Better measurement of the complementarity between UK public, charity and private medical research

Getting benefits from medical research in terms of preventing or treating illness, advancing scientific knowledge and generating economic wealth often, though not always, involves private industry. The private sector builds on and interacts with public and charity funded research and researchers; it conducts its own further research, develops and commercialises medicines and other technologies for use in health care. Theoretical and applied analyses so far published imply that public/charity funded medical research and private sector R&D are complements: extra spending on medical research stimulates extra private sector investment in R&D. But the only attempts to measure this complementarity so far have used US data and are somewhat out of date.

RAND Europe and the OHE therefore propose to develop the methodology for estimating how many £s of pharmaceutical industry R&D are stimulated by an extra £ of public or charity funded medical research. We will improve on the econometric analysis used in the earlier US studies and create a UK, and up to date, dataset of public, charity and private pharmaceutical industry research spending in the UK, disaggregated into therapeutic areas.

Time series data are available for total UK pharmaceutical industry R&D spending, but not broken down by therapeutic area. So we will construct estimated splits of spending by therapeutic area on the basis of two different proxies. First, we will use bibliometric analysis of peer reviewed research articles that have private sector authors based in the UK to estimate the relative weights given by the UK industry to different disease areas in its research. The necessary bibliometric database is available and permits analysis for each year from 1981 onwards. Second we will undertake a corresponding analysis of patents by therapeutic area, where that is evident, by year.

To allow for likely differences between companies in their willingness to publish and their desire to patent, and given that different companies may focus on different therapeutic areas, we will explore the development of indices of companies' propensity to publish and propensity to patent. The indices will then be used to weight the publications and patents produced by those companies when using those data to proxy the split of pharmaceutical industry R&D spending.

We will test the reasonableness of the proxies we develop for splitting total UK pharmaceutical industry R&D spend by therapeutic area via interviews with 12 senior company R&D managers in the UK (from the UK leading R&D companies), who will also be asked about the likelihood of, and the factors affecting, feedback from company R&D influencing public/charity research agendas.

We will improve on analyses to date by separately analysing publicly funded medical research and charity funded medical research, not merely aggregating the two.

We will also attempt to determine whether there is a measurable impact in the reverse direction, i.e. whether greater private pharmaceutical R&D investment stimulates greater charity or public medical research spending in similar areas.

We intend to create time series for the last 20+ years. Using the splits of public, charity and private research spending into 8-10 major therapeutic areas we will estimate models of the form:

(1) Change in Private R&D = f (public research, charitable research, control variables) + time lags of changes

(2) Change in Public or/and charitable research = g (private R&D, control variables) + time lags of changes

Three main challenges face the econometric work: the possibly long time-lags between shocks/interventions and their effects; the possible presence of stochastic time trends; and the possibly bi-directional causality between public/charity and private expenditures. Since the implications of the implied lagged response are important, the model itself should be subjected to stringent testing. Information contained in long-run trends is lost when only time-differenced data are analysed, and this second methodological opportunity and the directionality issue could both be addressed by studying the panel cointegration properties of the various expenditure and control variables, adapting the modelling strategy accordingly. In this approach, private R&D and public/charitable research are regarded as co-determined in the long run, while it is possible to allow for differences in their short-run dynamics.

We intend to break R&D expenditure down into therapeutic areas by developing techniques in bibliometric and patent analysis. We will use publications and patents as proxies for intensity of activity in those areas, controlling for the assumption that some fields are prone to more publication or patent activity than others.

Knowledge of the magnitude of public/charity/private research complementarities is an essential ingredient of the evidence base for policy decisions about public (and charity) funding of medical research. Currently we do not know whether, although it is widely assumed that, an extra pound of taxpayers' money (or charity) money spent on medical research generates additional private pharmaceutical industry R&D in the UK and consequent economic and scientific benefit to the UK. Our research is designed to answer that question.

The expected beneficiaries of our research in the UK are therefore the MRC and other medical research funders, central government in the form of the departments responsible for deciding on public research funding (BIS, HMT, DH), researchers, and the private sector organisations that benefit from the stimulus provided by public and charity funded research. All of these should benefit from greater understanding of the magnitude of public/charity/ private research complementarity. The principles of measuring complementarity should also benefit the equivalent bodies in non-medical areas of research where the private sector has a potentially important mediating role, although the magnitudes estimated for medical research cannot be assumed necessarily to apply in non-medical sectors. Furthermore, other nations with significant public and/or charity funded medical and other research sectors and private R&D based private sectors complementing them will also benefit from understanding improved ways of measuring that complementarity in their own cases.

These benefits should be realisable by all beneficiaries immediately after they are published, subject to effective dissemination of the published results to those respective audiences in the UK and internationally (see the 'pathways to impact' note).