Antibiotic Prescribing Pattern; Errors Perspective in Primary Health Care Centers in Riyadh City, Saudi Arabia

Objectives: To explore the medication error in prescribing antibiotics in primary health care centers in Riyadh City, Saudi Arabia. Methods: It is a retrospective analysis of antibiotics prescribing errors of randomly selected twenty-five primary health care centers in Riyadh city. One month’s prescription had been reviewed. The expert pharmacist revised the prescription and documented the errors. The medication error form of MOH had been used to document the errors. Results: During the study period, 18,031 prescriptions were recorded in all the included study settings. Of the total prescriptions, 3,879 (21.51%) contained antibiotics and were included in the current study. Of the total prescriptions, 1388 (35.78%) prescriptions detected medication errors were analyzed in the current study. The most errors detected in the prescription were height not written 1388 (100%), followed by diagnosis not found 535 (38.54%), and body weight not written 492 (35.45%). The antibiotics errors occurred with pediatrics and adolescents < or = 20 years 673 (50.08%) and adults > 20 years with 671(49.92%) with statically significant differences between all medication error types within pediatrics and adults ( p =0.000). The errors identified mainly were related to inappropriate drug selection that occurred in 748 (53.9%) of the prescriptions, followed by low therapeutic dose 277 (20.0%) and dose frequency 175 (12.6%), with statistically significant differences between all types ( p =0.000). The most Medications with errors were Amoxicillin 576 (41.5%), Amoxicillin/clavulanate 277 (28.1%), Fucidic Acid 123 (8.9%), and Azithromycin 116 (8.4%). Conclusion: One-third of primary healthcare centers prescriptions contained antibiotics and had at least one error. The most errors related to demographic information and most prescribed antibiotics at primary healthcare centers. Targeting electronic prescribing systems and implementing stewardship antimicrobial programs at primary healthcare centers are highly suggested.


INTRODUCTION
][3][4][5][6][7][8][9][10][11] Drug-related problems consist of eight problems based on the American Society of Health Care Pharmacists in the United States of America. 12,13Medication errors are a severe concern of drug-related problems. 14,15The error might occur when prescribing, dispensing, and administering medication.Prescribing errors are a significant concern today, clinically and economically.7][18][19][20][21][22][23][24] Antibiotic was the most implicated prescribing error at primary healthcare centers. 25,26[29][30][31][32] The pattern errors of antibiotic prescriptions are crucial to understanding the problem's magnitude and factors that might affect the mistakes and recommend preventive measures in future Primary health centers.The Primary Health Center (PHC) is accountable for treating and preventing diseases and is therefore treated as the base of the health care system.Thus, by determining medication errors in primary health centers and avoiding or reducing them, the healthcare system can improve health outcomes and decrease hospitalizations and costs. 25The pharmacist's role in improving clinical and economic outcomes by avoiding medication errors is well-established locally.
Antibiotics are frequently used in health service settings and should be utilized rationally to provide optimum benefits. 33Many studies stated that the overuse or misuse of Antibiotics might modify their efficacy of antibiotics. 34Antibiotics fail when prescribed for diseases not indicated or at low regimens. 35Detailed data on antibiotic use must be obtained to inhibit overprescribing or misuse of antibiotics. 36Unfortunately, very few studies investigated the issue of Antibiotic prescribing errors in PHCs in Saudi Arabia.The authors were unfamiliar with previous studies on antibiotic prescription errors at the primary healthcare center in Riyadh city or locally or in Gulf and Arabic countries about the topic.Therefore, the current study was conducted at a randomized selected primary healthcare center in Riyadh City to determine the pattern of antibiotic prescription errors.

METHODS
The study was a retrospective analysis of the antibiotic prescribing Errors in primary health care centers in Riyadh city.Riyadh City is served by almost 120 Primary Healthcare Centers (PHC) distributed across five health sectors.The study sample consisted of Antibiotic prescriptions.For this study, we selected 25 PHCs from a random Albahouth, et al.: Antibiotic Prescribing Pattern; Errors Perspective sample.These centers were typical of those across the city, offering services to a large population relatively representative of Riyadh City.The scope of primary health care services is explained in a previous study. 37he target population was all patients attending primary care centers.Antibiotic prescriptions given a month were recorded by a medication error form, certified by the general administration of pharmaceutical care from the Ministry of Health, as discussed before. 38That is to determine the medication error rate in antibiotic prescriptions in PHCs and the types of medication errors.The qualified pharmacists recorded the details in the medication error form, which consisted of information about the patient's age, gender, nationality, allergy, medication name, dosage form, type of error, and the specialty of prescribers.Data collection was performed every day except for weekends.Permission to conduct the research was obtained from the general administration for analyses and studies at the Ministry of Health (1440-1033702).A letter of approval was sent to the health sectors.Data was collected for one month through December 2018.The Statistical Package of the Social Sciences version 20 software analyzed the data.Descriptive statistics were used to calculate frequencies and percentages.Chi-Square was used to determine the difference between categorical variables.A p-value of 0.005 was set as statistical significance at a 95% confidence interval.

RESULTS
During the study period, 18,031 prescriptions were recorded in all the included study settings.Of the total prescriptions, 3,879 (21.51%) contained antibiotics and were included in the current study.Of the total prescriptions, 1388 (35.78%) prescriptions detected medication errors were analyzed in the current study.The description of patients with prescriptions containing antibiotics medication errors is shown in Table 1.The proportion of male patients was 648 (46.7%), while 740 (53.3%) were females, with statistically significant differences between them (p=0.001),while there are non-statistically significant differences between pediatrics (< or = 15 Years) and adults (> 15 years) ((p=0.575).The nationality distribution for 1270 (91.49%) patients was Saudi nationality with statistically significant differences between non-Saudi nationality (p=0.000), while there are non-statistically significant differences between pediatrics (< or = 15 Years) and adults (> 15 years) ((p=0.575)((p=239)).The patients' mean age (SD) recorded was 24.61 (19.92), and the weight was 10.54 KG (19.33).The patients' mean age (SD) recorded was 24.61 +/-(19.92),with the average age for pediatrics was 6.56 +/-(3.88),and for adults was 39.74 +/-(14.66),with statistically significant differences between pediatrics and adults (p=0.000).The number of medications per prescription was 2.77 within (1-6) per prescription, with a higher number of medications in pediatrics prescriptions with statistically significant differences between pediatrics and adults (p=0.000).The number of Antibiotics per prescription was 1.07 within (1-3) per prescription, with a higher number of antibiotics in adult prescriptions with statistically significant differences between pediatrics and adults (p=0.000).The Number of Non-Antibiotics per prescription was 1.7 within (0-5) medications per prescription, with a higher number of non-antibiotic medications in pediatric prescriptions with statistically significant differences between pediatrics and adults (p=0.000).The most documented diagnosis was Upper respiratory tract infection 216 (15.6%) and Tonsillitis 88 (6.3%) in pediatrics and adults, as explored in Table 2.

Patient characteristics errors
The most errors with patient characteristics were height not written in 1388 (100%) antibiotics prescriptions and diagnosis not written 535 (38.54%), followed by body weight not written 492 (35.45%) and age not written 44 (3.17%).The adult prescription had higher errors than pediatrics, with statistically significant differences between pediatrics (< or = 15 Years) and adults (> 15 years) ((p=0.000).The adult's prescriptions had more errors with writing height 731 (54.4%), body weight 388 (50.3%), and the diagnosis 304 (22.6%) than pediatrics prescriptions with statistically significant differences between pediatrics and adults (p<0.01).In contrast, the pediatrics prescription had more errors in writing age 44 (3.3%) than adults, with statistically significant differences between pediatrics and adults (p<0.000).However, there were no errors with writing gender or nationalities among pediatrics and adults, with non-statistically significant differences between pediatrics and adults (p>0.5), as explored in Table 2.

Antibiotics errors
The most type of route of administration errors thought orally 1197 (86.2%) with statically significant differences among them (p=0.000).

Prescriber errors
The number of prescribers who prescribed the antibiotics in this study was general practitioners (GPs) 943 (68.01%), specialists 439 (31.67%), with statically significant differences between them (p=0.000).The The general practitioner 1,694 (69.05%), and specialist physicians 755 (30.77%), had the most demographic errors, with height, diagnosis, and body weight not documented, with statistically significant differences between them (p=0.000).All specialties with the most errors in demographic information were height, body weight, diagnosis, and age, emphasizing that the pediatric physician had the highest error, 1086 (32.19%), followed by general physicians 1,083 (32%) and family medicine 748 (22.16%), with statistically significant differences (p=0.000), as in Tables 5 and 6.Most antibiotic errors were done by general physician 943 (76.93%) and specialists 439 (31.63%).The most common types of medication errors were inappropriate drug choice error type 320(66.9%)and low dosing 58(12.1%)by General Physicians.
In contrast, the family physician had more errors with inappropriate drug choice error type 162(43.8%)and 75(20.3%)dosing frequency.However, the dentist had an inappropriate drug choice error type 22 (47.8%) and duration of treatment 13(28.3%).There were statically significant variations among physician's specialties and types of mediation errors with statistically significant differences (p=0.000), as declared in Tables 7 and 8.

Factor associated with antibiotic error
Logistic regression multinominal analysis declared that The odds of antibiotics prescribing errors (Inappropriate drug choices) were seven times higher with the bactericidal antibiotics (OR = 7, 95% CI = 1.144-6.109,p-value = 0.000) compared to Bacteriostatic, other factors statistically significant differences (p>0.05).In antibiotics prescription (High doses), the odds of errors were 414      All other factors: gender, nationality, age levels, non-drops (nose, ear, eye) dosage form, route of administration, drug names, type of antibiotics, prescribing goals, rank of prescribers, prescriber specialties, and number of antibiotics per prescription were nonstatistically significant differences (p>0.05).

Inappropriate dosage form
All other factors; gender, nationality, age levels, dosage form, route of administration, drug names, type of antibiotics, prescribing goals, the rank of prescribers, prescriber specialties, and number of antibiotics per prescription were non-statistically significant differences (p>0.05).

number of antibiotics per prescription
All other factors; gender, nationality, age levels, dosage form, route of administration, drug names, type of antibiotics, prescribing goals, the rank of prescribers, prescriber specialties, and number of antibiotics per prescription were non-statistically significant differences (p>0.05).
locally.However, the zooming for prescription-contained antibiotics was not found locally.Thus, the required investigation on primary healthcare centers to explore the situation compared to hospitals.The current study was done at randomized selected primary healthcare centers with revised one-month regular non-electronic prescriptions.On those prescriptions, they choose the prescriptions containing antibiotics that had been prescribed and reviewed by the clinical pharmacist for antibiotic prescription errors.Those errors were in the demographics information errors and medication errors, as revealed in the discussion.27][28]30 In contrast, the percentage of errors was higher than in one study that used electronic prescribing. 29e study showed that most of the prescribing was for females, related to primary healthcare visiting, which is most suitable for females to visit the center beside their home, and non-significant between pediatrics and adults like previous study. 19Most patients were female, resembling significant differences (p>0.05).The odds of antibiotics prescribing errors (Treatment duration) were 3 B times higher with the Drops (nose, ear, eye) (OR = 3 B, 95% CI = 0.202 B-4.9 B, p-value = 0.034) compared to other dosage forms of antibiotics, other factors statistically significant differences (p>0.05).Inappropriate dosage forms and number of antibiotics per prescription are not associated with antibiotics prescribing errors, with non-statistically significant differences (p>0.05), as illustrated in Table 9.

DISCUSSION
National medication safety is one of the critical programs established at the Ministry of Health hospital. 41,42The proposed problem is preventing drug-related problems with an emphasis on medication errors.4][45][46][47] The pharmacist plays an essential role in the program and prevents medication errors and related issues.9][50][51][52][53] Various studies revealed antibiotic prescribing errors.6][27][28] Despite the majority of our study sample being female, there was a non-significant between pediatrics and adults in gender, which revealed that antibiotic errors might have occurred in both genders without differences.
Furthermore, most subjects were Saudi, which is expected because of the illegality process for Saudi and non-Saudi working at government agencies.There was a non-significant between pediatrics and adults in nationality, which declared equal occurrences of antibiotic errors in different nationalities.Most previous studies did not explore their nationality, [25][26][27][28] which was challenging to compare with previous studies.The primary pediatric age was around six years, while adults were around 40.That expected those age might exposed to viral infection and visit the nearest ambulatory care clinic.The pediatric prescription received higher antibiotics than adults without apparent reason.That is expected because most pediatric diagnostics cases were upper respiratory tract infections and more prescribed antibiotics.
Antibiotic prescribing errors might be divided into various types, including patient characteristics errors, antibiotic errors, and prescriber errors.In the patient characteristics errors, some elements were not found in the prescription, such as height, followed by diagnostic missed in one-third or prescribing and body weight not included in one-third of the primary healthcare sample emphasizing for adults prescriptions like previous study. 26The pediatric prescription had more errors in writing age than the adult prescription, with statistically significant differences.That is related to not using an electronic prescription, which found lower antibiotics prescribing errors in a previous study). 29In addition, the patient demographics information is critical, especially for pediatrics, because almost half of pediatric errors emphasize ages less than or equal to 15 years, which requires the body to calculate the appropriate dosing.Thus, most antibiotic errors were low dosing in the study because there was no information on patients to calculate the appropriate dosage.
Antibiotic errors are mostly found with oral route and dosage forms because most primary healthcare centers do not have emergency departments, and Parenteral medicine is unavailable.If there are any emergency cases, they should be transferred to the hospitals.Most errors occurred with an adult were tablets, while in pediatrics, it was a suspension that was expected because each dosage form required based on age level, either adults or pediatrics.
Most antibiotics had errors in using generic names, and for treatment, that is expected because the long duration of treatment might additionally be another error.The adult's prescription had more errors in the treatment than pediatrics, incorrect mistakes in the management period, and more extended time than pediatrics.However, there were no differences between pediatrics and adults in writing trade or generic because the prescriber did not differ in both populations with writing antibiotics.The most common antibiotic error was an inappropriate drug of choice in adults and a low dose in pediatrics.That was expected because one-third of prescriptions missed the diagnosis and body weight to revise the diagnosis and calculate the appropriate dosing for an antibiotic.Electronic prescription is the best choice to solve those problems). 29Besides, the implementation of medication safety programs in the primary health care center.Most antibiotics had errors, such as Amoxicillin and Amoxicillin/clavulanate, that were expected because most of the diagnostic in adults and pediatrics was upper respiratory tract infections, and those medications are commonly used in practice.
Most patient characteristic errors come from general physicians, which is expected because new staff need an orientation program of medication safety lectures.Unfortunately, most errors are made by pediatrics and general physicians.Thus, special attention should be given to physician prescribers for introductory medication safety lectures to prevent future medication errors.Most prescribers had errors in general practitioner followed by the specialist emphasizing general physician and family medicine that's most of the diseases they can be treated by then and sometimes need to prescribe the antibiotic.The general physician had more errors with adult prescriptions, while the family physician had more errors with pediatrics, which might be related to insufficient knowledge of pediatrics and not being treated by specialized pediatrics.
The most common antibiotic errors were inappropriate drug choices, insufficient doses, and frequent dosing, resembling previous studies. 19,27,30,54The current differed from the previous study, which used electronic prescriptions 29 such as wrong quantities, wrong doses, and duplicated antibiotics. 29That is expected because the essential information on the patient's character and the diagnosis was not found.Most of The physician's specialties had the same type of committing antibiotic errors with the inappropriate drug of choice and low doses.However, the dentist had a second significant type, and it was the treatment duration that was related to antibiotic policy that was not implemented at primary health care centers.The medication errors are undoubtedly due to the lack of scientific efficiency of prescribers, insufficient education, updates and training, and limited experience.In addition, Heavy patient loads and work pressure due to inadequate staffing may also be tentative reasons for the error.However, many primary health centers in Riyadh city suggested otherwise.Most antibiotics with errors were Amoxicillin and Amoxicillin/Clavulanate, classified as penicillin groups like previous studies in hospital settings. 26,29,54That is expected because the primary healthcare center is the first, quickest, and easiest healthcare service patients can contact if they suffer from infectious diseases.The antibiotic classes as penicillin differed from previous studies, 27 in which many errors with Cephalosporins might related to more complications and properly used penicillin before visiting the emergency section.
Several factors might associated with increased or reduced antibiotic prescription errors.Most factors, such as In the current study, age levels, dosage form, route of administration, drug names, prescribing goals, the rank of prescribers, prescriber specialties, and number of antibiotics per prescription, did not affect the type of most antibiotics prescribed errors.However, inappropriate drug choices might be affected by the type of antibiotics, bactericidal or bacteriostatic, related to choosing appropriate antibiotics for each diagnosis; that factor might be affected because the antimicrobial stewardship program was not implemented at primary healthcare centers.The high or low-dose errors might increase with pediatrics due to special attention in dosing calculations without writing the patient's body weight.There are more pediatric antibiotic errors in treatment duration with eye, ear, and nose dosage form errors because they need special attention for the dosage form and dosing calculation, particular administrations, storage conditions after opening the drops, and specialist consultations.The Inappropriate dosage form errors and number of antibiotics per prescription will not affect the errors that might related to a few medications per prescription.The medication errors are undoubtedly due to the lack of scientific efficiency of prescribers, insufficient education, updates and training, and limited experience.
In addition, Heavy patient loads and work pressure due to inadequate staffing may also be tentative reasons for the error.However, the large number of primary health centers in Riyadh suggests otherwise.That differed from previous studies due to different site study settings and different types of analysis, in which our study had more detail with each type of antibiotic error. 27,54

LIMITATIONS
The study had various advantages, including an appropriate sample size better than most previous studies, and the demographic information in gender, nationality, and diagnosis almost resemble adults and pediatrics.
The study explored a clear picture of antibiotic prescribing errors in pediatrics and adults and the factors associated with errors.However, the study has various limitations, including using a manual, not computerized, or alerting system in detecting antibiotic errors.Various factors were missed, such as height, body weight, and diagnosis, which were difficult to compare.

CONCLUSION
This study demonstrated a high prevalence of medication errors related to antibiotics.Improving how we prescribe and use antibiotics is critically effective in treating common infections and protecting patients from adverse drug reactions.The antibiotic errors revealed the absence of antimicrobial stewardship programs at primary healthcare centers.The antibiotics prescribing errors might lead to over or underuse antibiotic use and might lead to antibiotic resistance.Targeting to implement antibiotic guidelines emphasizing education and training is highly recommended.

Table 3 : Antibiotics errors. < or = 15 Years > 15 years All ages p-value Rout of administrations Response Count Response Percent Response Count Response Percent Total p-value (X2) Response Count Response Percent
continued...

Table 4 .
561 (40.7%) and dose frequency errors 94 (7%) mainly occurred in adults with statically significant differences among all types of medication errors and age levels(p=0.000).In contrast, low dosing 259 (19.3%) and Inappropriate drug choice 173 (12.9%) mainly occurred in < or = 15 with statically significant differences among all types of medication errors and age levels (p=0.000) as explored in Table3.

Table 9 : Logistic regression analysis of factors associated with Antibiotics prescribing errors.
; gender, nationality, age levels, dosage form, route of administration, drug names, prescribing goals, rank of prescribers, prescriber specialties, and number of antibiotics per prescription were non-statistically significant differences (p>0.05).
BacteriostaticAll other factors; gender, nationality, age levels, dosage form, route of administration, drug names, prescribing goals, rank of prescribers, prescriber specialties, and number of antibiotics per prescription were non-statistically significant differences (p>0.05).