Effectiveness of a behavioral treatment protocol for selective mutism in children: Design of a randomized controlled trial

Selective mutism (SM) is a relatively rare anxiety disorder, characterized by a child's consistent failure to speak in various specific social situations (e.g., at school), while being able to speak in other situations (e.g., at home). Prevalence rates vary from 0.2% to 1.9%. SM is usually identified between the ages of 3–5 years. It is often underdiagnosed and consequently children receive no or inadequate treatment, with negative consequences for school and social functioning. If left untreated, SM can result in complex, chronic anxiety and/or mood disorders in adolescence and impaired working careers in adulthood. Currently, no evidence-based treatment for SM is available in the Netherlands, therefore this study aims to [1] test the effectiveness of a treatment protocol for SM that is carried out at school, and to [2] identify baseline predictors for treatment success. This article presents the design of a randomized controlled trial into the effectiveness of a behavioral therapeutic protocol for selective mutism in children (age 3–18). The expected study population is n = 76. Results of the treatment group (n = 38) will be compared with those of a waiting list control group (WCG) (n = 38). Pre and post treatment assessments will be conducted at comparable moments in both groups, with baseline assessment at intake, the second assessment at 12 weeks and post-assessment at the end of treatment. If proven effective, we aim to structurally implement this protocol as evidence-based treatment for SM.


. Background
Se lec tive mutism (SM) is a se verely de bil i tat ing anx i ety dis or der, mostly oc cur ring in young chil dren [ 1 ]. It is char ac ter ized by con sistent fail ure to speak in sit u a tions where speak ing is ex pected (e.g., in school), while speak ing freely in other sit u a tions (e.g., at home). Preva lence rates vary be tween 0.2% and 1.9% [ 2 , 3 ], with a boys:girls ra tio of 1:2 [ 3 ]. SM is more com mon in bilin gual or mul ti lin gual children [ 4 ]. Not speak ing in so cial sit u a tions im pairs per sonal, so cial, and aca d e mic de vel op ment. If left un treated, SM can have a chronic char ac ter [ 5 , 6 ].
Cur rently, it is still un clear which mech a nisms af fect treat ment suc cess in chil dren with SM. For child hood anx i ety dis or ders in general, it is known that parental fac tors (par ent ing style, parental psychopathol ogy and pre vi ous health care con sump tion) can in flu ence treat ment suc cess [ 7 ]. A study of Oer beck, Stein [ 8 ] shows that children who are treated for SM at a younger age, show faster treat ment progress than older chil dren. These re sults were main tained on longer term [ 9 ]. In an other study of Oer beck, Over gaard [ 10 ], be sides age, higher sever ity at base line, fa mil ial anx i ety and com pli ance of par ents were pre dic tors of treat ment suc cess. Also the pre dic tive value of bilin gual ism, gen der, co mor bid ity and mi nor ity sta tus was in ves ti - gated. These re sults were less com pelling. As the au thors also note, pos si bly due to a rel a tive small sam ple size (N = 32) ef fect sizes were small. Es pe cially at fol low -up due to con se quences of drop -out dur ing the study. In our clin i cal in sti tu tion we see a large, het ero ge neous pop u la tion of chil dren with SM, with many bilin gual chil dren, as also de scribed by Elizur and Pered nik [ 4 ] and Top pel berg, Tabors [ 11 ]. Although sev eral ev i dence -based treat ments, based on (cog ni tive) behav ioral ther apy are avail able for anx i ety dis or ders in chil dren [ 12 ], stud ies in ves ti gat ing treat ment ef fects in chil dren with SM are still lim ited. A re cent non -randomized trial showed that be hav ioral techniques with ad di tional par ent child in ter ac tion ther apy im prove SM symp toms [ 13 ]. Three pre vi ous ran dom ized con trolled tri als (RCT) inves ti gated showed that be hav ioral ther apy was ef fec tive for se lec tive mutism [ 8 , 14 , 15 ]. These stud ies pro vided ther apy in dif fer ent set tings i.e. in the clin i cal set ting [ 14 , 15 ] and at home be fore start ing at school [ 8 ]. Un for tu nately sam ple sizes in these stud ies are rel a tively small (N = 21, N = 24, N = 29). Cur rently, no ev i dence -based proto col specif i cally for the treat ment of SM is avail able for the Dutch pop u la tion. Since such a pro to col is ur gently needed in clin i cal practice, in our in sti tu tion the pro to col 'Speak ing in school, a mat ter of do ing' was de vel oped [ 16 ]. A con sid er able strength com pared to most pre vi ous stud ies is that this treat ment is pro vided at school, where the prob lem is most pre sent. The ra tio nale is that treat ment in school makes it eas ier to gen er al ize speak ing be hav ior to set tings where the child still has dif fi culty speak ing. In this RCT we aim to in ves ti gate the ef fec tive ness of this be hav ioral ther a peu tic pro to col, com pared to a wait ing list con trol group. A ma jor method olog i cal ad van tage is that we aim to ex e cute a suf fi ciently pow ered RCT. Sec ondly, we aim to iden tify pu ta tive pre dic tors for treat ment suc cess in chil dren with SM.

. Methods/ design
De sign: this study is a sin gle -center, sin gle -blinded ran dom ized con trolled trial (RCT), com par ing pro to col ized be hav ioral ther apy for se lec tive mutism to a wait ing list con trol group.
In clu sion: el i gi ble are chil dren be tween the age of 3 -18 years old, with an es ti mated IQ of 85 or higher, who un der stand Dutch and are re ferred to our aca d e mic cen ter for di ag no sis and treat ment of se lective mutism be tween Sep tem ber 2018 and March 2020. This large age span cov ers all con sec u tive pa tients re ferred to our in sti tu tion. Though the vast ma jor ity of these chil dren is un der the age of 8, we also get re fer rals up to 18 years.
Ex clu sio n : in el i gi ble are chil dren who have pri mary di ag noses other than se lec tive mutism. Co mor bid ity is not ex cluded, but the primary and most im pair ing di ag no sis has to be se lec tive mutism. Dur ing the in take, it will be as sessed what the most ur gent need for help is and what type of care is most in di cated. In the case that treat ment is in di cated for a co mor bid di ag no sis at first, the child will be ex cluded from the study.

1 . Patient recruitment and procedure
Par ents of el i gi ble pa tients will be per son ally in formed about the re search pro ject by the psy chol o gist -researcher and will re ceive written in for ma tion about the re search pro ject. Writ ten in formed con sent will be ob tained from par ents/ guardians for pa tients un der the age of 12. From the age of 12 writ ten con sent will also be ob tained from them. Only pa tients (≥12 years), who have given in formed con sent, or whose par ents have signed in formed con sent will par tic i pate in this study.
Af ter in formed con sent is pro vided, chil dren are ran domly assigned to treat ment for se lec tive mutism (N = 38) or the wait ing list con trol group (WCG: N = 38).
Ran dom iza tion will be strat i fied for gen der, age (chil dren aged until 7 years old are strat i fied into the "young" group, and chil dren ages 8 and up are in the "old" group) and bilin gual ism (speak ing one language, i.e. Dutch, ver sus speak ing mul ti ple lan guages). Ran dom ization will be per formed by an in de pen dent re searcher, who will in form the ther a pists re gard ing the ran dom iza tion out come.

2 . Intervention
A stan dard ized, be hav ioral ther a peu tic pro to col was de vel oped in the Nether lands to treat SM [ 16 ]. Chil dren learn in dif fer ent steps, by us ing grad ual ex po sure to chal leng ing sit u a tions, shap ing and mod elling. The ad van tage and in no v a tive as pect of this pro to col is that treat ment is car ried out solely at school. Since at school the prob lem is pre sent most ev i dently, a ther a pist goes to school to prac tice in divid u ally with the child. This helps to gen er al ize the learned ex er cises to the school sit u a tion. Every week the child learns step by step to make sounds and to speak, af ter which the speak ing be hav ior is extended to var i ous places in school with as the fi nal goal: par tic i pat ing ac tively in the cir cle dis cus sion in the class room.
The pro to col is mainly based on be hav ioral tech niques, with as the pri mary aim step wise prac tic ing to speak. How ever, for chil dren from 12 years and older, ad di tional cog ni tive el e ments (e.g. help ing thoughts) which are used in tra di tional cog ni tive be hav ioral ther apy for anx i ety [ 17 ], are avail able to be used op tion ally. The ther a pist involves the teacher as a co -therapist in the treat ment process. Af ter each weekly treat ment ses sion at school, the ther a pist in structs the teacher how profit can be main tained and which ex er cises to prac tice dur ing the week. The teacher then con tin ues the ex er cises through out the week, find ing short mo ments in time dur ing the week to prac tice with the child. The ther a pist is in touch with the teacher face to face af ter each ther apy ses sion, by email or phone when nec es sary.
Pre vi ous clin i cal ex pe ri ence learns that the av er age treat ment dura tion with this pro to col is 22 weeks.

3 . Assessments
Par ents, chil dren older than 8 years and teach ers will com plete ques tion naires on line. Par ents will also be in ter viewed by the researcher.
The psy chol o gist -researcher who per forms the as sess ments will be blinded for par tic i pants' al lo cated in ter ven tion. Chil dren and par ents, as well as the treat ment -staff are ex plic itly asked not to talk about their ran dom iza tion con di tion with the psy chol o gist -researcher. Both groups are treated with the stan dard treat ment pro to col for se lec tive mutism.
As sess ments will be car ried out at the fol low ing time points [1]: T1, base line as sess ment dur ing in take, be fore in ter ven tion starts [2]; T2, 12 weeks af ter base line, ei ther af ter first 12 weeks of treat ment or 12 weeks of wait ing list pe riod [3]; T3, af ter last treat ment ses sion. If at any of the as sess ments or dur ing the wait ing list and/ or treat ment pe riod there is sud den need for im me di ate in ter ven tion (e.g. in the event of a cri sis), un blind ing will take place to en sure that the par tic ipat ing pa tient re ceives the nec es sary care.
To ac quire in for ma tion from all in for mants (par ents, guardians, teach ers, chil dren/ ado les cents), stan dard ized ques tion naires and inter views will be used with ad e quate psy cho me t ric prop er ties, ex cept for the Dutch trans la tion of the SMQ, which will be val i dated in this study.
In Table 1 , all as sess ments are listed.

4 . Primary outcomes
Se lec tive Mutism symp toms: to ob tain par ent re ports of se lec tive mutism symp toms in their child, a) The Dutch trans la tion of the Se lective Mutism Ques tion naire (SMQ) [ 18 ] and b) the se lec tive mutism sec tion of the Anx i ety Dis or ders In ter view Sched ule [ 19 ] will be used. For the pur pose of this study, the Dutch SMQ will be val i dated. The SMQ con sists of two scales, the symp tom scale (17 items, response cat e gories: "al ways", "of ten", "some times" and "never"), and the im ped i ment scale (6 items, re sponse cat e gories: "no", "mod er ate", "fairly" and "a lot"). Within the symp tom clus ter, sub scales for school, home and so cial en vi ron ments are dis tin guished [ 18 , 20 ]. A higher score in the SMQ in di cates more prob lems with speak ing.
Val i da tion of the SMQ will be based on the base line scores (T1) of all chil dren re ferred to our cen ter for di ag no sis and treat ment of se lective mutism.

5 . Secondary outcomes Anx i ety/ mood symp toms
In ter view: The Anx i ety Dis or ders In ter view Sched ule (ADIS -C) clin i cal in ter view is used to ob tain in for ma tion from par ents about pos si ble co mor bid prob lems.
Ques tion naires: Emo tional and be hav ioral prob lems will be assessed by the fol low ing ques tion naires, which have par al lel items.
The Child Be hav ior Check list (CBCL) [ 21 , 22 ] will be used to obtain stan dard ized par ent re ports of emo tional and be hav ioral problems in their child. The Youth Self Re port (YSR) [ 21 ] is the par al lel ver sion of the CBCL, but items are for mu lated in the child form. Children of 11 years and older will com plete the YSR.
The Dutch ver sion of the Teacher's Re port Form (TRF) [ 21 , 22 ] will be com pleted by the teacher of the child. The TRF as sesses prob lem be hav ior (at school). In the CBCL, YSR and TRF ques tion naires, a higher score in di cates more emo tional and/ or be hav ioral prob lems. In these ques tion naires, we are ex am in ing the in ter nal iz ing and ex ter nal -iz ing to tal scales as well as the sub scales. The items re lated to (so cial) com mu ni ca tion will also be ex am ined in di vid u ally.
Self -worth im age: The Dutch ver sion of the Self Per cep tion Pro file Child/ Ado les cent [ 23 , 24 ] is com pleted by chil dren of 8 years and older.
Qual ity of life: The Dutch ver sion of the Child Health Ques tion naire (CHQ) [ 25 , 26 ] is used to as sess qual ity of life of the child.
Pre dic tors Parental psy chopathol ogy: The Adult Self Re port (ASR) [ 27 ] will be used to screen for parental emo tional and be hav ioral prob lems.
Parental par ent ing styles: The Egna Min nen Be träf fande Upp fos tran -Par ent ver sion ques tion naire (EMBU -P) [ 28 ] will be used to as sess the par ent ing style par ents use for the child.
De mo graphic vari ables: De mo graphic vari ables such as age, gen der, so cio -economic sta tus and bilin gual ism will be as sessed dur ing in take, us ing a semi -structured in ter view used in care as usual.
Health care con sump tion: Pre vi ous phys i cal and men tal health consump tion will be as sessed dur ing in take, us ing a stan dard ized question naire [ 29 ].
Treat ment in tegrity To check treat ment in tegrity, all treat ment ses sions will be au dio/ video -recorded (if con sent of par ents and child). Record ings will be saved safely and coded. A ran dom 20% of all ses sions will be rated by two in de pen dent raters. To avoid pro to col drift ing, monthly su per vision will be given by a cer ti fied psy chol o gist. The psy chol o gists provid ing treat ment are trained in the pro to col.
Sam ple size cal cu la tion A pre lim i nary ran dom ized clin i cal trial into be hav ioral treat ment of se lec tive mutism showed an ef fect size of Co hen's d = 0.58 on speak ing in so cial sit u a tions [ 15 ]. To de ter mine ef fects on our primary out come (i.e. the pre ver sus post treat ment change (T1 -T3) on se lec tive mutism symp toms), based on an ef fect size of d = 0.58 (medium ef fect size), an al pha of 0.05 (two -tailed) and a power of 0.8, a sam ple size of 76 (38 per group) is needed [ 30 ].

1 . Possible difference between participants and drop -outs
Clin i cal ex pe ri ence in this pa tient pop u la tion shows us that dropout dur ing treat ment is rare. To be able to com pen sate for pos si ble drop -out dur ing the study (if needed), we aim to in clude 100 par tic ipants as a safety mea sure. A drop -out analy sis on the drop -out pop u lation will be con ducted: it will be stud ied whether there are pos si ble dif fer ences in vari ables (such as: gen der, age, so cio -economic sta tus, bilin gual ism, sever ity of se lec tive mutism at base line as sess ment and parental psy chopathol ogy) be tween the pop u la tion that par tic i pates un til the end of their treat ment and the pop u la tion that drops out. In this way it can be checked whether there is a se lec tive dropout. A selec tive dropout can lead to se lec tion bias, which could be sta tis ti cally ad justed for.

2 . Intention to treat
The sta tis ti cal analy ses will be based on the in ten tion -to -treat princi ple. If ap pro pri ate, sec ondary analy ses will be con ducted us ing a per -protocol ba sis.
Pri mary and sec ondary out comes: Pre -post changes (pre treat ment and post treat ment as sess ment of the treat ment ver sus WCG group) on the pri mary out come (symp toms of se lec tive mutism) will be tested with re peated mea sures analy ses (fac to r ial re peated mea sures ANOVA). To de ter mine pre -post change in treat ment out come (respon ders vs. non -responders) paired sam ple t -tests will be ex e cuted. The ef fec tive ness on sec ondary out come mea sures: symp toms of anx iety/ de pres sion, self -image, qual ity of life and care con sump tion, will be tested with the same analy ses. With lin ear and lo gis tic re gres sion pre dic tors (e.g., gen der, age, so cio -economic sta tus, bilin gual ism, parental psy chopathol ogy and par ent ing styles) for treat ment out come can be de ter mined (p < .05). In the analy ses, the num ber of treatment ses sions will be in cluded as a co vari ate.

3 . Validation of the Selective Mutism Questionnaire
The SMQ [ 18 ] is used in this study to as sess symp toms of se lec tive mutism. This ques tion naire has not yet been val i dated for the Dutch pop u la tion. In or der to val i date the SMQ, data will be used from baseline as sess ment of se lec tive mutism symp toms in the chil dren re ferred for di ag no sis and treat ment of se lec tive mutism. In ad di tion, data will be gath ered from 240 healthy school chil dren, re cruited from el e mentary and sec ondary schools and sports clubs, from the same re gional ar eas as the chil dren with se lec tive mutism. Con trol chil dren will be ad justed to pa tients with a 3:1 ra tion. To as sess the fac tor struc ture, con fir ma tory fac tor analy sis will be ap plied. Split -half re li a bil ity and in ter nal con sis tently (Chron bach's al pha) will be de ter mined. Itemrest and in ter -item cor re la tions will be de ter mined. The re li a bil ity of the items' re sponse scale is cal cu lated us ing Rasch as sess ment scale analy sis [ 31 ]. To as sess con cur rent va lid ity, re sults on the SMQ will be tested against out comes on the se lec tive mutism part of the Anx iety Dis or der In ter view for Chil dren (ADIS -C).

. Discussion
With the new clas si fi ca tion of the DSM -5 [ 1 ], se lec tive mutism is now cat e go rized as an anx i ety dis or der. Se lec tive mutism is of ten under diag nosed which causes chil dren with se lec tive mutism to re ceive no treat ment at all, in ad e quate or un timely treat ment. In the Netherlands and Dutch speak ing re gions of Bel gium, un til now no ev i dencebased treat ment for se lec tive mutism was avail able.
In ves ti gat ing the ef fec tive ness of the treat ment pro to col for se lective mutism is nec es sary to en sure that ev i dence -based treat ment is avail able as stan dard care within men tal health in sti tu tions for children and ado les cents. The pre sent treat ment pro to col has the ad vantage, that treat ment takes place at school, where the se lec tive mutism is of ten most pre sent. The ra tio nale is that once the child is able to speak at school, this makes it eas ier to gen er al ize speak ing be hav ior to other so cial set tings. The pro to col uses a be hav ioral ther a peu tic approach, which be longs to stan dard par a digm for the treat ment of anxi ety dis or ders, es pe cially in chil dren who are too young to profit from cog ni tive el e ments in treat ment.
This study has sev eral strengths. A stan dard ized treat ment pro tocol for be hav ioral ther apy is used, with ther a pists trained in this in terven tion, with weekly su per vi sion, and treat ment in tegrity be ing checked. Treat ment takes place in the school set ting, en hanc ing gener al iz abil ity of tech niques learned. Im por tantly, we aim to iden tify pre dic tors for treat ment suc cess, to get in sight into which chil dren ben e fit more or less form this treat ment. This study aims to con tribute to knowl edge, by gain ing in sight into why some chil dren im prove more than oth ers by in ves ti gat ing if de mo graphic vari ables (such as bilin gual ism), parental psy chopathol ogy, par ent ing styles and pre vious health care con sump tion mod er ate treat ment out comes.

1 . Limitations
Chil dren are treated by trained ther a pists of a sin gle cen ter. Because there are rel a tively few chil dren with se lec tive mutism, the intended pop u la tion might be het ero ge neous in terms of de mo graph ics. This cre ates both an in ter est ing op por tu nity for gen er al iz abil ity as well as a chal lenge in to what ex tent con clu sions can be made from a het ero ge neous study pop u la tion. We aim to con trol for this by care -fully as sess ing de mo graphic vari ables to gain in for ma tion about the char ac ter is tics of the pop u la tion.
Treat ing chil dren at school may pro vide lo gis tic chal lenges (such as ther a pist's travel dis tance) but also cre ates pos i tive op por tu ni ties in hav ing a di rect way of com mu ni ca tion with the teacher about the exer cises in the treat ment and to be able to strengthen the mo ti va tion for par tic i pat ing in the treat ment for busy teach ers with full classrooms.

2 . Implications for clinical practice
Since all re ferred chil dren be tween three and eigh teen years old will be screened, se lec tive mutism can be de tected early in de vel opment. Early treat ment for se lec tive mutism cre ates the po ten tial to pre vent fu ture prob lems (sec ondary pre ven tion), such as ed u ca tional prob lems and the de vel op ment of anx i ety in adult hood. This pro ject aims to iden tify an ef fec tive treat ment pro to col and un ravel some of the un der ly ing mech a nisms of treat ment suc cess in chil dren with selec tive mutism. Sub se quently this will lead to im prove ment of care. We ex pect that the re sults of the pre sented study will be im me di ately rel e vant to clin i cal prac tice and that there is po ten tial for large -scale roll -out across the Nether lands.

Funding
This re search pro ject is funded by Fonds Sticht ing Gezond hei dszorg Spaar neland (Grant ID: 2017284) . The fund ing source had no role in the de sign of the study, and will not have any role in its ex e cution, analy sis, in ter pre ta tion of the data, or de ci sion to sub mit re sults.

Availability of data and materials
Not ap plic a ble. This pa per pre sents the study pro to col and does not con tain any data or re sults.

Author's contributions
All au thors crit i cally re viewed the man u script for in tel lec tual content. All au thors have read and ap proved the man u script. CRP drafted the ini tial man u script and sub mit ted the man u script for pub li ca tion. CRP, JE and MG were re spon si ble for study con cept and de sign, which was su per vised by RL and EU. CRP, JE, RL and EU were respon si ble for fund ing. MdJ pro vided in tel lec tual in put and feed back on the study de sign. MG de vel oped the orig i nal treat ment pro to col and pro vided in tel lec tual in put for fund ing. KK was in volved as sta tisti cian and pro vided in for ma tion for the sam ple size cal cu la tion and the sta tis ti cal analy ses.

Trial registration
Dutch Trial Reg istry: NTR7534

Ethics approval and consent to participate
The Med ical Ethics Com mit tee of the Aca d e mic Med ical Cen ter approved this trial. This study will be con ducted ac cord ing to the Helsinki De c la ra tion and its later amend ments or com pa ra ble eth i cal stan dards. In formed writ ten con sents will be ob tained from the parents or guardians of the par tic i pat ing chil dren. This ar ti cle does not con tain any stud ies with an i mals per formed by any of the au thors.

Consent for publication
Not ap plic a ble. This pa per pre sents the study pro to col and does not con tain any data or re sults.

Declaration of competing interest
The au thors de clare that they have no con flict of in ter est.