The first round of the Delphi survey began wherein a panel of five speech-language pathologists (SLPs) with more than ten years of clinical and research experience in the domain of fluency disorders were contacted. All the experts were informed about the study's purpose, expected time commitment, participation criteria, and the need to stay involved throughout the research process. They were also told that their participation in the Delphi study was voluntary. After that, written consent was obtained from all the experts. Each participant who agreed to participate was given a unique code to protect their confidentiality and anonymity. Each expert was asked for their preferred mode of receiving the Delphi round questionnaires (either printed or mailed versions), and contact was established with them accordingly. The demographic characteristics of these SLPs are provided in Table 1.

The SLPs, during this phase, were asked to provide their opinion about three questions, i.e., “Whether they feel that there is a need for a screening test for quick identification of stuttering in their clinical practice with school-going children,” “What advantages would be there if such a test is available” and “What should be the nature of speech elicitation tasks in one such tool for the school going children.” The professionals were given a span of two weeks to answer the questionnaire, with a reminder sent during that time. The researchers analyzed the answers qualitatively received from the professionals. A summary of the results was then shared with the participants as controlled feedback to inform them of other participants' points of view and offer them a chance to clarify or modify their opinions during the Delphi procedure. The collective feedback from all the experts was gathered to support the development of screening tests for children and determine the speech elicitation stimuli based on the review of the relevant literature in this domain [25]. The suggestions made by the expert panel during this phase, reading, spontaneous speech tasks and picture description, emerged as the key speech elicitation tasks relevant to the school-going children; therefore, phase 2 focused on the construction of these tasks.

During this phase, the researcher reviewed the textbooks from the 1^st – 10^th standards to obtain the repertoire of words used in each standard. The most frequently occurring words from the language textbooks of each standard were included in the word list. A common theme of stringing the words was picked up for each standard, and reading passages were made for 4^th -10^th standards using the selected words from textbooks of their respective standards. Similarly, picture cards were made available for 1^st-3^rd standards from the respective wordlists since these students were still acquiring reading proficiency and describing a picture card was easier than reading. A common theme stringing the obtained word list was finalized to be represented in a series of events in a single picture card, and it was digitally depicted with the professional help of a technical expert. The third stimulus used for screening was a spontaneous speech task where the topic for each class was selected from the repertoire of words obtained from textbooks of each class. The topic was selected so the children could spontaneously talk about a theme using age-appropriate vocabulary. Likewise, three picture cards were determined for the first, second and third standard students, three reading passages each for fourth to tenth standard and three topics each from first to tenth standard. A sample of these constructed stimuli has been described in the results section.

The third phase involved content validation of the constructed passages, picture cards and topics of spontaneous speech tasks. Five Speech Language Pathologists, four teachers and one English professor affiliated with a university teaching college were approached to validate the stimuli for each task. The demographic details of these participants are shown in Table 2. These words were also approved by the teachers of the respective standards and observed to be readable by the students of those standards.

Teachers were given the passages so that they could check whether it was readable by the students. The passages were content validated by an English Professor teaching in a University affiliated college, and modifications were incorporated accordingly. Five Speech-Language Pathologists also were asked to validate the content based on the readability of the stimuli, vocabulary and grammatical appropriateness, the flow of information, and cultural adequacy for each standard. Participants rated the developed stimuli using the questionnaire developed for this purpose. The content-validated questionnaire which was circulated is attached in Appendix A. A quantitative and qualitative analysis of the outcomes from this round of the Delphi process was conducted. A threshold of 80% was used as the agreement standard, and the consensus level achieved was calculated. Any items that received agreement from more than 80% of participants were kept, while those with less than 20% agreement were removed. For items with agreement between 80% and 20%, the researchers reviewed ratings and comments to make revisions. The revisions continued until all the items reached the desirable consensus on the agreement.

Participants for phase IV included ten SLPs working in the field of fluency disorders with similar inclusion criteria as the one followed during the first phase. The expert panel in the fourth phase differed from Phases 1 and 3. The correspondence with these experts was done similarly during the first phase, and informed consent was obtained before their participation. The demographic details of these participants are shown in Table 3.

These 10 participants were contacted independently per their convenient mode of correspondence and asked to list the ten most critical differential diagnostic features in a ranked manner, which they use to differentiate stuttering from normally occurring disfluencies among school-going children. All the responses of the SLPs were tabulated and analyzed to obtain the features which the panelists agreed upon in descending rank order.

Once the content of the developed tool reached a satisfactory mark and the diagnostic criteria received the consensus, the Delphi procedure was terminated, and the results were shared with the expert panel with an acknowledgement note.

During the first round of Delphi, the expert panel comprising five SLPs was asked initially about the need for developing a screening test for stuttering. The results revealed that 100% of these panel members agreed to develop a test to screen stuttering among school-going children. The factors that were determined to help create a screening test included impartiality in screening, using consistent and appropriate stimuli for the Indian population's culture, and the availability of adequate time and personnel to screen a large number of children in their school surroundings. When asked about the nature of speech elicitation tasks in one such tool for the school-going children, the panel recommended the usage of picture cards for those who had not yet achieved adequate reading and writing skills, as in the first and second standard students; reading passages for the students between third to tenth standard; and a spontaneous speech task with a topic relevant to the age and stage of the child. The panel further emphasized that these picture cards, reading passages and topics for spontaneous speech should be determined to be appropriate to the Indian culture and incorporate the vocabulary appropriate to the given child’s language age.

The content and item selection for the screening test of stuttering was based on theoretical constructs related to the development of fluency skills and the opinions of the experts during the first phase. The three sections of the screening test considered were Picture description, reading and spontaneous speech tasks. In line with these suggestions, an initial list of words selected was from the school-going children's academic textbooks. These word lists were analyzed, and the frequently occurring content words were used to construct a reading passage. Reading passages were constructed using these words with the appropriate prepositions and conjunctions to connect the words. The sentences for the reading passage were constructed in such a way as to give a connected flow of information on a particular topic. The number of words selected from the textbooks for each standard, with examples of some of the chosen words and constructed sentences, have been provided in Table 4.

Per the expert panel's suggestions, reading activity was not considered for children in 1^st -3^rd standards. Picture description task was thus used instead of reading passages for these children. Picture cards were digitally created, depicting a theme consisting of words chosen from their textbooks. Three picture cards were digitally created for each standard based on the increased opportunities to talk from the available pictures. Three to four target words from their textbooks were included in each picture task, and those words could be elicited during this task. A sample of some of these pictures is shown in Fig. (1).

Similarly, a set of 3 topics were selected from the word list generated from their textbooks for the purpose of spontaneous speech tasks. The topics were selected so that there are opportunities to discuss the same. The explanations of the topics are already provided in the textbooks, and it was selected under the pretext that children would have learned about them in the classroom. Of the 3 topics selected, each student was free to choose and talk about it based on their knowledge. A sample of these topics for each of the standards has been given in Table 5.

Fig. (1). A sample of some of these pictures.

The content validity of the proposed screening test was conducted by asking the expert panel about the relevance of each subsection of the screening protocol and finding out the coherence of each section with respect to the age and class of the student being tested. The questions consisted of the relevance of separate sections and the appropriateness of the stimuli used under each section. A set of close-ended questions were included in the content validity questionnaire. All the experts either agreed or strongly agreed on the questions pertaining to the significance and suitability of the subsections. The expert panel proposed recommendations on the addition or removal of items. Feedback from the experts’ panel on the screening tool was used to modify the overall contents of the test. The university English professor corrected the grammatical errors in certain reading passages. The school teachers also commented on including age-appropriate words for the reading task according to each standard. Only the stimuli either agreed or strongly agreed by 80% of the panel members were included in the final test. The modified version of the proposed test was mailed back to panelists with the recommended changes. The panel reviewed the modifications and approved the test stimuli in this round of the Delphi survey.

The differential diagnostic criteria were developed based on the inputs obtained from the expert panel based on their clinical expertise. Each of their opinions was considered, and the feature most suggested was ranked top in the list of differential features, and likewise, the features were orderly arranged. The common outcome that emerged from all these helped give the screening test a better outline. All ten participants contributed to the Delphi process resulting in a 100% response rate. Many of the features which the experts identified were overlapping and thus removed. Table 6 represents each of the differential features and the consensus of the experts.

The first five diagnostic criteria (1-5 from Table 6), which received the highest consensus among the rankings provided by the experts from this list, were considered for identifying stuttering among school-going children.