Deprescribing in Older People Approaching End of Life

A Randomized Controlled Trial Using STOPPFrail Criteria

Denis Curtin, MB; Emma Jennings, MB; Ruth Daunt, MB; Sara Curtin, MSc; Mary Randles, MB; Paul Gallagher, PhD; Denis O'Mahony, MD


J Am Geriatr Soc. 2020;68(4):762-769. 

In This Article


Baseline Characteristics

Between March 27, 2018, and April 3, 2019, we randomized 130 participants to receive either usual pharmaceutical care or usual pharmaceutical care supplemented by individualized STOPPFrail-guided deprescribing advice. Recruitment ended before the sample size goal of 160 was reached because of a requirement, due to resource constraints, to complete follow-up before the planned trial closure date of June 30. Ten patients died before hospital discharge, 20 patients died before follow-up at 3 months, and one patient withdrew from the trial after enrollment (Figure 1). At baseline, no significant differences were found between the intervention (n = 65) and control (n = 65) groups in terms of age, sex, or measures of cognitive, functional, and comorbidity status (Table 2). The mean plus or minus standard deviation (SD) number of regular prescribed medications at baseline was 11.5 ± 3.0 in the intervention group and 10.9 ± 3.5 in the control group (P = .28). Significantly more participants in the intervention group, relative to the control group, were prescribed analgesic medications at baseline (75% vs 49.2%; P = .03).

Figure 1.

Recruitment and participation. CFS, Clinical Frailty Scale; NOK, next of kin; SQ, surprise question.

STOPPFrail Deprescribing Recommendations

At least one deprescribing recommendation was made for 90.8% of participants in the intervention group. A mean ± SD of 2.4 ± 1.4 medications per patient was targeted for discontinuation while .75 ± .73 medications per patient were targeted for dose reduction. Overall, 87.8% of deprescribing recommendations were accepted and implemented by the attending physicians. STOPPFrail criterion A2 (ie, Stop any drug without a clear clinical indication) was the most common recommendation triggered (44.4% of all recommendations; Supplementary Table S1 lists the most common drugs deprescribed using this criterion). Lipid lowering therapies (criterion B1), neuroleptic antipsychotics (criterion D1), proton pump inhibitors (criterion E1), antiresorptive/bone anabolic drugs (criterion G2), calcium supplementation (criterion G1), and multivitamin combination supplements (criterion J1) accounted for a further 40% of the deprescribing recommendations. The frequency of the individual STOPPFrail criteria is shown in Table S2. No potential adverse effects of deprescribing were reported to the research team during the conduct of the trial.

Primary Outcome

Data from 98 randomized participants were available for analysis for the primary outcome (Figure 1). Intervention arm patients (n = 51) and control arm patients (n = 47) received a mean (SD) of 11.5 (±2.7) and 10.9 (±3.6) regular prescription medications, respectively, at baseline. The mean (SD) change in the number of regular prescriptions at 3 months was −2.61 (±2.73) in the intervention group and −.36 (±2.60) in the control group (mean difference = 2.25 ± .54; 95% confidence interval [CI] = 1.18–3.32; P < .001; Figure 2). In the final analytical sample at 3 months, three drugs that were discontinued as a result of the intervention had been restarted.

Figure 2.

Change in number of regular prescriptions from baseline to 3-month follow-up. Mean (standard deviation) change in the number of regular prescriptions (for final analytic sample, n = 99) at 3 months was −2.61 (±2.73) in the intervention arm and −.36 (±2.60) in the control arm (mean difference = 2.25 ± .54; 95% confidence interval (CI) = 1.18–3.32; P < .001). Error bars are 95% CIs.

Secondary Outcomes

No statistically significant differences were found between the intervention and control groups for patient-related outcomes such as unscheduled hospital presentations, falls, fractures, or mortality (Table 3). QoL deteriorated significantly in both the intervention and control groups from baseline to 3-month follow-up, but no statistically significant differences were found in the mean change in QUALIDEM or ICECAP-O scores between groups from baseline to follow-up (Table S3).

Antipsychotic drugs were reduced or discontinued more often in intervention patients relative to control patients, but, again, the differences did not reach statistical significance (Table S4). At baseline, there were no statistically significant differences in the extrapolated mean (SD) monthly medication costs between the intervention and control groups ($267.04 ± $117.21 and $250.56 ± $140.64, respectively; P = .53). However, at 3-month follow-up, the mean change in monthly medication cost was significantly lower in the intervention group (ie, −$74.97 ± 148.32) compared with the control group (ie, −$13.22 ± $110.40) (mean difference = $61.74 ± $26.60; 95% CI = 8.95–114.53; P = .02).