Strengthening the reporting of observational studies in epidemiology (STROBE) guideline was used to design and report this epidemiological study [11] . Previous study reported that Ujung Pandang (now Makassar) was one of the most prevalent DM cities in Indonesia [12] , including for asymptomatic DM [13] . Thus, research setting was conducted in outpatient endocrine clinic, Wahidin Sudirohusodo hospital, a regional hospital in Makassar (834 beds), which considered representative in eastern Indonesia.

Sample size was calculated by using power analysis equation [14] , where P refers to the lifetime prevalence of DFU as 25% [15] . Thus, our calculated sample size was 288 participants. Study populations were all T2DM patients who attended the research setting from May 2013 to February 2014 for DM therapy. Inclusion criteria were T2DM patients who have ≥ 18 years old, this was our denominator and presence of DFU was numerator of study. T2DM patients who attended hospital other than endocrine outpatient clinic were excluded. Diagnosed T2DM patients are based on physician endocrine assessment and glycemic status according to American Diabetes Association (ADA) 2013 criteria [16] , which are written in hospital medical records.

Dependent variables were presence of risk and DFU. Presence of risk was neuropathy or angiopathy as proposed by IWGDF [6] . Neuropathy was evaluated by using 5.07/10 g Semmes-Weinsten Monofilament (SWM) at four points of each foot (dorsal hallux, metatarsal I, III and V) [15] , absence of one of the four sites is considered as neuropathy [17] . Meanwhile, presence of angiopathy was evaluated with Ankle Brachial Index (ABI) by using a hand held Doppler (Bidop ES-100V3, Hadeco-Kawasaki, Japan) both dorsal and posterior tibialis foot. An ABI ≤ 0.9 considered has peripheral ischemic [18] . Since there was inconsistent normal range of TBI, we reported as mean and standard deviation instead of categorical data [19] . DFU define as presence full thickness lesion on the skin as proposed by IWGDF [20] .

Independent variables were demography, clinical foot problems and foot care behaviors. Participants’ demographies were age, sex, occupation, religion, education, smoking status, DM therapy, duration of DM, body mass index (BMI), blood pressure and HbA_1C (at the time of visit or at least two months of previous data) was analyzed in hospital laboratories. Clinical foot problems were skin problems (dry skin, callus-corn-fissure and tinea pedis), nail problems (nail deformity and onychomycosis), foot deformity, ABI and Toe Brachial Index (TBI) status.

Foot care behaviors were interviewed by using questionnaire, including foot inspection, foot washing, nail trim, footwear inspection, and footwear practice. Participants’ cultural and spiritual belief related to foot care behaviors was also interviewed by using open questions (do you have spiritual or cultural belief related to your foot care?). These items were developed based on clinical experience in Indonesian setting. To control bias, all of foot assessments, including determination of DFU were done by one investigator (Wound Care Nurse). All of assessments were written into paper-pencil form by co-authors and stored by primary investigator. Participants’ anonimity was maintained by using identical number (ID number).

Both ordinal and nominal data were described as absolute values and percentages (n, %), while continuous data were reported as mean and standard deviation. Missing data were replaced with its group mean (11 TBI, 19 ABI, and 12 HbA_1C data).

Univariate data were analyzed by χ2 test or Fisher exact test for categorical and Independent t test for continuous data. Cut off P < 0.1 and logical reasons were used to select candidate of predictors into multivariate analysis. Possibilities for multicollinearity were diagnosed by reading variance inflation factor (VIF) by using regression linear. All of predictors with VIF < 5.0 were entered into logistic regression with Forward LR methods. The significance level was set up at 95% with P = 0.05 (two tailed). All data were analyzed by SPSS version 16.0 software (SPSS, Inc. Chicago, IL).

Ethical considerations were approved from Kanazawa University, Japan (Number: 438) and Hasanuddin University, Indonesia (Number:0866/H4.8.4.5.31/PP36-KOMETIK/2013). All of participants and family were received explanation and signed informed consent prior to data collection.

There are 280 T2DM patients who have registered in the research setting, 259 participants agree to participate (response rate 92.5%). At the end of the study, 10 participants are unable to complete assessment and remain 249 participants in the analysis. Participants without any risk are 90, with risk 112 (with neuropathy 14, neuropathy or deformity 64, and neuropathy or deformity or ischemic 34), history of DFU 14, history of amputation 3 and current DFU 30 of 249 participants. Overall, prevalence of presence of risk factors (excluding history and presence of DFU) is 55.4% (95% CI 53.7% - 57.0%) and prevalence for DFU (including history and presence of DFU) is 12.0% (95% CI: 10.3% - 13.6%).

In univariate analysis we evaluate candidate of predictor for presence of risk and DFU based on demographic, clinical foot assessment and foot care bahaviors. With regard to demographic factors, age is older (P = 0.021) for presence of risk, meanwhile religion (P = 0.032), education (P = 0.023), DM therapy (P = 0.020) and percentage of HbA_1C (P = 0.029) are associated with presence of DFU (Table 1). Based on clinical foot assessment, deformity (P = 0.000) and ABI (P = 0.000) are associated with presence of risk factors. However, only dry skin has an evidence role in presence of DFU (P = 0.001) (Table 2).

Related to participants’ foot care behaviors, foot inspection has been less frequently at risk group, including

^*Excluded presence history DFU amputation. Data are n(%) unless indicated by † are mean (±SD). p values determined by Chi Square x² or Fisher exact test for categorical data and Independent t test for continuous data.

p values determined by Chi Square x² or Fisher exact test for categorical data and Independent t test for continuous data. ^*Excluded presence history DFU amputation. ^†Forefoot Deformity including; hallux vagus, bunion, varsus deformity, hammer toe. ^‡ABI > 1.3 were ignored and read only TBI data if presence.

daily foot inspection (P = 0.003), entirely foot inspection (P = 0.002), and supported for foot inspection by family (P = 0.007). Related to the presence of DFU, daily five time foot washing or more (P = 0.014), supported for foot washing by family (P = 0.021) and using shoes as footwear (P = 0.003) are associated with presence of DFU. In addition, cultural belief (P = 0.036) and spiritual belief (P = 0.007) related DM status has an evidence role in presence of risk and presence of DFU, respectively (Table 3).

Based on multicollinearity diagnostic we excluded supported for foot inspection (VIF 6.84) and daily foot washing (VIF 6.99). In addition by using logical approach we also excluded systolic blood pressure and occupation (not related to foot problems), forefoot deformity, ABI and TBI (used to distinguish risk category) and religion (100% of DFU in group are Muslims).

Logistic regression was conducted to evaluate predictors of presence of risk and DFU. When all candidate predictors from univariate analysis entered together; age and daily foot inspection status were associated with presence risk, with an odds ratio (OR) 1.04 (95% CI: 1.005 - 1.074) and 0.36 (95% CI: 0.186 - 0.703), respectively. Meanwhile predictors for presence of DFU are DM therapy (insulin) OR 9.37 (95% CI: 2.240 - 39.182), type of footwear (shoes) OR 0.05 (95% CI: 0.007 - 0.294) and spiritual belief that DM was a disease OR 0.04 (95% CI: 0.004 - 0.326) and belief that DM was temptation from God OR 0.13 (95% CI: 0.027 - 0.598) (Table 4).

Prevalence of risk and DFU are higher in Indonesia. The current study found that prevalence of risk (neuropathy and angiopathy) in this study was 55.4%. These findings are within global prevalence of risk 40% - 70% [21] .

p values determined by Chi Square x² or Fisher exact test for categorical data and Independent t test for continuous data. ^*Excluded presence history DFU amputation.

This percentage still remains higher compared to India [22] . Meanwhile, current prevalence of DFU is 12%. These findings are higher compared to China (most populated DM country) [23] and in comparison with global prevalence 1.4% - 5.9% [21] . In addition, our previous study also confirmed high prevalence of DFU in home care setting 26.0% [24] . Prevalence of risk and DFU in Indonesia is potential to be higher since the prevalence of undiagnosed DM in Indonesia was higher [4] [5] , including prediction of prevalence DM in Indonesia [1] [2] . Thus, the findings indicate the importance of preventive strategies to prevent risk and DFU in Indonesia.

There are two predictors for presence of risk, age and daily foot inspection. Current study also finds that age is associated with presence of risk. These results are consistent with previous study that increasing age linear with increasing risk for neuropathy and angiopathy [25] . Regarding daily foot inspection status (OR: 0.36), participants who perform daily foot inspection were less prevalent to presence of risk factors. The findings indicate the importance of foot inspection to prevent presence of risk factors.

The strengths of this study are, besides insulin therapy, mainly predictors for presence of DFU related to foot care behaviors (using shoes as footwear and spiritual belief that DM was a disease or temptation from God). These predictors are preventable by increasing knowledge related to foot care.

Current study revealed that participants who used insulin were associated with presence DFU, similar with Europe [26] and in Asia study [23] . This finding might be explained by the fact that DM therapy reflect severity of glycemic status which increase risk for DFU. In addition, HbA_1C was associated with presence DFU in univariate analysis, but diminished in multivariate analysis as well as others studies [26] .

Another important result is that the spiritual belief is associated with presence of DFU. As we know, the role of spiritual belief against DFU has not been explained in previous studies. Current study indicates that spiritual belief of DM is a disease or temptation from God has less likely to presence of DFU compared to who has no spiritual belief. One of the potential answers is that the spiritual belief might give results to positive behaviors related to foot care. Previous study confirmed that spiritual belief associated with positive coping among T2DM patients [27] and has a role in controlling glycemic status [28] .

Another predictor is related to footwear practice. In this study, using shoes as footwear is less likely to presence of DFU compared to who use sandals. Most of the Indonesian people use sandals as footwear which only covers the plantar of foot area, as a result the foot remains at high risk from external trauma. Even though effectiveness of footwear to prevent DFU remains unclear [29] , these findings emphasize the important effect of shoes as footwear to prevent DFU compared sandals.