Chapter 2 50 be rated on a 3-point scale: 0, ‘no problem´ i.e. no need; 1, ‘no/moderate problembecause of continuing intervention´ i.e. met need; and 2, ´current serious problem, irrespective of any on-going intervention´ i.e. unmet need. The few cases where patients (N=21; 1.1 %) or staff (N=31; 1.6%) indicated that they did not know whether a need in a certain item existed the need was assigned ‘no need’ as no need was evident, to rate needs conservatively. 2.2.3 Statistical analysis Data were analyzed using the Statistical Package of the Social Sciences (SPSS, version 20.0; SPSS Inc., Chicago, IL). A significance level of 5% was applied. Differences between fully participating and partly participating patients were analysed with chi-square (χ²) statistics for categorical variables and a Mann-Whitney U tests for age, as age was considered not to be distributed normally. The frequency distributions of met and unmet needs, according to patient and staff, were determined. Comparisons between the total number of needs as rated by the patient and staff were performed with the Wilcoxon matched pairs signed-rank test, because the data were ordinal and skewed. To evaluate agreement on the presence of a need between patient and staff, Cohen Kappa coefficients (κ) were calculated. κ values between 0-0.20 indicate poor agreement, 0.21-0.40 fair, 0.41-0.60 moderate, 0.61-0.80 good, and 0.811.00 very good agreement. The percent agreement calculation was also documented, by dividing the number of cases in which both patients and staff agreed that there was a met or an unmet need by the total number of needs. Correlations between the patients characteristics and the number of needs was tested using the Spearman’s rank-order test. A Spearman’s rank correlation coefficient (rho) can vary between +1 and -1 (a perfect positive or negative correlation between the ranking of the two variables). There is no correlation with a coefficient of zero. 2.3 Results 2.3.1 Demographic and clinical characteristics The mean age of fully participating patients was 68.9 years (SD=7.8, range 61–98), and 51.3% were male (Table 1). Our sample had a balanced representation of bipolar I and II patients, predominantly with an onset of illness before age 50. Relevant current mood symptoms were present in 31 patients (40%), 14 with (hypo)manic symptoms (YMRS ≥7) and 21 with depressive symptoms (CES-D ≥16), including 3 patients with scores above threshold on both scales. Only a few patients had a very small network size (8%).
RkJQdWJsaXNoZXIy MTk4NDMw