Attitudes Towards Deprescribing: The Perspectives of Geriatric Patients and Nursing Home Residents

Carina Lundby MScPharm, PhD; Peter Glans MScPharm; Trine Simonsen Pharmaconomist; Jens Søndergaard MD, PhD; Jesper Ryg MD, PhD; Henrik Hein Lauridsen DC, MSc, PhD; Anton Pottegård MScPharm, PhD


J Am Geriatr Soc. 2021;69(6):1508-1518. 

In This Article


Setting and Participants

Nursing home residents were recruited from 27 long-term nursing homes in nine municipalities in the Region of Southern Denmark from November 2018 to March 2019. Geriatric in- and outpatients were recruited from the ward or outpatient clinic of the Geriatric Department at Odense University Hospital during October–November 2019. Potential participants from the nursing homes and geriatric ward were initially identified by nursing staff after which eligibility was assessed by one author (C. L., P. G., T. S.). No such initial selection was performed for the geriatric outpatients. Participants were eligible for study participation if they spoke and understood Danish, and had an Orientation-Memory-Concentration (OMC) score[14] of ≥8.


The rPATD questionnaire consists of four 5-item factors exploring patients' level of involvement in medication use, perceived burden of taking medication, belief in appropriateness of using medication, and concerns about stopping medication. It further includes two global questions exploring patients' satisfaction with medication and willingness to stop a medication if proposed by a physician. The questionnaire uses a 5-point Likert response scale (1–5 points) and is reported as a total score for each of the four factors (score range: 1–5). Higher total scores indicate more involvement, greater perceived burden, greater belief in appropriateness, and more concerns about stopping.[7]

The Danish version of the rPATD questionnaire has been validated in a cohort of nursing home residents and resulted in a model with a 4-factor structure similar to the original rPATD questionnaire.[7,13] However, to adjust the questionnaire to the Danish nursing home population and health care system, items 9 and 10 ("inconvenience of taking medication" and "medication expenses") were omitted.[13] In this study, responses for these items are only considered in supplementary analyses. Further, the Danish version was deemed suitable for use in the broader geriatric population since both populations are characterized by older people with frailty.[15,16]

Two additional previously validated instruments were used in the data collection: the Abbreviated Wake Forest Trust in Physician (Trust in Physician) Scale and the Beliefs about Medicines Questionnaire (BMQ) Specific-Concern Scale.[17,18] Both scales use a 5-point Likert response scale (1–5 points) and are reported as one total score (possible score range: 5–25). Higher scores indicate more physician trust[17] and more concerns about prescribed medication.[18]

Data Collection

The data collection has previously been described in detail.[13] In brief, one author (C. L., P. G., T.S.) went through all questions with the participants. Response options were presented on a paper in a large font. If participants did not provide an answer after having a question read three times, it was registered as missing. C. L. and T. S. recruited nursing home residents, while P. G. recruited all geriatric patients. Data was stored using REDCap.[19]


Validation of the Danish version of the rPATD questionnaire was based on a sample size of 162 nursing home residents (required to complete the exploratory factor analysis),[13] which was also considered sufficient for clinical interpretation. A similar sample size of geriatric patients was thus deemed feasible and sufficient for the aim of this study.

Participant characteristics and rPATD responses were reported using descriptive statistics. Individuals with two or more missing items within the same rPATD factor did not receive a total score. The same applied to the calculation of Trust in Physician and BMQ Specific-Concern scores. Scores were converted to a 0–100 scale using proportional recalculation.[20] rPATD factor scores were compared between participant groups (geriatric inpatients, geriatric outpatients, and nursing home residents) and across selected predefined participant characteristics (sex, age, OMC score, number of regular medications, Trust in Physician score, and BMQ Specific-Concerns score) using quantile regression for the median. A value of p < 0.05 was considered significant.


The study was registered in the Region of Southern Denmark's repository (approvals 18/46232 and 19/35570). The Regional Committees on Health Research Ethics waived registration (case numbers 20182000-129 and 20192000-129). Inclusion of participants was based on informed and written consent.