Smartphone-based abdominal inclination angles for assessing abdominal obesity: Convergent validity and reliability study

Abstract

Background

Abdominal obesity is associated with increased cardiometabolic risk, highlighting the need for regular monitoring. However, tape-based waist circumference measurements may be inconvenient for repeated self-monitoring and do not readily support digital recording. Given the widespread availability of smartphones, measuring abdominal inclination angles using a smartphone inclinometer (SI) may be a practical option for assessing abdominal obesity.

Objectives

This study aimed to investigate the convergent validity and reliability of SI-derived abdominal inclination angles obtained through examiner- and self-measurements among participants with obesity types ranging from normal weight to type II obesity.

Methods

Forty adults underwent abdominal inclination angle measurements using SI. Examiner- and self-measured upper abdomen (UA), lower abdomen (LA), and total abdomen (TA) angles were obtained. Convergent validity was evaluated using Spearman correlations with anthropometric indices. Intra- and inter-rater reliability were assessed using intraclass correlation coefficients, standard error of measurement, Bland–Altman analysis, and regression analysis for proportional bias and heteroscedasticity.

Results

Self-measured UA and TA angles showed stronger validity with anthropometric indices than examiner measurements. UA and TA angles demonstrated good-to-excellent intra-rater reliability, while inter-rater reliability was moderate. Bland–Altman analyses showed no proportional bias, although heteroscedasticity was observed in self-LA for intra-rater comparisons and in self-UA and self-LA for inter-rater comparisons.

Conclusions

SI measurements of abdominal inclination angles demonstrated acceptable validity and reliability, with consistent performance for UA and TA angles, whereas LA measurements showed greater variability and should be interpreted with caution. This method may provide a practical approach for monitoring abdominal obesity in daily life.

Keywords

abdominal obesity convergent validity measurement reliability smartphone inclinometer

Introduction

Obesity is a chronic condition characterized by excessive accumulation of adipose tissue and represents a major public health concern worldwide.¹ According to the global burden of disease, the global prevalence of general obesity based on body mass index (BMI) almost doubled between 1990 and 2021.^2,3 However, the prevalence of general obesity has decreased, while the prevalence of abdominal obesity continues to increase, as BMI does not reflect the accumulation of abdominal fat.^4–7 Abdominal obesity is also associated with a higher risk of diseases such as myocardial infarction, stroke, premature death, metabolic disorders, arterial stiffness, hypertension, and type 2 diabetes.^8,9 To predict the risk associated with abdominal obesity, the World Health Organization (WHO) recommends measuring abdominal obesity separately with BMI when assessing overall obesity.³ Given these health risks, regular self-monitoring of abdominal obesity is crucial in daily life for effective prevention and management.

For assessing abdominal obesity, computed tomography (CT) and magnetic resonance imaging (MRI) measurements are considered accurate methods in hospitals and clinical settings.⁹ However, these approaches are costly, time-consuming, and impractical for regular measurements of abdominal obesity.¹⁰ As a more accessible and cost-effective alternative method, waist circumferences (WC), hip circumferences (HC), and waist-to-hip ratio (WHR) measurements are widely used for abdominal obesity measurement using tape equipment, according to the guidelines of the WHO, in both clinical settings and at home.^3,9 However, tape measurement tends to underestimate WC when used as a self-measurement method, and individuals with abdominal obesity often experience difficulty in self-measuring WC and HC due to abdominal protrusion.^11,12

Recently, mobile health technology has been developed and introduced for regular measurement of abdominal obesity in daily life. As part of the mobile health technology field for self-assessment of abdominal obesity, a previous study demonstrated excellent validity and reliability of using a convolutional neural network-based smartphone camera for abdominal obesity measurement.^13–15 However, self-measurement of abdominal obesity using smartphone camera-based applications has several limitations, including the need to rearrange furniture and wear tight-fitting clothing during image capture, as well as privacy concerns related to photographing the body.^13,16 Additionally, to ensure accurate measurement, users should adjust their position and distance to ensure proper framing of the smartphone camera, which may be difficult in limited home spaces.^13,16 As one of the alternative digital healthcare technologies, the wearable sensor is still commonly used for abdominal obesity monitoring, without the limitations associated with computer vision artificial intelligence.¹⁷ A previous study developed a belt-type wearable device for monitoring abdominal circumference using a pressure sensor in the belt strap and estimating the inclined angle of the abdomen using an accelerometer in the belt buckle.¹⁷ Although this innovative approach provides self-monitoring for abdominal obesity management, this device has a lower penetration rate than smartphones in daily life, while smartphone inclinometers (SI) apps have the potential to measure the inclined angle of the abdomen, because previous studies also demonstrated that SI can measure the inclined angle of the body such as alignment or deformity of the spine and the range of motion (ROM) in the shoulder, hip, knee, and ankle.^18,19

Given that the SI app is commonly used to measure the inclination of the body, the present study applied the SI app to assess abdominal inclination angles from the side view as an indicator of abdominal protrusion. Abdominal inclination angles measured using SI may reflect external abdominal contour associated with central fat accumulation rather than directly quantifying internal adipose tissue volume. The application of SI may provide a practical method for examiners to assess abdominal obesity and also serve as a convenient, accessible alternative to tape measures or BMI for self-monitoring at home. To examine the relationship between SI-derived abdominal inclination angles and standard anthropometric indicators of obesity, WC, HC, WHR, and BMI were used as reference measures.³ Although BMI does not directly reflect abdominal fat distribution, it was included because it is routinely used in health assessments.³ The present study aimed to examine the convergent validity of examiner- and self-measurements of abdominal inclination angles using SI compared with WC, HC, WHR, and BMI, and to evaluate the intra- and inter-rater reliability of these measurements. This study hypothesized that measurements of abdominal obesity using SI would be valid and reliable in both examiner- and self-measurements.

Methods

Participants

The sample size was estimated using G*Power (3.1.9.7; Universität Kiel, Düsseldorf, Germany), with a two-tailed test, a significance level (α) of 0.05, and 80% power, resulting in a total sample size of 33 subjects. To account for potential dropouts, a total of 40 general individuals aged 20 to 40 years participated in the study. The age range was restricted to minimize age-related variability in body composition and fat distribution that could influence abdominal contour and inclination measurements. Participants were included from normal weight (18.5–24.9 kg/m²) to obesity type II (35–39.9kg/m²) based on the WHO criteria of the BMI.^3,20 The exclusion criteria for the study were as follows: 1) underweight (BMI < 18.5 kg/m²) or obesity type III (BMI ≥ 40 kg/m²); 2) participants who experienced a weight change greater than ±2 kg in the past month; 3) pregnant or breastfeeding women; 4) conditions affecting fat distribution (e.g., cushing’s syndrome and metabolic syndrome); 5) currently taking specific medications such as diet supplements; 6) participants who were unable to stand for abdominal obesity measurements due to postural imbalance or congenital spinal deformities.⁹ The general characteristics of the participants are summarized in Table 1. All participants gave written informed consent to the investigation, and the protocol was approved by the institutional review board of Yonsei university.

Table 1.

General characteristics and obesity classification of the participants.

	Total (n = 40)	Male (n = 25)	Female (n = 15)
Age (years)	26.51 ± 3.89^a	26.84 ± 4.33	25.8 ± 2.95
Height (cm)	169.95 ± 6.67	174.02 ± 4.04	162.74 ± 3.22
Weight (kg)	74.50 ± 18.24	85.41 ± 12.98	54.77 ± 4.82
BMI^b (kg/m2)	25.51 ± 4.81	28.15 ± 3.86	20.67 ± 1.63
18.5–24.9 kg/m2 (Normal)	20 (50%)^c	5 (20%)	15 (100%)
25–29.9 kg/m2 (Overweight)	15 (37.5%)	15 (60%)	0 (0%)
30–34.9 kg/m2 (Obesity Type I)	4 (10%)	4 (16%)	0 (0%)
35–39.9 kg/m2 (Obesity type II)	1 (2.5%)	1 (4%)	0 (0%)

^aMean ± Standard deviation.

^bBody mass index.

^cNumber of participants (percentage).

Instruments

The height and weight of the participants were measured using a stadiometer (Samhwa Instrument Factory, South Korea) and a digital weight scale (CAS, HE-70, South Korea), respectively. The current study used a smartphone application (Angle Finder, JRSoftWorx, Berlin) for the measurement of the SI angle, which is available on the iPhone app store. The smartphone (iPhone, Apple, Cupertino, CA) was used to measure the inclination angles of the upper abdomen (UA), lower abdomen (LA), self-upper abdomen (self-UA), and self-lower abdomen (self-LA). WC and HC were measured using a tape (SECA, Hamburg, Germany) to demonstrate the validity of the abdominal SI measurements.

Procedures

In this study, three examiners underwent a one-month training session prior to initiating data collection using tape and SI. Tape measurements (examiner 1) and SI measurements (examiners 2 and 3) were then performed on the same day.²¹ All examiners were blinded to each other’s results and conducted the measurements independently in separate spaces to minimize potential observer bias. The measurement procedures were as follows: 1) collection of demographic information (gender and age); 2) measurement of anthropometric variables (weight, height, and BMI) and tape-based parameters (WC, HC, and WHR) for validity assessment; and 3) examiner- and self-measurements of abdominal inclination angles (UA, LA, and total abdomen angle [TA]) to assess intra- and inter-rater reliability.

Measurements of body mass index and abdominal obesity using tape

Measurements of each SI angle were compared with BMI and tape measurement values to confirm validity. Participants were instructed to avoid wearing thick clothing and to remove any external accessories (e.g., waist belts and navel piercings) before measurements. To assess convergent validity, BMI was calculated as weight (kg) divided by height squared (m²).³ For tape measurements, participants stood in a relaxed posture of the trunk and both upper and lower limbs, and measurements were obtained at the end of a normal expiration without bending or twisting the body. To measure WC, the midpoint between the 12th rib and the top of both iliac crests was palpated and measured using a tape (Figure 1(a)).³ HC was measured by wrapping the tape around the widest portion of the buttocks (Figure 1(b)).³ WC and HC values were measured to the nearest 0.1 cm using a tape, following WHO guidelines. The WHR was calculated using the formula WC/HC.³ Each tape measurement conducted by examiner 1 was repeated three times at 10-second intervals within a single session, and the mean value was used as the reference for convergent validity analysis of the SI measurements.²²

Figure 1.

Tape measurement of the waist (a) and hip circumferences (b).

Examiner-measurements of abdominal inclination angle using a smartphone inclinometer

During all SI angle measurements, participants stood in a relaxed standing posture with their feet positioned approximately shoulder-width apart and knees fully extended, with the head facing forward. Measurements were obtained at the end of a normal expiration, following the same protocol as the tape-based measurements, while avoiding bending or twisting of the trunk and limbs. Prior to each measurement, the SI was calibrated to 0 degrees by placing it on a flat surface. The SI angle was automatically recorded when the smartphone was held steady for 2 seconds without any movement, with the value recorded to the nearest 0.1 degree. For examiner- and self-measurements, the smartphone was positioned at the UA and LA regions. To evaluate within-day reliability, each angle was measured three times at 10-second intervals during the first session.^21,22 After a 5-minute break, the same procedure was repeated in a second session. For each session, the mean of the three trials was calculated, and the session means were used for reliability analysis.^23,24 To assess inter-rater reliability, the mean of three measurements obtained by examiner 3 was compared with the mean of three measurements from examiner 2’s first session.^24,25

Measurement of upper and lower abdomen angles

The SI measurement of the UA inclination angle was set by placing the upper edge of the inferior border of the smartphone on the xiphoid process and simultaneously positioning the bottom surface of the charging port area in contact with the abdomen (Figure 2(a)). The SI measurement of the LA inclination angle was set by placing the lower edge of the inferior border of the smartphone on the midpoint between both anterior superior iliac spines (ASIS) and simultaneously positioning the entire bottom surface of the charging port area in full contact with the abdomen (Figure 2(b)).

Figure 2.

Measurements of the upper abdomen (a) and lower abdomen angles (b) using a smartphone inclinometer app and calculation of total abdomen angle (c).

Calculation of total abdomen angle

To calculate the total abdomen (TA) angle, the UA and LA angles were summated (Figure 2(c)), similar to the previous methods using SI for measuring spinal angle.²⁶ The calculated TA angle does not assume strict geometric linearity but was intended to reflect the combined contribution of upper and lower abdominal protrusion.

Self-measurements of abdominal inclination angle using a smartphone inclinometer for convergent validity and reliability

Self-measurements were performed using the same procedure as examiner measurements. Because the participants were not professionals, they were unfamiliar with the anatomical landmarks required for accurate measurement. The examiner explained smartphone positioning using simple descriptions rather than medical terminology. The xiphoid process was described as “the bottom of the breastbone”, and the midpoint between both ASIS was described as “the center point between the bony points at the front of the hips”. After these explanations, the examiner demonstrated the self-measurement procedure. Each participant performed three trials at 10-second intervals during two separate sessions. The mean of three trials per session was used to assess intra-rater reliability.^21–25

Statistical analysis

Convergent validity of abdominal inclination angles using a smartphone inclinometer

The Shapiro-Wilk test was used to assess the normality of the data. Except for self-UA, all variables showed P ≤ .05, indicating a violation of normality. Spearman correlation analysis was then performed to evaluate the validity of the average values of each SI angle variable (UA, LA, TA, self-UA, self-LA, and self-TA) obtained during the first session, compared to WC, HC, WHR, and BMI. Correlation coefficients of ρ < 0.25 were considered as mild; 0.25–0.50 as moderate; 0.50–0.75 as good; and > 0.75 as excellent.^27,28

Reliability of abdominal inclination angles using a smartphone inclinometer

Intra- and inter-rater reliability analyses of SI angle measurements were performed using the intraclass correlation coefficient (ICC).^27,29 Given the non-normal distribution of the data, 95% confidence intervals for ICC were computed using non-parametric bootstrap methods with 1,000 replications. To assess intra-rater reliability of both each examiner- and self-measurements, ICC (3, 3) was calculated using the average values from the first and second measurement sessions for each SI angle variable.²⁹ To assess inter-rater reliability of examiner-measurements, ICC (2, 3) was calculated by comparing the average values obtained by examiner 3 with those from the first measurement session of examiner 2.²⁹ ICC was considered as follows: poor, < 0.4; moderate, 0.4–0.75; good, 0.75–0.9, and excellent, > 0.9.^27,29 To quantify absolute measurement precision, the standard error of measurement (SEM) and minimal detectable change (MDC) at the 95% confidence level were calculated for each SI angle. Bland–Altman analyses were additionally performed to assess agreement between sessions (intra-rater) and between examiners (inter-rater), including the mean bias and 95% limits of agreement. To further evaluate potential proportional bias, linear regression of the differences against the means was performed for each Bland–Altman comparison. Heteroscedasticity was assessed using Spearman correlation between the absolute differences and the corresponding means. All statistical analyses were conducted using Google Colab.

Results

Participant characteristics, including obesity classifications based on BMI according to WHO criteria, were presented in Table 1.

Convergent validity of abdominal inclination angles using a smartphone inclinometer

The convergent validity of self-measurements with a good-to-excellent level was higher than that of examiner-measurements with a moderate-to-good level for the inclination angle of UA and TA, compared with WC, HC, WHR, and BMI (Figure 3). The lower abdomen (LA) angle, measured by both examiner-administered and self-administered methods, showed mild-to-good convergent validity (Figure 3).

Figure 3.

Heatmap showing the validity of abdominal inclination angle measurements using a smartphone inclinometer. UA: Upper abdomen, LA: Lower abdomen, TA: Total abdomen, S-UA: Self-measured upper abdomen, S-LA: Self-measured lower abdomen, S-TA: Self-measured total abdomen, WC: Waist circumference, HC: Hip circumference, WHR: Waist-to-hip ratio, and BMI: Body mass index.

Intra-rater reliability of abdominal inclination angles using a smartphone inclinometer

Intra-rater reliability demonstrated good-to-excellent reproducibility across all SI angles, with ICC values ranging from 0.86 to 0.95 for both examiner- and self-measured assessments (Table 2). The SEM values for examiner-measured angles were small (UA = 1.99°, LA = 2.40°, TA = 3.16°), indicating minimal absolute measurement error. In contrast, self-measured angles showed slightly larger SEM values (Self-UA = 2.21°, Self-LA = 2.85°, Self-TA = 4.01°). The MDC values ranged from 5.53° to 8.77° for examiner-measured angles, whereas self-measured MDC values ranged from 6.11° to 11.11° (Table 2). Bland–Altman analyses for examiner-based measurements showed small mean biases (UA: −0.06°, LA: −0.38°, TA: −0.44°) with 95% limits of agreement ranging from −5.59° to 5.47° for UA, −7.04° to 6.27° for LA, and −9.21° to 8.32° for TA. Self-measured angles demonstrated similarly small biases (Self-UA: 0.51°, Self-LA: 0.99°, Self-TA: 1.50°), although with broader but still acceptable limits of agreement (Self-UA: −5.61° to 6.62°, Self-LA: −6.89° to 8.88°, Self-TA: −9.61° to 12.61°) (Figure 4). Linear regression of the differences against the means showed no significant proportional bias across all intra-rater comparisons (p > 0.05). However, Spearman correlation analysis revealed significant heteroscedasticity only in self-LA measurements (ρ = 0.43, p < 0.01), whereas no significant heteroscedasticity was observed in the other comparisons.

Table 2.

Reliability of examiner- and self-measured abdominal obesity using a smartphone inclinometer.

		ICC^a (95% CI)†	SEM^b	MDC^c
Intra-rater reliability	UA^d	0.95 (0.92 to 0.97)	1.99	5.53
	LA^e	0.90 (0.82 to 0.95)	2.40	6.65
	TA^f	0.94 (0.88 to 0.97)	3.16	8.77
	Self-UA	0.93 (0.86 to 0.97)	2.21	6.11
	Self-LA	0.86 (0.74 to 0.94)	2.85	7.89
	Self-TA	0.91 (0.83 to 0.96)	4.01	11.11
Inter-rater reliability	UA	0.72 (0.50 to 0.89)	5.28	14.62
	LA	0.57 (0.23 to 0.78)	4.59	12.72
	TA	0.76 (0.57 to 0.89)	5.94	16.46

^aICC: Intraclass correlation coefficient.

^bSEM: Standard error of measurement.

^cMDC: Minimal detectable change.

^dUA: Upper abdomen.

^eLA: Lower abdomen.

^fTA: Total abdomen.

^†Calculated using a non-parametric bootstrap method.

Figure 4.

Intra-rater Bland–Altman plots for abdominal inclination angles. Bland–Altman plots show intra-rater agreement for examiner- and self-measured UA (upper abdomen), LA (lower abdomen), and TA (total abdomen) angles. Bias and 95% limits of agreement (±1.96 standard deviation) are shown for each comparison.

Inter-rater reliability of abdominal inclination angles using a smartphone inclinometer

Inter-rater reliability showed moderate agreement for UA (ICC = 0.72) and TA (ICC = 0.76), whereas LA demonstrated lower reliability (ICC = 0.57) with wider bootstrap confidence intervals (Table 2). SEM values were higher for inter-rater comparisons (UA = 5.28°, LA = 4.59°, TA = 5.94°) than for intra-rater analyses. The MDC values were also larger in inter-rater testing, ranging from 12.72 to 16.46° (Table 2). Bland–Altman analyses showed small mean biases (UA: −0.06°, LA: −2.28°, TA: −1.54°), but wide limits of agreement, indicating substantial between-examiner variability (UA: −14.68° to 14.56°, LA: −14.99° to 10.43°, TA: −17.99° to 14.93°) (Figure 5). Linear regression of the differences against the means demonstrated no significant proportional bias across inter-rater comparisons (all p > 0.05). However, significant heteroscedasticity was observed in UA (ρ = 0.43, p < 0.01) and LA measurements (ρ = 0.52, p < 0.01), whereas no significant heteroscedasticity was found in TA (ρ = 0.15, p = 0.34).

Figure 5.

Inter-rater Bland–Altman plots for abdominal inclination angles. Bland–Altman plots show agreement for UA (upper abdomen), LA (lower abdomen), and TA (total abdomen) angles between rater 2 and rater 3. The mean bias and 95% limits of agreement are presented for each comparison.

Discussion

This study investigated the convergent validity and reliability of abdominal obesity measurements using SI as a practical alternative to traditional tape-based methods and BMI. The key findings were that both UA and TA angles demonstrated good-to-excellent validity and intra-rater reliability, whereas LA measurements showed lower validity and greater variability. In addition, self-measured UA and TA angles showed stronger convergent validity with anthropometric indices than examiner measurements, while reliability estimates were generally comparable. Given the widespread availability of smartphones, SI-based assessment of UA and TA angles may provide a feasible and accessible method for routine home-based monitoring of abdominal contour associated with abdominal obesity.

An important aspect of this study is the use of SI as a novel method for assessing abdominal obesity, demonstrating good-to-excellent validity for UA and TA measurements, including self-measurements, compared to WC, HC, WHR, and BMI, which are widely used measures for assessing abdominal obesity in clinical and epidemiological studies.^3,30 Similarly, self-measured abdominal obesity using a smartphone camera with a CNN model to capture the abdominal outline has demonstrated excellent validity, as the predicted values of WC, HC, and waist-to-height ratio from the outline data showed strong agreement with the tape-based reference measurements and slightly higher validity than SI-based measurements.^14,15 The slight difference in validity between SI and smartphone camera-based methods may be attributed to differences in the measurement principles. While both tape and smartphone camera-based methods typically assess abdominal protrusion by measuring maximal horizontal distances in millimeters,¹³ SI quantifies the inclination angle of the abdominal region in the lateral view in degrees. This methodological difference may explain why smartphone camera-based methods have consistently demonstrated excellent validity, whereas SI-based measurements showed slightly lower, yet still good-to-excellent, validity. However, smartphone camera-based methods require a controlled environment, adequate lighting, and a fixed distance between the user and the device, and may also raise privacy concerns due to body image exposure.^13,16 In contrast, SI-based measurements are less affected by environmental factors and offer greater convenience in daily life, as users only need to place the smartphone over the abdomen. Given the widespread use of smartphones and the validated performance of SI demonstrated in this study, SI-based measurements may serve as a practical option for abdominal assessment in personal or home settings.

Examiner- and self-measurements of abdominal obesity using SI showed good-to-excellent intra-rater reliability, with SEM and MDC values indicating acceptable absolute measurement error, although slightly larger values were observed for self-measurements. In addition, regression analysis of the Bland–Altman differences against the means indicated no significant proportional bias across intra-rater comparisons, suggesting stable measurement agreement across the range of abdominal inclination angles. The intra-rater reliability reported in previous studies using SI for measuring the range of knee extension/flexion and shoulder external rotation has demonstrated good-to-excellent ICC levels.^19,30–32 Similarly, examiner-measurements of abdominal obesity in the present study demonstrated excellent ICC values when standardized measurement procedures were followed. In self-measurement using SI, self-UA and self-TA also showed excellent intra-rater reliability, whereas self-LA demonstrated a good level of reliability. Heteroscedasticity was identified in self-LA measurements, indicating increased variability at higher abdominal angles. This pattern likely reflects the greater curvature of the lower abdominal contour in individuals with increased abdominal protrusion. Compared with flatter abdominal surfaces, a curved contour may make it more difficult to position the smartphone consistently during self-measurement, thereby increasing measurement variability. Additionally, because measurements were performed over light clothing rather than directly on the skin, clothing folds, waistband compression, and fabric tension may alter the lower abdominal contour and introduce additional measurement variability. Standing posture may also have influenced the reliability estimates, as gravitational loading, postural alignment, and lumbar curvature can affect abdominal contour. Although a supine posture may reduce some gravitational effects, it does not represent functional abdominal protrusion during daily activities. Because standing is a natural and convenient posture for routine self-assessment, similar to using a weighing scale or tape measure, examiner-based SI measurements of UA and TA can be reliably performed in the standing position. For self-measurement, TA provided the most stable measurements, while UA also demonstrated acceptable reproducibility despite slightly greater variability than examiner-based assessments. In contrast, self-LA measurements exhibited heteroscedasticity in the intra-rater analysis, indicating increased variability at higher abdominal angles and suggesting that these measurements should be interpreted with greater caution.

In contrast, inter-rater reliability of SI-based measurements for abdominal obesity was lower than intra-rater reliability, although no proportional bias was observed. Larger SEM and MDC values and wider Bland–Altman limits of agreement in inter-rater comparisons indicate greater variability between examiners. Heteroscedasticity observed in UA and LA measurements further suggests that variability increased as abdominal angle values became larger. These findings are consistent with previous studies using smartphone inclinometers to assess thoracic kyphosis, which also reported lower inter-rater reliability due to differences in examiner proficiency and anatomical landmark identification.^18,24,25 Overall, these results indicate that SI-based abdominal angle measurements are most reliable when performed by the same examiner, supporting their use for repeated assessments within the same session.

This study has some limitations. First, the limited representation of individuals with severe obesity, particularly women, and the restriction of the sample to adults aged 20-40 years may limit the generalizability of the findings to populations with greater abdominal adiposity or older adults. Because abdominal contour and fat distribution patterns are known to differ according to sex, the relatively limited representation of women may further restrict the applicability of the findings to populations with different sex distributions. Second, the validity of SI measurements was not examined against imaging-based techniques such as CT, the gold standard for quantifying visceral adipose tissue. Instead, validity was assessed using anthropometric indices (WC and WHR), which are commonly used surrogate markers of abdominal obesity. Future studies should examine whether SI-derived abdominal angles can be used to estimate visceral fat mass and develop machine learning models for classifying abdominal obesity severity based on CT-derived labels. Third, measurements were performed over light clothing rather than directly on the skin, which may have affected abdominal contour and angular readings, particularly in the lower abdomen. In addition, self-measurements may be influenced by learning effects, postural variability, and minor differences in smartphone placement. Lastly, the TA angle should be interpreted as a simplified composite metric rather than a validated geometric model of abdominal morphology, as abdominal contour is influenced by nonlinear soft tissue characteristics and the additive structure of the TA angle may not fully capture the biomechanical complexity of abdominal protrusion. Future research may consider continuously measuring inclination angles along the abdominal contour from the UA to the LA under standardized conditions or directly on the skin. This approach may provide a more detailed abdominal contour profile and better capture the nonlinear characteristics of abdominal protrusion than a simplified two-point composite metric.

Conclusion

The present findings indicate that smartphone inclinometer measurements of abdominal inclination angles demonstrate acceptable validity and reliability for assessing abdominal contour associated with abdominal obesity. UA and TA angles showed consistent performance, whereas LA measurements demonstrated greater variability and should be interpreted with caution. SI-based assessment may serve as a practical option for monitoring abdominal obesity in daily and home settings, particularly when repeated by the same individual. Integrating this smartphone-based approach into digital health platforms may improve accessibility and user engagement in abdominal obesity monitoring without requiring additional devices.

Footnotes

Acknowledgements

We thank the participants of the study for their commitment and efforts.

ORCID iD

Kyuenam Park

Ethical considerations

This study was approved by the institutional review board of Yonsei university (7001988-202512-HR-2377-06).

Author contributions

YH Kwon and KN Park were responsible for conceptualization, study design, supervision, project administration, and preparation of the original draft of the manuscript, including all processes up to the initial submission. SR Seo and KN Park were responsible for manuscript revision, editing, and all subsequent revisions during the peer-review process. All authors read and approved the final version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Yonsei University Humanities and Social Sciences Field Creative Research Fund of 2025 (grant number No. 2025-22-0522).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Guarantor

Kyuenam Park.

References

Kopelman

. Obesity as a medical problem. Nature 2000; 404: 635–643. https://doi.org/10.1038/35007508

Global, regional, and national prevalence of adult overweight and obesity, 1990-2021, with forecasts to 2050: a forecasting study for the Global Burden of Disease Study 2021. Lancet 2025; 405: 813–838.

Consultation

. Waist circumference and waist-hip ratio. Report of a WHO Expert Consultation. Geneva: World Health Organization 2008; 2008: 8–11.

Wang

Liu

, et al. Epidemiology of general obesity, abdominal obesity and related risk factors in urban adults from 33 communities of Northeast China: the CHPSNE study. BMC Public Health 2012; 12: 967. https://doi.org/10.1186/1471-2458-12-967

Sun

Huang

Hua

, et al. Trends in general and abdominal obesity in US adults: Evidence from the National Health and Nutrition Examination Survey (2001-2018). Front Public Health 2022; 10: 925293. https://doi.org/10.3389/fpubh.2022.925293

Aranceta-Bartrina

Pérez-Rodrigo

Alberdi-Aresti

, et al. Prevalence of general obesity and abdominal obesity in the Spanish adult population (Aged 25-64 Years) 2014-2015: The ENPE Study. Rev Esp Cardiol (Engl Ed) 2016; 69: 579–587. https://doi.org/10.1016/j.rec.2016.02.009

Jeong

Jung

Yang

, et al. obesity fact sheet: prevalence of obesity and abdominal obesity in adults, adolescents, and children in Korea from 2012 to 2021. J Obes Metab Syndr 2024; 33: 27–35. https://doi.org/10.7570/jomes24012

Hofker

Wijmenga

. Apple or pear: size and shape matter. Cell metabolism 2015; 21: 507–508. https://doi.org/10.1016/j.cmet.2015.03.016

Browning

Mugridge

Chatfield

, et al. Validity of a new abdominal bioelectrical impedance device to measure abdominal and visceral fat: comparison with MRI. Obesity (Silver Spring) 2010; 18: 2385–2391. https://doi.org/10.1038/oby.2010.71

10.

Schmidt

Dinter

Reiser

, et al. The uses and limitations of whole-body magnetic resonance imaging. Dtsch Arztebl Int 2010; 107: 383–389. https://doi.org/10.3238/arztebl.2010.0383

11.

Song

Jensen

, et al. The perfect waist circumference measurement device–An easy and accurate self-measurement approach. Obesity 2022; 30: 117–118.

12.

Jaeschke

Steinbrecher

Pischon

. Measurement of waist and hip circumference with a body surface scanner: feasibility, validity, reliability, and correlations with markers of the metabolic syndrome. PLoS One 2015; 10: e0119430. https://doi.org/10.1371/journal.pone.0119430

13.

Farina

Orlandi

Lukaski

, et al. Digital single-image smartphone assessment of total body fat and abdominal fat using machine learning. Sensors (Basel) 2022; 22: 8365. https://doi.org/10.3390/s22218365

14.

Choudhary

Iyer

Smith

, et al. Development and validation of an accurate smartphone application for measuring waist-to-hip circumference ratio. NPJ Digit Med 2023; 6: 168. https://doi.org/10.1038/s41746-023-00909-5

15.

Graybeal

Brandner

Tinsley

. Evaluation of automated anthropometrics produced by smartphone-based machine learning: a comparison with traditional anthropometric assessments. Br J Nutr 2023; 130: 1077–1087. https://doi.org/10.1017/S0007114523000090

16.

Garbett

Degutyte

Hodge

, et al. Towards understanding people’s experiences of ai computer vision fitness instructor apps. In: Proceedings of the 2021 ACM Designing Interactive Systems Conference, 2021, pp. 1619–1637.

17.

Matsuda

Hasegaway

Arai

, et al. Waistonbelt 2: A belt-type wearable device for monitoring abdominal circumference, posture and activity. In: 2016 Ninth International Conference on Mobile Computing and Ubiquitous Networking (ICMU). , 2016, pp. 1–6.

18.

Elpeze

Usgu

Yiğit

. Reliability of the smartphone application inclinometer and flexicurve in measuring thoracic kyphosis. Cureus 2023; 15: e35886. https://doi.org/10.7759/cureus.35886

19.

Mourcou

Fleury

Diot

, et al. Mobile phone-based joint angle measurement for functional assessment and rehabilitation of proprioception. Biomed Res Int 2015; 2015: 328142. https://doi.org/10.1155/2015/328142

20.

Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19·2 million participants. Lancet 2016; 387: 1377–1396.

21.

Shin

Ro du

Lee

, et al. Within-day reliability of shoulder range of motion measurement with a smartphone. Man Ther 2012; 17: 298–304. https://doi.org/10.1016/j.math.2012.02.010

22.

Cools

De Wilde

Van Tongel

, et al. Measuring shoulder external and internal rotation strength and range of motion: comprehensive intra-rater and inter-rater reliability study of several testing protocols. J Shoulder Elbow Surg 2014; 23: 1454–1461. https://doi.org/10.1016/j.jse.2014.01.006

23.

Caña-Pino

Espejo-Antúnez

Adsuar

, et al. Test-retest reliability of an iPhone(®) inclinometer application to assess the lumbar joint repositioning error in non-specific chronic low back pain. Int J Environ Res Public Health 2021; 18: 2489. https://doi.org/10.3390/ijerph18052489

24.

Lewis

Valentine

. Clinical measurement of the thoracic kyphosis. A study of the intra-rater reliability in subjects with and without shoulder pain. BMC Musculoskelet Disord 2010; 11: 39. https://doi.org/10.1186/1471-2474-11-39

25.

van Baalen

Vanwanseele

Venter

. Reliability and validity of a smartphone device and clinical tools for thoracic spine mobility assessments. Sensors (Basel) 2023; 23: 7622. https://doi.org/10.3390/s23177622

26.

Takayama

Miyawaki

Nakamura

. Reliability of a measurement method for upper limb raised standing spinal alignment using a smartphone inclinometer application. Cureus 2024; 16: e76225. https://doi.org/10.7759/cureus.76225

27.

Portney

. Foundations of clinical research: applications to evidence-based practice. F. A. Davis, 2020.

28.

Khemthong

Packer

Passmore

. Classification of leisure participation in women with chronic conditions: a measurement tool. Annals of Leisure Research 2007; 10: 105–121. https://doi.org/10.1080/11745398.2007.9686757

29.

Trevethan

. Intraclass correlation coefficients: clearing the air, extending some cautions, and making some requests. Health Services and Outcomes Research Methodology 2017; 17: 127–143. https://doi.org/10.1007/s10742-016-0156-6

30.

Barrios

Martin-Biggers

Quick

, et al. Reliability and criterion validity of self-measured waist, hip, and neck circumferences. BMC Med Res Methodol 2016; 16: 49. https://doi.org/10.1186/s12874-016-0150-2

31.

Chew

Woon

Miao-Shi Low

, et al. Feasibility, reliability and validity of self-measurement of knee range-of-motion using an accelerometer-based smartphone application by patients with total knee arthroplasty. PLoS One 2024; 19: e0307219. https://doi.org/10.1371/journal.pone.0307219

32.

Mitchell

Gutierrez

Sutton

, et al. Reliability and validity of goniometric iPhone applications for the assessment of active shoulder external rotation. Physiother Theory Pract 2014; 30: 521–525. https://doi.org/10.3109/09593985.2014.900593