This study was conducted within the disability community organised under the Indonesian Association of Persons with Disabilities, Yogyakarta Branch, located in Sleman Regency, Yogyakarta. The community comprises individuals with various types of disabilities, including lower-limb amputation.

Participants were individuals with unilateral lower-limb amputation, either transtibial or transfemoral, who met specific eligibility criteria. Inclusion criteria were age 25–65 years, residence in Sleman Regency, ownership, and ability to operate a mobile phone, and willingness to participate until study completion by providing written informed consent. Exclusion criteria included severe comorbidities that could interfere with participation, communication difficulties, residence outside the study area or refusal to participate.

A quasi-experimental one-group pre-test–post-test longitudinal design was conducted. All eligible participants were allocated to the intervention group. Descriptive assessments were performed at four time points: baseline, 1 month, 2 months and 3 months (end of follow-up), to summarise participant characteristics and to illustrate trends in outcomes over time using graphical displays. Because of the small sample size and non-normal distribution, nonparametric methods were applied. Exploratory bivariate analyses were conducted at baseline to describe relationships between selected variables and quality of life. Owing to the small sample size and non-normal distribution of the data, nonparametric methods were applied. Changes in quality of life between baseline and the 3-month follow-up were evaluated using the Wilcoxon signed-rank test. No regression-based or generalised estimating equation (GEE) analyses were performed.

The study was carried out over 3 months, from September to December 2024. The initial 2 weeks were allocated for recruitment and participant screening. This was followed by the 3-month intervention period, which included baseline data collection (pre-test), implementation of the mHealth application, biweekly evaluations and post-test assessment at the end of the intervention.

A purposive sampling strategy was applied, selecting respondents according to the predefined inclusion and exclusion criteria. At the initial stage, 106 individuals with lower-limb amputation were screened. Eighty were excluded owing to refusal to participate, residence outside the study area, communication barriers (including difficulty using mobile applications) or severe comorbidities. A total of 25 participants met the criteria and received the intervention. During the study, three participants withdrew (two owing to work commitments and one owing to family issues), resulting in 22 participants included in the final analysis.

Participants were recruited and screened according to eligibility criteria, after which informed consent was obtained and a baseline assessment was conducted, including measures of attitude, perceived susceptibility, perceived severity, core muscle strength and quality of life. The mHealth intervention consisted of integrated educational and exercise content delivered through a smartphone application over a 3-month period. The exercise component comprised instructional videos demonstrating 11 core-strengthening movements, each accompanied by 45 min of step-by-step explanations (three replications each set) and increased by 15 min in the following month to facilitate correct and safe performance at home. Participants were required to complete at least two exercise sessions per week, following the video guidance. The educational component provided practical information on safe body positions and movement strategies for daily activities (e.g. sitting, lifting and transfers) customised for lower-limb amputees. To support adherence, the application included automated reminders for scheduled exercise sessions and prompts to review the educational materials on a regular basis. All participants received the mHealth intervention and were evaluated biweekly for 3 months, assessing their adherence to the exercise regimen, specifically the number of exercise sessions performed, with a final assessment at the end of the study period to evaluate overall outcomes. The intervention was designed to improve core muscle strength and provide guidance on safe movement strategies. From 25 enrolled participants, three dropped out, resulting in 22 participants included in the final analysis (Figure 1).

The intervention was delivered over a 3-month period with two supervised sessions per week, resulting in a total of 24 scheduled sessions. All sessions were conducted under the supervision of trained instructors. Participant attendance was monitored using signed attendance sheets at each session and cross-checked with application usage records. A total of 25 participants were initially enrolled. Three participants attended fewer than four sessions (range: 1–3 sessions) and were considered withdrawn because they did not meet the minimum adherence criteria required for effective application use. The final analysis therefore included 22 participants who completed the intervention.

Quality of life was assessed using the Indonesian version of the WHOQOL_BREF (translated by Catholic University Atma Jaya 2004; WHOQOL. 2012), which in this study comprised 26 items covering four domains: physical, psychological, social relationships and environment. The Indonesian version has been translated, culturally adapted and validated in previous studies. The Indonesian version of the WHOQOL-BREF has been widely used in previous Indonesian studies (Millenia et al. 2024; Resmiya & Misbach 2019) and has demonstrated good psychometric properties. Each item is rated on a five-point Likert scale. Domain scores were calculated according to the WHOQOL-BREF scoring guidelines and transformed to a 0–100 scale, with higher scores indicating better quality of life. In the primary analyses, quality of life scores were treated as continuous variables. For descriptive purposes, overall quality of life scores were also dichotomised into ‘high’ (≥ 60) and ‘low’ (< 60) categories based on previous studies (Silva et al. 2014; Silva et al. 2019). These categorical analyses were considered supplementary and exploratory. Baseline characteristics including age, duration of employment, education level, type of amputation and BMI were summarised descriptively.

Data were checked for completeness and consistency prior to entry, coded anonymously and verified using double entry. Data processing was performed with IBM SPSS Statistics version 29 (IBM Corp., Armonk, NY, USA). Given the quasi-experimental one-group pre-test–post-test longitudinal design, small sample size (n = 22) and non-normal distribution of outcome variables, nonparametric statistical methods were applied. Descriptive statistics and graphical displays were used to summarise participant characteristics and to illustrate trends in outcomes across four time points (baseline, month 1, month 2, and month 3). Pairwise changes between baseline and 3-month follow-up were analysed using Wilcoxon signed-rank tests. Bivariate associations between categorical predictors (attitude, perceived susceptibility, perceived severity and core muscle strength) and dichotomised quality of life were examined using chi-square or Fisher’s exact tests, as appropriate. Crude odds ratios (OR) and 95% confidence intervals (CI) were reported for descriptive and exploratory purposes only. Multivariable regression and GEE analyses were not performed because of the limited sample size and the high risk of overfitting. Adjusted analyses were considered statistically unreliable and were therefore avoided. Statistical significance was set at p < 0.05 (two-tailed).

The study protocol, including the use of the mHealth application and procedures for collecting application (app) usage data, was approved on 29 July 2024 by the Health Research Ethics Committee of Hasanuddin University (protocol no.: 3724093039). Before enrolment, all eligible participants received verbal and written information about the study objectives, procedures, potential risks and benefits, and each participant provided written informed consent specifically for participation in the study and for the use of the mHealth intervention.

A total of 22 participants were included in the study. The mean age was 46.3 ± 7.8 years (median 47; range 35–58 years). The majority were male (63.6%), and most had transtibial (below knee) amputation (68.2%). The mean duration since amputation was 6.8 ± 3.2 years (median 7; range 2–12 years). The mean BMI was 27.1 ± 3.9 kg/m² (median 27; range 21–34), indicating a trend towards overweight. According to World Health Organization (WHO) categories, more than half of the participants were overweight (54.5%), followed by obese (22.7%), whilst only 22.7% had normal BMI. Furthermore, educational attainment varied, with the majority (63.6%) having completed at least senior high school. The causes of amputation were varied, such as amputee (sports-related injuries and occupational accidents) and congenital limb deficiency. The majority of all participants used a transtibial or transfemoral prosthesis in daily life and wore their prosthesis for at least 8 h per day on weekdays. This makes the study population well-suited to the aims of the mHealth application, which was designed to provide education and core-strengthening exercises. Baseline characteristics are summarised in Table 1.

Baseline characteristics of participants are presented using descriptive analysis. Continuous variables are summarised as mean ± standard deviation or median (range), as appropriate, whilst categorical variables are presented as frequencies and percentages.

Of the 25 participants initially enrolled, 22 (88.0%) completed the full intervention and were included in the final analysis. Three participants (12.0%) were excluded because of insufficient attendance (≤ 3 sessions). Participants who completed the programme attended all scheduled supervised sessions and consistently used the mHealth application as prescribed.

Over the 3-month intervention period, both primary and secondary outcomes showed consistent within subject improvements. Descriptive analyses as shown in Figure 2 indicated significant overall changes across the four assessment points for quality of life, positive attitude and core muscle strength. The proportion of participants reporting good quality of life increased from 45.5% at baseline to 72.7% at month 3. Similarly, the proportion of participants with positive attitudes increased from 50.0% to 77.3%, and the proportion with good core muscle strength increased from 40.9% to 77.3%. Pairwise comparisons using Wilcoxon signed-rank tests showed significant improvements from baseline to month 3 in positive attitude and core muscle strength.

By contrast, perceived vulnerability and perceived seriousness did not demonstrate notable changes over the study period. This suggests that changes in quality of life occurred without corresponding changes in risk perception. Improvements in quality of life were observed concurrently with increases in positive attitudes and core muscle strength amongst lower-limb amputees using the mHealth application. Descriptive trends of these outcomes are presented in Figure 3.

Exploratory bivariate analysis using the chi-square test (Table 2) indicated that positive attitude and core muscle strength were associated with quality of life at the baseline. The result showed that participants with positive attitudes showed higher odds of reporting good quality of life (OR = 2.9; 95% CI: 1.0–8.6; p = 0.045), and those with good core muscle strength also showed higher odds (OR = 4.5; 95% CI: 1.3–12.4; p = 0.018). In contrast, perceived susceptibility (OR = 1.8; 95% CI: 0.7–5.0; p = 0.162) and perceived severity (OR = 1.2; 95% CI: 0.5–3.5; p = 0.287) were not significantly associated with quality of life. These findings suggest that, at the bivariate level, behavioural (positive attitude) and physical (core muscle strength) factors were associated with quality of life at the bivariate level.

The Wilcoxon signed-rank test was used to evaluate within participant changes between baseline and 3-month follow-up. Wilcoxon signed-rank analysis (Table 3) revealed that participant change across variables behavioural and physical. In the behavioural variables, positive attitude showed a statistically significant improvement with a W = 15.0, p = 0.007 and a large effect size (r = 0.77), indicating improvement across participants. High perceived vulnerability demonstrated a low W value (W = 0.0) and a large effect size (r = 0.88) but did not reach statistical significance (p = 0.083). Moreover, in the physical domain, good core muscle strength improved significantly (W = 13.0, p = 0.020) with a large effect size (r = 0.79), indicating substantial improvement between baseline and follow-up. In contrast, high perceived seriousness showed no significant change (W = 10.5, p = 0.981) despite a large effect size (r = 0.80), suggesting variability in individual responses.

This study has several important limitations. Firstly, the small sample size limited statistical power and generalisability. Secondly, dichotomising continuous WHOQOL-BREF scores into ‘good’ versus ‘poor’ quality of life likely reduced statistical power and may have masked more nuanced changes in quality of life. Thirdly, a quasi-experimental one-group pre-test–post-test longitudinal design lacking a control group precludes causal inference, and any observed benefits over time cannot be unequivocally ascribed to the mHealth intervention and may be partially as a result of natural adaptation or other unmeasured factors. Fourthly, although adherence was monitored using attendance records, detailed objective app usage metrics were not available. Several potentially relevant factors, including prosthesis type and prosthetic alignment or fit, were not measured in this study. These factors may influence functional outcomes and quality of life and should be considered in future studies.