2.0 IntroductionThe following chapter describes the methods and the protocols, which areundertaken to direct the systematic review. It also states the procedureinvolved in data collection, concluding with a description of how data wasprepared for analysis (Calabrese, 2009).
The procedure followed the guidelinesprovided by the Cochrane Collaboration Handbook for Systematic Review ofInterventions (Higgins & Green, 2011) where it is appropriate.2.1 ResearchMethodology and DesignThe type of research methodology,which is utilized in this study, is a Systematic review. A Systematic Review isa means of identifying, evaluating and drawing conclusions from availableresearch pertaining to a topic of interest (Kitchenham, 2004). The selection of the most appropriate study design relies on anunderstanding of the project’s purpose and complexities, and subsequentlychoosing the design which best accommodates these parameters (Kennedy N. Otwombe, 2014). Using inappropriate studies can lead to gathering of improper resultsand drawing of improper or biased conclusions.
The objectivity of the researchis to investigate the effectiveness of the chintuck manoeuvre in reducing or preventing aspiration in individuals withneurological impairment, in order to identify if this practice is justified ina clinical setting. Also, there continues to be a discrepancy in its use forneurologic population. Therefore the systematic review process provides amethodology through which the specific objectives of this study can beaddressed, and was thus the chosen methodology. 2.2 Criteriafor considering studies for this review 2.2.1 Types of studies Athoroughly designed systematic review limits inclusion criteria to randomizedcontrolled trials (RCTs) and Controlled Clinical Trials (CCTs) as they arethought to provide a higher level of unbiased information regardingdifferential intervention effects than other study designs(O’Connor, 2011). Allpublished and non-published RCTs and CCTs were evaluated for this systematicreview.
RCTswere defined as all trials that involved at least one test treatment aimed atimproving or eliminating aspiration and one control treatment; or no treatmentwith concurrent enrolment and follow up of the test and control treated groups;as well as trails in which the treatment to be administered is selected by arandom process, such as a random number tables (Lefebvre, 2011).CCTswere classed as all trials that involved at least one test treatment aimed atimproving or eliminating aspiration and one control treatment; or no treatmentwith a non-randomized but bias free method of assigning patients to the testtreatment (Lefebvre, 2011).As biasis the most serious threat to credibility of evidence, RCT studies have beenconsidered as having the highest level of evidence (Level 1).
The trials forinclusion were not restricted by language or date filter so as to make surethat all trials examining this intervention were found. Studies from past andtrials in different languages were not excluded as evidence from all studiesthat needs to be assessed. Studies from inception to the date of analysis ofresearch were included.2.2.2 Typesof participants Pre-specifyingcriteria for including and excluding studies in a review (eligibility criteria)is an important feature of systematic reviews (O’Connor, 2011). People above18 years of age were included in the review because of biological plausibilityand the existing controversy, which are present within adults (Terre? R, 2012;Macrae P, 2014; Solazzo A, 2012; Shanahan TK, 1993). Adults with a clinicaldiagnosis of any neurological disorder, who have symptoms and signs ofdifficulty swallowing and in whom aspiration or penetration has been confirmedby a full clinical bedside evaluation, video-fluoroscopy or fiber-opticexamination of swallowing (FEES) using valid reliable measures, whereavailable, such as the Penetration-Aspiration Scale (Rosenbek 1996).
Participants who suffer any type and severity of the neurological disorder wereincluded in the review. Also, Patient with head and neck cancer were excludedfrom the review because of the post-operative changes in the anatomy andphysiology, which results in the change of swallowing pattern (McCulloch,2006). Inclusion Criteria Exclusion Criteria Adults aged 18 and over (male and female) seen in any setting. Participants under 18 years of age. Adults with a clinical diagnosis of any neurological disorder. Adults with history of head and neck cancer.
Adults with any type and severity of neurological disorder Adults having history of radiation therapy. 2.2.3 Types of Interventions Theintervention included in this systematic review was chin tuck, which aimed to reduceor prevent aspiration.
This intervention was allowed to practice in any setting(generally in a medical setting for e.g. clinic, hospitals), and must have beendelivered by trained individual or team. Immediate, medium or long-termimprovements in drooling, as well as any positive or negative effects to chintuck intervention were taken into consideration.Theintervention considered in this review includes: 1. Interventionversus no intervention (i.e.
chin tuck versus normal swallow)2. Interventionvs. other intervention (i.e. Chin tuck versus other compensatory strategies fore.g. head back, head rotation) 2.
2.4 Types of outcome measures Theprimary outcome of this systematic review was to see the positive or negativechange in aspiration for individuals with dysphagia with neurologicalaetiology. Both the desirable and undesirable effects of these outcomes wereconsidered in relation to the current intervention. Loke, Price and Herxheimer(2011), reported that in order to achieve a balanced perspective; all reviewsshould consider the adverse aspects of intervention.
The secondary outcomes inthe review were to see the positive or negative changes to oral and pharyngealphase functions, which are measured using Video-fluoroscopic evidence. Also,positive or negative changes to lip, tongue and other structural movementsinvolved in swallowing.Primaryoutcome measures:1. Changein aspiration for individuals with dysphagia with neurological aetiologySecondaryoutcome measures:1.
Changesto oral and pharyngeal phase functions2. Changesto lip, tongue and other structures 2.3 Search methods for identification of studies The chosensearch strategy aimed to identify all available studies, published andunpublished. Search for studies were not restricted to electronic databases,but were expanded to other sources in order to maximize the accuracy of thesearch and minimize publication bias. This ensured that the search strategy wasthorough and detailed, covering all relevant studies.The Meshsearch words like “chin”, “dysphagia”, “dysphagic”, “deglutition”,”swallow”, “swallows”, “swallowing”, “swallowed”, “deglutition disorder”, “chindown”, “chin tuck”, “head forward”, “manoeuvre”, “posture”, “effectiveness”,”evaluation”, “video fluoroscopy”, “Neurological” were usedand these words were entered into the controlled vocabulary for indexing,specific to each database search. These terms were combined using standardBoolean operators “OR” and “AND”. It was seen that these search term cover therelevant studies that were needed for this review.
The main purpose of thedatabase indexing is to yield the accuracy of search results (Murphy, 2003). A comprehensive Title/Abstract searchincluding medical subject headings (MeSH) terms were completed in PubMed tobegin the search. MeSH terms are controlled, indexed vocabulary used to captureall possible interpretations of a concept (Cooper, 2014).2.
3.1 Electronicsearches Thefollowing twelve electronic databases search were included for the review;Academic Search Complete; AMED; CINAHL; the Cochrane Library; EMBASE; GoogleScholar; HSE Library; PsycINFO; Cochrane Central Register of Controlled Trials(CENTRAL); SCOPUS; MEDLINE; and Web of Science. Ongoing clinical trials weresearched on the clinical trials website (http://clinicaltrials.gov) andongoing controlled trials were searchedon (http://www.isrctn.com/). TheProQuest Dissertation and Thesis was reviewed for relevant abstracts.
Databasewere searched from inception up to and including May 2018. It was addressedthat biased conclusion from studies could result from reviews that omit studieswith delayed publication (Guyatt, 2011); a fact often termed “Time Lagbias” (Hopewell et al., 2007). Therefore, arepeated search in all the database was conducted just before the completion ofthis review. 2.3.
2 Searching other resources Asingle data search is not sufficient as it can introduce the risk of omittinglarge quantities of relevant literature. Additional search methods are alsorequired. Searches were not limited to electronic searches, to avoidpublication bias. Publication bias is defined as a selective publication ofstudies with a particular outcome, usually those that are statisticallysignificant, at the expense of null studies (Ferguson & Brannick, 2011). Grey literature, defined by Happe and Walker(2013) as information, which is not controlled by commercial publishing, it wasalso explored. These measures were taken as a part of the search strategy for thissystematic review in order to further minimize the risk of publication bias.Also, additional separate searches in many of the above sources were done toensure that we get the most up to date information. I.
For the review grey literature search were: · ISI webof knowledge conference proceedings · Indexto Theses· AustralasianDigital Theses. II. The following Journals were hand searched: · Dysphagia· BMCGeriatric· Americanjournal of Speech-Language Pathology· TheAmerican Journal of Occupational Therapy· OpenJournal of Therapy and Rehabilitation· ClinicalOtolaryngology· Journalof Medical Speech-Language Pathology· Journalof Oral Rehabilitation III. International Journal of Language and Communication Disorders.
IV. The “related article” function in PubMed wasused in order to enhance the search V. ProQuest Database was searched for unpublishedthesis and dissertations abstracts. VI.
Lastly, the reference list from all otherincluded studies was searched to identify other relevant studies. 2.4 Data collection and analysis 2.4.1 Selectionof studies Theresearcher initially screened all the titles and/or abstracts of studiesretrieved using the search strategy and those from additional sources werescreened independently by two review authors to identify studies thatpotentially meet the inclusion criteria outlined above.
Thesearch results were merged using reference management software and duplicatedrecords of the same report were removed. Results from the search werecategorized as either ‘relevant’, ‘potentially relevant’ or ‘not relevant’. Studiesthat were unclear from titles and abstracts of whether it should be included,the third reviewer was told to obtain copies of the study for further analysis.Any kind of disagreement on the selection of studies was resolved by consensusdiscussion. If theeligibility of the study was in question, we contacted the authors of the studyfor additional information.
The review team was not blinded to informationabout authors, institution, and journal of publication or results.ThePRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses;Moher, Liberati, Tetzlaff, & Altman, 2009) flow diagram was used to map outthe inclusion and exclusion of studies.2.4.2 Data extraction and management A specificallydesigned standardized data form was used to extract data from the includedstudies for assessment of studies quality and for evidence synthesis. Foreligible studies, two authors extracted the data using the standard form. Thedata was entered in Review manager software (Revman 2012); to check theaccuracy of the data. Two authors will extracted the data independently, anddiscrepancies were identified and resolved through discussion (with a thirdauthor where necessary).
Missingdata will be requested from study authors. If no response is received from theauthors, then the studies were eliminated from the meta-analysis but they wereretained for further discussion.Thisaimed to reduce risk of bias at the stage of data extraction. No attempts weremade to mask for authorship, journal name, or institution. 2.5 Assessment of risk of bias in includedstudies These five domains of bias were examined: selection bias,performance bias, attrition bias, detection bias and reporting bias. The “CochraneCollaboration’s Tool for Assessing Risk of Bias” (Higgins & Green, 2011)was used to assess the criteria for risk of bias for each individuals studywhich were included.
The “Downs and Black checklist” (Downs SH, Black N, 1998) wasused to assess the quality of the Non- Randomized Controlled Trial studies. This data was then entered into Review Manager (RevMan:Computer program Version 5.3, and a risk of bias table was generated. Anyfurther data analysis in this review was carried out using RevMan. 2.6 Measuresof treatment effect Measuringof treatment effect involves combining results from different studies to createone overall numerical estimate to increase statistical power and lead to moreprecise estimates of treatment effect (Lau, 1997).
It wasdecided that in this Systematic Review to create forest plots to display effectestimates for the outcomes considered in included studies using RevMan. 2.7 Unit ofanalysis issues The unitof analysis is the level at which researchers collect data and compare outcomes(Pieper, Mathes and Eikermann, 2014).
The unit of analysis, which was ofimportance in this study, were the individuals participants having Dysphagiawith Neurological origin. For studies where only group data was available, itwas planned to contact the authors of the studies, to provide the individualpatient data. 2.8 Dealingwith missing data A highpercentage of missing data may undermine the scientific credibility ofconclusions and decrease the potential statistical power (Little, 2012).
It wasdecided that the authors would be contacted in possible cases where there wasnoted to be missing data from included studies. 2.9 Assessmentof heterogeneity Statisticalheterogeneity occurs when the total variation across study results isbeyond that predicted by chance (Higgins, 2002). Authors of this article plannedto conduct heterogeneity test using the Chi² test (significance: p < 0.1).Also, with the estimated impact of variation due to heterogeneity calculatedusing the I² statistic.
This describes the percentage of variability in pointestimates caused by heterogeneity rather than sampling error (Higgins, 2002). Ifstatistical heterogeneity is identified, a number of analysis were used toaddress this. If there is evidence of heterogeneity, then which factor causedit were explored and a sensitivity analysis was performed based on the possiblereasons.Thetrials, which did not show heterogeneity a meta-analysis was conducted as,planned by authors. This involves meta-analysis of primary and secondary end-pointsvia calculation of risk ratio and 95% confidence intervals for dichotomousoutcomes, and standardized mean differences and 95% confidence interval forcontinuous outcomes.
2.10 Data synthesis Meta-analysisis used to systematically review and combine results of similar studies,producing an overall summary estimate (Munn Z, 2014). Meta-analysisis widely employed in systematic reviews of intervention effectiveness, whenthe intervention is well accepted and explained(Felix Achana, 2014). It wasintended to conduct a meta-analysis if a minimum of two sufficiently homogenousstudies were defined (Ryan & Cochrane Consumers and Communication ReviewGroup, 2013) in this review.
2.11 Subgroup analysis and investigation ofheterogeneity Subgroupanalysis allows exploration of heterogeneity within studies and can allowresearchers to answer specific clinical questions regarding particularintervention effects (Deeks, 2011). A subgroupanalysis considering the treatment effects was planned taking intoconsideration the aetiology of individuals with dysphagia.
Ifsubstantial heterogeneity (indicated by a Chi² test p < 0.1 or an I²value >50%) was found to be present. Then the researchers intended to explore possiblecauses of heterogeneity (e.
g. patient age). 2.
12 Sensitivity analysis TheSensitivity analysis of methodological quality allows researchers to assessstability and robustness of conclusion made during primary analysis (Deeks, 2011).Sensitivity analysis requires repeated calculations with inclusion/exclusion ofparticular subsets of studies to assess if this effects overall results (Taylor, 2009). The authorplanned to conduct a sensitivity analysis to explore potential influences ofmethodological quality on effect sizes, if appropriate. 2.13 ConclusionThis chapter presented the research methodology employed inthis systematic review.
Considerations of study design; data collection,synthesis and analysis were discussed. The following chapter presents thefindings of the systematic review.