Preferences for befriending schemes: a survey of patients with severe mental illness

BACKGROUND: Befriending has become a widely used method for tackling social isolation in individuals with severe mental illness (SMI), and evidence exists to support its effectiveness. However, patient preferences for befriending remain unclear. We aimed to determine whether patients with SMI want a volunteer befriender and, if so, the volunteer characteristics and character of the relationship they would prefer. METHODS: A survey of outpatients was conducted across London-based community mental health teams, for individuals diagnosed with affective or psychotic disorders. Questions consisted of measures of demographic characteristics, befriending preferences and social context, including measures of time spent in activities, number of social contacts, loneliness and subjective quality of life (SQOL). Binary logistic regressions were used to investigate potential predictors of willingness to participate in befriending. RESULTS: The sample comprised of 201 participants with a mean age of 43 years. The majority (58%) of the sample indicated willingness to participate in befriending. In univariable analyses this was associated with less time spent in activities in the previous week, higher level of loneliness and lower SQOL. When all three variables were tested as predictors in a multivariable analysis, only lower SQOL remained significantly associated with willingness to take part in befriending. Relative to other options presented, large proportions of participants indicated preference for weekly (44%), 1-hour (39%) meetings with a befriender, with no limits on the relationship duration (53%). Otherwise, patient preferences exhibited great variability in relation to other characteristics of befriending schemes. CONCLUSIONS: A substantial number of patients with SMI appear willing to take part in a befriending scheme. Patients with lower SQOL are more likely to accept befriending, so that befriending schemes may be a realistic option to help patients with particularly low SQOL. The large variability in preferences for different types of befriending suggests that there is no one-size-fits-all formula and that schemes may have to be flexible and accommodate different individual preferences.