Treatment variation by Cancer Alliance

This section describes an analysis of variation in treatment between the 21 Cancer Alliances defined for England in 2022 (see Appendix 3).

Consistent with the previous Geographic Treatment Variation report, stage 1 tumours were excluded from these analyses as the management of early-stage tumours is consistent throughout the United Kingdom: 95.2% (n=1,001) of stage 1 tumours were treated with either primary surgery only or surgery with adjuvant chemotherapy, with minimal variation in treatment pathways expected between specialist gynaecological cancer centres or Cancer Alliances (Figure 12). This exclusion left a cohort for analysis of 4,762 tumours. Patient demographics and tumour characteristics for this analytical sample are provided in Table 8 of the of the data downloads section.

Findings are presented below as both funnel plots and results tables. Each point on a funnel plot represents a geographical area (in this case, a Cancer Alliance). The standard error is shown on the horizontal axis and provides an indication of the number of tumours diagnosed within the Cancer Alliance. Estimates from Cancer Alliances with a greater number of tumours are more precise, appearing further to the right-hand side of the plot and represented by bigger red markers than Cancer Alliances with fewer tumour diagnoses. The percentage difference in the probability of a treatment or treatment combination is shown on the vertical axis relative to the population average (all sampled tumours combined). A Cancer Alliance with an estimate above the average (indicated by a solid black horizonal line) suggests that tumours within the geography were more likely to receive treatment than the population average, with estimates below the line indicating a lower probability.

Two pairs of dashed lines are included on each funnel plot that represent the bounds of statistical confidence around the average value. The inner set of dashed lines represents two standard deviations (SD) from the population average and the outer set represents three SD, being approximately equivalent to 95.0% and 99.7% confidence intervals, respectively. Any observation plotted outside of these dashed lines will have a confidence interval that does not include the average value and may therefore indicate a systematic deviation in clinical practice that warrants further investigation. However, some random variation in the probability of treatment is expected between regions such that some points will sit outside the dashed lines through chance alone. This should be taken into consideration when interpreting funnel plots (for example, five out of every 100 observations are likely to lie outside the two SD funnel).

Within each accompanying table, Cancer Alliances highlighted in blue had treatment probabilities that were significantly higher (p=<0.05) than the average, and those highlighted in red had significantly lower probabilities. These represent Cancer Alliances that fall outside the innermost pair of dashed lines in the corresponding funnel plot (two SDs).

Given the variation in treatment observed across a range of patient demographics and tumour characteristics (Table 8), three different models were developed, each with an increasing level of adjustment for these factors:

• Model 1: adjusts treatment probabilities within each Cancer Alliance for differences between regions in the distribution of the patient age at diagnosis.
• Model 2: adds adjustment for differences in tumour morphology and tumour stage between Cancer Alliances.
• Model 3: further adjusts for area deprivation and Charlson comorbidity score.

While funnel plots are only shown for the age adjusted (Model 1) and maximally adjusted (Model 3) models, findings from all three models are presented in a single table to allow comparisons according to differing levels of covariate adjustment. Note that these three models adjusted for confounding factors in a different order from that applied within the preceding report. For instance, Model 1 was formerly unadjusted rather than age-adjusted.

This first analysis looks at differences between Cancer Alliances in the proportions of tumours that received any treatment, defined here as surgery or chemotherapy either alone or in combination.

The weighted average probability of a stage 2-4 and unknown stage ovarian cancer having received any treatment during the first nine months following diagnosis was 73.7%, almost identical to the rate reported in the Treatment report (73.8%) for the preceding three years - Table 9 of the of the data downloads section.

Figure 16 Geographic variation in the probability of any treatment versus no treatment, excluding stage 1 disease, 2019

Figure 16 shows that geographic variation in treatment remained following adjustment for a range of factors associated with differences in the treatment pathway. The probability of any treatment fell more than two SDs below the population average for one Cancer Alliance, while two Alliances had probabilities more than two SDs greater than the average. These results indicate variation in overall treatment between Cancer Alliances in England to a degree that may not be explained by chance alone and warrants further investigation. Notwithstanding, comparison with the Geographic Treatment Variation report suggests that regional variation in access to any treatment was less marked in 2019 than in the previous three years. In the Treatment report maximally adjusted analysis of 2016-2018 diagnoses there were four Cancer Alliances with statistically significant low rates of ‘any treatment’, sitting below the -2SD line on the funnel plot. For 2019 data, only the Greater Manchester Cancer Alliance remained a low statistical outlier for this parameter. The treatment report also demonstrated 5 Cancer Alliances with 2016-2018 ‘any treatment’ rates more than 2SD above the national average. Comparison of data is complicated by adjustments of Cancer Alliance configuration in London, but for 2019 diagnoses just two Alliances (North West London and North East London) were demonstrated as high performing outliers on the maximally adjusted analysis. A complete table of coefficients is reported in Table 10 of the of the data downloads section.

The optimal management of ovarian cancer requires surgery in the large majority of cases to remove (debulk) the tumour. While the results above indicate variation in the receipt of any treatment, the analysis below looks specifically at the probability of surgery either alone or in combination with chemotherapy.

The average probability of stage 2-4 or unknown stage ovarian cancers being treated with surgery was 51.8% (Table 11 of the of the data downloads section). similar to the rate for diagnoses in 2016-2018 as reported in the Geographic Treatment Variation report (51.0%). The funnel plots in Figure 17 show large geographic variation in the delivery of surgery between Cancer Alliances, even after adjustment for other factors. Six Cancer Alliances had surgery probabilities two SDs above the average and four had probabilities two SDs below the average. Comparison with 2016-2018 data suggests that the marked regional variation demonstrated in the Treatment report remains equally profound in relation to cases diagnosed in 2019. All four Cancer Alliances with low 2019 surgery rates (South Yorkshire and Bassetlaw, East Midlands, West Midlands and Wessex) also had significantly low surgery rates for cases diagnosed 2016-2018 in the Treatment report. Whilst Cancer Alliance reconfigurations limits direct comparison between 2016-2018 data and 2019 data for some regions, it appears that surgery rates in London alliances, the North East and Surrey & Sussex have remained high in both analysis periods. A full table of coefficients is provided in Table 12 of the of the data downloads section.

Figure 17 Geographic variation in the probability of surgery versus no surgery, excluding stage 1 disease, 2019

The probability of treatment with surgery was particularly low in one Cancer Alliance and warrants investigation to determine whether the figure represents true regional variation in clinical practice or is the product of either poor treatment reporting or unadjusted differences in the type of tumour or patient being treated.

The majority of women with an ovarian cancer diagnosis should receive chemotherapy to treat and help manage the disease. Exceptions to this include some histological types and grades of stage 1 disease, where surgery alone may be adequate to provide a high chance of disease cure. Additionally, there are circumstances where risks from chemotherapy may outweigh the benefits, and where patients may decline chemotherapy treatments.

Of stage 2-4 and unknown stage ovarian cancers, the average probability of treatment with chemotherapy was 64.1%, compared to 66.5% demonstrated in the Treatment table for 2016-2018 diagnoses (Table 13 of the of the data downloads section).

Figure 18 Geographic variation in the probability of chemotherapy versus no chemotherapy, excluding stage 1 disease, 2019

Following adjustment for a broad range of factors, including patient age and tumour morphology, Figure 18 shows three Cancer Alliances with chemotherapy rates two SDs above the average and two with chemotherapy rates two SDs below the average. As previously reported for 2016 to 2018 diagnoses, there was markedly less regional variation in chemotherapy treatment rates throughout England compared to surgery rates. There were no Cancer Alliances with statistically significant low chemotherapy rates in both the original treatment report (diagnoses 2016 to 2018) and the current analysis (2019 diagnoses), with little evidence of consistently high chemotherapy treatment rates across the two reports except for the West Midlands Alliance and Alliances in the north of London. A full table of coefficients is provided in Table 14 of the of the data downloads section.

Chemotherapy is increasingly used prior to surgery (neoadjuvant). This approach is often applied in two circumstances specific to advanced disease. Firstly, if a patient is very unwell at the time of diagnosis, chemotherapy can start to treat the cancer and support improvement in the overall health and performance status of the patient before undergoing surgery. Secondly, if the multidisciplinary team (MDT) considers it unlikely that complete tumour resection (removal) will be feasible during primary surgery, neoadjuvant chemotherapy may be used to make the tumour more operable, reducing the risk of surgical complications and morbidity. ¹

The following analysis explores geographic variation in the probability of receiving primary surgery with adjuvant chemotherapy versus neoadjuvant chemotherapy with interval debulking surgery. Accordingly, the analysis was restricted to the 2,010 tumours within the cohort of stage 2-4 and unknown stage cancers that received one of these two treatment combinations.

Within this subsample, the probability of primary surgery and adjuvant chemotherapy was 49.6% on average (Table 15 of the data downloads section), compared to the rate of 49.4% for cases diagnosed 2016-2018, as described in the Geographic Treatment Variation report.

Figure 19 Geographic variation in the probability of primary surgery and adjuvant chemotherapy versus neoadjuvant chemotherapy with interval debulking surgery, excluding stage 1 disease, 2019

Figure 19 shows regional variation following adjustment for patient demographics and tumour characteristics associated with treatment, with three Cancer Alliances falling two SD above the average and one region falling two SDs below the average. The maximally adjusted probability (Model 3) was markedly higher in one Cancer Alliance than others. The former Cancer Alliance North East and Cumbria had primary surgery and adjuvant chemotherapy proportion that was 20.4 percentage points higher than the national average for the 2016-2018 diagnoses described in the Geographic Treatment Variation report, while the newly defined Northern Cancer Alliance remained the most marked outlier for diagnoses in 2019 (20 percentage points higher). Generally, there was markedly less variation in the proportion of cases having upfront surgery versus neoadjuvant chemotherapy compared to the extensive variation in actual surgical resection rates. A full table of coefficients from the underlying models can be viewed in Table 16 of the of the data downloads section.

Note:

1. Vergote I and others. Neoadjuvant chemotherapy versus debulking surgery in advanced tubo-ovarian cancers: pooled analysis of individual patient data from the EORTC 55971 and CHORUS trials. Lancet Oncology. 2018;19(12):1680–1687.

Table 17 of the data downloads section (also reproduced below) pulls together the maximally adjusted results from Table 9 (any treatment), Table 11 (any surgery) and Table 13 (any chemotherapy). Collectively, they indicate geographic variation in treatment delivery after accounting for differences in the regional distribution of various patient demographics and tumour characteristics.

Table 17 Summary of maximally-adjusted geographic variations in any treatment, surgery and chemotherapy (Model 3; diagnosed 2019)

To supplement these 2019 geographic variation in treatment data, Table 18 reports the variation in treatment over time for England as a whole. Specifically, the proportions of patients selected into each treatment category, stratified by year of diagnosis (2015 and 2019). This analysis suggests stable patterns and rates of treatment across England between 2015 and 2019. Very little variation in treatment was observed over time at a country level beyond a small fall in the proportion of patients who received any chemotherapy (2015: 65.2%; 2019: 64.1%).

To assist comparison against the refreshed age-standardised survival statistics presented earlier in this report for cases diagnosed 2015-2019, age-adjusted geographic variation treatment is presented in Table 19 of the data downloads section (also reproduced below) for tumours diagnosed during the same five-year period. This supplementary analysis excludes borderline and D391 cases, consistent with the survival cohort for the period.

In turn, Table 20 of the data downloads section presents the patient demographics for the patients in Table 19. It should be noted that the count of patients in the treatment tables (28,478 tumours for 28,351 patients) is slightly larger than presented in the survival tables for the period (28,189) patients. The reasons for this discrepancy are as follows:

• The survival method excludes patients with vital status data quality issues. Vital status was not a consideration for treatment variation analyses, and so this exclusion criterion was not applied.
• The survival method considers ovarian cancer patients aged 15-99 years at diagnosis, while the treatment analysis was restricted to patients aged ≥18 years at the time of diagnosis.
• Both sets of analyses selected on the earliest tumour documented per patient. However, while the treatment analysis applied this rule exclusively to the period 2015-2019, the survival method applied a rolling five-year analysis from 2001-2019. Thus, for patients with multiple diagnoses that were dated either side of 01/01/2015, there will be a discrepancy in how each analysis flagged the earliest tumour.
• The survival method selected on tumours documented with a behaviour code of 3 (i.e., confirmed as primary, invasive, and malignant), while the treatment analysis applied no such condition when selecting on tumours.
• The survival method excluded patients with missing or imputed gender, date of diagnosis, date of birth or age information.
• The survival method excludes ovarian cancers documented with a morphology of 8320, which is not a valid entry for ovarian cancer.