SUMMARY OF KEY POINTS AND RECOMMENDATIONS

—  AstraZeneca supports the objectives of NICE. There are core aspects of NICE that are fundamental to its working which should not be lost. For example, the commitment to raising standards of healthcare; the focus on areas of government health priority; the engagement with stakeholders; the overall transparency of process; the respect for commercial-in-confidence information, and; the ability to appeal decisions.

—  However, we believe there are areas for improvement in the way that NICE operates. These areas have undermined private and public confidence in NICE and must be addressed before any expansion in the role and remit of NICE (such as that proposed by Cooksey/recent OFT report) is to be countenanced.

—  This submission highlights the following shortcomings:

—  Over-reliance on cost per QALY as the tool for making rather than informing decisions.

—  Expectations of certainty in decision-making that cannot be achieved, particularly early in a product's lifecycle.

—  Quality and consistency in the work commissioned from Evidence Review Groups (ERGs), including important transparency issues.

—  These shortcomings underlie an increasing trend for companies to resort to an appeal process that is itself seriously flawed, and which has resulted in the past two years in no significant changes to any NICE guidance.

—  We believe these shortcomings are leading to a situation where English and Welsh patients may be permanently disadvantaged in their ability to benefit from absolute and incremental medical innovation, available as standard care in other countries, because the bar to positive recommendation is becoming unrealistically high.

—  NICE has recently started helping the NHS to implement its guidance. Many reports have indicated that implementation remains inconsistent and patchy within the NHS even after these focussed attempts. Reasons for inconsistent implementation include poor local financial planning, clinical resistance and ineffective sanctions.

WHY NICE'S DECISIONS ARE INCREASINGLY BEING CHALLENGED

NICE Decision-Making—Cost per QALY

  2.  AstraZeneca adds its support to the ABPI view that over-reliance on the cost per QALY within decision-making within NICE is detrimental to the quality of NICE decision-making overall. Whilst the use of cost per QALY cost-effectiveness arguments are key to informing decisions regarding a technology, they are not the only component of a product's value. For example the following items comprise a (non-exhaustive) list of other considerations that can inform on the value of a product:

—  Is the condition severe or life-threatening.

—  Are there any other direct treatment alternatives.

—  Does the medicine confer benefits/savings beyond those impacting health and social services, eg social security, carer benefits, ability to return to work.

—  Factors that are identified by patients/carers to be of most value.

—  Is the patient of an age where benefits may be wider than that conferred to the individual concerned.

—  Does the medicine reverse rather than stabilise the condition.

—  Timescale within which the benefits are anticipated.

—  Overall budget impact (or lack thereof).

—  Issues of health "equalisation" and equity.

  We believe that these "other considerations" are overshadowed in the NICE decision-making process by the single "cost per QALY" value.

  3.  There are a number of limitations to the application of the QALY which have been fully addressed in the ABPI submission. However, it is useful to highlight the following points:

—  Cost-effectiveness is a fairly new area of science which provides a single point estimate together with a range of values based on "likely assumptions". Modelling results in wide-ranging estimates of cost-effectiveness depending on the methodology and assumptions used. As such, quite often the most important question is not, "What is the cost per QALY?" but more, "How certain are we that that cost per QALY value provided is true?" With the complex nature of modelling this subtlety can be lost in interpretation in favour of a single point estimate.

—  The measurement of the quality of life component of the "QALY" is subject to a high degree of subjectivity and with a range of differing methodological approaches, the values garnered are often wide-ranging and hence contain a great deal of uncertainty. Additionally the generic approach to the measurement of quality of life, as required by NICE in the EQ-5D instrument, is relatively insensitive to incremental changes in quality of life and does not fully reflect the range of benefits that are important to patients.

  4.  In practice, these limitations mean that it is generally more difficult for medicines which deliver quality rather than quantity of life to demonstrate a large enough incremental benefit to be deemed cost-effective, eg in treatments at the end of life, for the elderly, and for long-term conditions, where apparently small quality-of-life benefits can make a significant difference to patients.

  5.  The perspective of cost-effectiveness taken by NICE is limited to that of the NHS and Personal Social Services; this excludes important benefits such as the impact on carers or enabling people to return to work and contribute to society.

  6.  Patients with rare diseases treated with orphan medicines[40] are disadvantaged by the NICE process. Orphan medicines are apparently expensive because development costs are relatively high on a "per-treated-patient" basis and an adequate return is necessary to sustain research. The combination of limited data at launch and relatively high per-patient prices means it is challenging for orphan medicines to meet NICE cost/QALY criteria. The current methodology requires re-evaluation, with distinct decision rules focusing more on unmet need, innovation, clinical effectiveness and budget impact, and less on cost/QALY, to ensure that the NHS principle of equal access based on clinical need is maintained for patients with rare diseases.

Consistency and Quality of Assessment

  7.  In line with the ABPI submission, AstraZeneca has concerns with the apparent consistency in approach and quality of the assessments carried out by the independent evidence review groups (ERGs).

  8.  A key issue is how the ERGs handle uncertainty, which is greatest early in a product's lifecycle. Given that effectiveness and cost-effectiveness data are derived from highly structured clinical trials and economic models based on a wide range of assumptions and inputs, it is inevitable that different conclusions will be reached by different assessors. It is how these uncertainties are managed and the transparency with which they are addressed that are important. Rather than seeking to constructively outline and address them, some ERGs present them as fatal flaws in the manufacturer's case.

  9.  Within the STA (Single Technology Appraisal) process there is an obvious omission in quality assurance which places it at odds with the MTA (Multiple Technology Appraisal) process. Within the STA process, no opportunity is afforded to stakeholders to comment on the technical accuracy of the report before the Appraisal Committee uses it to make recommendations. This technical accuracy aspect is particularly important in an STA as the ERG report is based solely on the company's submission. Factual errors that are not rectified are translated into draft and final guidance leaving appeal as the only route to make corrections.

  10.  The situation is exacerbated by the inability to challenge the methods and models developed by the ERGs. Whilst the ERGs have access to the full working models produced by manufacturers, manufacturers frequently do not have access to the models created by the ERGs, the results from which form the primary evidence upon which the Appraisal Committee bases its decisions. If the model is made available, it is "locked", which significantly hinders the ability of manufacturers to understand how it works and ultimately to challenge the working of the ERGs.

  11.  AstraZeneca support the ABPI's wish for a more collaborative approach with the ERGs, based on constructive dialogue between the ERG, the NICE Executive and the company from the outset of the appraisal, allowing discussion and debate on methods, data quality and sources used. This would reduce inaccuracies, address unnecessary misunderstandings, speed up the process and reduce the number of appeals.

Workings of the Appraisal Committee

  12.  It is not clear how relevant clinical expertise, particularly from those who have hands-on experience of managing patients with the condition, is used in Appraisal Committee decision-making.

  13.  The discussion in the Appraisal Committee is almost solely focused on the economic case, presented by an economist, with the over-riding consideration being the cost/QALY. As stated previously, we believe there should be much greater balance between this and other considerations in the discussion and subsequent decision-making.

  14.  NICE has consistently rejected the industry view (particularly for STAs) that the Appraisal Committee should give proper consideration to the lack of maturity of the evidence for products early in their lifecycle. Unreasonable expectations of clinical and economic certainty for these products then result. This is dangerous for two reasons: (1) patients are being denied access to important treatment advances; and (2) denying access so early in the life of the product means that learning and incremental innovation through normal clinical experience will not happen.

  15.  For example, anastrozole is an anti-oestrogen used to prevent breast cancer recurrence. It was first launched in 1995 and subsequently use has evolved through 2nd line for advanced breast cancer (ABC), 1st line ABC, tamoxifen-intolerant patients with early breast cancer (EBC) (2002), to 1st line EBC (2005). The clinical benefits are different in each of these settings, and therefore cost-effectiveness will vary correspondingly.

  16.  The landmark ATAC study published in 2005 showed that after five years' treatment, anastrozole increased DFS (disease free survival) compared with tamoxifen, in women with EBC. DFS is used to model longer-term benefits (overall survival, life years gained and QALYs). This study of over 9,000 women commenced recruitment in 1996 and it will be several years before we are likely to see a statistically significant difference in survival (it took seven years to show a survival advantage of tamoxifen over placebo).

  17.  These challenges will be common to any cancer treatment where the prognosis is good, and survival rates high. The two most common cancers (breast in women, prostate in men) fall into this category.

  Therefore:

—  Estimating value (cost per QALY) at launch from surrogate markers will always be subject to a high level of uncertainty.

—  Outcomes data cannot be generated until many years after launch.

—  The value of drugs like anastrozole is likely to evolve with the patient population treated.

—  A strict adherence to purely cost per QALY messages at launch may discourage the NHS from ever investing in products for chronic conditions. This will place UK patients at a disadvantage compared with their counterparts in other countries.

  18.  If an appraisal of anastrozole had been undertaken at launch, it is unlikely that it would have been found to be cost-effective due to the high level of uncertainty inherent in the data available at launch. A decision into its continued use on the NHS at this time could have lead to access to a valuable product being blocked for UK patients whilst its usage became standard of care in other countries.

  19.  In the case of MTAs, the assumption in "class" appraisals is that all products are essentially similar. This is done to enable cost comparison, but fails to take into account that some products may have more robust evidence than others. This opposes NICE's fundamental principle of evidence-based assessment and can lead to patients being denied access to a technology which has the best evidence, merely because it is more expensive. This approach is anti-innovative because the innovator entrant demonstrates clinical and cost-effectiveness only to have their evidence translated across the class, negating their advantage on pricing alone. Longer term, this is likely to dis-incentivise UK investment in research into new treatments.

APPEAL PROCESS

  20.  AstraZeneca supports the ABPI's view on the current Appeals process with NICE. The main concerns include lack of independence, inconsistency of approach at appeal hearings, absence of health economic expertise on the Appeal Panel, and stakeholders' ability to understand and challenge new evidence disclosed for the first time in the FAD. We believe the more pragmatic system adopted by the SMC and AWMSG (which allows for appeals on "scientific grounds" as opposed to the more legalistic criteria employed by NICE) should be utilised for NICE.

IMPLEMENTATION OF NICE GUIDANCE

  21.  NHS implementation of NICE guidance remains slow and patchy, in spite of considerable efforts by the NICE Implementation Team for the past four years to improving it, and the inclusion of implementation in NHS core and developmental standards, against which local NHS organisations are assessed by the HCC. A number of factors contribute to this, the most important being poor NHS financial management, clinical resistance, and lack of sanctions against poor implementation.

  22.  Implementation of NICE guidance is not a HealthCare Commission priority in its Annual Health Check. Assessment takes the form of ensuring that processes are in place rather than measuring tangible evidence of implementation. Indeed implementation of NICE Clinical Guidelines remains merely a best practice standard to aspire to rather than a requirement for the NHS. AstraZeneca believes that implementation of guidance should involve rigorous HCC measurement and inspection together with joined-up financial incentives/penalties for lack of implementation. A system for improving this implementation could involve inclusion of an implementation component in the QOF or automatic inclusion of NICE approved medicines on local formularies. (We are aware of the potential difficulties for PCT funding flows that may be attached to this latter recommendation. We recommend further advanced planning for PCTs, with NICE providing suitable warning to the NHS regarding ongoing appraisals for the year and hence potential budget implications.)

REVIEWS OF TECHNOLOGY APPRAISALS (TAS)

  23.  TAs are accompanied by mandatory funding within three months of guidance being published. There is no such mandatory funding for Clinical Guidelines. There has been a recent trend for reviews of TAs to be carried out "within the context of a clinical guideline", with express information from NICE that mandatory funding will be removed, even if the guideline continues to recommend use.

  24.  The reason given by NICE is that the average length of time for a TA to be reviewed (about three years) should provide sufficient time for the NHS to embed its use into clinical practice. However, as we discuss above, implementation is patchy and slow, and three years is simply not sufficient for use of a (new) intervention to have become routine. An example of this situation is the atypical antipsychotics for schizophrenia. A clinical guideline is currently underway which will update the existing appraisal which recommended use of these products within schizophrenia. AstraZeneca is aware that use of atypicals is not standard across the country. Removal of the mandatory aspect of the current TA will lead to fewer patients gaining access to these important drugs, even if the guideline subsequently recommends their use. This appears incongruous for the area of mental health which has been highlighted by the government as a priority area for England and Wales.

  25.  AstraZeneca supports the ABPI call for continued mandatory funding for a technology whose use remains recommended when reviewed within the context of a clinical guideline.

RECOMMENDATIONS FROM ASTRAZENECA UK LTD

—  Recognition that economic modelling (cost per QALY) should comprise one element of decision-making rather than the single over-riding factor.

—  NICE decision-making should include broader elements of "value"—for example those of most relevance to patients plus societal benefits.

—  NICE to become directly accountable for the quality and consistency of the assessment/ERG reports.

—  Mandatory funding for medicines, previously positively assessed within HTA but where reviews of technology appraisals have been achieved within the context of a clinical guideline, to become applicable where positive recommendation remains within the clinical guideline.

—  Dialogue, from the outset of an appraisal, between manufacturers, NICE and ERGs on methods, data quality and sources used.

—  A far-reaching review of the appeal process, with appeals allowed on the basis of scientific interpretation.

—  A review of NICE guidance implementation with recommendations that result in tangible action, eg greater HCC attention; "joined-up" financial incentives; NICE approved medicines automatically included in local formularies.

AstraZeneca

March 2007