Audit of Prescribing Patterns in Saudi Primary Health Care: What Lessons Can be Learned?

Introduction

For the practice of family medicine to continuously improve in quality and productivity, and justify an increased share of resources, it is essential that practice audits are periodically undertaken.¹ In their absence, decisions about family medicine practice will rely on people from outside the discipline who are liable to introduce flawed remedies in the current atmosphere of cost containment, accountability and doubts about quality.²

The prescription of medicines and drugs is one of the most important factors in the rising costs of health services in both developed and developing countries.³ Western studies reveal that a high percentage of consultations end with prescriptions, regardless of appropriateness or necessity to prescribe.^4,5 In a Saudi health center, 75% of consultations ended with a prescription, while health education took place in only 7.6% of the consultations.⁶

The English prescribing analysis and cost (PACT) scheme provides general practitioners (GPs) with information about their prescribing numbers and costs related to therapeutic groups, compared with local and national averages. Prescriptions are collected by a national authority, and information is fed back to GPs on a quarterly basis, either in a simple “headline” format, or in more detail when this is requested, or when a practice’s costs are substantially greater than the local average.⁷ The motives for using PACT include curiosity for information, desire to reduce cost, and a wish to improve the quality of prescribing.

Previous studies in Saudi Arabia have tackled some aspects of the prescribing habits of physicians. One of these studies was hospital-based,⁸ and another used a limited sample in mostly rural areas.⁹ The present study is the most extensive in a large urban area, aiming to assess the prescribing pattern of the primary health care (PHC) physicians in the studied PHC centers in the city of Riyadh, the capital of Saudi Arabia.

PATIENTS AND METHODS

Eight health centers were chosen to represent the geographical parts of Riyadh city by a stratified random sample.¹⁰ Sixteen doctors were recruited from the study health centers by simple random sampling, one male and one female doctor from each health center. Prescriptions issued by the study doctors during one week in October 1994 and one week in April 1995 were collected using systematic random sampling. The pharmacies at health centers are required to retain all prescriptions and so little difficulty was encountered in their collection afterwards. The doctors were not aware that their prescriptions would later be studied.

The information recorded on standard prescription forms includes patient’s age, sex and nationality, in addition to drug name, dose, frequency and route of administration. The Ministry of Health physicians are requested to allocate clinical diagnoses in the prescription forms in 25 broad diagnostic categories, e.g., “0.24” indicated disorders of the stomach, esophagus and small intestine (dyspepsia, peptic ulcer, gastritis, appendicitis). The data were processed and presented using the EP Info Statistical Program version 5.

RESULTS

The study population was, in general, young, with approximately 23% below five years, 24% between 5-14 years, 43% aged 15-44 years and only 10% above 45 years. Seventy-three percent of the patients were Saudi, and males constituted 52% of the patients for whom prescriptions were issued. Male patients represented 40.8% of the total number of patients who visited the study centers during the study period. The majority of the prescriptions (92%) had complete documentation of information, including patient’s demographic characteristics, provisional diagnosis, drug name, dose, route and frequency of intake.

A total of 17,067 prescriptions were written by the study physicians. The 20 most frequently prescribed drugs were identified and classified into nine drug categories (Table 1). Antihistamines represented 25% of the prescribed medications, followed by paracetamol (20.3%) and antibiotics (14.7%).

Table 1. The most frequently prescribed drug classes.

Upper respiratory tract infection (URTI) was the diagnosis recorded in 53.4% of all prescriptions (Table 2). Other diagnoses were stomach disorders (5.3%), musculoskeletal problems (4.9%), pregnancy (4.4%) and skin diseases (4.0%). In about 8.8% of all prescriptions, the diagnosis did not fit any of the MOH diagnostic codes.

Table 2. The diagnostic labels most frequently used by physicians.

The most frequently prescribed drugs for the four most common diagnostic categories were grouped into drug classes (Table 3). Paracetamol was prescribed for 43% of patients with URTI, followed by antihistamines (28%) and antibiotics (26%). Hyoscine N-butylbromide and antacids were prescribed for stomach disorders, while nonsteroidal anti-inflammatory drugs (NSAIDs) and vitamin B complex were prescribed for musculoskeletal problems and iron and folic acid for pregnancy (Table 3).

Table 3. The drugs most frequently prescribed for the four most common diagnoses.

DISCUSSION

Assessment and audit studies have the potential to make objective evaluation and analysis of health professionals’ work, and provide them with feedback to stimulate thinking about their practice and looking for ways to improve their own performance. Audit studies can and should become a method of increasing job satisfaction and means of education for health professionals, rather than being perceived as a threat or another bureaucratic burden. ¹¹

It must be acknowledged that the study period of two weeks may not present the whole prescribing pattern. However, the fact that the two weeks were six months apart and the study included 17,067 prescriptions issued by 16 randomly selected PHC physicians working in eight health centers, is expected to provide a not insignificant idea of the prescribing pattern of Riyadh city health centers.

Although more female patients attended the health centers (59%), a smaller proportion of them (48%) received prescriptions. This is in line with other studies in Saudi Arabia.^8,9 It is well known that women suffer more ill health,^12,13 use the health services more, and consult more frequently than men.^13,14 Possible explanations for a lower female prescription rate include the following: 1) a proportion of female visits are for preventive reasons (e.g., antenatal visits) that may not require prescriptions; and 2) men may exert greater pressure on their physicians to issue prescriptions. The consultation time was found to be significantly longer with male compared with female patients in Riyadh health centers.⁶ The rate of prescriptions with complete documentation of information in this study (e.g., age, sex, nationality, provisional diagnosis and frequency of medication) was comparable with a study in Saudi PHC centers,⁹ but better than that reported from hospital ambulatory care.^8,15

The predominance of URTI diagnosis (53%) at the expense of chronic illness indicates a great need for patient health education and promotion. Much of the patients’ help-seeking behavior regarding acute self-limiting illnesses has probably been learned from doctors. Prescribing an antibiotic or an antihistamine for URTI may reinforce the patients’ belief in the necessity of such treatment every time he develops such symptoms. A review of Kentucky Medicaid claims in the United States has shown that physicians prescribe antibiotics for approximately 60% of patients with simple colds.^16,17 Modification of patients’ help-seeking behavior by encouraging them towards self-treatment is time-consuming, but a long-term investment. For example, patients should be educated to treat diarrhea by oral rehydration solution (ORS) and influenza by antipyretics and rest. The limited contribution that the doctor or an antibiotic can make to many self-limiting conditions such as URTI should be emphasized. Furthermore, patients should be warned about the risk of chronic illnesses and the benefit of their treatment and good control.¹⁸ Several patient education methods have been shown to be associated with fewer consultations for self-limiting conditions. These include audiovisual means,¹⁹ educational leaflets,²⁰ or booklets.²¹ In Iceland, publicity campaigns directed at the problem of pneumococcal resistance and its relationship to antibiotic use resulted in a decrease in sales of antimicrobial agents and a concomitant decrease in the prevalence of resistant pneumococcal isolates.²²

The most frequently issued drugs in the study were antihistamines (25%), paracetamol (20%) and antibiotics (15%). This finding is consistent with that of similar Saudi studies in the PHC setting,⁹ and in other developing countries as well.²³ The rate of antibiotic prescriptions in hospital outpatient clinics (16.0%)¹⁵ was similar to the current study figure.

The frequency of issuing prescriptions of paracetamol (43%), antihistamines (28%) and antibiotics (25.6%) for URTI is similar to the findings of the Al-Baha study.⁹ Prescribing an antibiotic for one-quarter of patients with URTI is a matter of concern. Besides the drug costs, antibiotic use is not benign; it increases the risk of colonization with resistant organisms,²⁴ and side effects occur relatively frequently.²⁵ Controlled trials of antimicrobial treatment of the common cold have consistently failed to show that treatment changes the course or outcome,¹⁷ or has a preventive effect on lower respiratory infections.²⁶Subsequent to viral URTI, antibiotics were not found to prevent or decrease the severity of bacterial complications.²⁶ It was shown that 20% of adults with rhinovirus colds continue to cough,²⁷ and more than 30% of children and adolescents with viral URTI cough have mucopurulent rhinitis¹⁷ (discolored nasal discharge) for more than two weeks. This indicates that cough (and other lingering symptoms) is an expected part of uncomplicated viral URTI. A meta-analysis of six randomized, placebo-controlled antibiotic trials for bronchitis among adult patients concluded that there is no evidence to support the use of antibiotic treatment for acute bronchitis.²⁸ If unnecessary antibiotic use can be curtailed, there are indications that the community, as well as the individual patient, will benefit. In Japan, the resistance of group A streptococcal isolates to erythromycin, which was 62% in 1974, fell to less than 2% in 1988 when macrolides use was reduced from 22% of all antibiotics prescribed to only8%.²⁹

There have been many forms of intervention aimed at changing physicians’ prescribing behavior. These have included audit studies,³⁰ group discussion and feedback,³¹ introduction of general practice drug formulary³² or hospital formulary,³³ guidelines for antibiotics³⁴ and nonsteroidal anti-inflammatory drugs,⁷and a letter from a licensing body combined with group education.³⁵ Regardless of the type of study, the majority have shown positive effects on savings. The benefits of the intervention studies, namely the use of fewer and cheaper prescriptions, are shown to disappear over time,^36,37 which suggests the need for repeating the intervention at frequent intervals. Rational prescribing messages should be promoted at national and local medical meetings and the input from local practitioners should be considered.³⁸ In addition, endorsement by major professional organizations, as well as regional and local opinion leaders, may be useful.

In the study, antihistamines were found to be the most frequently prescribed drugs and the second most frequent (after paracetamol) for URTI. The literature offers little support for the use of antihistamines for the common cold.³⁹ While a few studies supported the use of cold preparations to alleviate symptoms of sneezing and runny nose in adolescents and adults,^40-42 their use for children younger than five years is not evidence based.⁴¹ As the under-five group constituted a large proportion of patients with URTI for whom antihistamine was prescribed, this implies that a huge amount of costly drugs were unnecessarily prescribed. In the West, the situation is probably not much better. An audit study from the charts of patients with URTI in the US found that 33% of patients younger than five years were given a prescription. Non-evidence-based and unnecessary medications accounted for almost 60% of the total prescription costs.⁴¹ There is a need to increase doctors’ awareness of the lack of proven benefits and the definite cost and side effects of many prescriptions for self-limiting illness.

Another matter of great concern was that although it has been repeatedly demonstrated that a large proportion of patients seen in PHC have a clinically significant mental disorders (MD),^34,43,44 no psychological diagnosis was used. Two organizational issues have probably contributed to this striking finding. First, the diagnostic code of the Ministry of Health for the commonly encountered diseases does not include MD. And second, PHC physicians are prohibited from writing psychotropic drug prescriptions. These issues need to be corrected before other actions are undertaken. A more appropriate prescribing system that allows PHC physicians to prescribe psychotropic drugs that are safe, cheap, effective and unlikely to cause dependence should be adopted.⁴⁵ Physicians’ low identification index and lack of sufficient training in this area were documented.^44,46 Short intensive courses on the detection and management of MD have been recommended to improve doctors’ detection capacity.^44,45