‘The Health of the Nation’ – an unattainable skin cancer target predictably remains beyond reach

Background – the target

In 1992 the UK government strategic document, The Health of the Nation, listed numerous targets to improve health. 1 Of the hundreds of targets listed, there was just one of dermatological relevance, which was to ‘halt the year on year increase of skin cancer by the year 2005’. Remarkably, the impossibility of achieving this target did not generate any real attention at the time of the publication, possibly because ‘target overload’ was developing, or because dermatology was clearly such a minor consideration in the document. An interim analysis did not consider skin cancer, and no formal evaluation has been performed. However, almost half (47%) of 390 general practitioners surveyed in 1996 believed that the desired halt in the year-on-year increase in skin cancer could be achieved by 2005, even though most believed that most of the other targets were not achievable (for example, only 25–35% believed that specific targets set for reduction in death rates due to coronary artery disease, stroke, lung cancer or accidents would be achievable). 2 The authors suggested that the reason for this relative optimism about skin cancer targets might be due to existing campaigns and public awareness at the time.

Method and outcomes

The 2005 Cancer Registration statistics have now been published. 3 The data available spans 1992–2005, i.e. the exact dates from publication of the ‘Health of the Nation’ document through to the proposed timescale for the skin cancer target, thus providing a rationale for performing the analysis presented.

The statistical publication for each year throughout this time period was examined to extract the number of registrations for skin cancer (the years 1995–1997 are published together but each year is identifiable in separate tables; some years have more than one publication, in which event the final version was used). 4 Melanoma of the skin is recorded separately (172 in ICD9, C43 in ICD10) from ‘other malignant neoplasms of skin’ (173 in ICD9, C44 in ICD10). In situ carcinomas of skin were not considered.

The results of the annual registrations are shown in Figure 1. For simplicity, the term non-melanoma skin cancer (NMSC) has been used to describe the ‘other malignant neoplasms of skin’; in common use, this essentially equates to basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) of the skin, although there are small numerical contributions from rarer primary skin tumours, and some (also rare) tumours that affect the skin (e.g. cutaneous T-cell lymphoma) are considered separately by Registries.

Figure 1

Registrations of melanoma and non-melanoma skin cancer (NMSC) in England 1992–2005 (data from www.statistics.gov.uk publications, 3 under licence)

Registrations for melanoma increased from 4151 to 8025 between 1992 and 2005 (a 93% increase), and for NMSC from 38,745 to 67,068 (a 73% increase). This is consistent with the worldwide increase in NMSC, estimated at 3–10% per annum, 5, 6 and an approximate doubling in melanoma incidence every 10 years. The available UK figures for NMSC do not allow evaluation of the registration numbers for BCC compared with SCC, as the two are considered together, but generally worldwide the crude incidence of melanoma is about 8–10% that of NMSC, and BCC accounts for about 85% of NMSC. 6, 7

Registration issues

Skin cancer registration data are subject to some acknowledged inaccuracy, other than the inevitable fact that all Registry data take a finite time to be acquired, collated and reported. Under-reporting of NMSC is a well-documented problem that arises for various reasons, and is certainly not unique to the UK; NMSC is not systematically recorded in many other European countries. 5 Even in the USA, the National Cancer Institute has not recorded data on squamous cell carcinoma of the skin, despite a lifetime risk of 9–14% in men and 4–9% in women. 8 As registration has generally relied on pathological data, there are situations where, even with willingness to record such data, some tumours may never come to the attention of Cancer Registries (for example, small basal cell carcinomas diagnosed clinically and treated with cryotherapy in an outpatient clinic). The magnitude of this problem is unclear, but a study in Wales estimated that the actual numbers of NMSC were 60% higher than previous estimates in the UK, a significant level of under-reporting. 9 Compounding this problem, considering BCC and SCC together takes no account of their different relationship with sunlight and other aetiologies, different biological behaviour and different mortality outcomes.

Although UK Cancer Registries can potentially document all tumours, the standard format for international comparisons only considers the first NMSC even if a patient has many such tumours, such that data pertaining to additional NMSC are superfluous to requirements. There has therefore been no impetus to consider the potential importance of full recording, although some important reasons to do so are discussed later. By comparison, all melanomas (including second primary lesions) are recorded. Many organizations (such as the World Health Organization and the International Agency for Research on Cancer) omit NMSC from their published figures, and others may record them but omit them from analyses.

Comparative data

To extrapolate conclusions regarding skin cancer registrations from internationally reported figures requires caution, as sunlight and skin colour have such an important role in the epidemiology. Even within the same country, there may be different incidences at different latitudes. 5 However, studies in other countries as well as other UK studies show trends similar to those reported here. In Wales, the crude 10-year increase in NMSC per 100,000 population between 1988 and 1998 was 53%, 10 and the Northern Ireland Cancer Registry documented a 62% increase in skin cancers (including melanoma) over the period 1993–2004. 11 Notably, this study found that the increase in numbers of patients only increased by 20%, consistent with the fact that patients often have more than one NMSC. In Scotland, there was an annual estimated increase in incidence of SCC by 2.1% from 1992–2003, and an equivalent increase for BCC of 1.4% per annum (although the latter was based on first occurrence of BCC only for any individual). 12 In the Netherlands, BCCs increased 2.4% per annum in men and 3.9% in women from 1973–2000; 13 this paper cites NMSC as accounting for over one-third of all cancers in the USA, and for twice as many cancers as all others combined in Australia.

Biological issues

While melanoma is undoubtedly the most important skin cancer in terms of mortality, it accounts for less than 10% of skin cancers in temperate climates on the basis of reported figures, and presumably less than this if the under-reporting issue of NMSC is acknowledged. Thus, to alter the year-on-year increase in skin cancer requires measures that alter the incidence of NMSC, especially BCC; altering the incidence of melanoma is crucial for improvements in deaths from skin cancer but is irrelevant to the stated government target of reducing the number of skin cancers. It may be that the importance of melanoma and its mortality was uppermost in the consideration when suggesting the target about halting progressive increases in skin cancer, but that this became generalized to ‘skin cancer’ rather than specific to ‘melanoma’; the rate of increase of melanoma was falling 12, 14 and had stabilized in women in Scotland by 1994, 15 but this was not documented for other skin cancers at the time. Reduction in SCC by sun avoidance has been confirmed subsequently, as discussed below, but only in active study populations.

The main factors that influence development of NMSC are ultraviolet radiation (UVR; mainly as sunlight), skin pigmentation and age. However, several lines of evidence suggest that the pattern of UVR exposure differs for BCC and SCC. Although sunlight within the preceding decade may have a disproportionate effect, the bulk of the evidence for squamous cell carcinoma (SCC) is that it is relatively simply related to long-term cumulative sunlight exposure 5, 8 although this may not be the case if pigmentary factors are taken into account. 16 While their degree of importance as an SCC precursor is debated, actinic keratoses (AKs) occur at an earlier age and are the lesion underlying at least some SCCs; the time between developing baldness and developing scalp AKs in the UK averages 35 years. 17 However, for BCC, the position of BCC on the face does not correlate with the site-specific UVR exposure, 18 and migration studies and sun exposure questionnaire studies suggest that childhood sunburn and recreational UVR exposure in the 0–19 years age group appear to be critical for later BCC risk. 19, 20 Thus prevention needs to be employed for several decades to have a significant impact. Even though strict sunscreen use had a detectable effect on SCC incidence in Australia after 4.5 years, 21 within the context of a controlled study rather than in the general population, the magnitude of any such impact in the UK is likely to be small and no such data exists for an influence on incidence of BCC which is numerically the greatest contributor to skin cancer incidence. Indeed, an additional eight years of follow-up in the Australian SCC study showed almost 40% reduction in SCCs and although there was a slight decrease in BCC in this interventional trial population, this was reported as non-significant. 22

Are the increases in NMSC real?

It is feasible that the increased recording of NMSC is due to improved ascertainment rather than a true increase in the number of cases. This is intuitively unlikely, the main increase more likely being due to increased longevity and a historical lack of sunlight avoidance measures over the timescale that contributes to risk. Equally, the undoubted increase in NMSC related to therapeutic immunosuppression, while contributing disproportionate numbers and mortality, is inadequate to explain the annual increases in numbers discussed above.

More direct proof of a true increase comes from, albeit short-term, studies of patients with a known previous NMSC. A meta-analysis of risk of subsequent NMSC in patients with a history of NMSC showed that the mean 3-year risk of a further SCC was 18% after a previous SCC, and for BCC was 44% after a previous BCC (both at least 10-fold increased above general population levels). 23 The mean cumulative three-year risk of any NMSC after a previous NMSC was 38% for those with <3 previous NMSCs, but 93% for those with 3–9 previous NMSCs. Many of the studies considered were prospective cohort studies and thus were not reliant on registry data.

Why record all NMSC?

NMSC can no longer be viewed as a non-serious minor nuisance in the field of cancer. It constitutes a major workload for dermatologists and others, is increasingly important in the field of organ transplantation, and has major public health impact. Even if under-registration is only around 30%, approximately 100,000 cases will occur each year in the UK; in the USA, the figure is well over 1 million. 5, 7, 9 As documented above, lifetime risk of SCC may be 5–15%, and BCC is four or five times commoner. Without accurate data, the differing epidemiologies of melanoma and the main NMSCs will remain poorly understood, and the impact of prevention campaigns will be impossible to assess accurately. Unfortunately, the sheer volume of cases may preclude achieving this aim, lack of time being the major limiting factor in documenting all cases of skin cancer during piloting of the National Cancer Data project; the current proposal is to record a full data entry for the first NMSC but to at least record the number and type of further lesions treated. 24

Conclusion

It should therefore be no surprise that the target to halt the year on year increase in skin cancer by 2005, starting from 1992, was an unattainable dream, not because various educational campaigns are not efficient or commendable, but because the lag-time for development of NMSC, and especially for BCC, the major issue in incidence if not in mortality, exceeds the 13-year timescale of the target document by several decades.

Thus, the pump has probably been primed for at least three or four times as long as the target for reducing the incidence has allowed. This is of some importance, not only for patients of the future, but also because healthcare in all developed countries is increasingly political and financial; the need for financial support of long-term educational efforts is clear, but at the same time there should be an awareness that unachievable short-term targets should not be part of political or local contracting quality standards. It also underlines the fact that the biology and natural history of skin cancers is poorly understood, both at political level and amongst non-specialists, perhaps identifying an additional educational need. Individually, we are all at risk; to cite a Netherlands study: ‘There is no indication of an end to this trend in BCC.’ 15

Footnotes

DECLARATIONS

Competing interests None declared

Funding None

Ethical approval Not applicable

Guarantor NHC

Contributorship NHC is the sole contributor