A predominantly descriptive survey design was used in this study. Given the absence of work addressing this specific topic, a descriptive method was chosen to allow for a more comprehensive understanding of the research question (Hafsa 2019; Swatzell & Jennings 2007). The questionnaire obtained quantitative data using closed-ended questions, supplemented by a few free-text responses and one open-ended question, rendering qualitative data, aimed at providing a comprehensive understanding of the subject matter (Hafsa 2019).

Data were collected via an electronic questionnaire circulated using Qualtrics XM. The online format was selected to allow for a larger reach across professionals in the geographical region. All individuals providing communication and swallowing rehabilitation services were invited to participate.

Participants were required to have a university degree (Bachelor’s or Master’s) in speech-language therapy, or have a university degree (Bachelor’s or Master’s) in a closely related field (education, special education, psychology, rehabilitation sciences), followed by SLT training. This broader criterion was selected to capture the diverse workforce of professionals providing communication and swallowing services due to service gaps. Participants were required to be rendering in-person SLT services, astele-rehabilitation was not the focus of this study. To ensure up-to-date and accurate information, all participants had to be currently practising and have a minimum of 6 months of experience providing SLT services. Because the questionnaire was in English, all participants also had to be proficient in English and have internet access to complete the questionnaire online.

The questionnaire was self-developed, drawing from previously published literature (Wylie et al. 2016, 2018) and was modified to capture information from SLTs who only practice in Kenya with appropriate justifications (Online Appendix 1). Permission to use and adapt the questionnaire was requested and obtained from the authors. To improve face and content validity, the questionnaire was pre-tested with three SLTs working in a similar region, and participants were asked to provide feedback on the content, wording and flow of the questions, multiple-choice answers where applicable, along with ease of filling the questionnaire and acceptable time taken to fill the survey. Based on their feedback, changes were made to the questionnaire regarding wording and answer formats.

In addition, reporting of the online survey followed the CHERRIES (Checklist for Reporting Results of Internet E-Surveys) (Eysenbach 2004) guidelines to ensure transparency in survey design, administration and reporting. The completed checklist is provided as Online Appendix 2. The questionnaire comprised of 30 closed-ended questions, including yes or no format questions, multiple-choice items, Likert scale ratings and opportunities to add free text. The questionnaire concluded with one open-ended question that explored participants’ views about the current services and support available in the country. The survey took approximately 20 min to complete and was accessible for six weeks on the Qualtrics XM software platform. A copy of the questionnaire is available in Online Appendix 3.

Participants were selected via purposive sampling, from the ASLTK group in Kenya who forwarded the cover letter and electronic link to their official membership list. Additionally, snowball sampling was also adopted through social media platforms (such as WhatsApp and Facebook) and professional networks where SLTs were encouraged to disseminate the questionnaire. This approach was necessary due to the absence of a national registry of SLT providers and the dispersion of professionals across diverse work settings. Although ASLTK reported 28 SLTs, this figure only represents SLTs that are affiliated with this group and does not represent the total number of SLTs in Kenya. Potential participants received an introduction to the survey via email, and remaining distributions were via social media channels.

A total of 67 responses were received, of which 21 responses had to be eliminated for not meeting the inclusion criteria (n = 6) and for not completing the survey (n = 15). Of the six respondents who did not meet the inclusion criteria, two individuals were not practising currently, one provided service only via tele-therapy, and the remaining three did not have the correct qualifications. A total sample of 46 responses was then analysed and reported on.

All included participants were required to be practising SLTs. Most participants were female, within the age range of 21 years to 40 years. Most participants were based within Nairobi county, the country’s capital. Further demographic details are reported in the Results section.

The quantitative data collected were on Qualtrics XM and analysed using Statistical Package for the Social Sciences (SPSS) version 30.0.0. Because all variables were ordinal and not continuous, the nonparametric Spearman correlation (rs) was used to identify statistically significant correlations between variables rather than the parametric Pearson correlation (r). For the strength of the relationship, the recommendations by Téllez, García and Corral-Verdugo (2015) were used: rs < 0.10 (weak), 0.10 < rs < 0.30 (weak-to-moderate), rs = 0.30 (moderate), 0.30 < rs < 0.50 (moderate-to-strong) and rs ≥ 0.50 (strong).

Given the exploratory and descriptive nature of this study, corrections for multiple comparisons, such as Bonferroni, were not applied. These adjustments make the interpretation of results contingent on the number of tests performed, introducing arbitrary variability – two authors examining the same hypothesis could reach different conclusions solely based on how many comparisons they conduct (Perneger 1998). Furthermore, Bonferroni corrections are known to be overly conservative, dramatically increasing the risk of Type II errors and potentially obscuring meaningful associations that warrant further investigation (Barnett et al. 2022). Owing to the investigative focus of this study, all correlation analyses are reported with exact p-values and interpreted cautiously rather than mechanically adjusted.

For the free text responses and one open-ended question, qualitative content analysis was applied using open thematic coding. Collected responses were transcribed into a spreadsheet and the raw data were first familiarised by the first author (Mayring 2021). Since most responses were in long paragraphs, shorter key phrases were categorised manually, and repeating ideas were sorted into themes (Puppis 2019). These were then grouped into broader categories through an iterative process. Given the minor volume of qualitative data, coding was only conducted by a single individual (first author), and inter-coder reliability was not undertaken. Representative quotations were selected to illustrate themes based on their clarity and how strongly they replicated participant experiences. No qualitiative analysis software was used.

To enhance the validity and reliability, the questionnaire was pre-tested before being sent out. This ensured the questionnaire questions were unambiguous, enhancing face validity. Content validity was also captured by these peers during the pre-test, ensuring the questionnaire did meet all aspects aptly (Taherdoost 2016). More details on the pre-test are given in the Material section. All participants received identical questionnaires, derived from two previously published studies (Wylie et al. 2016, 2018). Results from the mentioned studies are aligned with the current literature exploring speech-language therapy practices in the region, though no studies have been carried out specifically in Kenya (Ndiema et al. 2024). Since the questionnaire predominantly consisted of closed-ended questions, this increased the validity by aiding the criteria for suitable responses to the questions. Since the questionnaire was online and anonymous, there was no pressure on the participant, reducing researcher bias (Brink, Van der Walt & Van Rensburg 2018). The open-ended questions effectively allowed for a deeper analysis of the findings, making them more robust (Swatzell & Jennings 2007).

Institutional ethical clearance was obtained from the Faculty of Humanities’ Research Ethics Committee at the University of Pretoria (reference number HUM025/1124). Voluntary participation was requested from participants. Participants had the right to discontinue the questionnaire at any point before submitting their responses. No incentives were given to encourage participation. The electronic link was accompanied by a cover letter describing the aim of the study, and an informed consent declaration was obtained before beginning the study. To ensure confidentiality of data collected, identifying features such as names of workplace settings or contact details were not collected in the survey. IP addresses were deleted prior to data analysis.

Thirty-six (78.4%) of the 46 participants were female. Almost half (n = 22, 47.8%) of the participants were between 31 years old and 40 years old, followed by 10 (21.7%) participants between 21 years old and 30 years old (Table 1). Most participants (n = 36, 78.3%) were working in Nairobi county and city, Kenya’s capital, with several participants indicating they worked in more than one county (Table 1). Fourteen (31.1%) spoke English as their first language, and all participants reported proficiency in at least one other language. Numerous participants (n = 24, 52.2%) indicated they did not typically speak the same language as their clients. Participants listed most of their clients as bilingual and multilingual (Table 1).

The majority of participants (n = 43, 93.5%) reported having formal training in the field of speech-language pathology and/or therapy, and nearly half (n = 21, 45.7%) were registered and licensed with the KHPOA. It is to be noted that individual licensing was self-reported though not verified, as a publicly accessible record is not currently available. Most participants held a Master’s degree (n = 27, 58.7%) or Bachelor’s degree (n = 12, 26.1%), while a few were still awaiting their Master’s certification (n = 3, 6.5%). More than half of the participants (n = 26, 56.5%) were employed full-time, primarily in schools (n = 18, 39.1%) and private practice clinics (n = 12, 26.1%), among a variety of other settings like home health and hospitals (Table 2). More than half of participants (n = 26, 56.5%) self-identified as working full time. However, many of these participants reported providing services across multiple settings and counties, reflecting a consultancy-based employment rather than single-site positions.

Approximately three-quarters of the participants (n = 35, 76.1%) also had prior training in other fields; for example, nine of the 35 (25.7%) had prior training in special needs education. Experience levels varied, with half of the participants having between 1 year and 5 years of experience, and eight participants having between 6 years and 10 years of experience. The remaining participants had over 10 years of experience. Participants worked in various settings, and some also had prior training in other fields (Table 2).

Almost half of the participants (n = 22, 47.8%) saw up to six clients a day. Approximately one-quarter (n = 12, 26.1%) had a caseload of up to three clients per day. Two participants (4.3%) reported seeing 10 to 15 clients or more daily. Figure 1 shows the frequency with which client populations across various ages were seen. For example, more than half (n = 31, 67.4%) of the group reported they always saw pre-school-aged children. The majority of participants reported working with paediatric clients, with approximately two-thirds (n = 31, 67.4%) always seeing pre-schoolers and around a half (n = 24, 52.2%) always working with school-aged children. Inferential statistics via the Spearman test revealed there was no statistically significant correlation between the number of years of experience the therapist had and the age of clients they serve, as the value of p was larger than 0.05.

The participants’ caseloads are shown in Table 3. The main paediatric services rendered were to children with autism (n = 43, 93.5%), articulation or phonological disorders (n = 39, 84.8%), need for early communication intervention (n = 37, 80.4%) and fluency disorders (n = 31, 67.4%). Service delivery to adults included management of aphasia (n = 29, 63.0%), dysarthria (n = 20, 43.5%), stroke-related speech and communication difficulties (n = 25, 54.3%), fluency disorders (n = 24, 52.2%) and voice disorders (n = 22, 47.8%). Based on their training, support and professional development, approximately half of the participants (n = 23, 52.3%) felt confident in taking on the caseloads that they indicated. A third of participants (n = 14, 31.8%) felt very confident, whereas a smaller group (n = 6, 13.6%) felt somewhat confident. Two participants (4.3%) did not respond.

For these ordinal values, the nonparametric Spearman correlation was conducted to identify any correlation between the number of years of experience and the level of confidence the participant had (rs = 0.20, p = 0.201). There was no statistically significant correlation seen between these variables. The Spearman correlation was also applied to explore any correlation between the level of confidence of the participant and the types of caseloads they served. There was a statistically significant moderate positive correlation between the level of confidence for a dysphagia caseload (rs = 0.30, p = 0.048) and a moderate-to-strong positive correlation for cerebral palsy (rs = 0.35, p = 0.019). There was no statistically significant correlation between the level of confidence and any other caseloads (rs = 0.20, p = 0.200), since the p-value was greater than 0.05.

For the number of hours per week, including direct and indirect speech-language therapy services (e.g. planning, administration, report writing, supervision, educating and collaborating with team members), nearly one-third of participants (n = 14, 30.4%) reported working between 25 and 40 h weekly, while just over a quarter (n = 12, 26.1%) reported working more than 40 h per week. Eleven participants (23.9%) reported working between 15 and 25 h a week, and the remaining (n = 9, 19.6%) reported working less than 15 h weekly.

Most participants (n = 41, 89.0%) reported working in departments with only one to three SLTs. The remaining five participants (10.9%) indicated there were four to six SLTs at the workplace. The majority of participants (n = 30, 65.2%) felt that the number of SLTs in their workplace was insufficient to cover the current caseload demands. Participants cited contributing factors of high caseloads, time constraints and low number of SLTs, as attested to in the following statements:

In some areas, only one SLT in the surrounding geographical region was reported:

Conversely, more than a third (n = 16, 34.8%) reported having an adequate number of SLTs for their caseloads. The reasons reported for this were a lower caseload and an increased number of providers in the capital city.

Primary work functions for SLTs included assessment (n = 41, 89.1%) and treatment (n = 44, 95.7%) of clients. (Table 4). Professional collaboration included working with occupational therapists (n = 43, 93.5%), teachers (n = 33, 71.7%), physiotherapists (n = 27, 58.7%), medical practitioners (n = 25, 54.3%), audiologists (n = 24, 52.2%), psychologists (n = 20, 43.5%), dieticians (n = 17, 37.0%), nurses (n = 14, 30.4%), social workers (n = 11, 23.9%) and behavioural technicians (n = 1, 2.2%). Over half of the participants (n = 25, 54.3%) also rendered group therapy as part of their services, where two participants (4.6%) indicated mainly facilitating groups for child language therapy.

Regarding SLT service provision, 19 (41.3%) participants reported having in-house clinical protocols, such as standard operating protocols, assessment and screening tools. In contrast, 15 participants (32.6%) did not have any workplace guidelines, and the remaining were unsure (n = 9, 19.6%). Some participants (n = 3, 6.5%) noted that protocols did not apply to their specific setting.

The majority of participants (n = 37, 84.0%) agreed that Kenyan SLTs required national clinical guidelines for the assessment and treatment of specific disorders. In the free-text section of this question, multiple themes were identified: the need for standard, evidence-based guidelines, culturally and linguistically appropriate assessment protocols, and improved practice protocols. Barriers to service delivery and the need for affordable continuing professional educational opportunities were also expanded on.

Participants identified the need for standard guidelines for the profession, as well as the need for more specific, evidence-based guidelines outlining specific disorders:

Participants felt that guidelines would be beneficial, especially to guide the newly qualified professionals in the country:

A need for culturally and linguistically appropriate assessment protocols was identified:

Participants also recognised the need for improved standards of practice and for curbing unqualified professionals from performing the SLTs scope:

The need for uniformity for service provision amongst SLTs was also identified:

The requirement for structure and quality of care also emerged:

However, three participants were of the opinion that current local guidelines needed refinement rather than adopting ready guidelines from developed countries, due to a varied population:

The key barriers to service delivery participants recognised were as follows: lack of resources (n = 32, 69.6%), a high caseload (n = 24, 52.2%) and travelling across different geographical locations (n = 21, 45.7%) as indicated in Table 5. In an open-ended section that probed further, one participant reported that:

Speech-language therapists reported a lack of an adequateworkforce for the high caseload of clients. An uneven urbandistribution was seen:

Access to services was limited for clients due to financial factors:

Speech-language therapists also reported:

All participants (n = 46, 100%) reported regularly attending continued professional development (CPD) and training to improve their skills and knowledge. Participants preferred webinars (n = 39, 84.8%), short courses (n = 37, 80.4%) and in-person workshops (n = 30, 65.2%). Participants also noted self-directed study, specific treatment approaches or technique courses, professional groups and mentorship with colleagues as part of their development. The increased access and opportunities for CPD provided by groups such as the ASLTK were acknowledged by some participants (n = 12). However, access to these CPDs was limited by time and funding:

There was a call for more specialised and apt continuing professional development opportunities:

Culturally and linguistically appropriate learning opportunities were also mentioned:

Responses to the open-ended question explored views regarding current services and support available in the country: specific needs, continued education opportunities, support from professional bodies, and opportunities to collaborate and learn. Many participants (n = 6) recognised the growth of the SLT field in Kenya over the last few years:

There has been an improvement in the number of training programmes in the country; however, there is a lack of a well-established, government-supported regulatory body:

There is also need for mentorship:

Additionally, some participants (n = 10) reported the need for more peer support, mentorship and interdisciplinary collaboration to combat segregation and improve service quality:

This leads to isolation in the profession:

To further improve support one participant noted:

Participants also recognised the increased support from both regional and global groups of SLTs, such as the African Connections Group and the International Association of Communication Sciences and Disorders.

This study used purposive and snowball sampling to recruit participants. This resulted in a disproportionately urban-based group, and the authors acknowledge that selection bias could be present. As such, the findings reflect the experiences of the current sample rather than the entire national workforce, potentially limiting their generalisability to the wider population of Kenyan SLTs (Andrade 2021). It should also be noted that participants with diplomas in speech therapy were not included in the study, as these programmes were not known to the authors during the initial phase of disseminating the survey. The truthfulness of the results is reliant on how accurately the participants self-interpreted and responded to survey questions. However, the study did not independently verify professional qualifications or licensure status, and these factors may contribute to variability in professional status within the sample. Data were self-reported and may be influenced by recall or social desirability bias. Although the inherent limitations of self-administered online surveys remain, adherence to CHERRIES reporting standards enhanced transparency regarding survey administration and response patterns.

The survey instrument was adapted from existing literature but did not undergo formal psychometric validation. Similarly, the qualitative component was based on a limited number of open-ended survey responses, and inter-coder reliability was not calculated, thus qualitative findings should therefore be interpreted as illustrative. Even though the number of SLTs in this study is a fair sample size, a post-hoc power analysis using G*Power software (Faul et al. 2007) indicated that 84 respondents would be required to achieve a power of at least 0.80 for correlation analyses at a significance level of 0.05 to detect at least a medium effect size. This study included only 46 respondents, which means the statistical power was lower than recommended. As such, correlation findings should be interpreted cautiously, particularly where effect sizes were small. Despite these limitations, this study offers an insight into practices and challenges SLTs face in Kenya.