Table 3.
Psychotropic medication, β (SE), P value | Increasing number of psychotropic medications β (SE), P value | Antipsychotic medication, β (SE), P value | Antidepressant medication, β (SE), P value | Benzodiazepine medication, β (SE), P value | |
---|---|---|---|---|---|
Quality of life | |||||
DEMQOL proxy scores (n = 535) | −0.020 (0.013), 0.10 | −0.012 (0.006), 0.04 | − 0.030 (0.014), 0.03 | − 0.018 (0.011), 0.11 | − 0.008 (0.013), 0.53 |
DEMQOL scores of the residents (n = 228) | − 0.024 (0.017), 0.16 | − 0.009 (0.009), 0.32 | −0.040 (0.025), 0.11 | − 0.009 (0.017), 0.61 | −0.016 (0.018), 0.37 |
EQ-5D-5L scores (resident or proxy) (n = 531) | −0.018 (0.023), 0.44 | −0.024 (0.011), 0.03 | − 0.031 (0.026), 0.22 | 0.003 (0.021), 0.88 | − 0.059 (0.024), 0.01 |
Behaviour | |||||
NPI (n = 534) | 1.571 (0.506), 0.002 | 1.072 (0.229), < 0.001 | 2.525 (0.541), < 0.001 | 0.495 (0.463), 0.29 | 1.443 (0.514), 0.005 |
Cognitive function | |||||
PASCog (n = 537) | 0.979 (0.577), 0.09 | 0.752 (0.261), 0.004 | 2.650 (0.584), < 0.001 | −0.401 (0.499), 0.42 | −0.136 (0.560), 0.81 |
The reference for prescribed one or more psychotropic medications is being prescribed no psychotropic medications. The number of psychotropic medications was analysed as a continuous score. The models are adjusted for facility-level confounding factors: model of care, location, number of direct care hours, facility cost and size of facility and resident-level characteristics: age, sex, marital status, PAS-Cog scores (except where the outcome is PASCog scores), number of co-morbid conditions, NPI scores (except where the outcome is NPI scores), social ties and Barthel Index
Abbreviations: DEMQOL Dementia quality of life questionnaire, EQ-5D-5L EuroQol five dimensions questionnaire, NPI neuropsychiatric inventory, PASCog Psychogeriatric Assessment Scales – Cognition