Excess Treatment Costs (ETC): Estimating the magnitude and distribution of ETC in England to inform policy formulation

Background: There is anecdotal evidence that the payment of excess treatment costs (ETC), i.e. the treatment costs incurred in clinical studies that are in excess of standard care, are a point of friction between providers and commissioners and that many stakeholders feel that barriers to gaining support for these can increase research costs, delay commencement and completion of clinical studies, and alienate trusts and patients from participation in these. There is currently no single source that can be used to obtain an overview of the magnitude or distribution of ETC incurred in clinical practice in England. This document describes the results of a ‘fact finding exercise’ in this area jointly commissioned by the DH, NHS England and PH England.

Objective: The primary objective was a) to determine the overall value (£) and the distribution of ETC and Excess Treatment Savings, and determine the variability of the distribution over (ideally) 4 years in England, subgrouped by commissioner type, and geographical area, and b) to establish the proportion of non-commercial studies that incur ETC and ETS, and determine if there is an association between the total ETC incurred by a study and the total research funding awarded. Secondary objectives include: get a quantitative sense for the unrest/’noise’ in the system and provide a benchmark for evaluation any potential interventions/solutions, and identify and evaluate the different models/processes used to obtain ETC funding, their key components and in what context they would be transferable to other regions.

Method: Several online surveys and personalised data requests were designed for a broad range of stakeholders including: funders of non-commercial clinical studies (e.g. NIHR, UK charities), commissioners (i.e. Clinical commissioning groups (CCG), Specialised Commissioning, DH Subvention, Public Health England, Local Authorities) and providers (NHS Trusts). These requested both summary and detailed information of clinical studies funded and ETC incurred in clinical practice over the previous four years. Open-text questions relating to local processes used to handle ETC were requested from both CCGs and providers. These were designed to explore the primary objective and to support the secondary objectives. Evidence from the funders provide information on the ETC and ETS that are estimated at the point of grant application (terminology used: FETC and FETS) and due to a multiple of reasons (e.g. differences in estimated and actual recruitment, differences in standard care across individual provider sites, delays in study commencement, or the time horizon of studies etc), the FETC and FETS may not match the true ETC and ETS values that are incurred in clinical practice.

A series of semi-structured interviews (N=32) were conducted. Interviewees were selected to represent a broad geographical, and stakeholder perspective. The interviews were transcribed and analysed using NVivo11 using framework analysis.

Results: Evidence from funders indicates total FETC range from £10.9m (2015/16) to £32.5m (2016/17) while total FETS range from £9.9m (2013/14) to £47.6m (2016/17). The proportion of non-commercial funded studies that may incur ETC range from 23% (2013/14) to 55% (2015/16) for studies funded by NIHR and between 43% (2015/16) and 72% (2016/17) for studies funded by charities. The proportion of studies with FETS funded by NIHR is more constant at approximately 6% but varies between 3% (1/34) in 2013/14 and 10% (4/40) in 2014/15 for charities. Based on individual study level data provided by NIHR, there is a positive relationship (rho=0.49) between the total funding allocated and the FETC incurred. These data do not represent the total number of clinical studies awarded funding over the four year period due to low response rates to invitations to take part in the exercise. For example, only 3/6 of the funders provided returns and only 2 UK charities provided data. It is not possible to extrapolate from these data to calculate an indicative value as we do not have any information that could be used to determine what proportion was not captured.

Evidence from all commissioners and providers that responded to the invitation to take part in the exercise suggests that the total ETC funded in clinical practice range from £1.9m (2016/17) to £3.1m (2015/16). However, it is believed that these values are a gross underestimation of the likely true value of ETC incurred in clinical practice for several reasons. Data from Specialised Commissioning was not available, and this could constitute a non-negligible proportion of total ETC incurred. The numbers and robustness of returns from CCGs and from Trusts was extremely poor. Only 98/199 CCGs provided at least one annual value for ETC funded and many of the returns were clearly ballpark estimates. The response rate for Trusts was even lower with just 25/221 values for ETC absorbed by the organisation provided for 2016/17 (less returns in earlier years). Many of the supporting comments from Trusts who submitted blank returns clearly indicated that they did actually absorb the costs associated with ETC internally using funds such as department budgets, investigator research funds, or other ‘commercial income’. The vast majority of respondents for CCGs and Trusts who did not provide the information requested reported that these data were not routinely collected.

Notwithstanding the fact that FETC will not match ETC exactly, the total values for ETC (£1.9m in 2016/17) collected from the commissioners and providers do not have any face validity when compared to the incomplete FETC data (£32.5m in 2016/17) provided funders of clinical studies. They also lack face validity when compared to the total ETC funded in Wales where ETC are reported to be £2.1m for a patient population of 3.1m in 2016/17. Extrapolating the total ETC reported by CCGs and Trusts using the patient population covered by the returns to the population of England gives an estimate of £7.1m for a population of 55.3m. This is not comparable to the Welsh data even when adding the values provided by the DH Subvention (£0.4m in 2016/17) and PHE (£98.9k in 2016/17).

The returns from CCGs and Trusts were not considered sufficiently robust to determine a geographical distribution of ETC, and the other commissioners could not allocate the total amounts to geographical areas within the timescales of the project. Likewise, it was not possible to determine a distribution by commissioner type (primary care versus secondary care, Specialised Commissioning versus local commissioners) due to the incompleteness of the returns and the lack of alternative information.

The annual ETC returns from CCGs and providers were both positively skewed with a small proportion of the organisations funding or absorbing relatively large amounts respectively. Half of the total ETC reported to be funded by CCGs who provided returns in 2016/17 was funded by just 3 CCGs. The data from Trusts was even more skewed with 76% (19/15) of CCGs reporting annual values of less than £20k and three reporting values between £85k and £323k in 2016/17. Looking at the annual returns for the individual CCGs who reported at least annual 3 returns (N=19), if a budget was set at 50% higher than personal average historical value, 78% of CCGs would have at least one return that was higher than the budget. The data from Trusts also fluctuated over time for the individual Trusts, and setting a budget of 50% greater than the personal historical average, 71% of Trusts would have at least one return greater than the budget.

The returns for ETS absorbed by Trusts or passed on to CCGs was even poorer than for ETC. The values provided by the Trusts were negligible and no CCGs provided a value. Again the lack of evidence was reported to be because the data are not currently recorded within the organisations (for Trusts) or they never have any (both Trusts and CCGs).