Development and validation of sunlight exposure measurement questionnaire (SEM-Q) for use in adult population residing in Pakistan

In recent years, exposure to sunlight has been a subject of epidemiological interest both due to its beneficial as well as adverse effects on human health. It is an important etiological factor in the development of skin cancers and sun-related eye problems when exposure is high, or Vitamin D Deficiency (VDD) [1, 2]. However, the exposure to solar ultraviolet radiation (UVR) is an essential step for the production of Vitamin D, and also the main source of vitamin D in human body [3, 4]. Ultraviolet B (UVB) rays in the solar UVR spectrum produce vitamin D in the human body. UVB rays penetrate uncovered skin and converts cutaneous 7-dehydrocholesterol to pre-vitamin D₃, which in turn becomes vitamin D₃[5, 6]. The enteral route is not considered a good source of vitamin D as foods of animal origins e.g. fish, beef, eggs, milk, cheese etc., generally lack the required amount of vitamin D unless fortified [7].

VDD used to be considered rare in those parts of the world that had plenty of sunshine all year round but WHO now estimates that globally one billion people have VDD or insufficiency [1, 8]. Geographically, Pakistan lies in a region with adequate sunlight throughout the year [9] with UV index ranging between 5 and 11 during different seasons. However, despite its geographical location, the prevalence of VDD is reported to be high, ranging from 20% to 83% [10‐12].

A total of fifty four participants were enrolled in the study. Among the three exposure groups taken, low exposure group accounted for 35.2% of the total sample (n = 19), moderate for 33.2% (n = 18) and high exposure group for 31.5% (n = 17). Male participants constituted 53.7% of the sample. 83.3% of the total participants were literate with 79.6% of the study participants having 10 years or more of formal education. Predominant language spoken was Urdu followed by Punjabi. Muslims were 75.9% of the study participants. Table 3 shows the descriptive analysis of the study participants.

Serum 25(OH) vitamin D₃ levels were assessed for 50 study participants. 98% (n = 49) of the participants were either vitamin D deficient or insufficient. The mean vitamin D level (SD) of our study participants was 12 ng/ml (5.9), with levels of 9.8(4.7), 11(4.6) and 17(6.5) for low, moderate and high SE groups respectively. Table 4 shows the distribution of main variables according to the three SE groups.

The correlation coefficient between average time (minutes) spent outdoors over the four days for ST SEM-Q, and average readings of UV dosimeter over four days was 0.60 (P < 0.01). Spearman’s rank correlation between average score of ST SEM-Q and serum vitamin D levels was found to be 0.36 (P = 0.01), while Pearson’s correlation between serum vitamin D levels and average LT SEM-Q score in summers was 0.43 (P = 0.01) and 0.48 (P = 0.01) in winters. Table 5 shows the correlations among the main variables of interest.

Correlation between LT and ST questionnaires was assessed. It was found that the correlation coefficient between average time (minutes) spent outdoors as captured by LT and ST was 0.85 (P < 0.01). Similarly, the correlation coefficient between the scores for LT and ST SEM-Q was observed to be 0.82 (P < 0.01).

We found a good correlation between our ST SEM-Q and dosimeter readings (r_s = 0.60 (p < 0.01)) and fair correlation of 0.36 (P = 0.01) between ST SEM-Q and serum vitamin D levels. Similarly, a good correlation of 0.58 (P = 0.01) and 0.60 (P = 0.01) was observed between dosimeter readings and LT SEM-Q summer and winters respectively. The correlation between vitamin D levels and LT SEM-Q score for previous year summer and winter season was found to be 0.43 (P = 0.01) and 0.48 (P = 0.01) respectively.

Through recent literature search, this study is the first of its kind in South Asian region to develop a questionnaire to assess long and short-term sunlight exposure and validate it against a gold standard objective measure using dosimeter badges. It is also the first study to assess the correlation between the SEM-Q and serum vitamin D levels in Pakistan. Our research strength is that we developed the questionnaire based on the factors that are of importance in a non-western society. It is also a culturally sensitive and acceptable questionnaire in our setting.

The correlation we found between our ST-SEMQ questionnaire and the dosimeter readings (r_s = 0.60; P < 0.01) were higher than the Australian pilot study, which compared the objective measurement and self-reported ultraviolet radiation exposure during outdoor activities among children and mother pairs and life guards (r = 0.32) [29]. The main Australian study later showed fair to good correlations between the sunlight diaries and dosimeters for lifeguards (r = 0.38–0.57), parents (r = 0.28–0.29) and children (r = 0.18–0.34) and these findings were still lower than our study findings [30]. Similarly, our results were higher than previously published data on mothers and children [31], school children [32] and volunteers recruited from recreational organizations [25]. A possible reason for better correlation of our questionnaire could be the comprehensive nature of the questionnaire and inclusion of items relevant for our population. Moreover, our study population consisted of adults only, working for a university, which could be a reason for better response. Also, most of the studies done previously had included factors in relation to skin cancer development and had taken the time of day when the ultraviolet radiations are the most intense (11 am to 3 pm) as reference period. We included larger time of sun exposure, which most of the other studies had not taken into consideration.

The results of our study were comparable to some of the other studies that compared personal UVR exposures using PS dosimeters and diary or questionnaire entries among adult volunteers (r = 0.69) [33] and school children (r = 0.68) [34].

Our correlation between serum Vitamin D and average score for ST SEM-Q was found to be 0.36 (P = 0.016), 0.43 (P = 0.01) for LT SEM-Q score in summers and 0.48 (P = 0.01) for LT SEM-Q score in winters. We did not come across any study that developed scoring for sun exposure questionnaires to assess the correlation between the scores and vitamin D levels hence we are unable to compare our results with any other study. One study that did measure Vitamin D levels and related it to UVR behavior, correlated the mean daily hours measured by dosimeter in September and Feburary (r = 0.64, P = 0.001) and (r = 0.53; P = 0.007) [35], however, the study did not look into correlation between Vitamin D and scores for sunlight diaries or questionnaires.

There were a few limitations to our study. The fair correlations of 0.34, 0.43 and 0.48 observed between the serum vitamin D levels and scores for SEM-Qs were probably because of the smaller sample size and hence the inability to find a variation in terms of vitamin D levels, as majority of study sample was vitamin D insufficient or deficient. Even though our estimated sample size was larger (n = 203), due to budget constraints and time limitations, the study was conducted on approximately 25% of the sample size. We expect that with the larger sample, we would have achieved a larger variation in terms of vitamin D levels and probably a better correlation with the computed score. However, the sample size of 54 with a post hoc power of 83% was sufficient for meeting the main objective of our study.

We developed and validated two questionnaires i.e. ST and LT SEM-Q, in order to measure and assess both short term sunlight exposure and long term SE over previous 1 year period. Our study assessed sunlight exposure in context with Vitamin D levels, unlike the recent other studies that have assessed sunlight exposure in relation to skin cancer [36]. Both the questionnaires showed good correlation with dosimteres and fair correlations with serum vitamin D levels and, therefore, we can say that both the questionnaires can be used to assess the SE in adults. However, LT SEM-Q might be preferred due to convenience of its use and reflection of a longer duration of sun light exposure inadequacy which may be more relevant for epidemiological work.

The UV dosimeters, although gold standard for measuring UVR, have a few limitations which makes the questionnaires preferred source for assessing sunlight exposure. UV dosimeters are expensive tools and are not readily available in developing countries and hence need to be imported for use. Secondly, the calibration and reading of dosimeters is a complex process and require sensitive equipment and expertise, which is not easily available especially in our settings. Lastly, the dosimeters measure the current exposure to sunlight and do not take into account the long term exposure to sunlight, which is an important factor for vitamin D levels in the body.

The main outcome of this study was development and validation of a culturally-appropriate and inexpensive tool for assessing sunlight exposure. We found a high correlation between self-reported sun exposure using our tool and the objective measure of sun exposure using the dosimeters and, therefore, SEM-Qs can be used in population based studies for assessing SE. This has great epidemiological value for work in the related vitamin D sunlight exposure measurement and for assessing SE in relation with other disease outcomes such as skin cancers.

Furthermore, this is the first study in our region for assessing the exposure to sunlight through a questionnaire, and which incorporates methods for scoring based on the factors that affect the UVB rays absorption in the body, factors that may directly affect the amount of pre-vitamin D₃ synthesis in the skin.