In this nationwide cohort study, potential differences in access to expensive anticancer medications funded through the French liste en sus system, which allows direct payment of the drug to the hospital on top of the DRG tariff, were evaluated. A significant reduction in access to medication as a function of social deprivation was observed. The proportion of patients prescribed a FICHCOMP drug ranged from 58.1% for patients living in the most socially deprived municipalities to 63.6% in the most socially privileged. The gradient in medication use across the SDI groups would argue against a spurious correlation. On the other hand, in the univariate analysis, no association of prescription of FICHCOMP medications with the population density of the municipality or with the type of hospital in which the patient was treated was observed. Associations were also observed between the likelihood of FICHCOMP drug prescription and age, gender and the presence of comorbidities. The lower likelihood of prescription in older patients and in patients with those comorbidities taken into account in treatment decisions (MCT meetings) may be explained by reticence to use these treatments in patients who are already frail. For some of these medications, the prescribing indication lists certain comorbidities and the elderly in the precautions for use or contra-indications of the medication. The reason why prescription rates are lower in men than in women is unclear. We have previously shown using the PMSI database that survival is better in women with incident metastatic lung cancer than in men [14], which may indicate that their treatment or disease trajectories are different to those of men.
This study was performed using data from the PMSI database, which is an exhaustive data on all patients hospitalised in France. Since diagnosis and management of cancer patients is exclusively hospital-based in France, it should be possible to identify all incident cases of lung cancer in this database at a national level. Indeed, the number of such cases of that we identified (41715) is close to the number of incident cases in France for 2012 documented by the National Cancer Institute (~ 40,000) [15]. The contemporary quality of coding in the PMSI database is considered to be high and a recent comparison of standardised incidence ratios for different types of cancer determined from the PMSI and from local cancer registries has shown that the two sources provide very similar estimates [16].
Inequalities in access to anticancer medications have been reported previously for several other countries including the United Kingdom [17], Australia [18], Canada [19], and the USA [20], but we are not aware of specific data relating to this issue from France. Two earlier studies of access to chemotherapy for lung cancer in England, one performed in the wealthy South-East [21] and the other in relatively poorer Yorkshire [22], reported differences in access as a function of social deprivation of a similar magnitude to our own study. In the former study [21], these differences were smaller than differences associated with age (higher prescription rates in younger patients) and notably with the cancer network responsible for the area in which the patient lived [23]. In our study, the health service catchment area appeared to be a less important determinant of access to medication. The North American studies have suggested that patients living in rural areas or far from hospitals have a lower access to treatment [24,25,26], a difference which was not observed in our study or in the British studies [17]. This difference probably reflects the much lower population density, and in consequence density of hospitals, in rural areas in North America compared to Europe. The absence of influence of rurality in access to innovative drug is also an important finding.
The present study cannot address whether the differences in access to FICHCOMP medications as a function of social deprivation has a relevant clinical impact. Nonetheless, in a previous analysis of the PMSI database [14], it was found that social deprivation was also associated with reduced survival following diagnosis both at the metastatic and non-metastatic stages. Likewise, the reason why patients living in socially deprived areas are less likely to be prescribed FICHCOMP drugs is also unclear. This is perhaps not due to a lower level of access to cancer care in general, since a recent observational study comparing socially vulnerable patients with lung cancer to less vulnerable patients found that the socially vulnerable declared consulting a general practitioner or an oncologist more often than non-vulnerable individuals [27].
This study also demonstrated that diffusion of innovative anticancer drugs in France was extensive, with around two-thirds of patients with metastatic disease in public hospitals being prescribed an anticancer treatment covered by extra-DRG funding. Prescription volumes of these drugs in France are several-fold higher than they are in the United Kingdom where a specific funding programme for expensive innovative anticancer drugs was introduced in 2010 [28]. This fund was aimed at addressing both the issue of restrictive reimbursement recommendations from the National Institute for Health and Care Excellence (NICE) and to correct for postcode prescribing. In France, before retrospective payment for expensive drugs on the expensive list was implemented, access to such drugs in the public sector was at the discretion of hospitals, which could choose whether or not to fund treatment on a case-by-case basis depending on their financial situation. The special funding mechanism for innovative drugs shelters physicians from the economic impact of their clinical choices on condition that they comply with a “good practice contract” signed with the regional health authorities. However, removing responsibility for the costs of care from the physician shifts that responsibility to the insurer (through drug coverage and management practices) and to the patient (through out-of-pocket cost sharing) [29].
The strengths of the present study include the population-based approach, with a cohort of all lung cancer patients managed in France in 1 year with two-years of follow-up. Nonetheless, the use of this data source presents certain drawbacks. Firstly, at the patient level, demographic variables were limited to gender and age. No information is available in the PMSI database on smoking status, tumour histology and staging, or functional performance. Histology is a major criterion for systemic treatment strategy and several FICHCOMP drugs are actually specific to non-squamous NSCLC. This is the case for the two most-widely used drugs in our study (pemetrexed and bevasizumab). The observed association between access to these drugs and social deprivation may thus be even larger if only non-squamous NSCLC had been considered. Secondly, socioeconomic variables could only be estimated at the level of the municipality of residence of the patient, which is only a proxy marker of individual socioeconomic status. Thirdly, it is also important to note that private hospitals, which account for one-fifth of lung cancer diagnoses in France did not enter data on the use of expensive drugs to the FICHCOMP database at the time of the study. For this reason, it is possible that the extent of use of these drugs and the determinants of use, may differ between the private and public sectors. Finally, it was also not possible to extent the analysis to access to oral tyrosine kinase inhibitors, since these are delivered in community medicine and are thus not available in the FICHCOMP database. At the time of this analysis only medications targeting EGFR were available. Molecular testing is carried out systematically on all patients with lung cancer in France [30] and EGFR mutations and ALK rearrangements are observed in around 16% of patients [31]. The proportion of patients with metastatic lung cancer who are eligible for treatment by tyrosine kinase inhibitors targeting EGFR is relatively low.
It should also be noted that treatment paradigms for metastatic lung cancer are rapidly evolving with the introduction of immunotherapies, the first of which was licensed in France in February 2016, now included in the FICHCOMP list. This development may have consequences for the pattern of use and access to FICHCOMP drugs for the treatment of metastatic lung cancer. The data collected in the present study will serve as a useful reference point to assess such changes in future studies.