepted that all cancer cases are discussed at an MDM [25, 26]. This is reflected in the OCPs and Victorian cancer policy [6, 27]. The proportion of OG patients discussed at MDMs increased between the two time periods, with the latter (2022) meeting the statewide target of 85%. Despite this, the proportion of MDM discussions was lower in regional areas. Furthermore, documentation of key information from MDM discussions, including staging information, performance status, and communication with GP, was often missing. Internationally, it is known that rising cancer incidence, without a comparable increase in the health workforce and escalating case complexity, has increased pressure on MDMs [28]. Observational studies have shown that the average time for each patient discussion in an MDM is 2–3 min [28] and these time pressures in conjunction with increasing caseloads affect the quality of decision‐making [29, 30]. Strategies like treatment protocols, pre‐MDT meetings, and streamlining have been proposed by clinicians to improve the quality of MDM discussions [31].

Supportive care forms an integral component of care [32]. The OCP recommends ongoing screening for supportive care needs using a validated tool like the National Comprehensive Cancer Network Distress Thermometer and Problem Checklist [6]. Recent research highlights variations in the use of supportive care services for OG cancer patients in hospitals [33]. Strategies to reduce these disparities include early, consistent electronic symptom monitoring and validated screening tools for all patients, regardless of disease stage or treatment [34, 35]. In Australia, the PROPatient trial is evaluating nurse‐led electronic symptom monitoring for OG cancer [36], while international studies show telehealth programmes improve patient experience and self‐management [37]. Nurse coordinators have also been effective in addressing unmet supportive care needs through improved care navigation and referrals [35, 38].

Our findings show low rates of dietitian or physiotherapist interventions for OG cancer patients, despite malnutrition being linked to poorer outcomes [39]. This is likely due to underdiagnosis, lack of supportive care screening, and no consensus on the best nutrition screening tool for pre‐operative patients [40]. The absence of evidence‐based nutrition guidelines for OG cancer has led to inconsistent practices among health professionals [41, 42]. Developing a standardised nutrition care pathway focused on screening, monitoring body composition, and nutrition status throughout treatment is crucial for improving care [39].

Exercise‐based multidisciplinary rehabilitation has been demonstrated to address physical, emotional, and social needs while reducing isolation for OG cancer patients [43]. In Victoria, Australia, a prehabilitation programme provides physical, nutritional, psychological, and medical support to improve outcomes before and after surgery [44]. This approach has been shown to reduce hospital stays, postoperative pain, and complications [45].

Palliative care is vital for improving quality of life, yet many OG cancer patients, especially those elderly or with co‐morbidities, receive limited palliative support [33]. Referral rates to acute palliative care within 3 months of death remain low, particularly for regional patients, non‐English speakers, and those with metastatic disease who forgo treatment [33]. Oncologists and GPs play key roles in ensuring timely referrals to specialist palliative care [46]. Strengthening data systems, streamlining referrals [14], and integrating telehealth [47] and home‐based palliative care [48] can address variations and improve access, especially in rural and underserved communities.

5. Study Strengths and Limitations

This study has several strengths and limitations. This study's strengths include its population‐based data, large patient cohort, and coverage of the full cancer journey from diagnosis to end‐of‐life care, supported by linked administrative datasets. The involvement of a multidisciplinary OG expert advisory group to review all data and findings allowed for the prioritisation of unwarranted variation for targeted action for improving the quality of OG cancer care in Victoria.

However, limitations include gaps in the linked dataset, such as missing information on oral chemotherapy, Medicare and Pharmaceutical Benefits data, and community‐based palliative care. Additionally, no clinical trials data for OG cancer patients were available.

6. Conclusion

This is the first study to provide an overview of unwarranted variations in OG cancer compared to the OCPs in Victoria, Australia across two time periods. While the study shows that there were improvements between the two time periods in some areas, for example, mortality, unwarranted variation was identified in time between diagnosis and treatment, survival rates in some ICS, and access to supportive care including dietetics and physiotherapy. Furthermore, these findings offer important insights for surgical audiences regarding variations in referral, staging, treatment timing, multidisciplinary care, and access to preoperative support—components that are critical to improving surgical outcomes but are not often reported in routine surgical research.

Targeted actions to facilitate timelier access to diagnostic and treatment modalities represent a critical opportunity for improving outcomes and optimising care for OG cancer patients. In addition, ensuring all OG patients, despite their location, are presented to an MDM and receive care coordination will be important to ensuring optimal care provision. The provision of early and ongoing supportive care screening and access to palliative care will enable unmet needs to be addressed and promote referral to dietetics and physiotherapy. Continuing to analyse and compare data and outcomes at time intervals promotes the implementation of the OG cancer OCP and evaluates the impact of improvement initiatives such as OG cancer prehabilitation.

Author Contributions

Frances Graham: writing – original draft, conceptualization, methodology, formal analysis, writing – review and editing, project administration, validation. Wanyu Chu: methodology, writing – review and editing, data curation, conceptualization, validation. Helena Rodi: writing – review and editing, conceptualization, validation. Paul Cashin: validation, writing – review and editing, conceptualization. David S. Liu: conceptualization, validation, writing – review and editing. Norah Finn: writing – review and editing, methodology, data curation, validation. Tommy Hon Ting Wong: writing – review and editing, validation, supervision, data curation, methodology. Linda Nolte: funding acquisition, conceptualization, supervision, validation, writing – review and editing, resources.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1. Victorian oesophageal and gastric cancer patient characteristics in the linked dataset (2012–2021) (N = 9868).

Table S2. Time from treatment to surgery for non‐metastatic OG cancer patient by ICS of treatment (N = 1483).

Data S1.

Graham F., Chu W., Rodi H., et al., “Understanding Unwarranted Variation in Care for Victorian Oesophagogastric Cancer Patients Using Linked Data: A Comparative Analysis Between 2012–2016 and 2017–2021,” ANZ Journal of Surgery 95, no. 9 (2025): 1677–1686, 10.1111/ans.70196.