*3.5. Finalisation of the Tool*

The tool was reviewed and finalized according to the four criteria described in the methods. After the review of items' performance, including the feedback from cognitive interviews, 100 hundred items were retained, 23 items were marked for deletion (Table 4) and 10 were retained as important non-scoring items (see below). Four of the twentythree items were revealed to have extensive overlap in the range of age of attainment, and two were therefore deleted. The main reasons for item deletion were, therefore: poor developmental trajectories, poor reliability, wide differences in age attainment between countries and issues with clarity of meaning from cognitive testing. These are listed together with item wording in Table 4. Two additional items were added for the higher ages to fill gaps: LAN19 "Does your child say at least six words?" and LAN22 "Does your child identify at least seven objects?" Agreement was reached by the team to finalize the IYCD. The final list of items is in the Supplementary Materials and has 100 items

with 90 of them in 3 domains of development (40 motor items, 30 language and cognitive items and 20 socio-emotional items). Ten behavioral items that showed no or very poor developmental progression on the socio-emotional scale were also retained as important but were not to be scored as part of the IYCD final score.


**Table 4.** Table of items in IYCD prototype removed and reasons why.


**Table 4.** *Cont.*

\* Please note the item numbers shown here are the item codes from the previous version of the tool.

#### *3.6. Validation of IYCD Score Using the DAZ Scoring System*

Graphs depicting the 10th, 25th, 50th, 75th and 90th centiles of attainment of each item within a domain were constructed (Figure 4).

Figure 5 shows a density plot of the DAZ scores across countries.

The plots show clear differences between countries in DAZ scores, with Brazil achieving slightly higher developmental scores overall than Pakistan and Malawi. Malawi in turn is slightly ahead of Pakistan in terms of the children in our samples. Mean (SD) DAZ scores by country were Brazil = 0.52 (0.86), Malawi = −0.13 (0.91), Pakistan = −0.43 (0.98), and these were found to be statistically significantly different (*F*(2,266) = 26.87, *p* < 0.001). Table 5 shows that DAZ scores were highly correlated with HAZ, WAZ, maternal education, SES and family environment (FCI). Some example density plots are shown in Figure 5b–d. Figure 5b shows the relationship between DAZ scores by WAZ tercile, which was significant (F(2,243) = 5.657, *p* = 0.004), mean (SD) DAZ score for lower tercile = −0.16 (1.10), middle tercile = −0.08 (0.86), and upper tercile = 0.31 (0.91). Figure 5c shows the significant relationship between SES tercile by DAZ scores (F(2, 260) = 27.00, *p* < 0.001), lower tercile = −0.38 (0.88), middle tercile = −0.21 (1.03), upper tercile = 0.56 (0.77). However, as expected the relationship between sex and DAZ was not significant (*t*(266) = −0.66, *p* = 0.51), mean (SD) DAZ score for males = −0.05 (0.98) and females = 0.03 (1.01) (Figure 5d).

**Table 5.** Correlation matrix of main variables.


Correlation involving an ordinal variable uses Spearman's correlation coefficient; all others use Pearson's. SEX: Male = 1, SEX: Female = 2; URB/RUR: rural = 1, URB/RUR: urban = 2. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001. DAZ development-for-age *z* score, HAZ—height for age *z* score, WAZ—weight for age *z* score, MAT\_ED—maternal education, SES—socioeconomic status, FCI—family care indicators.

The final version of the tool (IYCD version 1.2) is shown in File S2. The tool and all related training materials are available on the WHO IYCD website (https://ezcollab.who. int/iycd, accessed on 2 June 2021).

**Figure 4.** Domains of development with age succession of items with bars showing 10th, 25th, 50th, 75th and 90th centile for attainment across countries.

**Figure 5.** Density plots of development-for-age *z*-scores (DAZ) by country, WAZ, SES and Sex. (**a**) Development for age *z* score by country, (**b**) development for age *z* score varying by weight for age *z* scores broken into terciles, (**c**) development for age *z* score by socioeconomic status broken up into terciles and (**d**) development for age *z* score by sex.

#### **4. Discussion**

The WHO 0–3 indicators of Infant and Young Child Development (IYCD) tool is a parent report tool with 100 items (40 fine motor and gross motor, 30 language, 20 socioemotional and 10 non-scored behavior items) that shows very good reliability, clear developmental trajectories and minimal variability across countries. Through this process, we have established a set of items for children under 3 years of age that can reliably monitor children's achievement of developmental milestones at approximately the same age across countries. We have further demonstrated that these items can be reported reliably by parents and that the items reflect well the development of children in three countries on three different continents.

Our analyses showed clear developmental trajectories comparable across the three different sites for all items within the motor (gross and fine) and language domains of development. There were fewer items that followed clear developmental trajectories in the socio-emotional development domain. Moreover, some items, though deemed important for healthy development, did not follow the age-based developmental trajectory in the same way as items in the motor and language domain. In response to this, the socioemotional domain in the IYCD was limited to 20 items that met the developmental criteria. Ten remaining items, due to their relevance and associations with other critical variables, were retained in a separate scale whose scoring is not included in the development-for-age (DAZ) final score.

Recent developments in the field have demonstrated similar results to our own. For example, the results of the study on the Guide for Monitoring Child Development—a more detailed practitioner interview with a parent of a child under 3 years—demonstrated clear and comparable developmental trajectories on milestones across four countries [38] Moreover, research evidence from another new tool created for the same purpose as IYCD, the Caregiver Reported Early Development Instrument (CREDI), has also demonstrated developmental progression of young children's skills across countries [12] and strong association with family socioeconomic status [39]. This work was conducted concurrently to IYCD, and the teams are now harmonizing for future stages of work [40]. Even though the development of Ages and Stages Questionnaire has been largely based on US-based samples, and it combines observation with parent report, thus not directly comparable with our methodology nor as feasible for large-scale monitoring, results of cross-cultural use support our findings [41]. While all the tools mentioned above have specific advantages, to our best knowledge, none of them used a theoretical framework combined with such an extensive empirical database in the way that the IYCD has.

Despite the diversity of our sample, which came from three countries from three continents with diverse language, culture and customs, our analyses revealed item performance remarkably similar across the sites. We believe that the detailed operating procedures developed prior to the study and closely adhered to throughout supported this outcome. For example, the process of exact translation and back translation from the English to each local language to retain the meaning of items as closely as possible across sites, supported by the local investigators and at least two other team members for each country, ensured high quality control. In following the standard guidelines, we also used exactly the same very specific operating procedures across all sites in assessment of developmental milestones, anthropometry, socioeconomic status and family care. All difficulties and country-specific adjustments were discussed to consensus to ensure that any amendments were within the accepted margins. Our study demonstrated that it is possible to maintain such processes in the measurement of very early child development and that it leads to results comparable across countries. It is our intent to make the procedures available so that they can be utilized on a wider scale in the future.

Despite these strengths, our study had some important limitations. The sample for this study was limited to 269 participants. The next stage in this work would be to conduct a full large-scale validation study across numerous countries and settings to enable us to ensure validity and normal reference values across settings. With the recent nurturing care framework and strategic development goals promoting the need for measurement of early child development as an outcome for programmatic and population level studies [42,43], it is important that efforts to establish valid and comparable developmental monitoring tools move to the next stage. In view of the recently increased interest and volume of evidence on development of children in the first 3 years, the next stages in this work will be to compare and harmonize tools to create one version of a tool that can be upscaled, validated and standardized across multiple countries in order to provide clarity for the international community. This next step is presently underway through the harmonization of work conducted by the CREDI, IYCD and D-Score teams [40]. Our detailed study processes described here in this paper and elsewhere will support these next stages of work ([13,44] Lancaster et al., 2020).

Even though 93% of the original recruited sample was included in the final analysis, there were also data missing due to some inaccuracy in birthdate data collection. As this is so crucial to the creation of developmental trajectories, future studies should attempt to avoid this difficulty, for example, by checking the age and birthdate on several occasions during data collection as an added check to ensure accuracy. Another method to ensure accuracy of age, corrected gestational age and birthdate would be to utilize large birth cohorts for the next stage of work in conducting a wider validation of a tool for global use. We aim to use these types of cohorts for our next stages of work [40]. For this study, we used a paper version to collect data, only using the electronic sources for item demonstrations

(e.g., audio sounds or photos or videos illustrating items). In the process of creation of the IYCD, we developed a tablet version that is available for use and can be obtained through the IYCD research team.

#### **5. Conclusions**

The IYCD has shown excellent reliability and validity to be used in population level measurement of early child development in low-income settings. Further testing is required to ascertain its ability to detect the effects of intervention. Through the detailed process of its development, we created a blueprint for a global adaptation of parent-reported measure of development of children under 3 years of age.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/ijerph18116117/s1, File S1: Table demonstrating sampling framework for each country/. File S2: Final IYCD tool with 100 items.

**Author Contributions:** M.G. wrote the first draft of the paper and revised the paper. M.G., G.L., T.D., P.K., V.C., and M.J. conceived the study. All authors contributed to protocol development and refinement. All authors were involved with data collection, but M.A.R., C.R.L.A. and L.M. led data collection in each country. G.L. and G.M. led and conducted the analysis. All authors contributed to final analysis and conclusions from data. M.G. led the writing of the manuscript, but all authors discussed the results and contributed to the final manuscript. All authors have reviewed and consented for publication of the manuscript and have certified the authorship list and contributions. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study has been supported by Grand Challenges Canada and The Bernard van Leer Foundation—grants provided to the World Health Organisation (WHO). G.L. and G.M. had full access to the data, and all authors had final responsibility for the decision to submit for publication. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

**Institutional Review Board Statement:** The study protocol complied with the principles of the Helsinki Declaration. The study received ethical approval from the WHO ethical review committee (ERC 0002747) as well as in-country ethics approval from Aga Khan University, Pakistan (4139- Ped-ERC-16), Federal University of Minas Gerais, Brazil (CAAE: 53888416.6.3001.5132), the College of Medicine Research Ethics Committee University of Malawi (P.03/16/1916) and the McMaster University, Hamilton (Canadian) Ethics Review Board (HiREB 1613). Approval was also acquired from all local health organizations where recruitment took place. All families who were recruited were provided with verbal and written information regarding the study, and all caregivers signed the approved consent form.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The dataset supporting the conclusions of this article is available upon request from the WHO team and can be requested through Vanessa Cavallera; cavallerav@who.int.

**Acknowledgments:** We would like to thank all those children and families who gave their time to take part in this study. We would also like to thank the research assistants and teams from each of the three countries (Brazil, Malawi, Pakistan) who worked tirelessly to ensure that this study happened with as much detail as it did.

**Conflicts of Interest:** M.G., P.K., M.J., G.L., C.R.L.A., M.A.R., L.M. and V.C. were all provided with funding to support running of the study from the WHO as a central body for the grants provided by G.C.C. and Bernard van Leer. The authors alone are responsible for the views expressed in this article, and they do not necessarily represent the views, decisions or policies of the institutions with which they are affiliated.
