Measuring up: Ensuring Intra- and Interobserver Reliability in Stretched Penile Length with the SPLINT Technique Authors: Prabudh Goel, Prativa Choudhury, Vivek Verma, Shivani Phugat, Mani Kalaivani, Teg Rabab Singh, Kanika Sharma, Vishesh Jain, Devendra Kumar Yadav, Anjan Kumar Dhua, Sandeep Agarwala Published: 2024 ABSTRACT ABSTRACTBackground:A discrepancy between the true and measured value of stretched penile length (SPL) may be a result of errors that can either be systematic or random. Hence, it becomes important to focus on the quality of measurements to prevent any iatrogenic harm to the patients.Objective:The objective of this study was to assess the magnitude of intra- and interobserver variations in the measurement of SPL with the SPLINT technique.Materials and Methods:SPL was measured prospectively in a cohort of 449 boys aged 0–14 years including 68 infants (substratified into Group I: >4 years, Group II: 4–8 years, and Group III: >8 years) with the SPLINT technique by expert (E: E1 and E2) and trainee (T: T1 and T2) surgeons after completing a three-tiered training module. Intra- and interobserver variability was assessed through descriptive statistics, intraclass correlation (ICC), relative technical error of measurement (rTEM), and reliability or R (%).Results:Intraobserver variability: the mean difference between the two readings (E1 and E2) is 0.08 cm (95% confidence interval [CI]: 0.073–0.087), ICC was 0.998 (95% CI: 0.997–0.998), and intraobserver variability ≤0.1 cm in 85% of the participants (n = 370 of 433). The rTEM and reliability (%) were 1.82% and 98.1% (Group I), 1.65% and 98.9% (Group II), and 1.09% and 99.7% (Group III), respectively. The intraobserver variability was observed to be inversely proportional to the age of the participants (correlation coefficient = −0.56). Interobserver variability was calculated separately for expert versus trainee and trainee versus trainee (T-vs-T) measurements. For expert versus expert, ICC, rTEM, and reliability (%) were 0.984, 2.4%, and 96.8% (Group 1), 0.992, 2.07%, and 98.3% (Group 2), and 0.997, 1.38%, and 99.05% (Group 3), respectively. A similar pattern of variability was observed for T-vs-T measurements. The reliability (%) of the SPL by experts is consistently more than that of trainees across all age groups; however, the difference ameliorates with the age of participant.Conclusions:The study has validated the SPLINT technique by demonstrating a high level of intra- and interobserver reliability. The adequacy of the training modules for SPL measurements described in this study has also been established. Evidence that the SPL can be used as an objective marker of penile dimensions is herewith furnished. INTRODUCTION Measurement of the penis is important for multiple reasons varying from medical to social, scientific, and psychological. In children, it may be a crucial indicator of puberty or genital development. When kids present with aberrations of penile development such as hypospadias, micropenis or buried phallus, assessment of penile dimensions becomes necessary for baseline evaluation, monitoring response to treatment, and in follow-up.[1] Stretched penile length (SPL) is considered a more reliable indicator of penile size than flaccid penile length.[23] SPL is fairly constant and not vulnerable to fluctuations in penile size. SPL is a better approximation of erect penile length, the “true” length of the penis. A myriad of techniques has been described in the literature for the estimation of penile size including SPL. The commonly used instruments include a straight, flexible, or rigid ruler or a Vernier calipers.[1234] Various studies have tried to formulate age-specific nomograms to serve as yardsticks for the assessment of SPL.[15] The precision and accuracy of such nomograms are crucial since the nomograms serve as yardsticks for comparison in clinical settings. A discrepancy between the true value and the measured value may be a result of errors that can either be systematic (such as observational bias or procedural errors) skewing the results in a predictable direction or random (chance errors).[6] Herein, lies the importance of focusing on the quality of measurement data being generated to prevent any iatrogenic harm to the patient(s) and to harvest the maximal benefit of the resources invested. With this background, the lead author planned this study to assess the magnitude of intra- and interobserver variations in the measurement of SPL with the SPLINT technique.[6] MATERIALS AND METHODS This is a prospective study upon a cohort of 449 boys aged 0–14 years (mean age: 5.43 ± 4.39 years) [Figure 1a] including 68 infants (mean age: 5.09 ± 3.27 months) Inset in [Figure 1a] to validate the SPLINT technique to measure the SPL in children [Table 1].[6] (a) Age distribution of participants in the study cohort with age distribution of Infants in the study cohort in inset (b) Flow diagram for the study (c) Training module for stretched penile length measurement by the SPLINT technique Table 1Overview of the composition and age-based stratification of the study cohort and the respective stretched penile lengthsClassObserverSample size (n)SPL, mean±SD (cm)Minimum–maximumTotal (n=1764)E-14495.13±1.193.3–11.1E-24335.13±1.193.3–11.1T-14495.13±1.203.3–10.9T-24335.12±1.183.4–11Age Group 1 (<4 years) (n=732)E-11874.42±0.603.3–6.6E-21794.41±0.583.4–6.5T-11874.42±0.613.3–6.5T-21794.42±0.603.3–6.5Age Group 2 (4–8 years) (n=492)E-11244.86±0.763.5–7.4E-21224.85±0.783.4–7.4T-11244.87±0.783.4–7.2T-21224.85±0.793.5–7.5Age Group 3 (>8 years) (n=540)E-11386.34±1.194.2–11.1E-21326.32±1.184.3–11T-11386.34±1.164.3–10.9T-21326.32±1.194.3–11SD: Standard deviation, SPL: Stretched penile length Overview of the composition and age-based stratification of the study cohort and the respective stretched penile lengths SD: Standard deviation, SPL: Stretched penile length Participants Boys presenting to our pediatric surgery clinic for nonrelated conditions were included in this study. However, patients with anatomical aberrations of the penis such as that seen in hypospadias with chordee and chordee without hypospadias and those who refused consent were excluded from the study. Observers The readings were taken by three observers (one expert: “E” and two trainees: “T”). A regularly practicing pediatric surgeon with more than 10 years’ experience in pediatric urology and comfortable with the measurement of SPL was considered an “expert” observer for validation. The “trainee” observers were final year residents of MCh in pediatric surgery (both 6-year MCh students) who were trained into SPL measurement with this technique [Figure 1b]. Training The module was designed in a face-to-face meeting of five experts and comprised a three-step process [Figure 1c]. Data collection The following preliminary details were recorded for all patients enrolled for this study: hospital registration number, age (in completed years for those more than 1 year of age and in completed months for those under 1 year of age), and weight (kg, accurate to nearest 0.2 kg). The data were entered into paper forms designed specifically for this purpose by a member of the team who was not involved with taking measurements (herein called the “recorder”). The person taking the measurements (herein called the “measurer”) would take one measurement and call out the value to the recorder. The recorder would then call the number back to the measurer to confirm the correct reading. Then, the recorder would record the number on the paper form. Subsequently, at the end of the working day, the data were transferred to Excel charts (Microsoft® Excel® for Office 365 MSO Version 1809, Build 10827.20138 Click-to-Run) using the same methodology. Self or parent-reported measurements were not considered acceptable Measurement of stretched penile length SPL was essentially measured by the SPLINT technique which has been described by the authors previously.[6] A flexible ruler calibrated in centimeters was used for this purpose. The uncalibrated portion of the ruler before the “0” mark was premeasured accurately with the help of Vernier calipers and added to all measured values to account for “zero error.” The subject undergoing measurement of SPL was laid supine on the couch. The foreskin was retracted to expose the tip of the glans. In the case of younger babies wherein it was not possible to retract the prepuce completely, the foreskin was retracted only partially but enough to expose the meatus ensuring that the tip of the glans should be the highest point of penis when the penis was held vertical. Cases wherein this maneuver was not possible were excluded from this study. The penis was held gently between the thumb and the index finger of the right hand. The Ruler 1 was held with the right hand and aligned vertically along the penis with its base against the mons pubis. The ruler was pushed against the mons pubis to account for the pubic pad of fat, whereas the penis was stretched vertically upward. The reading corresponding to the highest point of the glans was recorded, whereas the ruler was held vertically against the penis at a right angle. The observers were alert to the commonly encountered errors of measurement and the following cautions were universally exercised: (1) there should be no angulation between the Ruler 1 and the stretched penis [Figure 2a]. Both the Ruler 1 and stretched penis should be held vertical and parallel to each other, (2) the recorded SPL should correspond to the highest point of the glans [Figure 2b]. Ruler 2 was used for identifying the highest point of the glans objectively, and (3) the observer’s eyes should be at the level of the highest point of the glans to avoid parallax [Figure 2c]. Technical nuances during measurement of stretched penile length with the SPLINT technique. (a) The ruler held vertically should be parallel to the stretched penis and there should be no angulation between the two. (b) The recorded stretched penile length should correspond to the highest point of the glans. (c) The observer’s eyes should be at the level of the highest point of the glans to avoid any parallax On the day of enrollment, the SPL for each patient was measured by two observers (each blinded to the readings of the other observer). The first observer was an expert: “E-1” (corresponding author for all cases in this study cohort). The second observer was a trainee: “T-1” (constant for all cases in this study cohort). These two readings constituted the data for calculating the interobserver variations between expert and trainee observers. During the second visit of the same patient (necessarily, within 1 week), a repeat observation was made by the same expert: “E-2” (to assess the intraobserver variability). The second reading was taken by another trainee observer: “T-2” (constant for all cases in this study cohort) to assess the interobserver variability. Sample size estimation The sample size was calculated using the ANOVA model with a single factor, which is classified into three age groups of children, namely, Group I (<4 years), Group II (4–8 years), and Group III (>8 years), whose means are to be compared. Statistical R software (version 4.0.2) was used for the sample size calculation. A sample of 113 per group would allow the detection of a difference between groups of at least 1.0 unit using the Tukey–Kramer (Pairwise) multiple comparison tests with 90% power to detect a difference at a 5% (two sided) significance level. Assuming a loss to follow-up of 10% and the common standard deviation (SD) within a group as 1.50, an effective sample size of 126 per group (minimum) is required to be recruited. Statistical analysis Data analysis was carried out using IBM SPSS Statistics for Windows (Version 29.0.2.0). Data were presented as mean ± SD (minimum–maximum) or median (interquartile range [IQR]). The intraobserver and interobserver variability was assessed by intraclass correlation (ICC), technical error of measurement (TEM), relative TEM (rTEM), and reliability (%). P < 0.05 was considered statistically significant. Bland–Altman plots were generated for all combinations. RESULTS The SPL (minimum–maximum: 3.25 cm–11 cm; median, IQR: 4.85 cm, 2 cm) increases with the age of the participant (correlation coefficient = 0.72) [Figure 3a]. (a) Correlation of stretched penile length (SPL) with age of participant (b) Correlation between 2 readings of SPL taken by same observer (intraobserver Variability) (intraclass correlation: 0.998 [95% confidence interval: 0.997–0.998]) (c) Intraobserver variations in the measurement of SPL (d) Correlation between intra-observer variations and age of participants Intraobserver variations in measurement of penile length (E-1 vs. E-2) Two readings by the same observer (Expert: E1 and E2) have been recorded for 433 participants; there was a loss to follow-up of 16 (of 449) participants (3.56%) during the second outpatient department visit. Those lost to follow-up did not belong to any particular age or age group. The mean SPL for 433 patients was 5.13 ± 1.19 cm (minimum–maximum: 3.3–11.1 cm) during the first observation (E1) and 5.12 ± 1.18 (minimum–maximum: 3.4–11 cm) during the second observation (E2). The mean of the difference between the two readings is 0.08 cm (95% confidence interval [CI]: 0.073–0.087). The ICC between 2 readings across the study cohort is 0.998 (95% CI: 0.997–0.998) [Table 2 and Figure 3b]. The two observations were the same (intraobserver variability = 0 cm) in 162 (of 433; 37.41%) participants and differed by 0.1 cm in 48.03% (208 of 433) participants. Overall, the readings were concordant or varied by 0.1 cm (intraobserver variability ≤0.1 cm) in 85% of the participants (n = 370 of 433; 85.5%). Based on these observations, the intraobserver difference [Figure 3c] was considered ignorable. Table 2Intraclass correlation for intraobserver and interobserver variability in measurement of stretched penile length with the SPLINT techniqueIntraobserver differences (E1 vs. E2) (n=433)Mean difference between readings (95% CI): 0.08 cm (0.073–0.087)ICC (95% CI): 0.998 (0.997–0.998) Sample size ICC 95% CI Group I1790.9900.987–0.993Group II1220.9950.992–0.996Group III1320.9900.987–0.993Interobserver difference - expert versus trainee (E1 vs. T1, E2 vs. T2) (n=882)Mean difference between readings (95% CI): 0.11−0.09 cm (0–0.29)ICC (95% CI): 0.996 (0.996–0.997) Sample size ICC 95% CI Group I2400.9840.980–0.987Group II1770.9920.989–0.994Group III2010.9970.996–0.998Interobserver difference - trainee versus trainee (T1 vs. T2) (n=433)Mean difference between readings (95% CI): 0.11−0.09 cm (0–0.28)ICC (95% CI): 0.997 (0.996–0.997) Sample size ICC 95% CI Group I1790.9840.979–0.988Group II1220.9910.988–0.994Group III1320.9980.997–0.998*E1 and E2 refer to two readings taken by the same observer while T1 and T2 refer to two readings taken by two trainees. CI: Confidence interval, ICC: Intraclass correlation Intraclass correlation for intraobserver and interobserver variability in measurement of stretched penile length with the SPLINT technique *E1 and E2 refer to two readings taken by the same observer while T1 and T2 refer to two readings taken by two trainees. CI: Confidence interval, ICC: Intraclass correlation Age-based stratification of study cohort Group I (<4 years) (number of participants: 179 [Table 1])Mean SPL: 4.44 ± 0.59 cm (E1) and 4.41 ± 0.58 cm (E2)The mean difference between the two readings is 0.09 ± 0.07 cm (95% CI: 0.08–0.10) with ICC of 0.990 (95% CI: 0.987–0.993) [Table 2]The intra-observer variability ≤0.1 cm in 83% of the participants (n = 149 of 179; 83.24%)The rTEM was 1.82% and reliability (%) was 98.1%.Group 2 (4–8 years) (number of participants: 122 [Table 1])Mean SPL: 4.86 ± 0.76 cm (E1) and 4.85 ± 0.78 cm (E2)The mean of difference between the two readings is 0.08 ± 0.08 cm (95% CI: 0.07–0.10) with an ICC of 0.995 (95% CI: 0.992–0.996) [Table 2]The intra-observer variability ≤0.1 cm in 85% of the participants (n = 104 of 122; 85.24%)The rTEM was 1.65% and reliability (%) was 98.9%.Group 3 (>8 years) (number of participants: 132 [Table 1])Mean SPL: 6.33 ± 1.19 cm (E1) and 6.32 ± 1.18 cm (E2)The mean difference between the two readings is 0.07 ± 0.07 cm (95% CI: 0.05–0.09) with an ICC of 0.998 (95% CI: 0.998–0.999) [Table 2]The intraobserver variability ≤0.1 cm in 88% of the participants (n = 117 of 132; 88.64%). Zero intraobserver variability was observed in 62% of participantsThe rTEM was 1.09% and reliability was 99.7%. Group I (<4 years) (number of participants: 179 [Table 1])Mean SPL: 4.44 ± 0.59 cm (E1) and 4.41 ± 0.58 cm (E2)The mean difference between the two readings is 0.09 ± 0.07 cm (95% CI: 0.08–0.10) with ICC of 0.990 (95% CI: 0.987–0.993) [Table 2]The intra-observer variability ≤0.1 cm in 83% of the participants (n = 149 of 179; 83.24%)The rTEM was 1.82% and reliability (%) was 98.1%. Group I (<4 years) (number of participants: 179 [Table 1]) Mean SPL: 4.44 ± 0.59 cm (E1) and 4.41 ± 0.58 cm (E2) Mean SPL: 4.44 ± 0.59 cm (E1) and 4.41 ± 0.58 cm (E2) The mean difference between the two readings is 0.09 ± 0.07 cm (95% CI: 0.08–0.10) with ICC of 0.990 (95% CI: 0.987–0.993) [Table 2] The mean difference between the two readings is 0.09 ± 0.07 cm (95% CI: 0.08–0.10) with ICC of 0.990 (95% CI: 0.987–0.993) [Table 2] The intra-observer variability ≤0.1 cm in 83% of the participants (n = 149 of 179; 83.24%) The intra-observer variability ≤0.1 cm in 83% of the participants (n = 149 of 179; 83.24%) The rTEM was 1.82% and reliability (%) was 98.1%. The rTEM was 1.82% and reliability (%) was 98.1%. Group 2 (4–8 years) (number of participants: 122 [Table 1])Mean SPL: 4.86 ± 0.76 cm (E1) and 4.85 ± 0.78 cm (E2)The mean of difference between the two readings is 0.08 ± 0.08 cm (95% CI: 0.07–0.10) with an ICC of 0.995 (95% CI: 0.992–0.996) [Table 2]The intra-observer variability ≤0.1 cm in 85% of the participants (n = 104 of 122; 85.24%)The rTEM was 1.65% and reliability (%) was 98.9%. Group 2 (4–8 years) (number of participants: 122 [Table 1]) Mean SPL: 4.86 ± 0.76 cm (E1) and 4.85 ± 0.78 cm (E2) Mean SPL: 4.86 ± 0.76 cm (E1) and 4.85 ± 0.78 cm (E2) The mean of difference between the two readings is 0.08 ± 0.08 cm (95% CI: 0.07–0.10) with an ICC of 0.995 (95% CI: 0.992–0.996) [Table 2] The mean of difference between the two readings is 0.08 ± 0.08 cm (95% CI: 0.07–0.10) with an ICC of 0.995 (95% CI: 0.992–0.996) [Table 2] The intra-observer variability ≤0.1 cm in 85% of the participants (n = 104 of 122; 85.24%) The intra-observer variability ≤0.1 cm in 85% of the participants (n = 104 of 122; 85.24%) The rTEM was 1.65% and reliability (%) was 98.9%. The rTEM was 1.65% and reliability (%) was 98.9%. Group 3 (>8 years) (number of participants: 132 [Table 1])Mean SPL: 6.33 ± 1.19 cm (E1) and 6.32 ± 1.18 cm (E2)The mean difference between the two readings is 0.07 ± 0.07 cm (95% CI: 0.05–0.09) with an ICC of 0.998 (95% CI: 0.998–0.999) [Table 2]The intraobserver variability ≤0.1 cm in 88% of the participants (n = 117 of 132; 88.64%). Zero intraobserver variability was observed in 62% of participantsThe rTEM was 1.09% and reliability was 99.7%. Group 3 (>8 years) (number of participants: 132 [Table 1]) Mean SPL: 6.33 ± 1.19 cm (E1) and 6.32 ± 1.18 cm (E2) Mean SPL: 6.33 ± 1.19 cm (E1) and 6.32 ± 1.18 cm (E2) The mean difference between the two readings is 0.07 ± 0.07 cm (95% CI: 0.05–0.09) with an ICC of 0.998 (95% CI: 0.998–0.999) [Table 2] The mean difference between the two readings is 0.07 ± 0.07 cm (95% CI: 0.05–0.09) with an ICC of 0.998 (95% CI: 0.998–0.999) [Table 2] The intraobserver variability ≤0.1 cm in 88% of the participants (n = 117 of 132; 88.64%). Zero intraobserver variability was observed in 62% of participants The intraobserver variability ≤0.1 cm in 88% of the participants (n = 117 of 132; 88.64%). Zero intraobserver variability was observed in 62% of participants The rTEM was 1.09% and reliability was 99.7%. The rTEM was 1.09% and reliability was 99.7%. The intraobserver variability was observed to be inversely proportional to the age of the participants (correlation coefficient −0.56) in the study cohort (the intraobserver variability was ≤0.1 cm 83%, 85%, and 88% for Groups I, II, and III respectively) [Figure 3d]. The larger the size of the penis (or indirectly, older the participant), the lesser will be the intraobserver variability. Interobserver variations in measurement of penile length The following data sets were available for calculating the interobserver variations: Expert versus trainee E-1 versus T-1: There were 449 participants which have been substratified into age Groups I (n = 187), II (n = 124), and III (n = 138) [Table 1]E-1 versus T-2: There were 433 participants which have been substratified into age-Groups I (n = 179), II (n = 122), and III (n = 132) [Table 1]Mean SPL: 5.132 ± 1.19 cm (E1) and 5.127 ± 1.19 cm (T1 or T2)The mean difference between the two readings is negligible: 0.11 ± 0.09 cm (95% CI: 0–0.29) with an ICC of 0.996 (95% CI: 0.996–0.997) [Table 2]The intraobserver variability ≤0.1 cm in 71% (n = 627 of 882) of the participantsThe trainee observations were more than the expert in 33.3% (294 of 882) and less in 36.17% (319 of 882) participants in the study group; the difference was insignificant (P = 0.211)There were only 2 participants for whom the readings differed by 0.4 cm between the expert and the trainee. E-1 versus T-1: There were 449 participants which have been substratified into age Groups I (n = 187), II (n = 124), and III (n = 138) [Table 1] E-1 versus T-1: There were 449 participants which have been substratified into age Groups I (n = 187), II (n = 124), and III (n = 138) [Table 1] E-1 versus T-2: There were 433 participants which have been substratified into age-Groups I (n = 179), II (n = 122), and III (n = 132) [Table 1]Mean SPL: 5.132 ± 1.19 cm (E1) and 5.127 ± 1.19 cm (T1 or T2)The mean difference between the two readings is negligible: 0.11 ± 0.09 cm (95% CI: 0–0.29) with an ICC of 0.996 (95% CI: 0.996–0.997) [Table 2]The intraobserver variability ≤0.1 cm in 71% (n = 627 of 882) of the participantsThe trainee observations were more than the expert in 33.3% (294 of 882) and less in 36.17% (319 of 882) participants in the study group; the difference was insignificant (P = 0.211)There were only 2 participants for whom the readings differed by 0.4 cm between the expert and the trainee. E-1 versus T-2: There were 433 participants which have been substratified into age-Groups I (n = 179), II (n = 122), and III (n = 132) [Table 1] Mean SPL: 5.132 ± 1.19 cm (E1) and 5.127 ± 1.19 cm (T1 or T2) Mean SPL: 5.132 ± 1.19 cm (E1) and 5.127 ± 1.19 cm (T1 or T2) The mean difference between the two readings is negligible: 0.11 ± 0.09 cm (95% CI: 0–0.29) with an ICC of 0.996 (95% CI: 0.996–0.997) [Table 2] The mean difference between the two readings is negligible: 0.11 ± 0.09 cm (95% CI: 0–0.29) with an ICC of 0.996 (95% CI: 0.996–0.997) [Table 2] The intraobserver variability ≤0.1 cm in 71% (n = 627 of 882) of the participants The intraobserver variability ≤0.1 cm in 71% (n = 627 of 882) of the participants The trainee observations were more than the expert in 33.3% (294 of 882) and less in 36.17% (319 of 882) participants in the study group; the difference was insignificant (P = 0.211) The trainee observations were more than the expert in 33.3% (294 of 882) and less in 36.17% (319 of 882) participants in the study group; the difference was insignificant (P = 0.211) There were only 2 participants for whom the readings differed by 0.4 cm between the expert and the trainee. There were only 2 participants for whom the readings differed by 0.4 cm between the expert and the trainee. Age-stratified groups (expert vs. trainee) The ICC between the two observations is 0.984 (95% CI: 0.980–0.987) for Group 1, 0.992 (95% CI: 0.989–0.994) for Group 2 and 0.997 (95% CI: 0.996–0.998) for Group 3 [Table 2]The interobserver variability was ≤0.1 cm for 65.6% (n = 240), 72% (n = 177), and 77.8% (n = 210) participants in Groups I, II and III, respectivelyGroup I: The rTEM between observers was 2.4% and reliability was 96.8% [Table 3]Group II: The rTEM between observers was 2.07% and reliability was 98.3%Group III: The rTEM between observers was 1.38% and reliability was 99.05%. The ICC between the two observations is 0.984 (95% CI: 0.980–0.987) for Group 1, 0.992 (95% CI: 0.989–0.994) for Group 2 and 0.997 (95% CI: 0.996–0.998) for Group 3 [Table 2] The ICC between the two observations is 0.984 (95% CI: 0.980–0.987) for Group 1, 0.992 (95% CI: 0.989–0.994) for Group 2 and 0.997 (95% CI: 0.996–0.998) for Group 3 [Table 2] The interobserver variability was ≤0.1 cm for 65.6% (n = 240), 72% (n = 177), and 77.8% (n = 210) participants in Groups I, II and III, respectively The interobserver variability was ≤0.1 cm for 65.6% (n = 240), 72% (n = 177), and 77.8% (n = 210) participants in Groups I, II and III, respectively Group I: The rTEM between observers was 2.4% and reliability was 96.8% [Table 3] Group I: The rTEM between observers was 2.4% and reliability was 96.8% [Table 3] Group II: The rTEM between observers was 2.07% and reliability was 98.3% Group II: The rTEM between observers was 2.07% and reliability was 98.3% Group III: The rTEM between observers was 1.38% and reliability was 99.05%. Group III: The rTEM between observers was 1.38% and reliability was 99.05%. Table 3Technical error of measurement, relative technical error of measurement, and reliability (%R) across the various categories of observers and different study groupsSerial numberGroupGroup I (<4 years)Group II (4–8 years)Group III (>8 years) TEM (cm)r (TEM) (%)Reliability (%)TEM (cm)r (TEM) (%)Reliability (%)TEM (cm)r (TEM) (%)Reliability (%)IExpert versus expert (intraobserver)0.081.8298.100.081.6598.900.071.0999.70IIExpert versus trainee versus trainee (interobserver: Three sets of readings)0.12.4096.800.12.0798.300.091.3899.05IIITrainee versus trainee (interobserver)0.122.4096.800.12.1097.800.081.2798.90TEM: Technical error of measurement, rTEM: Relative TEM Technical error of measurement, relative technical error of measurement, and reliability (%R) across the various categories of observers and different study groups TEM: Technical error of measurement, rTEM: Relative TEM Trainee versus trainee (T-1 vs. T-2) There were 433 participants which have been substratified into age Groups I (n = 179), II (n = 122), and III (n = 132) [Table 1]. Mean SPL: 5.132 ± 1.18 cm (T1) and 5.119 ± 1.19 cm (T2)The mean difference between the two readings is 0.11 ± 0.09 cm (95% CI: 0–0.28) with an ICC of 0.997 (95% CI: 0.996–0.997) [Table 2]The intraobserver variability ≤0.1 cm in 70% of the participants (n = 303 of 433). Mean SPL: 5.132 ± 1.18 cm (T1) and 5.119 ± 1.19 cm (T2) Mean SPL: 5.132 ± 1.18 cm (T1) and 5.119 ± 1.19 cm (T2) The mean difference between the two readings is 0.11 ± 0.09 cm (95% CI: 0–0.28) with an ICC of 0.997 (95% CI: 0.996–0.997) [Table 2] The mean difference between the two readings is 0.11 ± 0.09 cm (95% CI: 0–0.28) with an ICC of 0.997 (95% CI: 0.996–0.997) [Table 2] The intraobserver variability ≤0.1 cm in 70% of the participants (n = 303 of 433). The intraobserver variability ≤0.1 cm in 70% of the participants (n = 303 of 433). Age-stratified groups (trainee vs. trainee) The ICC between the observation by two trainees was 0.984 (95% CI: 0.979–0.988) for Group 1, 0.991 (95% CI: 0.988–0.994) for Group 2 and 0.998 (95% CI: 0.997–0.998) for Group 3 [Table 2]The interobserver variability was ≤0.1 cm for 64.8% (n = 116 of 179), 63.9% (n = 78 of 122) and 82.6% (n = 109 of 132) participants in Groups I, II and III, respectivelyGroup I: The rTEM between the two trainees was 2.4% and reliability was 96.8% [Table 3]Group II: The rTEM between the two trainees was 2.1% and reliability was 98.3%Group III: The rTEM between the trainees was 1.27% and reliability was 99.5%. The ICC between the observation by two trainees was 0.984 (95% CI: 0.979–0.988) for Group 1, 0.991 (95% CI: 0.988–0.994) for Group 2 and 0.998 (95% CI: 0.997–0.998) for Group 3 [Table 2] The ICC between the observation by two trainees was 0.984 (95% CI: 0.979–0.988) for Group 1, 0.991 (95% CI: 0.988–0.994) for Group 2 and 0.998 (95% CI: 0.997–0.998) for Group 3 [Table 2] The interobserver variability was ≤0.1 cm for 64.8% (n = 116 of 179), 63.9% (n = 78 of 122) and 82.6% (n = 109 of 132) participants in Groups I, II and III, respectivelyGroup I: The rTEM between the two trainees was 2.4% and reliability was 96.8% [Table 3]Group II: The rTEM between the two trainees was 2.1% and reliability was 98.3%Group III: The rTEM between the trainees was 1.27% and reliability was 99.5%. The interobserver variability was ≤0.1 cm for 64.8% (n = 116 of 179), 63.9% (n = 78 of 122) and 82.6% (n = 109 of 132) participants in Groups I, II and III, respectively Group I: The rTEM between the two trainees was 2.4% and reliability was 96.8% [Table 3] Group I: The rTEM between the two trainees was 2.4% and reliability was 96.8% [Table 3] Group II: The rTEM between the two trainees was 2.1% and reliability was 98.3% Group II: The rTEM between the two trainees was 2.1% and reliability was 98.3% Group III: The rTEM between the trainees was 1.27% and reliability was 99.5%. Group III: The rTEM between the trainees was 1.27% and reliability was 99.5%. The differences of mean (95% limits of agreement) between the two observers have been depicted as the Bland–Altman plot in Figure 4. The Bland-Altman plot depicting the bandwidth in the difference of penile length measurement. The red and green colors are representative of the lower and upper limits of the differences in measurement respectively. The readings falling above the green line or below the red line are outside the 95% limits of the agreement The reliability of the SPL readings taken by the experts is consistently more than that of trainees across all age groups; however, the difference in reliability ameliorates as the age or penile size increases [Figure 5a]. Similarly, the rTEM has shown a decline from Group I to Group II [Figure 5b]. (a) Bar graphs depicting the reliability (%) of observations pertaining to the stretched penile length measurement with the SPLINT technique between experts and trainees (b) Visual depiction of relative technical error of measurements of stretched penile length with the SPLINT technique between experts and trainees (Blue line: expert vs. expert, Red line: expert vs. trainee, Black line: trainee vs. trainee) DISCUSSION Measurement of penile dimensions is called for in multiple situations such as in patients with congenital aberrations affecting the male genitalia (hypospadias, exstrophy-epispadias complex, micro-penis, buried penis, etc.), early detection of hormonal disorders (hypogonadism or adrenal disorders), monitoring of growth and development, pubertal assessment (precocious or delayed puberty), psychological or social well-being (concerns related to penile length may affect the self-esteem of the patient and be a cause of concern for the parents), and as an essential component to the body of epidemiological or research data (population health, development norms, and cross-ethnic comparisons). Assessment of intra-and interobserver variability is indispensable to ensure the reliability and validity of any research that attempts to evaluate a new technique of measurement or for quality control of measurement data. Furthermore, it ensures consistency and accuracy in measurements and can be used as a surrogate marker for adequacy of training modules, instructional materials, or guidelines. The measurement of the SPL is a result of two subprocesses performed sequentially. The first step encompasses positioning the patient supine and the penis vertical. The prepuce should be retracted, at least enough so that the tip of the glans is the highest point. The penis is gently held, preferably at the corona between the thumb and the index finger of the right hand, and stretched vertically upward with a uniform and appropriate force. The subsequent step is “measurement of SPL” which involves aligning the ruler (with preassessed “zero- error”) vertically alongside the penile shaft at 12 o’ clock position. The ruler is pushed into the pubic pad of fat on the mons pubis. The reading corresponding to the highest point of the glans is registered as the respective SPL after adjusting for the “zero error.”[6] Both systematic and random (or chance) errors pertaining to either of these two steps may affect the quality of measurement data.[7] Measurement of SPL has to be precise and accurate. The precision of any measurement depends upon repeatability, reproducibility, and reliability[8] and is quantified by observer variability. Repeatability or intraobserver variability of SPL measurement refers to the ability of the same observer to produce the same SPL value when the measurement is repeated on the same patient under identical conditions. High intraobserver variability may be indicative of lack of standardization of the technique, suboptimal skill (or training) in measuring the SPL, parallax error, fatigue from multiple measurements, a significant time gap between the two measurements particularly in the age of “puberty” when the penis is experiencing growth spurt, erroneous data-entry, casual attitude during data acquisition or use of another ruler with a different “zero-error” at the time of repeat measurement. The technique as described previously has been used for SPL measurement with finer details as described above. The risk of parallax error was addressed at the time of training.[9] The maximal acceptable time gap between 2 readings on the same patient was set at 1 week. Data entry mechanisms were meticulous, elaborate, and standardized with a separate person recording the values. The “recorder” would call the measurement back to the “measurer” while he noted down the reading for confirmation (verification by “re-reading” method). The “zero-error” of the ruler has been accounted for. In this study, measurable intraobserver variability has been observed across 62.8% (271 of 433) participants despite all precautions. However, the magnitude of the variability (0.8 mm [95% CI: 0–0.22 mm]) with the technique was ignorable. The two readings were concordant or differed by ≤1 mm only in a significant majority (85%) of the participants. The maximal discordance between the 2 readings was 3 mm and observed in a handful (11 of 433) of participants only. This variability continued to be ignorable across the three age groups upon subgroup analysis too. TEM as a metric has been used in anthropometry to reflect the robustness of the measurement process by assessing the precision of measurements recorded.[10] A low TEM reflects positively upon the data quality; the measures are consistent and there is little variability due to measurement error. To facilitate the interpretation of TEM, it is converted to expressing TEM as a percentage of the mean measurement, rTEM. For intraobserver variability in anthropometry, a rTEM of more than 2% is indicative of the need for review and improvisation in measurement techniques.[11] The rTEM was consistently <2% across the three age groups when SPL is measured by the SPLINT technique. The percentage reliability was excellent (>98%) for all the three subgroups. Interestingly, it was observed that the intraobserver reliability as reflected by ICC, intraobserver variability (0.1 cm as cutoff), rTEM (≤2% as cutoff), and percentage reliability (98% as cutoff) has improved progressively from Group I to Group III (98.1%, 98.9%, and 99.7% for Groups I, II, and III, respectively). The intraobserver variation was inversely proportional to the penile size with a correlation coefficient of −0.56. The proportion of participants with intraobserver variability ≤0.1 cm increased from 83% in Group I to 88% in Group III (older), respectively. For participants in Group III, no intraobserver variability was noted for 62% of participants. This may be related to the privilege of eliciting cooperation from older children and the ease of handling the penis as well as the prepuce. Besides, older children have better anatomical landmarks, are less apprehensive and less sensitive to touch.[12] The minimal or acceptable intraobserver variability in the measurement of SPL may be considered a surrogate marker of one’s skill and familiarity with the SPLINT procedure and consequent eligibility to conduct the training of the other (trainee) observers.[13] However, this conclusion has to be interpreted in light of the known limitations of the concept of intraobserver variability. While low intraobserver variability reflects high consistency, it is vulnerable to systematic errors and does not necessarily reflect high precision.[14] Precision as a matrix reflects how closely the observed values are in agreement with the actual value of measurement. For instance, the observer may ignore the zero error on the measuring device; the repeated values with the same or similar ruler will be closely aligned to each other but systematically distant (quantitatively equal to the magnitude of the zero error) from the actual measurement. Contrarily, a low intraobserver variability largely eliminates the possibility of random or “chance” errors (emanating from variability in the measurement process) and validates the measurement technique.[15] Reproducibility of SPL measurement refers to the ability of different observers to produce the same SPL values when the measurement is repeated on the same patient under identical conditions (interobserver reliability). The interobserver variability is expected to be minimum when the first step (positioning the patient supine and the penis vertical) is predefined and the SPL is re-measured by the second observer. However, the ultimate test of inter-observer variability is when both the sub-processes are repeated by the second observer. This is sometimes referred to as “test–retest variability.”[16] In this study, the authors had protocolized that the whole process of SPL measurement will be repeated by both observers. In addition to the errors accounting for intra-observer variability, differences in expertise or training, personal biases or perceptual differences, differences in communication or documentation, and differences in attention or focus between the two observers can also result in interobserver variability. The key factors instrumental in reducing this unintended inter-observer variability (thereby improving the reproducibility of readings) include standardization of the process of measurement, defining each step of measurement objectively, appropriate training of observers, and use of relatively simple equipment (straight ruler in our case). Nonetheless, the importance of using “appropriate statistical methods” can never be underscored. While one reading was taken by an “expert” and the other by a “trainee,” the mean difference between the two readings was negligible (0.11 ± 0.09 cm) with an excellent ICC (0.996) (ICC is stratified as poor [>0.5], moderate [0.5≤ ICC <0.75], good [0.75≤ ICC <0.9], and excellent ≥0.9]).[17] The interobserver variability was ≤0.1 cm in 71% of the subjects. The maximal inter-observer variability was recorded to be 0.4 cm (in 2 participants only). The rTEM values varied from acceptable (≤1.5; Group III) to reasonable (1.5< rTEM <2.5; Groups I and II), whereas the reliability varied from 96.8% to 99.05%.[18] The observations were similar when the readings by the two trainees were compared. The high degree of inter-rater reliability between the expert and the trainee as well as between the two trainees, accounts for most of the limitations despite a high intraobserver reliability and validates the SPLINT technique for measurement of SPL. The findings of this study indicate that the SPL data generated with the SPLINT technique will be scientific and may be used for clinical, academic, and research purposes. The SPLINT technique may be used for the generation of anthropometric data which may be generalized to the population at large. Furthermore, the high levels of intraobserver and interobserver reliability support the integrity of the methodology described herein for imparting training in SPL measurement by the SPLINT technique. The methodology may be adopted for regular training of medical students and residents of pediatric urology or related specialties in SPL measurement. Last but not the least, the penis is very versatile organ, at least with respect to its size and the spectrum of dimensions from a completely flaccid state to a full-blown erection is wide. Measuring penile size for assessment, comparison, or follow-up in clinical settings needs objectivity and is sometimes challenging. Similar or comparable SPL measurements on the same patients performed by the same or different observers on the same or different occasions have been documented in this study; the use of SPL as an objective metric for establishing the penile size in clinical use is herewith reiterated. CONCLUSIONS The study has validated the methodology inbuilt within the SPLINT technique by demonstrating a high level of intra- and interobserver reliability. The adequacy of the training modules adopted in this study for the final year MCh trainees is also established. The objectivity and stability of the SPL as an objective metric to assess the penile dimensions is herewith established. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.