The analysis involved data collected from 3,044 patients (1,966 patients in the high deprivation group, and 1,078 patients in the low deprivation group). The mean patient age was 43.4 years and 46.5 years, in the high and low deprivation groups respectively. Female patients made up 61% of the high deprivation group and 65% of the low deprivation group. Full details of the two groups in terms of patient characteristics (see table ) and consultation quality markers (including enablement scores) have been published previously [16
]. The overall mean enablement score was 3.0 (SD 3.36, n = 2471) with a median of 2.0. Mean PEI scores were similar in high and low deprivation areas; high deprivation 3.1, (SD 3.44, n = 1531), low deprivation group 3.0, (SD 3.21, n = 940). The median for both areas was 2.0. The distribution of PEI scores in high and low deprivation settings are shown in Figure . As can be seen, the scores were not normally distributed.
Characteristics of Patients living in high and low deprivation areas
Distribution of PEI scores in high and low deprivation groups. X-axis: Enablement Score- PEI. Y-axis: Percentage.
In view of this non-linear distribution of PEI scores, these scores was analysed as a binary outcome (PEI score below 3.0 or 3.0 and above) thus overcoming assumptions of linearity in the analysis. We carried out binary logistic regression with PEI score as the dependant variable and a score of less than 3 as the reference category. Univariate analysis showed 7 factors to be significantly associated with PEI scores; (Table ).
Factors associated with patient enablement: univariate analysis n = 3044
However, after multilevel multivariate analysis, only 4 factors remained independently associated with PEI score; (see Table )
Factors associated with patient enablement: multi-variate analysis n = 3044
To determine whether these independent predictors of enablement operated in patients from both high and low deprivation areas, interaction effects of deprivation group on each variable were examined. The effect of the 4 independent predictors of enablement and the influence of deprivation are detailed below.
Reason for consulting
Patients who were consulting about a new problem had higher PEI scores (i.e., a higher percentage scoring 3 and above) than those consulting for a long-standing problem (or both new and long-standing). There was no significant interaction effect with deprivation (results not shown).
Patients who had significant psychological distress (GHQ caseness) had lower PEI scores (i.e., a higher percentage scoring 3 and above) than those who did not have caseness. However, there was a highly significant interaction effect with deprivation, which indicated that the effect of psychological distress on PEI score was only apparent in the high deprivation group. Please see Figure which shows the interaction between GHQ caseness and deprivation status.
Effect of emotional distress on patient enablement in high and low deprivation groups. X-axis: Emotional distress. Y-axis: 95% Confidence interval for mean PEI-enablement score.
Number of long-term conditions/general health
Patients who had 3 or more long-term conditions had lower PEI scores (i.e., a lower percentage scoring 3 and above) than those with less than 3 conditions. There was no significant interaction effect with deprivation (results not shown).
When the number of long-term conditions (multimorbidity) variable in the multi-variate model was replaced by self-reported general health over the last 12 months, a similar finding was observed - patients with worse general health had lower PEI scores. Again, there was no significant interaction effect with deprivation (results not shown).
The number of long-term conditions and general health were not included in the model together because of a significant association between these two variables. The association between number of long-term conditions and general health was highly significant (Pearson's r = 0.423, p < 0.0001).
Patients' perceptions of GP empathy (CARE score)
Patients who perceived their GP to be empathic (above average CARE measure score in the current study) had higher PEI scores (i.e., a higher percentage scoring 3 and above) than those who scored their GP as having average or below average empathy. There was no significant interaction effect with deprivation (results not shown).
Further examination of the relationship between GP empathy and patient enablement showed that full enablement (maximal PEI score of 12; n = 82) never occurred when GP empathy was low (lower quartile of CARE Measure scores), whereas 63.4% (n = 52) of full enablement consultations occurred when GP empathy was high (upper quartile of CARE Measure scores) and 36.6% (n = 30) when GP empathy was average (second and third quartiles of CARE measure scores). The findings were similar in both high and low deprivation areas (results not shown). Thus, although high empathy was not always associated with high enablement (presumably due to the other factors that negatively influence enablement such as emotional distress, multimorbidity, etc), low empathy was always associated with low enablement, suggesting that GP empathy is a basic pre-requisite for patient enablement.