The proposed systematic literature review will be conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) reporting methods to ensure accuracy, completeness and transparency in planning [11]. A PRISMA-P checklist is presented in an Additional file 1 [11]. The systematic literature review was registered with PROSPERO with the number CRD42019132620. The PRISMA flow diagram for the protocol is presented in Figure 1 below.
Search strategy
A search strategy aimed towards locating both published and unpublished sources will be developed. An information scientist will assist the researcher with development of the search strategy on the appropriate databases.
Primary search strategies will search the databases accessible to the University and the electronic platforms. EBSCOHost databases that include: MEDLINE with Full Text, Academic Search Ultimate, CINAHL with Full Text, CAB Abstracts, Health Source - Consumer Edition, Health Source: Nursing/Academic Edition, SPORTDiscus with Full Text, Africa-Wide Information, and MasterFILE Premier will be searched.
Searches will further be performed on Web-of-Science and Scopus to ensure coverage of all available evidence. An example of a database with the search strategy keywords is presented in Table 1. Secondary search strategy will include: grey literature and unpublished sources. Two independent reviewers will screen the reference lists of included full text articles for potential inclusion of further eligible studies.
Table 1 Search strategy
Search
|
Query
|
Records retrieved
|
#1
|
Database: CINAHL
((boxer* or metacarpal*) n2 fractur*)
and
(exercis* or program* or protocol* or “functional rehab*” or rehab* or advis* or advic* or educat* or splint* or immobili* or physiotherap* or “physical therap*” or “occupational therap*” or outcome*)
|
|
Limited to January 2008 to September 2018, limited to English
|
|
Eligibility criteria for studies
This review will include experimental study designs, including RCT and quasi-experimental studies, observational studies including retrospective, prospective cohort studies and case control studies for the last ten years from January 2008 to September 2018. A language restriction will apply. Only English resources will be included. Studies including adult human participants older than 20 years and younger than 59 years of age will be included. Participants younger than 20 years will not be included due to skeletal immaturity [12]. Participants older than 59 years of age will not be included because of possible osteoarthritis [13]. Studies reporting on post-surgical and conservative hand rehabilitation interventions for single or multiple second to fifth metacarpal fractures will be included. The three main intervention categories related to hand rehabilitation that will be included are; hand rehabilitation programs, home education and immobilization. Hand rehabilitation programs include functional and/or non-functional exercises and other hand therapy modalities/treatments. Home education includes; advice, home education and home exercises. Immobilization interventions will include, but not be limited to, plaster of paris and/or splinting utilized. Studies comparing hand rehabilitation programs or exercises to home education, self-care programs and different immobilization/splinting programs, will be included.
Studies excluded will now be stated. Studies investigating thumb metacarpal fractures due to the difference in hand performance in rehabilitation. Second to fifth metacarpal fractures with an associated tendon injury, nerve injury or pre-existing osteoarthritis or rheumatoid arthritis due to different physiological healing timeframes acting as complicating factors. Studies investigating second to fifth metacarpal fractures with a concurrent fracture of the phalangeal bones, carpal bones, distal radius and ulna due to different management and joint kinematic involvement, which could be confounding variables and can possibly affect the generalization of results. Infections will be seen as a complicating factor to fracture healing and management in the instance where it is documented.
Outcomes measures
This review will consider studies that include primary outcomes of hand function, HRQoL and disability. Secondary outcomes included will be, digital ROM, grip strength and fine motor dexterity. Outcome measures per outcome may include but will not be limited to: HRQoL measured with the EQ-5D questionnaire, disability measured with the quickDASH or Disabilities of the Arm, Shoulder and Hand (DASH) questionnaires, hand function measured with the Jebson Taylor Hand Function Test or Michigan Hand Outcome questionnaire, grip strength measured with a Jamar hydraulic dynamometer, fine motor dexterity measured with a Nine-Hole-Peg-Test, Purdue Pegboard test or Bruininks-Oseretsky Test of Motor Proficiency and digital ROM measured with a Rolyan finger goniometer. When other outcome measures are identified in the studies, they will also be included.
Data management
Title and abstract screening, review of full-texts and data extraction using a piloted modified Cochrane document will be undertaken.
Study selection
Following the search phase, the citations of all identified sources will be imported into Endnote® X8 (Clarivate Analytics, PA, USA.) where duplicates will be removed. The two independent reviewers will screen all identified articles. The screening of all identified articles, against the inclusion criteria, will be based on both the title and abstracts of articles eligible for inclusion. Excluded studies will be documented in a table with detailed reasons for exclusion [14]. Full-text articles for potentially relevant article/document, which adhere to the inclusion criteria, will be obtained followed by a thorough screening of the article/document by two independent reviewers. Disagreements during any phase of the study selection process, will be resolved through discussion by the two reviewers, and a third reviewer with experience in systematic literature reviews will act as the deciding vote in the instance where reviewers’ discussion could not resolve the issue. Reporting of the results will be presented in the final systematic review with the PRISMA flow diagram to foster transparency in reporting of the results [15].
Data extraction
Data extraction from the included studies will be independently performed by two reviewers with a modified and piloted Cochrane extraction document [14]. Data extracted will include: population, participant demographics, participant numbers, participant’s characteristics, interventions, comparisons, controls, outcome measures, study methods, study design, sampling details, socioeconomic status, fracture level and fracture types. Disagreement will be resolved through discussions among the reviewers, or a third reviewer when no consensus can be reached.
Quality assessment
Quality assessment of all included studies will be undertaken by two independent reviewers at the outcome level, with a third reviewer acting as a deciding vote in disagreement. The JBI critical appraisal checklists for Experimental studies (RCT), Quasi-experimental studies, Controlled observational studies (including cohort studies) and Case control studies will be used [16]. The critical appraisal results will be reported in a narrative form and in a table format. The appraisal tools assess: selection bias, allocation bias, reporting bias, detection risk, performance bias and attrition bias. Response options on the appraisal checklists are: ‘yes’ (criteria appropriately applied), ‘no’ (criteria not applied), ‘unclear’ (criteria not described clearly) and ‘not applicable’. After reviewing each included study according to the criteria, the studies will be classified as follows: low risk of bias: criteria are all met, moderate risk of bias: one or more criteria not clearly reported and high risk of bias: one or more criteria not met.
Data synthesis and assessment of heterogeneity
A narrative summary report of the finding of included studies and a quantitative summary using meta-analysis, will be provided if the included studies are sufficiently homogeneous. An expression of either odd ratios for dichotomous data or weighted mean differences for continues data will be utilized and the confidence interval of 95% will apply for analysis. Heterogeneity is expected due to the diversity of included studies and thus will be evaluated by visual examination of forest plots and poor overlap between confidence intervals. The Chi2 (0.1 significance level) and I2 statistical test (0%-40% low importance, 30%-60% moderate, 75%-100% high heterogeneity) will be used (11). The X2 test of heterogeneity will determine whether the differences between results are due to chance alone. A random-effects model will be used when it has been decided to pool the results as some degree of heterogeneity is expected between studies. Where sufficient data is synthesized on variability in hand rehabilitation or immobilization/splinting effects across different participants or study types, a subgroup analysis will be performed. A sensitivity analyses on different ranges of variables will be conducted to test the outcomes of decisions made, when deemed appropriate.