Observer Variability in Head Circumference Measurement Using Routine Versus Non-Stretchable Tapes in Children

Address for correspondence Dr. Pareshkumar A Thakkar Associate Professor Department of Paediatrics Sir Sayajirao General Hospital & Medical College, Vadodara, Gujarat, India. E-mail: drpareshthakkar123@gmail.com Phone no: 91-9426032150 1Dr. Pareshkumar Thakkar, MBBS. MD, Associate Professor, Department of Paediatrics, Medical College Baroda and SSG Hospital, Vadodara. 2Dr. Kinjal Yagnik, MBBS. DCH, Resident, Department of Paediatrics, Medical College Baroda and SSG Hospital, Vadodara, 3Dr. Niyati T Parmar, MD, Tutor, Department of Community Medicine, Medical College Baroda and SSG Hospital, Vadodara, 4Dr. Rashmi Ranjan Das, MD, Assistant Professor, Department of Paediatrics, All India Institute of Medical Sciences, Bhubaneswar, India, 5Dr. Ukti P Thakkar, BPT, Physiotherapist, VIP Physiotherapy Clinic, Vadodara, Gujarat. Abstract


Introduction
H ead circumference (HC) or occipito-frontal circumference (OFC) monitoring of children is very crucial, and is a basic tool for early detection of neuro-developmental disorders 1,2,3 .An abnormal rate of head growth suggest a pathological disorder, and would require further evaluation 4 .Hence, there is need of a valid, precise, accurate, and reliable tool for OFC measurement in children.
Currently used measuring tapes for taking anthropometric measurements do not give accurate readings after being used for few months because of their stretchability 1 .In addition, there are no recent guidelines on when to replace the tape.Besides these, inter-observer (diff erent observer) and intra-observer (same observer) variations are very common because of diff erent forces (stretch) applied while taking OFC measurements 1,5,6 .Observer variations in anthropometric measurements of > 0.5 cm are common.This might result in missing an at-risk child or might be a false alarm and subject a child to unnecessary work-up and impose signifi cant stress among parents.This emphasizes the need of a tool which gives accurate OFC measurement with no or minimal variability, and is able to provide reliable measurement even after repetitive use.
Though it is known since long that signifi cant observer variability exists in anthropometric measurements, and stretchable tapes are not reliable after being used over a period of time, there are no published studies on how to minimize this variability.Non-stretchable exclusive tapes for OFC measurement have recently become available 7 .The aim of this study was to fi nd whether signifi cant observer variability exists and whether use of non-stretchable tape can minimize the observer variability or not.We also intended to study whether the stretchable and non-stretchable tapes are able to provide accurate readings after being used repeatedly over a period of time.

Material and Methods
This cross-sectional study was carried out over a period of six months (November 2013 to April 2014) at a tertiary care teaching hospital of Western India.All stable children in the age group of one month to fi ve years whose parents were willing to participate and ready to give consent, were included in this study.All the children meeting the inclusion criteria or sample selection criteria were enrolled by the observers (post-graduate students) on a particular day, the limit being up to fi ve children per day (this fi gure was chosen to prevent observer fatigue on the given day).The study was approved by Scientifi c Review Committee and Institutional Ethics Committee on Human Research.Participants were enrolled in the study after obtaining consent from the parents.

Outcome variables and their measurement:
The outcome variables were: (a) the intra-observer and inter-observer variability in OFC measurement by routinely used stretchable and non-stretchable tapes, and (b) accuracy of readings of old stretchable OFC measurement tapes.
Initially, we assessed the accuracy of various tapes by comparing them against the standard steel rule (scale).We took two new stretchable, two new non-stretchable and ten old stretchable OFC tapes.The stretchable tape (Butterfl y Brand Tape, China) has the following specifi cations: material (soft fi ber glass plus plastic cover), measuring range/length (0 -150 cm), width (1.5 cm), graduation (measuring rod: 1 mm/0.1 cm), and weight (60 g).The non-stretchable tape (Ibis Medical, India) has the following specifi cations: material (fl exible PVC), measuring range (0 -55 cm), graduation (measuring rod: 1 mm/0.1 cm), and measure (W x H x D: 25 x 1 x 663 mm).The old tapes selected were the tapes which had been used repeatedly for at least 3 months by the resident doctors in the paediatric wards, clinics and outpatient departments.Measurement of OFC was carried out by the new stretchable and non-stretchable tapes.The non-stretchable exclusive OFC tapes were also periodically (monthly) assessed for their reliability and accuracy during the study period.
Two observers were required for testing of the inter-observer variations.At the start of the study, both the observers (post-graduate students) were trained for taking OFC measurement accurately by the guide (faculty).They were taught about the important landmarks for measurement of OFC.Both the observers used the same stretchable and non-stretchable exclusive tapes for OFC measurements.
The stretchable tapes were discarded after their use for a period of one month, and thereafter a new stretchable tape was used.Non-stretchable exclusive OFC tapes were assessed every month for their reliability and accuracy against the standard steel rule (scale) for OFC measurement.It was decided, if repeated use of non-stretchable exclusive OFC tape was not able to provide accurate reading, then it would be discarded and a new tape would be used.Necessary measures and precautions were taken to avoid possible bias which could have arose while taking OFC measurements.
There was a possibility of bias if observer-2 had access to the readings taken by observer-1.To avoid this, observer-2 was blinded to the readings taken by observer-1.Observer-1 checked the inclusion criteria, enrolled the patient, entered the primary details in Proforma, and took fi rst reading of OFC.Observer-1 provided observer-2 just the list of participants in a separate sheet.Obsever-2 did not have access to the readings taken by observer-1.Observer-2 took independent measurement while observer-1 was not present, and entered in a separate sheet.Observer-2 used the same measuring tapes which were used by observer-1.The readings were taken on the same day, but the timings were diff erent.
For studying intra-observer variability, the observer-1 took three readings of OFC of same patients at diff erent times of the same day.This also possibly could lead to bias in two ways.First, if the observer-1 has access to his/ her previous readings, he/she might give closer readings.To decrease/eliminate this bias, observer-1 entered the fi rst reading in the Proforma, and subsequent two readings (entered in separate sheets) were handed over to the guide (faculty) for entry into the Proforma.Second, was the possibility of recall bias.To decrease/eliminate this, observer-1 enrolled nearly ten patients on a given day.The subsequent readings were taken after signifi cant lapse of time of at least 4-6 hrs.Both the observers took all the readings by using both stretchable and non-stretchable exclusive OFC tapes.All the readings were entered into the Proforma by guide (faculty).
Three diff erent readings taken by the same observer were analyzed for intra-observer variability.For analyzing inter-observer variability, the fi rst reading out of three readings taken by observer-1 was compared with one reading taken by observer-2.Then we compared the diff erence in observer variability with the use of stretchable and non-stretchable exclusive OFC tapes.
Sample size was calculated keeping the allowable error alpha 5% and power of study 80%.The required sample size was 130(N=4pq/L 2 ).We have enrolled total 301 patient which gives power of 96% to our study.Data were entered into Microsoft Excel Sheet (version 10.0).The data was checked for accuracy by two persons involved in the study.MedCalc for windows software (version 12.5.0)was used for statistical analysis.The data was analysed quantitatively using paired t-test, unpaired t-test, analysis of variance (ANOVA), and posthoc tests.The data was analysed for inter-observer and intra-observer variability.The diff erences between the measurements, their correlation, and agreement were analysed using the correlation coeffi cient (Pearson) and intra-class correlation coeffi cient.Inter-observer and intra-observer measurements were compared by ANOVA, with the F value ≥ F critical indicating signifi cant diff erence 8,9 .

Results
Three-hundred and one children aged one month to fi ve years were enrolled in this study: 42.5% were in the age group of 1 to 6 months (128/301).The male (55.1%): female (44.9%) sex ratio was 1.23:1.

Observations on accuracy of various measuring tapes
At the beginning of the study, we assessed accuracy of all available stretchable tapes which were used repeatedly over diff erent period of time.We looked at the diff erence in the reading by comparing various tapes against standard steel (scale) rule at 35cm and 50 cm level.When we compared accuracy of new stretchable tape and new non-stretchable tape, we found no diff erence in readings of both the tapes at 35 cm and 50 cm against the steel (scale) rule.We took ten old stretchable tapes which had been used for a period ranging from 3 to 8 months.The tapes which had been used for a period 3-4 months showed a diff erence of 0.1 to 0.4 cm at the level of 35 cm and 50 cm.Whereas, the tapes which had been used for 5 to 8 months showed a diff erence of 0.5 to 1 cm (Table 1 and Figure 1).It is obvious that use of such old tapes would lead to signifi cant diff erence in the observations, even under ideal conditions.Such a big diff erence or inaccuracy in the reading is clinically not acceptable.To avoid this, in the present study, we used new stretchable tapes, and changed them every month.The non-stretchable exclusive OFC tapes were assessed for their accuracy at monthly interval, during the entire study period (Table 2).The non-stretchable exclusive OFC tapes gave consistently accurate readings even after being used repeatedly over a period of six months.

Intra-observer variability
The fi rst observer took head circumference measurements at three diff erent times with both stretchable and non-stretchable exclusive OFC tapes.Observer was masked to own previous readings.ANOVA and post-hoc tests were applied to see whether

Stretchable tape:
The measurements obtained by the same observer using stretchable tape at three diff erent times were consistent without any signifi cant diff erence (p = 0.999, F value = 0.001).The post-hoc analysis compared each one of the three readings with the rest of the two readings, but didn't show signifi cant diff erence.

Non-stretchable tape:
There was no signifi cant intra-observer variability with the use of non-stretchable tape (p = 0.998, F value = 0.002).The post-hoc analysis compared each one of the three readings with the rest of the two readings, but didn't show signifi cant diff erence.
Good correlation was observed for intra-observer variability for both stretchable and non-stretchable tapes (Table 2).

Comparison of stretchable vs. non-stretchable tape:
When intra-observer variability was checked for stretchable and non-stretchable tapes, the diff erence observed was highly signifi cant (p < 0.001) (Table 3).This suggests that intra-observer variability is signifi cantly less with the use of non-stretchable exclusive OFC tape.This means, multiple readings obtained by the same observer are more consistent with the use of nonstretchable compared to stretchable tape.

Observed variability in clinical context:
The intraobserver variability in the measurement of ≥0.5 cm using stretchable tapes was seen in 79 (26%) cases and using non-stretchable tapes was seen in 30 (9.9%) cases, the diff erence being signifi cant (p < 0.001).The diff erence of ≥0.5 cm for any case is not acceptable clinically (Table 4).

Inter-observer variability
OFC measurements were taken by two diff erent observers independently using both stretchable and non-stretchable exclusive OFC measuring tapes.Independent t-test was used to analyse the variability in the measurements obtained by two diff erent observers.No signifi cant diff erence was observed with the use of stretchable (p = 0.86) and non-stretchable (p = 0.81) tapes.It is worth mentioning that, the stretchable tapes used were new tapes, and both the observers were trained for measurement of OFC.Good correlation was observed for inter-observer variability for both stretchable and non-stretchable tapes (Table 2).

Comparison of stretchable vs. non-stretchable tape:
The mean diff erence observed in the measurement of OFC between two observers using stretchable tape was 0.27 cm, and using non-stretchable tape was 0.23 cm.The diff erence observed was statistically signifi cant (p = 0.038) (Table 3).This suggests, with the use of nonstretchable tape, observation remains more consistent between two observers, as compared to stretchable tape.

Observed variability in clinical context:
The interobserver variability in the measurement of ≥0.5 cm using stretchable tapes was seen in 79 (26%) cases and using non-stretchable tapes was seen in 48 (15.9%) cases, the diff erence being signifi cant (p = 0.003).The diff erence of ≥0.5 cm for any case is not acceptable clinically (Table 4).

Discussion
In the present study, we found good correlation in the measurements taken by two observers, and three diff erent measurements of the same observer, correlation co-effi cient being > 0.99.However, with the stretchable tape, 26% subjects had intra-observer and inter-observer variability of ≥ 0.5 cm, which was 9.9% and 15.9% with the use of non-stretchable tape, respectively.Intra-observer variability and inter-observer variability were signifi cantly less with the use of nonstretchable tape compared to stretchable tape.Similar fi nding have been observed by a previous study using that compared the stretchability between the old and new tape 1 .The authors observed the intraobserver variability to be statistically insignifi cant, but in 5% measurements it was clinically signifi cant (diff ered by ≥ 0.65 cm).Similar was their fi nding with inter-observer variability, which was statistically insignifi cant, but in 9% measurements it was clinically more signifi cant (diff ered by ≥ 1 cm) 1 .
In another study using blank tape measures for anthropometric measurements (including OFC) in term infants, the intra-examiner mean absolute diff erence in OFC measurements was lesser than that of interexaminer measurements (intra-examiner: no diff erence, and inter-examiner: 0.01 cm diff erence) 6 .This means repeated measurements by same individual was more accurate than measurements by diff erent individuals.However, we could not fi nd any study comparing observer diff erence using non-stretchable exclusive OFC tape and any study comparing stretchable versus non-stretchable tapes.
In a similar study done previously, the authors made a statement that; their study was not designed to address the issue of timing of replacement of the tapes 1 .
In the present study, stretchable tapes which have been used over a period of 3-4 months showed diff erences of 0.1 to 0.4 cm, and tapes used over 5 to 8 months showed the diff erences of 0.5 to 1 cm.Non-stretchable tapes gave consistently accurate readings even after being used repeatedly over a period of 6 months.In routine clinical practice, stretchable measuring tapes are widely used, and only one reading is taken.In a previous study, the authors applied normal range of force required in clinical practice (calculated using Newton meters) to the new as well old stretchable tapes.They found that, in the clinical force range, the old tapes stretched signifi cantly compared to new tapes 1 .
It is easy to understand that stretchable tapes after repeated use are not reliable as they will not give accurate readings.This element of inaccuracy due to stretchability added to the element of observer variability would make the measurements highly unreliable.It is obvious that use of such old tapes will result in signifi cant diff erence in the observation even under an ideal condition.Such a large diff erence or inaccuracy is clinically not acceptable.In the present study, we indirectly found the duration till which stretchable tapes give reliable readings.However, there is no specifi c guideline regarding after what duration/period, a stretchable tape should be discarded, and a new tape should be used.Based on the fi ndings of the present study, we should either use non-stretchable tapes or, if we are using stretchable tapes, they should be new, periodically assessed for their reliability, and replaced The strengths of the present study are following: adequate sample size, use of non-stretchable OFC tapes and comparison with stretchable tapes, assessment of both intra-and inter-observer variability in same setting with robust methodology, and data from a tertiary care setting of a developing country (can be generalised to such settings).However, the study is not without limitations.First, there was no true value of measurements which could have served as gold standard for comparing the inter-observer and intraobserver variations.Second, we did not try to quantify the observer variability from the "true value".Third, did not aim to study after what duration a stretchable tape do not give accurate reading or the time when it should be discarded.
Future research should be directed to address the issue after what duration stretchable tapes should be discarded, and how to minimize the observer variability.Taking average from the multiple readings would give value near to the "true value".Future research should also try to address the question regarding "minimum" number of readings taken to fi nd an average value that is near to the "true value", and the reading is clinically acceptable.

Conclusions
Variability in head circumference (OFC) measurement is signifi cantly less with the use of nonstretchable exclusive tapes compared to routinely used stretchable tapes.The intra-observer and interobserver variability though statistically insignifi cant, was signifi cant in the clinical context.Stretchable tapes used repeatedly over certain periods (5 to 8 months) were not able to provide accurate readings, while non-stretchable tapes could provide accurate readings at the end of study period of six months.

Fig 1 :
Fig 1: Old stretchable tapes showing diff erence of 1 cm when compared to standard steel rule at 35 and 50 cm.

Table 1 :
Accuracy of used stretchable tapes

Table 2 :
Pearson correlation for the observer variability

Table 3 :
Comparison of observer variability -stretchable versus non-stretchable