Association of inappropriate prescribing with outcome variables
Table 2 reports the associations between the occurrence of at least one PIM or PPO or PIP identified at geriatric rehabilitation discharge with patient and disease characteristics, and outcomes. Patients with at least one PIM or PIP were significantly more likely to be discharged on higher number of medications, or to have greater than 10 discharge medications (hyper polypharmacy) and were mainly admitted due to cardiovascular diseases. Patients with at least 1 PIM, PPO or PIP were more common in patients with opioid, anti-inflammatory, Immunomodulatory and analgesic prescriptions, and with a greater burden of co-morbidities. Patients with at least 1 PPO were most likely admitted due to musculoskeletal disorders and were using walking aids. PIMs were less prevalent in older age, in patients with cognitive impairment and with longer hospitalisation.
Table 3 provides adjusted and crude ORs of the associations between inappropriate prescribing identified at geriatric rehabilitation discharge and outcome variables. On a univariate analysis, patients with at least 1 PIM and PIP were associated with an increased odds of hospital readmissions at longer follow up (i.e., 3 month), but not at 30 days of follow-up. A similar association was also evident for these outcomes when considering the numbers of PIMs and PIPs as a continuous variable. In addition, PIPs and PPOs as continuous variables also demonstrated a significant association with mortality data collected at 3 month and 12 months of follow-up.
After adjustment for confounding variables, exposure to at least 1 PIM, PPO and PIP did not show a significant association with post-discharge health outcomes. However, the number of discharge medications were significantly associated with hospital readmission at 30-days and 90-days follow-up. The prescription of each additional medication at discharge from geriatric rehabilitation increased the odds of hospital readmission by 1.04 (AOR 1.04; 95% CI 1.01─1.07) and 1.06 (AOR 1.06; 95% CI 1.03 ─1.09) at 30-days and 90-days post-discharge, respectively. Specifically, patients discharged with at least 1 cardiovascular medication showed a significant increased odds of hospital readmissions at 30-days (AOR 1.43; 95% CI 1.03─1.98) and 90-days of follow-up (AOR 1.45; 95% CI 1.10─1.89). Conversely, patients whose cardiovascular PPOs were omitted at discharge from geriatric rehabilitation were associated with 12-month mortality (AOR 1.34; 95% CI 1.00─1.78) although this was not evident for mortality and other outcomes collected at shorter follow-up (Table 4). PPOs related to CVD medications were mainly due to failure to prescribe or continue use of beta-blockers in patients with ischaemic heart disease, angiotensin converting enzyme inhibitors (ACEIs) in patients with systolic heart failure and/or documented coronary artery disease and statin therapy in patients with a documented history of coronary, cerebral or peripheral vascular disease with substantial life expectancy. On the other hand, PIMs related to CNS/psychotropics and fall risk increasing medications were significantly associated with the 30-day hospital readmission (AOR 1.53; 95% CI 1.09 ─ 2.15).
Linear regression analyses of the association between discharge PIMs or PPOs and the longitudinal changes of ADLs/IADs from geriatric rehabilitation discharge to 90-days post-discharge are shown in Table 5. There was no significant association between PIMs and gains in ADLs/IADLs scores but PPO use, and specifically PPOs inclusive of vaccines negatively affected ADL/IADL gains (PPO ≥ 1: β = -0.232, 95% -0.445 to -0.019; Number of PIPs (including vaccines): β = -0.107, 95% CI -0.194 to -0.021).