We conducted a retrospective formative feasibility evaluation using deidentified foot imaging and compliance data obtained from a pilot RPM program for high-risk patients with DFUs operated by Empo Health. Records collected between April 2024 and November 2024 from adult patients enrolled in remote monitoring through 3 podiatric clinics in the Baltimore metropolitan area, Maryland, were retrospectively evaluated. This analysis included up to 7 months of RPM data per participant, consisting of a 1-month device break-in period, 3 months of structured daily use with formal compliance tracking, and an additional optional 3-month period of unstructured use.

This study was designed as a formative evaluation to assess device usability, patient compliance, and real-world feasibility; it was not designed or powered to evaluate definitive clinical outcomes such as ulcer incidence, hospitalization rates, or amputation prevention. The dataset included patient demographics, device use dates and daily compliance metrics, transmitted foot scans and their podiatrist-rated clinical usability, time-stamped clinical notes from podiatrist follow-up, and documentation of clinical interventions prompted by image review.

Monitoring services were provided following podiatrist recommendation and clinical assessment of patient eligibility. Eligibility criteria included a documented history of DFU, high-risk status, and confirmed internet connectivity. Each participant was provided with an internet-connected, full-color imaging scale (Empo Footprint) for in-home use (Figure 1). The device is a full-color, high-resolution, internet-connected foot imaging scale designed for a patient to stand on every day to capture high-resolution scans of the soles of their feet, as well as to measure their weight. These scans are automatically and securely transmitted to health care providers via a HIPAA (Health Insurance Portability and Accountability Act)-compliant web portal for remote clinician data review as part of the monitoring program.

Participants were monitored for up to 7 months as part of the RPM program from which this retrospective dataset was derived. During the first 4 months, participants were asked to use the scale daily. The first month served as a break-in period, followed by 3 months of formal compliance tracking for usability assessments. The scale was designed to be “senior ready,” requiring no advanced technical skills or device configuration by the user.

Remote visual assessments were conducted asynchronously by the participating podiatrists, who reviewed all transmitted foot scans on a daily basis. Image review was performed by the treating podiatrist for each participant’s site, ensuring clinical familiarity with each patient’s baseline foot condition. Consistent with real-world RPM practices, reviewers were not blinded to participant medical history or prior imaging. No independent adjudication or interrater reliability assessment was performed as this formative evaluation was designed to assess real-world clinical feasibility rather than standardized diagnostic accuracy. Clinical findings identified during image review were corroborated through in-person examination where applicable and documented in participant medical records.

As part of the image review process, podiatrists were asked to complete an optional standardized questionnaire for each foot scan confirming whether (1) the image quality was sufficient for remote foot assessment, (2) there were any changes compared to prior scans, and (3) any additional clinical actions were needed beyond continued monitoring. For any concerns (eg, poor image quality, observed changes, or recommended actions), the clinicians were asked to provide a brief explanation. This feedback was used both for evaluation purposes and to enable over-the-air firmware improvements during the course of the usability period.

Following completion of the initial 4-month usability period, participants were offered the option to continue using the device for an additional 3 months of free, unstructured monitoring. Of the 21 participants, 20 (95.2%) elected to continue device use during this extended period for the fifth month, and 18 (85.7%) elected to continue using the device for the remaining 2 months.

Usability was assessed descriptively through participant compliance with daily foot scan capture and podiatrist-rated image usability. Compliance was defined as the proportion of days in which participants successfully completed at least one foot scan relative to the total number of expected monitoring days during the compliance tracking period. The proportion of participant-months meeting or exceeding 16 days of device use per 30-day period was also assessed as a secondary compliance metric. The 16-day threshold was selected based on contemporaneous CMS RPM reimbursement standards, which at the time of the study required a minimum of 16 days of monthly device use to qualify for RPM billing [27]. Although this program was not reimbursed under CMS at the time, this threshold represented the most widely accepted benchmark for clinically meaningful RPM engagement and was therefore adopted as a reference standard for this evaluation. Image usability was defined as the proportion of foot scans deemed sufficient for clinical assessment based on podiatrist responses during image review.

Following completion of device use (up to 7 months per participant), a retrospective analysis was performed using the complete dataset that had been deidentified prior to access by the research team. This analysis incorporated data from both the formal usability period (months 2-4) and the extended monitoring period (months 5-7) to assess longer-term compliance patterns and instances of clinical action associated with image review.

This retrospective analysis of the complete 7-month dataset was reviewed and determined to be exempt under Title 45 of the Code of Federal Regulations part 46.104(d)(4) by the Advarra Institutional Review Board. The exemption determination was based on the analysis of existing, deidentified data collected during routine clinical care. In accordance with research ethics, all patient information, including foot scans, was deidentified before analysis.

The retrospective analysis included data from 21 patients aged 43 to 86 years (median 65 years, IQR 59-74; mean 65.3, SD 10.6 years; n=17, 81% male and n=4, 19% female) enrolled in remote monitoring (Table 1).

During the 3-month required use phase of the study, podiatrist remote image review identified 7 unique indicators of early ulcers among 33.3% (7/21) of the participants, leading to clinical actions. Table 2 summarizes all detected findings by type, clinical response, and outcome. These findings included bleeding (n=4), callus color changes prompting in-person evaluation (n=2), foreign body impaction (n=1), and subcutaneous ulcer development (n=1). All identified findings prompted clinical follow-up, ranging from telephone-based guidance to expedited in-person evaluation.

During the subsequent 3-month optional continuation period, compliance remained high, with device use on 82.2% (1554/1890) of possible days. In our findings, we present 3 selected cases (participants 1, 14, and 27) to illustrate the spectrum of findings and clinical responses facilitated by daily remote visual monitoring.

Compliance was assessed using rolling 30-day participant-months, with 3 participant-months analyzed per participant (N=21) following the initial break-in period. The 21 participants as a population adhered to near-daily use of their in-home foot monitoring scales, as shown in Figure 5. Across the study period, participants used the device on 82.5% (3638/4410) of all possible days. Median device use was 27 out of 30 days per participant-month, and mean use was 23 out of 30 days over the 6-month study period.

A total of 88% (111/126) of the 30-day participant-months met or exceeded the target device use of 16 days for 6 months (corresponding to a device use of more than half of the days in each month). Of the 7 instances in which participants did not meet the 16-day monthly use target, 3 reflected near-threshold compliance, with device use of 14 to 15 days. These metrics included any time participants were unavailable to step on their device due to travel, vacation, or other medical issues.

After 4 months of participation, 95.2% (20/21) of the participants completed a user experience survey in which they reported high levels of satisfaction with and ease of use of the remote foot monitoring scale. In total, 4.8% (1/21) of the participants were unable to complete the survey due to an unrelated hospitalization. Using a 7-point Likert scale where 1 represented the worst score and 7 represented the best score, most participants reported that they found the device physically easy to use (median 7.0, IQR 7-7; mean 6.9, SD 0.45), were willing to use it daily (median 7.0, IQR 7-7; mean 6.9, SD 0.31), and trusted the system’s ability to detect foot-related issues (median 7.0, IQR 7-7; mean 6.9, SD 0.32). Table 3 summarizes the survey results.

All missed device use days recorded during the study period are summarized in Table 4. Missed days were categorized based on whether a reason for nonuse was communicated to the study team in advance (expected) or occurred without prior notification (unexpected). Expected missed use days included participant-reported travel or medical incidents that prevented device use. Unexpected missed use days included travel or medical incidents that were identified after the missed day occurred and were not communicated to the study team in advance.

Missed use days for which no reason was documented were categorized as “no reason documented.” No missed use days were attributed to device failures or technical malfunctions during the study period. Transmission errors leading to missing images were not classified as missed device use days as participants used the device and weight data were captured; these events reflected incomplete data capture (missing image) rather than nonuse or device failure.

During the study period, podiatrists completed surveys for 3295 total foot scans and identified 3220 (97.7%) as usable (Table 5). The 2.3% (n=75) of the scans that were unusable were primarily attributed to patient-related factors or image quality issues, such as patients wearing socks or shoes or movement during the foot scan (n=53, 1.6%), and technical instances in which a patient stepped on the device but the scanning mechanism did not capture an image, resulting in only weight data being recorded (n=22, 0.7%).

Device performance was monitored throughout the study period. No hardware failures, safety incidents, or data security breaches were reported. Transmission failures were uncommon, observed in just 0.7% (22/3295) of instances. In these cases, the system recorded only weight data without capturing a foot scan, likely due to brief connectivity interruptions or improper user positioning. These events were captured in the image usability analysis and did not result in any gap in participant monitoring as daily scanning continued uninterrupted. No device replacements were required during the study period.

This retrospective formative study highlights the usability and real-world feasibility of daily remote foot monitoring via full-color images and its potential role in supporting early clinical assessment of foot changes in a high-risk diabetic foot population. Across 21 participants over 7 months, the device achieved an overall compliance rate of 82.5% (3638/4410; median 27/30, IQR 21-30 days per month), with 97.7% (3220/3295) of submitted images rated as clinically acceptable for remote assessment. Patient-reported usability ratings were consistently high, with participants reporting ease of use and willingness to continue daily monitoring. Notably, compliance remained high despite the study being conducted during the summer months, a period typically associated with disruptions to daily health routines. Podiatrists identified 7 clinically concerning findings across 7 unique participants, all of which prompted clinical follow-up ranging from telephone-based guidance to expedited in-person evaluation. These aggregate findings collectively support the feasibility and acceptability of integrating this RPM modality into routine care for high-risk patients.

Prior to the RPM program enrollment, podiatrists reported that these participants were typically noncompliant with performing their recommended diabetic foot care routines, including routine visual foot inspections, which contributed to delayed identification of skin changes and a history of ulceration and, in some cases, amputation. By enabling high-quality daily visual inspections, the scale-provided foot scans facilitated earlier detection of skin changes that may have otherwise progressed without timely intervention.

Three illustrative cases demonstrate the range of clinical scenarios in which daily visual monitoring facilitated timely clinical response: (1) enabling detection of acute changes (foreign body and skin tears) within 24 to 48 hours rather than at the next scheduled visit (typically 61 days apart under Medicare guidelines), (2) facilitating asynchronous review to distinguish findings requiring urgent in-person evaluation from those manageable through telehealth guidance, and (3) providing serial documentation to identify patterns (eg, recurrent skin breakdown) that may warrant care protocol modifications. However, as a single-arm observational study, these findings represent preliminary observations of feasibility rather than evidence of clinical effectiveness, which would require controlled comparison.

These cases also highlight the limitations of traditional periodic in-office examinations and reliance on in-home unassisted patient self-inspections. Foot ulcers more commonly develop in individuals with physical and sensory limitations, including impaired vision, obesity, limited mobility, and peripheral neuropathy [1]. Thus, unassisted patient self-inspections are often unreliable [26,28], as demonstrated in our case series: participant 14 was unaware of the foreign object in their foot, whereas participant 27 would not have appreciated the subtle changes in callus color. Additionally, poor access to care and communication with health care providers in between in-person visits also leave substantial gaps in care [1]. In our case series, participants were typically scheduled for in-office visits every 61 days, meaning that injuries detected via RPM may not have otherwise been identified until several weeks later. Daily visual monitoring provided by the imaging scale may help address several critical limitations of existing approaches by enabling visualization of acute changes within 24 to 48 hours rather than weeks between clinic visits, facilitating asynchronous clinical review and supporting bidirectional communication between patients and health care providers.

Acknowledging the inherent limitations of a single-arm, sponsor-funded feasibility study, the compliance rate of 82.5% (3638/4410) overall (median 27/30 days per month) in this study nonetheless compares favorably to that of prior RPM studies in chronic disease management, which have reported compliance rates ranging from 45.9% to 82.8% [20-23,29]. This level of engagement in our study was supported by strong patient ratings for ease of use and willingness to continue using the device. These data suggest the acceptability of integrating an RPM device into daily routines even for older adults or those with mobility challenges, as in our case series. Notably, this study was conducted during the summer months, a period typically associated with lower compliance to prescribed care routines due to travel, vacations, and disruptions to daily schedules [30,31]. Despite this, participants demonstrated high compliance with daily monitoring.

The clinical acceptance rate of 97.7% (3220/3295) of scanned foot images in this study was also encouraging and compares well to a prior diabetic foot surveillance study that used mobile phone images, where clinical adequacies (adequacy of images collected via self-photo, selfie stick, and another party) were 74.1%, 83.7%, and 92.6%, respectively [12].

The findings from this formative evaluation suggest that full-color image–based RPM may be a feasible and usable approach for supporting remote foot surveillance in high-risk patients. By enabling consistent, home-based visual assessment of the plantar surface, this modality addresses a known gap in existing RPM strategies, which have largely focused on indirect physiological measures rather than direct visualization of skin integrity. Prior studies of RPM interventions in diabetic foot care have demonstrated that earlier identification of concerning changes and timely clinical response may reduce ulcer progression and hospital admissions, with substantial cost savings [25,32]. Moreover, effective remote foot monitoring may contribute to maintaining patients’ mobility and independence, improving their quality of life, and reducing the risk of emotional and psychological impacts associated with chronic foot ulcers and amputation.

In this context, full-color image–based RPM may offer a complementary approach by facilitating remote clinical assessment of early skin changes that might otherwise go undetected. Although this study was not designed to evaluate clinical outcomes, the observed adherence, image reliability, and illustrative cases of early clinical action suggest that this approach warrants further prospective investigation. Future studies should evaluate whether integration of daily visual monitoring into routine care can improve clinical outcomes, reduce health care use, and enhance patient quality of life.

This study has several limitations. As a retrospective formative feasibility evaluation, it provides descriptive insights into the use of RPM in a small patient population. The relatively small sample size and lack of sufficient gender and racial diversity limit the generalizability of the results. The absence of a control group also limits the ability to directly compare outcomes and determine the full impact of the device. This study was conducted over 210 days, with the first 30 days being the break-in period for device use, and longer studies are necessary to confirm outcomes and demonstrate long-term usability. Although the compliance rates were high for this relatively short study, longer follow-up periods are also crucial to assess sustained compliance, particularly in less motivated patients. Additionally, most missed use days (514/810, 63.5%) had no documented reason, which is consistent with real-world RPM programs where participants are not required to report nonuse, but this limits the ability to distinguish device-related barriers from lifestyle factors.

Another limitation of the study design is that image review was performed by treating podiatrists who were not blinded to participant history, reflecting real-world RPM workflows in which clinical assessments are made with access to patient context. However, this approach introduces the potential for confirmation bias in clinical assessments. The absence of independent review and interrater reliability measurement limits the reproducibility of the findings. Future studies should include larger and more diverse populations and use more rigorous controlled study designs. Blinded, independent image adjudication and formal interrater reliability assessment should also be incorporated to enhance the validity, reproducibility, and generalizability of the clinical findings.

Prior studies have demonstrated that RPM for diabetic foot care may support earlier identification and management of skin changes. Daily screening of foot scans may enable health care providers to identify and address subtle changes in foot health before they escalate into more serious complications such as infections, hospitalizations, or amputations. In this study, high adherence, clinical image usability, and illustrative cases of early clinical action suggest that this approach is feasible and acceptable for home use. However, as this study was not designed to evaluate clinical outcomes, these findings should be interpreted as preliminary. Given the financial relationships between the study sponsors and most contributing authors, these preliminary findings should be interpreted with caution and are best viewed as hypothesis generating. Independent replication by investigators without financial ties to the technology is a critical next step. Further prospective studies are needed to determine whether integration of daily visual monitoring into routine care can improve clinical outcomes, reduce health care use, and enhance patient quality of life.