The cost of oropharyngeal cancer in England: A retrospective hospital data analysis

To estimate the total costs of treating head and neck cancers, specifically oropharyngeal, laryngeal and oral cavity cancer, in secondary care facilities in England during the period 2006/2007 to 2010/2011.


| Ethical considerations
Approval for use of Hospital Episode Statistics (HES) data was granted by the Health and Social Care Information Centre. The individual HES records extracted contained no sensitive data and were pseudonymised, preventing the true identification of patients. Analyses pertaining to HES records adhered to published regulations. 15 The Health Research Authority decision tool stated no ethical approval was required for this research. 16 2.2 | Participants, setting and study design  Table 1.

| Data analysis and statistical methods
A large proportion of payments made to secondary healthcare providers contracted by commissioners in England are governed by the Payment by Results (PbR) framework, 17  Some key therapies in head and neck cancers do not have a National Tariff due to wide regional variation in costs and practice.
These therapies include (amongst others) chemotherapy, radiotherapy, palliative care and rehabilitation. Payments for these therapies are the product of local negotiations between commissioners and providers. For these costs, the relevant HRGs from HES were crossreferenced with the National Reference Costs and inflated using the pay cost index. 18 It was not possible to represent regional variation in these particular costs and for the purposes of presentation they were disaggregated from the core HRG to compartmentalise this uncertainty. The costs associated with the spell of care included all other costs associated with the initial diagnosis (if encountered in a hospital), surgical procedures and medical treatments.
Outpatient costs were estimated by grouping consultations by treatment specialty based on Treatment Function Codes (TFCs) and whether the consultation was the first of a series or a follow-up. As with the inpatient data, all activities for which reimbursement rates are locally negotiated, such as outpatient chemotherapy and radiotherapy sessions, were disaggregated from the core HRG.
To confirm whether there is evidence that costs have not increased uniformly with patient numbers, that is higher costs for newly diagnosed oropharyngeal patients, a difference in differences (DID) approach was utilised, with the t-test used to assess whether the means of the two groups were statistically different from each other. Oropharyngeal patients were taken to be the "treated" group, year and that seen in 2009/10 (the latest year for which complete data were available) was then estimated using ordinary least squares. respectively, over the course of the study, with inpatient numbers also increasing, by 372 (13%) and 502 (15%). Table 3 presents the mean values for the variables used in the DID analysis of the aggregate inpatient data. The results show there were no significant differences in the baseline period between the oropharyngeal and combined oral cavity and laryngeal groups in terms of total costs (P = .93) and total patients (P = .903). The results from the DID analysis can be found in Table 4. Model 1 included a dummy variable for the male cohort, on which the coefficient was found to be positive and a strong predictor of higher costs (P < .01), and showed some indication of differences between the two groups (results not shown). In Model 2, the natural logarithm of total patient numbers was also included and found to be a significant predictor of higher total costs (P < .01) along with being male (P = .059), while also providing some evidence of a real difference between the two groups in the follow-up period (P = .053) (results not shown). However, there was no statistically significant evidence of a real change in differences between total costs for oropharyngeal and laryngeal/oral cavity cancer in either oropharyngeal cancer did not increase at a faster rate than for other head and neck cancer sites, we did observe a rapid increase in the number of patients being treated for this cancer (1420 more patients admitted to hospital in the final year of the study compared to the first, of which more than 1000 were male). This is consistent with data published elsewhere. 13 Our data also help highlight that head and neck cancers tend to disproportionately impact the male population who accounted for almost 75% of the total economic burden.

| Strengths and limitations of the study
This research is the first to quantify the cost of treating head and neck cancers in England over time. The results provide important insights into how the cost burden for these cancers has changed over recent years, particularly for inpatient care where the data are most comprehensive.
The Audit Commission has stated that in 2012/13, only 4.1% of all outpatient attendances had a known primary diagnosis. 20 As this was our primary means of extracting information on patients with head and neck cancers, the figures for outpatient activity will not cover all consultations relating to the cancers under investigation and may be underestimated. Underestimation of inpatient costs is also likely to a lesser extent.
The HES database is heavily reliant on correct clinical coding and, as a result, data gaps have arisen. Although it was possible to assess a patient as having multiple hospital admissions from the data, it was not possible to distinguish between an initial or recurrent patient. The data did also not permit analysis of disease stage at presentation, which may in turn influence resource use.  The use of National Tariff precludes analysis of regional variation and contract negotiation and underestimates radiotherapy costs.
High-cost drugs are excluded, and no indication of hospital prescribing data are available within HES. The associated costs could therefore not be included in this analysis. Definitions for HRGs may alter year upon year and inaccurately matched codes and definitions could have potentially led to under-or overestimations.
There has doubtless been an increased attention on the increasing incidence of oropharyngeal cancer, and parallel interest in the role of HPV as a major aetiological factor. As a consequence of this, it may be that the accuracy of coding, for example, site accuracy between oral cavity and oropharynx, may have increased over time but the effect of this on the relative incidence of oropharyngeal cancer is not quantifiable. Together

| Clinical applicability of findings
The costing estimates presented here highlight that the cost burden of head and neck cancers is growing in England. The increasing pressure on NHS resources requires strategic investment for both treatment and prevention of these diseases. The observation that in 4 years, the number of patients with oropharyngeal cancer receiving some form of inpatient care has increased by more than half and the associated costs have increased by over three quarters reinforces the argument for prevention and early detection strategies to help contain these cost and health burdens.
The results of the present analyses suggest that the use of chemotherapy and radiotherapy in patients with oropharyngeal cancer has increased in recent years. It is not possible from the current data to determine why this is but greater use of chemotherapy and radiotherapy may have significant impacts on workforce and resource planning in head and neck oncology centres. It also highlights a need for better evidence from clinical trials in how to best manage this oropharyngeal cancer epidemic.