El exilio como tiempo de silencio fecundo
7. La fecundidad del exilio
As described in the previous section, systematic reviews are distinguished by their rigorous methods. A study protocol is required to outline the methods for searching, study inclusion, data extraction and analysis (Moher et al., 2015).
3.3.2.1 Protocol registration
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Liberati et al., 2009; Moher et al., 2015) advocates that all systematic review protocols should be registered. This is to reduce duplication and increase transparency. The study protocol was registered with the PROSPERO International prospective register of systematic reviews (see appendix 1) and is published online at: http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42015019361. The methodology for this systematic review conforms to the PRISMA guidelines (Moher et
3.3.2.2 Search strategy
The search strategy aimed to find both published and unpublished studies. It was designed in consultation with an information specialist and involved three facets, i.e. ’conceptual grouping of related search terms’ namely ‘homelessness’, ‘leaving’ and ‘tenancy sustainment’ as well as their synonyms. Terms such as housing stability,
housing tenure, residential stability and housing maintenance were included to get a
comprehensive understanding of the literature. Multiple versions of search terms were tested and modified in line with each database and the providers indexing system. This search strategy was designed to be sensitive to find relevant studies and the search was carried out in May 2015 and updated in October 2016. Databases searched were AMED (EBSCO), CINAHL (EBSCO), EMBASE (OVID), MEDLINE (OVID), OTseeker, PsycINFO (ProQuest), Social Care Online, SocIndex , Web of Science Social Sciences Citation Index, Cochrane Library, Campbell Library and Joanna Briggs Institute. The following sources were searched for grey literature: relevant third sector and government websites (see appendix 2), EThOS, ProQuest Thesis Index, ProQuest Conference Papers Index, Web of Science Conference Proceedings Citation Index, Google Scholar and Google. Only studies published in English were included and there was no date restriction. Please see appendix 3 for the electronic search strategies. All references were imported into EndNote.
3.3.2.3 Inclusion and exclusion criteria
Two reviewers independently screened all titles and abstracts. Full text papers of any titles and abstracts that were considered relevant by either reviewer were obtained where possible. The relevance of each study was assessed according to the following inclusion and exclusion criteria stated in table 3.4. Tenancy sustainment was the primary outcome as the sustaining of tenancies is one of the primary means of minimising
homelessness (see section 1.3.4).
Inclusion criteria Study design All study types
Population Individuals (aged 16 years and over) who were sleeping rough, using homeless service accommodation or formerly homeless, now in their own tenancy, regardless of length of time.
Supporters and/or homeless service staff of participants.
Intervention All interventions that support individuals to leave homelessness and sustain a tenancy
Outcomes Tenancy sustainment following homelessness Table 3.4. A summary of the inclusion criteria for the systematic review
Studies were excluded if they focused on families only. As leaving homelessness as part of a family can have different influencing factors and any interventions or services should fit a family-centred model (Kilmer, Cook & Crusto, 2012), this review focused only on evidence for individuals. In addition, studies in which participants moved between emergency homeless accommodations were excluded as the phenomenon under review was leaving homelessness.
3.3.2.4 Screening
Two authors independently reviewed all full text articles and any discrepancies were resolved by consensus. Inter-rater reliability was considered good as the kappa coefficient assessed at full text stage (n=458) was 0.698. Whether a study had tenancy sustainment as its primary outcome was the primary issue of discussion amongst the reviewers where there was no agreement. Figure 3.3 presents the PRISMA flow diagram detailing the number of articles at each stage of the process.
392 full-text articles excluded
Not primarily leaving homelessness =123 Not primarily tenancy
sustainment = 113 Non research =94 Unable to source =46 Population not relevant =
15 Language =1
Figure 3.3. PRISMA flow diagram summarising the outcome of searching and assessment of relevance and quality (Moher et al., 2009)
1 survey 12,241 records identified through database searching in May 2015 &
Sept 2016
277 additional records identified through other sources in May 2015
& Sept 2016
5,974 duplicates removed
6,544 records screened at title & abstract stage
6,063 records excluded based on title & abstract
485 full-text articles assessed for eligibility 4 additional articles sourced through citation searching (n = 4) 93 articles included in qualitative assessment 43 articles included in review 15 qualitative studies (reported in 17 articles) 7 Quasi experimental 5 secondary analysis of data 1 RCT 7 cohort studies 2 mixed methods (reported in 3 articles) 2 literature reviews
3.3.2.5 Quality assessment process
The methodological quality of included articles was assessed using QualSyst, a validated quality appraisal tool developed by Kmet, Lee and Cook (2004). This tool facilitates the evaluation of both quantitative and qualitative studies, using two checklists of relevant questions (see appendix 4). Each study was scored against each question in the appropriate checklist, as guided by the manual. A score of 2 was assigned if the quality criterion is met, 1 if partially met and 0 if not met. Criteria include the choice of appropriate study design, definition of outcomes and exposures, reporting of bias and confounding as well as sufficient reporting of results and limitations. The QualSyst score is calculated by the sum of ratings divided by the maximum score (the maximum score is 1). Each study was independently assessed by two reviewers and any disagreements were discussed until consensus was reached. QualSyst does not specify a cut-off point below which papers should be discarded. It designates 0.55 as a relatively liberal quality score below which papers should be discarded but to optimise quality given the large number of articles eligible, a 0.65 cut off was agreed upon. This threshold allowed a balance between efficiency, inclusiveness and quality. The most common reasons quantitative study designs were excluded using QualSyst were that the study aims were not clearly identified, analytical methods were not reported or there was incomplete control of confounding variables. Qualitative articles were excluded mainly due to limited or no reporting of data analysis or lack of evidence of reflexivity.
3.3.2.6 Data collection process
A bespoke data extraction sheet was developed to gather the data within the studies relevant for the review questions (see appendix 5). Data was extracted into Excel using the headings: study aim, design, sample, data collection and findings. Two reviewers extracted the data to ensure accuracy.
3.3.2.7 Data management
EndNote was used and found to be effective at all stages of the screening process including: storing references, removing duplicates, using smart groups to filter references and as a data management system. The included studies (43 articles) were imported into Excel and extracted data added to the worksheet. The results of the systematic review will be discussed in chapter seven. The methodology of the second study of the thesis, the constructivist grounded theory, follows.