European melanoma incidence: a response to Professor Shaw\'s melanoma editorial

THE NEW ZEALAND MEDICAL JOURNAL Journal of the New Zealand Medical Association

European melanoma incidence: a response to Professor Shaw's melanoma editorial The editorial by James Shaw (Melanoma in New Zealand: a problem that is not going away. NZMJ. 2008;121(1279):6–9. http://www.nzma.org.nz/journal/121-1279/3187) provides commentary on our paper1 published in the same volume of the Journal. Although we all agree that melanoma is a significant problem in New Zealand, and Shaw is supportive of our work, some reported differences need to be clarified and explained. Shaw quotes his published2 incidence rate of melanoma in Caucasians living in Auckland as 78 per 100,000 in 1995 (525 new cases) and questions the difference between his figure and our national age-standardised incidence rate (ASR) of 40.3 per 100,000. Apart from the two reasons given in his paper (the accuracy of 12 [sic] years of national data collection versus 1 year of local data, and the latitude gradient of melanoma incidence), there are several other reasons for this discrepancy which need to be clarified: •

•

•

•

•

•

The 1995 incidence rate for Auckland of 78 per 100,000 is a crude rate. The ASR for Auckland quoted in the Jones et al paper2 was 56.2 per 100,000, and that of Martin and Robinson3 35.6 per 100,000 for 1995–1999, whereas our national ASR for 1995 was 40.3 per 100,000. We did not provide regional figures. Both these other studies2,3 used Segi’s standard population for direct standardisation whereas our rates were standardised to the WHO standard population. Population denominators vary according to the data used. Population counts enumerate people on census night. Population estimates provide population data between census dates using the most recent census data as a base and are usually higher than the census night counts. If population counts were used in the denominator, this would result in an inflation of the incidence rate. Statistics New Zealand produces national population estimates for the usually resident population. Prior to 1996 the population estimates related to the de facto population. Use of the usually resident population results in slightly higher population estimates (and thus lower incidence rates) than the de facto population. NZHIS uses the estimated mean usually resident population for the year ended 31 December XXXX. Jones et al2 restricted the denominator population counts to Caucasians or Europeans. If ethnic prioritised counts were used to calculate the population data, this undercounts the total Caucasian/European population and results in an inflation of the incidence rate. Changes to the census ethnicity questions have resulted in some data that are not consistent between 1991 and 1996, particularly for the European and

NZMJ 7 November 2008, Vol 121 No 1285; ISSN 1175 8716 URL: http://www.nzma.org.nz/journal/121-1285/3359/

Page 141 ©NZMA

Māori ethnic groups. These 2 years were used by Jones et al2 to calculate the 1995 Auckland Caucasian population. Table 1 in Shaw’s editorial also contains some important issues which need to be clarified: The collection and categorisation of ethnicity data in the census has changed in the 3 periods given. In particular, in 2006 ‘New Zealanders’ (11.1% of the total) were not included in the ‘European’ category as in the previous censuses and have been omitted from the table, resulting in a reduced denominator and thus higher rates for Europeans. • Changes to the wording of the ethnicity question in 1996 resulted in a greater than expected increase in Māori and a decrease in Europeans. The wording in 2001 was similar to that of 1991. • Historically, population data have been published by Statistics NZ as total responses and prioritized counts. In 2006, only total population responses were published so they should not sum to 100%. The proportion of people choosing multiple ethnicities has increased over time. For comparative purposes it is important to use the same population data consistently over time. • The overall melanoma incidence figures provided in Table 1 for 1996 and 2001 appear to be WHO ASR from our paper. The provisional WHO ASR for 2006 from NZHIS is 39.2 (22/10/2008). It is unclear where the 2006 overall incidence estimate of 37 per 100,000 comes from. • As the overall melanoma incidence figures quoted appear to be ASR it is inappropriate to divide by an overall population fraction to get a figure for Europeans. Therefore the figures in the line ‘European melanoma incidence’ do not measure what is claimed. Mary Jane Sneyd •

Research Fellow, Hugh Adam Cancer Epidemiology Unit

Brian Cox Associate Professor, Hugh Adam Cancer Epidemiology Unit

Anthony Reeder Director, Social & Behavioural Research Unit Department of Preventive and Social Medicine University of Otago, Dunedin

Ann Richardson Community and Public Health Canterbury District Health Board, Christchurch

References: 1.

2. 3.

Richardson A, Fletcher L, Sneyd M, Cox B, Reeder A. The incidence and thickness of cutaneous malignant melanoma in New Zealand 1994-2004. NZMJ. 2008;121(1279):18–26. http://www.nzma.org.nz/journal/121-1279/3189 Jones W, Harman C, Ng A, Shaw J. Incidence of malignant melanoma in Auckland, New Zealand: highest rates in the world. World Journal of Surgery. 1999;23:732–5. Martin R, Robinson E. Cutaneous melanoma in Caucasian New Zealanders: 1995-1999. Australian & New Zealand Journal of Surgery. 2004;74:233–7.

NZMJ 7 November 2008, Vol 121 No 1285; ISSN 1175 8716 URL: http://www.nzma.org.nz/journal/121-1285/3359/

Page 142 ©NZMA