Hospitalisation was significantly reduced in intervention versus control 510 n2239 P002Selfmanagement tools

Hospitalisation was significantly reduced in intervention versus control 5%/10%, n=22/39. P=0.02.Self-management tools delivered during intervention workshop may explain why participants reported an ‘impression’ of improvement despite no symptom improvement due to autonomy gained by self-management. Researcher acknowledged that responsiveness of questionnaires may differ in older people who perceive ‘no loss’ of autonomy positively, due to attitudes about ageing (Fritel, 2018). This study highlights that educating and empowering patients helps autonomy. This features significantly in self-care models used by CMs (Department of Health (DoH),2005).Recall bias is possible as participants could change responses over time due to reduced motivation or difficulty in recalling workshops. Spencer et al (2017) says this occurs when participants cannot remember previous events; confounded in longitudinal studies or ageing. Although, Tolmie et al, (2004) counter that older people are highly motivated to participate and remain in trials when they believe they will receive health-monitoring. Systematic reviewA somewhat dated systematic review (SR) by Martin et al (2006), was used due to lack of recent evidence. Parahoo (2014) says that SR top the hierarchal pyramid for best evidence focusing on striving for evidence-based care; partly due to meta-analysis enabling conclusions to be drawn from large literature collections.Critical appraisal tool by CEBM (2014), for critique. The topic appeared broad, examining invasive (urodynamic) and non-invasive assessments for UI (Martin et al, 2006). This will demonstrably affect heterogeneity as multiple variables were included. This review selected 129 articles from 6009 identified. Grey literature and non-English articles appeared excluded; this could affect results; leading to publication bias if only favourable articles are used (Bowling,2014). This can be detected by funnel plots highlighting asymmetry (Bruce et al, 2018). Shaneyfelt (2013) maintains, SR’s are only as reliable as studies chosen.Detailed breakdown of evidence is provided. Meta-analysis model used with 95% CI. However, Martin et al (2006) admitted that many papers were unsuitable to measure specificity or sensitivity which increased heterogeneity; severely impeding the use of meta-analysis to produce meaningful data. Bruce et al (2018) explained that if results differ too widely, pooling them can lead to inaccuracies. Translating forest plots has proved challenging due to multitude of variables used and reader ineptitude. Generalisability is limited as review concerns predominantly female participants, 40+, whereas population of interest is mostly frail older people. However, review can be applied to practice as results demonstrate bladder diaries with clinical history and ultrasound-scanning are simple cost-effective interventions for UI, most appropriately performed in primary care (Martin et al,2006). This highlights how economic and impactive simple interventions are; easily performable by community services. QualitativeVaismoradi et al (2013) say qualitative research endeavours to understand phenomenon from the experience and perspectives of those experiencing it. Studies were evaluated using the CEBM critical appraisal of qualitative studies tool (CEBM, 2019). (Study summaries: Appendix 2). Hägglund (2010) and Ferdinand (2018) studied continence care delivery from the perspective of nurses. Whereas, Vethanayagam et al (2016), studied help-seeking behaviour of Continence patients. Hägglund (2010) noted that some participants ignored invitations; missing the views of these non-responders. This is responder burden if the perceived effort of completing questionnaires was too much Rolstad et al (2011). Conversely, study could be biased if surveys are completed because participants are interested in the topic. Whereas, responses may differ if participants had no opinion. Similarly, to Ferdinand’s (2018) study, where participants are nurses enrolled on a continence module; they were interested in continence so would not represent ‘all nurses’ views. Spencer and Bracey(2017), describe this as ascertainment bias as it seeks the views of ‘nurses knowledge’ regarding continence education (Ferdinand, 2018) but findings are based on continence students. This reduces the relevance to practice.Vethanayagam et al, (2016) commended their research for exploring help-seeking behaviour in varied settings. However, as participants were existing care-recipients they did not represent those who hadn’t sought help, therefore their ‘help-seeking’ behaviour cannot be compared to non-service users. Furthermore, study was not transferable to community caseloads as it excluded housebound participants. Transferability would have improved if frailer participants were included as evidence is only transferable to similar populations. This was arguably achievable as interviews mostly occurred in participant’s homes (Vethanayagam et al, 2016).Data collection Hägglund (2010) and Vethanayagam et al (2016) collected data from open-ended questionnaires which were transcribed and checked verbatim. Hägglund’s (2010) participants completed questionnaires in privacy so answers would be their own. However, Parahoo (2014) argued; questionnaires prevent researchers from clarifying or elaborating responses. Vethanayagam et al(2016) said research assistants conducted semi-structured, face-to-face interviews, until data saturation was reached. This strengthened reliability and validity as Pope and May (2009) concur that participants can respond differently knowing that interviewers are healthcare professionals. Ferdinand(2018) used focus-group based discussions using a case-study methodology. Pope and Mays (2009) acknowledge that groups can facilitate discussion. However, participants may answer dishonestly due to social desirability bias and may respond as they think is expected (Bowling,2014). Furthermore, groups could display the ‘Hawthorne affect’ where, as Spencer and Mahtani (2017) mention, participant’s behaviour may change knowing they are being observed. This