Impact of Dysphagia Rehabilitation in Adults on Swallowing Physiology Measured With Videofluoroscopy: A Mapping Review

Abstract

Purpose:

The research aims of this review were to (a) map swallowing rehabilitation approaches to specific swallowing impairments using the Modified Barium Swallow Impairment Profile (MBSImP) to develop evidence maps, (b) match desired rehabilitation treatment targets to treatment approaches, and (c) identify gaps in the literature and highlight which rehabilitation approaches require further investigation to support accurate mapping of interventions to physiologic change.

Method:

A mapping review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses–Scoping Review extension framework. The databases searched were CINAHL, Ovid Medline, and Ovid Embase. Data extracted included swallowing rehabilitation approach details via the Rehabilitation Treatment Specification System framework, study characteristics, and resulting change in swallowing physiology. The resulting change in swallowing physiology was mapped onto MBSImP components, where applicable, and effect sizes were reported where data were available. Extracted data were summarized in the form of evidence maps.

Results:

Forty-three unique articles met the inclusion criteria for this review and were divided into single-approach and multi-approach exercise studies. Within single-approach studies, 13 different exercise approaches were investigated, and 117 outcome measures could be mapped to MBSImP components. Within multi-approach studies, 13 different combinations of exercise approaches were investigated and 60 outcome measures could be mapped to MBSImP components.

Conclusions:

This review supports speech-language pathologists in incorporating current best evidence into their practice, as it found there is potential for improvement in many MBSImP components by using rehabilitative exercises. In the future, more intervention studies are needed to ensure that recommended rehabilitation approaches are beneficial for improving the targeted swallowing physiology.

Dysphagia (i.e., swallowing dysfunction) results from a variety of medical etiologies and is especially prevalent in older adults (Bhattacharyya, 2014). It is an independent predictor of decreased quality of life, increased length of hospitalization, and serious health concerns, such as pneumonia, that can lead to increased risk of mortality (Altman et al., 2010). Thus, the development and use of effective treatment methods for dysphagia is a priority in the field of rehabilitation research (Robbins et al., 2002).

When treating dysphagia, it is recommended that decisions are based on the three pillars of evidence-based practice: current best evidence, clinical experience, and patient values (Dollaghan, 2019; Sackett et al., 1996; Wheeler-Hegland et al., 2009). In clinical practice, integrating these three pillars is complex, and although evidence has shown that speech-language pathologists (SLPs; see Appendix A) understand and value evidence-based practice, they may struggle to integrate scientific evidence when making clinical decisions (Wheeler-Hegland et al., 2009; Zipoli & Kennedy, 2005).

Despite the importance of evidence-based decision making, there is a lack of consensus in the literature regarding which intervention approaches are most appropriate for targeting a specific swallowing impairment or set of impairments (Vose et al., 2018). This may in turn lead to difficulties in the application of evidence-based practices during clinical decision making. The lack of consensus in the literature may also be leading to a lack of consensus among practicing clinicians. Two separate studies have identified a lack of consistency among clinicians when identifying swallowing impairments and selecting appropriate treatment approaches for specific physiologic impairments (Carnaby & Harenberg, 2013; Vose et al., 2018). Carnaby and Harenberg (2013) surveyed 254 clinicians who were provided with a hypothetical swallowing study and found that 96 different combinations of treatment techniques were recommended after viewing the same swallow study. No single combination was exactly repeated across respondents, and only 4% of responses were derived based on a specific physiologic abnormality (Carnaby & Harenberg, 2013). They also found that more than 58% of recommended techniques did not match the dysphagia-related symptoms specific to the hypothetical patient, and only 13% of proposed techniques were exercise-based (Carnaby & Harenberg, 2013). A study by Vose et al. (2018) reported similar results, in that, when examining treatment selection across respondents, each of the 21 treatment options available in the survey were selected by at least one respondent for each case (except one for the complex video). They also found that 73% of respondents did not mention physiology in their rationale for treatment choice after reviewing a complex case (Vose et al., 2018). Given these studies, it is essential that clinicians are provided with a resource to facilitate better use of the available evidence. This is critical given that treatment selection not informed by the patient's physiological impairments has potential implications for patient outcomes, as positive treatment outcomes are typically dependent upon clinicians understanding the swallowing pathophysiology and using that knowledge to choose the most appropriate treatment (Easterling, 2017).

Given current gaps in understanding the linkage between intervention approach and physiologic outcomes (Easterling, 2017; Wheeler-Hegland et al., 2009), clinicians may be relying heavily on clinical expertise and patient values (Carnaby & Harenberg, 2013; Vose et al., 2018). A survey by McCurtin and Healy (2016) revealed that clinicians report their choice of dysphagia interventions is primarily based on patient suitability and clinical experience (McCurtin & Healy, 2016). Many clinicians also develop treatment plans based on bolus-related outcomes (i.e., aspiration, penetration, residue, and esophageal clearance), rather than specific physiologic impairments (Vose et al., 2018). This focus on bolus-specific outcomes is limited to addressing the signs of dysphagia rather than the impairment underlying these outcomes. The emphasis on bolus-specific outcomes may be at least partially explained by the high number of intervention studies that report these outcomes, as opposed to measuring change in swallowing physiology (Vose et al., 2018). However, it is important that the focus of swallowing rehabilitation be shifted back towards underlying physiologic impairments by improving our understanding of how dysphagia interventions directly target swallowing physiology.

To encourage the use of scientific evidence in decision making regarding dysphagia intervention, this review article endeavored to answer the following question: “How do swallowing rehabilitation approaches impact specific swallowing impairments in adults with dysphagia?” Given the breadth of this research question, this review article addresses the following aims:

1. Map swallowing rehabilitation approaches to specific swallowing impairments using the Modified Barium Swallow Impairment Profile (MBSImP) to develop evidence maps.

2. Match desired rehabilitation treatment targets to treatment approaches.

3. Identify gaps in the literature and highlight for clinicians and researchers which rehabilitation approaches require further investigation to support accurate mapping of interventions to physiologic change.

Method

This mapping review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses–Scoping Review extension, which applies to both scoping reviews and evidence maps (Peters, Godfrey, et al., 2020; Peters, Marnie, et al., 2020; Tricco et al., 2018). Mapping reviews aims to categorize, classify, characterize patterns, trends, or themes, allowing a deeper thematic understanding of pre-existing research on a particular topic including any gaps that can be addressed by future research (Grant & Booth, 2009).

Eligibility Criteria

Studies were included if they met the following criteria: (a) prospective or retrospective published research study in English; (b) included adults 18 years of age or older; (c) participants with swallowing difficulties; (d) swallowing physiology measures derived from videofluoroscopic swallow study (VFSS) images; and (e) at least one outcome measure could be mapped to an MBSImP component describing swallowing physiology. The MBSImP is a standardized approach to instruction, assessment, and reporting of physiologic swallowing impairments based on observations obtained from the VFSS (Martin-Harris et al., 2008). More specifically, it outlines 17 components that represent key factors required for safe and efficient swallowing. While widely adopted, one limitation of the tool is that it does not include bolus-specific measures, pixel-based measures, or timing parameters. However, it was selected for the purposes of this mapping review given that it is a widely accepted framework for describing swallowing physiology, and is used by over 5,000 clinicians in 50 states and at least 14 countries (Martin-Harris et al., 2017). Components describing residue (5 and 16) were not included in this review, as they describe consequences of impaired physiology rather than the physiologic impairment itself.

Exclusionary criteria for this review included (a) conference abstracts; (b) only healthy participants; (c) studies of compensatory approaches for impaired swallowing (e.g., postural changes to redirect the bolus or changes to bolus viscosity); (d) studies focused on proactive intervention (e.g., prior to patients undergoing chemoradiation treatment for head and neck cancer); (e) studies focused on nonactive exercise approaches; (f) studies involving individualized intervention across participants; (g) studies utilizing clinical or other instrumental assessments to measure physiological change, such as Fiberoptic Endoscopic Evaluation of Swallowing or manometry, for which there is not a validated tool comparable to the MBSImP for characterizing specific swallowing impairments; and (h) studies that only evaluated Components 5 and 16 (residue) and did not measure swallowing physiology.

Search Strategy

The search strategy was developed by four authors (M.R., J.S., E.W., and C.Q.) in consultation with a McMaster University librarian and two experts in the field (N.R.P. and A.N.M.). The databases searched were CINAHL, Ovid Medline, and Ovid Embase from inception until March 24, 2021. Refer to Appendix B for full search strategies employed in all databases.

Selection of Sources of Evidence

Following the search, all identified citations were uploaded into Covidence for systematic review, and duplicates were removed. Titles and abstracts were divided among a team of four (C.Q., M.R., J.S., and E.W.) and each citation was independently screened by two reviewers for assessment against inclusion and exclusion criteria. Potentially relevant sources were retrieved for full-text review by two independent reviewers. Reasons for exclusion of sources at all stages of the review were recorded and reported. Any disagreements were resolved through consensus and/or review by an additional reviewer.

Data Extraction

Data were extracted from studies that met all criteria to be included in the mapping review by two independent reviewers. Extraction was double-blinded. Data extracted included the swallowing rehabilitation approach employed, study characteristics, and the resulting change in swallowing physiology. Studies were divided into either single or multi-approach exercises. A study was defined as a single approach if the differences between groups could be attributed to one exercise or protocol or if only one exercise or protocol was completed across all participants. A study was defined as multi-approach when multiple exercises were included, and results could not be attributed to a single exercise alone. The resulting changes in swallowing physiology were mapped to at least one of the 17 MBSImP components, where applicable. Unfortunately, many studies included in this review did not utilize the MBSImP tool as an outcome measure; therefore, it was decided that mapping would be performed based on consideration of the name of each component, rather than the operational definition, to allow for broad applicability. After the research team developed the mapping plan, two reviewers (C.Q., M.R., J.S., and/or E.W.) independently mapped outcome measures to MBSImP component(s). Any discrepancies were resolved through team discussions to achieve consensus.

The Videofluoroscopic Dysphagia Scale (VDS), also referred to as the New VFSS Scale and the Functional Dysphagia Scale, was commonly used as an outcome measure (Byeon, 2020; Jang et al., 2019; Kang et al., 2012; Kim et al., 2015; A. Park et al., 2021; H.-S. Park, Oh, et al., 2019; J.-S. Park, Hwang, et al., 2019; J.-S. Park, An, et al., 2018; J.-S. Park et al., 2016; J.-S. Park, Oh et al., 2018; Toyama et al., 2014). The VDS was developed as a scale to quantify functional dysphagia in stroke patients (Han et al., 2001). The VDS has 14 parameters across the oral and pharyngeal phases of the swallow. Each parameter is scored individually, and composite scores can be derived for oral and pharyngeal phases, as well as a total VDS score. Table 1 was created to demonstrate which VDS measures correspond to MBSImP components, to ensure consistency across all reviewers.

Table 1.

Mapping the Videofluoroscopic Dysphagia Scale (VDS) to Modified Barium Swallow Impairment Profile (MBSImP) components.

VDS components	MBSImP components
Lip closure	1. Lip closure
	2. Tongue control during bolus hold
Bolus formation Mastication	3. Bolus prep/mastication
Apraxia	4. Bolus transport/lingual motion
Residue in oral cavity	5. Oral residue
Triggering of pharyngeal swallow Premature bolus loss	6. Initiation of the pharyngeal swallow
	7. Soft palate elevation
Laryngeal elevation (or laryngeal elevation and epiglottic closure)	8. Laryngeal elevation
	9. Anterior hyoid excursion
Laryngeal elevation and epiglottic closure	10. Epiglottic movement
	11. Laryngeal vestibular closure
	12. Pharyngeal stripping wave
	13. Pharyngeal contraction
	14. Pharyngoesophageal segment opening
	15. Tongue base retraction
Vallecular residue Pyriform sinus residue Coating of pharyngeal wall	16. Pharyngeal residue
	17. Esophageal clearance
VDS components not mappable to MBSImP
Tongue-to-palate contact
Oral transit time
Pharyngeal transit time
Aspiration

Note. When more than one VDS component mapped onto one MBSImP component, only one of the VDS components was required to show a significant difference for the corresponding MBSImP component to be considered as changed.

As explained by Gaeta and Brydges (2020), we often rely on p values to determine the statistical significance of the difference between groups. However, it is important to understand that p values are often insufficient when trying to determine if a meaningful difference exists, particularly in the presence of small sample sizes. As such, effect size can be used to quantify the magnitude of difference/association between groups/variables (Cohen, 1992). To consider how small sample sizes may have skewed results, effect sizes were either extracted or calculated when the necessary data were available for single approach studies. Effect size was not included for multi-approach studies given that it is not clear if the combination of exercises or a single approach within the combination was critical in improving swallowing outcomes. Per Cohen's d (Cohen, 1992) and Hedges' g (Rosenthal et al., 1994), small, medium, and large effect sizes are generally interpreted to be 0.2, 0.5, and 0.8, respectively. That is, a value of 0.2 indicates that there is a difference of 0.2 SD between mean scores of two groups. However, given that this has been found to be underestimated in published research in audiology and speech-language pathology (Gaeta & Brydges, 2020), we used updated effect size interpretations as suggested by Gaeta and Brydges (2020), where Cohen's d/Hedges' g of 0.25, 0.55, and 0.95 can be interpreted as small, medium, and large effect sizes, respectively,

To ensure transparency in reporting the studies included in this review and provide context regarding exercise approaches, all information regarding exercise-based approaches from the studies included was extracted in accordance with the Rehabilitation Treatment Specification System (RTSS) framework (Hart et al., 2019). Using the RTSS, rehabilitation treatments are specified using three elements: targets, ingredients, and mechanisms of action (Turkstra et al., 2016). Targets describe what aspects of the patient's functioning the clinician is attempting to directly change through treatment. They can be further subdivided into organ functions, skills and habit, and representations. According to Van Stan et al. (2019), organ function targets “are concerned with changes in the efficiency, functioning, or replacement of an organ or organ system (e.g., exercise, habituation, prosthetics)” (p. 147). Skill and habits are targets involved in “modifying mental or behavioral skills through providing ingredients such as practice, repetition, feedback, etc.” (p. 147). Last, representations “are intended to change mental representations related to cognition, affect, motivation, and volitional behavior” (p. 147). Ingredients consist of observable and measurable actions that are delivered in therapy by the clinician to produce changes in the target. Finally, the mechanism of action refers to the known or hypothesized process by which the treatment's ingredients generate change on the target. Considering relationships among elements of the RTSS can aid clinicians in choosing rehabilitation approaches that best fit with their goals for therapy with their patients (Turkstra et al., 2016). As such, the ingredients were mapped (i.e., specific features that make up each exercise-based rehabilitation approach) to specific targets (i.e., MBSImP components).

Two out of three possible reviewers (M.R., J.S., and/or E.W.) independently determined the level of evidence for each study in accordance with the definitions by the American Society of Plastic Surgeons (as cited in Burns et al., 2011). Any discrepancies were resolved through team discussions to achieve consensus. Group discussions (M.R., J.S., and E.W.) were used to determine grade recommendations for each exercise approach, in accordance with the Centre for Evidence-Based Medicine's (CEBM, 2009) levels of evidence for therapeutic studies. The grading system is an important component in evidence-based medicine and assists in clinical decision making. Extracted data were summarized in the form of an evidence map. Between-group results were reported when available, otherwise, within-group results were reported to provide an overall picture of the best available evidence across varying study designs.

Results

The literature search yielded 6,920 unique articles (see Figure 1). After reviewing titles and abstracts, 261 full-text articles were further reviewed for possible inclusion. Interrater reliability was 97% at the title and abstract review stage and 94% at the full-text review stage. Reasons for exclusion of full texts are listed in Figure 1. A primary reason for exclusion was that outcomes could not be mapped to any MBSImP component. This was common when articles used the Penetration–Aspiration Scale (PAS). The PAS reports on airway invasion (i.e., penetration or aspiration) as a consequence of impaired physiology and does not refer to any specific physiologic deficits leading to the airway invasion. The VDS was another commonly used measure; however, outcomes could not be mapped to the MBSImP when only aggregate oral and pharyngeal scores were reported rather than information specific to individual components. There were 43 unique articles (i.e., references; see Tables 2 and 3) that met the inclusion criteria for this review. Four articles (Byeon, 2020; Choi et al., 2020; Kim et al., 2015; Lin et al., 2011) included designs that employed two different exercise approaches for two different groups of participants. To account for multiple exercise approaches within a single article, this review includes reports of 47 studies, of which 35 were reported as a single-approach and 13 as multi-approach.

Figure 1.

Flow diagram of screening and review stages, including reason for exclusion at the full-text stage. VFSS = videofluoroscopic swallow study; MBSImP = Modified Barium Swallow Impairment Profile.

Table 2.

Characteristics of single-approach studies.

Single approach
Study (primary author, year)	Approach	# of participants	Average age	Gender	Ethnicity	Primary medical diagnosis	Identification of swallowing difficulties
Byeon, 2020 a [1]	Mendelsohn maneuver	15	63.5 years	8 M, 7 F	Not reported	Cerebral infarction	Diagnosed with swallowing disorders due to cerebral infarction within the past 6 months
Byeon, 2020 a [1]	NMES	13	65.1 years	7 M, 6 F	Not reported	Cerebral infarction	Diagnosed with swallowing disorders due to cerebral infarction within the past 6 months
Carnaby-Mann, 2010 [2]	McNeill Dysphagia Therapy Program (MDTP)	24 (8 experimental, 16 control)	58.9 years	Experimental: 2 M, 6 F Control: 4 M, 12 F	Not reported	Head/neck cancer & stroke	Functional Oral Intake Scale, Mann Assessment of Swallowing Ability, confirmed via VFSS
Choi, 2020 a [3]	Jaw Opening Exercise (JOE)	11	63.5 years	5 M, 6 F	Not reported	Stroke	Confirmed via VFSS
Choi, 2020 a [3]	Shaker/Head Lift Exercise (HLE)	10	61.2 years	4 M, 6 F	Not reported	Stroke	Confirmed via VFSS
Crary, 2012 [4]	McNeill Dysphagia Therapy Program (MDTP)	9	55.9 years	7 M, 2 F	Not reported	6: oropharyngeal cancer, 2: neurological deficit, 1: combination of both	Functional Oral Intake Scale, Mann Assessment of Swallowing Ability, confirmed via VFSS
El Sharkawi, 2002 [5]	Lee Silverman Voice Treatment (LSVT)	8	69.8 years	6 M, 2 F	Not reported	Parkinson's disease	Confirmed via VFSS
Hegland, 2016 [6]	Expiratory Muscle Strength Training (EMST)	14	64.5 years	8 M, 6 F	Not reported	Stroke	Exclusion criteria included dysphagia secondary to impairment other than stroke
Hutcheson, 2018 [7]	Expiratory Muscle Strength Training (EMST)	26 (3 participants withdrew before completing entire 8-week program)	67 years	23 M, 3 F	Not reported	Head/neck cancer survivors with chronic radiation-associated aspiration	Confirmed via VFSS
Juan, 2013 [8]	Tongue Resistance Training ([TRT] Isometric Progressive Resistance Oropharyngeal [I-PRO] Therapy)	1	59 years	1 F	Not reported	Stroke	Confirmed via VFSS
Kim, 2015 a [9]	Proprioceptive neuromuscular facilitation (PNF)–based short neck flexion exercises	13	63.2 years	8 M, 5 F	Not reported	Stroke	Dysphagic symptoms for over 6 months
Kim, 2015 a [9]	Shaker/Head Lift Exercise (HLE)	13	63.6 years	7 M, 8 F	Not reported	Stroke	Dysphagic symptoms for over 6 months
Koyama, 2016 [10]	Dry Swallow After Suction (DSAS)	1	68 years	1 M	Not reported	Bickerstaff's brainstem encephalitis (BBE)	Confirmed via VFSS
Koyama, 2017 [11]	Modified Jaw Opening Exercise (MJOE)	12 (6 experimental, 6 control)	Experimental: 66 years Control: 71.8 years	10 M, 2 F	Not reported	Stroke	Confirmed via VFSS
Langmore, 2016 [12]	Neuromuscular Electrical Stimulation (NMES)	168 (116 active NMES, 52 sham)	61.9 years	114 M, 24 F	121 White (72%), 19 Black (11.3%), 14 Asian (8.3%), 13 Hispanic (7.7%), 1 Mixed (0.6%)	Head/neck cancer	Confirmed via VFSS
Lin, 2011 b [13]	Neuromuscular Electrical Stimulation (NMES)	10 in the functional electrical stimulation (FES) group	52.3 years	6 M, 4 F	Not reported	Nasopharyngeal carcinoma	The Dysphagia Outcome and Severity Scale
Logemann, 2009 [14]	Shaker/Head Lift Exercise (HLE)	19 (8 experimental, 11 control; however, final data was only collected for 11 participants - 5 experimental, 6 control)	Experimental: 63.1 years Control: 70.9 years	16 M, 3 F	Experimental: 5 White, 3 Black, 0 Asian Control: 9 White, 1 Black, 1 Asian	Stroke & head/neck cancer	Confirmed via VFSS
Mano, 2015 [15]	Shaker/Head Lift Exercise (HLE)	6	64.7 years	6 M	Japanese	Spinal and bulbar muscular atrophy	Subjective reports
Martin-Harris, 2015 [16]	Respiratory Swallow Training (RST)	30 (Fifteen 1-month follow-up)	61 years	26 M, 4 F	26 non-Hispanic White, 4 non-Hispanic Black	Head/neck cancer	PAS, MBSImP
Matsubara, 2018 [17]	Jaw Opening Exercise (JOE)	21	74 years	10 M, 11 F	Japanese	Symptoms suggestive of swallowing disorders	Symptoms suggestive of swallowing disorders
McCullough, 2013 [18]	Mendelsohn maneuver	18	70.2 years	11 M, 7 F	Not reported	Stroke	Confirmed via VFSS
Namiki, 2019 [19]	Tongue Pressure Resistance Training (TPRT)	18	76.8 years	11 M, 7 F	Not reported	Presbyphagia	Subjective complaints of swallowing problems
A. Park, 2021 [20]	Lee Silverman Voice Treatment (LSVT)	13 (6 idiopathic Parkinson's disease [IPD], 7 multiple system atrophy cerebellar type [MSA-C]	IPD: 66.7 years MAS-C: 63.9 years	Not reported	Not reported	Parkinsonism & complaint of difficulty in swallowing	Subjective complaints of swallowing difficulty
H.-S. Park, 2019 [21]	Effortful swallow	24 (12 experimental, 12 control)	Experimental: 66.5 years Control: 64. 8 years	Experimental: 6 M, 6 F Control: 5 M, 7 F	Not reported	Stroke	Confirmed via VFSS
J.-W. Park, 2012 [22]	Neuromuscular Electrical Stimulation (NMES)	18 (9 experimental, 9 control)	Experimental: 68.7 years Control: 62 years	16 M, 2 F	Not reported	Stroke	Confirmed via VFSS
J.-S. Park, 2020 [23]	Resistive Jaw Opening Exercise (RJOE)	40 (20 experimental, 20 control; however final data consists of only 29 participants - 15 experimental, 14 control)	Experimental: 62.1 years Control: 61.8 years	Experimental: 9 M, 6 F Control: 8 M, 6 F	Not reported	Stroke	Confirmed via VFSS
J.-S. Park, An, et al., 2018 [24]	Chin Tuck Against Resistance (CTAR)	25 (13 experimental, 12 control; however, 3 participants withdrew from the study and final results only include results from 22 participants - 11 experimental, 11 control)	Experimental: 61.2 years Control: 58.4 years	Experimental: 6 M, 5 F Control: 4 M, 7 F	Not reported	Stroke	Confirmed via VFSS
J.-S. Park, Oh, et al., 2018 [25]	Neuromuscular Electrical Stimulation (NMES)	18 (9 experimental, 9 control)	Experimental: 63.4 years Control: 54.7 years	8 M, 10 F	Korean	Parkinson's disease	Confirmed via VFSS
J.-S. Park, 2017 [26]	Shaker/Head Lift Exercise (HLE)	27 (13 experimental, 14 control)	Experimental: 59.3 years Control: 61.6 years	17 M, 10 F	Korean	Stroke	Confirmed via VFSS
J.-S. Park, 2016 [27]	Neuromuscular Electrical Stimulation (NMES)	50 (25 experimental, 25 control)	Experimental: 54 years Control: 55.8 years	26 M, 24 F	Not reported	Stroke	Confirmed via VFSS
Shaker, 2002 [28]	Shaker/Head Lift Exercise (HLE)	27 participants completed the “real exercise (with 7 participants completing sham exercise before continuing on complete 6 weeks of real exercise afterward)	72 years	25 M, 2 F	Not reported	Abnormal UES opening due to various causes	Not reported
Sia, 2015 [29]	McNeill Dysphagia Therapy Program (MDTP)	8	54.5 years	6 M, 2 F	Not reported	Head/neck cancer & stroke	Confirmed via VFSS
Toyama, 2014 [30]	Neuromuscular Electrical Stimulation (NMES)	26 (12 experimental, 14 control)	Experimental: 63.6 years Control: 67.2 years	22 M, 4 F	Japanese	Stroke	Confirmed via VFSS
Troche, 2010 [31]	Expiratory Muscle Strength Training (EMST)	68 (33 experimental, 35 control; however, final data collected only for 60 participants)	Experimental: 66.7 years Control: 68.5 years	47 M, 13 F	Not reported	Idiopathic Parkinson's disease	Reports of coughing with meals & increased eating duration; confirmed via VFSS
Wada, 2012 [32]	Jaw Opening Exercise (JOE)	8	70.5 years	7 M, 1 F	Not reported	Stroke, myelitis, angina, & gastritis	Confirmed via VFSS

Note. VFSS = videofluoroscopic swallow study; PAS = Penetration–Aspiration Scale; MBSImP = Modified Barium Swallow Impairment Profile; UES = upper esophageal sphincter.

^a Article is mentioned twice in this chart.

^b Article is also mentioned in the multi-approach chart.

Table 3.

Characteristics of multi-approach studies.

Multi-approach
Study (primary author, year)	# of participants	Average age	Gender	Ethnicity	Primary medical diagnosis	Identification of swallowing difficulties
Argolo, 2013	15	59.2 years	10 M, 5 F	Not reported	Parkinson's disease	Confirmed via VFSS
Balou, 2019	9	75.3 years	3 M, 6 F	Not reported	None (healthy, but radiographically confirmed dysphagia)	Confirmed via VFSS, PAS
Byeon, 2020	43 (15 Mendelsohn maneuver [MM], 13 NMES, 15 Mendelsohn + NMES)	MM: 63.5 years NMES: 65.1 years MM + NMES: 65 years	MM: 8 M, 7 F NMES: 7 M, 6 F MM + NMES: 9 M, 6 F	Not reported	Cerebral infarction	Diagnosed with swallowing disorders due to cerebral infarction within the past 6 months
Cha, 2010	1	25 years	1 M	Not reported	Spinal muscular atrophy Type II	Swallowing problems related to nonoral feeding
El-Tamawy, 2015	30 (15 experimental, 15 control)	Experimental: 61.5 years Control: 61.3 years	Not reported	Not reported	Stroke	Beside assessment of swallowing
Holmes, 2012	3	27.3 years	3 F	Not reported	Anorexia nervosa	Beside swallow evaluation & VFSS
Jang, 2019	36 (18 experimental, 18 control)	Experimental: 67.3 years Control: 71.2 years	Experimental: 10 M, 8 F Control: 9 M, 9 F	Not reported	Subacute stroke with velopharyngeal incompetence	Confirmed via VFSS
Jiang, 2017	1	26 years	1 F	Not reported	Anastomotic stricture following esophageal replacement with the colon	Clinical evaluation, VFSS
Kang, 2012	50 (25 experimental, 25 control)	Experimental: 66.7 years Control: 68.3 years	34 M, 16 F	Not reported	Stroke	Confirmed via VFSS
Kraaijenga, 2017	17 (2 participants withdrew from study and were not included in final results)	65 years	13 M, 4 F	Not reported	Head/neck cancer	Confirmed via VFSS and/or by a severely limited intake of normal diet
Lazarus, 1994	1	51 years	1 M	White	Oral cancer	Confirmed via VFSS
Lin, 2011 a	10 in the home rehabilitation program (HRP) group	56.1 years	6 M, 4 F	Not reported	Nasopharyngeal carcinoma	The Dysphagia Outcome and Severity Scale
J.-S. Park, 2019	10	60.1 years	5 M, 5 F	Not reported	Stroke	Confirmed via VFSS

Note. VFSS = videofluoroscopic swallow study; PAS = Penetration-Aspiration Scale.

^a Article is also mentioned in the single approach chart.

Study Characteristics

Articles included in this review span the period from 1994 to 2021, with the majority (n = 37/47; 79%) published within the last 10 years (refer to Tables 2 and 3). With all articles combined, this review included individualized data from 975 individual participants. Almost half of the studies (n = 22/47; 47%) included participants with a primary medical diagnosis of stroke, with other diagnoses including head and neck cancer, Parkinson's disease, and other conditions impacting swallowing such as encephalitis, acquired brain injury, myelitis, gastritis, anorexia nervosa, anastomotic stricture, and a broad designation of “neurological deficit.” Papers involving participants with neurodegenerative disease focused on improvement as opposed to maintenance, meaning that their aim of improvement matched with all other studies included in this review. Most studies (n = 33/47; 70%) identified dysphagia using VFSS, and all studies employed VFSS to assess swallowing physiology. Other methods of identifying dysphagia across studies included bedside swallow evaluations, the Dysphagia Outcome and Severity Scale (O'Neil et al., 1999), complaints of swallowing difficulties, and/or limited intake as reported by participants. In total, 15 (32%) articles employed a randomized controlled trial (RCT) design (see Tables 4 and 5): five for Neuromuscular Electrical Stimulation (NMES; Byeon, 2020; Langmore et al., 2016; J.-W. Park et al., 2012; J.-S. Park et al., 2016; J.-S. Park, Oh, et al., 2018), three for Shaker/Head Lift Exercise (HLE; Logemann et al., 2009; J.-S. Park et al., 2017; Shaker et al., 2002), two for Jaw Opening Exercises (JOE; Koyama et al., 2017; J.-S. Park et al., 2020), one for Chin Tuck Against Resistance (CTAR; J.-S. Park, An, et al., 2018), one for effortful swallow (H.-S. Park, Oh, et al., 2019), one for Expiratory Muscle Strength Training (EMST; Troche et al., 2010), one for Mendelsohn maneuver (Byeon, 2020), and three for multi-approach (Byeon, 2020; El-Tamawy et al., 2015; Jang et al., 2019). However, for data extraction purposes, between-group data were not available for Troche et al. (2010) and J.-W. Park et al. (2012), as statistical analyses were not reported for between-group outcome measures that could be mapped to MBSImP components.

Table 4.

Statistically significant improvements in Modified Barium Swallow Impairment Profile (MBSImP) components for single-approach studies.

Approach	Study (primary author, year)	RCT	MBSImP components investigated	Between-group improvements (MBSImP component number)	Within experimental group improvements (MBSImP component number)
CTAR	J.-S. Park, An, et al., 2018 [24]	✓	1, 3, 6, 7, 8, 10	8, 10	8, 10
DSAS	Koyama, 2016 [10]		7	N/A	7
Effortful swallow	H.-S. Park, 2019 [21]	✓	1, 3, 4, 6, 8	No significant differences	8
EMST	Hegland, 2016 [6]		6, 7, 8, 9, 10, 11, 12, 14, 15	N/A	No significant differences
	Hutcheson, 2018 [7]		7, 8, 9, 10, 11, 12, 13, 14, 15	N/A	No significant differences
	Troche, 2010 [31]	✓ c	9	NR	No significant difference
JOE	Choi, 2020 a [3]		9	N/A	9
	Koyama, 2017 [10]	✓	9	9	9
	Matsubara, 2018 [17]		9, 14	N/A	9, 14
	J.-S. Park, 2020 [23]	✓	9	No significant difference	9
	Wada, 2012 [32]		9, 14	N/A	14
LSVT	El Sharkawi, 2002		2, 3, 6, 7, 8, 9, 11, 15	N/A	NR
	A. Park, 2021 [20]		1, 3, 4, 6, 8	N/A	MSA-C: 6 IPD: 6, 8
MDTP	Carnaby-Mann, 2010 [2]		8, 9, 15	N/A	NR
	Crary, 2012 [4]		8, 9	N/A	8
	Sia, 2015 [29]		8, 9	N/A	8
Mendelsohn maneuver	Byeon, 2020 a , b [1]	✓	1, 3, 6, 7, 8	N/A	NR
	McCullough, 2013 [18]		9, 14	N/A	No significant differences
NMES	Byeon, 2020 a , b [1]	✓	1, 3, 6, 7, 8	N/A	NR
	Langmore, 2016 [12]	✓	9	No significant difference	9↓
	Lin, 2011 b [13]		9	N/A	No significant differences
	J.-W. Park, 2012 [22]	✓ c	8, 9, 14	NR	8
	J.-S. Park, 2016 [27]	✓	9	9	9
	J.-S. Park, Oh, et al., 2018 [25]	✓	9	9	9
	Toyama, 2014 [30]		8, 9	8, 9	8, 9
PNF	Kim, 2015 a [9]		1, 3, 4, 6, 8, 10	N/A	6, 8, 10
RST	Martin-Harris, 2015 [16]		1, 2, 4, 6, 7, 8, 9, 10, 11, 12, 14, 15	N/A	7, 11, 15
Shaker/HLE	Choi, 2020 a [3]		9	N/A	9
	Kim, 2015 a [9]		1, 3, 4, 6, 8, 10	N/A	6, 8, 10
	Logemann, 2009 [14]	✓	8, 9, 14	No significant differences	14
	Mano, 2015 [15]		7	N/A	No significant differences
	J.-S. Park, 2017 [26]	✓	8, 9	No significant differences	8, 9
	Shaker, 2002 [28]	✓ d	8, 9, 14	No significant differences	14, 16
TRT	Juan, 2013 [8]		14	N/A	NR
	Namiki, 2019 [19]		9, 14	N/A	9, 14

Note. RCT = randomized controlled trial; CTAR = Chin Tuck Against Resistance; DSAS = Dry Swallow After Suction; N/A = not applicable; EMST = Expiratory Muscle Strength Training; NR = not reported; JOE = Jaw Opening Exercises; LSVT = Lee Silverman Voice Treatment; MSA-C = Multiple system atrophy-cerebellar type; IPD = Idiopathic Parkinson’s disease; MDTP = McNeill Dysphagia Therapy Program; NMES = Neuromuscular Electrical Stimulation; PNF = Proprioceptive Neuromuscular Facilitation Based Short Neck Flexion Exercises; RST = Respiratory Swallow Training; HLE = Head Lift Exercise; TRT = Tongue Resistance Training. ↓ = component got worse/declined.

^a Article is mentioned in more than one location in chart.

^b Article is mentioned in the multi-approach chart as well.

^c Study design was an RCT however no between-group statistical analyses were reported for outcome measures mappable to the MBSImP.

^d Study design was originally an RCT and then participants from the control group joined the exercise group.

Table 5.

Statistically significant improvements in Modified Barium Swallow Impairment Profile (MBSImP) components for multi-approach studies.

Study (primary author, year)	RCT	MBSImP components investigated	Between-group improvements (MBSImP component number)	Within experimental group improvements (MBSImP component number)
Argolo, 2013		2, 4	N/A	2
Balou, 2019		1, 2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15	N/A	6, 8
Byeon, 2020	✓	1, 3, 6, 8, 10,	8, 10	NR
Cha, 2010		1, 3, 4, 6, 7, 8, 11	N/A	NR
El-Tamawy, 2015	✓	8, 14	8	NR
Holmes, 2012		4, 6, 8, 9, 10, 11, 14, 15	N/A	NR
Jang, 2019	✓	7	7	7
Jiang, 2017		7, 9, 14	N/A	NR
Kang, 2012		1, 3, 4, 6, 8, 10	3, 4	1, 3, 4, 6
Kraaijenga, 2017		9	N/A	No significant differences
Lazarus, 1994		4, 8, 11, 14, 15	N/A	NR
Lin, 2011 a		9	N/A	9↓
J.-S. Park, 2019		1, 3, 4, 6, 8	N/A	No significant differences

Note. RCT = randomized controlled trial; N/A = not applicable; NR = not reported. ↓ = component got worse/declined.

^a Article is mentioned in single approach chart as well.

Frequency of MBSImP Components Occurring Across Studies

During the data extraction process, outcome measures across studies were mapped to certain MBSImP components more frequently than others (see Figure 2). The most included components in single-approach studies were anterior hyoid excursion (Component 9; [n = 25/35; 71%]), and laryngeal elevation (Component 8; [n = 19/35; 54%]). No studies had outcome measures that mapped to esophageal clearance (Component 17) and very few included measures that mapped to pharyngeal contraction (Component 13; [n = 1/35; 3%]), pharyngeal stripping wave (Component 12; [n = 3/35; 9%]), and tongue control during bolus hold (Component 2; [n = 2/35; 6%]).

Figure 2.

Frequency of each Modified Barium Swallow Impairment Profile (MBSImP) component occurring across single approach studies and multi-approach studies. Percentage is calculated based on the frequency each component was measured divided by the total number of single or multi-approach studies. Color scale highlights components occurring least to most often. PES = pharyngoesophageal segment.

Frequency of Exercise Approaches

Within single-approach studies, 13 different exercise approaches were investigated (see Table 4). NMES and Shaker/HLE had the most studies supporting the approaches (n = 8/35 or 23% and n = 6/35 or 17%, respectively), followed by JOE (n = 5/35; 14%), EMST (n = 3/35; 9%), and McNeill Dysphagia Therapy Program (MDTP; n = 3/35; 9%). Refer to Table 6 for a full description of each approach. Within multi-approach studies, each study included a different combination of exercises. Refer to Table 7 for a full description of approaches in each study.

Table 6.

Single-approach studies described according to the Rehabilitation Treatment Specification System (RTSS) framework.

Rehabilitation Treatment Specification System: Single Approach
Approach and description	Study (primary author, year)	Physiologic target(s)	Mechanism of action	Treatment ingredients
				Ingredients	Dosage
Chin Tuck Against Resistance (CTAR): CTAR device facilitating isometric and isotonic exercise	J.-S. Park, An, et al., 2018	Strengthen the suprahyoid muscles in the pharyngeal phase of swallowing	Increase activation of suprahyoid muscles	CTAR device; provide instruction; demonstrate exercise; emphasize correct posture. Conventional dysphagia treatment (CDT): orofacial muscle exercises; thermal tactile stimulation; compensatory maneuvers.	5×/week, 4 weeks: -Isometric: chin tuck 3×/day, 60 s each -Isotonic: 30 reps/day -CDT: 30 min/day
Dry Swallow after Suction (DSAS): Straw held in mouth, air inhaled through blocked straw, followed by dry swallow	Koyama, 2016	Decrease velopharyngeal inadequacy	Increase velopharyngeal closure	4mm x 180mm straw; provide instruction.	8 reps, 1×/day, 5×/week, 6 weeks
Effortful swallow: Press tongue onto hard palate while squeezing neck muscles and swallowing forcefully	H.-S. Park, 2019	Improve tongue strength and activation of the suprahyoid muscles	Increase posterior motion of the tongue base toward posterior pharyngeal wall	Provide instruction; visual observation and palpation to confirm accuracy of exercise, and expression and contraction of muscles; supervision and encouragement during training; small spray of water to induce swallowing as necessary; rest provided as necessary. CDT: compensatory techniques (e.g., chin tuck, head tilting and rotation) and therapeutic techniques (e.g., orofacial muscle exercises, thermal tactile stimulation, expiratory training).	5×/week, 4 weeks: -CDT: 30 min/day -Effortful swallow: 10 reps/session, 3 sessions/day
Expiratory Muscle Strength Training (EMST): EMST device facilitating forceful breathing that is adjusted over time	Hegland, 2016	Improve expiratory muscle strength, cough strength, and oropharyngeal swallowing	Increase strength and coordination of expiratory and submental muscle	EMST device; EMST calibration over time; provide instruction; practice with supervision before independent use.	5 sets of 5 reps/day, 5×/week, 5 weeks
	Hutcheson, 2018	Improve airway protection	Strengthen subglottic expiratory pressure generating forces	EMST device; EMST calibration over time; training load was set weekly; practice with supervision before independent use; written instructions.	25 reps, 5×/week, 8 weeks
	Troche, 2010	Improve hyoid bone displacement	Increase strength and contraction of expiratory and submental muscles	EMST device; weekly checkups; weekly EMST calibration; reminders on use of device; provide instruction; feedback to ensure accuracy; tracking of training compliance.	5 sets of 5 reps, 20 min/day, 5×/week, 4 weeks
Jaw Opening Exercises (JOE): Isometric and isotonic exercises completed by opening jaw (with and without resistance)	Choi, 2020 a	Increase hyoid bone movement	Increase thickness of digastric and mylohyoid muscles	Resistance bar. CDT: oral-facial massage, thermal-tactile stimulation, and various compensatory training exercises.	5×/week; 6 weeks: -Isometric: 3×/day, 10 s each -Isotonic: 3 sets of 30 reps/day -CDT: 30 min/day
	Koyama, 2017	Increase anterior displacement of the hyoid bone	Suppress eccentric contraction of the suprahyoid muscles	Biofeedback equipment (surface electrodes attached to suprahyoid muscles); provide instruction; clinician present every session; clinician provided resistance during exercises; visual feedback.	4 sets of 5 reps, 5× day, 6 weeks
	Matsubara, 2018	Improve UES opening	Increase suprahyoid muscle strength	Set metronome; provide instruction; record and submit training history.	2× daily, 4 weeks −2 sets of 20 reps/day, open every 2 s, with 10s interval
	J.-S. Park, 2020	Improve function of oropharyngeal muscles while swallowing	Suprahyoid muscle strengthening	Resistance bar; participants sit and hold device. CDT: orofacial muscle exercise, thermal tactile stimulation, and compensatory maneuvers.	5×/week, 4 weeks: -Isometric: 30 s 3×, with 30–60 s interval -Isotonic: 2–3 s reps, 3 sets of 10 reps, 30 s rest. -CDT: 30 min/day
	Wada, 2012	Increase hyolaryngeal elevation, resulting in increased UES opening	Increase strength of suprahyoid muscles	Instruct in safety; feedback on strength of muscle contraction; record sheet provided to participants.	2 sets of 5 reps, 7×/week, 4 weeks
Lee Silverman Voice Treatment (LSVT): Speech-behavior therapy protocol; including voice exercises (e.g., sustained vowel phonation, speech loudness drills, speaking louder)	El Sharkawi, 2002	Improve function in both the oral tongue and the tongue base	Improve neuromuscular control of the aerodigestive tract	Provide instruction; training to use and accurately judge louder voice; daily homework and carryover exercises.	50–60 min/session, 4×/week, 4 weeks
	A. Park, 2021	Improve pharyngeal phase	Increase voice strength, sound quality, and maximum phonation time elicit improvement in swallowing function	Provide instruction using standardized program.	4×/week, 4 weeks 5- to 10-min hmwk/day, (10- to15-min hmwk on untreated days)
McNeill Dysphagia Therapy Program (MDTP): Swallowing protocol with predetermined hierarchy altering bolus volume, consistency, and rate	Carnaby-Mann, 2010	Advance safe oral intake and improve strength and coordination of swallow	Progressive strengthening, development of movement patterns, and refinement of muscular coordination	Provide instruction; predetermined advancement in program hierarchy; record swallow attempts, bolus types, and events of coughing, throat clearing, expectoration, and/or other clinical signs of struggle; home practice and dietary records; review of participant compliance; identify difficulties with materials.	1 hr/day, 5×/week, 3 weeks
	Crary, 2012	Improve movement and strength of swallow	Physiological changes	Provide instruction; determine advancement or regression in program hierarchy based on participant performance; reinstruct as needed; record successful swallow attempts (absence of expectation or clinical signs of aspiration); review dietary records; review of participant compliance; identify difficulties with materials; additional activities.	1 hr/day, 3 weeks Hmwk each night
	Sia, 2015	Improve airway protection and increase duration of UES opening	Increase hyolaryngeal excursion, velocity, and duration	Provide instruction; standard protocol with predetermined hierarchy.	1 hr/day, 5×/week, 3 weeks
Mendelsohn maneuver: Prolongation of hyolaryngeal elevation during swallow by swallowing “long and strong” with squeeze at peak of swallow	Byeon, 2020	Improve swallowing function	Increase movement of larynx, hyoid bone, and swallowing in the pharyngeal stage	Provide instruction; administration of a small amount of food; practice time.	30 min/session, 15–20 reps, 8 weeks
	McCullough, 2013	Increase duration of hyolaryngeal movement and UES opening	Increase duration of hyolaryngeal movement and UES opening	Provide instruction; SEMG biofeedback; cervical auscultation and laryngeal palpation to confirm swallow; provide dental swabs dipped in ice water prior to each swallow; coach and correct maneuver.	45-min to 1-hr sessions, 30–40 reps, 2×/day, 2 weeks
	McCullough, 2013	Improve strength or coordination of swallow	Improve superior/anterior hyoid movement, width of UES opening	Provide instruction; SEMG biofeedback; provide dental swabs dipped in ice water prior to each swallow; visual and verbal feedback of SEMG feedback.	45 min/session, 30–40 reps, 2× day, 2 weeks
Neuromuscular Electrical Stimulation (NMES): Electrodes placed on target muscles to stimulate muscle contraction combined with active exercise	Byeon, 2020	Improve swallowing function	Increase movement of larynx, hyoid bone, and swallowing in the pharyngeal stage	VitalStim electrical stimulation therapy; provide instruction; practice time.	30 min/session, 8 weeks
	Langmore, 2016	Increase swallow strength	Facilitate muscle contraction and push through of fibrotic tissue caused by radiation damage	BMR NeuroTech 2000 electrical stimulation device; Daily logs to document compliance; training sessions prior to home-based protocol; participant re-assessment.	2×/day, 6×/week, 12 weeks −60 swallows/session: 10× super-supraglottic swallows; 10× regular swallows; 10× Mendelsohn swallows; 10× regular swallows; 10× effortful swallows; 10× regular swallows
	Lin, 2011 b	Improve elevation of the hyoid bone and the larynx system	Stimulate suprahyoid muscle group	VitalStim electrical stimulation therapy; intensity of current personally adjusted.	15× 60 min/session, 1–3× per week
	J.-W. Park, 2012	Increase hyoid bone displacement	Muscle strengthening & prevention of muscle atrophy	VitalStim electrical stimulation therapy; provide instruction; intensity of current personally adjusted.	30 min/session, 5×/week, 6 weeks
	J.-S. Park, 2016	Improve anterior hyoid bone displacement	Increase swallowing muscle strength and sensory awareness	VitalStim electrical stimulation therapy. CDT: orofacial muscle exercises; thermal tactile stimulation; and therapeutic or compensatory maneuvers.	30 min, 5×/week, 4 weeks -CDT 30 min after
	J.-S. Park, Oh, et al., 2018	Increase in the degree of hyoid elevation	Trigger swallow reflex with electrical stimulation	VitalStim electrical stimulation therapy; provide instruction; administration of a teaspoon of water and thickened water for swallowing safety.	3 sets of 20 min sessions/week, 4 weeks (1 set: 2× 10 min exercises (swallow every 10 s), 2 min intervals)
	Toyama, 2014	Improve anterior and superior movement of larynx and hyoid bone	Increase muscle activation and strength	VitalStim electrical stimulation therapy; adjust intensity; confirm displacement of hyoid bone and larynx using VFSS; participant seated in upright position; place marks on neck to identify motor points. CDT: tongue exercise, thermal tactile stimulation with intensive repetition of dry swallow, Mendelsohn maneuver, etc.	40 min/day, 5×/week, 8 weeks -Thermal-tactile stimulation with dry swallow task: three 10-min sets + 2 min rest
Proprioceptive neuromuscular facilitation (PNF)–based short neck flexion exercises	Kim, 2015 a	Increase the nerve-root mechanism's response	Proprioceptive stimulation; activating, strengthening, and relaxing muscle groups	Provide instruction; physically manipulate and support participants during exercise; provide physical support when necessary.	30 min, 3×/week, 6 weeks
Respiratory-Swallow Training (RST): Protocol training participants to coordinate swallowing with respiration	Martin-Harris, 2015	Increase airway protection and swallowing function	Improve coordination of respiration and swallowing using exhale–swallow–exhale respiratory pattern	Provide instruction; instruction manual; feedback; monitor and measure progress on goals; verbal and visual instruction; graphic illustrations, printed and on computer; record nasal airflow via the Swallow Signals Lab with the KayPENTAX Digital Swallowing Workstation; and record respiratory movements via respiratory inductance plethysmography	1 hr/day, 2×/week until mastery (max. 4 weeks). Range: 4–8 sessions 3 learning modules with 21 goals
Shaker/Head Lift Exercise (HLE): Isometric and isokinetic exercises performed with participants lifting head in supine position	Choi, 2020 a	Increase hyoid bone movement	Increase thickness of digastric and mylohyoid muscles	HLE: not reported CDT: oral-facial massage, thermal-tactile stimulation, and various compensatory training exercises.	5×/week, 6 weeks Isometric: 3 reps, 10 s head raise Isotonic: 3 sets of 30 reps CDT: 30 min/day
	Kim, 2015 a	Strengthen the suprahyoid and infrahyoid muscles and facilitate the opening action of the UES	Contraction of thyrohyoid, mylohyoid, geniohyoid, and the anterior belly of digastric muscles	Provide instruction; conduct exercise programs.	30 min/day, 3×/week, 6 weeks Isometric: 3 reps, ~1 min head raise + 1 min rest Isotonic: ~30 reps
	Logemann, 2009	Increase UES opening and hyolaryngeal movement	Strengthen muscles and increase opening width of the UES	Provide instruction; document compliance, record of practice.	3×/day, 6 weeks Isometric: 3 reps 1 min head raise + 1 min rest Isokinetic: 30 reps
	Mano, 2015	Strengthen the floor-of-mouth muscles	Increase in muscle fibres	Provide instruction; participants take photos/videos for compliance.	3×/day, 6 weeks Isometric: 3 reps ~1 min head raise + 1 min rest Isokinetic: 30 reps
	J.-S. Park, 2017	Increase hyolaryngeal movement	Improve strength and endurance of the submental muscles	HLE: not reported CDT: orofacial muscle exercises, thermal tactile stimulation, and therapeutic maneuvers.	5×/week, 4 weeks Isometric: 3 reps 1 min head raise + 1 min rest Isokinetic: 30 reps
	Shaker, 2002	Increase UES opening and hyolaryngeal movement	Strengthen muscles and increase opening width of the UES	Provide instruction; written exercise instructions; exercise log.	3×/day, 6 weeks Isometric: 3 reps 1 min head raise + 1 min rest Isokinetic: 30 reps
Tongue Resistance Training (TRT): Tongue strengthening exercises with or without adjunct materials	Juan, 2013	Increase isometric and swallowing pressures	Repetition and overloading	Madison Oral Strengthening Therapeutic (MOST) device: training to use device; calculate therapy targets; verbal encouragement; biofeedback.	3×/day, 3×/week, 8 weeks 10× each: anterior lingual reps and posterior lingual reps Detraining: 5 weeks Maintenance: 10 anterior & posterior reps, 3×/day, 1×/week, 9 weeks
	Namiki, 2019	Enhance tongue and suprahyoid muscle function	Pushing tongue against palate to produce pressure	Provide instruction; observe and check hyoid elevation via VFSS during tongue pressure resistance training; record and submit training history.	5 reps (10 s each), 2×/day, 4 weeks

Note. CDT = Conventional Dysphagia Treatment; SEMG = Surface Electromyography; UES = Upper Esophageal Sphincter; hwmk = homework; VFSS = Videofluoroscopic Swallow Study.

^a Article is mentioned in more than one location in chart.

^b Article is mentioned in the multi-approach chart as well.

Table 7.

Multi-approach studies described according to the Rehabilitation Treatment Specification System (RTSS) framework.

Rehabilitation treatment specification system: Multi-approach
Study (primary author, year)	Exercise description	Physiologic target(s)	Treatment ingredients
			Clinician ingredients	Dosage
Argolo, 2013	Sustained vowel phonation of /a/; pushing of plosive phonemes in a forceful manner to increase glottic closure; vertical range of motion of the larynx and posterior tongue to soft palate contact; sucking of wet gauze; swallowing while holding tongue; modified supraglottic maneuver; ascending and descending gliding phonations of the vowels /a/ and /u/; rotating each side of the tongue in the oral vestibule	Increase strength and range of motion of the mouth, tongue, larynx and pharyngeal structures; improve oral control of bolus and tongue-to-palate contact; improve coordination between breathing and swallowing; increase airway protection	Illustrated booklet to mark each exercise; supervision.	2×/day, 5×/week, 5 weeks 5 reps each ascending and descending gliding phonations of the vowels /a/ and /u/; 3 sets of 5 reps for each side of rotating the tongue in the oral vestibule; 10 reps of all other elements
Balou, 2019	Effortful swallow, Masako maneuver, supraglottic swallow, Shaker exercise, Mendelsohn maneuver, effortful pitch glides, daily homework (mix of unspecified swallowing and nonswallowing exercises)	Improve swallowing physiology manifesting in improved swallowing safety and efficiency	Provide instruction; modify dosage as necessary (incremental increase); homework.	45 min, 1×/week, 8 weeks 10 reps/set effortful pitch glide, 20 reps/set all other elements (total swallow exercises per treatment session = 110) Hmwk: 3 sets (330 swallow reps/day total)
Byeon, 2020	Combined NMES and Mendelsohn maneuver	Increase movement of larynx, hyoid bone, and swallowing in the pharyngeal stage	VitalStim electrical stimulation therapy; provide instruction; administration of a small amount of food; practice time.	30 min (2× 15 min of each)/session, 8 weeks
Cha, 2010	Flexibility exercises for the neck and temporomandibular joint with gentle joint mobilization, tactile oral stimulation and gentle massage, passive and active exercise of the oropharyngeal and laryngeal muscles, and supraglottic swallowing maneuvers	Establish acceptable conditions necessary for safe swallowing	Diagnostic examinations and observations to develop specific treatment strategy for participants.	30 min/day, 3×/week, 7 months
El-Tamawy, 2015	NMES + therapeutic exercise program (Maintain correct normal position of the head and neck with the chin tucked toward the chest. Push the tongue against resistance; round and stretch lips; open jaw widely, hold and move to both sides against resistance; open mouth and elevate the tongue, then hold the tongue and release; tilt head forward until initiation of swallow; manipulate large bolus in mouth; put tongue on alveolar ridge and swallow with upward–backward push of the tongue; extend tongue, hold and pull backward; take a deep breath, hold and release through a cough; vibrate laryngeal musculature from under the chin, downward to the sternal notch; push the neck and head in flexion, extension, and lateral flexion against resistance; sour and cold bolus; grasp tongue wrapped in gauze and pull forward, then stroke firmly down the middle of tongue with the edge of a tongue blade)	Strengthen and stimulate the elevator muscles of the larynx above and below the hyoid bone	VitalStim electrical stimulation therapy; provide instruction; provide physical resistance, physical manipulation of tongue, practice time.	70–75 min/session (therapeutic exercise program L 45 min, NMES: 30 min), 3×/week, 6 weeks 10 reps/exercise/session
Holmes, 2012	NMES in conjunction with swallowing therapy tasks, including strengthening exercises and compensatory strategies	Target the mylohyoid, digastric, and thyrohyoid muscles, and potentially the hyoglossus and upper and middle pharyngeal constrictors, if adequate intensity is reached	Patient 1: VitalStim electrical stimulation therapy; NMES paired with tongue-hold maneuver and effortful swallow. CDT: safety instruction; diet modification.	~15 min/session, 12 sessions (9 with NMES), 16 days
			Patient 2: VitalStim electrical stimulation therapy; NMES paired with effortful swallow; provide instruction; diet modification. CDT: NMES paired with chin tuck and multiple swallows.	~19 min/session, 9 sessions, 12 days
			Patient 3: VitalStim electrical stimulation therapy; NMES paired with effortful swallow; provide instruction. CDT: instruct in safety of oral diet; diet modification. Electrode placement 3a with average current of 4.5 mA.	~19 min/session, 7 sessions (6 with NMES), 8 days
Jang, 2019	Mechanical inspiration and expiration (MIE) exercises; individual feeding therapy and dysphagia treatment based on the results of VFSS. CDT: oral motor and sensory stimulation, neuromuscular electrical stimulation of the suprahyoid muscle, and oral and lingual exercises)	Improve coughing ability and strengthen oropharyngeal muscles	CNS-100 Cough Assist; adjust and titrate CNS according to participant condition; provide instruction.	5×/week, 2 weeks -MIE: 1×/day, 30 min/session -CDT: 2×/day, 30 min/session
Jiang, 2017	TRI-BALL breathing exercises, Sounder training, super-supraglottic swallow exercise, training of the vocal fold motion, tongue resistive training, Masako maneuver, effortful swallow exercise, NMES, swallowing maneuver modification (tightly held deep inhalation, consumption of food, intense swallowing while lowering head and holding breath), training and medication to prevent reflux	Improve oropharyngeal bolus propulsion and the anterior hyolaryngeal traction	TRI-BALL breathing exerciser device; VitalStim electrical stimulation therapy; acid inhibitor and gastro-kinetic agent; provide instruction.	40 days: -TRI-BALL: 50 exhalations and inhalations/session, 2×/day; -Sounder training: 20 vocal sounds/session, 2×/day; -NMES: 20 min/session, 2×/day, 5×/week -Swallowing maneuver modification: 90 min/session, 10×/week
Kang, 2012	Oral exercises: lips, tongue, and jaw exercises. Oral-pharyngeal exercises: tongue movement; soft palate exercise; Shaker exercise. Laryngeal exercises: airway closure, vocal cord adduction, and breathing exercises. Respiration exercises: effortful swallowing and supraglottic swallowing	Increased volume and strength of muscles and enhanced cooperation of the affected swallowing muscles	Provide explanation of purpose; provide training to nurses; video recordings of the swallowing exercise; daily check-in of implementation by nurses.	Tactile-thermal stimulation: 30 min/day, 5×/week, 2 months Exercise program: 1 hr/day, 2 months
Kraaijenga, 2017	CTAR, jaw opening against resistance (JOAR) exercise, and effortful swallow exercise using the Swallow Exercise Aid (SEA) device.	Improve hyolaryngeal elevation, improve amount of upper esophageal sphincter opening, improve time for pharynx passage, increase tongue base retraction, and decrease pharyngeal residue	Swallow Exercise Aid (SEA) device; written instruction sheet; instructional visits; progression of intensity/resistance based on interim swallowing strength measurements and self-perceived exertion; exercise log track performance.	3×/day, 6–8 weeks CTAR and JOAR exercises = Isokinetic: 30 reps, 1 s/contraction Isometric: 3 reps, 60 s each with 60 s rest Effortful swallow exercise = 10 reps
Lazarus, 1994	Oral tongue range of motion and strengthening exercises, base of tongue movement exercises, the supraglottic swallow, and the Mendelsohn maneuver	Not reported	Provide instruction	Inpatient: 1×/day Outpatient: 1×/week
Lin, 2011 a	Range of motion exercises (hold lips/tongue maximum extent of movement for 5 s); resistance exercise (push lips/tongue against a tongue blade for 5 s); tongue-hold exercise (swallow while keeping tip of tongue nipped softly between the teeth); effortful swallow; Shaker exercise	Improve tongue and pharyngeal muscle strength	Provide instruction; phone calls to participants to ensure maintenance of the exercise program.	Each exercise: +10×/session, 2×/day
J.-S. Park, 2019	NMES combined with electromyographic biofeedback (EMG-BF)	Strengthen muscles associated with swallowing, facilitate reflex swallowing by sensory stimulation, and increase rate of motor learning	VitalStim electrical stimulation therapy; calculate set threshold; customize treatment parameters; provide instruction; set intensity level; visual and auditory feedback.	30 min/day, 5×/week, 4 weeks

Note. NMES = Neuromuscular Electrical Stimulation; CDT = Conventional Dysphagia Treatment.

^a Article is mentioned in single approach chart as well.

Statistical Analyses and Study Designs

In the majority of articles, swallowing physiology outcomes were measured objectively and analyzed statistically. However, seven of the 43 (16%) articles reported subjective observations as outcomes (Carnaby-Mann & Crary, 2010; Cha et al., 2010; El Sharkawi et al., 2002; Holmes et al., 2012; Jiang et al., 2017; Juan et al., 2013; Lazarus et al., 1994). For example, Carnaby-Mann and Crary (2010) reported changes descriptively and did not report conducting statistical tests of measures that were mappable to MBSImP components. Of the studies that reported statistical test results, 22 of 35 single-approach studies (68%) and nine of 13 multi-approach studies (69%) did not have study designs that permitted between-group comparisons and relied on pre–post test measures within the experimental group (see Tables 4 and 5). In this review, between-group results were reported when available and within-group results were only reported when between-group comparisons were unavailable.

Single Approach Results

Across all single-approach studies, 117 outcome measures could be mapped to MBSImP components. Refer to Table 4 for results of each exercise approach and to Figures 3 and 4 for summaries of the evidence.

Figure 3.

Evidence map showing statistically significant improvements by Modified Barium Swallow Impairment Profile (MBSImP) component and approach. The numerator of each fraction refers to number of times component improved, and the denominator refers to number of times component was measured. The green boxes represent unanimous support for that approach resulting in improvement in the MBSImP component. Yellow boxes represent conflicting evidence, where at least one study found an improvement for a particular component while other studies using the same approach did not. The red boxes represent no reported evidence of improvement. The blank white boxes represent the components that were not investigated by any study within an approach. CTAR = Chin Tuck Against Resistance; DSAS = Dry Swallow After Suction; EMST = Expiratory Muscle Strength Training; JOE = Jaw Opening Exercises; LSVT = Lee Silverman Voice Treatment; MDTP = McNeill Dysphagia Therapy Program; NMES = Neuromuscular Electrical Stimulation; PNF = Proprioceptive neuromuscular facilitation; RST = Respiratory Swallow Training; HLE = Head Lift Exercise; TRT = Tongue Resistance Training.

Figure 4.

Table of reported and calculated effect sizes, where effect sizes in red indicate smalls effects, those in orange indicate medium effects, and those in green indicate large effects. Asterisks (*) indicate effect sizes associated with statistically significant results, degree symbols (°) indicate between-group results, and periods (.) indicate where effect sizes could not be calculated but outcomes were measured. The blank white boxes represent the components that were not investigated by any study within an approach. The studies associated with each effect size are indicated in square brackets, corresponding back to Tables 2 and 4. MBSimP = Modified Barium Swallow Impairment Profile; CTAR = Chin Tuck Against Resistance; DSAS = Dry Swallow After Suction; EMST = Expiratory Muscle Strength Training; JOE = Jaw Opening Exercises; LSVT = Lee Silverman Voice Treatment; MDTP = McNeill Dysphagia Therapy Program; NMES = Neuromuscular Electrical Stimulation; PNF = Proprioceptive neuromuscular facilitation; RST = Respiratory Swallow Training; HLE = Head Lift Exercise; TRT = Tongue Resistance Training.

Multi-Approach Results

Across all multi-approach studies, 60 outcome measures could be mapped to MBSImP components and there was statistically significant improvement observed for 13 components. There was not unanimous support for improvement in any one component (see Table 5 and Figure 3).

Decline in Measures of Swallowing Function

There were two instances of statistically significant worsening of anterior hyoid excursion following rehabilitative exercise. One instance was following NMES (Langmore et al., 2016) and the second following a multi-approach that involved range of motion exercises, resistance exercises, tongue-hold exercises, effortful swallow, and Shaker/HLE (Lin et al., 2011; see Tables 4 and 5).

Summary and Evidence Maps

A summary of results presented as evidence maps can be found in Figures 3 and 4. The first evidence map in Figure 3 shows all the statistically significant improvements by MBSImP component and exercise approach. When between-group results were available, these were used to report improvements, otherwise within-group results were used. Of the 15 MBSImP components evaluated for the purposes of this review, nine (53%) were mapped to studies that demonstrated an improvement in swallowing physiology postintervention for that component. In general, MBSImP Components 8 (laryngeal elevation) and 9 (anterior hyoid excursion) were most often reported to improve across all studies (reported 11 and 9 times, respectively); however, those components were also most frequently measured. When accounting for frequency of each component occurring across all studies, laryngeal elevation (Component 8) remained the most frequently improved but was followed by epiglottic movement (Component 10) and then initiation of pharyngeal swallow (Component 6). As can be ascertained from Tables 7 and 8, although these components showed improvement postintervention, they were not explicitly stated as the intended targets. Lip closure was reported in 21% (n = 10/47) of studies (Balou et al., 2019; Byeon, 2020; Cha et al., 2010; Kang et al., 2012; Kim et al., 2015; Martin-Harris et al., 2015; J.-S. Park, An, et al., 2018; H.-S. Park, Oh, et al., 2019; J.-S. Park, Hwang, et al., 2019; A. Park et al., 2021), yet no evidence for improvement was found and no studies explicitly stated that lip closure was the target of the intervention. Very few studies included outcome measures that could be mapped to pharyngeal stripping wave (Component 12; [n = 5/47; 11%]) and pharyngeal contraction (Component 13; [n = 2/47; 4%]), and no evidence for improvement was found across the studies targeting these aspects of physiology.

Table 8.

Levels of evidence and grades.

Approach	Study (primary author, year)	Level of evidence (CEBM, 2009)	Grade practice recommendation (American Society of Plastic Surgeons, as cited in Burns et al., 2011)
CTAR	J.-S. Park, An, et al., 2018	1B	C
DSAS	Koyama, 2016	4	C
Effortful swallow	H.-S. Park, 2019	1B	D
EMST	Hegland, 2016	2B	D
	Hutcheson, 2018	2B
	Troche, 2010	2B
JOE	Choi, 2020 a	4	B
	Koyama, 2017	2B
	Matsubara, 2018	2B
	J.-S. Park, 2020	2B
	Wada, 2012	2B
LSVT	El Sharkawi, 2002	2B	C
	A. Park, 2021	2B
MDTP	Carnaby-Mann, 2010	4	C
	Crary, 2012	2B
	Sia, 2015	4
Mendelsohn maneuver	McCullough, 2013	2B	D
NMES	Langmore, 2016	2B	C
	Lin, 2011 a	4
	J.-W. Park, 2012	2B
	J.-S. Park, 2016	1B
	J.-S. Park, Oh, et al., 2018	1B
	Toyama, 2014	4
PNF	Kim, 2015 a	2B	C
RST	Martin-Harris, 2015	2B	C
Shaker/HLE	Choi, 2020 a	4	C
	Kim, 2015 a	2B
	Logemann, 2009	2B
	Mano, 2015	2B
	J.-S. Park, 2017	2B
	Shaker, 2002	2B
TRT	Juan, 2013	4	C
	Namiki, 2019	2B
Multi-approach	Argolo, 2013	2B	C
	Balou, 2019	4
	Byeon, 2020	2B
	Cha, 2010	4
	El-Tamawy, 2015	2B
	Holmes, 2012	4
	Jang, 2019	1B
	Jiang, 2017	4
	Kang, 2012	2B
	Kraaijenga, 2017	2B
	Lazarus, 1994	4
	Lin, 2011 a	4
	J.-S. Park, 2019	2B

Note. Centre for Evidence-Based Medicine = CEBM; CTAR = Chin Tuck Against Resistance; DSAS = Dry Swallow After Suction; EMST = Expiratory Muscle Strength Training; JOE = jaw opening exercises; LSVT = Lee Silverman Voice Treatment; MDTP = McNeill Dysphagia Therapy Program; NMES = Neuromuscular Electrical Stimulation; PNF = Proprioceptive Neuromuscular Facilitation Based Short Neck Flexion Exercises; RST = Respiratory Swallow Training; HLE = head lift exercise; TRT = Tongue Resistance Training. 1B = Individual RCT (with narrow confidence intervals). 2B = Individual cohort study (including low quality RCT, e.g., < 80% follow-up). 4 = Case series (and poor quality cohort and case–control study). B = “Recommendation (Qualifying Evidence: Levels II, III, or IV evidence and findings are generally consistent; Implications for Practice: Generally, clinicians should follow a recommendation but should remain alert to new information and sensitive to patient preferences)” (American Society of Plastic Surgeons, as cited in Burns et al., 2011). C = “Option (Qualifying Evidence: Levels II, III, or IV evidence, but findings are inconsistent; Implications for Practice: Clinicians should be flexible in their decision making regarding appropriate practice, although they may set bounds on alternatives; patient preference should have a substantial influencing role)” (American Society of Plastic Surgeons, as cited in Burns et al., 2011). D = “Option (Qualifying Evidence: Level V evidence: little or no systematic empirical evidence; Implications for Practice: Clinicians should consider all options in their decision making and be alert to new published evidence that clarifies the balance of benefit versus harm, patient preference should have a substantial influencing role)” (American Society of Plastic Surgeons, as cited in Burns et al., 2011).

^a Article is mentioned in more than one location in chart.

Effect sizes that were either extracted directly from the articles or were calculated based on the data presented in the single-approach articles can be found in a second evidence map within Figure 4. Some studies did not include effect sizes and did not provide the information necessary to calculate effect sizes; therefore, no effect size data were available for MDTP, Respiratory Swallow Training or Tongue Resistance Training. Effect sizes were also unavailable for any studies that measured tongue control (Component 2), laryngeal vestibule closure (Component 11), pharyngeal stripping wave (Component 12), pharyngeal contraction (Component 13), and tongue base retraction (Component 15). As shown in the figure, some studies measured a single component in more than one way. In these cases, effect sizes were calculated for each unique measurement. For example, J.-S. Park et al. (2016) calculated both horizontal and vertical hyoid displacement when evaluating the efficacy of NMES, so two effect sizes were calculated. In total, 68 effect sizes were calculated, and over 50 were not reported or could not be calculated based on the data presented in the original manuscript.

Per the criteria set out by Gaeta and Brydges (2020), 19 of the 68 effect sizes (28%) can be interpreted as large, and 40 of the 68 effect sizes (59%) can be interpreted as small. These 19 large effect sizes represented 14 unique studies. When looking across MBSImP components, anterior hyoid excursion (Component 9) had the greatest number of large effect sizes (n = 8/19; 42%), followed by laryngeal elevation (Component 8; [n = 5/19; 26%]). When looking across exercise approaches, Shaker/HLE had the greatest number of large effect sizes (n = 6/19; 32%), followed by JOE (n = 3/19; 16%). Of the 19 large effect sizes, 13 (68%) also reported a statistically significant change and six (32%) did not. For example, Hegland et al. (2016) reported no statistically significant changes in hyoid excursion (Component 9) after completing EMST, but there was a large effect size for this measure.

Four of the 68 effect sizes (6%) can be interpreted as medium effects with statistically significant change, and five of the 68 effects sizes (7%) can be interpreted as small effects with statistically significant change. For example, the study examining effortful swallowing by H.-S. Park, Oh, et al. (2019) reported statistically significant changes in laryngeal elevation (Component 9) between groups postintervention, but the effect size was very small effect.

Levels of evidence were determined for each study and ranged from 1B, an individual RCT, to 4, a case series (see Table 8; CEBM, 2009). The most common level of evidence was 2B, representing an individual cohort study (n = 28/46; 61%). Grades were assigned to each exercise approach based on the number of studies evaluating the approach, the levels of evidence, and the associated statistically significant improvements in swallowing physiology (Burns et al., 2011). Single-approach study grades ranged from B (recommendation) to D (option). All multi-approach studies received a C grade.

RTSS

Tables 6 and 7 document the RTSS elements for each study included in this review (i.e., physiological targets, mechanism of action, and relevant ingredients). Table 6 presents single approach studies, and Table 7 presents multi-approach studies.

The consistent therapeutic target category across all studies was organ function targets. Treatment ingredients, including dosage (frequency, intensity, and duration) and treatment elements (clinician and participants activities), were extracted from studies as available. Most studies (16/35 [46%] single approach studies and 9/13 [69%] multi-approach studies) used devices to facilitate treatment, such as the Iowa Oral Performance Instrument, CTAR device, or NMES device. Other common ingredients seen across studies included providing instruction, calibration and/or adjustment of adjunct materials, practice, feedback, and training journals or logs. In general, studies of the same approaches had similar descriptions. For example, five out of the six Shaker/HLE studies (Kim et al., 2015; Logemann et al., 2009; Mano et al., 2015; J.-S. Park et al., 2017; Shaker et al., 2002) included the same dosage and had similar descriptions of the dosage and treatment elements.

There were no reported theories or models explicitly supporting treatment mechanisms of action. Although specific mechanisms of action were not explicitly mentioned in any study reviewed, rationale and discussion of cause and effect for at least one outcome measure were reported in all studies. In addition studies that included Lee Silverman Voice Treatment and MDTP lacked the transparency necessary to replicate any of the studies (Carnaby-Mann & Crary, 2010; Crary et al., 2012; El Sharkawi et al., 2002; A. Park et al., 2021; Sia et al., 2015). Detailed information regarding dosage (i.e., intensity, repetitions, and frequency) was also not consistently reported. For example, Lazarus et al. (2014) stated that the therapy was carried once per day without details regarding the length of the sessions or number of repetitions. While frequency was reported across all studies, number of repetitions carried out was reported in 25 out of 35 (71%) single approach studies and five out of 13 (38%) multi-approach studies. Intensity was only reported for the EMST studies (Hegland et al., 2016; Hutcheson et al., 2018; Troche et al., 2010).

Discussion

This mapping review explored the existing literature on dysphagia rehabilitation approaches and its impact on swallowing physiology, as measured based on VFSS images. The primary aim of this study was to generate evidence maps to identify which swallowing rehabilitation approaches improve which aspects of swallowing physiology per the MBSImP components. This evidence maps can be found in Figures 3 and 4. To use the evidence maps, a participant must first have a VFSS completed and analyzed according to the MBSImP protocol. Once the areas of impaired physiology have been identified, the evidence maps can be consulted to find an exercise approach with the best evidence for improving that aspect of swallowing physiology. For example, if pharyngoesophageal segment (PES) opening is the identified impairment, the clinician would refer to the evidence maps and note that JOE were consistently found to support improvement in PES opening (see Figure 3) and has three studies that demonstrated large effects for improving hyoid excursion (see Figure 4). The clinician can then refer to the results tables (see Tables 4 and 5) to identify the specific studies that reported this improvement. Next, the clinician can find the corresponding studies in the RTSS tables (see Tables 6 and 7) to learn what specific ingredients led to the improvement. This addresses the second aim of this review article, which was to use the RTSS framework to extract relevant information from the exercise approaches to support clinicians in understanding how to implement the approaches in practice. The third aim was to highlight the rehabilitative approaches that require further research to map physiologic improvements. All the approaches could benefit from further research, especially since many approaches only had one or two studies involving patients with dysphagia who underwent a VFSS; many studies only had one or two outcome measures that could be mapped to the MBSImP components; and many had mixed evidence supporting their effectiveness. On the basis of the effect sizes presented in Figure 4, we see that few outcomes had large effect sizes and mixed results (i.e., large effect sizes with nonsignificant statistical results or small effect sizes with statistically significant results) are not uncommon. Many of the included studies that reported nonsignificant results consisted of relatively small sample sizes; however, some of these studies also reported medium or large effect sizes, which may suggest meaningful change in swallowing impairment. In other words, even in the absence of statistical significance, it may be that certain swallowing interventions are positively impacting outcomes despite small sample sizes.

Determining which approach or approaches are best overall is challenging as it depends on which aspect of swallowing physiology is impaired and requires consideration of the state of the evidence. Therefore, to help clinicians select an approach from among the many options that may lead to improvement in a specific aspect of impaired physiology, the following should be considered:

• impaired physiology, ideally per an instrumental assessment;

• the evidence supporting each approach;

• the resources available, per facility;

• the training available to implement the approach, per what was learned in graduate school and currently available continuing education courses; and

• the patient's personal factors (e.g., diagnosis, goals, and values).

Previous systematic reviews have evaluated the effects of various swallowing interventions in patients with dysphagia (Ashford et al., 2009; McCabe et al., 2009), which have helped to guide treatment decisions. However, our review is unique in that it directly maps swallowing physiology using the MBSImP framework to rehabilitative intervention approaches, with consideration of the magnitude of change that we can expect. Furthermore, this review is not specific to medical etiology; rather, it considers the possible applications of interventions more broadly to assist clinicians and researchers in better understanding the benefits of each intervention approach.

There are also other more recent narrative reviews that have discussed dysphagia treatment approaches across medical etiologies (Langmore & Pisegna, 2015; Vose et al., 2014). In a paper discussing an overview of exercises and compensatory approaches, Vose et al. (2014) summarized treatment approaches and the physiologic impairments that they target. While some of the findings agree with those reported in this review, the majority conflict in terms of improvements in swallowing function observed with certain treatment techniques. This may be because some of the literature included in review by Vose et al. focused on healthy participants, which were excluded in this review. For example, Vose and co-authors (2014) reported that the effortful swallow improves initiation of pharyngeal swallow (Component 6), laryngeal elevation (Component 8), anterior hyoid excursion (Component 9), epiglottic movement (Component 10), laryngeal vestibular closure (Component 11), PES opening (Component 14), tongue base retraction (Component 15), and esophageal clearance (Component 17). Only two of these components were measured in the single effortful swallow study included in this review (H.-S. Park, Oh, et al., 2019); however, neither initiation of pharyngeal swallow (Component 6) nor laryngeal elevation (Component 8) were found to improve. Similar discrepancies were found for both the Shaker/HLE and the Mendelsohn maneuver. These discrepancies may have been impacted by individual study design, specifically participants included, outcome measures used, and/or variations in the exercise protocol. For example, a hypothetical study involving people with a neurodegenerative disease, such as dementia, may have goals and outcomes focused on maintenance of function, versus a study involving stroke patients, which would likely focus on improvement of function. These discrepancies highlight the need for careful consideration of study design to help explain the differences in results found and reported across studies.

Langmore and Pisegna (2015), in their narrative review, summarized a selection of articles that met their criteria of both swallowing and nonswallowing exercises, the majority of which are included in this review. While there was some overlap in the findings, results were not entirely consistent because, due to the scope of this review, changes in swallowing kinematics not represented by a MBSImP component were not captured. For example, Langmore and Pisegna (2015) discussed the Mendelsohn maneuver, reporting on the results of a single article published by McCullough and Kim (2013), also included in this review. Duration of hyoid movement was the only measure for which significant improvement posttreatment was observed but this could not be mapped to a component of the MBSImP given that the related component refers to amount of movement of the hyoid, rather than duration of movement. Another consistent treatment included in both this review and that of Langmore and Pisegna (2015) is EMST. Langmore and Pisegna (2015) reported on the results of a single article published by Troche et al. (2010), also included in this review, reporting improvement in participants' PAS scores and physiologic measures of swallowing. Like the Mendelsohn maneuver, statistically significant results of the EMST article discussed by Langmore and Pisegna (2015) could not be mapped to the MBSImP (i.e., PAS scores, duration of hyoid elevation, hyoid displacement measured at specific moments of the swallow); therefore, these improvements were not captured in this review. When using the evidence maps developed through this review, clinicians should consider that there may be aspects of physiology that are not captured; however, this does not mean these impairments cannot be addressed through swallowing rehabilitation and the approaches discussed in this review.

Limitations

This review had several limitations that require discussion. First, we limited the search to articles written in English and those that have been peer reviewed. As such, we likely missed strong research papers written in other languages and/or descriptions of studies within gray literature. Furthermore, while combining exercises and compensatory measures in swallowing therapy reflects clinical practice, it does not allow for clear extraction of data for the purposes of this review (i.e., specifying which ingredient was responsible for change). Therefore, it was impossible to understand the mechanism of action resulting in improvement—or in some cases decline—for multi-approach studies. Second, only outcome measures observed through VFSS were examined, and this review article does not consider other assessment approaches, such as endoscopy and/or mechanical or functional measurements (e.g., tongue pressure generation, temporal measurements, and manometry). Additionally, by limiting included studies to those that reported physiologic measures that could be mapped to the MBSImP framework, we have excluded evidence pertaining to other improvements in swallow physiology that may have relevance but could not be captured in this review. It is also important to note that changes in outcome measures were often measured in alternative ways to the MBSImP across studies (e.g., using measures involving frame-by-frame analysis rather than MBSImP scales). As an example, total oral and total pharyngeal VDS scores could not be mapped to the MBSImP. As such, future research needs to consider using standardized terminology and physiologic measures, rather than limiting outcomes reporting to safety and efficiency measures, to allow for the aggregation of data and meta-analyses. If there are questions regarding the exact outcome measures used within the studies included in this review, readers should refer to the original manuscripts detailing the approach. It is also important to note that etiology/medical diagnoses are likely to have varying impacts on swallowing physiology and more specifically, on MBSImP scores. Therefore, the influence of heterogeneity of disorders on the implementation of a singular clinical guide must be acknowledged. We also could not report effect sizes for several studies given the lack information available in the manuscripts, so it is possible a larger proportion of studies report large effect sizes. Additionally, we did not report whether studies included reliability assessments of measure reported. This is important to consider as inadequate reliability when deriving measures from VFS studies may impact the validity of results reported. Last, other types of targets, such as those relating to quality of life, were not investigated. As a result, there is the possibility that quality of life or other functional measures may have improved in participants even when physiology did not.

Future Directions for Research

Contributing to challenges in dysphagia treatment is the lack of standardization in assessing the presence and degree of swallowing impairments (Martin-Harris et al., 2008). When clinicians implement self- or facility-developed outcome measures that lack rigorous validation, this may lead to inappropriate selection of treatment goals and techniques (Carnaby & Harenberg, 2013). An important consideration is the clinician's degree of training and familiarity with available assessments, as this can influence which protocol is used, sometimes without proper evidence to support development of a treatment plan. Focus on standardization of physiologic assessment in research, such as the MBSImP (Martin-Harris et al., 2008), may benefit clinicians in accurately identifying impaired physiology. However, clinicians must realize and acknowledge that there may be other measures of swallowing physiology that are not captured within the MBSImP but are still important considerations when identifying impairments and developing a treatment plan.

When using this review in clinical practice, it is important to note that it includes evidence at all different levels. To create more robust and generalizable evidence maps, this field will need more well-designed, high-quality evidence that considers rigor and transparency. This includes research with larger sample sizes and that reports effect sizes. This will help to elucidate if swallowing rehabilitation is effective at improving additional components of swallowing physiology over and beyond the nine shown to improve within Figure 3 (per the green boxes).

This review highlighted that decisions pertaining to dysphagia rehabilitation are made primarily based on low levels of evidence (most often level 2B) that may prevent clinicians from being certain if a particular rehabilitation approach is more beneficial than the current standard of care. This review also found that many published articles test the efficacy of dysphagia exercises in healthy participants, which means the results may not be generalizable or applicable to clinical populations with dysphagia. This conclusion was supported by Langmore and Pisegna's (2015) review. We encourage future research to consider the clinical application of exercises for patients with dysphagia, and include more detail regarding aims, targets, mechanisms of action, and clinician ingredients, per the RTSS. Common missing items that may have contributed to the rigor and transparency of studies included rationale, treatment targets, consistent dosage, methods of recording and/or tracking participant adherence, and description of instructions provided to participants. Moving forward, research should support identification of dysphagia based upon the impairments in physiology rather than focusing solely on bolus-specific outcomes (e.g., aspiration and oropharyngeal residue). This allows clinicians to address the cause of dysphagia as opposed to only addressing the symptoms.

Conclusions

When treating dysphagia, clinicians should be making decisions based on the three pillars of evidence-based practice: current best evidence, clinical experience, and patient values (Dollaghan, 2019). This review supports clinicians in incorporating the pillar of current best evidence into their practice, as it found that there is a potential for improvement in many MBSImP components by using rehabilitative exercises. Discretion and careful consideration are required when balancing the weight of each evidence-based practice pillar and the validity, applicability, and generalizability of dysphagia rehabilitation approaches. It is important to recognize, however, that any evidence is better than no evidence. Results tend to look at groups as a whole and not the individuals involved, and therefore it is still important to consider the pillars of individual patient values and clinical experience when making clinical decisions.

This review was used to develop evidence maps to be used in conjunction with all aspects of evidence-based practice to help clinicians select recommended exercises to their patients' physiologic impairments. We intend for continued evolution of this review with emerging research findings. To perform a more robust and generalizable mapping review, more intervention studies are needed to ensure that any given rehabilitation approach is beneficial for improving swallowing physiology. Further consideration should also be made for the aspects of swallowing physiology that could not be mapped to MBSImP components and were consequently not discussed in this review. This should serve as a call to the dysphagia research community to consider how standardization of terminology, reporting of physiologic measures, and reporting of effect sizes could benefit our ability to aggregate data and perform meta-analyses. Standardization would allow us to truly determine the effect of our treatment approaches on swallowing despite small samples sizes.

Author Contributions

Ashwini Namasivayam-MacDonald: Conceptualization (Lead), Formal analysis (Supporting), Methodology (Lead), Resources (Lead), Supervision (Lead), Writing – original draft (Supporting), Writing – review & editing (Lead). Megan Rapley: Formal analysis (Lead), Methodology (Equal), Writing – original draft (Lead), Writing – review & editing (Lead). Josephine Stewart: Formal analysis (Lead), Methodology (Equal), Writing – original draft (Lead), Writing – review & editing (Equal). Eryn Webster: Formal analysis (Lead), Methodology (Equal), Writing – original draft (Lead), Writing – review & editing (Equal). Christina Quon: Formal analysis (Lead), Methodology (Equal), Writing – original draft (Equal). Nicole Rogus-Pulia: Conceptualization (Lead), Methodology (Lead), Supervision (Lead), Writing – review & editing (Lead).

Appendix A

Acronyms Used in This Mapping Review

Acronyms Used in Paper

CDT = Conventional Dysphagia Treatment CTAR = Chin Tuck Against Resistance DSAS = Dry Swallow After Suction EMST = Expiratory Muscle Strength Training HLE = Head Lift Exercise IOPI = Iowa Oral Performance Instrument JOE = Jaw Opening Exercises LSVT = Lee Silverman Voice Treatment MBSImP = Modified Barium Swallow Impairment Profile MDTP = McNeill Dysphagia Therapy Program MIP = Maximum Isometric Pressures NMES = Neuromuscular Electrical Stimulation PAS = Penetration–Aspiration Scale

PES = Pharyngoesophageal Segment PNF = Proprioceptive Neuromuscular Facilitation Based Short Neck Flexion Exercises RCT = Randomized Controlled Trial RST = Respiratory Swallow Training RTSS = Rehabilitation Treatment Specification System SLP = Speech-Language Pathologist TPPT = Tongue Pressure Profile Training TPSAT = Tongue Pressure Strength and Accuracy Training TRT = Tongue Resistance Training UES = Upper Esophageal Sphincter VDS = Videofluoroscopic Dysphagia Scale VFSS = Videofluoroscopic Swallow Study

Appendix B

Search Strategies for Included Databases

CINAHL	Ovid Medline	Ovid Embase
MH “Deglutition Disorders” “dysphagia*” TX (deglut* OR swallow*) N2 (impair* OR disorder* OR difficult* OR problem*) 1 or 2 or 3 MH “Fluoroscopy+” “videofluoroscop*” “modified barium*” “dynamic swallow stud*” “cookie swallow*” “videofluorograph*” 5 or 6 or 7 or 8 or 9 or 10 4 and 11	Deglutition Disorders/ “dysphagia*”.mp. ([deglut* OR swallow*] adj2 [impair* OR disorder* OR difficult* OR problem*]).mp. 1 or 2 or 3 Fluoroscopy/ “videofluoroscop*”.mp. “modified barium*”.mp. “dynamic swallow stud*”.mp. “cookie swallow*”.mp. “videofluorograph*”.mp. 5 or 6 or 7 or 8 or 9 or 10 4 and 11	Dysphagia/ “dysphagia*”.mp. ([deglut* OR swallow*] adj2 [impair* OR disorder* OR difficult* OR problem*]).mp. 1 or 2 or 3 Fluoroscopy/ “videofluoroscop*”.mp. “modified barium*”.mp. “dynamic swallow stud*”.mp. “cookie swallow*”.mp. “videofluorograph*”.mp. 5 or 6 or 7 or 8 or 9 or 10 4 and 11

Funding Statement

This work was supported in part by National Institutes of Health (NIH) Grant 1K76AG068590 (awarded to N.R.P.). The article was partially prepared at the William S. Middleton Veteran Affairs Hospital in Madison, WI; GRECC manuscript XXX-2022. The views and content expressed in this article are solely the responsibility of the authors and do not necessarily reflect the position, policy, or official views of the Department of Veteran Affairs, the U.S. government, or the NIH.