Ulceration of the foot is a major complication of diabetes mellitus, and optimal self-care may help prevent its development. Research suggests that mobile applications (apps) may affect behavioral change. OBJECTIVE: The purpose of this study was to develop the Mobile Diabetic Foot Personal Care System (m-DAKBAS) and evaluate its effectiveness for patients with diabetes. METHOD: During Phase 1, a mobile app that included communication features, remote patient monitoring, and information was developed and pilot-tested among 10 patients. The Phase 2 study, conducted from June 2017 to April 2018, used a 2-group, pre-test/post-test design to evaluate the effect of the app on patients’ knowledge, behavior, and self-efficacy scores when used for 6 months. Both the experimental (app) and control groups participated in 1 education session at the start of the study. RESULTS: Of 106 patients who enrolled, 88 completed the study (44 in the experimental group and 44 in the control group). Only 6 patients had received education about foot care previously. The average age of all participants was 51.63 years (SD = 8.08). There were significantly more women in the experimental group than in the control group (65% vs. 45.5%; P = 0.5). Each participant used the app for 24 weeks, and the data entry rate was 72.9%. Throughout the study, participants had 1977 data entries (blood glucose and foot observation) in total. Differences between pre- and post-intervention test scores were significantly higher for knowledge, behavior, and self-efficacy in both groups, but the difference was greater in the experimental group (P < .05). Only post-test knowledge scores were significantly higher in the experimental compared with the control group (P < .05). Compared to the start of the study, the proportion of participants with cracked/dry skin and inappropriate footwear was significantly lower in the experimental group but not in the control group. CONCLUSION: In this study, education and follow-up via the mobile app and verbal-only instruction increased the knowledge, behavior, and self-efficacy scores of patients in both groups. Post-study knowledge scores were significantly higher in the experimental group than in the control group. Patient education remains a crucial component of optimal care, and further development, refinement, and testing of mobile applications to improve self-efficacy and reduce the risk of diabetic foot are warranted.

Diabetic foot, one of the major complications of diabetes mellitus, is defined as infection, ulceration, or destruction of tissues of the foot with peripheral arterial disease and/or neuropathy in the lower extremities of persons with diabetes.¹ Diabetic foot ulcers are reported in 2% to 4% of patients with diabetes and are common health problems in developing countries. Only two-thirds of foot ulcers will eventually heal, and up to 28% may result in some form of lower extremity amputation.² For patients who do not require amputation and in whom the ulcers heal, 40% will experience a recurrence within 1 year, 65% within 5 years, and > 90% within 10 years.³ Diabetic foot not only creates a physical problem for the patient but also affects family relationships, financial situations, and the larger society.²

The International Diabetes Federation’s 2019 guideline states that diabetic foot complications and amputations can be decreased by 85% with good foot care, patient education, and treatment by a multidisciplinary team.⁴ A Cochrane Review by Dorresteijn et al⁵ examined studies on the impact of patient education. There was insufficient robust evidence that limited patient education reduced the occurrence of diabetic foot ulcers or amputations.

The World Health Organization reported that solutions to health problems require an innovative point of view, a community that protects and manages its own health, and integrated, patient-centered care that records and processes medical data from birth to death.⁶ Globally, 5.1 billion people have mobile phones, and approximately 4 billion of them use the internet. In Turkey, 77% of 59 million mobile phone users have smartphones, and 51.5 million people use mobile internet.⁷ This supports the idea of using mobile technology as an important tool for improving health.

For patients with diabetes, it can be difficult to create a behavioral change.⁸ Mobile health technologies have the potential to educate patients and encourage them to be active in maintaining their health by providing reminders, personalized data, and social interactions that can enable behavior changes, especially for patients with chronic diseases; in addition, these technologies have direction, interaction, and facilitation features that are always available.^9–11 The American Diabetes Association recommends technology-assisted methods for the prevention or delay of type 2 diabetes.¹² The related literature involves studies that show the effects of the internet and mobile phone usage on self-care activities of persons with diabetes¹³; management of glycemic control¹⁴; knowledge, behavior, and self-efficacy development¹⁵; and diabetes prevention.¹⁶ 

There are many studies in the literature regarding diabetic foot ulcers and mobile technologies.^17–19 An interpretive description study was performed by Kolltveit et al¹⁷ (n = 33) to explore health care professionals’ experience in the initial phase of introducing telemedicine technology into the care of persons with diabetic foot ulcers. Their results showed that using a telemedicine intervention enabled health care professionals to approach these patients with more knowledge, better wound assessment skills, and heightened confidence.

A systematic review (n = 140) of the literature by Nordheim et al¹⁸ assessed the effect of telemedicine follow-up care on clinical, behavioral, and organizational outcomes among patients with leg and foot ulcers. The authors showed that there were no statistically significant differences in outcomes between patients receiving telemedicine and those receiving traditional follow-up. A review and analysis of a database of 5795 patients from a mobile wound healing center in France discussed dermal thermography, hyperspectral imaging, digital photographic imaging, and audio/video/online communication.¹⁹ 

To the authors’ knowledge, there are a limited number of studies evaluating the effects of mobile applications (apps) on foot self-care activities in patients with diabetes. A randomized controlled study by Orhan²⁰ (65 patients in the experimental group, and 65 patients in the control group) evaluated the efficacy of an education app on patient foot care knowledge, self-efficacy, and behavior levels in patients with diabetes. In that study, knowledge, behavior, and self-efficacy level scores were higher in the experimental group than in the control group (P < .01). 

A feasibility study by Hassan²¹ (N = 224) tested the effectiveness of text messaging to reinforce learning and foot care practices in persons with diabetes. Hassan’s study showed that 51 patients (23%) performed good foot care in the pre-test; in the last test, 149 patients (66%) were performing good foot care. In addition, the percentage of patients performing poor foot care decreased from 170 (76%) to 2 (0.9%), and this difference was found to be significant (P = .000). 

A qualitative study by Ogrin et al²² reported on a co-designed app that included information on amputation risk and self-care practices to prevent serious foot complications. The app was piloted in a convenience sample of adults with diabetes (N = 40) from 1 community health service for 12 weeks. Use of the app was low, with 18 participants using the app for any period of time. Qualitative interviews or focus groups were undertaken with 31 participants. Generally, the information was perceived as highly useful and worth pursuing by patients using the app.²²

Despite the positive effects of mobile health practices in the management of diabetes reported in the literature, there are a limited number of studies on this issue, indicating a need for new studies to provide strong evidence. The aim of the current study was to develop the Mobile Diabetic Foot Personal Care System (m-DAKBAS) and evaluate the effects of this mobile app on knowledge, behaviors, and self-efficacy for patients with diabetes. The hypotheses were as follows:

H1: Use of m-DAKBAS increases the knowledge level of patients with diabetes related to foot care.

H2: Use of m-DAKBAS encourages patients with diabetes to form positive behaviors related to foot care. 

H3: Use of m-DAKBAS increases the self-efficacy levels of patients with diabetes related to foot care.

Participants. A 2-group, pre-test/post-test design was used. CONSORT’s extension guidelines for randomized pilot and feasibility trials were followed. The study was conducted from June 2017 to April 2018 in a diabetes outpatient clinic of a university hospital with a 1050-bed capacity (clinical trial registration no. NCT04029103). Power analysis was performed. To find a 30% (50–80%) success difference between the 2 groups, the minimum number of subjects required in each group was determined as 44 (Type I Error = 0.05; Power of the Test, 0.80). In the research process, more samples were selected against the possibility of exclusion of patients. The sample included 52 patients in the experimental group (3 with type 1 diabetes, and 49 with type 2 diabetes) and 54 patients in the control group (3 with type 1 diabetes, and 51 with type 2 diabetes). The convenience sampling selection was made among the patients who visited the diabetes outpatient clinic. Patients who met the research criteria were referred to the researcher by a physician who was not involved in the research. After providing informed consent to participate, the researcher divided the patients into the experimental and control groups according to a computer-based randomization list (MedCalc version 18). 

Inclusion criteria were as follows: being 18 years of age and older, having type 1 or type 2 diabetes and a diagnosis history of  greater than 1 year, having a mobile phone with an Android or IOS operating system, no visual impairment, no hand skill problem, no communication difficulty or mental health limitations, and not having a foot ulcer. 

Data collection tools

Demographic and clinical characteristics form. This form consists of 2 parts and was completed at the start of the study. The first part includes 5 demographic variables including sex, age, education level, employment status, and smoking status. The second part includes the following 8 clinical variables: type of diabetes, duration of diabetes, history of training on foot care, average hemoglobin A1c%  (HbA1c) levels, body mass index (BMI), foot skin condition (skin cracks, calluses, corns, between-toes maceration, and fungal infection), nail cutting (straight across or not), and appropriate footwear (plantar pressure-relieving effect and shape of the foot). Patient knowledge, behavior, and self-efficacy regarding diabetes foot care was evaluated using the forms/scales described in the following section. Data collection forms were completed by the researcher at the beginning and end of the study.

Knowledge form. The Diabetic Foot Knowledge Form (DFKF) was developed by the researchers to measure patients’ knowledge about diabetic foot and foot care.^1,2,23–25 It consists of 20 questions about the following topics: peripheral vascular signs; nerve damage signs; wearing appropriate shoes; nail cutting; callus medication and tape; checking the inside of the shoe; walking in bare feet; using a heater or hot water bag to warm the feet; drying between the toes; checking the sole of the feet; moistening the feet; sock selection; foot examination by a physician; control of the bath water temperature; amputation; the effect of smoking on diabetic foot ulcers; use of alcohol and tincture of iodine in wound dressings;  effects that walking has on blood circulation; effects of massage on blood circulation; and  effects of high blood sugar on the feet. The questions have 3 options (true, false, and I don’t know). The form contains 9 false and 11 correct statements. Each correct answer is scored as 1 point. Higher scores indicate higher knowledge levels. Cronbach’s α coefficient was α = 0.72 in this study. 

Behavior form. The Foot Care Observation Guide was developed by Borges and Ostwald in 2008.²⁶ Turkish language validity of the scale was performed by Biçer and Enç, and it was adapted as the Foot Self Care Behaviours Scale (FSCBS). The scale is composed of 15 items that are assessed as 1 = never, 2 = rarely, 3 = sometimes, 4 = frequently, and 5 = always. Scores range between 15 and 75. Higher scores indicate better self-care behaviors. Cronbach’s α coefficient was α = 0.76.²³  

Self-efficacy form. The Diabetic Foot Care Self-Efficacy Scale (DFCSES) was developed by Quarles in 2005.²⁷ Turkish validity and reliability was performed by Biçer and Enç. The scale had 9 items rated on an 11-item visual scale that ranged from 0 (I find it totally insufficient) to 10 (I find it totally sufficient). Scores range between 0 and 90. Higher scores indicate higher self-efficacy. Cronbach’s α coeeficient was 0.83.²⁴ 

m-DAKBAS evaluation form. This form was developed by the researcher to assess the views of patients with diabetes about ease of  use of the mobile app (1 point = very difficult to 10 points = very easy), practicality, recommendations to other patients with diabetes (yes or no), and benefits and suggestions (open-ended questions) about the application. The m-DAKBAS assessment form is composed of 6 questions.

The study was conducted in 2 phases (Figure 1). 

Phase 1: app development. m-DAKBAS was formed within the theoretical framework of knowledge, behavior, and self-efficacy. That is, patients with diabetes should demonstrate positive care behaviors to protect their foot health. If a decrease in ulcer and amputation rates is desired, optimal foot care behavior should be realized. Individuals’ self-efficacy needs to increase for self-care knowledge to become a behavior.^28,29 Self-efficacy is defined as “the belief of an individual’s ability in executing and realizing courses of action to obtain some attainments.”³⁰ Individuals having these beliefs and attitudes tend to demonstrate behaviors appropriate toward meeting a goal.²⁸

The text content of the app is in Turkish and was prepared in line with the latest evidence and guidelines. The content was presented to 6 experts (a lecturer in the Department of Endocrinology, 4 nurse lecturers who conducted research about diabetes and diabetic foot, and a physician in the Department of Undersea and Hyperbaric Medicine). 

Software process. Mobile apps on the web were analyzed. There was a limited number of diabetes-related Turkish apps on the web. The researcher, who is a diabetes nurse specialist, is familiar with the profile of patients with diabetes in Turkey. For design and use convenience, the patient profile was taken into consideration when interfaces were developed. Service was bought for the software of the application, and the development process took 3 months. During that time, 3 meetings were held with the information technologies (IT) specialist. In the first meeting, the researcher informed the IT specialist about the text content and other features that were desired. The domain name and hosting were bought for the researcher for 1 year. In the second meeting, the IT specialist presented the interface design of the app. In the third meeting, the administrator and user panels were introduced by the IT specialist. In this meeting, misunderstandings about the expectations of the tool were clarified, and final revisions were made accordingly.

Framework of the m-DAKBAS. 

• The application was designed as a web-based mobile app. 

• The system has admin (health professional) and user (patients with diabetes) panels. 

• The app can be used on mobile phones with IOS and Android operating systems. 

• The app has a password login system for the storage and safety of personal information. 

• The user panel has a home screen interface, 8 interfaces, and 8 subinterfaces.  

• There are “get information” and “prevention” interfaces that provide the text that contains information about the definition of diabetic foot, risk factors, and proper foot care for prevention. 

• There is a messaging interface for administrator communications. Personal messages with motivational and informative content were prepared, including the following: “Dear …. your blood glucose values seem to be very good. Congratulations; It is great!” “You have a moisturizer for your feet,” “Healthy diet, regular exercise, and medicine are important for blood glucose control in diabetes,” “HbA1c test gives the 3-month blood glucose average,” and “It is good to keep HbA1c values under 7% to avoid the adverse effects of diabetes.”

• The section on foot observation allows patients to enter data. There are checkboxes for dryness, swelling, ulcer, warmth, callus, fissure, tingling, burning, unusual sensation, and pain assessment in the foot observation section. Patients can check the boxes of their own findings and send them to the system administrator. 

• The blood glucose section includes text boxes where morning, noon, and evening fasting and postprandial blood glucose levels can be recorded. 

• The “test yourself” interface enables patients to test their knowledge about foot care and enables them to improve their knowledge.

• In the “photo share” interface, patients can take and send a photo of their foot.

Health personnel/system administrators can confirm users’ records from the admin panel, update information content, monitor negative findings reported by the patient and enable timely intervention, and send diabetes-related reminders and informative messages. They can also monitor users’ blood sugar levels, regular and irregular data entry, and foot health status. 

Pilot study. The pilot study was conducted with 10 patients with diabetes who met the sample group research criteria. Informed consent was obtained for the pilot study. Patients with diabetes used the mobile app for 10 days; they were asked to send their blood glucose level each time they measured it and to send their foot observations daily. During this period, some revisions were made in line with the feedback provided on the device. For example, initially patients could not see their right or wrong answers in the “test yourself” interface or messages from the admin; these deficiencies were corrected. 

Phase 2: app evaluation. After the pilot study, the physician in the diabetes outpatient clinic referred patients who met the research criteria to the researcher. The patients who agreed to participate in the study were included after their written consent was received. The researcher assigned patients to the experimental or control groups according to the order in a computer-based randomization list previously created. The convenience sampling selection was made among the patients who visited the diabetes outpatient clinic. All participants were given training about personal foot care and foot observations in the patient training room by the researcher (a diabetes nurse specialist). Participants in the control group were given a training by the researcher through verbal instruction about the information in the content of m-DAKBAS (definition of diabetic foot, risk factors, protective precautions, and daily foot care). The training was given once and was not reviewed again. Patient training took about 30 minutes. Forms were completed in 15 minutes.  

For participants in the experimental group, the m-DAKBAS app was downloaded to their mobile phone by the researcher, and participants were given a username and password. Participants were instructed how to use the application by the researcher after a sufficient number of trials performed together (eg, patients were asked to send sample blood glucose values and foot observation data). The duration of the use of the app was 6 months (24 weeks). 

Throughout the 24 weeks, participants were asked to use the app to send their blood glucose levels each time they were measured and also to send foot observations daily, because it is recommended that patients with diabetes observe their feet every day.^1,4,12 The researcher communicated with participants in the experimental group through the app. Using the admin panel, the researcher followed the participants’ frequency of using the app, their blood glucose levels, and foot care, and also tried to find solutions to the medical problems encountered. The data sent by the participants were analyzed by the researcher on a weekly basis and reviewed to monitor for potential patient problems/concerns. The participants were told that they could share photos if they had a problem with their feet. Use of the “test yourself” interface was optional, and data from this field were not evaluated by the researcher. In the case of abnormal findings, the researcher communicated with the participant through messaging, phone, or face-to-face interviews. Based on these records, participants were provided feedback about foot observations and blood glucose values. Participants were sent SMS reminders if the tasks were not completed. Participants who completed the research period were given an appointment for the post-test after 24 weeks. Post-test forms (DFKF, FSCBS, DFCSES, m-DAKBAS evaluation form, and foot care findings,) were completed by the researcher. 

Statistical analysis. Data from the questionnaires were manually entered into the SPSS for Windows version 24.0 package program. P < .05 was accepted as statistically significant. Normality distribution of the data was tested using the Shapiro-Wilk test, which indicated that the data were not distributed normally (P > .05).  

Average ± standard deviation for numerical variables (age, HbA1c, duration of diabetes, and BMI) as well as number and percentage values for categorical variables (sex, educational level, employment status, smoking status, and m-DAKBAS evaluation) used frequency and description. The Wilcoxon test was used to compare the scores of the information form (DFKF) as well as the behavioral (FDCBS) and self-efficacy (DFCSES) scales at different times. The Mann-Whitney U test was used to compare independent groups. 

Chi-square test was used to test the normal distribution of categorical variables between experimental and control groups (sex, educational level, employment status, and smoking status). The t test was used to test the normal distribution of numerical variables between the experimental and control groups (age, HbA1c, duration of diabetes, BMI). McNemar test was used to determine if there were differences on a dichotomous dependent variable between 2 related groups (foot examnination findings).

Ethics review and approval. In line with the Declaration of Helsinki, written approval was obtained from the Clinical Studies Ethics Committee and the department where the study was conducted (2017/69 decision number dated February 27, 2017). Patients were expected to use their own internet packages when sending data. The patients had no concern about this. Written informed consent was obtained from all participants. 

A total of 106 patients who met the research criteria were included in the study. Some patients were excluded from the experimental and control groups during the study. Of these, 7 patients in the experimental group (2 with type 1 diabetes and 5 with type 2 diabetes) and 9 patients in the control group (2 with type 1 diabetes and 7 with type 2 diabetes) were excluded because they could not be reached by phone, their phone broke, they did not participate in the post-test, or they withdrew the study. Because the number of patients with type 1 and type 2 diabetes was not homogeneously distributed, 1 participant with type 1 diabetes in both groups was not included in the evaluation. The sample included 44 experimental and 44 control group patients with type 2 diabetes (88 in total) (Figure 1). 

In the experimental group, 65.9% (n = 29) of the participants were women, 56.8% (n = 25) were primary school graduates, 65.9% (n = 29) were unemployed, and 65.9% (n = 29) were non-smokers. As for the control group, 54.5% (n = 24) of the participants were male, 63.7% (n = 28) were primary school graduates, 70.5% (n = 31) were unemployed, and 68.2% (n = 30) were non-smokers. Although none of the participants in the experimental group had previously received training on foot care, 14% (n = 6) of participants in the control group did (P = .007) (Table 1).

The average age of participants in the experimental group was 51.16 (SD = 8.27) years, the average duration of diabetes was 7.36 (SD = 5.43) years,  BMI was 30.91 (SD = 4.7) kg/m², and the average HbA1c  level was 8.10% (SD = 2.05).  The average age of the participants in the control group was 52.11 (SD = 7.96) years, average diabetes duration was 7.34 (SD = 6.48) years, BMI was 32.58 (SD = 6.57) kg/m², and HbA1c was 8.03% (SD = 1.69) (Table 2). Demographic and clinical characteristics of the participants were not statistically significant (P > .05) except for having previously received training about foot care (P < .05) (Table 1 and Table 2). 

In the pre-test, the knowledge score of the control group was 13.61 (SD = 3.65) and the knowledge score of experimental group was 12.89 (SD = 4.34). There was no significant difference (P = .58) between the groups regarding knowledge scores, but both were above average (range, 0–20). In the post-test, the knowledge score of the experimental group (16.73 [SD = 1.56])  was significantly higher than the knowledge score of the control group (15.05 [SD = 2.17]; P = .000). In addition, knowledge scores increased significantly in both groups in the post-test  (P = .013) (Table 3).

Overall, both groups’ behavior scores were above average (range, 15–75) (Table 3). In the pre-test, the behavior score of the control group was 52.61 (SD = 8.75) and behavior score of  experimental group was 51.02 (SD = 9.78). Behavior scores were distributed homogeneously between groups (P > .05). In the post-test, behavior scores increased significantly in both groups (P = .000). Although the behavior score of the experimental group (62.59 [SD = 7.76]) was higher than the behavior score of the control group (59.45 [SD = 10.53]) in the post-test, there was no significant difference between the groups (P = .23) (Table 3).

In the pre-test, the self-efficacy score of the control group was 65.59 (SD = 17.45) and the behavior score of the experimental group was 65.59 (SD = 16.68). Overall, both groups’ self-efficacy scores were above average (range, 0–90). In the post-test, self-efficacy scores increased significantly in both groups (P = .000 and P = .010, respectively). The self-efficacy score of the experiment group (74.16 [SD = 13.46]) was higher than that of the control group (71.27 [SD = 12.97]) in the post-test, but there was no significant difference between the groups (P = .019) (Table 3).

Participants in the experimental group reported that they liked the system because it enabled them to communicate with health personnel (52.3%), it was informative (40%), and it had a remote monitoring system (38.6%); participants reported that they wanted to recommend the mobile app to other patients with diabetes (93.2%) and that they thought it contributed  to foot health  (93.2%).  The ease of use score was 8.54 ± 1.78 (range, 3–10) (Table 4). 

Table 5 shows the foot examination findings before and after the intervention. In the beginning, 14 participants in the experimental group had made non-appropriate footwear selections; 10 of these 14 participants had appropriate footwear selections in the post assessment. This difference was found to be statistically significant (P = .010).  In the pre-test, 13 participants in the experimental group had skin cracks on their feet. In the post assessment, 3 participants had foot skin cracks (P = .004). As for the post assessment in the control group, although positive changes were detected in patients’ footwear selections (P = .343) and foot skin cracks (P = .97), the results were not significant.

 The data entry rate was 72.9%. In total, participants had 1977 (blood glucose and foot observation) data entries from the beginning (June 2017) until the end of the study (April 2018) (Figure 2). The application was used for 24 weeks. Participants did not send foot observations daily. Only 5 participants shared photos to show their foot care. None of the participants reported foot problems during the study. The participants contacted the researcher only in case of poor glycemic control. Two (2) participants were hospitalized and followed-up because they had poor glycemic control. 

Diabetes is a chronic disease with serious complications. The increase in the duration of diabetes, HbA1c values, and BMI levels are risk factors for microvascular and macrovascular complications.⁴ Clinical characteristics of experimental and control groups were evaluated in the current study. The mean BMI was > 30 kg/m², mean HbA1c level was > 8%, and duration of diabetes was > 7 years (Table 2). From these data, it can be concluded that the patients did not have adequate metabolic control and were therefore at risk for chronic complications. In addition, these results reflect those of the TURDEP II study conducted in Turkey.³¹

The International Working Group on the Diabetic Foot¹ recommends education about appropriate foot self-care for those with diabetes who are at risk for foot ulceration. In the current study, 6 participants in the control group previously received training on foot care (Table 1). The presence of these 6 patients may have negatively affected the targeted results in the experimental group. 

Overall, the knowledge scores of patients in the experimental and control groups were above average (range, 0–20) (Table 3). In the pre-test, it was seen that the mean knowledge score was homogeneously distributed between the experimental and control groups (P > .05). In the post-test, the knowledge score of the experimental group was statistically significantly higher than the control group (P = .000). In addition, knowledge scores increased significantly in the experimental and control groups in the post-test (P = .000 and P = .013, respectively). Verbal training given to the control group increased patient knowledge, but the app was more effective (Table 3). This result was expected because it had been shown that apps provided ease of access to information.^10,11

In the pre-test, the mean knowledge score was homogeneously distributed between the experimental and control groups (P > .05), and the behavior scores in both groups were above average (range, 15–75). In the post-test, although the behavior score of the experimental group was higher than the behavior score of the control group, there was no significant difference (P = .23). This result was considered to be due to the limitations of the mobile device used in the study (eg, participants’ inability to see statistical graphic data on the mobile device). Also, diabetes is a lifelong chronic disease. Individuals need to cope with many factors to adapt, and this can be a long process for some patients.³²

Knowledge alone cannot change behavior; however, self-efficacy (belief in one’s own abilities, specifically the ability to meet a challenge) can change behavior.^28,30 In the post-test, although the behavior score of the experimental group was higher than the behavior score of the control group, there was no significant difference between the groups (P = .19). In addition, self-efficacy scores increased significantly in both groups in the post-test (P < .05) (Table 3). This result is not surprising. Self-efficacy can be improved through education, but it can also be influenced by environmental and personal factors.^28,30 Initially, both groups’ self-efficacy scores (range, 0–90) were above average (Table 3). This indicates that participants were ready to create a behavior change. To the authors’ knowledge, there is limited research regarding the time necessary to achieve a desired result.

To assess the effect of short messaging via mobile phones on the knowledge and behaviors of patients with diabetes, Hassan²¹ sent short messages to 224 patients throughout 12 weeks about foot care practices. Results of that study indicated significant increases in foot care knowledge and behaviors. The percentage of the patient group that was considered to have good knowledge about foot care was 23% in the pre-test and increased to 66% in the post-test. The proportion of patients with poor foot care decreased from 76% to 0.9% in the post-test. This difference was found to be significant (P = .000). The significant increase in the knowledge score was in line with the present study, but the behavior score was not (Table 3). 

Orhan²⁰ assessed a mobile diabetic foot care education app and reported a significant increase in the knowledge about diabetic foot, behavior, and self-efficacy scores of the experimental group (n = 65) in comparison to the control group (n = 65) post test. Enhancing knowledge levels as well as behavioral changes is highly important for the management of diabetes and its complications. This requires a change in lifestyle, which is multifactorial and often difficult for patients. Guo et al¹⁵ aimed to improve and increase self-care activities in patients with diabetes; therefore, they developed a mobile app and evaluated its efficiency. Although there was a significant increase in knowledge (17%; P = .001) and behavior (22%; P = .015) levels, the increase in self-efficacy was not significant (P = .065).¹⁵ Although similar results were obtained in terms of knowledge and self-efficacy levels in the current study, the same result was not obtained in behavior (Table 3). 

In a study that involved 150 patients with type 1 and type 2 diabetes, Seyyedrasooli and Parvan³³ assessed the effects of individual and group training methods in foot care on self-efficacy. The participants were randomly divided into 2 intervention groups (group training and individual training) and 1 control group. There were no significant differences among the 3 groups in terms of self-efficacy mean scores, but the patients’ self-efficacy increased within the overall group in the post-test. This is similar to the current study in that although there was no difference in self-efficacy scores between groups in the post-test, an increase was observed within the overall group. 

Ogrin et al²² (N = 40) developed a mobile app for early diagnosis of foot complications in patients with diabetes and evaluated its effectiveness. The pilot study was conducted with 31 participants and lasted 12 weeks. In that study, 18 of 31 patients used the app regularly. In the current study, 44 of 52 participants in the experimental group used the mobile app until the end of the study period. However, not every participant sent data at the desired frequency. Throughout the study, participants sent blood glucose and foot observation data (n = 1977) (Figure 2). Participants did not send foot observation data daily. On average, each patient sent blood glucose measurements once a week and foot observation data once a week. This result may be due to the participants’ use of their own internet packages while sending data. As in the present study, Ogrin et al²² also reported that the app they developed was beneficial and worth optimizing.

Patients at risk should understand the importance of foot monitoring on a daily basis, the proper care of the foot, nail cutting, skin care, and the selection of appropriate footwear.³⁴ During the post-test assessment in the current study, participants in the experimental group were wearing more appropriate footwear, and the proportion of persons with dry/cracked skin decreased significantly in the experimental group (P < .05) (Table 5). In a similar study, Hassan²¹ found that poor foot care, which was present in 76% of patients (n = 170) at the beginning of the study, decreased to less than 1% (n = 2) after 12 months.

Mobile health apps may be advantageous for patients with busy working lives.³⁵ Studies have reported that patient satisfaction with mobile apps is high^{13,15,20,36,37} and that the most-favored features include being portable, accessible, and educational in nature; having audiovisual richness; and offering the facilitator role for learning.^{15,18,36–38} These results are in line with patient views related to the m-DAKBAS (Table 4). 

The aim of the current study was to develop the m-DAKBAS and evaluate its effects on knowledge, behaviors, and self-efficacy for patients with diabetes. Although the present study did not find a significant difference in post-study behavior and self-efficacy scores between participants who did or did not use the app, there was a significant increase in knowledge level, behavior, and self-efficacy scores in both the experimental and control groups. In addition, the difference in post-study knowledge scores was signficant between the control and experimental groups. According to the control group, the percentage of persons with dry/cracked feet was also significantly lower at the end of the study. Moreover, those who completed the study were satisfied with the app and 80% wanted to keep using it. The results suggest that, in persons with diabetes mellitus, education about diabetic foot increases knowledge, affects behavior, and improves self-efficacy. It also suggests that further development, refinement, and testing of mobile apps to improve self-efficacy and reduce the risk of diabetic foot are warranted.

Limitations of this study include that it was conducted in 1 center with patients who could use a mobile phone, had diabetes for at least 1 year, and did not have severe vision problems, hearing loss, or foot ulcer. Patients used the application for only 6 months (24 weeks). The device did not create statistical data for each participant per day/week/month. Participants were unable to see their own statistics. Another limitation to be considered was not performing intent-to-treat analysis.

It is recommended that mobile app designs be supported with behavioral theory models. Designs should also be applied to larger groups of patients at higher risk of diabetic foot. More long-term monitoring and results (eg, about amputation and recurrent foot ulcer) need to be assessed.

The authors express their gratitude to the Vehbi Koç Foundation for providing financial support; Professor Ersin Akarsu, RN, PhD; the diabetes polyclinic staff for their collaboration and support; and all the study participants.

Dr. Kiliç is an assistant professor, Department of Nursing, Sanko University, Gaziantep, Turkey. Dr. Karadağ, is a professor, Koç University School of Nursing, Istanbul, Turkey. Addres all correspondence to: Meryem Kilic, RN, PhD, Sanko University, Faculty of Health Sciences, Department of Nursing, post code: 27090 Gaziantep/Turkey; tel: 90 (342) 21165 83; fax: 90 (342) 211 65 81; email: meryemcal@gmail.com.

Financial support of this research was provided by Koç University, Vehbi Koç Foundation, Turkey (foundation support no.  DK.28.00012.09). This study was presented at the 5th National and 1st International Basic Nursing Care Congresses, December 6–8, 2019, Antalya, Turkey. More detailed information about this study can be found at https://tez.yok.gov.tr/UlusalTezMerkezi/tezSorguSonucYeni.jsp. Thesis number 532446.