2. The Methodology
2.1 The Sample
There have been 656 referrals to the Sources of Support Service from its operational start at March 2011 to 31st Dec 2014. The cases reported represented a purposive sample of 100 cases that completed working with the Link Workers. The table below shows the distribution of the sample of referral/discharge sample:
|SOS 100 Referral Year||SOS 100 Referral No of Cases||Discharge Year||SOS 100 Discharge No of Cases*|
- Note 5 cases were still open as at end Jan 2015
80% of the SOS 100 completed cases were referred in 2013/14, with 73% of discharges taking place within the same time-frame.
In terms of the sample distribution between Link Workers and practices, the table below shows the sample construction - that is an equal number from each Link Worker and a practice distribution that reflects the practices the Link Workers had most commonly received their referrals from. Note that one practice had always referred more patients to the service than the other practices and that is reflected in the purposive sample.
|Link Worker||Practice 1||Practice 2||Practice 3||Practice 4||TOTAL|
2.2 Data Collection
The data collection mirrored the new SOS Excel data system providing more in-depth information on existing cases than previously held, particularly in relation to outcomes. It also ensured that the data for the cases the Link Workers worked with prior to this evaluation was maximised. Three case based telephone interviews were held with each Link Worker, who was provided with a pro-forma beforehand to facilitate the data collection. Each Link Worker had the relevant files to hand to enable them to consult their notes and refresh their knowledge.
2.3 Data Analysis and Presentation
The quantitative data (Tables can be found in Appendix 1) was analysed at the level of descriptive statistics and statistical significance (Confidence Intervals). As per statistical conventions, only those results that were significant would be reported. It should be noted that where the analysis related to the full 100 cases, the proportion would be identical to the number and therefore no percentage has been given. Where the analysis has been undertaken on 95 or fewer cases, the key proportions have been given within the tables and highlighted by use of shading. In line with statistical convention, no proportions have been given for any values under ten with the exception of sub divisions where the value under ten is the main result (i.e. no values above ten). Key tables have been transformed into 'figures' to make it easier to follow the results. The qualitative data has been analysed with regard to emergent themes and each theme makes reference to specific case/s evidence.
Email: Naureen Ahmad