ECAD Study: Evaluating agreement degree among paediatricians in hip dysplasia assessment by Graf ultrasound method

doi:10.21203/rs.3.rs-4670003/v1

Download PDF

Research Article

ECAD Study: Evaluating agreement degree among paediatricians in hip dysplasia assessment by Graf ultrasound method

https://doi.org/10.21203/rs.3.rs-4670003/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 24 Aug, 2024

Read the published version in European Journal of Pediatrics →

You are reading this latest preprint version

Introduction:

developmental dysplasia of the hip (DDH) is the most common musculoskeletal anomaly of childhood. This study analyses paediatrician’s concordance performing hip ultrasound according to Graf's method.

Materials and methods

an analytical cross-sectional study is designed. Patients’ hips are evaluated by Graf´s ultrasound method at 4–6 and 12–16 weeks of age. Demographic characteristics, pregnancy history, factors related to DDH, hip type, and α and β angles are collected. To assess agreement degree of hip diagnosis and hip types Kappa index (I. kappa) is used. Intraclass correlation coefficient (ICC) evaluates concordance of α and β angles. A p value < 0.05 is considered to be statistically significant. All results are represented with their 95% confidence interval (95% CI).

Results

467 neonates are recruited and meet inclusion criteria. 3182 images were analysed. Agreement between paediatricians according to hip diagnosis and different types ranges from moderate to almost perfect I. kappa 0.6–1.0 (95%CI 0.5–0.7; 95%CI 1.0–1.0). α angle ICC for paediatricians is between 0.75–0.88 (95%CI 0.61–0.86; 95%CI 0.80–0.92). β angle value agreement degree among paediatricians at both visits is moderate.

Discussion

Graf method is the most popular ultrasound technique for hip assessment. In our setting, paediatricians carry out children follow-up, therefore, they should perform universal screening. In this study agreement degree between paediatricians varies between substantial and almost perfect. As conclusion, Graf ultrasound procedure performed by paediatricians is a reliable test and can be used for DDH screening.

hip dysplasia

paediatrics

ultrasound

screening

Developmental dysplasia of the hip (DDH) includes the spectrum of abnormal anatomical relationships between the femoral head and the acetabulum (dysplasia, subluxation or dislocation) (1). It is the most common congenital musculoskeletal anomaly of childhood. Early diagnosis and treatment improves DDH prognosis. Detection failure of minor forms can lead to osteoarthritis development (2).

Ultrasound is currently the diagnostic tool of choice for hip evaluation in children (3, 4). Graf introduced the first method of neonatal hip ultrasound assessment in 1978. This method proposes hip analysis in a transverse, static, and standardized plane, being the most widely used worldwide (Fig. 1). Graf defines four types of hips according to their bone and cartilage coverage, represented by the value of α and β angles (Table 1). α and β angles values, obtained from the intersection of the baseline, bony acetabular roof line, and cartilaginous roof determines hip type (Fig. 2).

Table 1

Types of hips according to *Graf's* classification
HIP TYPE	Bone cover Angle α	Cartilaginous cover Angle β	Acetabular promontory
Type I MATURE	Good α ≥ 60°	Wide Ia β < 55° Ib β > 55°	Pointy/blunt
Type IIa Age < 12 weeks PHYSIOLOGICALLY IMMATURE	Enough α = 50–59°	Wide	Rounded
Type IIb Age > 12 weeks OSSIFICATION DELAY	Deficient α = 50–59°	Wide	Rounded
Type IIc CRITICAL • STABLE • UNSTABLE	Severely deficient α = 43–49°	• In the face of stressful manoeuvres β < 77° • In the face of stressful manoeuvres β ≥ 77°	Rounded/flattened
Type D DESCENTRING	Severely deficient α = 43–49°	Compressed β ≥ 77°	Rounded/flattened
Type III DISLOCATED	Poor α < 43°	Compressed Up	Flattened
Type IV DISLOCATED	Poor α < 43°	Compressed down	Flattened

DDH universal screening is based on performing ultrasound on all newborns. It has been part of the Austrian and German National Surveillance Programme since 1991 and 1996, respectively (5–7). And, it is also being implemented in some settings of other countries such Ireland, Greece, Italy, Hungary, China, and the United Arab Emirates (8). It allows early diagnosis, reduction of surgical interventions, decrease of long-term complications and, it is cost-effective (9).

Graf's standardized ultrasound method is the most widely used in universal screening programs (10). This method is easy to perform and has little variability between observers. However, all studies reviewed compare results between technicians, radiologists, or traumatologists. In our setting, dysplasia screening is prescribed by paediatricians, and ultrasound could be performed bedside as part of the child follow-up. This study analyses paediatrician’s concordance performing hip ultrasound according to Graf's method, and also the possibility of its use as a screening tool for DDH in our setting.

An analytical cross-sectional study is designed. All newborns in the province of Guadalajara (Spain) born from March 29 to May 15, and from August 15 to December 31 of 2019 are included. Patients’ hips are evaluated by ultrasound according to Graf´s method regardless their physical examination or risk factor for hip dysplasia.

The study is approved by the hospital's Ethics and Clinical Research Committee. Data is collected on an anonymized database. Analysis is carried out in compliance with ethical principles for research on human subjects proposed in Helsinki Declaration by the World Medical Association (WMA).

Patients which the Universal Hospital of Guadalajara is their reference centre, and born in the recruitment period are included. Their legal representatives have to signature the informed consent and the study protocol has to be complained to be part of the study. Exclusion criteria are absence of acceptance and signature of informed consent, revocation of consent, study protocol failure or major musculoskeletal malformations (bone dysplasia, phocomelia, congenital defect of a limb, etc.).

Study protocol involves performing two hip ultrasounds at 4–6 and 12–16 weeks of age. Patients assessed at the first visit with less than 12 weeks and who do not attend the second ultrasound appointment are considered as missing follow-up and are excluded from the study. If the age is 12 weeks or more at the first visit, a second examination may not be performed, since at 12 weeks, hip has undergone maturation changes necessary for an accurate diagnosis. Ultrasound images that do not meet Graf standardized image criteria are also excluded from the analysis (Fig. 3).

During the first month of the study, examination of both hips is performed by three different paediatricians, and from then on, the technique is performed by two paediatricians. In addition, an external expert on Graf`s method reviews all images obtained during the first two months of the study. Later, it only reviews those in which there is no agreement among paediatricians. Each paediatrician obtains two images of each hip, then selects the one that strictly complies with Graf "standard", on which realizes measurements of the α and β angles. All images are converted to digital format, and a report is prepared on each of the visits.

All paediatricians who perform the ultrasounds take a theoretical and practical course with a total duration of 40 hours.

Ultrasound scans are performed using the "Logiq F8 6002958WX0" ultrasound system and the "L6-12-RS Linear Array Probe 727414WX5" probe, both from GE Healthcare®. To keep patients in the proper position, Graf's "Rapid Echo Ultrasound Positioning Device: SonoQuick" cradle and Gio®'s "Eco Support Arm Guide: SonoGuide" ultrasound probe arm are used.

Data collected includes demographic characteristics (sex, race, gestational age, birth weight, and skeletal malformation), pregnancy history (type of delivery, gestation unique or multiple, primiparity, maternal age, number of previous pregnancies), factors related to development and diagnosis of hip dysplasia (presentation type, family history of DDH, and age at first and second ultrasounds), hip type, and values of α and β angles.

Sample size is calculated using Excel Real Statistics add-in. Degree of diagnostic agreement according to Graf method is estimated to range between 65% (p1) and 91% (q1). To obtain an almost perfect agreement degree (0.81–0.9), a sample size of 247 observations is required, considering a study power of 80% and a confidence level of 95%.

Measures of central tendency (mean and median) and dispersion measures (standard deviation and interquartile range) are used for quantitative variables. Frequencies and percentages are used to describe qualitative variables. For univariate analysis, Chi² or Fisher test for categorical variables, and T-student for continuous variables are applied. To assess agreement degree of two categories (healthy or pathological hip), Kappa index (I. kappa) is used. Reproducibility according to pathological degree (I, IIa, IIb, IIC, C, III, and IV) is performed by weighted Kappa index. Intraclass correlation coefficient (ICC) is calculated to evaluate concordance of α and β angles, considering a mixed-effects model. In each case, standard error, confidence interval and index value categorization are established as a measure of reproducibility strength according to Landis and Koch. The following levels of agreement are considered: almost perfect when the Kappa coefficient oscillates between 0.81–1, substantial between 0.61–0.8, moderate between 0.41–0.6, reasonable 0.21–0.4, minimum 0–0.2 and null for values < 0. A value of p < 0.05 is considered to be statistically significant, all results are represented along with their 95% confidence interval (95% CI). StataIC 16 and SAS System version 9.4 are used for statistical analysis.

573 newborns are recruited. 467 patients met the inclusion criteria. At the end of the study, 3752 images were obtained, of which 3182 were analysed (Fig. 3). No patients with major musculoskeletal malformations were born during de study period. The clinical characteristics of patients are shown in Table 2. 5 patients were diagnosed of developmental dysplasia of the hip (DDH) during the study period, with no statistically significant differences between them and the healthy patients (Table 2).

Table 2

Patients demographic and clinical characteristics. Comparison between patients characteristics diagnosed with DDH versus healthy patients in the sample studied.
Patients’ demographic and clinical characteristics
	N (number of cases)		%
Sex Boy Girl	244 223		52.2 47.8
Race Caucasian Latin American North African Sub saharan Asian	362 37 51 8 7		77.5 7.9 10.9 1.7 1.5
Full-term newborn (37–42 weeks gestational age))	446		95.5
Preterm newborn (< 37 weeks gestational age)	21		4.5
Birth Hospital University Hospital of Guadalajara Outside	447 21		95.7 4.5
Birth weight Low weight (< p10) Suitable weight (p10-p90) High weight (> p90)	55 369 43		11.8 79 9.2
Skeletal malformation	0		0
Type of delivery Vaginal Cesarean	375 92		80.3 10.7
Gestation Unique Multiple	453 14		97 3
Primiparity	156		33.5
Presentation Cephalic Breech	457 10		97.9 2.1
Family history of dysplasia First or second degree Third Grade	27 16/27 11/27		5.8 59.3 40.7
Developmental Dysplasia of the Hip	5		1
	Median		IR
Birth weight (kg)	3.2		2.9–3.5
Maternal age	33		30–37
Number of previous pregnancies	2		1–3
Age (weeks) first ultrasound	5		4.6–6
Age (weeks) second ultrasound	15.4		14–16.3
Characteristics of patients diagnosed with DDH versus healthy patients in the sample studied
	DDH	Healthy
	N (%)	N (%)	p
Female	3 (60)	220 (47.6)	0.5
Caucasian race	3 (60)	359 (82)	0.3
Gestational age > 37 weeks	5 (100)	441 (95.5)	0.6
Vaginal delivery	3 (60)	372 (80.5)	0.3
Multiple gestation	0 (0)	12 (2.6)	0.9
Family history	1 (20)	26 (5.6)	0.4
Antenatal position: Breech Cephalic	0 (0) 5 (100)	40 (19.5) 372 (97.8)	0.7
	Median (IR)	Median (IR)	p
Weight at birth	3.2 (2.9–3.5)	3.5 (3.1–3.9)	0.9
Maternal age	33 (30–37)	32 (26.6–37)	1

Agreement between paediatricians according to healthy or pathological hip diagnosis at the first visit varies between substantial and almost perfect with a kappa index 0.6–1.0 (95% CI: 0.4–0.9; 95% CI: 1.0–1.0) (Table 3). Agreement degree between the external expert and paediatricians at the first visit is more variable. In this case, minimum agreement degree among observers is moderate (I. kappa 0.5; 95% CI: 0.2–0.7) and maximum is almost perfect (I. kappa 1.0; 95% CI 1.0–1.0) (Table 3). After 2 months of study, it was observed that agreement between paediatricians continued to be substantial: I. kappa 0.7 (95% CI: 0.5–0.9); distancing themselves from the expert's opinion, although within limits considered reasonable: I. kappa 0.3–0.4 (95% CI: 0.1–0.4; 95% CI: 0.1–0.7) (Table 3).

Table 3

Concordance among diagnosis of healthy or pathological hip between paediatricians and the expert at visit 1 and 2
Table 3 - Visit 1
Right Hip	Healthy hip	Dysplasic hip	Total	Kappa	Standard Error	CI95%	Reproducibility Strength (Landis and Koch)
Rows = Paediatrician 1 and Columns = Paediatrician 2
Healthy hip	63	0	63
Dysplasic hip	0	1	1	1.0	0.0	1.0–1.0	Almost perfect
Total	63	1	64
Rows = Paediatrician 1 and Columns = Paediatrician 3
Healthy hip Dysplasic hip Total	458 1 459	3 5 8	461 6 467	0.8	0.1	0.5–1.0	Substantial
Rows = Paediatrician 2 and Columns = Paediatrician 3
Healthy hip Dysplasic hip Total	63 0 63	0 1 1	63 1 64	1.0	0.0	1.0–1.0	Almost perfect
Rows = Paediatrician 1 and Columns = External expert
Healthy hip Dysplasic hip Total	129 1 130	0 3 3	129 4 133	0.9	0.1	0.6–1.0	Almost perfect
Rows = Paediatrician 2 and Columns = External expert
Healthy hip Dysplasic hip Total	63 0 63	0 1 1	63 1 64	1.0	0.0	1.0–1.0	Almost perfect
Rows = Paediatrician 3 and Columns = External expert
Healthy hip Dysplastic hip Total	129 1 130	1 2 3	130 3 133	0.6	0.3	0.3–1.0	Moderate
Left hip	Healthy hip	Dysplasic hip	Total	Kappa	Standard Error	CI95%	Reproducibility Strength (Landis and Koch)
Rows = Paediatrician 1 and Columns = Paediatrician 2
Healthy hip Dysplasic hip Total	60 1 61	0 3 3	60 4 64	0.9	0.1	0.6–1.0	Almost perfect
Rows = Paediatrician 1 and Columns = Paediatrician 3
Healthy hip Dysplasic hip Total	455 3 458	4 5 9	459 7 467	0.6	0.1	0.4–0.9	Moderate
Rows = Paediatrician 2 and Columns = Paediatrician 3
Healthy hip Dysplasic hip Total	61 0 61	0 3 3	61 3 64	1.0	0.0	1.0–1.0	Almost perfect
Rows = Paediatrician 1 and Columns = External expert
Healthy hip Dysplasic hip Total	125 6 131	0 2 2	125 8 133	0.5	0.2	0.2–0.7	Moderate
Rows = Paediatrician 2 and Columns = External expert
Healthy hip Dysplasic hip Total	61 1 62	0 2 2	61 3 64	0.8	0.2	0.4–1.0	Substantial
Rows = Paediatrician 3 and Columns = External expert
Healthy hip Dysplasic hip Total	128 3 131	0 2 2	128 5 133	0.7	0.2	0.4–1.0	Substantial

Table 4shows agreement degree according to different types of hips at the first and second visits among paediatricians and the external expert. In this case, agreement degree ranges from moderate to substantial at the first visit, with a minimum observed of 0.6 (95% CI: 0.3–0.9) and a maximum of 0.8 (95% CI: 0.6–1.0). Agreement degree at the second visit among paediatricians was also moderate to substantial: I. kappa 0.7 (95% CI: 0.4–0.8). At the second visit comparison of results between external expert and paediatricians decreases to reasonable, I. kappa 0.3 (95% CI: 0.1–0.6).


Table 3 - Visit 2
Right Hip	Healthy hip	Dysplasic hip	Total	Kappa	Standard Error	CI95%	Reproducibility Strength (Landis and Koch)
Rows = Paediatrician 1 and Columns = Paediatrician 2
Healthy hip	424	5	429
Dysplasic hip	2	8	10	0.7	0.1	0.5–0.9	Substantial
Total	426	13	439
Rows = Paediatrician 1 and Columns = External expert
Healthy hip Dysplasic hip Total	25 5 30	1 1 2	26 6 32	0.3	0.1	0.1–0.6	Reasonable
Rows = Paediatrician 2 and Columns = External expert
Healthy hip Dysplasic hip Total	22 8 30	1 1 2	23 9 32	0.3	0.1	0.1–0.4	Reasonable
Left hip	Healthy hip	Dysplasic hip	Total	Kappa	Standard Error	CI95%	Reproducibility Strength (Landis and Koch)
Rows = Paediatrician 1 and Columns = Paediatrician 2
Healthy hip Dysplasic hip Total	420 4 424	7 8 15	427 12 439	0.7	0.1	0.5–0.9	Substantial
Rows = Paediatrician 1 and Columns = External expert
Healthy hip Dysplasic hip Total	24 5 29	1 2 3	25 7 32	0.3	0.2	0.1–0.7	Reasonable
Rows = Paediatrician 2 and Columns = External expert
Healthy hip Dysplasic hip Total	22 7 29	0 3 3	22 10 32	0.4	0.2	0.1–0.7	Reasonable

Table 4: Concordance among hip types between paediatricians and the expert at visit 1 and 2

Table 4 - Visit 1
Right Hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR³	CI95%	Reproducibility Strength
R¹ Paediatrician 1 and C² Paediatrician 2
I IIA IIB IIC D III IV Total	42 8 0 0 0 0 0 50	4 9 0 0 0 0 0 13	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 1 0 1	0 0 0 0 0 0 0 0	46 17 0 0 1 0 0 64	0.7	0.1	0.4–0.9	Substantial
R¹ Paediatrician 1 and C² Paediatrician 3
I IIA IIB IIC D III IV Total	315 41 0 0 0 0 0 356	38 66 0 1 0 0 0 105	0 0 1 0 0 0 0 1	0 0 0 1 0 0 0 1	0 2 0 0 1 0 0 3	0 0 0 0 0 1 0 1	0 0 0 0 0 0 0 0	353 109 1 2 1 1 0 467	0.7	0.1	0.5–0.7	Substantial
R¹ Paediatrician 2 and C² Paediatrician 3
I IIA IIB IIC D III IV Total	41 4 0 0 0 0 0 45	9 9 0 0 0 0 0 18	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 1 0 0	0 0 0 0 0 0 0 0	50 13 0 0 0 1 0 64	0.6	0.1	0.3–0.9	Moderate
Left hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR³	CI95%	Reproducibility Strength
R¹ Paediatrician 1 and C² Paediatrician 2
I IIA IIB IIC D III IV Total	28 11 0 0 0 0 0 39	4 17 0 0 1 0 0 22	0 0 0 0 0 0 0 0	0 0 0 1 0 0 0 1	0 0 0 0 1 0 0 1	0 0 0 0 0 1 0 1	0 0 0 0 0 0 0 0	32 28 0 1 2 1 0 64	0.7	0.1	0.5–0.9	Substantial
R¹ Paediatrician 1 and C² Paediatrician 3
I IIA IIB IIC D III IV Total	285 50 0 0 0 0 0 335	33 89 0 2 2 0 0 126	0 0 0 0 0 0 0 0	0 1 0 1 2 0 0 4	0 0 0 1 0 0 0 1	0 0 0 0 0 1 0 1	0 0 0 0 0 0 0 0	318 140 0 4 4 1 0 467	0.6	0.04	0.5–0.7	Moderate
R¹ Paediatrician 2 and C² Paediatrician 3
I IIA IIB IIC D III IV Total	33 5 0 0 0 0 0 38	6 17 0 0 0 0 0 23	0 0 0 0 0 0 0 0	0 0 0 1 0 0 0 1	0 0 0 1 0 0 0 1	0 0 0 0 0 1 0 1	0 0 0 0 0 0 0 0	39 22 0 1 1 1 0 64	0.7	0.1	0.6–0.9	Substantial

Table 4 - Visit 1
Right Hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR¹	CI95%	Reproducibility Strength
Rows = Paediatrician 1 and Columns = External expert
I IIA IIB IIC D III IV Total	96 16 0 0 0 0 0 112	1 16 0 1 0 0 0 18	0 0 0 0 0 0 0 0	0 0 0 1 0 0 0 1	0 0 0 0 0 0 0 0	0 0 0 0 0 1 0 1	0 0 0 0 0 0 0 0	97 32 0 2 0 1 0 132	0.7	0.1	0.6–0.9	Substancial
Rows = Paediatrician 2 and Columns = External expert
I IIA IIB IIC D III IV Total	49 5 0 0 0 0 0 54	1 8 0 0 0 0 0 9	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 1 1	50 13 0 0 0 1 0 64	0.8	0.1	0.6–1.0	Substancial
Rows = Paediatrician 3 and Columns = External expert
I IIA IIB IIC D III IV Total	96 18 0 0 0 0 0 113	2 14 0 1 1 0 0 18	0 0 0 0 0 0 0 0	0 1 0 0 0 0 0 1	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 1 0 1	98 33 0 1 1 1 0 132	0.6	0.1	0.4–0.8	Moderate
Left hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR¹	CI95%	Reproducibility Strength
Rows = Paediatrician 1 and Columns = External expert
I IIA IIB IIC D III IV Total	67 28 0 0 0 0 0 95	7 22 0 3 3 0 0 35	0 0 0 0 0 0 0 0	0 0 0 1 0 0 0 1	0 0 0 1 0 0 0 1	0 0 0 0 0 1 0 1	0 0 0 0 0 1 0 1	74 50 0 4 3 1 0 132	0.6	0.1	0.4–0.7	Moderate
Rows = Paediatrician 2 and Columns = External expert
I IIA IIB IIC D III IV Total	35 10 0 0 0 0 0 45	4 12 0 0 1 0 0 17	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 1 0 0 0 1	0 0 0 0 0 0 0 0	0 0 0 0 0 1 0 1	39 22 0 1 1 1 0 64	0.6	0.1	0.4–0.8	Moderate
Rows = Paediatrician 3 and Columns = External expert
I IIA IIB IIC D III IV Total	79 16 0 0 0 0 0 95	3 29 0 3 0 0 0 35	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 1 0 0 1	0 0 0 0 0 0 0 0	0 0 0 0 0 1 0 1	82 45 0 3 1 1 0 132	0.7	0.1	0.6–0.8	Substancial
1.Rows. 2. Columns. 3. Standard error

TABLE 4 - Visit 2
Right Hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR	CI95%	Reproducibility Strength
Rows = Paediatrician 1 and Columns = Paediatrician 2
I IIA IIB IIC D III IV Total	421 0 0 2 0 0 0 423	3 1 0 0 0 0 0 4	4 0 0 8 0 0 0 12	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	428 1 0 10 0 0 0 439	0.7	0.1	0.5 – 0.9	Substantcial
Left hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR	CI95%	Reproducibility Strength
Rows = Paediatrician 1 and Columns = Paediatrician 2
I IIA IIB IIC D III IV Total	414 2 3 0 1 0 0 420	1 1 0 0 0 0 0 2	6 0 9 0 1 0 0 16	0 0 0 0 0 0 0 0	0 0 1 0 0 0 0 1	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	421 3 13 0 2 0 0 439	0.7	0.1	0.4 – 0.8	Substantcial

Agreement degree of α and β angles values is represented in Table 5. It should be noted that intraclass correlation coefficient (ICC) of α angle at the first visit for paediatricians is between 0.75 (995% CI: 0.59–0.85) and 0.88 (95% CI: 0.80–0.92), ranging from substantial to almost perfect. On this first visit, agreement with external expert was also high, with ICC ranging from 0.76 (995% CI 0.60–0.85) to 0.82 (95% CI 0.75–0.88). At second visit, α angle value agreement degree among paediatricians continues to be substantial, ICC 0.74–0.76 (95% CI: 0.69–0.79; 95% CI: 0.72–0.81). However, during this visit, agreement degree with external expert decreased, ranging from minimum to reasonable: ICC 0.18 (95% CI: 0.05–0.60) and ICC 0.45 (95% CI: 0.20–0.75). Compared to the results described above, β angle value agreement degree among paediatricians of both visits is moderate, with an ICC between 0.41 and 0.58 (95% CI: 0.04–0.64; 95% CI: 0.50–0.65). Agreement between paediatricians and external expert decreases already at the first visit, with the ICC being lower in this assessment 0.23 (95% CI: 0.14–0.50) and in the second visit of 0.04–0.21 (95% CI: 0.01–0.43; 95% CI: 0.10–0.54).

TABLE 4 - Visit 2
Right Hip	I	IIA	IIB	IIC	D	III	IV	Total	Kappa	SR	CI95%	Reproducibility Strength
Rows = Paediatrician 1 and Columns = External expert
I IIA IIB IIC D III IV Total	25 5 0 0 0 0 0 30	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	1 1 0 0 0 0 0 2	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	26 6 0 2 0 1 0 32	0.3	0.1	0.2 – 0.6	Reasonable
Rows = Paediatrician 2 and Columns = External expert
I IIA IIB IIC D III IV Total	19 3 8 0 0 0 0 30	0 0 0 0 0 0 0 0	1 0 1 0 0 0 0 2	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	20 3 9 0 0 0 0 32	0.3	0.1	0.2 – 0.5	Reasonable
Left Hip		IIA	IIB	IIC	D	III	IV	Total	Kappa	SR	CI95%	Reproducibility Strength
Rows = Paediatrician 1 and Columns = External expert
I IIA IIB IIC D III IV Total	22 2 5 0 0 0 0 29	0 0 0 0 0 0 0 0	1 0 2 0 0 0 0 3	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	23 2 7 0 0 0 0 32	0.3	0.1	0.2 – 0.6	Reasonable
Rows = Paediatrician 2 and Columns = External expert
I IIA IIB IIC D III IV Total	20 2 7 0 0 0 0 29	0 0 0 0 0 0 0 0	0 0 2 0 1 0 0 3	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	0 0 0 0 0 0 0 0	20 2 9 0 1 0 0 32	0.3	0.1	0.1 – 0.6	Reasonable

1.Roaws. 2. Columns. 3. Standard error

Table 5: Agreement among α and β angle values between paediatricians and the external expert at both visits

VISIT 1
	ICC*	CI 95%	p	Reproducibility Strength
Right Hip Angle α
Paediatrician 1 - Paediatrician 2	0.77	0.61 – 0.86	< 0.01	Substantial
Paediatrician 1 - Paediatrician 3	0.82	0.78 – 0.85	< 0.01	Almost Perfect
Paediatrician 2 - Paediatrician 3	0.75	0.59 – 0.85	< 0.01	Substantial
Paediatrician 1 - Expert	0.81	0.71 – 0.88	< 0.01	Almost Perfect
Paediatrician 2 - Expert	0.76	0.60 – 0.85	< 0.01	Substantial
Paediatrician 3 - Expert	0.82	0.75 – 0.88	< 0.01	Almost Perfect
Left Hip Angle α
Paediatrician 1 - Paediatrician 2	0.80	0.68 – 0.88	< 0.01	Substantial
Paediatrician 1 - Paediatrician 3	0.80	0.76 – 0.83	< 0.01	Substantial
Paediatrician 2 - Paediatrician 3	0.88	0.80 – 0.92	< 0.01	Almost Perfect
Paediatrician 1 - Expert	0.78	0.68 – 0.85	< 0.01	Substantial
Paediatrician 2 - Expert	0.79	0.67 – 0.88	< 0.01	Substantial
Paediatrician 3 - Expert	0.82	0.75 – 0.88	< 0.01	Almost Perfect
Right Hip Angle β
Paediatrician 1 - Paediatrician 2	0.48	0.16 – 0.68	0.03	Moderate
Paediatrician 1 - Paediatrician 3	0.58	0.50 – 0.65	<0.01	Moderate
Paediatrician 2 - Paediatrician 3	0.49	0.18 – 0.69	0.03	Moderate
Paediatrician 1 - Expert	0.31	0.01 – 0.52	0.02	Reasonable
Paediatrician 2 - Expert	0.23	0.14 – 0.50	0.05	Reasonable
Paediatrician 3 - Expert	0.50	0.16 – 0.70	<0.01	Moderate
Left Hip Angle β
Paediatrician 1 - Paediatrician 2	0.41	0.04 – 0.64	0.02	Moderate
Paediatrician 1 - Paediatrician 3	0.50	0.40 – 0.58	< 0.01	Moderate
Paediatrician 2 - Paediatrician 3	0.55	0.28 – 0.73	< 0.01	Moderate
Paediatrician 1 - Expert	0.42	0.15 – 0.60	< 0.01	Moderate
Paediatrician 2 - Expert	0.48	0.13 – 0.68	0.01	Moderate
Paediatrician 3 - Expert	0.40	0.07 – 0.60	< 0.01	Reasonable
VISIT 2
	ICC	CI 95%	p	Reproducibility Strength
Right Hip Angle α
Paediatrician 1 - Paediatrician 2	0.74	0.69 – 0.79	< 0.01	Substantial
Paediatrician 1 - Expert	0.37	0.16 – 0.67	0.05	Reasonable
Paediatrician 2 - Expert	0.45	0.20 – 0.75	0.02	Moderate
Left Hip Angle α
Paediatrician 1 - Paediatrician 2	0.76	0.72 – 0.81	< 0.01	Substantial
Paediatrician 1 - Expert	0.18	0.05 – 0.60	0.2	Minimum
Paediatrician 2 - Expert	0.37	0.15 – 0.68	0.06	Reasonable
Right Hip Angle β
Paediatrician 1 - Paediatrician 2	0.53	0.43 – 0.61	<0.01	Moderate
Paediatrician 1 - Expert	0.04	0.01 – 0.43	0.40	Minimum
Paediatrician 2 - Expert	0.20	0.10 – 0.53	0.05	Minimum
Left Hip Angle β
Paediatrician 1 - Paediatrician 2	0.47	0.36 – 0.57	<0.01	Moderate
Paediatrician 1 - Expert	0.11	0.05 – 0.48	0.30	Minimum
Paediatrician 2 - Expert	0.21	0.10 – 0.54	0.06	Minimum

*ICC: intraclass correlation coefficient

Graf method is the most widely used ultrasound static technique for neonatal hip assessment (2, 11). It requires standardised imaging and has been accepted by health systems of several European countries. The extensive demonstration of the validity of the method, as well as the good results in other European countries, makes us select this technique as the most appropriate for our study (12). This screening method is performed in most European centres by trained personnel specialized in traumatology and orthopaedics (13). In our environment, we consider that, since examination and follow-up of the newborn and healthy child is carried out by paediatricians, they are the most suitable for performing the technique and it allows to implement universal screening, without overloading other specialties.

Different studies have evaluated reproducibility of Graf's ultrasound method, however, in none of them this technique is performed by paediatricians. Karakus studies degree of agreement between traumatologists (14). In the study led by Chang et al., orthopaedic surgeons performed three ultrasound assessments according to Graf method on 248 newborns. In his study, the most representative images of each hip are selected by three observers with different levels of experience (15). In our study agreement degree among paediatricians between healthy or pathological hip diagnosis, hip type of Graf's classification and α - β values are compared.

The current study analyzes results obtained between paediatricians and between paediatricians and an external expert. It is observed that results between paediatricians are constant in both visits. Agreement degree with external expert, however, is very high in the first month, decreasing throughout the study. This lower correlation can be explained by the fact that, as the study progresses, the expert only reviews images in which there is no agreement among observers. These images are possibly more difficult to interpret and this could be a source of bias.

The study presented, stands out for being one of the few that compare results between healthy and pathological hips. Roover’s study, as in our case, is one of the few published that evaluates agreement degree between observers according to healthy or pathological hip. In their study, 200 images were evaluated by five observers (one radiologist and five X-ray technicians) obtaining a slightly lower degree of agreement than in our study, varying between moderate and substantial. In terms of using Graf ultrasound as a screening method, this comparison seems to be the most useful. When hip is evaluated, it is necessary to discern initially whether it is healthy or pathological, and once a pathological hip is detected, the most accurate evaluations should be carried out by expert radiologists or traumatologists.

Reinhard Graf formulated guidelines for diagnose the different types of neonatal hips, eliminating inter-observer variability of the results (16). This has been corroborated by several authors to date. Karakus finds an almost perfect degree of agreement in all his comparisons and Pedrotti et al, published in 2020 an agreement between observers up to 91% (14, 17). Among all studies on topic, to date, this study compares the largest number of ultrasonography scans analyzing 3128 images, and the degree of agreement obtained is very good. It should also be noted that compared to the rest of the published series, in our case we compared two or three ultrasound scans performed by different paediatricians on the same patient, whereas, in most of the previously published studies, a single image is interpreted by several observers. This aims to demonstrate a greater scientific rigor to extrapolate our findings to daily clinical practice.

Finally, our study compares agreement between α and β angles values. As in the previous results, agreement degree among paediatricians for α angle value is almost perfect. Chang et al. found an ICC of 0.67 (95% CI: 0.38–0.78) for the value of this same angle, Simon et al. obtained 90% agreement, and in the study led by Pedrotti and Copuroglu the ICC 0.72–0.82 (15, 17, 18). Therefore, α angle measurement, according to Graf method is very consistent and can be obtained by paediatricians for DDH screening reliably.

β angle measurement, however, is less solid. In our study, we observed that agreement among paediatricians in the assessment of this angle is moderate. This not so great agreement in β value has already been expressed by various authors. Copuroglu finds similar results to ours with an ICC 0.47 (18) Omeroglu et al. report a 10.1° variation in β angle value between different observers and Karakus finds a minimum degree of agreement (14, 19). This disparity in β angle may be due to the difficulty presented by sonographers in correctly identifying the labrum. β angle is determined by the intersection of two lines, the baseline and the cartilaginous roofline and this second line is tracing from the labrum to the promontory. Adequate training of observers in anatomical structures recognition could reduce this results disparity. Even so, since the aim of this study is to test the use of Graf ultrasound as a screening method, the determination of α angle would be sufficient. The usefulness of the β angle lies in differentiate type IIc joint stability, exceeding the scope for which Graf ultrasound is intended to be studied in our case.

As a conclusion, we can establish that Graf ultrasound performed by paediatricians is a reliable screening test. A low variability between different observers was demonstrated both for the diagnosis of healthy or pathological hip, between different types of hips and α value. Agreement degree of β angle is moderate, but sufficient for the use of the Graf method as a screening. Ultrasound according to Graf method by paediatricians can be used for DDH screening because it is a standardized method with strong agreement among observers.

Ethical Approval and Consent to participate.

The study is approved by the hospital's Ethics and Clinical Research Committee of Guadalajara University Hospital. The data are collected on an anonymized basis and the analysis is carried out in compliance with the ethical principles for research on human subjects proposed in the Declaration of Helsinki by the World Medical Association (WMA)

Consent for publication.

All authors of the study give their active consent to the publication of the study. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Availability of supporting data

All data from the study are available in a coded database and available to those who wish to review it.

Competing interests

The authors do not present any conflicts of interest.

No financial or non-financial benefits have been received or will be received from any part related directly or indirectly to the subject of this article.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit-sectors.

Acknowledgements

Special thanks to the paediatricians who are part of the Guadalajara University Hospital department where this study was carried out. As well as to primary care paediatricians in this same region who continue screening for this pathology outside the hospital setting. Special mention to the radiology department of our centre that collaborates with paediatricians in the diagnosis of hip dysplasia and to MD. PhD. Miguel Torralba who always collaborates with enthusiasm in the design of the studies and the interpretation of the data in our centre. Finally, our recognition to MD. Ines Osiniri Kippes for collaborating in this project, instructing us, and introducing us to child hip ultrasound. She helps us with the design and approach of the study and collaborates throughout the course of the study reviewing our images.

Chen J, Zhang WB, He JZ, Zhang R, Cao YQ, Liu X. Developmental dysplasia of the hip: A special pathology. Chin J Traumatol. 2018;21(4):238-42.
Duarte ML, Motta GGB, Rodrigues NVM, Chiovatto ARS, Chiovatto ED, Iared W. Ultrasound techniques for the detection of developmental dysplasia of the hip: a systematic review and meta-analysis. Sao Paulo Med J. 2023;141(2):154-67.
Treiber M, Korpar B, Sirše M, Merc M. Early neonatal universal ultrasound screening for developmental dysplasia of the hip: a single institution observational study. Int Orthop. 2021;45(4):991-5.
Wilf–Miron R, Kuint J, Peled R, Cohen A, Porath A. Utilization of ultrasonography to detect developmental dysplasia of the hip: when reality turns selective screening into universal use. BMC Pediatr. 2017;17(1):136.
Biedermann R, Riccabona J, Giesinger JM, Brunner A, Liebensteiner M, Wansch J, et al. Results of universal ultrasound screening for developmental dysplasia of the hip: A Prospective Follow-Up of 28 092 Consecutive Infants. Bone Jt J. 2018;100-B(10):1399-404.
Radiology, OLVG Hospital, Netherlands, Kilsdonk I, Witbreuk M, Orthopedic Surgery, OLVG Hospital, Netherlands, Van Der Woude HJ, Radiology, OLVG Hospital, Netherlands. Ultrasound of the neonatal hip as a screening tool for DDH: how to screen and differences in screening programs between European countries. J Ultrason. 2021;21(85):e147-53.
Krysta W, Dudek P, Pulik Ł, Łęgosz P. Screening of Developmental Dysplasia of the Hip in Europe: A Systematic Review. Children. 2024;11(1):97.
O’Beirne JG, Chlapoutakis K, Alshryda S, Aydingoz U, Baumann T, Casini C, et al. International Interdisciplinary Consensus Meeting on the Evaluation of Developmental Dysplasia of the Hip. Ultraschall Med - Eur J Ultrasound. 2019;40(04):454-64.
von Kries R, Ihme N, Oberle D, Lorani A, Stark R, Altenhofen L, et al. Effect of ultrasound screening on the rate of first operative procedures for developmental hip dysplasia in Germany. The Lancet. 2003;362(9399):1883-7.
Hsu KH, Chang WC, Feng CK, Su YP. Implementing the AAOS Guidelines for Screening of Developmental Dysplasia of the Hip Before the Age of 6 Months in Taiwan. J Pediatr Orthop. 2023;43(6):e416-20.
Schams M, Labruyère R, Zuse A, Walensi M. Diagnosing developmental dysplasia of the hip using the Graf ultrasound method: risk and protective factor analysis in 11,820 universally screened newborns. Eur J Pediatr. 2017;176(9):1193-200.
Thallinger C, Pospischill R, Ganger R, Radler C, Krall C, Grill F. Long-term results of a nationwide general ultrasound screening system for developmental disorders of the hip: the Austrian hip screening program. J Child Orthop. 2014;8(1):3-10.
Bucher F, Görg C, Weber S, Peterlein CD. Paediatric hip ultrasound: uncertainties in examination and choice of treatment. J Child Orthop. 2021;15(1):42-7.
Karakus O, Karaman O, Sari AS, Orak MM, Muratli HH. Is it difficult to obtain inter-observer agreement in the measurement of the beta angle in ultrasound evaluation of the paediatric hip? J Orthop Surg. 2019;14(1):221.
Chang WC, Hsu KH, Lo IF, Liao KH, Su YP. Interobserver agreement and clinical disparity between the Graf method and femoral head coverage measurement in developmental dysplasia of the hip screening: A prospective observational study of 198 newborns. Medicine (Baltimore). 2021;100(24):e26291.
graf R. Hip sonography update. Quality-management, catastrophes – tips and tricks. Med Ultrason. 2013;15(4):299-303.
Pedrotti L, Crivellari I, Degrate A, De Rosa F, Ruggiero F, Mosconi M. Interpreting neonatal hip sonography: intraobserver and interobserver variability. J Pediatr Orthop B. 2020;29(3):214-8.
Copuroglu C, Aykac B, Saridogan K, Ozcan M, Tuncer B. Reliability of ultrasonographic measurements in suspected patients of developmental dysplasia of the hip and correlation with the acetabular index. Indian J Orthop. 2011;45(6):553.
Õmeroğlu H. Use of ultrasonography in developmental dysplasia of the hip. J Child Orthop. 2014;8(2):105-13.

No competing interests reported.

Download PDF

Journal Publication

published 24 Aug, 2024

Read the published version in European Journal of Pediatrics →

Editorial decision: Revision requested
13 Jul, 2024
Reviewers agreed at journal
08 Jul, 2024
Reviews received at journal
07 Jul, 2024
Reviewers agreed at journal
06 Jul, 2024
Reviewers invited by journal
06 Jul, 2024
Editor assigned by journal
05 Jul, 2024
Submission checks completed at journal
05 Jul, 2024
First submitted to journal
01 Jul, 2024

You are reading this latest preprint version

ECAD Study: Evaluating agreement degree among paediatricians in hip dysplasia assessment by Graf ultrasound method

Status:

Journal Publication

Version 1

Abstract

Introduction:

Materials and methods

Results

Discussion

Figures

Introduction

Materials and methods

Results

Discussion

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1