Skin cancers are a significant public health problem in New Zealand (NZ). In 2009, cutaneous malignant melanoma (melanoma) was the fourth most commonly registered cancer and resulted in 326 deaths.1There are significant sex differences with the male age-standardised registration rate 27% higher than the female rate, and the male death rate more than twice that recorded for females (7.2:3.3 per 100,000).1Non-melanoma skin cancers, estimated to exceed 67,000 cases per annum2 and resulting in 97 reported deaths in 2009,1 are not required to be notified to the New Zealand Cancer Registry, but place a substantial burden on the health system and society.2,3Most skin cancers are potentially preventable through the avoidance of excessive, harmful exposure to ultraviolet radiation (UVR),4 with sunburn associated with increased risk.5For these reasons, national and regional SunSmart health promotion programs have been initiated in NZ since 1988, aimed at increasing public awareness of skin cancer, particularly melanoma, and reducing excessive UVR exposure.6As it was important to evaluate these efforts, the Cancer Society of New Zealand Inc. (CSNZ) and the Health Sponsorship Council (HSC) - now the Health Promotion Agency (HPA), commissioned the Triennial Sun Protection Survey (Sun Survey) series.From 1994 to 2006, cross-sectional surveys were carried out every three years among the NZ urban population. Data collected included sociodemographic information, outdoor sun exposure time, associated activity types and sunburn experience.To date, two peer-reviewed publications have reported descriptive Sun Survey data: one focused on respondents, 15-69 years, in 1994,7 and the other on the youngest age group, 12-17 years, in 1997.8The aim of the present paper is to extend the knowledge gained from these two reports by including data for three additional summers (1999/00, 2002/03 and 2005/06). As previously, outdoor sun exposure (activity type and duration) and sunburn experience for the recall weekends will be described, by sex, age group and skin type.Method Sample selection Respondents, 15-69 years inclusive, were residents of five surveyed metropolitan areas (Auckland, Wellington, Hamilton, Christchurch, and Dunedin) which represented approximately 55% of the total NZ population in the 2006 Census. Data were collected during summer, based on procedures used for a survey administered in Victoria, Australia.9 Participants were recruited either from randomly selected households using random digit dialling in the predetermined areas (1994 and 1997), or through telematching from electoral rolls, 1999-2006. Overall, data usable for analysis were obtained from 6,195 respondents with the breakdown by year as follows: 1994 n=1,243, 1997 n=1,188, 1999/00 n=1,250, 2002/03 n=1,250, and 2005/06 n=1,264. Given a primary prevention focus, interview protocols prioritised younger household members, but a quota system ensured approximately equal numbers by sex and city of residence. Procedures and instrument Data were obtained from computer-assisted telephone interviews (CATI), usually conducted on either the Monday or Tuesday evening, following selected survey weekends. Meteorological data were used to select appropriate survey weekends, with the main criterion being that the weather had been fine enough for potentially harmful sun exposure to have occurred.7 TNS was the contracted data collection agency in 2006, 2003 (then known as NFO New Zealand), and 2000 (then known as CM Research). The MRL Research group (1997) and Roy Morgan Research (1994) were contracted for previous survey years. The surveys used much common questionnaire content to facilitate comparisons across time. Sociodemographic information sought included sex, age and self-defined ethnicity (prioritised according to Level 1, the highest, of the NZ Ministry of Health ethnicity and data protocols).10 Self-reported skin type was based on a modified Fitzpatrick classification of skin sun-reaction: Type I (always burn, never tan), Type II (usually burn, tan with difficulty), Type III (sometimes burn, tan moderately), Type IV (rarely burn, tan easily).11 The exposure days selected varied by year. Data for both Saturdays and Sundays were recorded in 1994, however, due to possible respondent fatigue from survey length and cost issues, Sunday was preferentially selected in 1997 (unless the respondent was only outdoors and sunburnt on Saturday) so 1994 data were treated equivalently. Thereafter the recall day was selected using a standardised procedure.12 Information on sunburn (defined as any amount of reddening of the skin after being in the sun) and the body areas affected were recorded. Those who had spent longer than 15 minutes outdoors on the exposure day, 11am-4pm during daylight savings time (NZ Standard Time +1 hour), provided information on the type and duration of outdoor activities. Activities were classified slightly differently in 1994,7 therefore it was necessary to reclassify and collapse categories, to permit comparisons by year. The six activity categories formed were active recreation (e.g. sports), passive recreation (e.g. reading), water-based recreation (e.g. swimming), paid work, unpaid work (e.g. gardening), or unspecified recreation. The duration of the main outdoor activity (in minutes) was collected from respondents estimates. Analyses Chi-squared tests for association were used to identify significant between-group differences in sunburn patterns, including between years, sexes, age groups, skin sensitivities, and activities. Linear regression models were used to examine associations with duration outdoors. All data were analysed without sampling or post-stratification weights and using SAS 9.1.313 and Stata v12.1 software,14 with two-tailed tests used for the regression models. Statistical significance was determined by p<0.05 in all cases. Ethical approval Participation was taken as informed consent. Participants had previously been notified of the survey by mail from the commissioned market research agency. The proposed project analyses, reported here, were reviewed and ethical approval granted at the Departmental level, following University of Otago Human Ethics Committee procedures. Results Respondent characteristics Overall, 68.8% (n=4,264/6,195) of respondents reported meeting the outdoor status criterion (\u226515 minutes outdoors, 10am-4pm on recall weekend) (Table 1). Table 1. Respondent characteristics by outdoor status, 1994-2006 inclusive Variable All respondents n=6,195 Outdoor respondents n=4,264 P value n % n % of all outdoors % within category outdoors Sex Male Female 3,084 3,111 49.8 50.2 2,273 1,991 53.3 46.7 73.7 64.0 <0.001 Age group (years) 15-19 20-29 30-39 40-49 50-59 60-69 756 1,270 1,416 1,109 999 645 12.2 20.5 22.9 17.9 16.1 10.4 551 907 1,002 748 658 398 12.9 21.3 23.5 17.5 15.4 9.3 72.9 71.4 70.8 67.4 65.9 61.7 <0.001 Skin type* I II III IV Missing data 1,494 3,432 1,109 84 76 24.4 56.1 18.1 1.4 1,014 2,404 773 41 32 23.8 56.4 18.1 1.0 67.9 70.0 69.7 48.8 <0.001 City Auckland Hamilton Wellington Christchurch Dunedin 1,254 1,237 1,230 1,242 1,232 20.2 20.0 19.9 20.1 19.9 891 876 857 808 832 20.9 20.5 20.1 19.0 19.5 71.1 70.8 69.7 65.1 67.5 0.005 Priority ethnicity NZ European M\u0101ori Pacific Asian All other Missing data 5,326 405 123 231 55 55 86.7 6.6 2.0 3.8 0.9 3,676 288 79 140 42 39 86.2 4.6 1.9 3.3 1.0 69.0 71.1 64.2 60.6 76.4 0.026 Survey year 1994 1997 1999/2000 2002/2003 2005/2006 1243 1188 1250 1250 1264 20.1 19.2 20.2 20.2 20.4 854 921 802 808 879 20.0 68.7 <0.001 21.6 77.5 18.8 64.2 18.9 64.6 20.6 69.5 * Fitzpatrick sun-sensitivity scale, modified. Percentages may not total 100% due to rounding. All respondent characteristics were statistically significantly associated with being outdoors, with males, younger people, those with skin types I-III, Maori and those responding in 1997 more likely to report being outside than females, older people, those with skin type IV, Pacific and Asian respondents and those responding in other survey years. The following results are based on the 4,264 respondents who reported \u226515 minutes of outdoor activity, 11am-4pm, during the previous weekend. Sub-group analyses by ethnicity, while of potential interest, were not undertaken because of relatively small numbers in some categories. Sunburn experience Overall, 20.7% (n=882) of 4,259 outdoor respondents with valid datareported sunburn on the recall day (Table 2). Overall, sunburn experience was more common among males, younger respondents and those with vulnerable skin types than females, older respondents and those with less vulnerable skin types. Sunburn rates varied by year between 17.0% and 23.9%, but without a consistent pattern despite being somewhat higher in the most recent year than in the baseline survey. Table 2. Sunburn experience by sex, age and skin type* All Sunburned P value n %

Skin cancers are a significant public health problem in New Zealand (NZ). In 2009, cutaneous malignant melanoma (melanoma) was the fourth most commonly registered cancer and resulted in 326 deaths.1There are significant sex differences with the male age-standardised registration rate 27% higher than the female rate, and the male death rate more than twice that recorded for females (7.2:3.3 per 100,000).1Non-melanoma skin cancers, estimated to exceed 67,000 cases per annum2 and resulting in 97 reported deaths in 2009,1 are not required to be notified to the New Zealand Cancer Registry, but place a substantial burden on the health system and society.2,3Most skin cancers are potentially preventable through the avoidance of excessive, harmful exposure to ultraviolet radiation (UVR),4 with sunburn associated with increased risk.5For these reasons, national and regional SunSmart health promotion programs have been initiated in NZ since 1988, aimed at increasing public awareness of skin cancer, particularly melanoma, and reducing excessive UVR exposure.6As it was important to evaluate these efforts, the Cancer Society of New Zealand Inc. (CSNZ) and the Health Sponsorship Council (HSC) - now the Health Promotion Agency (HPA), commissioned the Triennial Sun Protection Survey (Sun Survey) series.From 1994 to 2006, cross-sectional surveys were carried out every three years among the NZ urban population. Data collected included sociodemographic information, outdoor sun exposure time, associated activity types and sunburn experience.To date, two peer-reviewed publications have reported descriptive Sun Survey data: one focused on respondents, 15-69 years, in 1994,7 and the other on the youngest age group, 12-17 years, in 1997.8The aim of the present paper is to extend the knowledge gained from these two reports by including data for three additional summers (1999/00, 2002/03 and 2005/06). As previously, outdoor sun exposure (activity type and duration) and sunburn experience for the recall weekends will be described, by sex, age group and skin type.Method Sample selection Respondents, 15-69 years inclusive, were residents of five surveyed metropolitan areas (Auckland, Wellington, Hamilton, Christchurch, and Dunedin) which represented approximately 55% of the total NZ population in the 2006 Census. Data were collected during summer, based on procedures used for a survey administered in Victoria, Australia.9 Participants were recruited either from randomly selected households using random digit dialling in the predetermined areas (1994 and 1997), or through telematching from electoral rolls, 1999-2006. Overall, data usable for analysis were obtained from 6,195 respondents with the breakdown by year as follows: 1994 n=1,243, 1997 n=1,188, 1999/00 n=1,250, 2002/03 n=1,250, and 2005/06 n=1,264. Given a primary prevention focus, interview protocols prioritised younger household members, but a quota system ensured approximately equal numbers by sex and city of residence. Procedures and instrument Data were obtained from computer-assisted telephone interviews (CATI), usually conducted on either the Monday or Tuesday evening, following selected survey weekends. Meteorological data were used to select appropriate survey weekends, with the main criterion being that the weather had been fine enough for potentially harmful sun exposure to have occurred.7 TNS was the contracted data collection agency in 2006, 2003 (then known as NFO New Zealand), and 2000 (then known as CM Research). The MRL Research group (1997) and Roy Morgan Research (1994) were contracted for previous survey years. The surveys used much common questionnaire content to facilitate comparisons across time. Sociodemographic information sought included sex, age and self-defined ethnicity (prioritised according to Level 1, the highest, of the NZ Ministry of Health ethnicity and data protocols).10 Self-reported skin type was based on a modified Fitzpatrick classification of skin sun-reaction: Type I (always burn, never tan), Type II (usually burn, tan with difficulty), Type III (sometimes burn, tan moderately), Type IV (rarely burn, tan easily).11 The exposure days selected varied by year. Data for both Saturdays and Sundays were recorded in 1994, however, due to possible respondent fatigue from survey length and cost issues, Sunday was preferentially selected in 1997 (unless the respondent was only outdoors and sunburnt on Saturday) so 1994 data were treated equivalently. Thereafter the recall day was selected using a standardised procedure.12 Information on sunburn (defined as any amount of reddening of the skin after being in the sun) and the body areas affected were recorded. Those who had spent longer than 15 minutes outdoors on the exposure day, 11am-4pm during daylight savings time (NZ Standard Time +1 hour), provided information on the type and duration of outdoor activities. Activities were classified slightly differently in 1994,7 therefore it was necessary to reclassify and collapse categories, to permit comparisons by year. The six activity categories formed were active recreation (e.g. sports), passive recreation (e.g. reading), water-based recreation (e.g. swimming), paid work, unpaid work (e.g. gardening), or unspecified recreation. The duration of the main outdoor activity (in minutes) was collected from respondents estimates. Analyses Chi-squared tests for association were used to identify significant between-group differences in sunburn patterns, including between years, sexes, age groups, skin sensitivities, and activities. Linear regression models were used to examine associations with duration outdoors. All data were analysed without sampling or post-stratification weights and using SAS 9.1.313 and Stata v12.1 software,14 with two-tailed tests used for the regression models. Statistical significance was determined by p<0.05 in all cases. Ethical approval Participation was taken as informed consent. Participants had previously been notified of the survey by mail from the commissioned market research agency. The proposed project analyses, reported here, were reviewed and ethical approval granted at the Departmental level, following University of Otago Human Ethics Committee procedures. Results Respondent characteristics Overall, 68.8% (n=4,264/6,195) of respondents reported meeting the outdoor status criterion (\u226515 minutes outdoors, 10am-4pm on recall weekend) (Table 1). Table 1. Respondent characteristics by outdoor status, 1994-2006 inclusive Variable All respondents n=6,195 Outdoor respondents n=4,264 P value n % n % of all outdoors % within category outdoors Sex Male Female 3,084 3,111 49.8 50.2 2,273 1,991 53.3 46.7 73.7 64.0 <0.001 Age group (years) 15-19 20-29 30-39 40-49 50-59 60-69 756 1,270 1,416 1,109 999 645 12.2 20.5 22.9 17.9 16.1 10.4 551 907 1,002 748 658 398 12.9 21.3 23.5 17.5 15.4 9.3 72.9 71.4 70.8 67.4 65.9 61.7 <0.001 Skin type* I II III IV Missing data 1,494 3,432 1,109 84 76 24.4 56.1 18.1 1.4 1,014 2,404 773 41 32 23.8 56.4 18.1 1.0 67.9 70.0 69.7 48.8 <0.001 City Auckland Hamilton Wellington Christchurch Dunedin 1,254 1,237 1,230 1,242 1,232 20.2 20.0 19.9 20.1 19.9 891 876 857 808 832 20.9 20.5 20.1 19.0 19.5 71.1 70.8 69.7 65.1 67.5 0.005 Priority ethnicity NZ European M\u0101ori Pacific Asian All other Missing data 5,326 405 123 231 55 55 86.7 6.6 2.0 3.8 0.9 3,676 288 79 140 42 39 86.2 4.6 1.9 3.3 1.0 69.0 71.1 64.2 60.6 76.4 0.026 Survey year 1994 1997 1999/2000 2002/2003 2005/2006 1243 1188 1250 1250 1264 20.1 19.2 20.2 20.2 20.4 854 921 802 808 879 20.0 68.7 <0.001 21.6 77.5 18.8 64.2 18.9 64.6 20.6 69.5 * Fitzpatrick sun-sensitivity scale, modified. Percentages may not total 100% due to rounding. All respondent characteristics were statistically significantly associated with being outdoors, with males, younger people, those with skin types I-III, Maori and those responding in 1997 more likely to report being outside than females, older people, those with skin type IV, Pacific and Asian respondents and those responding in other survey years. The following results are based on the 4,264 respondents who reported \u226515 minutes of outdoor activity, 11am-4pm, during the previous weekend. Sub-group analyses by ethnicity, while of potential interest, were not undertaken because of relatively small numbers in some categories. Sunburn experience Overall, 20.7% (n=882) of 4,259 outdoor respondents with valid datareported sunburn on the recall day (Table 2). Overall, sunburn experience was more common among males, younger respondents and those with vulnerable skin types than females, older respondents and those with less vulnerable skin types. Sunburn rates varied by year between 17.0% and 23.9%, but without a consistent pattern despite being somewhat higher in the most recent year than in the baseline survey. Table 2. Sunburn experience by sex, age and skin type* All Sunburned P value n %