The results of our nationwide study generally reinforce the findings of the NSW statewide survey , but further studies in other countries would be useful to assess the need for guidelines and their evaluation. We were able to substantially extend the NSW findings by including in our analyses consideration of potential latitude band effects and investigation of plausible statistical predictors of confidence about vitamin D knowledge and the provision of recommended advice about sun protection and vitamin D, both for the general population and those at increased risk of skin cancer or vitamin D deficiency. We also investigated statistical predictors of agreement/disagreement with a range of statements relating to perceptions regarding vitamin D and skin cancer.
Reading of specific information sources and confidence about vitamin D knowledge
Overall, 20% of GPs reported having read the CSNZ position statement , similar to the 24% in NSW who had read comparable Australian guidelines . However, our finding that almost five times more NZ than NSW GPs reported being ‘not at all confident’ about their vitamin D knowledge indicates that lack of confidence is more pressing in NZ. Practising in a major metropolitan centre with a medical school was positively associated with confidence, giving support to the hypothesis that such locations may provide better access to educational opportunities. Both NZ and NSW GP’s almost unanimously agreed that clear clinical guidelines about vitamin D ‘would be useful’, providing a very clear indication of what they wanted. A NZ Consensus Statement on Vitamin D and Sun Exposure has since been published , in part, as a response to preliminary findings of the present study, thereby providing future opportunities to assess whether or not such a resource is associated with change in confidence and the advice provided.
Estimated summer sun exposure times required to achieve adequate vitamin D
For patients with high sun sensitivity (defined as Fitzpatrick Skin Types I & II), the mean summer sun exposure time of the unprotected face, hands and arms that GPs perceived would be required (before 10 am, after which hour current NZ guidelines recommended routine sun protection when the UVI is ≥ 3) in order to obtain adequate vitamin D was approximately 15 minutes. Assuming that the UVI was no higher than 3, sufficient vitamin D should be able to be produced in that period while erythema could be avoided . Such a mean exposure time may, broadly, be considered compatible with the recommendation of ‘a few minutes of sunlight on either side of the peak UVR periods’ . For those with low sun sensitivity (defined as Skin Types V-VI) the perceived mean time was considerably longer (27 minutes), consistent with the longer exposure time required for darker skin types. However, for both skin type groups the reported range was wide, with some GPs providing estimates of 2 hours and 4 hours as appropriate for those of low and high sun sensitivity, respectively. This should be of concern and more conservative estimates should be a target for information strategies to achieve. It was not possible to compare these NZ estimates with those reported for NSW as in that study the estimated period of exposure was during peak UVR , a behaviour incompatible with existing NZ recommendations not to seek ‘deliberate exposure at peak UVR times’ . Even so, 22% of Australian GPs were reported as believing 30 minutes during peak UVR would be required for a person of average sun sensitivity to achieve adequate vitamin D, whereas the Australian guidelines indicate that only 6 to 8 minutes would be required at 10 am in Sydney during summer. While acknowledging the challenges, we agree that there is ‘a need for an easier and quicker way for doctors to calculate safe UV exposure and for determining risk status to help them provide tailored advice’, such as ‘desktop decision aids, with computer algorithms that take into account the complexities of skin type, weather and location’ .
Sun protection/exposure advice
Most NZ GPs (70%) and more than in NSW (55%) advised the currently recommended summer sun protection strategy (‘to use sun protection at all times during peak UV’), although 17% (NSW 26%) recommended sun protection ‘most of the time during peak UV, but to receive some direct sunlight during that time.’ Winter advice was less restrictive, with 36% (NSW 33%) advising patients ‘not to use sun protection outside of peak UV times and receive direct sunlight during this time’, and 13% (2% in NSW) advising patients ‘not to use any sun protection at any time’ during winter. For patients ‘at increased risk of vitamin D deficiency’ the latter advice was more commonly provided than for the general NZ population (24% vs 13%), otherwise the winter advice provided by NZ GPs to these groups did not differ markedly.
When asked ‘How much information about vitamin D have you received in the last 12 months?’ almost the same percentages in NZ (45%) as NSW (46%) indicated ‘more than usual.’ This was despite the surveys having been conducted in different years (NSW: Aug-Dec. 2009; NZ: Oct-Nov. 2010), indicating virtually no difference in perceptions of the balance of information available to GP’s during the two time periods in the two geographical areas. Most NZ GP’s (59%) had not changed their advice as a result of information received during the past 12 months, which was consistent with, but less stable than found in NSW (68%). We found a somewhat stronger shift in NZ than NSW towards recommending less winter protection (29% vs 20%) and, furthermore, 10% reported a shift towards recommending less protection all year round (a response option not reported for NSW).
Reading of information sources and other factors associated with confidence about vitamin D knowledge and ‘quality’ of advice
We investigated plausible statistical predictors of (1) confidence about vitamin D knowledge and (2) the provision of ‘correct’ advice. GPs who trained outside NZ/Australian/other ‘western’ centres were more confident about their vitamin D knowledge, but less likely to advise routine sun protection at times of high summer UVR. This finding is consistent with possibly less awareness about the seasonally extreme UVR levels in NZ, which can be almost 50% higher than at comparable northern hemisphere latitudes in summer . The provision of specific information about this significant difference may, therefore, be of value during on-going clinical education. However, among all participants, completion of a skin cancer training course was associated neither with confidence nor provision of ‘quality’ advice. GPs in practice longer were also less likely to advocate sun protection at times of high UVR in summer months, consistent with possibly increasingly emphasis on protective strategies during recent medical training regimes. This would seem to reinforce the need for training updates in skin cancer prevention. When we examined latitude gradient of Medical Council Register address against confidence in knowledge about vitamin D, the lowest confidence levels were in the 40-42°S latitude band which includes the capital city, Wellington (North Island) and the Nelson region (northern South Island). There seems no clear explanation for this association with proximity to the national political capital. In a multivariable context, although the reading of some specific information sources was positively associated with confidence and lower perceived exposure times for sufficient vitamin D in summer among those with high sun sensitivity, there was no association with the categorical ‘correctness’ of advice provided. These findings provide partial confirmation of the potential value of such resources.
Responses to statement matrix and expression of greater concern about vitamin D than skin cancer
Responses to the matrix of statements (Table 4) indicate widespread concerns about vitamin D deficiency and the potentially negative impact of skin cancer prevention messages on vitamin D status. A very similar pattern was found in NSW, although more NZ (81%) than NSW (68%) GPs indicated agreement with the statement that ‘skin cancer prevention messages’ ‘contribute to the development of vitamin D deficiency.’ Bonevski et al.  found that females were more likely to express such concerns and we confirmed this among NZ GPs for two of the three related questionnaire items. However, we also found evidence of statistically significant associations with having been trained in S.E. Asia, residence in a city with a medical school, being in practice for a greater number of years and expressing greater confidence about vitamin D knowledge. Consideration should be given to each of these factors when targeting educational interventions.
The potential for these perceptions to undermine appropriate sun protection messages in the context of sometimes extreme NZ summer UVR levels should be of concern. As Bonevski et al. note, ‘although vitamin D plays an important role in bone health, the evidence regarding the other health benefits of vitamin D remains inconclusive’  and this situation continues . Nevertheless, to pre-empt perceptions of division within the scientific community and inconsistency in public health messages, it remains important for sun protection messages to take into account vitamin D issues, in particular, known lower vitamin D levels in NZ associated with living at high latitude, non-European ethnicity and more highly pigmented skin . This may be challenging, but follows a predicted pattern of the need to develop more targeted messages for specific population groups , and reinforces the call for decision aids .
Study strengths and limitations
Our procedures meant that we were unable to employ all of the recommended strategies for improving response rates in surveys of physicians, in particular, the use of financial incentives . However, our national survey drew on the two most relevant professional organisations (RNZCGP and MCNZ) to provide contact with GPs, and the participation rate in our study was 32%, slightly higher than obtained for the comparable study in the Australian state of New South Wales which used financial incentives . Those authors argue that their reported level of response was comparable to other practitioner surveys and the literature suggests that there may be only a weak association between response rates and bias, if any . Participants in our survey differed from the national population of practitioners in terms of higher participation by females than males and part-time than full time GPs, similar to the Australian findings . Other demographic data were not accessible to permit valid comparisons between respondents and the NZ practitioner population. These factors should be taken into account when extrapolating to the GP population. There were similarities and differences between the NZ national and NSW state samples. Somewhat fewer in NZ had been practising for longer than 20 years, in both cases most had graduated within the country of survey, but many more in NZ than Australia reported a college fellowship as their highest medical qualification, more had received their highest medical qualification since 2000 and many fewer before 1980; 17% had completed a skin cancer course whereas only 9% of the NSW sample had either enrolled in or completed such a course. Many fewer NZ GPs practiced fulltime, although the definition of full time employment differed, in NZ being based on the reported number of sessions worked, but dichotomously self-reported in Australia.
The results of our study add substantially to knowledge, not only largely confirming the findings of similar prior NSW research, but also extending that research by investigating, in a multivariable context, factors associated with GPs’ perceptions and the advice they provide. These multivariable analyses allowed us to identify some significant differences in perceptions and advice according to gender, location of medical graduation, number of years in practice, confidence about vitamin D knowledge, residence in a city with a medical school and information sources read. Among the other strengths of our study was the inclusion of randomisation in the order of presentation of lists of response options, something which was not reported for the NSW survey, but which adds confidence to the findings. The nation-wide reach of our study also permitted investigation of potential latitudinal differences in perceptions and advice. That none were found indicates homogeneity, although there is justification for some variation, given significant regional differences in seasonal UVR levels.