The group of FiLWHEL study investigators is composed of professionals in the fields of medicine, nutrition, and epidemiology. In addition, the study has Filipino assistant researchers and volunteer interviewers acquainted with the ways and common lifestyle practices of the Filipinos and who could communicate with the participants without any language barrier.

The first aim of our study is to explore how health-related behaviors, including diet, metabolic factors, quality of life factors such as physical, mental, and social domains, and acculturation influence the progression of cardiovascular disease and diabetes among Filipino women in Korea. Secondly, we examine how lifestyle, environment, and genetic factors play a role in the health of Filipino women.

FiLWHEL is a cohort of Filipino women married to Korean men in Korea.

Study locations included Seoul, Incheon, Daejeon and Nonsan, and parts of Gyeonggi Province (Fig. 1). We administered surveys at universities, university hospitals, community centers, and churches. Study enrollment started in March 2014, and it is ongoing. Our target number of participants is 500 Filipino women. A Filipino woman is eligible to join if she has ever been married to a Korean man. Active Filipino community leaders played a key role in the recruitment of participants. They disseminated information at their respective areas, encouraged their members to participate, and arranged the venue for the study. In addition, participants were also recruited through personal contacts and social media. To gather more participants, the survey team, including physicians, phlebotomist, interviewers, and other staff, visited several coordinated locations. Prior to the on-site visit, interested participants were contacted and given specific instructions so they could prepare themselves and bring things needed for the interview (e.g. medicine or/and supplement bottles). Detailed instructions about fasting were also given. Those who agreed to participate in our study signed the informed consent and were enrolled. This study was approved by the Institutional Review Board of Sookmyung Women's University (reference number SMWU-1311-BR-012). The data collection had three parts: answering of the questionnaire, physical examination, and biospecimen collection (Fig. 2). Total time spent in all procedures was around 3 hours.

We asked questions on demographic, socioeconomic, acculturation, quality of life, mental health, and other health-related factors (Table 1). The health-related questions included weight and body size history, nutritional behavior and dietary intake, smoking, physical activity, sleeping pattern, cellular phone and social networking use, medical history comprising gynecologic and obstetrical history as well as past or present medication use, family medical history, and sun exposure. Furthermore, our questionnaire used validated instruments: short format of the International Physical Activity Questionnaire (IPAQ) [31], World Health Organization Quality of Life-BREF (WHOQOL-BREF) [32], and EQ-5D-3L health questionnaire [33]. We assessed mental health by using the DSM-5 self-rated level 1 cross-cutting symptom measure [34]. Additionally, we asked questions regarding the participants' children's medical history and health behavior, including birth history (birth weight and length as well as presence of infections, anomalies, or abnormalities), after birth medical history, vaccination record, dietary habits and supplement use, use of video games and/or television, and physical activity.

All parts of the questionnaire were written in English. Due to its length, we gave participants several options so they could answer the questionnaire at their convenience. Prior to visiting the site, they can opt to answer the questionnaire online, receive a printed form, phone interview, or just answer it on site. Those who answered on site were interviewed, and some parts were self-administered and if needed, assistance was given. Most participants answered questions through on site face-to-face interviews. Answers were double-checked and if there were any discrepancies, we asked it on site or over the phone.

Dietary assessment was conducted through an in-person interview using the 24-hour recall method. The participants provided a detailed description of what they ate and drank on the last full day (midnight to midnight) before the interview. Information gathered included place and time a particular meal was eaten, types and amounts of all foods and beverages, meal composition and, if obtainable, its recipe. Further, information on nutritional supplements or medicine was obtained. The amount consumed was quantified by means of food miniatures, photographs, household measures, weight/volume, and standard units and portions. Each participant was interviewed for approximately 30 minutes. Participants were encouraged to provide the correct estimation of all consumed foods or beverages, if possible. Collected data were edited for completeness, coded, and entered using the CAN-pro 4.0 (Computer Aided Analysis Program 4.0 for professionals, Korean Society of Nutrition, Seoul, Korea). If the food item, in particular Filipino food item, was not included or the nutrient content was different in the existing software, new food codes were generated. The nutrient information of the new food codes were taken directly from the manufacturer's nutritional information.

Using a non-stretchable tape measure, height and circumferences of waist and hip were measured to the nearest 0.1cm with the participant standing. Waist circumference was measured at the midpoint, between the lower border of the rib cage and the iliac crest, and hip circumference was measured at the largest circumference. Body composition was measured by the bioelectric impedance analysis machine (InBody 620, Biospace Co. Ltd, Seoul, Korea). The participants were barefoot, wore light clothing, and accessories were removed. The machine measured body fat mass, soft lean mass, fat free mass, skeletal muscle mass, and percent body fat.

Blood pressure (BP) was measured using a sphygmomanometer in accordance with the standard BP measurement principles while the participant was seated calmly at least after a 5-minute rest. Two BP readings were obtained using the same arm with a 5-minute interval after the first measurement.

Collected biospecimens included toenail clippings and fasting blood samples. We asked the participants to clip their toenails using stainless steel clippers and place them in small envelopes. However, if toenails were polished, we instructed them, if possible, to remove the polish at home and save their toenail clippings every time they clip and send them to us within 3 months (from the study entry) by using our prepared envelopes that had return stamps and our office address. We stored their toenails at room temperature for future mineral analysis. For the blood sample, we asked them to fast for more than 8 hours, and fasting was carefully defined as not including food and beverages such as coffee or tea. In addition, we instructed them to refrain from smoking within the fasting period as well as alcohol intake 2 days before the blood sample collection. Blood samples were drawn through a venipuncture procedure, and a maximum amount of 19 ml was taken from each participant. Whole blood samples together with the centrifuged blood serum were stored at 2-8℃ until they were transferred to the laboratory for analysis and aliquoting (Seegene Medical Foundation, Korea). All procedures were conducted within 24 hours from blood collection. Aliquots of serum and plasma and DNAs were stored at -86℃. The analyzed serum biomarkers include creatinine, fasting blood glucose (FBS), aspartate transaminase/serum glutamic oxaloacetic transaminase (AST/SGOT), alanine transaminase/serum glutamic-pyruvic transaminase (ALT /SGPT), total cholesterol, triglyceride, high-density lipoprotein (HDL) cholesterol, 25-OH-Vitamin D, insulin, high-sensitivity C-reactive protein (hs-CRP), glycated haemoglobin (HbA1c), and complete blood count with differential (CBC+Diff.).

We mailed the blood test results, body composition analysis, and dietary assessments to each participant. In this way, they were encouraged to improve or maintain a healthy lifestyle. Participants with concerns regarding their blood test results were advised to see their personal physicians or make appointments with the FiLWHEL physician.

Personal interviews were conducted by the Filipinos who could speak the Filipino language, thereby reducing the chance of interviewer-interviewee misunderstandings, and answers for sensitive questions were easily obtained. Detailed instructions and guidelines were given to the interviewers prior to conducting the interview. In cases wherein the participants preferred to self-administer the questionnaire, except for the 24-hour recall that required interviewers, assistance was offered to the participants whenever they had any difficulties.

Physical examination procedures and biospecimen collection were carried out by trained professionals. BP measurements were repeated two times. Quality control measures for blood samples include repeated emphasis of the importance of fasting and the choice of laboratory for analysis. Interested participants were carefully instructed and reminded several times about the fasting procedures before the actual blood collection. On site, each participant was asked about the details of their last food or beverage intake. Moreover, blood samples were analyzed and aliquoted by the Seegene laboratory, an international company that is compliant with high quality control procedures. The coefficients of variation (CV %) of duplicate analysis was also obtained.

In addition, the study process at all sites followed the same protocol, and all study days were supervised. Lastly, before data entry, all parts of the questionnaires were checked, and seemingly unclear or missing answers and other observed inconsistencies were clarified on site or over the phone. All entries were double-checked.

Analysis were performed using SAS version 9.3 software package (SAS Institute Inc., Cary, NC, USA). Estimations based on the incidence of pre-diabetes and diabetes in the Philippines led to the conclusion that enrollment of 500 participants with 100 cases occurred during the follow-up would provide 80% power to detect a relative risk of 0.5 comparing the top quartile with the bottom quartile [35].