This study found that DRPs exist in prescriptions dispensed in the community (86.5%, 346/400) and community pharmacists identified and resolved some of them (36.4%, 441/1211, and 33.5%, 406/1211 respectively). Community pharmacists identified wrong doses, frequencies, dosage forms and durations of drugs written on prescriptions. However, they overlooked DRPs related to missing information, drug duplications, and drug interactions. Pharmacists also resolved most of the DRPs identified (406/441) using self-judgement (366/406). Patients were sent back to prescriber (14/406) and refused dispensation (9/406) in some cases.
During the study period, we reported a DRP rate of 86.5% which is considerably higher than DRPs reported in other studies. Although most DRPs may not end up in immediate clinical consequences, they could lead to poor compliance and sub-therapeutic effects in the long run. In a study done in Europe: Austria, Denmark, Germany, The Netherlands, Portugal and Spain, Paulino et al., reported a total of 451 DRPs in 277 (out of 435) prescriptions (63.7%) [19] and a study conducted in India reported 90 DRPs among 215 prescriptions (41.8%) [1]. It is difficult to directly compare findings of this study with other international studies due to the explicit DRP definitions we used to capture even trivial issues in prescriptions. However, the fact that more than 80% of prescriptions had at least one DRP is similar to results reported from the West (18% to 88% of patients) [13]. However, it should be noted that the number of identified DRPs depend on multiple factors including study design, type of settings, study population, classification system used and the denominator used for statistical analysis.
There were 17.9% (217/1211) of drugs where the strength was not indicated on the prescriptions and was prescribed as ‘1 tab’ or ‘2 tab’ and the community pharmacist resolved to dispense the lowest strength available. Although unnecessary overdosing and toxicities may be avoided, selecting the lowest strength option could result in therapeutic failure if the prescriber had intended a higher dose. Drug selection was the next most common problem identified in this study of which 10 out of 15 inappropriate combinations and drug duplications identified were potentially harmful. A previous prospective, cohort study conducted in two Sri Lankan medical wards to evaluate medication appropriateness, also reported a rate of 1.7% drug duplications [20]. In this study the most likely reason for drug duplications was prescribing in brand names. Pharmacists need to be advocated on identification of DRPs, even the potential issues that could result in harm, and on prioritizing for corrective actions.
There were inappropriate abbreviations observed in this sample of prescriptions, but these were not included as DRPs in the analysis. Inappropriate abbreviations were used to indicate frequency (‘d’ to denote daily, ‘m’ to denote mane, ‘n’ to denote nocte), drug name (HCT to denote hydrochlorothiazide, CBZ to denote carbamazepine and HCQ to denote hydroxychloroquine) and dose units of drugs (mcg to denote microgram and IU to denote International Unit). Previous studies too have reported the prevalence of non-standard abbreviations and incomplete units used in Sri Lankan prescriptions [6].
It is encouraging to note that, community pharmacists in this study had taken corrective actions for 92% of the DRPs detected by them. Other published studies indicate that, counselling and reassuring patients, followed by practical instructions, were by far the most common types of interventions used by pharmacists when resolving DRPs [19,27]. This practice was not frequently seen in our study but self-resolving of DRP was frequently observed instead. It is important to give clear instructions to patients on dose timing and strength, for drugs like thyroxine, warfarin, alendronate, methotrexate, proton pump inhibitors and anti-diabetes drugs are often associated with significant drug-drug and food-drug interactions. We observed that most of these instructions were not clearly indicated in prescriptions. Often important drug related instructions were given on the dispensing label even if it was missing in the prescription (n=364; 89.6%). Patients were sent back to prescriber to clarify DRPs (n=14; 3.4%) in a few instances, mostly when they encountered ambiguous names of medicines that were illegible (n=8; 1.9%), when frequencies of medicines were missing (n=3; 0.7%), and when too high drug doses were prescribed (n=3; 0.7%). Pharmacists refused to dispense drugs on nine occasions (n=9; 2.2%) which were related to outdated prescriptions, and missing prescriber credentials on prescription. Pharmacists also made every effort to resolve issues by discussing with patients (n=8; 1.9%), checking with recent written medical histories of patients (n=4; 0.9%), and through discussion with fellow pharmacists (n=5; 1.2%) which is acceptable if DRPs are resolvable beyond doubt through these measures. It is noteworthy that none of the community pharmacists consulted written references or decision support software in this study possibly due to the unavailability of such resources. Ezeuko et al., too reported this observation in resource limiting settings [3].
To our knowledge, there are no studies that have focused on the prevalence of a wide-range of DRPs in prescriptions brought for dispensation to community pharmacies, the ability of the community pharmacists to detect such DRPs, and their practices when resolving these DRPs using prescription review and direct observation methods as in our study. However, there are some limitations to our study which needs to be considered. We carried out the study only in one community pharmacy in Sri Lanka. Therefore, our sample does not represent the total study population of this country. However, this study provides evidence on a very important but seldom researched area. A direct observation was done to assess if pharmacists identified and took corrective action for DRPs, but it is possible that a ‘Hawthorne Effect’. Nevertheless, it is concluded that behavior of observes are seldom affected by direct observation [5].