321 BMI and sun exposure is associated with increased risk of non-melanoma skin cancer: A prospective study from the Women’s Health Initiative J Noguti 1 , AA Chan 1 , Y Pak 2 , RT Chlebowski 3 and DJ Lee 4 1 Dermatologya/Immunology, Torrance, CA, 2 Los Angeles Biomedical Research Institute, Torrance, CA, 3 Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, Torrance, CA and 4 Dermatology/Immunology, Torrance, CA Incidence of nonmelanoma skin cancer (NMSC) exceeds all other types of cancer combined. Cumulative and intermittent sun exposure are known risk factors for the development of NMSC. Since obesity has been shown to contribute to cancer incidence, we hypothesized that heavier individuals (BMI > 25 kg/m 2 ) have increased the risk of NMSC when accounting for sun exposure. Using the Womens Health Initiative (WHI) cohort, we investigated the risk of NMSC with sun exposure and anthropometric measures. We analyzed the incidence of NMSC associated with sun exposure: reported duration (behavioral) and intensity (geographic location), as well as two anthropometric measures: body mass index (BMI) and waist to hip ratio (WHR). We tested whether the effect of sun exposure on NMSC depended on anthro- pometric measures by a two-way interaction table. From 66,822 postmenopausal women eligible for inclusion in this study, 9,317 (13.9%) participants developed NMSC. We found that the more time spent outdoors, the higher the risk of NMSC and subjects in geographic locations with higher sun intensity had a higher risk of NMSC. While BMI > 25 kg/m 2 showed lower risk of NMSC by univariable analysis, we found that the risk of NMSC for women with BMI > 25 kg/m 2 vs. normal weight increases with increased sun exposure (pval: 0.024). Therefore, the risk of NMSC from sun exposure is affected by BMI. 322 Skin cancer risk among adults with atopic dermatitis exposed to topical calcineurin inhibitors J Feng 1 , A Tsai 2 , D Raimondo 3 , CP Quesenberry 2 and M Asgari 4 1 Dermatology, Massa- chusetts General Hospital, Boston, MA, 2 Kaiser Permanente Northern California, Oakland, CA, 3 Valeant Pharmaceuticals International, Bridgewater, NJ and 4 Massachusetts General Hospital, Boston, MA Atopic dermatitis (AD), a chronic inflammatory skin disease, affects between 2-10% of the adult population in industrialized nations. AD is primarily treated with topical emollients and anti-inflammatory agents such as topical corticosteroids (TCS) and topical calcineurin in- hibitors (TCI), though more severe cases require systemic immunosuppressive agents. TCIs currently carry a black box warning label for a possible association with cutaneous malig- nancy, including melanoma and keratinocyte carcinomas (KCs), defined as basal cell carci- noma (BCC) and squamous cell carcinoma (SCC). The relationship between TCI exposure and skin cancer remains poorly understood as large-scale post-marketing surveillance studies in real-world settings are lacking. We will examine the association between TCI exposure and the development of cutaneous melanoma and KCs in adults age 40 with a physician- rendered AD diagnosis in the Kaiser Permanente Northern California health plan(N¼96,672). Electronic pharmacy data was used to determine pharmacy-dispensed TCI exposure (n¼1,538). Electronic pathology records were used to identify incident melanoma (n¼395) and KCs (n¼7,218) from cohort entry (January 1, 2002) until the end of the study period (December 31, 2017). Cox proportional hazards modelling will estimate melanoma and KC hazard ratios and 95% confidence interval. The primary objective of this study is to assess the risk of developing KC overall and by type (BCC or SCC), in adult subjects aged 40 years with AD exposed to TCIs as compared to those exposed to TCS and untreated AD subjects. The secondary objective is to assess TCI dose, intensity and duration of use in relation to mela- noma and KC risk. The unique study setting of an integrated healthcare delivery system allows for complete capture of exposure, outcome, and control of screening bias, enabling an in- depth exploration of the association of TCI exposure with risk of melanoma and KCs. 323 Correlates of low sun-protection factor sunscreen users in 2000-2015: A population based study B Kahn 1 , SC Chen 2 and H Yeung 3 1 Department of Dermatology, Emory University School of Medicine, Atlanta, GA, 2 Dep. of Dermatology, Emory University School of Medicine; Div. of Dermatology, Atlanta VAMC, Atlanta, GA and 3 Department of Dermatology, Emory University, Atlanta, GA Frequent use of sun-protection factor (SPF) >15 sunscreen is an important skin cancer pre- vention strategy, but little is known about users of SPF <15 sunscreen. We correlated soci- odemographic characteristics and sun protective behaviors of sunscreen users age 18+ who report usual use of >15 SPF versus <15 SPF sunscreens in the nationally representative, population-based, National Health Interview Surveys, 2000-2015. Among 93,250 adult re- spondents who report any sunscreen use, 4,772 (5.1%) reported usual use of SPF <15 sun- screens while 78,429 (84.1%) used SPF >15 sunscreens. SPF <15 sunscreen users had significantly lower educational attainment, were more likely to smoke, binge-drink, and lack health insurance, but less likely to be overweight or obese or have a personal history of skin cancer (each P < 0.05). SPF <15 sunscreen users reported significantly lower prevalence of frequent sunscreen use (39.5% vs. 55.2%), seeking shade (20.8% vs. 35.7%), wearing long- sleeved shirts (9.0% vs. 11.7%), wearing long pants (18.6% vs. 26.6%), wearing a wide- brimmed hat (16.2% vs. 20.0%), wearing a cap/visor (23.9% vs. 35.1%), and having ever received a skin cancer screening exam in their lifetime (19.4% vs 25.9%) or in the past 12 months (8.7% vs. 12.8%, each Bonferroni-corrected P < 0.001). SPF <15 sunscreen users reported significantly higher prevalence of indoor tanning (23.8% vs. 13.7%) and frequent indoor tanning (21.4% vs. 11.8%, each Bonferroni-corrected P < 0.001) in the past 12 months. Prevalence of sunburns was not significantly different between SPF <15 and SPF >15 sunscreen users. Usual use of SPF <15 sunscreen is a novel marker for high behavioral risks for skin cancers. Targeted education on skin cancer risks from indoor and outdoor tanning among SPF <15 sunscreen users is warranted. 324 Opportunities for improving keratinocyte cancer care in primary and specialist care in the Netherlands: A retrospective descriptive cohort study M Wakkee 1 , S van Egmond 1 , K Ramdas 1 , P Bindels 2 , J van der Lei 3 , M Louwman 4 , T Nijsten 1 and L Hollestein 1 1 Dermatology, Erasmus MC, Rotterdam, Zuid-Holland, Netherlands, 2 General Practice, Erasmus MC, Rotterdam, Zuid-Holland, Netherlands, 3 Medical Infor- matics, Erasmus MC, Rotterdam, Zuid-Holland, Netherlands and 4 Research, Netherlands Comprehensive Cancer Center, Utrecht, Netherlands Background: The high keratinocyte carcinoma (KC) incidence in Western countries puts pressure on healthcare systems and urges to identify areas for improvement. Aim: To describe current KC care based on routinely collected data in primary and secondary care in the Netherlands. Methods: A retrospective cohort study using the Integrated Primary Care In- formation database and for specialist care the nationwide Netherlands Cancer Registry. In primary care 1736 records of patients with suspicious KC lesions were included (2009-2013). In specialist care 1006 confirmed KC cases were included (2010-2015). Results: Primary care: 1373 patients were seen by the GP and of 363 patients only a letter of the medical specialist was retrieved from the GP records. KC was included in the differential diagnosis in 63% (877/ 1373) of patients seen by the GP. If the GP suspected KC, the GP referred the patient in 58% (511/877), took a biopsy in 9% (80/877), performed a direct diagnostic excision in 13% (110/ 877), performed a wait and see policy in 11% (99/877) and used another type of treatment in 9% (77/877). Specialist care: the medical specialist treated KC in 92% (923/1006) by exci- sion, 3% (30/1006) by Mohs surgery, 5% (53/1006) by another treatment. Two years after diagnosis, 27% (157/534) of squamous cell carcinoma patients did not visit the dermatologist. Although follow-up is not recommended for low risk basal cell carcinoma, 74% (65/88) received follow-up care. Conclusion: In primary care strengthening the diagnostic pathway for KC and in secondary care optimizing follow-up schedules seem potential areas for improving KC care. 325 Kaposis sarcoma severity, treatment and survival in a large community-based HIV health care network in Kenya E Freeman 1 , N Busakhala 2 , F Asirwa 2 , S Regan 3 , M Wenger 4 , K Chelidze 3 , A Semeere 5 , D Seth 3 , K Wools-Kaloustian 6 , I Bassett 3 and J Martin 7 1 Massachusetts General Hospital (MGH), Boston, MA, 2 Moi University, Eldoret, Rift Valley, Kenya, 3 MGH, Boston, MA, 4 U. California San Francisco (UCSF), San Francisco, CA, 5 Infectious Diseases Institute, Kampala, Uganda, 6 Indiana U., Indianapolis, IN and 7 UCSF, San Francisco, CA Background: Kaposis sarcoma remains highly incident in sub-Saharan Africa with poor sur- vival, despite antiretroviral therapy (ART). Our goal was to assess KS epidemiology in the modern African setting, including stage at diagnosis, real-world use of chemotherapy, and the relationship between treatment and survival.Methods: We identified all patients newly diagnosed with HIV-related KS from 2009-2012 in the 50+ clinic AMPATH network in Kenya. Through chart review, we ascertained disease severity at diagnosis and treatment. Chemo- therapy indications were AIDS Clinical Trial Group T1 stage and/or severe disease from WHO KS guidelines. Survival was determined by medical record review and patient tracking.- Results: Of 674 patients diagnosed with KS, 587 had evaluable charts. Median age was 35, 61% were male, and median CD4 was 158 cells/ml. At diagnosis, most patients (58%) had severe disease with at least one chemotherapy indication. 92% of patients received ART. Within 1 year of KS diagnosis, 48% received chemotherapy. Median time to first chemo- therapy was 28 days. Cumulative mortality at 1 year, with loss to follow-up as a competing event, was 39% for patients who had a chemotherapy indication at diagnosis but never received chemotherapy, 31% for patients who received chemotherapy, and 28% for those with mild disease who did not have a chemotherapy indication. Receiving chemotherapy was associated with lower mortality (hazard ratio 0.75, 95% CI 0.57-0.99). Conclusion: KS pa- tients in Kenya have high mortality, with the majority diagnosedwith severe disease. Patients with chemotherapy indications who did not receive chemotherapy fared worst, suggesting that increasing access to chemotherapy could be beneficial in this setting. 326 Utilization and cost of actinic keratosis destruction in Medicare Part B fee-for- service beneficiaries in 1998-2016 H Yeung 1 , MLH Baranowski 2 , RA Swerlick 3 , SC Chen 4 , J Hemingway 5 , DR Hughes 5 and R Duszak 6 1 Department of Dermatology, Emory University, Atlanta, GA, 2 Emory Uni- versity School of Medicine, Atlanta, GA, 3 Dept. of Dermatology, Emory University School of Medicine; Div. of Dermatology, Atlanta VAMC, Atlanta, GA, 4 Dep. of Dermatology, Emory University School of Medicine; Div. of Dermatology, Atlanta VAMC, Atlanta, GA, 5 Harvey L. Neiman Health Policy Institute, Reston, VA and 6 Department of Radiology, Emory University School of Medicine, Atlanta, GA Actinic keratosis has the potential to progress to keratinocyte carcinoma and was estimated to impose $1.6 billion in medical costs in the United States in 2013. The changing incidence of actinic keratosis and its associated treatment burden and cost are not well understood. We aimed to evaluate temporal trends in actinic keratosis destructions in the Medicare Part B fee -for-service population and associated treatment costs. Actinic keratosis destruction pro- cedures were aggregated from provider billing claims using Healthcare Common Procedure Coding System codes 17000-17004 in the Medicare Physician/Supplier Procedure Summary Master Files 1998-2016. Treatment prices were estimated from Medicare national non-facility allowable fees and adjusted to 2017 dollar values. Over 16.5 million actinic keratosis lesions were treated in 1998, increasing to 29.7 million in 2007 and 37.1 million in 2016. The average number of treated actinic keratosis lesions nearly doubled from 545 to 1,079 lesions per 1,000 Medicare Part B fee -for-service beneficiaries from 1998 to 2016, while treatment costs plateaued from $13,330 to $17,908 per 1,000 beneficiaries due to successive price cuts. The proportion of actinic keratosis lesions treated by independently billing non-physician providers increased from 0.1% in 1998 to 4.0% in 2007 to 15.3% in 2016. Actinic keratosis imposes high incidence and increasing levels of treatment burden in the aging U.S. popu- lation. Fee -for-service price cuts have been used to control rising treatment costs. Critical research is warranted to optimize access and value for actinic keratosis treatment and kera- tinocyte carcinoma prevention. Clinical Research: Epidemiology of Skin Diseases | ABSTRACTS www.jidonline.org S55