Identification and solution of drug-related problems in the neurology unit of a tertiary hospital in China

Background The prevalence and characteristics of drug-related problems (DRPs) and factors associated with the occurrence of DRPs in the neurology unit in China remain unknown. This study aimed to determine the prevalence, characteristics and severity ratings of DRPs and identify factors associated with the occurrence of DRPs in the neurology unit of a tertiary care and academic teaching hospital in China. Methods A retrospective study of DRPs and pharmacists’ interventions for neurology patients was performed during a non-consecutive 24-month study period. Patient demographics and clinical characteristics, and pharmacist’s intervention records were collected. The characteristics and severity ratings of DRPs were categorized using the Pharmaceutical Care Network Europe (PCNE) DRP classification tool V9.00 and the National Coordinating Council for Medication Error Reporting and Prevention (NCC-MERP) classification respectively. Results A total of 242 DRPs were detected for 974 admitted patients, an average of 0.25 DRPs per patient. Treatment safety was the major type of DRPs (106;43.8%) followed by treatment effectiveness (78;32.2%). The primary causes of DRPs were drug selection (124;44.1%) and dose selection (92;32.7%). Clinical pharmacists provided 525 interventions, and most interventions occurred at the prescriber level (241;45.9%). A total of 91.4% of these interventions were accepted, contributing to solving 93.0% of the identified problems. The majority of DRPs (210;86.8%) were rated at severity categories B to D (causing no patient harm). Multiple logistic regression showed that creatinine clearance, number of medications used, nasogastric feeding, diabetes, and infectious diseases were associated with more frequent DRPs (p < 0.05). Conclusions DRPs are relatively common in the neurology unit in China, with primary causes of drug and dose selection, and clinical pharmacists can effectively reduce and prevent DRPs to optimize medication therapy.

2005 [6]. DRPs occur for many reasons, for example, inappropriate combination of medications, inappropriate medication form, and medication dose too high, leading to an increase in morbidity, mortality, and medical costs [4][5][6].
The prevention, identification and solution of DRPs constitutes the core of pharmaceutical care in which clinical pharmacists, together with the multi-disciplinary team (MDT), make an effort to improve therapeutic outcomes and the quality of life in patients [7][8][9]. Clinical pharmacists are involved in the entire treatment process from the beginning of patient hospitalization including medication reconciliation, participating in clinical ward rounds, providing medication consultation and so on. By monitoring DRPs and carrying out the appropriate interventions, clinical pharmacists play an increasingly important role in improving the efficacy and safety of medication therapy [10][11][12][13][14].
Studies from different countries discovered an average of 0.29-1.45 DRPs per patient admitted into the neurology unit [15][16][17][18]. These studies further indicate that clinical pharmacists can effectively identify and resolve DRPs in patients with neurological diseases. However, the prevalence and characteristics of DRPs in patients admitted into the neurology unit in China and factors associated with the occurrence of DRPs in this population are largely unknown.

Setting and study design
This retrospective study was performed in the 21-bed neurology unit on patients hospitalized from Apr 1, 2018 to Sep 31, 2020 at the Southern District of Beijing Tongren Hospital affiliated to Capital Medical University, a 1759-bed tertiary care, teaching and researching institution. Patients admitted into the neurology unit were cared for by a MDT, including a clinical pharmacist. The clinical pharmacist had obtained clinical pharmacy training certificates from the China Ministry of Health and had 5 years of hospital practice experience.

Data collection
Patients who required at least one overnight stay in the neurology unit were recruited during the study period. This period from Jan 1, 2020 to Jun 30, 2020 was not included in the study considering the effect of the novel coronavirus disease 2019 (COVID-19) on hospitalization rates. The clinical activities of the clinical pharmacist were: (1) reviewing medication orders to identify DRPs and proposing clinical interventions to resolve the DRPs identified, (2) visited the patients within 24-72 h of patient admission to the ward and medication reconciliation, (3) participating in daily MDT ward round and providing therapy advice to the MDT, (4) detecting and reporting adverse drug reactions, (5) providing discharge education. All DRPs identified by clinical pharmacists were documented, categorized, and entered into a data collection sheet. The sheet included patient clinical characteristics, types and causes of DRPs, pharmacists' interventions, outcomes of interventions, and screenshots of DRPs. The following patient demographics and clinical information were collected: gender, age, smoking and drinking habits, body mass index (BMI), length of hospital stay, patient admission diagnosis, concomitant diseases, creatinine clearance, types of medications used, nasogastric feeding, number of medications used.
The identified DRPs, causes, interventions, and outcomes were categorized and characterized using the Pharmaceutical Care Network Europe (PCNE) DRP classification (version 9.00) system, which was last updated in 2019 [4]. The PCNE DRP classification system is a validated DRP classification used in a variety of settings, and it includes five domains: problems (P), causes (C), planned interventions (I), intervention acceptance (A), and status of the DRP (O). Although one problem may have multiple causes and lead to more than one intervention, it leads to only one outcome. The severity ratings of the outcomes of DRPs were categorized using the National Coordinating Council for Medication Error Reporting and Prevention (NCC-MERP) classification [19]. This classification comprises four categories according to ascending severity of the patient outcome: (1) circumstances or events that have the capacity to cause error (no error, subcategory A); (2) MEs occurred without posing harm to patients (subcategories B, C and D); (3) MEs caused harm to patients (subcategories E, F, G and H); (4) MEs resulted in a patient's death (subcategory I). The severity ratings of DRPs were performed by two pharmacists independently.

Statistical analysis
We divided the patients into DRP and non-DRP groups. A descriptive analysis was performed on the patient's demographics, clinical characteristics, identified DRPs, causes of DRPs, and types and outcomes of interventions. The Student's t-test for continuous variables was used to compare means between groups when continuous variables conformed to the homogeneity of variance and normal distribution. Otherwise, the Mann-Whitney U test was used, represented by medians and interquartile ranges (IQRs) (25-75th percentiles). Categorical variables were represented by frequencies and percentages, and between-group differences were analyzed using the Chi-square test and Fisher's exact test if necessary. Variance inflation factor (VIF) values were calculated to measure the degree of multicollinearity among the variables that were significant in the univariate analysis (p < 0.1). A VIF of > 10 was considered indicative of multicollinearity and excluded from the logistic regression analysis. Based on the univariate analysis and VIF values, variables that were significant (p < 0.1) were included in the multiple logistic regression analysis to identify factors associated with the occurrence of DRPs. All statistical analyses were carried out using SPSS (Version 27.0). p values < 0.05 were considered statistically significant.

Patient characteristics
During the 24-month study period, 1225 patients were admitted to the neurology unit, 251 patients of them were excluded including 154 patients with less than one night, 95 patients admitting from Jan 2020 to Jun 2020, and 2 patients with no medication therapy. A total of 974 patients were eligible, and 198 (20.3%) patients had at least one DRP requiring pharmacist interventions. The median age of the study patients was 62.0(54.0,70.0) years and 65.9% were males. According to the International Classification of Diseases 10 (ICD-10), admitted patients suffered mainly from cerebrovascular disease, disorders of the optic nerve and visual pathways, episodic and paroxysmal disorders. The most common comorbidities were hyperlipidemia (82.9%), hypertension (70.2%), diabetes (35.8%), gastrointestinal diseases (28.6%), hyperhomocysteinemia (15.8%). Tables 1 and 2 provide the baseline characteristics of all patients and DRP status.

Identified drug-related problems
A total of 242 DRPs were identified, an average of 0.25 per patient (Table 3). Treatment Safety P2 was the major type of DRPs (106; 43.8%) followed by treatment effectiveness P1(78;32.2%). Within the treatment effectiveness P1 category, the effect of drug treatment not optimal P1.2 was the dominant category. Unnecessary drugtreatment P3.2 was the major category of Others P3.

Causes of drug-related problems identified
A total of 281 DRPs causes were identified (Table 4). Drug selection C1 was the primary cause of DRPs (124; 44.1%) followed by dose selection C3(92;32.7%). Within the dose selection C3, dosage regimen too frequent C3.4 was the dominant subcategory followed by drug dose too high C3.2. Inappropriate drug according to guidelines/formulary C1.1was the major subcategory in the drug selection domain C1.
Pharmacists' interventions to solve the drug-related problems A total of 525 interventions were suggested; an average of 2.2 interventions per DRP identified (Table 3). Most interventions occurred at the prescriber level I1 (241; 45.9%) followed by at the drug level I3(238;45.3%). At the prescriber level, intervention proposed to prescriber I1.3 was the major subcategory; and instructions for use changed to … I3.4 was the major subcategory at the drug level, followed by at the drug paused or stopped I3.5. Total 480(91.4%) interventions were accepted and fully implemented by prescribers or patients, while 45(8.6%) interventions were not accepted by prescribing physicians. 225(93.0%) DRPs were totally solved, and 17(7.0%) DRPs were unresolved (Table 3). Among the not solved O3 domain, lack of cooperation of prescriber O3.2were the major causes of failed intervention outcome.

Severity ratings of DRPs
The severity ratings of DRPs from low to high were category B (107;44.2%), category C (79;32.6%), category D (24;9.9%), category E (25;10.3%) and category F (7;2.9%). A total of 86.8% of the DRPs were rated at severity categories B to D (causing no patient harm). None of DRPs was implicated to cause death (subcategory I). Medications that caused harm mainly included antihypertensive agents, insulin, diuretics, anti-infective drugs, and nonsteroidal anti-inflammatory drugs. All-cause mortality in DRP and non-DRP groups were 0.

Analysis of factors associated with the occurrence of DRPs
Patients with DRPs had a higher median age in comparison with those without DRPs (p < 0.005). The median length of hospital stay was 14.0(11.0,15.0) days, and patients with DRPs had a longer duration of hospital stay, but no statistically significant difference (p = 0.063) ( Table 1).
Before proceeding to multiple logistic regression analysis, 16 variables (p < 0.1) in the univariate analysis (including age, smoke, creatinine clearance, number of medications used, nasogastric feeding, diabetes, cerebrovascular diseases, hypertension and so on) were assessed for multicollinearity. The results showed that all VIF values were less than 10, indicating the absence of multicollinearity. In multiple logistic regression analysis, factors of creatinine clearance, number of medications used, nasogastric feeding, diabetes, infectious diseases associated with more frequent DRPs (p < 0.05) ( Table 5).

Discussion
This is the first retrospective study conducted in China to categorically evaluate DRPs using PCNE classification, and to identify factors associated with the occurrence of DRPs in the neurology unit of a tertiary care and academic teaching hospital in China. DRPs were relatively common in the neurology unit, an average of 0.25 per patient. This number was close to that in two studies, 0.39 from one study conducted in the neurology ward in Switzerland by pharmacists [18], and 0.29 from the other study done in the inpatient ward and intensive care unit of the department of neurology in Egypt [16]. However, the average number of DRPs per patient in this study was lower than the following two studies: (1) the Brazilian study (1.26 DRPs per patient) using the trigger DRP classification [15], (2) the Iran research (1.37 DRPs per patient) using the modified PCNE DRP classification V5.01 [17]. These studies showed a significantly different mean number of DRPs per patient in different hospitals. Study population size and setting, characteristics of the patient population, study duration, definition of DRPs, types of classification systems, patterns of medication use, and differences of clinical guidelines may contribute to the variations of DRPs prevalence among studies [20][21][22].
Multiple logistic regression showed that creatinine clearance, number of medications used, nasogastric feeding, diabetes, and infectious diseases were factors associated with DRPs. The findings that age and length of hospital stay were not factors predicting the occurrence of DRPs in our analysis were different from previous studies [15,16,23]. It was worth noting that whether patients were fed through nasogastric feeding was an important factor associated with DRPs. Since a significant proportion of patients admitted to the neurology unit may require nasogastric feeding, the use of modifiedrelease medication products can be problematic [24]. Modified-release drug products such as nifedipine controlled-release tablets, felodipine sustained-release tablets, and aspirin enteric-coated tablets when crushed may alter the release profile and location, which may affect its effectiveness and safety. However, because the doctors lacked knowledge on differences of various types of dosage forms, patients that were fed through nasogastric feeding were more likely to have DRPs [21]. In our study, treatment safety was the major type of DRPs. The result was different compared to the Brazilian and Egypt study, where the untreated condition was the major type of DRP identified. The result indicates the unique role that clinical pharmacists in China play in ensuring the safe use of medications for patients in the  [15][16][17][18]25]. The highest rate of acceptance so far demonstrates that pharmacist interventions were highly useful for physicians and a sufficient relationship of trust between physicians and pharmacists. The fact that 86.8% of DRPs were rated at severity categories B to D (causing no patient harm) further proves the importance of clinical Abbreviations: C cause pharmacists in preventing medication errors and ensuring medication safety. Our study has the following limitations: (1) this was a single-center study and patient populations admitted into the neurology unit could not include all neurologic diseases due to specializations of the department, so our findings may not be generalizable to other hospitals and neurology units in China, and (2) medication review was performed by one clinical pharmacist, and (3) our study did not assess the relationship between patients' longterm outcomes and the resolution of DRPs.

Conclusion
This study indicates that the prevalence of DRPs is relatively common in Chinese neurology patients. Treatment safety is the major type of DRPs. Improving clinical pharmacy services in neurology unit could contribute to rational medication use and ensure patient safety.