Three hundred and fifty patients were interviewed, of whom 15 were excluded due to lack of data. Consequently, the final sample was composed of 335 patients (Fig. 1), who were predominantly female (65.7%) and white (76.7%). The mean age was 44.9 years (SD=19.2). Most (61.4%) had elementary education and were from the metropolitan area of Porto Alegre (94.9%) (Table 1).

Of the 335 patients, 106 (31.6%; CI95%: 26.7–36.9) sought emergency care due to DRPs. A total of 123 DRPs were observed, representing 1.2 DRPs per patient, since 17 (16%) patients presented two DRPs. Univariate analysis showed that age, educational level and the amount of medication were significantly associated with the development of DRPs (Table 1). Multiple logistic regression was performed to evaluate the impact of age, educational level, and the number of drugs on the presence of DRPs. Age showed no correlation with the development of DRPs (OR=1.14; CI95%: 0.65–2.11). A high school educational level was a protective factor against the development of DRPs (OR=0.32; CI95%: 0.16–0.64). The use of five or more drugs contributed independently to the occurrence of DRPs (OR=2.21; CI95%: 1.3–3.9) (Table 2).

The most frequent DRPs were ineffectiveness and unsafe use, respectively (Table 3). Ineffectiveness was related mainly to inadequate dosage regimens (22 cases; 17.9%) and refractoriness (15 cases; 12.2%). Adverse drug reactions were the principal cause of safety problems (35 cases; 28.5%).

In 21 DRPs (17.1%) the patient was identified as the main causative factor. The most common reasons were lack of adherence to treatment and self-medication. The drug itself was linked to 36 DRPs (29.3%). The remaining 66 DRPs (53.7%) could be attributed to the health system itself or to healthcare professionals.

Of the 71 drugs related to DRPs, the most frequently involved were those acting on the cardiovascular system (24.3%) and on the nervous system (18.6%). These drug groups contain agents such as captopril (reported in 11 DRPs) and lithium carbonate (reported in 6 DRPs). Over-the-counter drugs formed a group of nine drugs (12.7%), which were reported in 16 DRPs (13%).

The reasons for non-adherence to the treatment prescribed were cost or probably a drug supply problem in the public system during the study period, which lowered affordability difficult and affected the treatment of these patients.20,21

Self-medication was identified in a few cases and involved over-the-counter drugs. Responsible practice of self-medication should be emphasized, both by the user and by the professional recommending the medication.

The high frequency of effectiveness-related problems (54.5%) agrees with that reported by Baena et al2 (62.7%). Most of these DRPs involved drugs of continuous use that require monitoring by the healthcare team. In our system, continued and multiprofessional care is not always provided due to deficiencies in the public health service. Consequently, the patient may resort to seeking emergency care to solve any problems that appear during the treatment.

Refractoriness to the treatment was considered as a possible cause of unsuccessful chronic therapy in 15 DRPs (12.2%). The other most common cause of effectiveness-related DRP was an inadequate dosage regimen (17.9%). This result may be related to the frequency of appointments with a physician, resulting in an inadequate dosage regimen due to the absence of follow-up. In patients requiring the addition of more drugs to monotherapy, which usually occurs during the treatment of hypertension, closer monitoring by health professionals could reduce these DRPs. These patients may not have had access to the required number of consultations with a physician, or may not have kept scheduled consultations, and could have maintained the same therapeutical scheme for longer than recommended. Physicians see a large number of patients, hampering assessment of signs and symptoms that optimize the therapeutic plan. Furthermore, there is a lack of continuing education programs for prescribers and other health professionals.

The possibility of non-adherence to treatment should not be excluded, as this is one of the factors most closely correlated to ineffectiveness.5 According to O’Connor,22 treatment failure is due to medical inertia in about 75% of cases and to patients’ resistance to adhering to the therapeutic plan in 25%. In our study, some of the patients did not take the medication as recommended, suggesting that instruction on correct drug use requires more attention from health professionals, as simple oral instruction is often insufficient. In some cases, patients forgot to take some drug dose(s). Factors such as the patient's age and the number of drugs prescribed should also be considered.

Ineffectiveness was attributed to bacterial resistance in five DRPs. Stricter monitoring by health professionals and more adequate patient counseling could have prevented these patients from needing to seek ER care.

The remaining DRPs indicated prescription problems, such as dosages below the recommended level and lack of adequate assessment of drug interactions, which could have been avoided.

Many of the patients (57.6%) came from the Hospital de Clínicas de Porto Alegre outpatient clinic and, due to the high frequency of effectiveness-related DRPs, the use of the ER is probably inappropriate, i.e., non-urgent cases or cases that could be managed in the primary or secondary levels of care are treated by this service. As other authors report, this result causes concern, as people that use the ER inappropriately may be losing quality care. In addition, service to genuinely urgent patients is delayed, with overcrowding of the ER and other services that are normally used, such as laboratories, radiology departments, etc.23,24

Thirty-five cases (97%) of safety-related DRP were suspected of being ADR and, in one patient, two DRPs were suspected of being ADR, totaling 36 suspected cases. According to Naranjo's algorithm,13 20 DRPs suspected of being ADR were classified as probable and 16 as possible. No DRPs suspected of being ADR were classified as definite or doubtful. Considering only the probable suspected case, we found a 6% frequency of seeking ER care due to ADRs, which is similar to the frequency range (1%-5%) reported by other authors.25,26 The clinical manifestations observed ranged from mild to severe.

Importantly, the ER medical team recorded the suspected ADRs in only 30.5% of the cases. The pharmacovigilance system in Brazil is relatively new and hence few health professionals perceive drugs as a potential cause of health problems.27 Additionally, there is a lack of educational measures to improve the safety of drug therapy, which may be related to the high rate of ADRs.

In 21 DRPs (17.1%), the patient was identified as the main causative factor. The most common reasons were lack of adherence to treatment due to neglect of one or more doses of medicines for chronic use or the decision not to continue treatment and self-medication.

The drug itself was linked to 36 DRPs (29.3%). Access to drug information from the healthcare team could help to reduce predictable ADR. Thus, pharmacovigilance programs, structuring of drug information centers and involvement of professionals specialized in drug use in the healthcare teams should be encouraged.

The remaining 66 DRPs (53.7%) can be attributed to the health system itself or to healthcare professionals, specifically to the prescriber, since in Brazil the participation of other professionals is still low. We chose to group these cases into a single category. The study design did not allow data on previous visits to the ER to be collected for a more detailed analysis. In many cases, medical inertia was evident, but whether this was intrinsic or due to the organization of the health system is unclear.

Chronic patients, who take drugs every day, are expected to show a higher number of DRPs than patients who take drugs sporadically. The frequent association of drugs from these classes with DRPs has already been reported.28

Over-the-counter drugs should present no risk of use, but their possible use without proper instruction might have resulted in a health problem that required the ER service. This possibility may be related to the failure of the pharmacists to take responsibility for rational self-medication in Brazil.

One of the main limitations of this study is that it is based on reported data, leading to considerable uncertainty on the possible causes of the DRPs identified. This uncertainty was reduced by collecting information such as diagnosis and outcome of medical examinations from medical records.

Another issue is that multiple causes can be involved in DRPs and each of the factors listed as important may be underestimated, since only the most likely cause for each DRP was given. To estimate the responsibility of professionals, patients or the organization of the health system and to evaluate the contribution of each factor to the development of a specific DRP, longitudinal studies are required, as well as adjustments to the original questionnaire. Despite these limitations, we believe that the data may be useful for designing strategies to reduce the percentage of DRPs found.