Perceived Confidence and Barriers of Family Physicians Toward Deprescribing Potentially Inappropriate and Preventive Medications in Older Adults: A Cross-Sectional Survey in Riyadh, Saudi Arabia

Abstract

Background

Polypharmacy is common among older adults and increases the risk of inappropriate medication use and adverse outcomes. Deprescribing is a potential strategy to optimize treatment. This study evaluated family physicians’ confidence and perceived barriers toward deprescribing in primary care in Riyadh, Saudi Arabia.

Methods

We conducted a cross-sectional survey from January 1 to December 31, 2024, among family physicians working in Riyadh primary care centers. Physicians with < 6 months of experience were excluded. Data were collected using paper-based or electronic questionnaires. Descriptive statistics summarized responses. Group comparisons used chi-square, Student’s t-test, and one-way ANOVA, as appropriate. Associations between confidence and attitude/barrier items were examined using Pearson’s correlation coefficients (r) with two-sided p-values (< 0.05 considered significant). A total of 321 physicians were included, representing 83.6% of the planned sample (321/384).

Results

Overall, 64.2% of physicians reported confidence in deprescribing, but only 28.0% reported deprescribing at least weekly. Commonly reported barriers included lack of solid evidence, time constraints, and concern about withdrawal effects. Confidence was positively associated with support for deprescribing preventive medications in patients with limited life expectancy (r = 0.259, p < 0.001), willingness to deprescribe medications initiated by another physician (r = 0.508, p < 0.001), and perceived ease of motivating patients/caregivers (r = 0.506, p < 0.001).

Conclusion

Family physicians in Riyadh reported substantial confidence in deprescribing, yet regular implementation was limited, indicating a confidence–practice gap. Addressing key barriers through practical guidance, workflow support, and targeted training may improve deprescribing uptake and medication safety for older adults.

Introduction

Older adults commonly live with multimorbidity and persistent symptoms, making long-term medication use a central component of care [1, 2]. Medication burden rises with age; by 70 years, nearly three-quarters of individuals use five or more medications daily [3]. At the same time, population aging is accelerating in Saudi Arabia, with adults aged ≥ 60 years projected to increase substantially in the coming decades [4]. This demographic shift is expected to increase the number of older adults exposed to polypharmacy in routine primary care.

Although pharmacotherapy can improve disease control, function, and quality of life [5, 6], medication use in older adults is also associated with adverse drug reactions, drug–drug interactions, medication-related hospitalizations, falls, cognitive decline, and higher healthcare utilization [7–10]. Polypharmacy (often operationalized as the concurrent use of ≥ 5 medications) affects approximately 30–60% of adults aged ≥ 65 years in high-income settings [11–13]. Clinically, this burden often includes cardiometabolic preventive therapies (e.g., antihypertensives, lipid-lowering agents, antiplatelets), glucose-lowering medications, psychotropics, sedative-hypnotics, analgesics, and gastrointestinal drugs, where cumulative risk may exceed benefit in frail or complex older patients [9, 19]. Polypharmacy is driven by multimorbidity, guideline-driven disease-specific prescribing, fragmented care across providers, prescribing cascades, and limited routine medication review [9, 14–16].

Deprescribing-the planned dose reduction or discontinuation of medications when potential harm outweighs expected benefit-is a key strategy to reduce inappropriate polypharmacy [17, 19]. It is particularly relevant for high-risk, low-value, or preference-incongruent medications and should align with patient goals, functional status, life expectancy, and values [18, 19]. Evidence suggests deprescribing can reduce medication-related harm and improve safety outcomes [19, 22].

Despite these benefits, implementation remains inconsistent. Physicians report barriers including limited consultation time, uncertainty about risk–benefit balance, communication challenges with patients/caregivers, and lack of practical discontinuation guidance [18, 20, 21, 23]. In this context, physicians’ confidence and attitudes are not peripheral—they are central determinants of whether deprescribing is initiated, discussed, and sustained in practice [18, 20, 21].

Family physicians are a priority group for study because they provide longitudinal, first-contact, and comprehensive care for older adults, reconcile prescriptions from multiple specialties, and are often best positioned to lead shared deprescribing decisions in primary care [18, 20, 21]. However, evidence from Saudi Arabia—especially on family physicians’ confidence and perceived barriers in everyday primary-care workflows—remains limited. To our knowledge, no published Saudi study has comprehensively examined this issue in family physicians in Riyadh.

Accordingly, this study aims to assess perceived confidence and barriers toward deprescribing among family physicians working in primary care centers in Riyadh, Saudi Arabia. We hypothesized that family physicians would report moderate confidence in deprescribing, but that substantial system-, clinician-, and patient-level barriers would remain common and would influence deprescribing practice.

Methods

Study Design and Setting

This was a descriptive, questionnaire-based, cross-sectional study conducted in Riyadh, Saudi Arabia. Eligible participants were family physicians working in primary care centers with at least 6 months of continuous practice in their current primary-care setting. We used this threshold as an operational criterion to ensure participants had sufficient exposure to routine prescribing workflows, follow-up patterns, and deprescribing decisions in daily practice.

A multistage sampling approach was used. Riyadh was divided into five main regions (North, South, East, West, and Central) according to the Ministry of Municipal and Rural Affairs’ map. These regions were further divided into 15 administrative districts containing multiple primary care centers. Riyadh includes approximately 169 neighborhoods, some sparsely populated. A list of primary care centers across neighborhoods was generated using Google^® Maps, and stratified random sampling (Microsoft Excel^®) was used to select centers.

The sample size was calculated using the single population proportion formula for cross-sectional studies [1], assuming a 50% expected prevalence (maximizing required sample), 95% confidence level, and 5% margin of error, yielding a minimum target of 384 physicians. Due to time and logistical constraints, 321 physicians were recruited. With n = 321 and p = 0.50, the achieved precision is approximately ± 5.5% at the 95% confidence level, indicating slightly reduced precision compared with the planned ± 5.0%, but still acceptable for descriptive and exploratory analyses.

Questionnaire Development

All nine Likert-scale items (1–7) shown in Table 2 were adapted, with permission, from a similar published study, “How Confident are Physicians in Deprescribing for the Elderly and What Barriers Prevent Deprescribing” [2]. Additional attitude/belief items presented in Figs. 1, 2 and 3 were adapted, with permission, from “Doctors’ Attitudes and Beliefs Towards Deprescribing Medications” [3].

Table 2.

Attitudes and Barriers to deprescribing as reported by the enrolled physicians (n = 321)

	Strongly disagree					Strongly agree
	1	2	3	4	5	6	7
Attitudes
Q1) From a clinical standpoint, I feel confident with deprescribing in my elderly patients.	3.4%	5.0%	8.7%	18.7%	25.9%	22.7%	15.6%
Q2) When the life expectancy of my elderly patients no longer justifies potential benefits, I am in favor of deprescribing preventive medications.	7.8%	6.5%	8.7%	20.9%	20.6%	16.2%	19.3%
Q3) In elderly patients with poor life expectancy, it would be appropriate to consider deprescribing therapeutic medications even though they are recommended by guidelines.	16.8%	14.0%	13.4%	22.1%	16.2%	10.3%	7.2%
Q4) In my elderly patients, I have no hesitation in deprescribing medications initially prescribed by another physician.	6.2%	11.8%	16.8%	19.3%	13.7%	21.2%	10.9%
Q5) I have no problem in deprescribing medications even if my elderly patients and/or caregivers believe continuation is needed.	6.2%	11.5%	14.6%	22.4%	16.8%	15.9%	12.5%
Q6) I have no difficulty motivating my elderly patients and/or caregivers to engage them in the process of deprescribing medications.	2.8%	4.7%	6.2%	18.7%	17.4%	24.3%	25.9%
Barriers
Q7) In elderly patients, a lack of robust evidence in favor of continuation or cessation of preventive medications prevents me from deprescribing.	2.5%	3.4%	10.0%	34.0%	21.8%	18.4%	10.0%
Q8) I do not have the necessary time to spend with my elderly patients and/or caregivers to effectively undertake the process of deprescribing medications even though I consider it important.	13.4%	9.7%	13.1%	17.8%	17.1%	13.4%	15.6%
Q9) Although in certain situations I may consider appropriate deprescribing medications in my elderly patients, I do not consider it for fear of adverse drug withdrawal effects.	8.4%	12.5%	12.8%	29.0%	19.6%	12.1%	5.6%

Fig. 1.

How often do you de-prescribe medication?

Fig. 2.

Medications commonly deprescribed by family physicians in Riyadh, Saudi Arabia

Fig. 3.

Enablers for deprescribing medications by family physicians in Riyadh, Saudi Arabia

To support content validity, the questionnaire was reviewed by a Geriatric/Family Medicine consultant and a Family Medicine consultant. Face validity and clarity were pilot-tested in 20 family physicians before formal data collection.

For interpretability and comparability with the source approach [2], Likert responses were dichotomized for selected analyses: 5–7 as “agree” and 1–4 as “disagree/neutral.” This approach improves interpretive clarity for practice-focused reporting but may reduce response granularity. Therefore, findings should be interpreted as directional rather than fully ordinal effects.

Although the adapted tools were previously used [2, 3], formal internal consistency testing (e.g., Cronbach’s alpha) was not calculated in the current dataset. This is acknowledged as a methodological limitation and should be addressed in future psychometric validation work.

Data Collection

Participants provided written informed consent after receiving an explanation of study objectives and their right to withdraw at any time. Two trained medical interns conducted data collection at selected centers. Training included role-play and a standardized protocol for participant approach, consent procedures, and interview administration. When multiple eligible physicians were available in one center, interviews were conducted individually after separate written consent.

Statistical Analysis

Data were analyzed using SPSS. Descriptive statistics (means, standard deviations, frequencies, percentages) summarized participant characteristics and survey responses. Bivariate analyses used chi-square tests, Student’s t-tests, one-way ANOVA, and Pearson’s correlation, as appropriate to variable type and distribution. Statistical significance was set at p < 0.05 with 95% confidence intervals.

In Fig. 3, consultants/registrars were grouped as “Seniors,” and family medicine residents as “Juniors.” For barrier and attitude items, seven-point Likert responses were dichotomized (Table 2): 5–7 as “agree,” and 1–4 as “disagree/neutral.”

Given the exploratory cross-sectional design, analyses focused primarily on bivariate associations. We recognize that residual confounding (e.g., age, years of experience, professional role) may influence observed relationships; therefore, results should be interpreted cautiously, and future studies are recommended to apply multivariable regression to identify independent predictors.

Results

A total of 321 family physicians participated in the study. As shown in Table 1, participants were predominantly younger physicians (most aged ≤ 35 years), with a near-equal gender distribution (49.2% male, 50.8% female). As shown in Table 1, registrars constituted the largest professional group, followed by residents and consultants. As shown in Table 1, nearly half of participants had < 5 years of experience, while about one-quarter had 5–10 years of experience.

Table 1.

Baseline sociodemographic characteristics of the enrolled family physicians

Characteristic
Age
1. < 30	122 (38%)
2. 31–35	93 (29%)
3. 36–40	28 (8.7%)
4. 41–45	29 (9%)
5. 46–50	18 (5.6%)
6. > 50	31 (9.7%)
Gender
1. Male	158 (49.2%)
2. Female	163 (50.8%)
Current Position
1. Consultant	88 (27.4%)
2. Registrar	123 (38.3%)
3. Resident	110 (34.3%)
Years of Experience
1. < 5	149 (46.4%)
2. 5 to 10	84 (26.2%)
3. > 10 to 15	29 (9%)
4. > 15 to 20	18 (5.6%)
5. > 20	41 (12.8%)

Figure 1 depicts the frequency of medication deprescription among family physicians. The results show that the largest proportion of physicians deprescribe medication at least once a week (28%). This is followed by those who deprescribe at least once a month (19.3%). Approximately 27.1% deprescribe at least once every six months, and 21.5% deprescribe at least once every three months. A small minority, 4%, reported that they never deprescribe medications.

As shown in Fig. 2, proton pump inhibitors were the most commonly deprescribed medications (58.6%), followed by vitamins/supplements (54.8%), antihistamines (48.3%), and statins (45.5%). Mid-range frequencies were observed for antihypertensives (37.1%) and antiplatelets (34.9%), whereas psychotropics and opioids were less commonly deprescribed. As shown in Fig. 2, the least frequently deprescribed classes were antianginals (7.2%) and anticonvulsants (6.9%).

As shown in Fig. 3, the most commonly reported facilitator in both groups was the availability of medication-specific deprescribing guidelines (85.3% of seniors; 72.7% of juniors). Team-based support was also frequently endorsed, particularly pharmacist involvement (55.0% of seniors; 50.0% of juniors), followed by educational lectures (51.2% of seniors; 52.7% of juniors). Pharmacist-led alerts were less frequently selected, and a strong departmental focus on deprescribing was reported at similar levels across groups. Only a small minority in either group reported that none of the listed factors were facilitators.

As shown in Tables 2, 64.2% of physicians agreed/strongly agreed that they were confident in deprescribing for older patients (Q1), and 56.1% agreed/strongly agreed with deprescribing preventive medications when life expectancy is limited (Q2). In contrast, views were more divided for deprescribing guideline-recommended therapies (Q3), with only 33.7% agreeing versus 47.4% disagreeing/neutral.

As shown in Table 2, the most prominent barriers were lack of solid evidence (Q4: 65.8% agree/strongly agree), time constraints (Q6: 53.0%), and fear of adverse withdrawal effects (Q8: 49.6%). Willingness to deprescribe medications initiated by another physician (Q5) was moderate (46.7% agree/strongly agree). For patient/caregiver-related items, 53.0% agreed that deprescribing remained acceptable even when continuation was perceived as necessary (Q7), and 67.0% reported little difficulty motivating patients/caregivers to participate in deprescribing discussions (Q9).

As shown in Table 3, confidence in deprescribing (Q1) was positively correlated with favorable deprescribing attitudes, including deprescribing preventive medications in limited life expectancy (Q2: r = 0.259, p < 0.001), willingness to deprescribe medications initiated by another physician (Q5: r = 0.508, p < 0.001), and perceived ease of motivating patients/caregivers (Q9: r = 0.506, p < 0.001). Confidence was also positively associated with deprescribing despite perceived continuation need by patients/caregivers (Q7: r = 0.342, p < 0.001).

Table 3.

The Association between the question regarding physicians’ confidence in deprescribing (Q1) and the questions concerning physicians’ attitudes and barriers (Q2–Q9)

Item	Correlation coefficient	P-value
Q2) When the life expectancy of my elderly patients no longer justifies potential benefits, I am in favor of deprescribing preventive medications.	0.259**	0.000
Q3) In elderly patients with poor life expectancy, it would be appropriate to consider deprescribing therapeutic medications even though they are recommended by guidelines.	0.047	0.402
Q4) In elderly patients, a lack of robust evidence in favor of continuation or cessation of preventive medications prevents me from deprescribing.	0.147**	0.008
Q5) In my elderly patients, I have no hesitation in deprescribing medications initially prescribed by another physician.	0.508**	0.000
Q6) I do not have the necessary time to spend with my elderly patients and/or caregivers to effectively undertake the process of deprescribing medications even though I consider it important.	−0.109	0.052
Q7) I have no problem in deprescribing medications even if my elderly patients and/or caregivers believe continuation is needed.	0.342**	0.000
Q8) Although in certain situations I may consider appropriate deprescribing medications in my elderly patients, I do not consider it for fear of adverse drug withdrawal effects.	−0.009	0.879
Q9) I have no difficulty motivating my elderly patients and/or caregivers to engage them in the process of deprescribing medications.	0.506**	0.000

**Significant at the significance level (α ≤ 0.05)

As shown in Table 3, weaker associations were observed with perceived lack of robust evidence (Q4: r = 0.147, p = 0.008), while no significant correlations were found with guideline-related therapeutic deprescribing (Q3: r = 0.047, p = 0.402), fear of withdrawal effects (Q8: r = − 0.009, p = 0.879), or time constraints (Q6: r = − 0.109, p = 0.052).

Discussion

This cross-sectional study assessed family physicians’ confidence, attitudes, and perceived barriers toward deprescribing for older adults in Riyadh primary care settings. Overall, physicians reported moderate-to-high confidence; however, routine implementation was less frequent, indicating a clear confidence–practice gap. This pattern is clinically important because confidence may reflect perceived readiness, but it does not necessarily translate into real-world implementation when organizational and contextual barriers persist [7, 12, 13, 23, 24].

A key finding was that many physicians reported confidence in deprescribing, yet a substantially smaller proportion reported regular deprescribing practice. This discrepancy should be interpreted cautiously because practice frequency was measured using a single self-reported item, which may not fully capture behavior across clinical settings. Self-reported outcomes are vulnerable to recall and social desirability bias, and perceived confidence is not equivalent to clinical competence or sustained implementation [23–26]. Future studies should complement survey data with objective measures, including electronic medical record (EMR) audits, medication discontinuation indicators, and longitudinal prescribing trajectories [25–27].

Our findings align with international evidence showing that physicians may conceptually support deprescribing while still encountering operational barriers in day-to-day care [7, 12, 13, 28, 29]. Similar studies from Italy reported higher proportions of physicians confident in deprescribing and greater comfort with deprescribing preventive medications [7]. In contrast, our estimates were comparatively lower in these domains, particularly for deprescribing against guideline-driven continuation. Differences across settings may be explained by variation in workload intensity, clinical governance, prescribing culture, training exposure, and availability of practical deprescribing tools [28–31].

The epidemiological relevance of this issue remains substantial. Polypharmacy in older adults is associated with adverse drug events, hospitalization, functional decline, and increased mortality risk [4, 11, 32, 33]. At the same time, evidence suggests that carefully selected deprescribing—especially when medication benefit becomes limited relative to life expectancy or frailty burden—can be feasible and safe in appropriate patients [9, 10, 34]. However, evidence remains mixed for some hard endpoints (e.g., mortality), partly due to heterogeneity in study designs and intervention intensity [5, 34, 35]. These findings support individualized, shared-decision deprescribing rather than a uniform de-intensification approach [18, 19, 35].

The barrier profile observed in our study is consistent with prior work: limited consultation time, concern about withdrawal-related adverse effects, uncertainty in risk–benefit balance, communication challenges, and limited practical guidance [12, 13, 20, 21, 30, 31]. Importantly, several barriers are structural rather than purely clinician-level. Therefore, increasing confidence alone may be insufficient to change practice. Effective implementation likely requires protected medication-review time, decision-support tools, interdisciplinary collaboration (including pharmacists), and locally applicable deprescribing pathways [6, 29–31, 36].

Because family physicians provide first-contact, comprehensive, and longitudinal care for older adults, these findings are highly relevant to Saudi primary care. Family physicians commonly reconcile prescriptions initiated across specialties and are often central to aligning treatment intensity with patient goals and functional status [18, 20, 21, 24, 36]. Strengthening deprescribing capacity in this group may therefore improve medication safety, continuity, and appropriateness at the system level.

From a policy perspective, our findings support development of nationally adapted deprescribing guidance and implementation infrastructure in primary care. While chronic disease frameworks are generally strong in treatment initiation and escalation, they are less explicit regarding de-intensification in older adults with multimorbidity [14–16, 23, 24, 36]. Embedding deprescribing competencies in CME, residency curricula, and quality indicators could support Saudi healthcare transformation and Vision 2030 priorities related to quality, patient safety, and value-based care [23, 24, 36].

External validity should be considered carefully. The study was limited to Riyadh, where service organization, staffing models, patient flow, and workload pressures may differ from other Saudi regions. In addition, the relatively young physician cohort may influence confidence reporting and practice patterns compared with more experienced national samples [28, 30, 31, 36]. Accordingly, broader national extrapolation should be cautious, and multi-region studies are warranted.

Overall, this work should be viewed as exploratory and hypothesis-generating. It identifies an implementation gap that appears modifiable across clinician-, patient-, and system-level domains. Future research should use multicenter designs, objective prescribing outcomes, and multivariable longitudinal modeling to identify independent determinants of deprescribing behavior and to evaluate implementation strategies in Saudi primary care [25–27, 34–36].

This study has several limitations. First, the planned sample size of 384 physicians was not fully achieved; the final sample included 321 physicians. This shortfall modestly reduced estimate precision (approximately ± 5.5% vs. ± 5.0% at 95% confidence) and may have limited statistical power to detect smaller associations. Second, the study was conducted only in Riyadh and included a relatively young physician cohort, which may limit generalizability to other Saudi regions and international contexts. Third, outcomes were self-reported, and deprescribing practice was captured using a single self-reported frequency item; therefore, recall and social desirability biases cannot be excluded [23–26]. Fourth, formal internal consistency testing (e.g., Cronbach’s alpha) was not calculated for the adapted questionnaire in this population. Fifth, dichotomization of 7-point Likert responses (including grouping neutral responses with disagreement) may have reduced granularity and influenced interpretation [2, 26]. Finally, because analyses were primarily bivariate within a cross-sectional design, residual confounding cannot be excluded and causal inference is not possible. Future multicenter longitudinal studies using objective prescribing indicators (e.g., EMR audits) and multivariable modeling are needed [25–27, 34–36].

Conclusion

In conclusion, this study addressed its primary aim by characterizing family physicians’ confidence and perceived barriers toward deprescribing older adults in Riyadh primary care. Although many physicians reported confidence, regular deprescribing practice was less frequent, indicating a confidence–implementation gap. The main barriers identified were limited consultation time, fear of adverse withdrawal effects, uncertainty when evidence is limited, communication challenges with patients/caregivers, and lack of clear deprescribing guidance. To translate confidence into consistent practice, Saudi primary care would benefit from structured deprescribing guidance, workflow-supported medication review, interdisciplinary collaboration (especially pharmacist-supported review), and targeted training within CME and residency programs. Given the study’s geographic scope and reduced sample size, future nationwide multicenter studies using objective prescribing data are needed to confirm these findings and guide scalable policy implementation.

Author contributions

All authors contributed significantly to the conception, design, and execution of the study. Abdullah A. Alrasheed was the principal investigator, overseeing all aspects of the study, including conceptualization, methodology, and manuscript preparation. Majed Ali Almasaoud and Abdulelah Ibrahim Alrasheed contributed to data collection and analysis. Abdullah Khaled Alqahtani and Khalid F. Alsadhan provided critical insights into the study design and supervised the data analysis process. Haytham I. AlSaif and Saad M. Alsaad contributed to drafting and revising the manuscript. Norah A. AlShehri was involved in the interpretation of findings and final approval of the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

This research was funded by the Vice Deanship of Scientific Research Chairs at King Saud University, Saudi Arabia, through the Prince Faisal Bin Bandar Chair for Geriatric Research, which was established in collaboration with the Kebar Association, Riyadh, Saudi Arabia.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethical Approval and Consent to Participate

This study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki. Ethical approval was obtained from the Institutional Review Board (IRB) of King Saud University, College of Medicine (IRB Project No. E-24-8562). All participants were fully informed about the purpose, procedures, and potential risks of the study, and written informed consent was obtained prior to their participation. The study adhered to all relevant national and international ethical guidelines.

Consent for Publication

Not applicable. This study does not involve identifiable individual data from participants.

Competing Interests

The authors declare no competing interests.