Abstract
Purpose:
Most research on language acquisition and impairments is neutral to work setting; however, work settings (e.g., schools, health care) are expected to differ in alignment with overlaid workplace models (e.g., education, medical). These differences may affect clinical service provision for individuals with specific language impairment (SLI). This article evaluates potential effects of work setting on top-down advocacy initiatives and clinical service provision for children with symptoms of SLI.
Method:
Speech-language pathologists serving pediatric populations in health care–based (n = 140) and school-based (n = 423) work settings completed a three-part survey: (a) participant demographics, (b) report of case/workload and practice patterns, and (c) clinical vignettes and eligibility belief. Data analysis included descriptives and chi-square tests.
Results:
The work setting groups reported differences in eligibility terminology, eligibility criteria, and practice patterns from the point of referral through discharge. The reported differences aligned with overlaid workplace models. As compared to the school-based group, health care–based participants reported fewer eligibility restrictions in the workplace, agreed more often with a belief in less restrictive eligibility criteria, and reported more sensitive clinical decisions when operating under neutral workplace circumstances. Despite these findings, health care–based participants reported a smaller proportion of individuals with language impairment only on their caseload.
Discussion:
Work setting variations influence the underidentification of individuals with SLI for speech-language pathology services. Differences in responses by workplace indicate the need for unique and targeted advocacy efforts. Shifting diagnostic terminology and criteria will be insufficient in closing the gap unless advocacy efforts also address speech-language pathologists' workplace realities.
Children with specific language impairment (SLI) are likely to be overlooked for clinical services, a phenomenon well documented in the school setting (e.g., Johnson et al., 1999; Tomblin et al., 1997; Zhang & Tomblin, 2000). The professional and scientific literature intermittently links this underidentification to diagnostic labels, with a recent call to change the label along with a shift in diagnostic criteria (i.e., developmental language disorder [DLD]; Bishop et al., 2017). The hope is that a new diagnostic label and criteria, with enthusiastic promotion, will increase identification rates for individuals with language impairments arising in childhood with no known cause. The resulting SLI/DLD debate (Green, 2020; Volkers, 2018a, 2018b) is important for research replication and validity as well as advancing knowledge of clinical conditions. While the broad terminology and diagnostic criteria of DLD may aid advocacy efforts with the general public (Bishop et al., 2017; McGregor et al., 2020), the narrow diagnostic criteria of SLI have allowed for scientific replicability (Rice, 2020). The precise specification of SLI avoids confounds with a broader definition.
Thus far, the primary voices in the SLI/DLD debate have been researchers. When considering such a research change, which also has clinical implications, all stakeholders need a voice. Speaking for university educators, Victorino and Magaldi (2020) explained that the shifting terminology interrupts a critical evidence-based practice step—the literature search—and for applying research to practice. Speaking for school-based speech-language pathologists (SLPs), Murza and Ehren (2020) explained that the terminology debate may not be worthwhile because larger issues exist for school-based advocacy work, such as reducing caseload sizes. We have yet to hear from health care–based SLPs in the United States, who also serve individuals with SLI. Furthermore, the SLI/DLD debate and calls to action have focused on child-centered factors (e.g., clinical symptoms; Bishop et al., 2017) and SLP-centered factors (e.g., recommended assessment practices; Bishop et al., 2016) that affect service provision, with less attention on workplace realities, which may constrain the autonomy of the SLP. For example, an SLP may recognize an individual's need for services but may also be thwarted by workplace limitations. In this report, we center SLPs within the complex systems where they work (i.e., public education, health care) to consider how work setting affects eligibility terminology, eligibility criteria, beliefs on eligibility restrictiveness, decision-making behaviors, clinical practice patterns, and caseload compositions.
SLPs as Agents for Change in Complex Workplace Systems
SLPs operate within complex systems, which are systems with many discrete yet interacting components where no clear boundaries exist (Braithwaite et al., 2018). Complex systems exist everywhere, ranging in size from quantum physics to galaxies in the universe. The United States' public education and health care systems are also complex, with SLPs placed in the middle in terms of capacity for widespread change. Above them in the systems are policy-makers, regulatory bodies, administrators, and researchers; below them in the systems are the populations they serve such as individuals with SLI.
When attempting to change complex systems, top-down initiatives are common. Within the context of SLI underidentification, the proposed shift in diagnostic labels and criteria to DLD is a top-down approach. These top-down approaches aim to transmit information from “knowledge producers” (e.g., policy-makers, researchers) through “knowledge users” (e.g., SLPs, university educators) to the target population who will benefit most from systemic change (e.g., individuals with SLI and their families; Braithwaite et al., 2018). Top-down approaches have limitations in their effectiveness within complex systems, however, since they assume linearity and predictability, which the concept of a research-to-practice “pipeline” suggests. Complex systems are neither linear nor predictable, but, rather, they are multidimensional, multidirectional, nonlinear, and adaptive (Braithwaite et al., 2018). For example, the multidimensional nature of the complex public education and health care systems means that SLPs must balance policies, demands, and guidelines from multiple top-down entities when engaging in clinical decision-making (e.g., evidence-based guidelines, federal and state eligibility criteria, health insurance definitions of medical necessity).
The complex nature of these systems leaves various stakeholders with finite influence in systemic change. As Parag and Janda (2014) reviewed, in a top-down approach, knowledge producers have the capacity for widespread change, but they lack the agency to ensure that the top-down changes are implemented. For example, researchers have the capacity to create guidelines for the profession of speech-language pathology, such as the proposed shift to DLD and calls for evidence-based practice. They do not have the agency to guarantee that practicing SLPs will choose or have the necessary resources and supports to follow those guidelines, which weakens the effectiveness of top-down approaches. In a bottom-up approach, the target population, who will benefit most from change, has the agency to make certain demands or requests, but they lack the capacity to ensure the demands or requests are met. For example, individuals with SLI and their families have the agency to request speech-language pathology services, but they cannot guarantee that the SLP will place them on their caseload.
A third and overlooked approach to systemic change exists—middle-out (Parag & Janda, 2014). In this approach, the stakeholder groups in the middle of complex systems are the agents for advocacy and systemic change. Because of their position in the complex system, they influence the success of top-down and bottom-up approaches since they act as a bidirectional conduit connecting knowledge producers with the target population. This middle stakeholder group has the potential to initiate and accelerate change since they can implement the top-down initiatives (e.g., policy, research), but they also have the potential to delay, stall, or resist change depending on their circumstances. For example, the middle stakeholder group may agree with a top-down initiative but may not implement it because of limited support and resources in the workplace.
Whether middle stakeholder groups accelerate change, stall the possibility for change, or actively resist change depends on three factors—conditions, concerns, and capacity. Per Parag and Janda (2014), conditions refer to the workplace realities that may hinder or facilitate the needed change, concern refers to one's awareness of the issue needing change, and capacity refers to the knowledge and skills needed for action. Here, practicing school- and health care–based SLPs represent the stakeholder groups of interest since they are in between the knowledge producers (e.g., policy-makers, researchers) and the target population benefiting from systemic change—individuals with SLI and their families. Without considering SLPs and their position within the complex systems, top-down approaches may have limited effectiveness. Because SLPs work in two different complex systems (i.e., public education, health care), we consider in this study how workplace conditions may vary across work settings, how conditions may interact with SLPs' concerns and capacity, and how all three factors may influence their practice patterns and the composition of their caseloads.
American Speech-Language-Hearing Association Membership by Work Setting
Exploring work setting effects is timely because not all SLPs serving individuals with SLI work in the schools, and workforce trends are changing in the American Speech-Language-Hearing Association (ASHA). As of 2020, 42% of ASHA members reported work in health care facilities (e.g., hospitals, private physician's office, SLP's office; ASHA, 2021). In a separate non–mutually exclusive question, 10% of ASHA members reported working full time in a private practice, with another 12% working part time. These health care and private practice numbers combine SLPs working with pediatric and adult populations; however, in a survey of health care–based SLPs, 40% reported serving pediatrics, with language and literacy skills accounting for the largest portion of service time (ASHA, 2019). ASHA workforce trends show that the number of SLPs working in health care has increased by 5% from 2007 to 2019 while those working in school-based settings decreased by 4% (ASHA, 2020). Although these trends may seem small, they indicate a potential shift of SLPs moving from school-based to health care–based employment positions. This shift will likely continue as those entering the workforce—undergraduate and graduate speech-language pathology students—report a preference for health care–based work settings (Leonard et al., 2016). In addition, whether the novel 2019 coronavirus (COVID-19) will affect these workforce trends is unknown. Depending on why the shifting trends exist, COVID-19 effects on workplace conditions may accelerate the changing ASHA workforce. Taken together, health care–based SLPs represent a notable and growing portion of the ASHA membership base, and the health care–based research literature must expand accordingly.
Potential Work Setting Effects in Clinical Service Provision
Workplace Conditions: Eligibility Terminology, Eligibility Considerations, and Funding Mechanisms
The United States' public education and health care systems have unique workplace realities, which create different conditions for SLPs' to operate within when making clinical decisions. Both work settings start with broad but different eligibility definitions, and the work settings' eligibility processes differ too. In the schools, students become eligible for language-based speech-language pathology services under the speech or language impairment or specific learning disability categories of the Individuals with Disabilities Education Act (IDEA, 2004). The former captures all communication disorders, including language impairment, which IDEA does not define, and has no exclusionary criteria. The latter is a language disorder that may manifest across academic skills. Exclusionary criteria specify that learning problems are not because of sensorimotor deficits (e.g., hearing loss), intellectual disability, emotional disturbance, or socioeconomic disadvantage. Some states, such as Kansas (Kansas Administrative Regulations, 2008), use the federal definitions, whereas others refine the IDEA definitions with varying restrictiveness. For example, the Office of Special Education defines language impairment in Missouri (Missouri State Board of Education Special Education Regulations, 2020) as scoring (a) 1.75 SD s or more below the mean, (b) on two or more composite standardized language assessments, (c) affecting both receptive and expressive language function, (d) alongside adversely affected educational performance, (e) and as based on lack of response to evidence-based general education curriculum interventions. In the United States, school-based SLPs provide free services based on the legal rights of students with disabilities to a free and appropriate education (IDEA, 2004). School-based SLPs must serve all students deemed eligible based on documentation of academic impact. What constitutes academic impact, however, varies in interpretation and may be difficult to see when the impact is covert and protracted, such as lower rates of graduation or postsecondary enrollment.
In contrast, health care–based SLPs in the United States operate on a fee-for-service model, do not need to demonstrate academic impact, and have the option to turn prospective clients away or place on a waitlist once their caseload is full. Under this model, after the SLP provides services, either the patient pays out of pocket or the SLP submits a claim to the patient's insurance company for reimbursement. To submit a claim for insurance, the SLP assigns the relevant codes from the International Statistical Classification of Diseases and Related Health Problems, 10th Edition (ICD-10; World Health Organization [WHO], 2016). The relevant codes for individuals with SLI are expressive language disorder, receptive language disorder, or mixed receptive–expressive language disorder. All three categories define the language disorder as language ability below expectations benchmarked to mental age. Exclusionary criteria are acquired aphasia with epilepsy, autism spectrum disorder, selective mutism, dysphasia and aphasia, elective mutism, language delay because of deafness, and mental retardation. Beginning in 2022, the label will change to DLD with subcategories of mainly expressive, receptive and expressive, and mainly pragmatic (WHO, 2019). The definition further specifies that the disorder arises in childhood and benchmarks language ability to age and intellectual functioning. Exclusionary criteria are acquired aphasia with epilepsy, autism spectrum disorder, selective mutism, dysphasia, disease of the nervous system, and deafness.
Like the federal IDEA categories, the ICD-10 categories are broad and span beyond SLI and DLD. They also do not specify a performance threshold for eligibility (e.g., standard score cutoff). Unlike IDEA, individual states do not refine the ICD-10 definitions or impose more restrictive criteria. Rather, third-party payors, including insurance companies, determine eligibility in the health care work setting. Insurance companies base eligibility on whether the service is medically necessary and whether the individual's insurance plan includes benefits to cover the provided service and diagnosis. Both private health insurance companies (i.e., a for-profit enterprise) and federal/state-funded health insurance programs (e.g., Medicare, Medicaid) are motivated to protect financial margins when deciding what services and diagnoses to cover. Seeking third-party financial coverage for speech-language pathology services can generate a cycle of denials and appeals as the burden of proving medical necessity falls to the SLP and the family. If the insurance company approves services, they will reimburse the SLP, but, oftentimes, the patient still contributes to the overall cost through a copayment, co-insurance, or deductible. The SLP must continue to document that medical necessity is present. If the insurance company denies services, the patient is then responsible for the full cost, but the eligibility criteria for medical necessity no longer exist. The patient can continue using the services and pay the SLP out of pocket even with an insurance denial. In fact, some allied health care practitioners do not accept insurance at all and run their entire practice as private or cash pay. Under this business model, the SLP has full autonomy in identifying individuals as eligible for service.
Based on eligibility and funding mechanisms, the conditions for identifying individuals with SLI as eligible for speech-language pathology differ between the public education and health care systems. When considering these complex systems, SLPs in both work settings may view an individual with SLI as warranting their services based on their awareness (i.e., concern) and knowledge (i.e., capacity) that SLI is present and requires specialized language intervention; however, the SLP must also consider the top-down eligibility criteria (i.e., conditions) within their work setting. Both school- and health care–based work settings operate with eligibility labels that would allow individuals with SLI to receive speech-language pathology services, but eligibility criteria vary across (i.e., IDEA vs. ICD-10) and within (e.g., state-level differences in IDEA) these work settings. For example, an SLP may determine that an individual warrants speech-language pathology services based on their knowledge and skills of diagnostic criteria, but the individual's deficits may not be severe enough for specific workplace eligibility criteria. The IDEA requirements of a given state may stipulate a lower performance threshold on a norm-referenced assessment, or the individual's insurance plan may exclude the ICD codes relevant for SLI. Overall though, SLPs working in the health care–based setting experience fewer eligibility constraints (i.e., better conditions); medical necessity and insurance plan coverage do not depend on performance thresholds, and patients may continue services at their own cost despite a denial.
Awareness of the SLI Profile and Eligibility Beliefs
Although workplace conditions may influence eligibility based on policies and guidelines, SLPs may still have concern—referred to as awareness or beliefs moving forward—for individuals with SLI. That is, despite potentially restrictive workplace eligibility criteria, SLPs may be aware that individuals with SLI warrant their services even if the SLP cannot deem them eligible based on workplace conditions. In this case, when asked to disregard workplace conditions, SLPs would still report that they believe or agree that individuals meeting diagnostic criteria for SLI should be eligible for services. However, the conditions of SLPs' workplaces regarding eligibility considerations may affect their beliefs around eligibility restrictiveness. Personal and professional experiences shape individuals' beliefs (Kritikos, 2003), and SLPs identify beliefs as a core aspect of therapy effectiveness (Kamhi, 1994). For example, Kamhi (1995) found that SLPs believe interpersonal factors (e.g., rapport) supersede technical factors (e.g., diagnosis) in therapy effectiveness. Previous experience in a work setting with few eligibility restrictions may foster that same belief; in contrast, previous experience in a work setting with many eligibility restrictions may foster a belief that restrictive eligibility practices are necessary. These previous experiences and beliefs then affect the SLPs' practice patterns, including how much clinical significance to place on various aspects of a child's clinical profile, the value they place on certain aspects of the assessment process, and their interpretation of assessments (Kritikos, 2003). Taken together, given their previous professional experiences, health care–based SLPs may believe that fewer factors should restrict eligibility. In turn, such a belief may also lead to less restrictive assessment practices.
Clinical Decision-Making
In parallel to awareness and beliefs, SLPs may have the capacity—referred to as clinical decision-making moving forward—for identifying individuals with SLI despite potentially restrictive workplace eligibility criteria. In this case, when asked to disregard workplace conditions, SLPs would identify individuals with SLI for speech-language pathology services at higher rates than occur in actual practice. In a previous report of the same questionnaire study presented here, Selin et al. (2019) designed vignettes portraying profiles of children with SLI. The vignettes removed the workplace conditions that confound our understanding of the knowledge and skills that SLPs bring to clinical decision-making. First, the vignettes assumed a referral source brought the individual to the attention of the SLP, removing the confound of low referral rates potentially driving underidentification. Second, the vignettes provided assessment data, which removed the confound of SLPs using insensitive language measures. Third, the vignettes instructed the SLPs to ignore workplace conditions (e.g., caseload and time constraints, mandated policies) when making clinical decisions. Free from these confounds and workplace conditions, SLPs identified individuals with SLI as needing speech-language pathology services at higher rates than are occurring in actual practice, but the rates varied across profiles. Omnibus language standard score severity drove decision-making, which other researchers have found as well (Fulcher-Rood et al., 2018; Records & Tomblin, 1994). When the vignettes presented with standard scores lower than 78, SLPs identified the individual as needing services 90% of the time or more. However, when vignettes presented with borderline standard scores (i.e., 83–85), the rate dropped to 28%–50% depending on the profile. In the previous report (Selin et al., 2019), we did not consider whether these rates differed across work settings, which we do here. On the one hand, no differences may emerge because the SLPs made decisions free from workplace conditions. On the other hand, if work setting affects SLPs' awareness and beliefs around eligibility restrictiveness, then those may trickle down to also affect clinical decision-making, even when free from workplace conditions.
Practice Patterns
Thus far, the influence of workplace on conditions, awareness and eligibility beliefs, and clinical decision-making have focused on eligibility; however, eligibility is not the only relevant clinical decision SLPs make regarding access to services for individuals with SLI. Prior to eligibility, individuals with SLI must come to the attention of the SLP through referral sources. The SLP then applies their practice patterns in assessment, including which assessments to use. Once deemed eligible, the SLP determines service delivery models (e.g., individual or group therapy sessions), interprofessional practice, and treatment planning. Furthermore, an SLP decides when an individual is ready for discharge from services, which likely ties back into eligibility considerations.
There is no research base establishing whether these practice patterns vary across work settings. In theory, SLPs in both school- and health care–based work settings have equal access to ASHA's (2004) preferred practice patterns and to evidence-based guidelines in the research literature. Yet, workplace conditions may constrain the feasibility of selecting certain practice patterns while the SLPs' eligibility beliefs and clinical decision-making may influence whether and how they implement certain practice patterns. For example, SLPs often use language assessments with insufficient psychometric properties for diagnosis (Betz et al., 2013), but the reasons for this practice are unclear. SLPs may select assessments with insufficient psychometric properties because of (a) limited awareness that psychometric properties are important to consider, (b) limited knowledge of how to interpret psychometric properties when making clinical decisions, or (c) constraining workplace conditions may influence assessment selection based on resource availability (e.g., funding for assessments, sharing assessments with multiple SLPs) and eligibility criteria (e.g., tests must report standard scores for receptive and expressive language or syntax and semantics). All three potential reasons are not mutually exclusive and may vary across work settings. Knowing how SLPs across work settings engage in various practice patterns and the underlying reasons for those practice patterns will inform researchers as to workplace-specific deviations from evidence-based practice for individuals with SLI.
Caseload Compositions
We have presented multiple factors that may affect whether an individual with SLI receives speech-language pathology services and what clinical service provision looks like across work settings. Do these factors integrate to affect the SLP's caseload composition? Previous research considers how caseload size affects clinical practices (Brandel & Frome Loeb, 2011; Dowden et al., 2006; Hutchins et al., 2009), but, in order to close the underidentification gap, it is critical to also understand how workplace factors drive who is—and who is not—on clinical caseloads and how that varies across work settings. The vulnerabilities to underidentification occur across service provision from the point of referral to discharge. The SLP or a referral source must first identify that an individual with SLI warrants an SLP's attention. Then, the SLP must agree that an assessment is warranted based on their beliefs, their workplace's eligibility criteria, and their caseload size. Then, during annual and triennial reviews or evaluations, the SLP must determine whether the individual with SLI remains eligible for services. In addition, SLI is a persistent disorder arising early in childhood and continuing into adulthood (Johnson et al., 1999; Rice, 2012; Rice & Hoffman, 2015); however, it is unknown whether access to SLPs varies across age ranges for prevention, early identification, and treatment—infant and toddler, preschool, elementary, middle school, and high school. These vulnerabilities likely differ across work settings, which this study explores.
Aims
This study aimed to evaluate work setting effects by comparing SLPs' reports across the two complex systems of health care and public education in a national survey. We do not consider whether the work settings' reports align with evidence-based practice. Our focus on documenting potential work setting differences is agnostic to evidence-based practice, which can lead to and clarify future research questions of whether practice patterns and decision-making align with the evidence both across and within work settings. Our hypotheses were that the work setting groups would report different conditions based on their underlying diagnostic models and funding mechanisms. We also expected different eligibility beliefs given a link between previous professional experiences and clinical beliefs. For clinical decision-making, a work setting effect may or may not emerge based on the reviewed theory. Finally, we hypothesized that variations across work setting for conditions, eligibility, and clinical decision-making would associate with reported differences in practice patterns and caseload composition. In this study, we address the following research questions.
How do reports of eligibility terminology and criteria (workplace conditions) vary across work settings?
How do beliefs of eligibility restrictiveness vary across work settings?
How does clinical decision-making vary across work settings?
How do reported practice patterns vary across work settings?
How do caseload compositions vary across work settings?
Method
Ethics
The University of Kansas Institutional Review Board approved this study. Participants read an information statement and consent form on the first page of the survey. The survey collected no identifiable personal information, and participants received no incentives for completion. The participants consented to the study by selecting “I agree to participate” at the end of the information statement.
Survey Development, Design, and Administration
The web-based questionnaire was descriptive and exploratory, including a wide scope of 97 questions. The survey included 17 pages with a range of two to 10 questions per page, depending on the participants' individual response pattern. Adaptive questioning was used to reduce the number of items per participant. For example, participants who reported working in health care–based settings did not see questions relevant only to the school-based participants. Participants could complete the survey across multiple sessions, but they could not return to previously answered questions so they could not change their responses based on later questions.
Item generation occurred following a thorough review of the literature and development of a theoretical framework. As outlined in Selin et al. (2019), the overarching theoretical framework includes three constructs that may influence SLPs' clinical decision-making: workplace characteristics (including work setting), child characteristics (i.e., clinical symptoms), and practitioner characteristics (e.g., cumulative years of experience). Each item focused on one construct of interest with a neutral question stem and with as few words as possible. Responses for each item were closed. Answer options were binary (e.g., yes, no), nominal (e.g., school, clinic, or hospital), or ordinal (e.g., none, 0%–33%, 34%–66%, 67%–100%). Given that the order of provided responses may influence selection, nominal response choices were randomized for all items across participants. The order of items and question blocks were not randomized, however.
Prior to disseminating the survey, 11 certified SLPs reviewed a pilot version of the survey and provided feedback about usability, time to complete, content, and clarity. Based on the feedback provided, we removed redundant items/responses and edited confusing items/responses. These SLPs worked in school-based settings, as this survey was originally designed to investigate school-based issues. However, we decided to expand the scope and include SLPs working in health care–based work settings. The first author has extensive clinical experience in a variety of health care–based work settings (e.g., Early Intervention, hospital outpatient clinics, private practices), and the preliminary findings from the pilot survey did not align with her workplace experiences, thereby providing a rationale to expand the scope. Few content-related changes were made to the survey. The primary work setting question was added as well as questions and response choices related to health insurance.
Qualtrics (2005) hosted the web-based survey and captured participants' responses. Upon closing the survey, the data were transferred to SPSS (IBM Corp., 2020) for data analysis. The three section blocks were (a) participant demographics and characteristics, (b) reported work/caseload and practice patterns, and (c) clinical decision-making vignettes and eligibility opinion. Although Dillman et al. (2014) recommended that demographic information occur last in surveys, the authors placed them at the beginning for two reasons. First, Burns et al. (2008) reported that including demographic questions in the beginning may ease participants into the survey as they are simpler questions. Second, including demographic questions in the beginning ensured that we could judge the representativeness of the sample and examine differences between participants who did or did not complete the survey. For Section 2, questions targeted the following categories of practice patterns: referral, eligibility determination, assessment practices, service delivery, collaboration, treatment practices, and considerations for discharge and monitoring.
Part 3 included six vignettes that portrayed school-age children with SLI (see Appendix). Instructions indicated that participants should disregard workplace circumstances (i.e., caseload and time constraints, mandated policies) when making clinical decisions. After reading each vignette, participants made a series of clinical decisions. Here, we investigated the effect of work setting on the first clinical decision: recommendation for (continued) speech-language pathology services. The operational definition of SLI for the vignettes was performing one or more standard deviations below the age mean on an omnibus language assessment, adequate performance on nonverbal intelligence, and the presence of finiteness marking impairment. These dimensions of affectedness were then varied across the vignettes to determine whether SLPs used these clinical symptoms in their decision-making.
Although a 1-SD cutoff score is the source of continued scientific debate, it captures the broadest definition of SLI in the research literature, which allows for investigating who is at the highest risk of underidentification. Any cutoff score selection is arbitrary (e.g., Spaulding et al., 2006) and must be interpreted in consideration of the standard error of measurement (i.e., a measure of how much observed test scores are spread around a “true” score). We prioritized the inclusivity that the broadest definition provides. In addition, the vignettes included data from language sample analysis (e.g., clinical symptoms of finiteness marking impairment) and clinical observations (e.g., descriptions of functional impacts such as difficulty learning new material from texts), which support the validity of the SLI portrayed in the vignettes. We direct readers to the previous report (Selin et al., 2019) for more detailed information on the selection of these diagnostic criteria. In summary, the −1-SD diagnostic threshold is within the standard error of measurement of the −1.14-SD threshold used in the EpiSLI diagnostic system (Tomblin et al., 1996), aligns with SLPs' clinical judgments (Records & Tomblin, 1994), differentiates affected from unaffected children across a variety of language measures over time (e.g., Rice & Hoffman, 2015), and coincides with the presence of secondary symptoms such as social consequences (e.g., Gertner et al., 1994). Based on these findings in the research literature, all of the children portrayed in the vignettes meet our operational definition of SLI and are at risk for persistent language impairment and a host of negative long-term outcomes. However, how SLPs make clinical decisions when presented with these same children under workplace-neutral conditions was unknown prior to our 2019 report.
As a summary of Selin et al. (2019), the survey respondents recommended the vignettes with omnibus and vocabulary standard scores less than 78 significantly more often than the vignettes with borderline standard scores (83–85). Practitioner characteristics largely did not affect recommendation rates (i.e., cumulative years of experience, finiteness marking knowledge, verb tense knowledge, grammar training); however, their eligibility belief for SLI did, which is a variable of interest in the current study. The survey respondents who selected “yes” when asked whether “all students with a language impairment (i.e., 1 standard deviation below the mean) [should] be regarded as eligible for speech-language pathology services” were more likely to recommend the vignettes for speech-language pathology services, regardless of the omnibus language or vocabulary standard score. We did not previously look at how workplace characteristics may affect this clinical decision-making, which we report on here.
Recruitment
Following precedence in our literature, we used purposive sampling, which allows for targeted recruitment based on meeting specific criteria (Burns et al., 2008). Here, the specific criterion was occupation status as a pediatric SLP in the United States at the time of survey completion. Participating organizations for recruitment were the ASHA Significant Interest Group (SIG) 1: Language Learning and Education, ASHA SIG 16 School-Based Issues, and state speech-language-hearing association organizations. Of the 51 state organizations, 30 agreed to distribute the survey to their members via e-mail, website posting, or their social media profiles. The participating state organizations represented a variety of regions across the country. Participants did not report residency to protect privacy.
Data Analysis
We used chi-square tests with a between-groups variable of work setting. Because of the large number of analyses, we a priori selected the p value of < .01 and considered effect sizes, relative risk (RR), and confidence intervals (CIs). Chi-square tests reveal whether an association exists between work setting and the participants' survey responses. Cramer's V effect sizes were interpreted to be large when greater than 0.5, medium when between 0.3 and 0.5, and small when between 0.1 and 0.3 (Cohen, 1988). We also reported RR to provide a measure for CIs. RR values represent the probability of one group endorsing a survey response as compared to the other group. To calculate RR, the percentage of health care–based participants endorsing a survey response was divided by the percentage of school-based participants endorsing the same survey response. An RR value of 1.0 represented no difference between the two groups. The health care–based work setting is the reference group for the chi-square analyses; hence, RR values greater than 1.0 represented a higher proportion of health care–based participants endorsing a survey response, whereas RR values less than 1.0 represented a higher proportion of school-based participants. CIs at the 95% level are provided for the RR values as well. If the CI included the value of 1.0, the difference between the two groups was considered nonsignificant. Note that RR and CI values are only available for chi-square analyses when both variables have two levels (i.e., 2 × 2 tables).
Results
Participants
Participant demographics mirrored those of the ASHA membership base (ASHA, 2021), as shown in Table 1. Most participants identified as female and Caucasian, and most participants reported holding ASHA certification and a master's degree. The group variable of work setting included the school-based participants (69.27% of sample) and the health care–based participants (24.86% of sample). The health care–based group included SLPs working in medical (e.g., hospital outpatient clinics; n = 45) and private practice (n = 95) settings. These groups were combined into a single health care–based group given the relatively smaller sample sizes compared to the school-based group. Although these groups are distinct, they both share the same underlying funding mechanisms (e.g., health insurance) and diagnostic models (e.g., ICD-10). While the medical work setting includes SLPs working for hospitals and outpatient clinics, the private practice work setting includes SLPs working for practices unaffiliated with a medical entity. Often, private practices are owned by an SLP or another allied health professional (e.g., occupational therapist). The private practice may include only the owner, who is the sole clinician, or it may include multiple employed or contracting SLPs. Private practice does not represent contracting companies that provide SLPs with contract positions in the schools; these SLPs would select schools as their primary work setting.
Table 1.
Participant demographics and characteristics.
| Characteristic | Health care |
Schools |
Total |
ASHA
a
|
|||
|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | % | |
| Total | 140 | 24.87 | 423 | 75.13 | 563 | 100 | — |
| Sex | |||||||
| Female | 131 | 93.57 | 400 | 97.32 | 531 | 94.32 | 95.5 |
| Male | 9 | 6.43 | 11 | 2.68 | 20 | 3.55 | 4.5 |
| Ethnicity: Hispanic | 1 | 0.71 | 14 | 3.41 | 15 | 2.66 | 6.1 |
| Race | |||||||
| American Indian/Eskimo/Aleut | 1 | 0.71 | 2 | 0.49 | 3 | 0.53 | 0.3 |
| Asian or Pacific Islander | 0 | 0.00 | 3 | 0.73 | 3 | 0.53 | 3.2 |
| Black/African American | 6 | 4.29 | 10 | 2.43 | 16 | 2.84 | 3.6 |
| White | 132 | 94.29 | 389 | 94.65 | 521 | 92.54 | 91.5 |
| Other | 2 | 1.43 | 9 | 2.19 | 11 | 1.95 | — |
| More than one race | 0 | 0.00 | 3 | 0.73 | 3 | 0.73 | 1.4 |
| Don't know | 1 | 0.71 | 2 | 0.49 | 3 | 0.53 | — |
| Certification: Yes | 133 | 95.00 | 385 | 93.67 | 518 | 92.01 | 99.8 |
| Education level | |||||||
| Bachelor's | 3 | 2.14 | 6 | 1.46 | 9 | 1.60 | — |
| Master's | 126 | 90.00 | 381 | 92.70 | 507 | 90.05 | — |
| PhD | 11 | 7.86 | 10 | 2.43 | 21 | 3.73 | — |
| SLPD | 0 | 0.00 | 5 | 1.22 | 5 | 0.89 | — |
| EdD | 0 | 0.00 | 9 | 2.19 | 9 | 1.60 | — |
| Cumulative years of experience | |||||||
| < 5 | 31 | 22.14 | 67 | 17.18 | 98 | 18.49 | — |
| 6–20 | 52 | 37.14 | 169 | 43.33 | 221 | 41.70 | — |
| 21–35 | 33 | 23.56 | 107 | 27.44 | 140 | 26.42 | — |
| 35+ | 24 | 17.14 | 47 | 12.05 | 71 | 13.40 | — |
| Grammar training | |||||||
| Extremely well | 12 | 21.05 | 49 | 18.22 | 61 | 18.71 | — |
| Very well | 18 | 31.58 | 109 | 40.52 | 127 | 38.96 | — |
| Moderately well | 20 | 35.09 | 91 | 33.83 | 111 | 34.05 | — |
| Slightly well | 6 | 10.53 | 17 | 6.32 | 23 | 7.06 | — |
| Not well | 1 | 1.74 | 3 | 1.12 | 4 | 1.23 | — |
Note. Em dashes indicate data no reported. PhD = Doctor of Philosophy; SLPD = Doctor of Speech-Language Pathology; EdD = Doctor of Education.
Membership data from 2020 (ASHA, 2021). These numbers include the full membership base (i.e., speech-language pathologists and audiologists who work with pediatrics or adults). ASHA = American Speech-Language-Hearing Association.
Of the 563 participants who started the survey, 401 completed Parts 1 and 2, and 326 completed the entire survey. Visual inspection of the data revealed that participants without ASHA certification discontinued participation more often than those with certification. Because of the attrition throughout the survey, the sample size will vary across the results reported in the text and tables. In addition, the sample size will vary because not all questions required the participant to answer before moving forward with the survey. Participants with missing responses were not included in chi-square analyses because of list-wise deletion. To maximize the overall sample size, we did not exclude participants with missing data; hence, each question will vary in the total number of participants who responded.
Conditions Differ Across Work Settings
As hypothesized, workplace conditions differed across the two work settings in ways that align with the underlying diagnostic models and funding mechanisms.
Eligibility Terminology
Eligibility terminology differed across work settings in ways that align with the overlaid eligibility frameworks, as shown in Table 2. Health care–based participants reported expressive language impairment/disorder, receptive language impairment/disorder, and mixed receptive–expressive language impairment/disorder. In contrast, school-based participants reported speech and or language impairment/disorder and specific learning disability significantly more often than health care–based participants. No difference appeared for language impairment/disorder; however, health care–based participants reported SLI significantly more often than school-based participants.
Table 2.
Terminology across work setting.
| Terminology | Health care |
School |
Χ2(1) | V | RR | 95% CI |
|||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | LL | UL | ||||
| Mixed expressive–receptive language impairment/disorder | 85 | 87.63 | 117 | 34.72 | 84.75** | 0.442 | 2.52 | 2.14 | 2.98 |
| Expressive language impairment/disorder | 75 | 77.32 | 121 | 35.91 | 52.16** | 0.347 | 2.15 | 1.80 | 2.58 |
| Receptive language impairment/disorder | 71 | 73.20 | 115 | 34.12 | 46.96** | 0.329 | 2.15 | 1.77 | 2.60 |
| Specific language impairment | 35 | 36.08 | 76 | 22.55 | 7.24* | 0.129 | 1.60 | 1.15 | 2.23 |
| Speech or language impairment/disorder | 60 | 61.86 | 259 | 76.85 | 8.70* | 0.142 | 0.80 | 0.68 | 0.95 |
| Specific learning disability | 19 | 19.59 | 119 | 35.31 | 8.59* | 0.141 | 0.55 | 0.36 | 0.85 |
| Language impairment/disorder | 41 | 42.27 | 122 | 36.20 | 1.18 | 0.052 | 1.16 | 0.89 | 1.53 |
Note. The bolded percentages represent the work setting with the higher-than-expected responses. CI = confidence interval; RR = relative risk; LL = lower limit; UL = upper limit.
p < .01.
p < .001.
Eligibility Considerations
Eligibility considerations differed across work setting in expected ways as well, as shown in Table 3. Health care–based participants reported the following eligibility considerations more often than school-based participants: presence of a medical diagnosis, performing below a particular cutoff score on one standardized assessment, health insurance, clinical judgment alone (without requirement of a standardized assessment), and mismatch between an individual's mental and chronological age. In contrast, school-based participants reported the following eligibility considerations more often than health care–based participants: impact on academic performance, previous individualized education program, performing below a particular cutoff score on two standardized assessments, and performing below grade level. No differences appeared for the considerations of informal assessment, a mismatch between an individual's cognitive and language abilities, or participants' report on required cutoff score.
Table 3.
Eligibility considerations.
| Survey question | Health care |
School |
Χ2(1) | V | RR | 95% CI |
|||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | LL | UL | ||||
| Eligibility considerations | |||||||||
| Presence of medical diagnosis | 71 | 72.45 | 152 | 44.84 | 23.19** | 0.230 | 1.62 | 1.36 | 1.91 |
| Below cutoff on 1 assessment | 70 | 71.43 | 149 | 43.95 | 22.96** | 0.229 | 1.63 | 1.37 | 1.93 |
| Health insurance | 34 | 34.69 | 3 | 0.88 | 112.13** | 0.506 | 39.20 | 12.30 | 124.92 |
| Clinical judgment alone | 34 | 34.69 | 51 | 15.04 | 18.74** | 0.207 | 2.31 | 1.59 | 3.34 |
| Mismatch mental & chronological age | 28 | 28.57 | 43 | 12.68 | 14.10** | 0.178 | 2.25 | 1.48 | 3.43 |
| Impact on academic performance | 65 | 66.33 | 327 | 96.46 | 74.73** | 0.414 | 0.69 | 0.60 | 0.79 |
| Previous IEP | 43 | 43.88 | 290 | 85.55 | 72.78** | 0.408 | 0.51 | 0.41 | 0.64 |
| Below cutoff on 2 assessments | 41 | 41.84 | 212 | 62.54 | 13.36** | 0.175 | 0.67 | 0.52 | 0.86 |
| Below grade level | 27 | 27.55 | 170 | 50.15 | 15.68** | 0.189 | 0.55 | 0.39 | 0.77 |
| Mismatch cognitive & language abilities | 43 | 43.88 | 127 | 37.46 | 1.32 | 0.055 | 1.17 | 0.90 | 1.52 |
| Informal assessment | 67 | 68.37 | 272 | 80.24 | 6.16 | 0.119 | 0.85 | 0.74 | 0.98 |
| Cutoff scores | |||||||||
| No standardized assessment required a | 6 | 4.29 | 1 | 0.24 | — | — | — | — | — |
| No particular cutoff score required | 37 | 26.43 | 80 | 18.91 | 3.61 | 0.058 | 1.40 | 1.00 | 1.96 |
| Standard score of 85 | 30 | 21.43 | 67 | 15.84 | 2.30 | 0.129 | 1.35 | 0.92 | 2.00 |
| Standard score of 81 | 7 | 5.00 | 35 | 8.27 | 1.63 | 0.201 | 0.60 | 0.27 | 1.33 |
| Standard score of 76 | 8 | 5.71 | 53 | 12.53 | 5.56 | 0.025 | 0.46 | 0.22 | 0.94 |
Note. Em dashes indicate that data were unable to be calculated. The bolded percentages represent the work setting with the higher-than-expected responses. CI = confidence interval; RR = relative risk; LL = lower limit; UL = upper limit; IEP = Individualized Education Program.
More than 25% of cells had less than five expected counts, invalidating the chi-square test.
p < .001.
When asked whether low performance on language assessments alone would qualify an individual for speech-language pathology services, health care–based participants agreed significantly more often than school-based participants, χ2(1) = 68.73, p < .001, V = 0.400 (health care: 62.24%, school: 19.17%; RR: 3.25; 95% CI [2.48, 4.24]). In parallel, when asked whether low performance must coincide with affected academic performance, health care–based participants agreed significantly less often than school-based participants, χ2(1) = 97.41, p < .001, V = 0.472 (health care: 34.34%, school: 84.37%; RR: 0.41; 95% CI [0.31, 0.54]).
Eligibility Beliefs Differ Across Work Settings
As hypothesized based on a theoretical link between previous professional experiences and beliefs, health care–based participants reported less restrictive beliefs for eligibility. They responded yes to the question, “Should all students with a language impairment (i.e., 1 SD below average) be regarded as eligible for speech-language pathology services?” significantly more often than school-based participants, χ2(1) = 24.28, p < .001, V = 0.273 (health care: 59.65%, school: 26.12%; RR: 2.28; 95% CI [1.70, 3.06]).
Clinical Decision-Making Differs Across Work Settings
Whether a work setting effect would emerge for clinical decision-making was unknown based on prior literature. The results revealed that health care–based participants were more likely than school-based participants to recommend children in the vignettes for speech-language pathology services, although the effect size was negligible, as shown in Table 4. Post hoc chi-square tests revealed that work setting associated with recommendation for speech-language pathology services for Vignettes B and F. Health care–based participants also recommended Vignette E more often than school-based participants, but the group difference was nonsignificant. No effect of work setting emerged for Vignettes A, C, and D. Severity level of omnibus and vocabulary standard scores is the one dimension of affectedness that separates Vignettes B, E, and F from A, C, and D.
Table 4.
Clinical decision-making.
| Survey question | Health care |
School |
Χ2(1) | V | RR | 95% CI |
|||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | LL | UL | ||||
| Recommendation for (continued) SLP services | |||||||||
| Total | 272 | 73.51 | 1033 | 62.57 | 15.83** | 0.089 | 0.85 | 0.79 | 0.91 |
| Vignette A | 63 | 95.45 | 250 | 88.34 | 2.93 | 0.092 | 1.08 | 1.01 | 1.16 |
| Vignette B | 29 | 45.31 | 61 | 21.63 | 15.20** | 0.21 | 2.09 | 1.48 | 2.97 |
| Vignette C | 61 | 98.39 | 274 | 98.92 | 0.12 | 0.019 | 0.99 | 0.96 | 1.03 |
| Vignette D | 51 | 83.61 | 235 | 86.08 | 0.25 | 0.027 | 0.97 | 0.86 | 1.10 |
| Vignette E | 29 | 48.33 | 88 | 32.96 | 5.04 | 0.124 | 1.47 | 1.07 | 2.00 |
| Vignette F | 39 | 68.42 | 125 | 46.47 | 9.07* | 0.165 | 1.47 | 1.18 | 1.83 |
Note. The bolded percentages represent the work setting with the higher-than-expected responses. CI = confidence interval; RR = relative risk; LL = lower limit; UL = upper limit; SLP = speech-language pathologist.
p < .01.
p < .001.
Next, we examined how work setting associated with response rate within each severity level: standard scores below 78 and borderline standard scores between 83 and 85. When vignettes presented with omnibus and vocabulary standard scores lower than 78, no effect of work setting emerged. Both health care–based and school-based participants recommended (continued) speech-language pathology services often, χ2(1) = 0.43, p = .51, V = 0.020 (health care–based: 92.59%, school-based: 91.12%; RR: 1.02, 95% CI [0.97, 1.06]). When vignettes presented with borderline standard scores (between 83 and 85), a small work setting effect emerged, χ2(1) = 25.63, p < .001, V = 0.160. Health care–based participants recommended speech-language pathology services around chance rates, whereas school-based participants did not often make this recommendation (health care–based: 53.59%, school-based: 33.50%; RR: 1.60; 95% CI [1.35, 1.89]).
Practice Patterns Differ Across Work Settings From Referral to Discharge
As hypothesized, work setting effects appeared across reported practice pattern areas. For referral sources, nearly all health care–based participants reported physicians as a referral source, which was significantly more often than school-based participants, χ2(1) = 64.21, p < .001, V = 0.383 (health care–based: 90.82%, school-based: 45.13%; RR: 2.01; 95% CI [1.76, 2.30]). In contrast, school-based participants reported teachers as referral sources significantly more often than health care–based participants, χ2(1) = 98.25, p < .001, V = 0.474 (health care–based: 95.89%, school-based: 58.13%; RR: 0.61; 95% CI [0.51, 0.72]). This pattern emerged for school screenings as well, χ2(1) = 25.35, p < .001, V = 0.241 (health care–based: 37.76%, school-based: 66.08%, RR: 0.57; 95% CI [0.44, 0.75]). No difference emerged for parents as a referral source (health care–based: 88.78%, school-based: 91.74%; RR: 0.97; 95% CI [0.90, 1.05]).
For assessment practices, a few differences between groups appeared, as shown in Table 5. Health care–based participants reported using the Preschool Language Scales (Zimmerman et al., 2002, 2011) significantly more often than school-based participants; however, they were less likely to report using the Comprehensive Assessment of Spoken Language (Carrow-Woolfolk, 1999, 2017), the Expressive One-Word Picture Vocabulary Test (Martin & Brownell, 2000, 2011), and Expressive Vocabulary Test (Williams, 2007, 2018) than school-based participants. No differences emerged for the Clinical Evaluation of Language Fundamentals (Wiig et al., 2003, 2013), the Peabody Picture Vocabulary Test (D. M. Dunn, 2018; L. M. Dunn & Dunn, 2007), the Receptive One-Word Picture Vocabulary Test (Brownell, 2000, 2010), the Test of Language Development (Newcomer & Hammill, 2008, 2019), the Oral and Written Language Scales (Carrow-Woolfolk, 1995, 2011), or the Test of Early Grammatical Impairment (TEGI; Rice & Wexler, 2001). Health care–based participants also reported using the informal assessment method of parent interviewing more often than school-based participants while reporting classroom observations and teacher interviews less often than school-based participants. No differences emerged for language sample analysis, clinical observation, or dynamic assessment.
Table 5.
Assessment practices.
| Survey question | Health care |
Schools |
Χ2(1) | V | RR | 95% CI |
|||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | LL | UL | ||||
| Norm-referenced assessments | |||||||||
| Preschool Language Scales | 67 | 72.84 | 186 | 55.36 | 9.12* | 0.146 | 1.32 | 1.12 | 1.54 |
| Comprehensive Assessment of Spoken Language | 25 | 27.17 | 16 | 49.70 | 14.82** | 0.186 | 0.55 | 0.38 | 0.78 |
| Expressive One-Word Picture Vocabulary Test | 25 | 27.17 | 141 | 41.96 | 6.65* | 0.125 | 0.65 | 0.45 | 0.93 |
| Expressive Vocabulary Test | 14 | 16.30 | 106 | 31.55 | 8.28* | 0.139 | 0.52 | 0.32 | 0.84 |
| Clinical Evaluation of Language Fundamentals | 72 | 78.26 | 269 | 80.06 | 0.14 | 0.018 | 0.98 | 0.87 | 1.10 |
| Peabody Picture Vocabulary Test | 34 | 36.96 | 164 | 48.81 | 4.08 | 0.098 | 0.76 | 0.57 | 1.01 |
| Receptive One-Word Picture Vocabulary Test | 22 | 23.91 | 108 | 32.14 | 2.31 | 0.074 | 0.74 | 0.50 | 1.11 |
| Test of Oral Language Development | 23 | 25.00 | 121 | 36.01 | 3.92 | 0.096 | 0.69 | 0.47 | 1.02 |
| Oral and Written Language Scales | 33 | 35.87 | 130 | 38.69 | 0.24 | 0.024 | 0.93 | 0.68 | 1.26 |
| Test of Early Grammatical Impairment a | 2 | 2.17 | 1 | 0.30 | 3.65 | — | — | — | — |
| Informal assessments | |||||||||
| Parent interview | 93 | 93.94 | 242 | 71.60 | 21.36** | 0.221 | 1.31 | 1.21 | 1.43 |
| Classroom observations | 31 | 31.31 | 275 | 89.87 | 91.37** | 0.457 | 0.38 | 0.29 | 0.52 |
| Teacher interview | 46 | 46.46 | 301 | 89.05 | 84.93** | 0.441 | 0.52 | 0.42 | 0.65 |
| Language sample analysis | 82 | 82.83 | 284 | 84.02 | 0.08 | 0.014 | 0.99 | 0.89 | 1.09 |
| Clinical observation | 91 | 91.92 | 282 | 83.43 | 4.41 | 0.100 | 1.10 | 1.02 | 1.19 |
| Dynamic assessment | 35 | 35.35 | 98 | 28.99 | 1.46 | 0.058 | 1.22 | 0.89 | 1.67 |
Note. The bolded percentages represent the work setting with the higher-than-expected responses. CI = confidence interval; RR = relative risk; LL = lower limit; UL = upper limit.
More than 50% of cells had less than five expected counts, invalidating the chi-square test.
p < .01.
p < .001.
For service delivery, significant differences between work settings appeared, as shown in Table 6. Almost all health care–based participants reported treating patients individually, whereas school-based participants reported treating up to a third of their students individually. In contrast, most school-based participants reported treating their students in groups of two to four and some treating students in groups of five to six. Few participants in either group reported treating patients or students in groups of seven or more.
Table 6.
Service delivery practices.
| Survey question | Health care |
School |
Χ2(3) | V | ||
|---|---|---|---|---|---|---|
| n | % | n | % | |||
| Individual | ||||||
| None | 0 | 0.00 | 25 | 7.18 | 264.81** | 0.436 |
| 1%–33% | 74 | 5.17 | 216 | 70.11 | ||
| 34%–66% | 17 | 10.34 | 95 | 14.66 | ||
| 67%–100% | 12 | 84.48 | 2 | 8.05 | ||
| Groups of 2–4 | ||||||
| None | 78 | 67.24 | 12 | 3.45 | 250.97** | 0.427 |
| 1%–33% | 28 | 24.14 | 70 | 20.11 | ||
| 34%–66% | 7 | 6.03 | 116 | 33.33 | ||
| 67%–100% | 3 | 2.59 | 150 | 43.10 | ||
| Groups of 5–6 | ||||||
| None | 107 | 92.24 | 247 | 70.98 | 22.85** | 0.128 |
| 1%–33% | 9 | 7.76 | 77 | 22.13 | ||
| 34%–66% | 0 | 0.00 | 20 | 5.75 | ||
| 67%–100% | 0 | 0.00 | 4 | 1.15 | ||
| Groups of 7+ | ||||||
| None | 111 | 95.69 | 322 | 92.53 | 2.28 | 0.041 |
| 1%–33% | 4 | 3.45 | 21 | 6.03 | ||
| 34%–66% | 1 | 0.86 | 2 | 0.57 | ||
| 67%–100% | 0 | 0.00 | 3 | 0.86 | ||
Note. The bolded percentages represent the work setting with the higher-than-expected responses. Relative risk and confidence intervals are not provided because the chi-square table is 2 × 4.
p < .001.
For treatment, a difference appeared across work settings for the number of goals per treated individual, χ2(1) = 54.57, p < .001, V = 0.363. Most health care–based participants reported four or more goals (one goal: 2.17%, two goals: 5.43%, three goals: 28.26%, four or more goals: 64.13%). In contrast, only 25.39% of school-based participants reported four or more goals per student (one goal: 7.12%, two goals: 31.27%, three goals: 36.22%). However, goal selection did not vary based on work setting, as shown in Table 7.
Table 7.
Treatment practices.
| Survey question | Health care |
School |
Χ2(1) | V | RR | 95% CI |
|||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | LL | UL | ||||
| Vocabulary development | 83 | 91.21 | 304 | 94.12 | 0.99 | 0.049 | 0.97 | 0.90 | 1.04 |
| Listening comprehension | 75 | 82.42 | 287 | 88.85 | 2.68 | 0.080 | 0.93 | 0.84 | 1.03 |
| Morphological development | 68 | 74.73 | 230 | 71.21 | 0.44 | 0.032 | 1.05 | 0.91 | 1.20 |
| Sentence formulation | 79 | 86.81 | 288 | 89.16 | 0.39 | 0.031 | 0.97 | 0.89 | 1.06 |
| Syntax development | 75 | 82.42 | 288 | 89.16 | 2.99 | 0.085 | 0.92 | 0.83 | 1.02 |
| Narrative development | 60 | 65.93 | 184 | 56.97 | 2.36 | 0.076 | 1.16 | 0.97 | 1.38 |
| Pronoun development | 82 | 90.11 | 288 | 89.16 | 0.07 | 0.013 | 1.01 | 0.93 | 1.09 |
| Verb tense | 81 | 89.01 | 292 | 90.40 | 0.15 | 0.019 | 0.98 | 0.91 | 1.07 |
| Preposition | 74 | 81.32 | 259 | 80.19 | 0.06 | 0.012 | 1.01 | 0.91 | 1.13 |
| Verb vocabulary | 60 | 65.93 | 194 | 60.06 | 1.03 | 0.050 | 1.10 | 0.92 | 1.30 |
| Question formulation | 64 | 70.33 | 244 | 75.54 | 1.01 | 0.050 | 0.93 | 0.80 | 1.08 |
| Complex sentences | 61 | 67.03 | 223 | 69.04 | 0.13 | 0.018 | 0.97 | 0.83 | 1.14 |
Note. The bolded percentages represent the work setting with the higher-than-expected responses. CI = confidence interval; RR = relative risk; LL = lower limit; UL = upper limit.
p < .01.
p < .001.
For collaboration, differences appeared across work settings, as shown in Table 8. Health care–based participants reported less frequent collaboration with teachers (i.e., general education, reading, special education) as compared to school-based participants. They also reported less frequent collaboration with psychologists than school-based participants. Conversely, health care–based participants reported collaborating with parents frequently (i.e., daily, 2–3 times per week), whereas most school-based participants reported collaborating with parents monthly and some collaborating weekly.
Table 8.
Collaboration practices.
| Survey question | Health care |
School |
Χ2(5) | V | ||
|---|---|---|---|---|---|---|
| n | % | n | % | |||
| Classroom teacher | ||||||
| Daily | 7 | 8.05 | 61 | 18.94 | 91.27** | 0.472 |
| 2–3 times per week | 14 | 16.09 | 111 | 34.47 | ||
| Weekly | 17 | 19.54 | 106 | 32.92 | ||
| Monthly | 14 | 16.09 | 8 | 2.48 | ||
| Never | 26 | 29.89 | 12 | 3.73 | ||
| Not applicable | 9 | 10.34 | 24 | 7.45 | ||
| Reading teacher | ||||||
| Daily | 4 | 4.60 | 23 | 7.14 | 51.96** | 0.336 |
| 2–3 times per week | 3 | 3.45 | 44 | 13.66 | ||
| Weekly | 7 | 8.05 | 116 | 36.02 | ||
| Monthly | 19 | 21.84 | 45 | 13.98 | ||
| Never | 51 | 58.62 | 80 | 24.84 | ||
| Not applicable | 3 | 3.45 | 14 | 4.35 | ||
| Special education teacher | ||||||
| Daily | 4 | 4.60 | 85 | 26.40 | 161.30** | 0.628 |
| 2–3 times per week | 6 | 6.90 | 116 | 36.02 | ||
| Weekly | 24 | 27.59 | 50 | 15.53 | ||
| Monthly | 13 | 14.94 | 1 | 0.31 | ||
| Never | 31 | 35.63 | 9 | 2.80 | ||
| Not applicable | 9 | 10.34 | 61 | 18.94 | ||
| Psychologist | ||||||
| Daily | 1 | 1.15 | 85 | 26.40 | 28.64** | 0.265 |
| 2–3 times per week | 4 | 4.60 | 116 | 36.02 | ||
| Weekly | 31 | 35.63 | 50 | 15.53 | ||
| Monthly | 18 | 20.69 | 1 | 0.31 | ||
| Never | 26 | 29.89 | 9 | 2.80 | ||
| Not applicable | 7 | 8.05 | 61 | 18.94 | ||
| Parents | ||||||
| Daily | 14 | 16.47 | 7 | 2.24 | 200.60** | 0.711 |
| 2–3 times per week | 48 | 56.47 | 21 | 6.73 | ||
| Weekly | 8 | 9.41 | 74 | 23.72 | ||
| Monthly | 5 | 5.88 | 205 | 65.71 | ||
| Never | 0 | 0 | 3 | 0.96 | ||
| Not applicable | 10 | 11.76 | 2 | 0.64 | ||
Note. The bolded percentages represent the work setting with the higher-than-expected responses. Relative risk and confidence intervals not provided because the chi-square table is 2 × 6.
p < .01.
p < .001.
Finally, for discharge and monitoring, differences appeared across work setting. Most participants reported they discharge an individual when they perform within the normal range of an assessment, although health care–based participants reported this less often than school-based participants, χ2(1) = 8.82, p = .003, V = 0.147 (health care–based: 79.07%, school-based: 90.68%; RR: 0.87; 95% CI [0.78, 0.98]). Most participants also reported that they discharge an individual once they meet treatment objectives (health care–based: 63.95%, school-based: 74.84%; RR: 0.85; 95% CI [0.72, 1.01]) or no progress is occurring (health care–based: 59.30%, school-based: 50.93%; RR: 1.16; 95% CI [0.95, 1.42]), with no significant difference between groups. A significant difference occurred for the discharge indicator of reaching grade level; less than a third of health care–based participants reported this (31.40%) while 66.15% of school-based participants reported this, χ2(1) = 33.84, p < .001, V = 0.288 (RR: 0.47; 95% CI [0.34, 0.65]). Just under half of participants reported that they continue monitoring individuals after discharge with no significant difference between groups (health care–based: 44.58%, school-based: 49.05%; RR: 0.91; 95% CI [0.70, 1.18]).
Caseload Composition Differs Across Work Settings
As hypothesized, differences across work setting appeared for participants' reported caseload composition, as shown in Table 9. Most health care–based participants reported caseloads of less than 20 patients, whereas most school-based participants reported caseloads of over 40 students. Despite the significant difference in caseload size, school-based participants were no more likely to report paraprofessional support. When considering caseload composition, effects of work setting also appeared. Health care–based participants reported serving infant-, toddler-, and preschool-age children more often than school-based participants. Both groups reported serving elementary-, middle school–, and high school–age students at equivalent rates, although markedly fewer participants in both groups served middle- and high school-age students. Health care–based participants reported serving fewer students with speech impairment only and language impairment only than school-based participants. They reported serving more students with speech and language impairment and those with medical diagnoses (e.g., autism spectrum disorder, cerebral palsy, attention deficit hyperactivity disorder) than school-based participants. Finally, no difference between health care–based (M = 16.77, SD = 24.47) and school-based (M = 19.08, SD = 17.56) participants appeared for the reported number of individuals with SLI on the their caseloads, t(324) = 0.84, p = .403. The 95% CI around the difference of 2.31 is (−3.12, 7.74).
Table 9.
Caseload composition.
| Survey question | Health care |
School |
df | Χ2 | V | RR | 95% CI |
|||
|---|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | LL | UL | |||||
| Caseload size a | ||||||||||
| Less than 20 | 67 | 53.60 | 32 | 8.70 | 2 | 139.73 ** | .532 | — | — | — |
| Between 20 and 40 | 44 | 35.20 | 118 | 32.07 | ||||||
| Greater than 40 | 14 | 11.20 | 218 | 59.25 | ||||||
| Hours worked a | ||||||||||
| Less than 20 | 31 | 22.14 | 29 | 7.44 | 2 | 22.34 ** | .205 | — | — | — |
| Between 20 and 40 | 63 | 45.00 | 201 | 51.54 | ||||||
| Greater than 40 | 46 | 32.86 | 160 | 41.03 | ||||||
| Paraprofessional support: yes | 16 | 12.80 | 45 | 12.23 | 1 | 0.03 | .008 | 1.05 | 0.61 | 1.78 |
| Ages served | ||||||||||
| Infant and toddler | 85 | 73.91 | 24 | 6.88 | 1 | 216.27 ** | .683 | 10.72 | 7.20 | 16.05 |
| Preschool | 94 | 81.74 | 202 | 57.88 | 1 | 21.32 ** | .214 | 1.41 | 1.25 | 1.60 |
| Elementary | 89 | 77.39 | 276 | 79.08 | 1 | 0.14 | .018 | 0.99 | 0.87 | 1.10 |
| Middle school | 59 | 51.30 | 140 | 40.11 | 1 | 4.42 | .098 | 1.28 | 1.03 | 1.59 |
| High school | 43 | 37.39 | 93 | 26.65 | 1 | 4.82 | .102 | 1.40 | 1.05 | 1.88 |
| Speech impairment only a | ||||||||||
| None | 23 | 19.83 | 25 | 7.18 | 3 | 20.10 ** | .204 | — | — | — |
| 0%–33% | 74 | 63.79 | 216 | 62.07 | ||||||
| 34%–66% | 17 | 14.66 | 95 | 27.30 | ||||||
| 67%–100% | 12 | 3.45 | 2 | 1.72 | ||||||
| Language impairment only a | ||||||||||
| None | 18 | 15.52 | 14 | 4.02 | 3 | 18.91 ** | .202 | — | — | — |
| 0%–33% | 59 | 50.85 | 183 | 52.59 | ||||||
| 34%–66% | 29 | 25.00 | 107 | 30.75 | ||||||
| 67%–100% | 10 | 8.62 | 44 | 12.64 | ||||||
| Speech & language impairment a | ||||||||||
| None | 2 | 1.72 | 6 | 1.72 | 3 | 20.41 ** | .210 | — | — | — |
| 0%–33% | 38 | 32.76 | 196 | 56.32 | ||||||
| 34%–66% | 47 | 40.52 | 98 | 28.16 | ||||||
| 67%–100% | 29 | 25.00 | 48 | 13.79 | ||||||
| Medical diagnosis a | ||||||||||
| None | 2 | 1.72 | 10 | 2.87 | 3 | 11.14 * | .155 | — | — | — |
| 0%–33% | 47 | 40.52 | 181 | 52.01 | ||||||
| 34%–66% | 43 | 37.07 | 123 | 35.34 | ||||||
| 67%–100% | 24 | 20.69 | 34 | 9.77 | ||||||
Note. Em dashes indicate that data were unable to be calculated. The bolded percentages represent the work setting with the higher-than-expected responses. CI = confidence interval; RR = relative risk; LL = lower limit; UL = upper limit.
Relative risk and confidence intervals were not provided because the chi-square tables were 2 × 3 or 2 × 4.
p < .01.
p < .001.
Discussion
The overarching aim of this project combined with our previous report (Selin et al., 2019) was to explore how SLPs use the SLI research literature based on their reported practice patterns and clinical decision-making. To do so, we examined the effects of practitioner characteristics (e.g., years of experience), child characteristics (e.g., clinical symptoms), and workplace characteristics (e.g., policies). This report focused on one workplace characteristic—work setting—and the primary findings are that SLPs operate within complex systems, and effects of work setting are pervasive across reported practice patterns and clinical decision-making in ways that may influence clinical service provision for individuals with SLI. A differential work setting effect appeared in SLPs' workplace conditions (e.g., eligibility considerations), eligibility beliefs, and clinical decision-making for identifying individuals with SLI for services when standard scores are borderline. These work setting differences appeared even when instructed to disregard workplace conditions, which points to deeper issues than top-down initiatives can address alone.
In parallel to Murza and Ehren's (2020) claims, more pressing issues exist besides shifting diagnostic terminology and criteria. Advocacy efforts without consideration for the SLP and their workplace conditions will have limited effectiveness. Furthermore, without also addressing these workplace conditions and SLPs' roles within the complex systems, public campaigns aimed at increasing referral rates for speech-language pathology services will further clog an already saturated service provision system, especially in the schools. In the sections to follow, we discuss the findings from this study through a middle-out lens—how work setting variations influence SLPs' workplace conditions, eligibility beliefs, and clinical decision-making as relevant for systemic change. We then review how those factors influence reported practice patterns and caseload composition.
Work Setting Variations Influence Clinical Service Provision
Work setting variations influence speech-language pathology service provision as relevant to individuals with SLI. First, SLPs reported differing eligibility terminology and criteria across work settings in ways that aligned with specific diagnostic and funding models of the overlaid complex system. Health care–based SLPs reported fewer eligibility restrictions than school-based SLPs. Second, health care–based SLPs reported a less restrictive belief around eligibility for individuals with clinical profiles aligning with SLI. Third, health care–based SLPs showed more sensitivity to clinical profiles aligning with SLI when making clinical decisions in the vignettes. Fourth, reported practice patterns from referral to discharge differed between health care– and school-based SLPs. Finally, despite the above, health care–based SLPs reported fewer individuals with language impairment only on their caseloads than school-based SLPs.
In all, the following findings point to the complexity that SLPs must navigate in the workplace. Importantly, these findings should not be interpreted as health care–based SLPs having inherently superior ability or performance in identifying individuals with SLI. Rather, health care– and school-based SLPs work in separate but equally complex workplace systems with different challenges. In addition, the variation in workplace conditions, eligibility beliefs, and clinical decision-making that influence clinical service provision are likely not independent of one another when looking at reported differences across the two work setting groups. This complexity is important for researchers to consider when contextualizing their results so they can identify targeted advocacy initiatives based on work setting to empower SLPs as agents for change.
Workplace Conditions: Eligibility Criteria Drive Clinical Practice
In this study, the workplace condition of interest was eligibility, which might affect the clinical service provision for individuals with SLI, and how eligibility differs across the health care and public education work settings. A primary finding is that eligibility drives clinical practice in different ways across work settings, affecting both terminology and criteria. As expected, the SLPs in this study reported terminology consistent with the classification systems used in their work setting. For health care–based SLPs, reported terminology coincided with ICD-10 codes (WHO, 2016) used for insurance reporting. For school-based SLPs, reported terminology coincided with the IDEA (2004) classification system. Some SLPs in both groups reported using the diagnostic terminology of SLI and language disorder; however, SLPs reported these terms markedly less often than the eligibility terminology. Health care–based SLPs reported using the diagnostic term SLI more often than school-based SLPs.
Perhaps more alarming than limited use of diagnostic terms was how SLPs in this study reported use of diagnostic criteria. Again, eligibility superseded diagnostics, and SLPs reported eligibility criteria consistent with their workplace. SLPs reported eligibility criteria that may interfere with identifying individuals with SLI for speech-language pathology services, creating potentially restrictive workplace conditions. These workplace eligibility practices are of great concern. Individuals meeting the broadest research-based definitions of SLI (e.g., −1 SD below the age mean on omnibus language assessments) are at risk for an array of negative long-term outcomes, including persistent language impairment (Johnson et al., 1999; Rice & Hoffman, 2015), academic underachievement (Records et al., 1992), peer victimization (Redmond, 2011), emotional and behavioral disorders (Conti-Ramsden & Botting, 2008), incarceration (Snow & Powell, 2011), and sexual assault (Brownlie et al., 2007). Based on reported workplace conditions and eligibility practices reported in this study, these individuals are unlikely to meet eligibility criteria, despite meeting the broad research-based diagnostic criteria for SLI. The SLP has limited control over workplace conditions and eligibility practices when they need to work in alignment with legal educational requirements or insurance company policies. SLPs will benefit from top-down advocacy efforts (e.g., policy, legislation, lobbying) aimed at removing and reducing these potentially restrictive workplace conditions associated with eligibility criteria.
Overall, health care–based SLPs reported less restrictive eligibility criteria. Health care–based SLPs operate within a fee-for-service model, with insurance reimbursement as a potential gatekeeper. Insurance companies impose fewer and less restrictive criteria than in some school-based settings. Neither of the health care–based eligibility criteria (i.e., medical necessity, insurance coverage) generally impose a performance threshold above and beyond the broadest research-based diagnostic criteria associated with SLI and other language disorders: performing 1 SD below the age-referenced mean. However, most health care–based SLPs (65%) in this study did not report health insurance as an eligibility criterion. Perhaps this sample included many SLPs who run cash- or private pay–only practices, where eligibility criteria do not exist and the SLP has full autonomy. Alternatively, health care–based SLPs may engage in fully autonomous clinical decision-making, and nonclinical staff handle issues related to insurance reimbursement or denials. Regardless, a third of health care–based SLPs reported full autonomy in eligibility such that their clinical judgment alone may suffice, and 62% of health care–based SLPs reported that low performance on language assessments alone qualifies an individual for speech-language pathology services. Taken together, an individual with SLI will experience fewer barriers to service in the health care setting since the SLP's workplace conditions are less restrictive around their clinical decision-making.
School-based eligibility criteria, on the other hand, include a two-prong approach such that an individual with SLI must first meet eligibility for a disability and then also show adverse educational impact (IDEA, 2004). The school-based SLPs in this study reported the need to follow such an approach for eligibility, as consistent with federal legislation. Definitions for both prongs vary across states in terms of restrictiveness. In the study presented here, we did not ask which state SLPs worked in; thus, we report on the average restrictiveness in the school setting. Overall, school-based SLPs reported the need for adverse educational impact and more restrictive performance thresholds such as often needing to meet a particular cutoff score on two assessments. Furthermore, almost all school-based SLPs in this study reported academic impact as an eligibility consideration and that academic impact must coincide with low language performance for eligibility. This study did not investigate how SLPs interpret adverse academic impact or how an SLP determines whether an individual meets this criterion. Given that individuals with SLI are at risk for academic underachievement (Records et al., 1992), one would expect individuals with SLI to meet this eligibility criterion. However, the research is clear that individuals meeting research-based diagnostic criteria for SLI are not often on clinical caseloads (e.g., Johnson et al., 1999; Tomblin et al., 1997; Zhang & Tomblin, 2000). Future research is needed regarding which prong of school-based eligibility places individuals with SLI at greater risk for underidentification for services. In all, an individual with SLI will experience more barriers to service in the public education setting since school-based SLPs' workplace conditions are more restrictive around their clinical decision-making.
Health Care–Based SLPs Agree More Often With Less Restrictive Eligibility
Health care–based SLPs not only reported less restrictive eligibility criteria in the workplace, but they also agreed more often than school-based SLPs that any individual performing 1 SD below the age mean on a language assessment should receive services. The variation across and within groups on this question highlights that clinicians differ in their beliefs around arbitrary cutoff scores. We are not asserting that all individuals scoring 1 SD below the age-mean warrant speech-language pathology services; however, the response variation for this question highlights that clinicians' beliefs differ between work settings in ways that may influence who is or is not identified for services. Given that beliefs influence practice patterns (Kritikos, 2003), this between-groups difference in eligibility restrictiveness belief may protect or place an individual with SLI at further risk for underidentification within work settings that have restrictive eligibility criteria. Specifically, beliefs on eligibility restrictiveness may influence important assessment practices such as assessment selection, which aspect of a clinical profile is most important, and how to interpret assessment results (e.g., standard score severity).
This difference between health care– and school-based SLPs also underscores that the field faces larger issues than diagnostic terminology. Ultimately, researchers utilize diagnostic criteria that align with the goals and methods of their program of research, the need for valid and replicable findings, and clear interpretations for clinical practice across various practice settings. Clinicians may not implement what researchers identify as best practice, given the difference in the research aims and outcomes and the realities of various clinical settings. Possible overlap of research outcomes and clinical practice settings and objectives could accelerate systemic change. The questionnaire's items about eligibility restrictiveness beliefs and in practice provide an informative example. When asked to disregard workplace conditions, many SLPs, especially in the schools, did not agree that clinical profiles aligning with the broadest research-based diagnostic criteria of SLI warrant services. This finding may point to a lack of awareness around the research that documents the negative long-term outcomes and the importance of speech-language pathology services for individuals with SLI even when standard scores are borderline or when occurring in the absence of other symptoms and diagnoses. If so, targeted advocacy may focus on preservice education that illustrates how children with mild profiles remain at risk for a host of negative outcomes in adulthood if undiagnosed and untreated. Alternatively, workplace conditions specific to the public education system may influence school-based SLPs' beliefs around eligibility for individuals with SLI. School-based SLPs may instead believe that individuals with mild impairment are better served through a response-to-intervention or multitiered systems of support because of their experiences with state guidelines or mandates around this workplace practice.
Note that this study cannot determine whether this eligibility belief predicates certain practices (e.g., how SLPs interpret assessment data for eligibility) or whether workplace conditions (e.g., mandated policies, caseloads) create previous experiences from which SLPs form their beliefs. Either way, top-down initiatives aiming to increase referrals to the SLP will have limited effectiveness if SLPs do not agree with researchers about who warrants their services. Future research is needed regarding SLPs' belief systems around eligibility, which also may influence their tolerance for various workplace conditions such as eligibility restrictiveness.
Health Care–Based SLPs Recommend Services More Often for Borderline Language Impairment in Vignettes
When instructed to ignore workplace conditions, an effect of work setting appeared for decision-making. In this study, health care–based SLPs identified vignettes portraying children with SLI as warranting speech-language pathology services at higher rates than school-based SLPs, especially when the clinical profile had borderline omnibus language standard scores. If workplace eligibility criteria alone explained low rates of identification for individuals with SLI in actual practice, this finding would be unexpected. One theory is that health care–based SLPs possess stronger knowledge and skills for identifying individuals with SLI; however, an alternative theory is more likely—SLPs may not be able to disregard workplace conditions when making clinical decisions even when instructed to do so. In other words, workplace conditions create previous professional experience that influence SLPs' clinical decision-making, which parallels the finding that previous professional experience influences SLPs' practice patterns (Kritikos, 2003).
Indeed, an interesting association appeared when comparing health care– and school-based SLPs' reported cutoff score requirements (i.e., a workplace condition) with their clinical decision-making when instructed to disregard workplace conditions. Approximately half of health care–based SLPs reported a required cutoff score of 85, no standard score required, or no standardized assessment required for eligibility. When making clinical decisions under neutral workplace conditions, health care–based SLPs identified the vignettes with borderline omnibus language standard scores (Vignettes B, E, and F) as warranting speech-language pathology services approximately 50% of the time. Thus, health care–based SLPs' workplace conditions (i.e., required cutoff scores) paralleled their decision-making when instructed to disregard workplace conditions. For school-based SLPs, approximately a third reported those same cutoff score requirements, and they identified the vignettes with borderline omnibus language standard scores as warranting speech-language pathology services approximately a third of the time. The reduced rates reported by the school-based SLPs may stem from a workplace condition influencing their decision-making—implementing response to intervention, as reviewed above. School-based SLPs may have recognized the vignettes with borderline performance presented with language difficulties, but they may also have perceived the difficulties as mild and addressable through Tier 1 or 2 services. Taken together, an SLP's clinical decision-making as relevant for closing the SLI underidentification gap may be tied to their workplace conditions; advocacy efforts must consider both elements to be effective.
SLPs' Practice Patterns Vary Across Work Settings
SLPs' workplace conditions as well as their eligibility beliefs and clinical decision-making for identifying SLI differ across work setting. These differences influence their reported practice patterns from referral to discharge; however, it is unclear which of the three factors affect each individual practice pattern, which future research will need to address. When considering implementation science, researchers should consider that certain work settings may require targeted evidence-based practice campaigns based on setting-specific workplace conditions, SLPs' awareness of needed change, and their previous knowledge and skills for decision-making in that clinical area. For example, a previous report found that SLPs deviate from best practice in assessment selection for individuals with SLI (Betz et al., 2013). Rather than using morphosyntactic domain-specific follow-up testing (e.g., TEGI; Rice & Wexler, 2001), most SLPs reported using single-word vocabulary tests. Single-word vocabulary tests lack sensitivity to the specific deficits of individuals with SLI (Betz et al., 2013) and lack sufficient content validity given they assess only one domain of the multidimensional language construct. However, that study included only school-based SLPs. In our study, health care–based SLPs used single-word vocabulary tests less often than school-based SLPs; however, it is unknown whether this differing practice pattern is because of workplace conditions, the SLPs' awareness of the importance of psychometric properties, or their knowledge and skill in interpreting psychometric properties. For example, state interpretation of IDEA (2004) may require that school-based SLPs administer two standardized assessments for qualification. The selection of a follow-up vocabulary assessment may not stem from a lack of awareness or limited knowledge that the test lacks sensitivity, but, rather, they may select that assessment because it is fast and easy to fulfill the state mandate requirement in an efficient manner. Researchers wanting to improve clinical assessment selection will need to explore all three setting-specific factors—conditions, concerns (i.e., awareness and beliefs), and capacity (i.e., knowledge, skills, and clinical decision-making)—and determine the root issue.
Intervention is another example of work setting effects requiring targeted evidence-based campaigns. Much of the clinical research available for morphosyntactic and syntactic intervention focuses on individually delivered services and parent coaching. Both intervention types will be difficult for school-based SLPs to deliver because of workplace conditions. First, over 75% of school-based SLPs reported delivering individual intervention less than a third of the time; thus, most individuals with SLI receive intervention in groups, likely because of large caseload sizes (Amir et al., 2021; Brandel & Frome Loeb, 2011). Second, most school-based SLPs reported collaborating with parents monthly (65.71%), limiting the ability to coach parents on language intervention strategies. Health care–based SLPs do not encounter these barriers for intervention. In contrast, most health care–based SLPs collaborate with parents 2–3 times per week (56.47%) but rarely collaborate with their clients' teachers, which may influence generalization into the classroom environment. In this report, school-based SLPs reported frequent collaboration with classroom teachers; however, the nature of the collaboration is unclear (see previous report on barriers to collaboration in the school setting; Pfeiffer et al., 2019).
Not all practices varied across work settings. No differences emerged in goal selection and certain assessment practices, such as language sample analysis, clinical observations, and dynamic assessment. In all, clinical research may not generalize to both work settings equally given their unique opportunities and barriers. One work setting may need additional advocacy for the implementation of evidence-based practices to ensure effective service provision for individuals with SLI, and that advocacy may be best directed at workplace conditions such as restrictive eligibility criteria and unmanageable caseload sizes.
SLPs Report Few Individuals With Language Impairment on Caseloads
Despite the prevalence and negative long-term outcomes associated with SLI, most SLPs in this study reported that individuals with language impairment only comprised less than a third of their caseloads. Given the discussion thus far, perhaps this is not surprising. Many systemic barriers face individuals with SLI before an SLP deems them eligible for services. First, a referral source must notice that the individual warrants the SLP's attention. Second, they must meet potentially restrictive eligibility criteria and, in the schools, must also present with adverse educational impact. Third, their SLP may not believe an individual with their profile warrants services, or they may not be sensitive to their profile if the standard score is not severe enough. Fourth, SLPs have a responsibility to manage their caseload size, which likely influences who they deem eligible for services. School-based SLPs, for example, may determine that individuals with SLI have their language needs met in the classroom or resource room, whereas the SLP is the professional best suited for an individual with complex communication needs.
With health care–based SLPs reporting fewer eligibility restrictions and increased beliefs and sensitivities to less severely impaired clinical profiles, do they have more individuals with SLI on their caseload? When considering language impairments more broadly, health care–based SLPs report fewer individuals on their caseload with language impairment only. Instead, they report more individuals with speech and language impairment and medical diagnoses than school-based SLPs. However, when asked about SLI, health care– and school-based SLPs report equivalent numbers. Thus, despite having fewer individuals on their caseload overall and fewer individuals with language impairment only, health care–based SLPs do not report fewer individuals with SLI on their caseload. This may be the case because SLI can present as mild or borderline, which may not qualify for school-based services so families with health insurance coverage or financial resources for out-of-pocket pay may seek private health care–based services.
Finally, the ages of people served varied across work settings. SLI occurs across the life span, such that it emerges in early childhood and persists into adulthood (Rice, 2012). Ideally, individuals with SLI are identified during the early childhood years and continue to receive services throughout the academic years. This report revealed that, in general, infants, toddlers, and preschool-age children are more likely to receive speech-language pathology services from health care–based participants. Although a diagnosis of SLI is premature in the infant and toddler years, early red flags emerge during this time (Rudolph, 2017). By the age of 3 years, SLPs have assessment options to make a clinical diagnosis of SLI, including the TEGI (Rice & Wexler, 2001). Researchers investigating prevention and early identification of SLI will want to focus dissemination of evidence-based practices toward health care–based SLPs. Elementary-age children are more likely to receive speech-language pathology services in both work settings. However, middle- and high school–age students are less likely to receive speech-language pathology services overall, regardless of work setting. Advocacy is needed to increase the number of middle and high school individuals receiving speech-language pathology services given that SLI is unlikely to be remediated. Furthermore, as individuals with SLI enter middle and high school, they are at risk for high-stakes outcomes such as high school graduation, postsecondary enrollment, and entering the workforce.
Implications
When centering SLPs in the complex systems through a middle-out lens, the implication is that top-down stakeholder groups (e.g., researchers, policy makers, administrators, regulatory bodies) must consider that workplace conditions influence not only SLPs' reported practice patterns but also their beliefs around eligibility and their clinical decision-making. Disentangling workplace conditions from SLPs' beliefs and clinical decision-making may not be possible in the context of closing the SLI underidentification gap. Without considering this complexity in the systems where SLPs operate, top-down initiatives, such as the diagnostic label and shift to DLD and the research-to-practice pipeline, will have limited effectiveness. Instead, advocacy must address all areas of the complex system including the SLPs' conditions, concerns, and capacity in the workplace, which together influence whether and how SLPs choose or are able to engage in initiatives for systemic change based on their available resources and supports.
Workplace conditions appear to have trickle-down effects on clinical practice, but SLPs have limited ability to change these conditions in their day-to-day work experiences. SLPs can and should engage in advocacy campaigns directed toward administrators and policy-makers to improve their workplace conditions; however, expecting SLPs to shoulder this responsibility is unreasonable given the extent of complex systemic barriers facing them (e.g., large caseload sizes, low health insurance reimbursement rates) and that their primary scope of work is clinical service provision. Rather, top-down stakeholder groups in the United States (e.g., researchers, national and state associations) must focus their advocacy work on removing these barriers so that SLPs can use their best professional judgment and effectively practice as the “autonomous professionals who are the primary care providers of speech-language pathology services” (ASHA, 2016, p. 4), as the ASHA Scope of Practice outlines. These barriers include, but are not limited to, restrictive or outdated eligibility criteria in state-level IDEA mandates, health insurance companies' limited interpretation of medical necessity, unmanageable caseload sizes, and a graduate program bottleneck where the field is limited in how many new SLPs they can train per year to meet the demand for services.
University educators should consider the eligibility beliefs and clinical decision-making findings of this study when working with preservice SLPs. Approximately 40% of health care–based SLPs and 75% of school-based SLPs did not agree with eligibility for speech-language pathology services when asked about a broad definition of language impairment (i.e., performing 1 SD or more below the age-mean). This parallels with approximately 46% of health care–based SLPs and 66% of school-based SLPs not recommending services for the borderline vignette profiles. Although both of these questions instructed SLPs to disregard workplace circumstances, the findings of this study suggest that workplace conditions may nevertheless influence SLPs' beliefs and clinical decision-making given the complex systems they work within. Regardless of whether or not researchers and clinicians believe anyone meeting this broad definition warrants services, these statistics reveal that some SLPs may lack the awareness, belief, or knowledge for identifying students with mild impairments who may benefit from services, likely secondary to workplace conditions (e.g., restrictive eligibility criteria, unmanageable caseloads, workplace mandates and guidelines). This creates specific and high risk for students with this borderline and isolated (i.e., no concomitant issues) profile. Increasing SLPs' awareness and knowledge around the risk for negative long-term outcomes when individuals with even mild language impairment are unidentified for services may attenuate the workplace condition effects. Alternatively, or in combination, university educators must also recognize how workplace conditions potentially shift preservice SLPs' perspectives on best practices once enculturated in different work settings. Perhaps preservice SLPs have the awareness and knowledge for identifying individuals with SLI even when the cases are mild, but their work setting conditions these perspectives over time. Either way, SLPs will benefit from preservice preparation regarding the seriousness of SLI, the coinciding risk for negative long-term outcomes, and the workplace realities they will encounter so they can better advocate for an environment that fosters evidence-based practice.
Clinical researchers and practicing SLPs may identify many clinical implications from the practice patterns reported in this study, but they must be contextualized within the complex systems framework. For example, 44% of health care–based SLPs and 37% of school-based SLPs reported that they use the outdated practice of cognitive referencing, but is this occurring because of the SLPs' workplace conditions (e.g., state-level IDEA eligibility mandate, ICD-10 criteria), SLPs' concern (e.g., unaware that this practice is outdated and exclusive), SLPs' capacity (e.g., limited knowledge for why this practice is problematic), or some combination of the three factors? Future research and advocacy work targeting this practice pattern, and others, will need to consider the complexity of these clinical issues.
Limitations and Future Directions
Although this is the first study to include health care–based SLPs, the sample size was relatively small compared to the school-based respondents. Future research should use targeted recruitment efforts to increase the health care–based sample size. In addition, the medical and private practice work settings were combined into a single health care–based group. Future research may consider investigating work setting effects across three groups—schools, private practices, and hospital outpatient clinics.
Another potential sampling limitation is that we asked the SLPs to select their primary work setting; some SLPs in the sample may have worked in both settings prior to or while completing the survey. Based on the results of this survey, SLPs working in both settings would make an interesting third group for future research since workplace conditions influence SLPs' practice patterns and clinical decision-making. Exposure to both settings may lead to a separate pattern of results. For example, perhaps they experience less restrictive eligibility in their health care–based position, which then influences their practice patterns and clinical decision-making when working in the schools.
Although this report focused on setting as the workplace condition of interest, other workplace condition variables exist and should be considered in future research. Caseload size is another workplace condition that may affect practice patterns and clinical decision-making. Other potential conditions include the ages and grades the SLP serves, the SLP's workplace role in the eligibility determination team, workplace cultures, stress, power dynamics, and time of year (e.g., beginning vs. end of school year). In addition, this report focused on documenting a snapshot of what is happening in clinical practice across work settings rather than how current practices are aligned with the evidence. Future research can explore a variety of workplace conditions to identify barriers to implementing evidence-based practice.
Finally, as with any nonmandatory survey (Eysenbach, 2004), self-selection and volunteer bias are risks. Potential participants who are members in the recruited groups and organizations, as well as potential participants who elect to participate in surveys, may differ from those who do not. Steps that were taken to avoid effects of such bias included attempts to ensure a high response rate, pilot testing, and post hoc comparison of participant demographics with those of ASHA members (Burns et al., 2008; Eysenbach, 2004).
Conclusions
The SLP's work setting influences a host of factors that impact identification rates of individuals with SLI for services (e.g., eligibility criteria, eligibility beliefs, clinical decision-making, practice patterns). Targeted advocacy is needed to address the barriers associated with each complex work setting—health care and public education. The scope of the SLI research literature must also expand to include health care–based SLPs given that findings from school-based research may not always generalize and the health care–based work setting has unique characteristics (e.g., insurance). This report provides a foundation for more research and more specific questions regarding work setting differences, which can provide a more robust basis for future advocacy to increase the identification rates of individuals with SLI for speech-language pathology services.
Acknowledgments
This work was supported by Grants T32DC000052 (awarded to Mabel L. Rice, director), R01DC001803 (awarded to Mabel L. Rice, principal investigator), T32HD101390 (awarded to Jennie Noll and Yo Jackson, co-directors), and T32DC017703 (awarded to Inge-Marie Eigsti and Emily Myers, co-directors). The authors acknowledge and thank the organizations who assisted with dissemination of the survey to potential participants (i.e., American Speech-Language-Hearing Association [ASHA] Significant Interest Group [SIG] 1, ASHA SIG 16, and participating speech-language-hearing state organizations), the speech-language pathologists who completed pilot testing, and the speech-language pathologists who completed the survey.
Appendix
Survey Questions
Demographics
Select the item that best describes you: Female, Male
Please select one response to item (a) and one or more responses to item (b) that best describes you.
Of Spanish-Hispanic origin: No, Yes, Do not know
Select the item(s) that best describe you (select all that apply): American Indian, Eskimo, Aleut; Asian or Pacific Islander; Black or African-American; White; Other; Do not know
What is your current level of education? Bachelors, Masters, Ph.D., Slp.D., Ed.D.
Do you currently have a professional certification in speech-language pathology? Yes, No
What is your primary work setting as a speech-language pathologist? Medical setting, private practice, school
Terminology
-
The following are possible diagnostic categories for students with speech and/or language impairments (disorders). Select only the one(s) that are used for a formal diagnostic label on IEP documentation or ICD-10 codes for the purposes of determining eligibility.
Expressive language impairment/disorder
Receptive language impairment/disorder
Mixed receptive–expressive language impairment/disorder
Language impairment/disorder
Specific learning disability
Speech and language impairment
Eligibility Considerations
-
How does a student qualify for speech-language pathology services (select all that apply)?
Performs below grade level despite implementation of strategies suggested by the speech-language pathologist for a set period of time
Performs below some cut-off point on 2 standardized assessment measures
Performs below some cut-off point on 1 standardized assessment measure
Demonstrates a discrepancy between their mental and chronological age
Demonstrates a discrepancy between their cognitive and language abilities
Speech and/or language abilities are impacting the child's academic success
Informal assessments (e.g., classroom observation, parent/teacher interview, etc.)
Medical diagnosis (e.g., autism spectrum disorder, cerebral palsy)
Previous services/came in with an IEP
Clinical judgment alone (i.e., no standardized assessment required)
Health insurance and/or third-party payer option
-
If using standardized/norm-reference assessments, is a particular cut-off score required? And, if so, what is that particular cut-off score?
I do not use standardized/norm-referenced assessments
No particular cut-off score is required
Required cut-off score: standard score of 85/16th percentile
Required cut-off score: standard score of 81/10th percentile
Required cut-off score: standard score of 76/5th percentile
Would performance below the required cut-off score on language assessments alone be sufficient in order to qualify for speech-language pathology services? Yes, No
Would qualification for speech-language pathology services require performance below the required cut-off score on language assessments and impacted academic performance? Yes, No
Eligibility Belief
Should all students with a language impairment (i.e., 1 standard deviation below average) be regarded as eligible for speech-language pathology services? Yes, No
Practice Patterns
Referral
How is a student identified as needing a speech-language pathology evaluation (select all that apply)? Parent referral, Physician referral, Teacher referral, School screening
Assessment
If using standardized/norm-referenced assessments, which do you typically utilize to determine eligibility (select all that apply)? CASL, CELF, EOWPVT, EVT, OWLS, PLS, PPVT, TEGI, ROWPVT, TOLD
What informal assessment measures do you typically utilize to determine eligibility (select all that apply)? Parent interviews, Teacher interviews, Language sampling, Informal observations, Classroom observations, Dynamic assessment (i.e., test, teach skill, re-test)
Service Delivery
What percentage of children are receiving speech-language pathology services:
| None | 1–33% | 34–66% | 67–100% | |
|---|---|---|---|---|
| Individually | ||||
| In groups of 2–4 students | ||||
| In groups of 5–6 students | ||||
| In groups of 7+ students |
Note. Participants selected one for each row. For example, they would select 1–33% for “individually.”
Treatment
On average, how many treatment objectives do you have per student each year? 1, 2, 3, 4+
Please indicate which of the following intervention targets you would list as a documented language goal (i.e., on IEPs or reports/Subjective-Objective-Assessment-Plan notes; select all that apply).
| ○ Vocabulary development | ○ Pronoun development |
| ○ Listening comprehension | ○ Verb tense |
| ○ Morphological development | ○ Preposition |
| ○ Sentence formulation | ○ Verb vocabulary |
| ○ Syntax development | ○ Question formulation |
| ○ Narrative development | ○ Complex sentences |
Collaboration
Imagine all of the children on your caseload. How much collaboration do you, on average, have with other professionals? Select the number rating that best applies to each category.
| Daily | 2–3 times per week | Weekly | Monthly | Never | Not Applicable | |
|---|---|---|---|---|---|---|
| Classroom teacher | ||||||
| Reading teacher | ||||||
| Special education teacher | ||||||
| Psychologist |
Note. Participants selected one for each row. For example, they would select daily for “Classroom teacher.”
Imagine all of the children on your caseload. How much collaboration do you, on average, have with the caregiver(s)? Daily, 2–3 times per week, weekly, monthly, never
Discharge/Monitoring
-
When are students discharged from speech-language services? Select all that apply
The child is performing at grade level in the classroom
The child has met the short- and long-term goals/objective
Assessment measures indicate the child is now in normal performance range for their age
The child has not progressed with intervention services over time
Insurance is no longer reimbursing for services
Are students monitored after they are discharged/dismissed/determined ineligible for services? Yes, No
Vignette Decision-Making (Selin et al., 2019)
Directions: The next section of the survey will present six case scenarios and ask you questions regarding recommendation of services and/or intervention. The number of questions presented per scenario is dependent on your responses. Please assume that each student (a) is a monolingual speaker of standard American English, (b) has passed a hearing screening, (c) does not use AAC, (d) has no history of nonverbal cognitive impairment, (e) does not have a diagnosed syndrome, such as Down Syndrome, and (f) does not have autism. Note: For this section, please use your professional judgement in order to identify whether the presented cases warrant services and/or intervention in an optimal and ideal setting. Disregard particular workplace protocols/policies for eligibility criteria and resource or time constraints for the purposes of these questions.
The first question displayed after reading the vignettes were analyzed in the current study; please see Selin et al., 2019, for additional questions.
In your clinical opinion, would you recommend [vignette's initials] for speech-language pathology services? Yes (i.e., direct intervention, teacher/caregiver consultation), No
Vignette A: S. C. is a 7-year 5-month-old boy who was referred for speech-language pathology services by his teacher. His teacher stated that S. C. got along well with his classmates although he was somewhat shy. His teacher also stated that S.C. had a quiet voice and that he frequently mumbled when he answered questions. S. C. demonstrated some difficult with past tense (e.g., inconsistently producing “digged” for dug or “ride” for rode). When assessed, S. C. scored a standard score of 78 on the Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4), a language quotient with 5 standard scores of 75 on the Test of Language Development–Primary: Fourth Edition (TOLD-P:4), and a score at the 33rd percentile on the Goldman-Fristoe Test of Articulation–Second Edition (GFTA-2). Assume that a nonverbal intelligence quotient is unavailable at this time. Additionally, a 200-utterance language sample was elicited and S. C.'s MLU was calculated to be 4.29.
Vignette B: C. C. is an 8-year 10-month-old girl who has received speech-language services for the past 3 years. Currently, C. C. sees the school speech-language pathologist two times per week for 30 minutes per session. During the sessions, C. C. works on pronoun use (e.g., she/her, she/he), present progressive –ing, and regular and irregular past tense verb forms. C. C.'s classroom teacher reports that she is very outgoing and has little difficulty making friends. C. C. is a hard worker and performs at or above grade level. C. C. does, however, demonstrate some difficulty with verb use, substituting a regular past tense form for an irregular past tense form on occasion (e.g., swimmed/swam). C. C.'s latest assessment indicates a Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4) standard score of 85, a Test of Language Development–Primary: Fourth Edition (TOLD-P:4) standard score of 85, and a 200-utterance MLU of 5.52.
Vignette C: E. C.'s teacher is concerned about his communication skills. He reports that E. C. avoids communicating with his classmates. When he does communicate, he “rambles” and has little sense of “getting to the point.” In addition, he lacks the skills for requesting, disagreeing, and role-playing that other second graders normally demonstrate. When E. C. is assessed by the speech-language pathologist, he receives a total language score of 69 on the Clinical Evaluation of Language Fundamentals– Fourth Edition (CELF-4) and a standard score of 71 on the Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4). His score on the Goldman-Fristoe Test of Articulation–Second Edition (GFTA-2) is in the 62nd percentile.
Vignette D: The parents of B. D., a girl in first grade, have requested that she receive a speech and language evaluation. They report that B. D. occasionally omits the –s in plural constructions, −s in third-person singular verb constructions, −ed in past tense verb constructions, and forms of auxiliary verb BE. On the language assessment, B. D. scores a Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4) standard score of 73 and a Clinical Evaluation of Language Fundamentals–Fourth Edition (CELF-4) total language standard score of 72. In addition, B. D.'s school records indicate that she received a standard score of 104 on the Columbia Mental Maturity Scales (nonverbal intelligence assessment), which revealed normal kindergarten readiness abilities, and normal physical development and gross/fine motor development.
Vignette E: A. P. is a 7-year 4-month-old boy who has received speech-language services for the past two years. Currently, A. P. sees the school speech-language pathologist once a week for 30 minutes per session. During the sessions, A. P. works on maintaining conversation with peers, using appropriate eye contact and body orientation, and improving awareness of his communication partner's needs (e.g., awareness of nonverbal gestures). During his IEP meeting this year, the speech-language pathologist reported that A. P. has made progress in his conversational exchanges with classmates as well as teachers. Additionally, he demonstrates a standard score of 85 on the Peabody Picture Vocabulary Test–Fourth Edition (PPVT-4), a total language score of 84 on the Test of Language Development–Primary: Fourth Edition (TOLD-P:4), and an MLU of 5.67 (based on a 100-utterance sample).
Vignette F: P. L. is an 11-year 7-month-old girl (fifth grade) who is receiving special education services through an IEP in math, language arts, physical/life sciences, and social sciences. In the classroom, her teachers report that she is quiet and works hard on her schoolwork. Her abilities to recognize and decode words are within normal limits. However, she exhibits difficulties with learning new material from texts. Teachers report that her responses on exams, conversational language, and written composition tend to be short, simple, and vague with missing curricular vocabulary. She demonstrates a standard score of 83 on the Clinical Evaluation Language Fundamentals–Fifth Edition (CELF-5) and a standard score of 102 on the Wechsler Intelligence Scales for Children–Fifth Edition (WISC-5).
Caseload Composition
How many hours per week do you work as a speech-language pathologist? 1–20 hours, 21–40 hours, 40+ hours
How many students are on your caseload? 1–10, 11–20, 21–40, 40+
Are you provided with assistance from a paraprofessional/speech-language pathology assistant? Yes, No
Which of the following groups do you currently serve (select all that apply)? Infants and toddlers, preschool, elementary school, middle/junior high school, high school
What percentage of children on your current caseload receive services for the following?
| None | 1–33% | 34–66% | 67–100% | |
|---|---|---|---|---|
| Speech impairment only | ||||
| Language impairment only | ||||
| Speech and language impairment |
Note. Participants selected one for each row. For example, they would select 1–33% for “speech impairment only.”
What percentage of children on your current caseload have a medical/educational diagnosis (e.g., autism spectrum disorder, cerebral palsy, ADHD, etc.)? None, 1–33%, 34–66%, 67–100%
Please consider the following definition of “specific language impairment” (SLI): “An unexpected failure to develop language normally despite intact general cognitive, emotion, and hearing abilities and typical opportunities to acquire language” (Archibald & Joanisse, 2009, p. 899). How many children on your caseload fit this description of SLI? [Free text response]
Funding Statement
This work was supported by Grants T32DC000052 (awarded to Mabel L. Rice, director), R01DC001803 (awarded to Mabel L. Rice, principal investigator), T32HD101390 (awarded to Jennie Noll and Yo Jackson, co-directors), and T32DC017703 (awarded to Inge-Marie Eigsti and Emily Myers, co-directors).
References
- American Speech-Language-Hearing Association. (2004). Preferred practice patterns for the profession of speech-language pathology. http://www.asha.org/policy [PubMed]
- American Speech-Language-Hearing Association. (2016). Scope of practice in speech-language pathology. https://www.asha.org/policy
- American Speech-Language-Hearing Association. (2019). Speech-language pathology health care 2019 survey. https://www.asha.org/
- American Speech-Language-Hearing Association. (2020). Supply and demand resource list for speech-language pathologists. https://www.asha.org
- American Speech-Language-Hearing Association. (2021). Profile of American Speech-Language-Hearing Association members and affiliates, year-end 2020. https://www.asha.org
- Amir, R. , Jones, S. E. , Frankel, D. , & Fritzsch, J. (2021). Job satisfaction of school-based speech-language pathologists in New York state as a function of workplace features. Perspectives of the ASHA Special Interest Groups, 6(2), 470–484. https://doi.org/10.1044/2020_persp-20-00196 [Google Scholar]
- Archibald, L. M. , & Joanisse, M. F. (2009). On the sensitivity and specificity of nonword repetition and sentence recall to language and memory impairments in children. Journal of Speech, Language, and Hearing Research, 52(4), 899–914. https://doi.org/10.1044/1092-4388(2009/08-0099) [DOI] [PubMed] [Google Scholar]
- Betz, S. K. , Eickhoff, J. R. , & Sullivan, S. F. (2013). Factors influencing the selection of standardized tests for the diagnosis of specific language impairment. Language, Speech, and Hearing Services in Schools, 44(2), 133–146. https://doi.org/10.1044/0161-1461(2012/12-0093) [DOI] [PubMed] [Google Scholar]
- Bishop, D. V. M. , Snowling, M. J. , Thompson, P. A. , Greenhalgh, T. , & CATALISE Consortium. (2016). CATALISE: A multinational and multidisciplinary Delphi consensus study. Identifying language impairments in children. PLOS ONE, 11(7). e0158753. https://doi.org/10.1371/journal.pone.0158753 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bishop, D. V. M. , Snowling, M. J. , Thompson, P. A. , Greenhalgh, T. , & CATALISE Consortium. (2017). Phase 2 of CATALISE: A multinational and multidisciplinary Delphi consensus study of problems with language development: Terminology. The Journal of Child Psychology and Psychiatry, 58(10), 1068–1080. https://doi.org/10.1111/jcpp.12721 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Braithwaite, J. , Churruca, K. , Long, J. C. , Ellis, L. A. , & Herkes, J. (2018). When complexity science meets implementation science: A theoretical and empirical analysis of systems change. BMC Medicine, 16(1), 63. https://doi.org/10.1186/s12916-018-1057-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brandel, J. , & Frome Loeb, D. (2011). Program intensity and service delivery models in the schools: SLP survey results. Language, Speech, and Hearing Services in Schools, 42(4), 461–490. https://doi.org/10.1044/0161-1461(2011/10-0019 [DOI] [PubMed] [Google Scholar]
- Brownell, R. (2000). Receptive One-Word Picture Vocabulary Test–Third Edition. Academic Therapy Publications. [Google Scholar]
- Brownell, R. (2010). Receptive One-Word Picture Vocabulary Test–Fourth Edition. Academic Therapy Publications. [Google Scholar]
- Brownlie, E. B. , Jabbar, A. , Beitchman, J. , Vida, R. , & Atkinson, L. (2007). Language impairment and sexual assault of girls and women: Findings from a community sample. Journal of Abnormal Child Psychology, 35(4), 618–626. https://doi.org/10.1007/s10802-007-9117-4 [DOI] [PubMed] [Google Scholar]
- Burns, K. E. A. , Duffett, M. , Kho, M. E. , Meade, M. O. , Adhikari, N. K. J. , Sinuff, T. , Cook, D. J. , & Academy Group. (2008). A guide for the design and conduct of self-administered surveys of clinicians. Canadian Medical Association Journal, 179(3), 245–252. https://doi.org/10.1503/cmaj.080372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carrow-Woolfolk, E. (1995). Oral and Written Language Scales–First Edition. AGS. [Google Scholar]
- Carrow-Woolfolk, E. (1999). Comprehensive Assessment of Spoken Language–First Edition. AGS. [Google Scholar]
- Carrow-Woolfolk, E. (2011). Oral and Written Language Scales–Second Edition. Pearson. [Google Scholar]
- Carrow-Woolfolk, E. (2017). Comprehensive Assessment of Spoken Language–Second Edition. AGS. [Google Scholar]
- Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Academic Press. [Google Scholar]
- Conti-Ramsden, G. , & Botting, N. (2008). Emotional health in adolescents with and without a history of specific language impairment (SLI). The Journal of Child Psychology and Psychiatry, 49(5), 516–525. https://doi.org/10.1111/j.1469-7610.2007.01858.x [DOI] [PubMed] [Google Scholar]
- Dillman, D. A. , Smyth, J. D. , & Christian, L. M. (2014). Internet, phone, mail, and mixed-mode surveys: The tailored design method. Wiley. [Google Scholar]
- Dowden, P. , Alarcon, N. , Vollan, T. , Cumley, G. D. , Kuehn, C. M. , & Amtmann, D. (2006). Survey of SLP caseloads in Washington state schools: Implications and strategies for action. Language, Speech, and Hearing Services in Schools, 37(2), 104–117. https://doi.org/10.1044/0161-1461(2006/013) [DOI] [PubMed] [Google Scholar]
- Dunn, D. M. (2018). Peabody Picture Vocabulary Test–Fifth Edition. Pearson. [Google Scholar]
- Dunn, L. M. , & Dunn, D. M. (2007). Peabody Picture Vocabulary Test–Fourth Edition. Pearson. [Google Scholar]
- Eysenbach, G. (2004). Improving the quality of web surveys: The checklist for reporting results of internet e-surveys (CHERRIES). Journal of Medical Internet Research, 6(3). e34. https://doi.org/10.2196/jmir.6.3.e34 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fulcher-Rood, K. , Castilla-Earls, A. P. , & Higginbotham, J. (2018). School-based speech-language pathologists' perspectives on diagnostic decision making. American Journal of Speech-Language Pathology, 27(2), 796–812. https://doi.org/10.1044/2018_AJSLP-16-0121 [DOI] [PubMed] [Google Scholar]
- Gertner, B. L. , Rice, M. L. , & Hadley, P. A. (1994). Influence of communicative competence on peer preferences in a preschool classroom. Journal of Speech and Hearing Research, 37(4), 913–923. https://doi.org/10.1044/jshr.3704.913 [DOI] [PubMed] [Google Scholar]
- Green, L. (2020). The specific language impairment/developmental language disorders forum: Fostering a discussion of terminology. Perspectives of the ASHA Special Interest Groups, 5(1), 3–5. https://doi.org/10.1044/2019_PERSP-19-00184 [Google Scholar]
- Hutchins, T. L. , Howard, M. , Prelock, P. A. , & Belin, G. (2009). Retention of school-based SLPs: Relationships among caseload size, workload satisfaction, job satisfaction, and best practice. Communication Disorders Quarterly, 31(3), 139–154. https://doi.org/10.1177/1525740109336870 [Google Scholar]
- IBM Corp. (2020). IBM SPSS statistics for Windows (Version 27.0).
- Individuals with Disabilities Education Act, Act 20 U.S.C. § 1400. (2004).
- Johnson, C. J. , Beitchman, J. H. , Young, A. , Escolar, M. , Atkinson, L. , Wilson, B. , Brownlie, E. B. , Douglas, L. , Taback, N. , Lam, I. , & Wang, M. (1999). Fourteen-year follow-up of children with and without speech/language impairments: Speech/language stability and outcomes. Journal of Speech, Language, and Hearing Research, 42(3), 744–760. https://doi.org/10.1044/jslhr.4203.744 [DOI] [PubMed] [Google Scholar]
- Kansas Administrative Regulations. (2008). 91–40-1(ooo).
- Kamhi, A. G. (1994). Toward a theory of clinical expertise in speech-language pathology. Language, Speech, and Hearing Services in Schools, 25(2), 115–118. https://doi.org/10.1044/0161-1461.2502.115 [Google Scholar]
- Kamhi, A. G. (1995). Defining, developing, and maintaining clinical expertise. Language, Speech, and Hearing Services in Schools, 26(4), 353–356. https://doi.org/10.1044/0161-1461.2604.353 [Google Scholar]
- Kritikos, E. P. (2003). Speech-language pathologists' beliefs about language assessment of bilingual/bicultural individuals. American Journal of Speech-Language Pathology, 12(1), 73–91. https://doi.org/10.1044/1058-0360(2003/054) [DOI] [PubMed] [Google Scholar]
- Leonard, M. V. , Plexico, L. W. , Plumb, A. M. , & Sandage, M. J. (2016). Emerging practice preferences of speech-language pathology students. Contemporary Issues in Communication Science and Disorders, 43(Fall), 285–298. https://doi.org/10.1044/cicsd_43_F_285 [Google Scholar]
- Martin, N. A. , & Brownell, R. (2000). Expressive One-Word Picture Vocabulary Test–Third Edition. Academic Therapy Publications. [Google Scholar]
- Martin, N. A. , & Brownell, R. (2011). Expressive One-Word Picture Vocabulary Test–Fourth Edition. Academic Therapy Publications. [Google Scholar]
- McGregor, K. K. , Goffman, L. , Van Horne, A. O. , Hogan, T. P. , & Finestack, L. H. (2020). Developmental language disorder: Applications for advocacy, research, and clinical service. Perspectives of the ASHA Special Interest Groups, 5(1), 38–46. https://doi.org/10.1044/2019_persp-19-00083 [Google Scholar]
- Missouri State Board of Education Special Education Regulations. (2020). Regulation III.B.
- Murza, K. A. , & Ehren, B. J. (2020). Considering the language disorder label debate from a school speech-language pathology lens. Perspectives of the ASHA Special Interest Groups, 5(1), 47–54. https://doi.org/10.1044/2019_PERSP-19-00077 [Google Scholar]
- Newcomer, P. L. , & Hammill, D. D. (2008). Test of Language Development–Primary: Fourth Edition. Pro-Ed. [Google Scholar]
- Newcomer, P. L. , & Hammill, D. D. (2019). Test of Language Development–Primary: Fifth Edition. Pro-Ed. [Google Scholar]
- Parag, Y. , & Janda, K. B. (2014). More than filler: Middle actors and socio-technical change in the energy system from the “middle out”. Energy Research & Social Science, 3, 102–112. https://doi.org/10.1016/j.erss.2014.07.011 [Google Scholar]
- Pfeiffer, D. L. , Pavelko, S. L. , Hahs-Vaughn, D. L. , & Dudding, C. C. (2019). A national survey of speech-language pathologists' engagement in interprofessional collaborative practice in schools: Identifying predictive factors and barriers to implementation. Language, Speech, and Hearing Services in Schools, 50(4), 639–655. https://doi.org/10.1044/2019_LSHSS-18-0100 [DOI] [PubMed] [Google Scholar]
- Qualtrics. (2005). https://www.qualtrics.com
- Records, N. L. , & Tomblin, J. B. (1994). Clinical decision making: Describing the decision rules of practicing speech-language pathologists. Journal of Speech and Hearing Research, 37(1), 144–156. https://doi.org/10.1044/jshr.3701.144 [PubMed] [Google Scholar]
- Records, N. L. , Tomblin, J. B. , & Freese, P. R. (1992). The quality of life of young adults with histories of specific language impairment. American Journal of Speech-Language Pathology, 1(2), 44–53. https://doi.org/10.1044/1058-0360.0102.44 [Google Scholar]
- Redmond, S. M. (2011). Peer victimization among students with specific language impairment, attention-deficit/hyperactivity disorder, and typical development. Language, Speech, and Hearing Services in Schools, 42(4), 520–535. https://doi.org/10.1044/0161-1461(2011/10-0078) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rudolph, J. M. (2017). Case history risk factors for specific language impairment: A systematic review and meta-analysis. American Journal of Speech-Language Pathology, 26(3), 991–1010. https://doi.org/10.1044/2016_AJSLP-15-0181 [DOI] [PubMed] [Google Scholar]
- Rice, M. L. (2012). Toward epigenetic and gene regulation models of specific language impairment: Looking for links among growth, genes, and impairments. Journal of Neurodevelopmental Disorders, 4(27), Article No. 27. https://doi.org/10.1186/1866-1955-4-27 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rice, M. L. (2020). Clinical lessons from studies of children with specific language impairment. Perspectives of the ASHA Special Interest Group, 5(1), 12–29. https://doi.org/10.1044/2019_PERSP-19-00011 [Google Scholar]
- Rice, M. L. , & Hoffman, L. (2015). Predicting vocabulary growth in children with and without specific language impairment: A longitudinal study from 2;6 to 21 years of age. Journal of Speech, Language, and Hearing Research, 58(2), 345–359. https://doi.org/10.1044/2015_JSLHR-L-14-0150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rice, M. L. , & Wexler, K. (2001). Rice/Wexler Test of Early Grammatical Impairment. University of Kansas. https://cldp.ku.edu/rice-wexler-tegi [Google Scholar]
- Selin, C. M. , Rice, M. L. , Girolamo, T. , & Wang, C. J. (2019). Speech-language pathologists' clinical decision making for children with specific language impairment. Language, Speech, and Hearing Services in Schools, 50(2), 283–307. https://doi.org/10.1044/2018_LSHSS-18-0017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Snow, P. C. , & Powell, M. B. (2011). Oral language competence in incarcerated young offenders: Links with offending severity. International Journal of Speech-Language Pathology, 13(6), 480–489. https://doi.org/10.3109/17549507.2011.578661 [DOI] [PubMed] [Google Scholar]
- Spaulding, T. J. , Plante, E. , & Farinella, K. A. (2006). Eligibility criteria for language impairment: Is the low end of normal always appropriate. Language, Speech, and Hearing Services in Schools, 37(1), 61–72. https://doi.org/10.1044/0161-1461(2006/007) [DOI] [PubMed] [Google Scholar]
- Tomblin, J. B. , Records, N. L. , Buckwalter, P. , Zhang, X. , Smith, E. , & O'Brien, M. (1997). Prevalence of specific language impairment in kindergarten children. Journal of Speech, Language, and Hearing Research, 40(6), 1245–1260. https://doi.org/10.1044/jslhr.4006.1245 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tomblin, J. B. , Records, N. L. , & Zhang, X. (1996). A system for the diagnosis of specific language impairment in kindergarten children. Journal of Speech and Hearing Research, 39(6), 1284–1294. https://doi.org/10.1044/jshr.3906.1284 [DOI] [PubMed] [Google Scholar]
- Victorino, K. R. , & Magaldi, N. M. (2020). Navigating terminology and promoting evidence-based practice for child language disorders: Challenges and solutions for educators. Communication Disorders Quarterly, 42(1), 58–64. https://doi.org/10.1177/1525740119857701 [Google Scholar]
- Volkers, N. (2018a). Diverging views on language disorders: Researchers debate whether the label “developmental language disorder” should replace “specific language impairment”. The ASHA Leader, 23(12), 44–53. https://doi.org/10.1044/leader.FTR1.23122018.44 [Google Scholar]
- Volkers, N. (2018b). Does an SLI label really restrict services? The ASHA Leader, 23(12), 54–61. https://doi.org/10.1044/leader.FTR2.23122018.54 [Google Scholar]
- Wiig, E. H. , Semel, E. , & Secord, W. A. (2003). Clinical Evaluation of Language Fundamentals–Fourth Edition (CELF-4). Pearson. [Google Scholar]
- Wiig, E. H. , Semel, E. , & Secord, W. A. (2013). Clinical Evaluation Language Fundamentals–Fifth Edition (CELF-5). Pearson. [Google Scholar]
- Williams, K. T. (2007). Expressive Vocabulary Test–Second Edition. Pearson. [Google Scholar]
- Williams, K. T. (2018). Expressive Vocabulary Test–Third Edition. Pearson. [Google Scholar]
- World Health Organization. (2016). International statistical classification of disease and related health problems (10th ed.). https://icd.who.int/browse10/2016/en
- World Health Organization. (2019). International statistical classification of diseases and related health problems (11th ed.). https://icd.who.int/
- Zhang, X. , & Tomblin, J. B. (2000). The association of intervention receipt with speech-language profiles and social-demographic variables. American Journal of Speech-Language Pathology, 9(4), 345–357. https://doi.org/10.1044/1058-0360.0904.345 [Google Scholar]
- Zimmerman, I. L. , Steiner, V. G. , & Pond, R. E. (2002). Preschool Language Scales–Fourth Edition. Pearson. [Google Scholar]
- Zimmerman, I. L. , Steiner, V. G. , & Pond, R. E. (2011). Preschool Language Scales–Fifth Edition. Pearson. [Google Scholar]