Skip to main content
. 2019 Apr 8;19:83. doi: 10.1186/s12906-019-2490-z

Table 5.

Univariate and multivariate analyses of the factors influencing preference for treatment

Univariate analyses Multivariate analyses
Conditions A1 to A4 credible interval credible interval
categorical variables Comparisons M SD 2.5% 97.5% Pr(OR > 1) M SD 2.5% 97.5% Pr(OR > 1)
frequency W vs. C 0.992 0.086 0.837 1.175 0.461 0.944 0.083 0.795 1.122 0.254
(high = 1, low = 2) W vs. S 1.157 0.076 1.017 1.315 0.987 1.088 0.071 0.957 1.236 0.901
W vs. A 1.352 0.147 1.096 1.671 0.997 1.091 0.130 0.866 1.374 0.769
intensity of symptoms W vs. C 0.985 0.086 0.830 1.166 0.429 0.917 0.082 0.770 1.091 0.164
(high = 1, low = 2) W vs. S 1.263 0.084 1.108 1.437 > 0.999 1.168 0.078 1.026 1.330 0.990
W vs. A 1.368 0.149 1.108 1.691 0.998 1.121 0.135 0.890 1.416 0.832
type of illness W vs. C 0.803 0.070 0.678 0.952 0.006 0.838 0.080 0.696 1.008 0.031
(mental = 1, somatic = 2) W vs. S 1.132 0.075 0.995 1.288 0.970 1.029 0.073 0.896 1.180 0.657
W vs. A 0.647 0.071 0.523 0.801 < 0.001 0.674 0.086 0.527 0.863 0.001
linear variables Comparisons M SD 2.5% 97.5% Pr(Beta> 0) M SD 2.5% 97.5% Pr(Beta> 0)
age W vs. C −0.024 0.009 −0.041 −0.006 0.003 − 0.016 0.006 − 0.029 − 0.003 0.007
W vs. S − 0.010 0.009 −0.027 0.007 0.129 0.002 0.006 −0.009 0.013 0.654
W vs. A − 0.029 0.009 − 0.047 − 0.011 0.001 − 0.026 0.007 − 0.041 − 0.012 < 0.001
distress W vs. C −0.105 0.037 −0.177 − 0.031 0.003 −0.171 0.040 −0.250 − 0.091 < 0.001
W vs. S −0.111 0.031 −0.172 −0.050 < 0.001 −0.079 0.032 −0.143 − 0.016 0.007
W vs. A −0.571 0.044 −0.658 −0.486 < 0.001 −0.518 0.049 −0.615 − 0.422 < 0.001
treatability W vs. C 0.147 0.071 0.010 0.286 0.982 0.138 0.069 0.003 0.274 0.977
W vs. S −0.002 0.057 −0.114 0.111 0.488 −0.084 0.054 −0.191 0.022 0.060
W vs. A −1.338 0.084 −1.501 −1.177 < 0.001 −1.266 0.082 −1.427 −1.106 < 0.001
social stigma W vs. C 0.151 0.053 0.049 0.254 0.998 0.141 0.060 0.023 0.260 0.991
W vs. S −0.171 0.043 −0.255 −0.087 < 0.001 −0.137 0.047 −0.230 −0.044 0.002
W vs. A −0.268 0.064 −0.394 −0.143 < 0.001 −0.106 0.079 −0.261 0.049 0.090
belief in efficacy W vs. C −0.033 0.083 −0.197 0.131 0.347
of complementary W vs. S −0.019 0.073 −0.163 0.123 0.395
treatments W vs. A 0.104 0.094 −0.081 0.289 0.865
belief in efficacy W vs. C −0.031 0.078 −0.182 0.121 0.344
of conventional W vs. S 0.052 0.066 −0.077 0.181 0.784
treatments W vs. A −0.079 0.090 −0.254 0.098 0.189
Conditions B1 to B4 Univariate analyses Multivariate analyses
categorical variables Comparisons M SD 2.5% 97.5% Pr(OR > 1) M SD 2.5% 97.5% Pr(OR > 1)
name of illness W vs. C 1.006 0.134 0.779 1.303 0.518 1.008 0.133 0.781 1.301 0.524
(absent = 1. present = 2) W vs. S 0.702 0.068 0.582 0.847 < 0.001 0.741 0.066 0.623 0.881 < 0.001
W vs. A 0.806 0.167 0.542 1.193 0.142 0.790 0.173 0.519 1.195 0.133

Note: Comparisons: W=Weak complementary (reference category), C = strictly Conventional, S=Strong complementary, A = Alternative

Results are presented as OR or Beta with a 95% CI, with the probability of the OR being above 1, Pr(OR > 1) or the Beta being above 0, Pr(Beta> 0)

Large Pr(OR > 1) or Pr(Beta> 0) values (e.g., > 0.95, or 0.99) indicate higher values for category 2 vs category 1 (see code of categorical variables). Conversely, small Pr(OR > 1) or Pr(Beta> 0) values (e.g., < 0.05, or 0.01), reflect higher category 1 vs category 2. Moreover, the probabilities Pr(OR > 1) or Pr(Beta> 0) can be interpreted as 1 - Pr(OR > 1) or as 1 – Pr(Beta< 0), respectively. Thus, probability values near 1 and 0 both indicate meaningful effects and are indicated in bold

Conditions A1 to A1 examined the influence of illness severity (in terms of frequency of relapse and intensity of symptoms)

Conditions B1 to B4 examined the influence of adding or not the name of illness in the clinical vignettes (comparing only the vignettes of illness with the highest degree of severity). Results of the uni- and multivariate analyses are not reported here and provided similar results as in Conditions A1 to A4