Abstract

The QSkin Sun and Health Study comprises a cohort of 43 794 men and women aged 40–69 years randomly sampled from the population of Queensland, Australia in 2011. The cohort was established to study the development of skin cancer and melanoma in the population with the highest reported incidence of these diseases in the world. At baseline, besides demographic items and general medical history, information about standard pigmentary characteristics (including hair and eye colour, freckling tendency, tanning ability and propensity to sunburn), past and recent history of sun exposure and sunburns, sun protection behaviours, use of tanning beds and history of skin cancer was collected by self-completed questionnaire. Participants have given their consent for data linkage to the universal national health insurance scheme and for linkage to cancer registries and pathology databases, thus ensuring complete ascertainment of all future skin cancer and melanoma occurrences and medical treatments and other cancer events. Linkage to these registers will occur at predetermined intervals. Approval to access QSkin data can be obtained on application to the study investigators and submission of a formal research plan that has previous approval from the human research ethics committee of the applicant’s institution.

Why was the cohort set up?

The QSkin Sun and Health study was established in 2011 to investigate prospectively the role of environment and host/genetic characteristics in the aetiology of cutaneous melanoma and other cancers of the skin. Funding was provided through a programme grant from the National Health and Medical Research Council of Australia to investigators at the Queensland Institute of Medical Research in Brisbane, Australia, a setting in which the reported incidences of cutaneous melanoma and skin cancer are the highest in the world (melanoma age-standardized incidence rates (ASR) 65.3 × 105; keratinocyte cancer ASR 1878 × 105).1

The primary aim is to derive measures of absolute and relative risk for melanoma and other skin cancers [notably basal cell carcinomas (BCC) and squamous cell carcinomas (SCC)—the most common cancers in humans] associated with phenotypic, lifestyle, clinical and environmental factors. Secondary aims are to use these prospective data to develop risk prediction tools for melanoma and other skin cancers, to measure the effects of protective behaviours as far as possible and to estimate the burden (e.g. doctor visits, direct and indirect costs, mortality) of skin cancer and melanoma in the Queensland population. Finally, the cohort aims to serve as a platform for nested studies exploring skin cancer and melanoma and other disease end points that may be associated with sun exposure.

The QSkin Sun and Health Study fulfils several research needs; whereas numerous previous studies of skin cancer have been conducted, the vast majority have been retrospective studies using case-control designs. Concerns about selection bias and misclassification of exposure (especially differential recall of sun exposure and other factors) have complicated the interpretation of the findings of those earlier studies. Another pressing issue, still unresolved, is to establish whether a ‘critical period’ in life exists during which exposure to sunlight is particularly harmful. A further challenge has been to explore the temporal relationships between patterns (as distinct from amount) of sun exposure and risks of melanoma, BCC and SCC. Such questions have proven impossible to answer satisfactorily with retrospective designs. Moreover, it is difficult to derive measures of absolute risk with which to develop risk prediction tools from case-control studies, especially when one wishes to predict the risks associated with combinations of risk factors. Few prospective studies have measured skin cancer end points; still fewer have measured key risk factors for skin cancer at baseline. We recently conducted systematic reviews/meta-analyses of the major established risk factors for melanoma including common and atypical naevi, pigmentary characteristics (eye and hair colour, skin colour/phototype, freckling) and family history.25 These reviews provide the most comprehensive estimates of melanoma risk associated with these factors; however, it was not feasible to use these estimates for real-life risk prediction, given that the various possible interactions between the risk factors cannot be derived from published literature to date. Of the currently available melanoma risk prediction tools, most are based on data from case-control studies,610 and all but one of these7 were conducted in the northern hemisphere. Currently, there is only one prediction tool based on prospective cohort data,11 again for a northern hemisphere population.

Who is in the cohort?

The study is being conducted in the state of Queensland (population 4.6 million), Australia, an extensive subtropical/tropical jurisdiction (latitude 29°S–10°S) of 1 727 000 km2, which is approximately seven times the size of Great Britain. Because of its nearness to the equator, average ultraviolet radiation (UVR) levels in spring and summer in Queensland are high to extreme; annual erythemally effective solar UVR for the south of the state is ∼11 800 standard erythema dose (SED) and ∼14 500 SED in the north.12 The majority of residents (66%) trace their ancestry to northern Europe.13 From within this population, we randomly sampled 96 644 men and 96 700 women (193 344 in total) aged 40–69 years from the Australian Electoral Roll. Registration on the electoral roll is compulsory in Australia, and at the time of the 2007 federal election, 92% of eligible people were enrolled.14 The age and sex distribution of the sample (in 5-year age-groups) matched the age distribution of the state of Queensland; sampling was random within each age-group. Between November 2010 and November 2011, potentially eligible individuals were mailed an invitation to take part. Included in the mail-out pack was an information sheet, the study survey and consent forms and a reply-paid envelope. The survey questions are provided in Table 1. Participants joined the study by completing the survey and consent forms and returning them in the reply-paid envelope. One consent form covered use of information provided in the survey and gave permission for data linkage to cancer registries, pathology laboratories and public hospital databases; a second consent form gave permission for data linkage to Australia’s universal national health insurance scheme (Medicare Australia) that records information on all medical services to Australian residents except for those conducted in public hospitals (covered by the first consent form). Participants were also given the option of completing their survey and consent forms online through the website www.qskin.qimr.edu.au. We sent a single reminder card to all participants who did not respond within ∼2.5 weeks of the initial letter. Most participants (87%) completed the hard-copy survey with only 13% completing the survey online.

Table 1

Data items collected in the baseline survey of QSkin Sun and Health Cohort

Domain Question 
Birth and residence 
  1. Age in years.

  2. Marital status.

  3. Place of birth.

  4. Age at migration to Australia.

  5. Residential history.

  6. Private health insurance status.

  7. Veterans health card holder.

  8. Ancestry/ethnicity (Australian census question).

 
Education and work 
  1. Education level (six categories).

  2. Current working status (seven categories).

 
Phenotype 
  1. Skin colour (four categories).

  2. Skin type (burning) (four categories).

  3. Skin type (tanning) (four categories).

  4. Eye colour (six categories).

  5. Hair colour (five categories).

  6. Freckling (four categories, with pictograms).

  7. Naevi (four categories, with pictograms).

 
Sun exposure and sun protection 
  1. Number of sunburns in three separate age periods (each six categories).

  2. Frequency of sunscreen use (four categories).

  3. Frequency of sunscreen and hat use (each four categories).

  4. Time spent outdoors, weekdays and weekends for four separate age periods (each four categories).

  5. Frequency of use of sunbeds or tanning beds (six categories).

 
Medical history 
  1. Self-rated health (five categories).

  2. Number of skin cancers excised (five categories).

  3. Number of sunspots or skin cancers destroyed (five categories).

  4. Frequency of medication use (separately for paracetamol, aspirin, anti-inflammatories, corticosteroids) (each four categories).

  5. History of melanoma in close blood relatives.

  6. Self-perceived risk of developing melanoma in the future (five categories, and percentage probability).

 
Height, weight and lifestyle 
  1. Height (cm or feet + inches).

  2. Weight now (kg).

  3. Weight at age 21 years (kg).

  4. Trouser/dress size.

  5. Smoking duration and average daily intake.

  6. Weekly number of alcoholic drinks (eight categories); number of days per week alcohol consumed (seven categories).

  7. Number of servings of fruit, fruit juice and vegetables per day.

  8. Hours of sleep per night (six categories).

  9. Self-rated stress in past year (10 categories).

 
Skin checks 
  1. Frequency of whole body skin examination in past 3 years by self, doctor, other person (each five categories).

 
Women only 
  1. Age at menarche (years).

  2. Menopausal status, age at menopause (years), type of menopause (four categories).

  3. Hormone replacement therapy use and duration.

  4. Oral/injected contraceptive use and duration.

  5. Parity.

  6. Endometriosis ever; basis for diagnosis (two categories).

 
Other diagnoses 
  1. Diagnoses of cancer other than skin cancer.

  2. Diagnoses of other serious disease requiring hospitalization.

 
Domain Question 
Birth and residence 
  1. Age in years.

  2. Marital status.

  3. Place of birth.

  4. Age at migration to Australia.

  5. Residential history.

  6. Private health insurance status.

  7. Veterans health card holder.

  8. Ancestry/ethnicity (Australian census question).

 
Education and work 
  1. Education level (six categories).

  2. Current working status (seven categories).

 
Phenotype 
  1. Skin colour (four categories).

  2. Skin type (burning) (four categories).

  3. Skin type (tanning) (four categories).

  4. Eye colour (six categories).

  5. Hair colour (five categories).

  6. Freckling (four categories, with pictograms).

  7. Naevi (four categories, with pictograms).

 
Sun exposure and sun protection 
  1. Number of sunburns in three separate age periods (each six categories).

  2. Frequency of sunscreen use (four categories).

  3. Frequency of sunscreen and hat use (each four categories).

  4. Time spent outdoors, weekdays and weekends for four separate age periods (each four categories).

  5. Frequency of use of sunbeds or tanning beds (six categories).

 
Medical history 
  1. Self-rated health (five categories).

  2. Number of skin cancers excised (five categories).

  3. Number of sunspots or skin cancers destroyed (five categories).

  4. Frequency of medication use (separately for paracetamol, aspirin, anti-inflammatories, corticosteroids) (each four categories).

  5. History of melanoma in close blood relatives.

  6. Self-perceived risk of developing melanoma in the future (five categories, and percentage probability).

 
Height, weight and lifestyle 
  1. Height (cm or feet + inches).

  2. Weight now (kg).

  3. Weight at age 21 years (kg).

  4. Trouser/dress size.

  5. Smoking duration and average daily intake.

  6. Weekly number of alcoholic drinks (eight categories); number of days per week alcohol consumed (seven categories).

  7. Number of servings of fruit, fruit juice and vegetables per day.

  8. Hours of sleep per night (six categories).

  9. Self-rated stress in past year (10 categories).

 
Skin checks 
  1. Frequency of whole body skin examination in past 3 years by self, doctor, other person (each five categories).

 
Women only 
  1. Age at menarche (years).

  2. Menopausal status, age at menopause (years), type of menopause (four categories).

  3. Hormone replacement therapy use and duration.

  4. Oral/injected contraceptive use and duration.

  5. Parity.

  6. Endometriosis ever; basis for diagnosis (two categories).

 
Other diagnoses 
  1. Diagnoses of cancer other than skin cancer.

  2. Diagnoses of other serious disease requiring hospitalization.

 

The overall participation fraction was 23%; of the 193 344 people invited, 43 794 (19 920 men and 23 874 women) returned completed surveys with a signed consent form, and 5990 were deemed ineligible (5972 because their invitations were returned marked ‘not at this address’ and 18 were deceased, as notified by family members). The exact participation fraction is difficult to specify, as other potential participants may not have received the invitation if their address details were incorrect in the Australian Electoral Commission database, although this is not considered to have led to substantial non-response. All completed hard-copy surveys were scanned electronically and stored as images and data following a rigorous series of checks and validation procedures by two data management operators. Of the invited participants, positive responders were more likely to be in the older age-group (55–69) than non-responders (54% vs 38%; P < 0.001) and were more likely to be women (59% vs 55%; P < 0.001) (Table 2). We compared the QSkin Cohort with the Queensland population for several key characteristics;15,16 the two populations were similar with respect to educational attainment, employment status and body mass index (BMI); QSkin participants were slightly less likely to be current smokers (Table 3).

Table 2

Comparison between positive responders and non-responders in the QSkin Sun and Health Cohort

Age-group (years) Males P value Females P value 
Responders (n = 19 920) Non-responders (n = 76 724) Responders (n = 23 874) Non-responders (n = 72 826) 
n (%) n (%) n (%) n (%) 
40–44 2192 (11.0) 17 445 (22.7)  3421 (14.3) 16 743 (23.0)  
45–49 2708 (13.6) 16 384 (21.4)  3994 (16.7) 15 519 (21.3)  
50–54 3525 (17.7) 14 120 (18.4)  4488 (18.8) 13 186 (18.1)  
55–59 4019 (20.2) 12 879 (16.8)  4625 (19.4) 11 956 (16.4)  
60–64 3847 (19.3) 9450 (12.3)  3996 (16.7) 8907 (12.2)  
65–69 3629 (18.2) 6446 (8.4)  3350 (14.0) 6515 (8.9)  
   <0.001   <0.001 
Age-group (years) Males P value Females P value 
Responders (n = 19 920) Non-responders (n = 76 724) Responders (n = 23 874) Non-responders (n = 72 826) 
n (%) n (%) n (%) n (%) 
40–44 2192 (11.0) 17 445 (22.7)  3421 (14.3) 16 743 (23.0)  
45–49 2708 (13.6) 16 384 (21.4)  3994 (16.7) 15 519 (21.3)  
50–54 3525 (17.7) 14 120 (18.4)  4488 (18.8) 13 186 (18.1)  
55–59 4019 (20.2) 12 879 (16.8)  4625 (19.4) 11 956 (16.4)  
60–64 3847 (19.3) 9450 (12.3)  3996 (16.7) 8907 (12.2)  
65–69 3629 (18.2) 6446 (8.4)  3350 (14.0) 6515 (8.9)  
   <0.001   <0.001 
Table 3

Comparison between the QSkin Sun and Health Cohort and Queensland population for selected characteristics

Variable Males Females 
QSkin (%) QLDa (%) QSkin (%) QLDa (%) 
Education 
    Post-school qualification 57 60.1 45.4 45.4 
Employment 
    Full-time 58.1 54.3 31.7 27.4 
    Part-time 8.3 8.9 24.6 24.2 
Smoking 
    Current smokers 10.6 17.1 13.9 
BMI (kg/m2) 
    <25 24.6 27 40.4 42.2 
    25–29.9 46.6 44.8 29.8 31.7 
    >30 26.2 28.1 25.4 26.1 
    Missing 2.6 – 4.5 – 
Variable Males Females 
QSkin (%) QLDa (%) QSkin (%) QLDa (%) 
Education 
    Post-school qualification 57 60.1 45.4 45.4 
Employment 
    Full-time 58.1 54.3 31.7 27.4 
    Part-time 8.3 8.9 24.6 24.2 
Smoking 
    Current smokers 10.6 17.1 13.9 
BMI (kg/m2) 
    <25 24.6 27 40.4 42.2 
    25–29.9 46.6 44.8 29.8 31.7 
    >30 26.2 28.1 25.4 26.1 
    Missing 2.6 – 4.5 – 

aNote that for Queensland (QLD) the age-group is 35–74 years for education and employment; 40–69 years for smoking and BMI.

The socio-demographic and phenotypic characteristics of the 43 794 members of the cohort are provided in Table 4. Overall, 46% of participants are male and 54% female. At the time of survey completion, mean age of participants was 56 years (57 for males and 55 for females). Most participants reported having White European ancestry (93%); 2% had Asian, 0.2% Aboriginal or Torres Strait Islander, 0.5% ‘other’ and 4% mixed ancestry. Based on self-reported measures, 59% had fair skin, 32% medium skin and only 8% olive/brown or black skin. Most participants (90%) had a skin type that burned after long-term sun exposure, and 53% reported freckling on the face in early adulthood. Thirty-nine percent of participants reported having had one or more skin cancers surgically removed, and 19% had had >10 skin cancers/solar keratoses burned or frozen; 23% of participants reported a family history of melanoma. Almost three quarters of participants (73%) had had their skin checked by a medical practitioner in the past 3 years, 43% had their skin checked by someone else, and 83% had checked their own skin. Women were more likely than men to check their own skin (P < 0.001), and men were more likely than women to have their skin checked by another person (P < 0.001). Twelve percent of participants reported having had no skin checks (doctor, other, self) in the past 3 years. Thirty-two percent of participants were in the healthy BMI range (18.5–24.9 kg/m2), 37% were overweight (25–29.9 kg/m2) and 26% obese (≥30 kg/m2). Overall, 10% of participants were current smokers and 35% ex-smokers.

Table 4

Baseline characteristics of the QSkin Sun and Health Cohort

Variables Invited participants 
Male (n = 19 920) Female (n = 23 874) Total (n = 43 794) 
n (%) n (%) n (%) 
Age (years) 
    40–44 2192 (11.0) 3421 (14.3) 5613 (12.8) 
    45–49 2708 (13.6) 3994 (16.7) 6702 (15.3) 
    50–54 3525 (17.7) 4488 (18.8) 8013 (18.3) 
    55–59 4019 (20.2) 4625 (19.4) 8644 (19.7) 
    60–64 3847 (19.3) 3996 (16.7) 7843 (17.9) 
    65–69 3629 (18.2) 3350 (14.0) 6979 (15.9) 
Ethnicity 
    White 18 675 (93.8) 21 952 (92.0) 40 627 (92.8) 
    Black 49 (0.3) 58 (0.2) 107 (0.2) 
    Asian 324 (1.6) 503 (2.1) 827 (1.9) 
    Aboriginal/Torres Strait Islander 43 (0.2) 53 (0.2) 96 (0.2) 
    Other 574 (0.3) 69 (0.3) 132 (0.3) 
    Mixed 192 (2.9) 981 (4.1) 1555 (3.6) 
    Missing 192 (1.0) 258 (1.1) 450 (1.00) 
Highest qualification 
    No school certificate 1518 (7.6) 1905 (8.0) 3423 (7.8) 
    School certificate 2388 (12.0) 4340 (18.2) 6728 (15.4) 
    Higher school certificate 2930 (14.7) 4999 (20.9) 7929 (18.1) 
    Trade/apprenticeship 3379 (17.0) 648 (2.7) 4027 (9.2) 
    Certificate/diploma 3264 (16.4) 4855 (20.3) 8119 (18.5) 
    University degree 4693 (23.6) 5694 (23.9) 10 387 (23.7) 
    Missing 1748 (8.8) 1433 (6.0) 3181 (7.3) 
Employment 
    Full-time worker 11 566 (58.1) 7567 (31.7) 19 133 (43.7) 
    Part-time worker 1651 (8.3) 5871 (24.6) 7522 (17.2) 
    Home duties 123 (0.6) 2969 (12.4) 3090 (7.1) 
    Unemployed 393 (2.0) 288 (1.2) 681 (1.6) 
    Student 90 (0.5) 177 (0.7) 267 (0.6) 
    Retired 4719 (23.7) 4742 (19.9) 9461 (21.6) 
    Other 677 (3.4) 783 (3.3) 1460 (3.3) 
    Missing 701 (3.5) 1479 (6.2) 2180 (5.0) 
Smoking status 
    Never smoked 9607 (48.2) 14 280 (59.8) 23 887 (54.4) 
    Ex-smoker 8075 (40.5) 7314 (30.6) 15 389 (35.1) 
    Current smoker 2105 (10.6) 2152 (9.0) 4257 (9.7) 
    Missing 133 (0.7) 128 (0.5) 261 (0.6) 
BMI a year ago (kg/m2) 
    <18.5 70 (0.4) 377 (1.6) 447 (1.0) 
    18.5–24.9 4834 (24.3) 9282 (38.9) 14 116 (32.2) 
    25–29.9 9277 (46.6) 7109 (29.8) 16 386 (37.4) 
    30–34.9 3871 (19.4) 3823 (16.0) 7694 (17.6) 
    35–39.9 963 (4.8) 1453 (6.1) 2416 (5.5) 
    >40 377 (1.9) 779 (3.3) 1156 (2.6) 
    Missing 516 (2.6) 1063 (4.5) 1579 (3.6) 
Parity 
    No children  2782 (11.7)  
    One child  2436 (10.2)  
    Two children  9105 (38.1)  
    Three children  6052 (25.4)  
    Four or more children  3248 (13.5)  
    Missing  251 (1.1)  
Family history of melanoma 
    Yes 4003 (20.1) 6065 (25.4) 10 068 (23.0) 
    No 12 406 (62.3) 14 637 (61.3) 27 043 (61.8) 
    Not known 3309 (16.6) 2884 (12.1) 6193 (14.1) 
    Missing 202 (1.0) 288 (1.2) 490 (1.1) 
Skin cancers surgically removed 
    None 11 196 (56.2) 15 031 (63.0) 26 227 (59.9) 
    One 2590 (13.0) 3458 (14.5) 6048 (13.8) 
    2–10 4742 (23.8) 4536 (19.0) 9278 (21.2) 
    11–20 710 (3.6) 424 (1.8) 1134 (2.6) 
    >20 546 (2.7) 227 (1.0) 773 (1.8) 
    Not known 13 (0.1) 17 (0.1) 30 (0.1) 
    Missing 123 (0.6) 181 (0.8) 304 (0.7) 
Skin cancers/solar keratoses burned/frozen 
    None 8526 (42.8) 11 359 (47.6) 19 885 (45.4) 
    1–5 4652 (23.4) 6762 (28.3) 11 414 (26.1) 
    6–10 1957 (9.8) 2121 (8.9) 4078 (9.3) 
    11–20 1757 (8.8) 1580 (6.6) 3333 (7.6) 
    21–50 1513 (7.6) 1117 (4.7) 2630 (6.0) 
    >50 1416 (7.1) 785 (3.3) 2201 (5.0) 
    Not known 11 (0.1) 19 (0.1) 30 (0.1) 
    Missing 92 (0.5) 131 (0.6) 223 (0.5) 
Skin colour 
    Fair 11 648 (58.5) 14 318 (60.0) 25 966 (59.3) 
    Medium 6321 (31.7) 7699 (32.3) 14 020 (32.0) 
    Olive/dark 1828 (9.2) 1685 (7.1) 3513 (8.0) 
    Black 21 (0.1) 13 (0.1) 34 (0.1) 
    Missing 102 (0.5) 159 (0.7) 261 (0.6) 
Skin type 
 Skin reaction to 30 minutes midday sun 
        No burns 2122 (10.7) 1901 (8.0) 4023 (9.2) 
        Burns a little 9273 (46.6) 9567 (40.1) 18 840 (43.0) 
        Burns moderately 6314 (31.7) 8274 (34.7) 14 588 (33.3) 
        Burns badly 2107 (10.6) 3967 (16.6) 6074 (13.9) 
        Missing 104 (0.5) 165 (0.7) 269 (0.6) 
 Skin reaction to several weeks in sun 
        No tan 990 (5.0) 1857 (7.8) 2847 (6.5) 
        Tan lightly 3682 (18.5) 5480 (23.0) 9162 (20.9) 
        Tan moderately 10 104 (50.7) 11 267 (47.2) 21 371 (48.8) 
        Tan deeply 5008 (25.1) 5053 (21.2) 10 061 (23.0) 
        Missing 136 (0.7) 217 (0.9) 353 (0.8) 
Natural hair colour at age 21 years 
    Black 2971 (14.9) 1394 (5.8) 4365 (10.0) 
    Dark brown 6313 (31.7) 8195 (34.3) 14 508 (33.1) 
    Light brown 7186 (36.1) 8910 (37.3) 16 096 (36.8) 
    Blond 2366 (11.9) 3723 (15.6) 6089 (13.9) 
    Red/auburn 993 (5.0) 1482 (6.2) 2475 (5.7) 
    Missing 91 (0.5) 170 (0.7) 261 (0.6) 
Eye colour 
    Blue 7370 (37.0) 7487 (31.4) 14 857 (33.9) 
    Grey 740 (3.7) 740 (3.1) 1480 (3.4) 
    Green 2384 (12.0) 3473 (14.6) 5857 (13.4) 
    Hazel 4229 (21.3) 6006 (25.2) 10 235 (23.4) 
    Brown 4730 (23.7) 5472 (22.9) 10 202 (23.3) 
    Other 105 (0.5) 197 (0.8) 302 (0.7) 
    Missing 362 (1.8) 499 (2.1) 861 (2.0) 
Freckles at age 21 years (face) 
    None 10 979 (55.1) 9465 (39.7) 20 444 (46.7) 
    A few 5756 (28.9) 7880 (33.0) 13 636 (31.1) 
    Some 2289 (11.5) 4522 (18.9) 6811 (15.6) 
    Many 774 (3.9) 1869 (7.8) 2643 (6.0) 
    Missing 122 (0.6) 138 (0.6) 260 (0.6) 
Moles at age 21 years (whole body) 
    None 6250 (31.4) 5822 (24.4) 12 072 (27.6) 
    A few 9819 (49.3) 12 815 (53.7) 22 634 (51.7) 
    Some 2748 (13.8) 3737 (15.7) 6485 (14.8) 
    Many 568 (2.9) 858 (3.6) 1426 (3.3) 
    Missing 535 (2.7) 642 (2.7) 1177 (2.7) 
Naevi—left upper arm 
    None 6118 (30.7) 6419 (26.9) 12 537 (28.6) 
    1–4 5947 (29.9) 7279 (30.5) 13 226 (30.2) 
    5–10 1946 (9.8) 2795 (11.7) 4741 (10.8) 
    >10 1592 (8.0) 2540 (10.6) 4123 (9.4) 
    Not known 346 (1.7) 218 (0.9) 564 (1.3) 
    Missing 3971 (19.9) 4623 (19.4) 8594 (19.6) 
Naevi >5 mm—left upper arm 
    None 10 628 (53.4) 12 172 (51.0) 22 800 (52.1) 
    1–4 2158 (10.8) 2872 (12.0) 5030 (11.5) 
    >5 250 (1.3) 330 (1.4) 580 (1.3) 
    Not known 282 (1.4) 168 (0.7) 450 (1.0) 
    Missing 6602 (33.1) 8332 (34.9) 14 934 (34.1) 
Variables Invited participants 
Male (n = 19 920) Female (n = 23 874) Total (n = 43 794) 
n (%) n (%) n (%) 
Age (years) 
    40–44 2192 (11.0) 3421 (14.3) 5613 (12.8) 
    45–49 2708 (13.6) 3994 (16.7) 6702 (15.3) 
    50–54 3525 (17.7) 4488 (18.8) 8013 (18.3) 
    55–59 4019 (20.2) 4625 (19.4) 8644 (19.7) 
    60–64 3847 (19.3) 3996 (16.7) 7843 (17.9) 
    65–69 3629 (18.2) 3350 (14.0) 6979 (15.9) 
Ethnicity 
    White 18 675 (93.8) 21 952 (92.0) 40 627 (92.8) 
    Black 49 (0.3) 58 (0.2) 107 (0.2) 
    Asian 324 (1.6) 503 (2.1) 827 (1.9) 
    Aboriginal/Torres Strait Islander 43 (0.2) 53 (0.2) 96 (0.2) 
    Other 574 (0.3) 69 (0.3) 132 (0.3) 
    Mixed 192 (2.9) 981 (4.1) 1555 (3.6) 
    Missing 192 (1.0) 258 (1.1) 450 (1.00) 
Highest qualification 
    No school certificate 1518 (7.6) 1905 (8.0) 3423 (7.8) 
    School certificate 2388 (12.0) 4340 (18.2) 6728 (15.4) 
    Higher school certificate 2930 (14.7) 4999 (20.9) 7929 (18.1) 
    Trade/apprenticeship 3379 (17.0) 648 (2.7) 4027 (9.2) 
    Certificate/diploma 3264 (16.4) 4855 (20.3) 8119 (18.5) 
    University degree 4693 (23.6) 5694 (23.9) 10 387 (23.7) 
    Missing 1748 (8.8) 1433 (6.0) 3181 (7.3) 
Employment 
    Full-time worker 11 566 (58.1) 7567 (31.7) 19 133 (43.7) 
    Part-time worker 1651 (8.3) 5871 (24.6) 7522 (17.2) 
    Home duties 123 (0.6) 2969 (12.4) 3090 (7.1) 
    Unemployed 393 (2.0) 288 (1.2) 681 (1.6) 
    Student 90 (0.5) 177 (0.7) 267 (0.6) 
    Retired 4719 (23.7) 4742 (19.9) 9461 (21.6) 
    Other 677 (3.4) 783 (3.3) 1460 (3.3) 
    Missing 701 (3.5) 1479 (6.2) 2180 (5.0) 
Smoking status 
    Never smoked 9607 (48.2) 14 280 (59.8) 23 887 (54.4) 
    Ex-smoker 8075 (40.5) 7314 (30.6) 15 389 (35.1) 
    Current smoker 2105 (10.6) 2152 (9.0) 4257 (9.7) 
    Missing 133 (0.7) 128 (0.5) 261 (0.6) 
BMI a year ago (kg/m2) 
    <18.5 70 (0.4) 377 (1.6) 447 (1.0) 
    18.5–24.9 4834 (24.3) 9282 (38.9) 14 116 (32.2) 
    25–29.9 9277 (46.6) 7109 (29.8) 16 386 (37.4) 
    30–34.9 3871 (19.4) 3823 (16.0) 7694 (17.6) 
    35–39.9 963 (4.8) 1453 (6.1) 2416 (5.5) 
    >40 377 (1.9) 779 (3.3) 1156 (2.6) 
    Missing 516 (2.6) 1063 (4.5) 1579 (3.6) 
Parity 
    No children  2782 (11.7)  
    One child  2436 (10.2)  
    Two children  9105 (38.1)  
    Three children  6052 (25.4)  
    Four or more children  3248 (13.5)  
    Missing  251 (1.1)  
Family history of melanoma 
    Yes 4003 (20.1) 6065 (25.4) 10 068 (23.0) 
    No 12 406 (62.3) 14 637 (61.3) 27 043 (61.8) 
    Not known 3309 (16.6) 2884 (12.1) 6193 (14.1) 
    Missing 202 (1.0) 288 (1.2) 490 (1.1) 
Skin cancers surgically removed 
    None 11 196 (56.2) 15 031 (63.0) 26 227 (59.9) 
    One 2590 (13.0) 3458 (14.5) 6048 (13.8) 
    2–10 4742 (23.8) 4536 (19.0) 9278 (21.2) 
    11–20 710 (3.6) 424 (1.8) 1134 (2.6) 
    >20 546 (2.7) 227 (1.0) 773 (1.8) 
    Not known 13 (0.1) 17 (0.1) 30 (0.1) 
    Missing 123 (0.6) 181 (0.8) 304 (0.7) 
Skin cancers/solar keratoses burned/frozen 
    None 8526 (42.8) 11 359 (47.6) 19 885 (45.4) 
    1–5 4652 (23.4) 6762 (28.3) 11 414 (26.1) 
    6–10 1957 (9.8) 2121 (8.9) 4078 (9.3) 
    11–20 1757 (8.8) 1580 (6.6) 3333 (7.6) 
    21–50 1513 (7.6) 1117 (4.7) 2630 (6.0) 
    >50 1416 (7.1) 785 (3.3) 2201 (5.0) 
    Not known 11 (0.1) 19 (0.1) 30 (0.1) 
    Missing 92 (0.5) 131 (0.6) 223 (0.5) 
Skin colour 
    Fair 11 648 (58.5) 14 318 (60.0) 25 966 (59.3) 
    Medium 6321 (31.7) 7699 (32.3) 14 020 (32.0) 
    Olive/dark 1828 (9.2) 1685 (7.1) 3513 (8.0) 
    Black 21 (0.1) 13 (0.1) 34 (0.1) 
    Missing 102 (0.5) 159 (0.7) 261 (0.6) 
Skin type 
 Skin reaction to 30 minutes midday sun 
        No burns 2122 (10.7) 1901 (8.0) 4023 (9.2) 
        Burns a little 9273 (46.6) 9567 (40.1) 18 840 (43.0) 
        Burns moderately 6314 (31.7) 8274 (34.7) 14 588 (33.3) 
        Burns badly 2107 (10.6) 3967 (16.6) 6074 (13.9) 
        Missing 104 (0.5) 165 (0.7) 269 (0.6) 
 Skin reaction to several weeks in sun 
        No tan 990 (5.0) 1857 (7.8) 2847 (6.5) 
        Tan lightly 3682 (18.5) 5480 (23.0) 9162 (20.9) 
        Tan moderately 10 104 (50.7) 11 267 (47.2) 21 371 (48.8) 
        Tan deeply 5008 (25.1) 5053 (21.2) 10 061 (23.0) 
        Missing 136 (0.7) 217 (0.9) 353 (0.8) 
Natural hair colour at age 21 years 
    Black 2971 (14.9) 1394 (5.8) 4365 (10.0) 
    Dark brown 6313 (31.7) 8195 (34.3) 14 508 (33.1) 
    Light brown 7186 (36.1) 8910 (37.3) 16 096 (36.8) 
    Blond 2366 (11.9) 3723 (15.6) 6089 (13.9) 
    Red/auburn 993 (5.0) 1482 (6.2) 2475 (5.7) 
    Missing 91 (0.5) 170 (0.7) 261 (0.6) 
Eye colour 
    Blue 7370 (37.0) 7487 (31.4) 14 857 (33.9) 
    Grey 740 (3.7) 740 (3.1) 1480 (3.4) 
    Green 2384 (12.0) 3473 (14.6) 5857 (13.4) 
    Hazel 4229 (21.3) 6006 (25.2) 10 235 (23.4) 
    Brown 4730 (23.7) 5472 (22.9) 10 202 (23.3) 
    Other 105 (0.5) 197 (0.8) 302 (0.7) 
    Missing 362 (1.8) 499 (2.1) 861 (2.0) 
Freckles at age 21 years (face) 
    None 10 979 (55.1) 9465 (39.7) 20 444 (46.7) 
    A few 5756 (28.9) 7880 (33.0) 13 636 (31.1) 
    Some 2289 (11.5) 4522 (18.9) 6811 (15.6) 
    Many 774 (3.9) 1869 (7.8) 2643 (6.0) 
    Missing 122 (0.6) 138 (0.6) 260 (0.6) 
Moles at age 21 years (whole body) 
    None 6250 (31.4) 5822 (24.4) 12 072 (27.6) 
    A few 9819 (49.3) 12 815 (53.7) 22 634 (51.7) 
    Some 2748 (13.8) 3737 (15.7) 6485 (14.8) 
    Many 568 (2.9) 858 (3.6) 1426 (3.3) 
    Missing 535 (2.7) 642 (2.7) 1177 (2.7) 
Naevi—left upper arm 
    None 6118 (30.7) 6419 (26.9) 12 537 (28.6) 
    1–4 5947 (29.9) 7279 (30.5) 13 226 (30.2) 
    5–10 1946 (9.8) 2795 (11.7) 4741 (10.8) 
    >10 1592 (8.0) 2540 (10.6) 4123 (9.4) 
    Not known 346 (1.7) 218 (0.9) 564 (1.3) 
    Missing 3971 (19.9) 4623 (19.4) 8594 (19.6) 
Naevi >5 mm—left upper arm 
    None 10 628 (53.4) 12 172 (51.0) 22 800 (52.1) 
    1–4 2158 (10.8) 2872 (12.0) 5030 (11.5) 
    >5 250 (1.3) 330 (1.4) 580 (1.3) 
    Not known 282 (1.4) 168 (0.7) 450 (1.0) 
    Missing 6602 (33.1) 8332 (34.9) 14 934 (34.1) 

How often will they be followed up?

The cohort will be followed passively through linkage of survey data from individual study participants with existing individual-level health databases. For melanoma end points (incidence, recurrence, second primary, metastasis), the primary registers with universal population coverage are the Queensland Cancer Registry and the National Cancer Statistics Clearing House (NCSCH). This linkage will be both retrospective and prospective. Australian states and territories are required by legislation to maintain a cancer registry of new cases of malignant cancer, and the NCSCH coordinates cancer statistics on a national basis. For other skin cancer end points, we will link to Medicare Australia to capture information on skin cancer treatments and prevention. Histological details of skin cancers identified through Medicare will be obtained through linkage to the pathology laboratories servicing the Queensland population. Finally, mortality data for the cohort will be obtained through linkage with the National Death Index that records the date and cause of all deaths occurring in Australia. Thus, passive follow-up of the cohort through comprehensive population registers will yield near-universal coverage for all primary end points in the study.

Currently, the study is not funded to collect biological samples; however, a proposal to collect DNA samples from participants for future genotyping studies is being prepared. We anticipate a second round of data collection to accompany sample collection. Thereafter, future collections will be targeted at nested samples to address specific additional hypotheses as they arise.

What is attrition like?

Since the beginning of recruitment, six members of the cohort have formally withdrawn, and three have died.

What has been measured?

In addition to baseline survey data (Tables 1 and 4), a repeatability and validation sub-study was undertaken on 114 participants; details have been submitted for publication separately (see summary ahead in the text). The planned timeline for follow-up through comprehensive health databases is outlined in Table 5. Planned biospecimen collection is contingent on successful applications for funding.

Table 5

Planned follow-up for the QSkin Sun and Health Cohort

Phase Measurements 
Baseline 2010-11 Survey only (self-completed). 
Sub-study (114 participants) for validation/repeatability. 
Follow-up 2012 Linkage to Queensland Cancer Registry and Medicare (MBS/PBS). 
Additional linkage to pathology registers to obtain histological data. 
Follow-up 2015 Linkage to Queensland Cancer Registry and the National Cancer Statistics Clearing House. 
Follow-up 2016 Linkage to Medicare (MBS/PBS). 
Additional linkage to pathology registers to obtain histological data. 
Follow-up 2019 Linkage to Medicare (MBS/PBS). 
Additional linkage to pathology registers to obtain histological data. 
Follow-up 2020 Linkage to Queensland Cancer Registry and the National Cancer Statistics Clearing House. 
Ongoing Linkage to the National Death Index to provide data on mortality of cohort members since baseline. 
Phase Measurements 
Baseline 2010-11 Survey only (self-completed). 
Sub-study (114 participants) for validation/repeatability. 
Follow-up 2012 Linkage to Queensland Cancer Registry and Medicare (MBS/PBS). 
Additional linkage to pathology registers to obtain histological data. 
Follow-up 2015 Linkage to Queensland Cancer Registry and the National Cancer Statistics Clearing House. 
Follow-up 2016 Linkage to Medicare (MBS/PBS). 
Additional linkage to pathology registers to obtain histological data. 
Follow-up 2019 Linkage to Medicare (MBS/PBS). 
Additional linkage to pathology registers to obtain histological data. 
Follow-up 2020 Linkage to Queensland Cancer Registry and the National Cancer Statistics Clearing House. 
Ongoing Linkage to the National Death Index to provide data on mortality of cohort members since baseline. 

What has it found? Key findings and publications

The QSkin Study completed recruitment of the cohort in December 2011. During the recruitment phase, two methodological sub-studies were performed that have been submitted for publication; full details of these and all future publications will be updated on the QSkin website (www.qskin.qimr.edu.au).

For the first sub-study, we resurveyed 114 participants at random to test the repeatability of survey items and to validate self-reported phenotypic measures against physician examinations. We observed high levels of agreement for phenotypic characteristics (e.g. eye colour κ = 0.87, 95% CI: 0.80–0.94; skin colour at age 21 years κ = 0.76, 95% CI: 0.64–0.88). Measures of past and recent sun exposure had slightly lower estimates of agreement. We found high levels of repeatability for medical and family history of skin cancer. Physician counts of naevi correlated well with categorical measures of self-reported naevus density at age 21 years, but only modestly with absolute naevus counts conducted by participants.

In the second sub-study, we tested the effectiveness of emotive leaflets included in mail-out packs to potential participants using a randomized trial design. We found no difference in participation fractions between the intervention and control groups, and conclude that emotive leaflets make no substantial difference to participation.

What are the main strengths and weaknesses?

The QSkin Study is the largest prospective study ever conducted specifically to address melanoma and skin cancer outcomes. We believe it will provide accurate and unbiased information about the risk factors for these cancers, and help to untangle the relationships between sun exposure, phenotype and the different histological types of skin cancer. The high rates of melanoma and skin cancer in the target population offer a unique opportunity to study these end points using a population-based prospective study. The study is adequately powered for melanoma, BCC and SCC as outcomes, with an expected 256 melanoma cases after 5 years of follow-up (Table 6). Another strength of the study is assurance of virtually complete follow-up for the primary end points through linkable, comprehensive health registers, even for participants otherwise lost to follow-up. The collection of biospecimens will further strengthen the study’s position as a valuable resource.

Table 6

Modelled incidence of melanoma, BCC and SCC with 2–10 years of follow-up

Year of follow-up Cumulative number of melanoma cases in the cohort Cumulative number of BCC cases in the cohort Cumulative number of SCC cases in the cohort 
97 2245 999 
148 3401 1527 
201 4651 2070 
256 5771 2619 
314 6962 3171 
373 8148 3734 
435 9328 4301 
497 10 496 4868 
10 562 11 646 5433 
Year of follow-up Cumulative number of melanoma cases in the cohort Cumulative number of BCC cases in the cohort Cumulative number of SCC cases in the cohort 
97 2245 999 
148 3401 1527 
201 4651 2070 
256 5771 2619 
314 6962 3171 
373 8148 3734 
435 9328 4301 
497 10 496 4868 
10 562 11 646 5433 

A limitation of the study is that all exposure data to date are self-reported by study participants. Although we have evidence that repeatability and validity are high to very high for most measures, we know that some items (notably naevus counts) suffer modest levels of misclassification. With unlimited resources, we would capture physician skin examinations on all participants. Another limitation is that the questionnaire is currently available only in English, limiting the participation of people from non-English speaking backgrounds. The cohort is funded until the end of 2014, which will enable several rounds of linkage with Medicare Australia and to the Cancer Registry. As with many other cohorts, the lack of long-term funding is a potential limiting factor, and survival of the cohort is dependent on the success of individual project grant applications to secure funding to conduct further record linkage and active follow-up of cohort members beyond our current grant funding period.

Can I get hold of the data? Where can I find out more?

The data are held by the QSkin Study research team. Information about study progress is available through the study website (www.qskin.qimr.edu.au). Potential collaborators should discuss ideas informally with the study investigators by email (David.Whiteman@qimr.edu.au). The exchange of ideas and proposals to add to the research are welcome. Formal approval to access data can then be obtained on application to the QSkin Study Data Access Committee, and it must be accompanied by a research plan and previous approval from the ethics committee of the applicant’s institution.

Chief Investigators: David C. Whiteman, Catherine M. Olsen and Adèle C. Green.

Associate Investigators: Rachel E. Neale and Penelope M. Webb.

Study Team: Rebekah A. Cicero, Lea M. Jackman, Suzanne M. O’Brien, Susan L. Perry and Barbara A. Ranieri.

Clinical Collaborators: Conrad J. Morze and Peter H. Soyer (University of Queensland).

Scientific Advisory Board: Professor Dallas English (University of Melbourne) and Professor Alison Venn (Menzies Research Institute, Hobart). Previously Lisa McFadyen (former CEO Melanoma Patients Australia); new appointment to Scientific Advisory Board pending.

Funding

The QSkin Study was supported by a programme grant from the National Health and Medical Research Council of Australia (no. 552429). C.O. is supported by a postdoctoral fellowship from the Xstrata Community Trust. D.W. is supported by a future fellowship from the Australian Research Council (FT0990987). A.G. is partly supported by a fellowship from the UK Medical Research Council (no. 89912). R.N. is funded by a career development fellowship from the National Health and Medical Research Council of Australia.

Acknowledgements

The QSkin Study is conducted by a team of researchers from the Queensland Institute of Medical Research. We are grateful to the National Health and Medical Research Council of Australia (NHMRC) for funding, and to the Queenslanders who have willingly given their time to take part. The authors gratefully acknowledge the valuable contributions of all staff, students and colleagues who have been associated with the project since its inception.

Conflict of interest: None declared.

Key Messages

  • The QSkin Sun and Health Study comprises a cohort of 43 794 men and women aged 40–69 years randomly sampled from the population of Queensland, Australia in 2011.

  • Passive follow-up of the cohort through comprehensive population registers will yield near-universal coverage for all episodes of melanoma and keratinocyte cancer in the study.

  • Most of the cohort is of White European ancestry (93%), with fair or medium skin colour, and a skin type that burns after long-term sun exposure; 23% reported a family history of melanoma, and 39% reported having one or more skin cancers removed surgically.

  • Although almost three quarters of participants had had their skin checked by a medical practitioner in the past 3 years, 12% of participants had not had their skin checked (by doctor, other or self) at all in that period.

References

1
AIHW
Cancer in Australia: An Overview, 2008
 , 
2008
Canberra, Australia
Australian Institute of Health and Welfare
2
Olsen
CM
Carroll
HJ
Whiteman
DC
Estimating the attributable fraction for cancer: a meta-analysis of nevi and melanoma
Cancer Prev Res (Phila Pa)
 , 
2010
, vol. 
3
 (pg. 
233
-
45
)
3
Olsen
CM
Carroll
HJ
Whiteman
DC
Familial melanoma: a meta-analysis and estimates of attributable fraction
Cancer Epidemiol Biomarkers Prev
 , 
2010
, vol. 
19
 (pg. 
65
-
73
)
4
Olsen
CM
Carroll
HJ
Whiteman
DC
Estimating the attributable fraction for melanoma: a meta-analysis of pigmentary characteristics and freckling
Int J Cancer
 , 
2010
, vol. 
127
 (pg. 
2430
-
45
)
5
Williams
PF
Olsen
CM
Hayward
NK
Whiteman
DC
Melanocortin 1 receptor and risk of cutaneous melanoma: a meta-analysis and estimates of population burden
Int J Cancer
 , 
2011
, vol. 
129
 (pg. 
1730
-
40
)
6
MacKie
RM
Freudenberger
T
Aitchison
TC
Personal risk-factor chart for cutaneous melanoma
Lancet
 , 
1989
, vol. 
2
 (pg. 
487
-
90
)
7
English
DR
Armstrong
BK
Identifying people at high risk of cutaneous malignant melanoma: results from a case-control study in western Australia
Br Med J (Clin Res Ed)
 , 
1988
, vol. 
296
 (pg. 
1285
-
88
)
8
Marrett
LD
King
WD
Walter
SD
From
L
Use of host factors to identify people at high risk for cutaneous malignant melanoma
CMAJ
 , 
1992
, vol. 
147
 (pg. 
445
-
53
)
9
Garbe
C
Buttner
P
Weiss
J
, et al.  . 
Risk factors for developing cutaneous melanoma and criteria for identifying persons at risk: multicenter case-control study of the Central Malignant Melanoma Registry of the German Dermatological Society
J Invest Dermatol
 , 
1994
, vol. 
102
 (pg. 
695
-
99
)
10
Fears
TR
Guerry
Dt
Pfeiffer
RM
, et al.  . 
Identifying individuals at high risk of melanoma: a practical predictor of absolute risk
J Clin Oncol
 , 
2006
, vol. 
24
 (pg. 
3590
-
96
)
11
Cho
E
Rosner
BA
Feskanich
D
Colditz
GA
Risk factors and individual probabilities of melanoma for Whites
J Clin Oncol
 , 
2005
, vol. 
23
 (pg. 
2669
-
75
)
12
Gies
P
Wright
J
Measured solar ultraviolet radiation exposures of outdoor workers in Queensland in the building and construction industry
Photochem Photobiol
 , 
2003
, vol. 
78
 (pg. 
342
-
48
)
13
Khoo
S
Lucas
D
Australian Census Analytic Program: Australians' Ancestries 2001
 , 
2004
Canberra, Australia
Australian Bureau of Statistics
14
 
AEC 2011.http://www.aec.gov.au/faqs/Electoral_Roll.htm (10th October 2011, date last accessed)
15
ABS. Cat. 1367
State and Territory Statistical Indicators. 2012
 
Canberra, Australia
Australian Bureau of Statistics
16
Queensland Health
Self Reported Health Status 2011
 , 
2011
Brisbane, Queensland, Australia
Queensland Health