Well, it’s that time of year— fall is upon us. At UT, that means new students (and likely their parents) driving down San Antonio (a one way street) in the wrong direction. So far, I’ve spotted one driver doing this and it’s sure to increase as students move in and as classes start in a couple of weeks.
For those of you who work in elementary schools or preschool settings, screening may be part of the fall routine. I remember I worked for a few years in Head Start and we would screen children every fall. My first year (1984) I remember we just made up a screener. The SLPs and I got together and came up with a form and a few questions that we would talk to kids about while we observed their speech and language. Later on as there was less money to spend on things like screening, we relied on teacher referral. I did notice that often teachers made referrals if children had articulation errors but not much else. And many of the 3 year olds were being referred for typical developmental errors. So, we went to a modified screening procedure where we asked teachers to complete a form that focused their attention on aspects of speech and language that might be problematic for given ages. We would sit together to then determine if the child had more typical developmental errors or if a referral was really warranted. It also helped us to pick up on children who might have language-based impairment as well.
Thirty years later, SLPs are still screening, still making up tests, and still taking referrals. And there’s nothing wrong with that but we want to make sure that the tools we use are accurate and that they will result in appropriate referral that will make a difference in the lives of the kids we are supposed to serve.
Probably, it’s not a good idea to just make up a screener without some level of validation of whether or not it works. What does “working” mean? Well, it should help you identify children who are likely to have speech-language impairment and not identify children who are not likely to have speech/language impairment. So, this means sensitivity and specificity need to be pretty good. For a screener it’s probably okay if it over-identifies children as possibly having the impairment (that is sensitivity is high but specificity is a little lower) because with follow up testing you can rule out the disorder.
I have also seen programs make up tests from bits and pieces of published assessment instruments. Publishers frown on this practice big time. First it is a violation of copyright. Second, you still don’t know if the bits and pieces selected are valid and reliable as a screener. Don’t do this.
There are a few published screeners. Unfortunately, many of them have high specificity but poor sensitivity. This means that you are likely to miss children with impairment. Those you do identify with impairment probably will have it, but it’s a lot of time spent if you’re missing a lot as well. So, read those manuals!!
We have a new paper published last week in AJSLP (Lugo Neris, Peña, Bedore & Gillam, 2015) on our screener– the BESOS. It’s not yet published but we’re hoping to do that soon. The screener has 90% sensitivity and 90% specificity for bilinguals. And it also works well with monolinguals (though we have to do the stats on that). In this paper, we were interested in whether our screener, given before kindergarten, would predict who was likely to have language impairment in first grade. Knowing who has speech-language impairment now helps us know who to test. Knowing who is likely to continue to present with language impairment a couple of years out, helps us to think about programming, coordination with schools (for kinder transition) and to prioritize assessment. We found that 95.2% and predictive specificity of 71.4%.
While the BESOS is not yet ready for clinical use (but feel free to inquire from time to time at AR-Clinical Publications), we do have the ITALK which is part of the BESA that can be utilized as a screener. It is a series of questions that we ask parents and teachers about children’s articulation, comprehension, sentence length, vocabulary, and grammar in each language. It is rated on a 1 to 5 scale and averaged. If the higher score (between Spanish and English) is below 4.2 we have found that children are likely to present with speech-language impairment. We like to include both teachers and parents because for bilinguals, parents are likely to know more about their Spanish and teachers are likely to know more about their English.
Anyway, have a good fall. Get ready to start working with all those kiddos!