Identification of Drug-Related Problems and Assessment of Health-Related Quality of Life in Elderly Multimorbid Inpatients with Stroke

Background

A growing global focus on population aging is evident. Projections show that from 2020 to 2030, the global population aged 60 and above will grow substantially from 1 billion to 1.4 billion, while its share of the world’s population is expected to double from 12% in 2015 to 22% by 2050.¹ The aging trend confronting China is particularly pronounced, as China is home to the largest elderly population globally. By the end of 2024, Chinese elderly population aged 60 years and older had reached 310 million, accounting for 22.0% of the total population.² It is expected that by 2030, the number of elderly people requiring care will increase by 14.02 million compared with 2020.³ Multimorbidity is commonly defined as the simultaneous occurrence of two or more chronic diseases in an individual.⁴ The incidence of multimorbidity among the elderly aged 60 years and older has reached 46.2%, constituting the largest proportion of all age groups.⁵ The gravity of the aging problem, compounded by the intricate and precarious state of the elderly patients, frequently marked by multiple chronic conditions, prolonged medication regimens, and heightened mental health concerns, along with alterations in pharmacokinetics and pharmacodynamics in this demographic,⁶ is poised to escalate the prevalence of drug-related problems (DRPs).

The Global Burden of Disease (GBD) study indicated that ischemic heart disease and stroke continue to be the predominant causes of mortality worldwide; however, advancements have been made in decreasing their age-standardized mortality rates on a global scale.⁷ Based on disability adjusted life years (DALY) counts, the top three non-communicable diseases were ischaemic heart disease, stroke and diabetes in 2023.⁸ In light of the above burden of disease data, the China Stroke Prevention Project Committee (CSPPC) published a plan to establish stroke centers as early as 2016.⁹ By 2025, the number of established Comprehensive Stroke Center (CSC) had reached 2,346. As one of CSCs, it is imperative to identify the population characteristics of elderly inpatients admitted to the hospital, the prevalence of medication issues, and their profound effect on clinical outcomes. This identification process is crucial for the optimization of management for this patient group.

DRPs are defined as any event or circumstance involving drug therapy that actually or potentially interferes with desired health outcomes. This is anticipated to adversely impact patient prognoses, diminish quality of life, and exacerbate the financial and economic burden on healthcare systems.^10,11 The survey revealed a marked increase in healthcare expenditures related to DRPs in hospitalized patients, from 1.6 billion dollars in 2000 to 4.0 billion dollars in 2008.¹² Numbers of characteristics exhibited by stroke center inpatients have been demonstrated to be associated with the occurrence of DRPs. It has been established that approximately 94.2% of patients with acute ischemic stroke have multiple multimorbidity, with majority of acute ischemic strokes exhibiting one to three multimorbidity.¹³ Elderly inpatients in CSC who have been on multiple medications for a protracted period due to multimorbidity and other factors are more likely to encounter medication-related issues. It is imperative to identify and intervene in this category, as this can reduce problems such as increased healthcare costs due to DRPs.

As members of the therapeutic team, clinical pharmacists have been demonstrated to exert a positive influence on patient outcomes, medication safety and medication adherence.^14,15 A series of studies have demonstrated that clinical pharmacists can identify and address some of the DRPs through interventions in different patient groups, thereby reducing the incidence of DRPs in patients and improving their prognosis. This has been observed in patients receiving antithrombotic therapy,¹⁶ patients with chronic renal disease,¹⁷ patients with oncology,¹⁸ and patients undergoing surgical procedures.¹⁹

There is a lack of specific studies assessing and evaluating the DRPs and therapeutic efficiency by pharmaceutical service among elderly multimorbid inpatients with stroke in CSC. This study aims to explore the prevalence and factors associated with DRPs as well as their potential impact on health-related quality of life (HRQoL) among elderly multimorbid inpatients with stroke. The findings of this study corroborate the efficacy of earlier identification and management of DRPs by clinical pharmacists in this high-risk hospitalized population.

Methods

Study Design and Participants

This retrospective observational study at Zhejiang Provincial People’s Hospital examined CSC inpatients from January 2022 to December 2023 (excluding patients admitted during the peak COVID-19 outbreak period from January to March 2023 to minimize confounding effects). Elderly (aged 60 years and older) multimorbid (defined as having two or more chronic health problems requiring long-term medical care) inpatients with stroke included those hospitalized among CSC with multimorbidity for over 24 hours. All procedures followed the Helsinki Declaration, and written informed consent for publication was obtained from patients for the case descriptions in Table S1.

Data Collection

Clinical pharmacists gathered demographic, clinical, and medication data from hospital information systems. Stroke-associated clinical scales, including the National Institutes of Health Stroke Scale (NIHSS), the modified Rankin Scale (mRS), and Activities of Daily Living (ADL), were routinely assessed by the medical team upon admission and recorded in the Electronic Medical Records (EMR). In contrast, the assessment of DRPs and HRQoL (using the EQ-5D-5L instrument) was not part of the routine EMR documentation. These were specifically evaluated by two trained clinical pharmacists for the purpose of this study. The clinical pharmacists reviewed and assessed electronic medical prescriptions and clinical notes using clinical guidelines, drug instructions, and the patient’s health status, considering factors like drug choice, indications, contraindications, dosage, interactions, effectiveness, and side effects to identify DRPs. Similarly, the HRQoL status was determined based on comprehensive admission records and patient interviews documented during the hospitalization.

DRPs Identification by Medical Review

DRPs were classified using the Pharmaceutical Care Network Europe version 9.1 (PCNE-DRP V9.1) system,²⁰ focusing on problems and causes. Severity was rated using the National Coordinating Committee for Medication Error Reporting and Prevention (NCC-MERP) classification,²¹ which has nine grades (A-I) grouped into five categories: (1) circumstances or events that could cause error (potential errors, category A); (2) medication errors (MEs) occurred without posing harm to patients (categories B and C); (3) MEs caused potential harm to patients (category D); (4) MEs caused harm to patients (of increasing severity, categories E, F, G, and H); (5) MEs resulted in a patient’s death (category I). Two clinical pharmacists independently assessed and graded the DRPs, resolving any differences through discussion. The drugs implicated in DRPs are standardized and categorized according to the Anatomical Therapeutic Chemical classification system code (ATC code).²²

HRQoL Assessment by Clinical Pharmacists

HRQoL of elderly multimorbid inpatients with stroke in CSC was applied in the assessment of quality of life using the European Quality of Life Five Dimension Five Level Scale Questionnaire (EQ-5D-5L) by clinical pharmacists for this study. This assessment captured the patients’ health status based on their initial records at hospital admission. Utility values were calculated based on the validated China EQ-5D-5L value set.²³

Stroke-Associated Clinical Scales Assessment by Medical Team

The professional neurologist assessed the NIHSS score and mRS score of each patient. ADL score of elderly multimorbid inpatients with stroke was assessed using the Barthel Index (BI). The ADL scale consists of ten items to determine the ADL ability of elderly multimorbid inpatients with stroke. These three scales (NIHSS, mRS, and ADL) were evaluated during initial hospital admission records.

Statistical Analyses

All data collected in this study were analyzed using SPSS version 26. Categorical variables were presented as frequencies and percentages, while continuous variables were shown as mean ± standard deviation (s) if normally distributed or as median and quartiles (quartile 1 (Q1), quartile 3 (Q3)) if not. Categorical variables were compared using χ² tests, and continuous variables with Student’s t-test for normal distributions or Mann–Whitney U-tests otherwise. Binomial logistic regression analyzed factors affecting DRPs and the relationship between dependent and independent variables. The Spearman correlation analysis was used for the analysis of correlation between the DRPs and HRQoL or disease related rating scale with p < 0.05 indicating statistical significance. Gephi 0.9.2 was used to visualize and analyze the multimorbidity network in elderly inpatients.

Results

Pharmaceutical Service at the CSC

A total of 1,049 elderly multimorbid inpatients with stroke who were receiving pharmaceutical services were initially recruited into the study. A total of 457 patients were excluded during the assessment period due to having fewer than two chronic diseases and being under 60 years of age. The data was collected and analyzed from 592 patients (Figure 1). The medication review, identification of DRPs which were present during the hospital stay and HRQoL assessment were conducted by a clinical pharmacist, while stroke-associated clinical scales were assessed by the medical team.

Figure 1 Flowchart of the implementation of pharmaceutical service in CSC. Abbreviations: CSC, comprehensive stroke center; DRPs, drug-related problems; PCNE, Pharmaceutical Care Network Europe; HRQoL, health-related quality of life; NIHSS, National Institute of Health stroke scale; mRS, modified Rankin scale; ADL, activities of daily living.

Multimorbidity Patterns in Elderly Inpatients in CSC

Among the patients enrolled from the CSC in this study, the majority had a primary diagnosis of stroke. The multimorbidity patterns network among the elderly inpatients (Figure 2) consists of 14 nodes, with a total of 76 edges connecting any two nodes. The top three chronic diseases in terms of node size are hypertension, stroke and diabetes. The weighted degrees of the nodes range from 8 to 833, with hypertension (833), stroke (578), and diabetes (525) having the top three weighted degrees, thus establishing these as the three most influential nodes in the multimorbidity patterns network among the elderly inpatients. The edge weights within the network ranged from 1 to 256. The top three multimorbidity patterns were as follows: co-existing hypertension and stroke (256), co-existing hypertension and diabetes (223), and co-existing diabetes and stroke (131).

Figure 2 Multimorbidity network for elderly inpatients at the CSC (n=592). Abbreviation: CSC, comprehensive stroke center. Notes: Nodes in the network represent types of chronic diseases, with node size proportional to the prevalence of the disease. Larger nodes indicate higher prevalence. Edges represent elderly inpatients with two chronic diseases, with edge thickness reflecting the frequency of co-occurrence of the two diseases. Thicker edges indicate higher co-occurrence frequency.

Demographic Characteristics and Incidence of DRPs

A total of 592 inpatients met the inclusion criteria for this study. The demographic and clinical characteristics of the study population are presented in Table 1. Of the patients included in the study, 62.3% (369/592) were male. The age of the patients ranged from 60 to 95 years old, with a mean age of 73.64 ± 7.96 years old. The mean body mass index (BMI) was 23.61 ± 3.58 kg/m². The median hospital length of stay (LOS) was 7(5,10) days. The median number of diagnoses of diseases was 8(6,10). A total of 176 patients (29.8%) reported a history of smoking, while 164 patients (27.7%) had a history of alcohol consumption. The proportion of patients with ≤ 9 years of education was 70.3% of patients (416/592), while 22.8% (135/592) had > 9 years of education. The median NIHSS score was 2 (0,4), the mRS score was 2 (1,4), and the ADL score was 80 (55,100). The three most prevalent multimorbidities were hypertension, affecting 501 (84.6%) patients, diabetes in 252 (42.6%), and coronary heart disease in 96 (16.2%) cases. Patients with DRPs were older, had longer hospital stays, more diagnoses, and a higher prevalence of atrial fibrillation (p<0.05). They also had higher NIHSS and mRS scores and lower ADL scores, indicating more severe neurological impairment and worse functional status.

Table 1 Patient Demographics and Disease Characteristics According to DRPs (n=592)

During the study period, a total of 15,099 medical prescriptions from CSC were reviewed by clinical pharmacists. Among the patients, 380 (64.2%) experienced at least one DRP, resulting in the identification of 576 DRPs in total. The mean number of DRPs per patient was 1.52 ± 0.96. Specifically, one DRP was identified in 251 (66.1%) patients, two DRPs in 97 (25.5%) patients, three DRPs in 13 (3.4%) patients, 10 (2.6%) patients were found to have four DRPs, and 9 (2.4%) patients were found to have five or more DRPs.

Types of DRPs

As demonstrated in Table 2, which is based on the PCNE-DRP V9.1. The most prevalent DRPs were categorized under “Other, P3” (54.3%; 313/576), followed by “Treatment effectiveness, P1” (30.6%; 176/576), and “Treatment safety, P2” (15.1%; 87/576). Within the “Other, P3” category, the predominant subcategory was “Unnecessary drug-treatment, P3.1” (44.1%; 254/576). Representative cases included the daily application of Traditional Chinese Medicine rub to prevent phlebitis, the administration of Gastrodin injection in patients lacking specific symptoms (eg, vertigo or headache), and the unnecessary continuation of potassium chloride supplementation after serum potassium levels had normalized. Regarding the primary subcategories of P1 and P2, “Effect of drug treatment not optimal, P1.2” (25.7%; 148/576) was observed in cases such as the recurrence of stroke symptoms following irregular adherence and self-discontinuation of secondary preventive medication (eg, antiplatelets), as well as the irregular use of long-term medications for multimorbidities, such as antihypertensives and hypoglycemic agents.” Adverse drug event (possibly) occurring, P2.1” (15.1%; 87/576) was exemplified by various adverse events, including hemorrhagic complications (eg, gum bleeding associated with dual antiplatelet therapy), allergic reactions (eg, rash and edema caused by iodixanol), and organ dysfunction (eg, elevated liver enzymes induced by atorvastatin).

Table 2 Types of DRPs According to the PCNE DRPs Classification Tool V9.1

Causes and Severity of the Identified DRPs

As illustrated in Table 3, “Drug selection, C1” emerged as the predominant cause of DRPs (54.0%; 311/576), followed by “Patient related, C7” (18.2%; 105/576). In the category “Drug selection, C1”, the most dominant subcategory was “No indication for drug, C1.2” (21.2%; 122/576), followed by “Too many different drugs/active ingredients prescribed for indication, C1.6” (13.5%; 78/576). In the category “Patient related, C7”, the most dominant subcategory was “Patient intentionally uses/takes less drug than prescribed or does not take the drug at all for whatever reason, C7.1” (10.9%; 63/576), followed by “Patient physically unable to use drug/form as directed, C7.9” (6.4%; 37/576).

Table 3 Identified Causes According to the Pharmaceutical Care Network Europe DRP Classification Tool V9.1 and Severity of DRPs Based on the NCC MERP Classification

The potential severity ratings of DRPs (n=576) were predominantly distributed in the B-H category, with 79.3% (457/576) falling within the low severity B-D category and 20.7% (119/576) falling within the high severity E-H category. In the category designated as “Drug selection, C1”, 87.5% (272/311) of DRPs were classified as the B-D category, while 12.5% (39/311) in the E-H category. Representative cases of severity ratings of DRPs included iodixanol injection induced serious adverse allergic reaction and having antiplatelet medication stopped without consultation (Table S1).

Medication Categories That Cause DRPs

According to the primary codes of the WHO ATC classification, common DRPs were caused by drugs acting on “N: Nervous System” (217; 34.2%) followed by “V: Various” (108; 17.0%) and “C: Cardiovascular System” (98; 15.4%) drugs, as shown in Table 4. Among the secondary codes, the drug class most frequently involved in DRPs was “N07 other nervous system drugs” (98; 15.4%), followed by “V07 all other non-therapeutic products” (90; 14.2%) and “B01 antithrombotic agents” (53; 8.3%). Specifically, regarding individual medications, the top five drugs most frequently implicated in DRPs were Traditional Chinese Medicine rub (n=90, 15.6%), followed by Ginkgo leaf extract and dipyridamole (n=27, 4.7%), butylphthalide sodium chloride (n=27, 4.7%), aspirin (n=27, 4.7%), and acetylcysteine (n=26, 4.5%) (detailed in Table S2).

Table 4 Classification of Drugs Involving DRPs According to the ATC Classification

Factors Influencing DRPs

As shown in Table 5, in the binomial logistic regression analysis identified LOS [OR=1.338, 95% CI (1.226, 1.461)], a continuous variable, and atrial fibrillation [OR=3.480, 95% CI (1.288, 9.397)], a categorical variable, as significant predictors of DRPs, with both factors achieving statistical significance (p<0.05). These results indicate that patients with atrial fibrillation had significantly higher odds of experiencing DRPs, and a longer LOS was also associated with increased odds of DRPs occurrence. In patients with atrial fibrillation, DRPs mainly involved standard treatments like warfarin and metoprolol, along with acute stroke medications such as Traditional Chinese Medicine rub, butylphthalide sodium chloride, aspirin, and Gastrodin injection. This indicates that DRPs were linked to both atrial fibrillation management and the complexities of acute stroke treatment.

Table 5 Analysis of Influencing Factors of DRPs

The Correlation Analysis Between HRQoL or Stroke-Associated Clinical Scales and DRP Numbers Among Elderly Multimorbid Inpatients with Stroke

Regarding the HRQoL analysis, the total HRQoL utility value was 0.63 ± 0.29. Patients with DRPs had a significantly lower utility value (0.57 ± 0.30) compared to those without DRPs (0.73 ± 0.25). Among the five health domains, mobility was the most severely affected, with 11.2% (66/592) of participants reporting being unable to walk around, with the proportion being higher among patients with DRPs (14.2%, 54/380) compared to those without DRPs (5.7%, 12/212). This was followed by usual activities, reported by 8.3% (49/592) of patients who were “unable to do my usual activities”, with this impairment being more prevalent among patients with DRPs (11.6%, 44/380) than among those without DRPs (2.4%, 5/212). Subsequently, self-care was reported by 7.1% (42/592) of participants as “unable to wash or dress myself”, which was also more common in patients with DRPs (10.0%, 38/380) than in those without DRPs (1.9%, 4/212) (Table 6).

Table 6 EQ-5D-5L Health Dimensions Report of Elderly Multimorbid Stroke Inpatients

Additionally, the number of DRPs showed a weak negative correlation with the HRQoL utility value (correlation coefficient = −0.291, p < 0.001) and with the ADL score (correlation coefficient = −0.241, p < 0.001), both of which were statistically significant. There was a moderate positive correlation with mRS (correlation coefficient = 0.304, p < 0.001) and NIHSS (correlation coefficient = 0.306, p < 0.001) and these were with statistical significance. The correlation analysis between these scale scores reported by study participants and the number of DRPs is shown in Table 7.

Table 7 The Correlation Analysis Between HRQoL or Stroke-Associated Clinical Scales and DRP Numbers Among Elderly Multimorbid Stroke Inpatients

Discussion

In this study, we conducted an inaugural investigation focusing on elderly multimorbid inpatients with stroke who were admitted to a CSC in China. A systematic evaluation was performed to determine the incidence and characteristics of DRPs in this demographic. Furthermore, we examined the relationship between the number of DRPs and HRQoL among these elderly multimorbid inpatients with stroke. A review of research indicates that most studies on DRPs have been limited to two primary areas: firstly, the examination of DRPs in patients hospitalized with a single condition or within broad departments, such as surgery; and secondly, there has been insufficient focus on special populations with complex medication regimens and their health outcomes in the context of multimorbidity. This study contributes additional evidence to the ongoing challenges of medication management for high-need populations within high-quality healthcare settings, particularly in the context of an aging population. It thereby establishes a foundation for future research and the development of improvement strategies.

Although several baseline characteristics showed statistically significant differences between patients with and without DRPs, their clinical significance warrants caution. Nonetheless, these findings suggest a potential association between greater clinical complexity and medication safety risks. In our univariate comparison, patients with DRPs were older and had a higher burden of multimorbidities. Advanced age is often accompanied by distinct pharmacokinetic changes, such as reduced renal clearance, and inevitably leads to polypharmacy, making patients susceptible to DRPs.⁶ Furthermore, the DRP group presented with higher NIHSS and mRS scores and lower ADL scores. Clinically, such functional dependency often necessitates nasogastric tube feeding, a frequent source of DRPs involving improper crushing of sustained-release formulations or drug-nutrient incompatibilities.²⁴ Regarding atrial fibrillation, its higher prevalence in the DRP group is clinically consistent: atrial fibrillation management necessitates high-alert anticoagulants (eg, warfarin or DOACs) with narrow therapeutic indices, thereby predisposing patients to dosing errors and adverse events.²⁵ Finally, the prolonged LOS in the DRP group reflects a bidirectional relationship: complex patients require longer hospitalization with increased error exposure, while DRP.

In this study, the incidence of DRP was 64.2%, with a mean DRP per patient of 1.52 ± 0.96, which is notably higher than the 20.3% incidence of DRP observed in a Chinese study of DRP in neurology inpatients at a tertiary care hospital, with a mean DRP of 0.25 per patient.²⁴ The existence of disparities in the study’s finding could be attributed to the inclusion of differing patient populations. The study population comprised elderly multimorbid inpatients with stroke in a CSC. This may have included patients with relatively less severe disease, a younger age structure, and fewer multimorbidity. These patients generally have less complex and more stable medication regimens, and consequently exhibit a significantly reduced risk of DRP. The most prevalent type of DRPs identified in this study was “other”. In an analytical study of adult patients in the neurology department of a Brazilian tertiary teaching hospital, the majority of the DRPs were categorized as “need for indication”.²⁶ In the present study, the “need for indication” category mainly refers to “unnecessary drug-treatment”. Despite the differences in the DRP categorization methods used in the two studies, the conclusions were similar, possibly because patients with neurological disorders face more consistent medication decisions, and the balance between therapeutic necessity and appropriateness is more difficult to manage in a clinical setting where treatments for the primary condition are prioritized over treatments for neurological multimorbidity, which tend to be simplified. It is imperative to enhance the capacity to systematically evaluate the comprehensive therapeutic requirements and risks of complex patients. One potential approach to optimize care is the implementation of a multidisciplinary collaborative management model. Around 10% of ambulatory elderly (out patients) have been shown to develop DRPs in a large Indian study, and the majority of DRPs in older adults are Adverse drug reactions.²⁷ A significant proportion of these DRPs can be prevented through the MISO approach in which drug omission (drug not needed). The MISO approach comprises monitoring (M), instructions (I), the start-low/go-slow approach (S), and omission (O).²⁸ This foundational research provides a significant impetus for our subsequent investigation into DRP management in elderly multimorbid inpatients with stroke. It offers a directly applicable and structured management model. By adopting and validating this systematic strategy, we can develop targeted interventions to optimize medication regimens, proactively mitigate risks, and ultimately improve medication safety outcomes in this vulnerable population.

The most prevalent cause of DRPs was identified as “drug selection”, followed by “patient-related” factors, which together accounted for 72.2% of all DRPs. The most prevalent subcategory of medication choice was “no indication for drug” which included the administration of Gastrodin injection to patients diagnosed with anxiety disorder, sleep disorder, or mixed dementia. The most prevalent patient-related subcategory was “patient intentionally using/taking less medication than prescribed or not taking medication at all for any reason”. It was also identified as an important reason for the effect of drug treatment not being optimal in this study. To illustrate, a patient with a 2-month history of cerebral infarction was discharged from hospital and returned to normal life, not taking aspirin, clopidogrel, or atorvastatin on a regular basis. Following the termination of medication after a period exceeding one week, symptoms of cerebral infarction recurred, resulting in readmission to the hospital. Another patient with an eight-year history of atrial fibrillation, who was not regularly taking the anticoagulant warfarin, was now hospitalized for cardiogenic cerebral infarction. The efficacy of secondary prevention strategies in reducing the incidence of recurrent stroke and mortality has been well documented. However, studies have revealed that a significant proportion of follow-up ischemic stroke patients exhibit inadequate adherence to secondary prevention measures within the first year following discharge, with only half of these patients adhering to prescribed secondary prevention medications.^29,30 Prolonged medication use has been identified as a contributing factor to this suboptimal adherence. The findings of this study demonstrate that suboptimal adherence to secondary prevention medications by patients resulted in an increased incidence of DRPs and a heightened susceptibility to disease recurrence. Furthermore, the presence of a high number of multimorbidity has been demonstrated to increase the incidence of DRPs.

According to the ATC classification, nervous system drugs and “various” agents accounted for over half of the DRP-implicated medications. Specifically, Traditional Chinese Medicine rubs, butylphthalide sodium chloride, and Ginkgo leaf extract and dipyridamole were the most frequent triggers (Table S2). The widespread use of these adjunctive therapies in China, often driven by expectations for rapid recovery, frequently leads to overuse, underscoring the need for pharmacist intervention.^31,32 Furthermore, aspirin was also a frequently implicated medication. Its long-term use for secondary stroke prevention poses a dual challenge: the risk of hemorrhagic complications and issues with non-adherence driven by patient concerns about adverse effects.

When adjusting for confounders using binary logistic regression, prolonged LOS and atrial fibrillation were identified as independent risk factors for DRPs (Table 5). This aligns with previous studies in neurological populations that established hospitalization duration and clinical complexity as key drivers of medication errors.^26,33 Interestingly, although age appeared significant in the baseline comparison and is commonly cited as a risk factor in broader literature, it did not emerge as an independent predictor in our multivariate model. This contrast is likely due to our study’s specific inclusion criteria (≥60 years), which homogenized the age-related physiological decline across the entire cohort. Consequently, our analysis suggests that in this specific geriatric population, specific clinical conditions, such as atrial fibrillation and stroke severity are stronger independent predictors of DRPs than biological age alone.

Our analysis revealed distinct patterns in the relationship between patient health status and medication safety. The correlation between the number of DRPs and ADL/HRQoL scores was found to be weak (Tables 6 and 7). This weak association is likely attributable to the fact that functional deficits in the acute phase are predominantly driven by the stroke event itself rather than by medication-related issues. In contrast, a moderate positive correlation was observed with stroke severity indices (NIHSS and mRS). This suggests that severe neurological impairment, which often necessitates complex therapeutic regimens and neuroprotective medications, is a key driver of the medication burden. Consequently, clinical pharmacist interventions should be prioritized for patients with high NIHSS and mRS scores to mitigate the risks associated with treatment complexity.

Limitation

This study has several limitations. First, as a single-center retrospective study with a limited sample size, selection bias may exist, restricting generalizability. Second, the prevalence of DRPs might be underestimated due to pre-existing interventions by the hospital’s rational drug use system and routine clinical pharmacist activities. Third, we did not stratify DRPs by stroke subtype (ischemic vs hemorrhagic) or adjust for confounders when evaluating the association between DRPs and clinical outcomes (HRQoL, ADL, NIHSS). Finally, although we excluded patients admitted during the peak COVID-19 outbreak period, vaccination history was unavailable in retrospective records. Despite China’s near-universal COVID-19 vaccination policy with high coverage rates, we were unable to adjust for potential vaccination-related effects on coagulation and stroke outcomes.

Conclusion

The presence of DRPs is a common occurrence among elderly multimorbid inpatients with stroke in CSC. The primary etiology of DRPs in this population is often attributed to challenges in drug selection problems and patient-related factors. Prolonged hospitalization days and the presence of atrial fibrillation were both found to be risk factors for DRPs incidence. A pharmacy review of medical orders has been demonstrated to be an effective method for the identification of DRPs and can serve as a foundation for the implementation of precise interventions. The findings of this study corroborate the efficacy of earlier identification and management of DRPs by clinical pharmacists in this high-risk hospitalized population.