The results of our study indicate that both models of care provided high quality anticoagulation management. The TTR was over 60% for both PC and UC. However, our results demonstrated that patients in the PC group spent more time in both the TTR and the expanded TTR compared to the UC group and these differences were statistically significant. The percentage of time patients' INR was < 1.5 was lower in the PC group versus the UC group and the percentage of time patients' INR was > 5 was higher in the PC group versus UC group. Low numbers of adverse clinical events were detected. Our current study's findings demonstrate that patients experience better anticoagulation control for the TTR and expanded TTR when managed by a pharmacist with expertise in anticoagulation management who applies a systematic, evidenced based approach to patient care.
Several other studies have compared pharmacist managed anticoagulation services to usual care, most often by a physician. Several of these studies support the current study findings while others indicate no difference [9–15]. The studies that support our findings include two randomized controlled trials and three observational studies [9–13]. In one Canadian study, researchers conducted a randomized controlled trial where patients were allocated to either anticoagulation clinics with a pharmacist in three tertiary hospitals (n = 112) or to their family physician practices (n = 109) . Patients managed by the anticoagulation clinics were within the expanded therapeutic range more than patients managed by family physicians (82% vs 76%, p < 0.05). High risk INR values (<1.5 or >5.0) were more often observed in patients managed by family physicians (49% vs 39%, p < 0.05). In another randomized control trial conducted in Hong Kong, patients were randomized to either a pharmacist managed anticoagulation clinic (n = 68) or physician managed service (n = 69) . Patients in the pharmacist managed group were within the TTR more than the physician managed group (64% vs 59%, p < 0.001). Of the three observational studies, one was conducted in Canada and two in the United States [11–13]. In one Canadian prospective cohort study, consecutive patients (n = 125) referred to the pharmacist Anticoagulation Management Service (AMS) with at least 4 months anticoagulation management prior to referral were included in a pre- and post-analysis of anticoagulation control . The anticoagulation control was greater in the AMS compared with standard care in the period before referral (66.5% vs 48.8%, p < 0.0001). Both observational studies conducted in the United States found pharmacist managed anticoagulation clinics significantly improved patients anticoagulation control as measured by the TTR [12, 13].
Two studies found no difference in anticoagulation control between a pharmacist managed anticoagulation clinic and usual care by a physician [14, 15]. In the randomized controlled trial conducted in Quebec, patients stabilized in a pharmacist managed anticoagulation service (PMAS) in a large community hospital were subsequently randomly allocated to either continue in the PMAS (n = 128) or be transferred to their physician for follow-up care (n = 122) . Researchers reported no difference in patients managed by the PMAS vs physicians in either the TTR (77.3% vs 76.7%) or expanded TTR (93.0% vs 91.6%). This finding could be explained by a potential selection bias. Patients spent an average of 11.3 weeks in the PMAS and were eligible for group allocation only if they were stable in their anticoagulation control. In addition, physicians were given the option of agreeing to participation for each patient individually or for all their patients at once. Close to 50% of the participating physicians chose patients on an individual basis to participate in the study. Consequently this study population may not have included patients with poorer anticoagulation control that would likely benefit from a pharmacist managed program. In the observational cohort study, patients in a family medicine clinic were cared for either through a pharmacist managed anticoagulation clinic or in a traditional care model by physicians . Overall, the authors found no difference in anticoagulation control between the two groups. However, their method of data analysis and reporting results were not standard and make comparisons with other studies difficult.
A prospective randomized control study is generally considered the gold standard for evaluating therapeutic interventions. However it may not be the optimal method to assess this type of intervention as it is difficult to minimize observation biases. As all groups are aware of the intervention, their behaviours may be influenced accordingly in a prospective randomized trial. For example, physicians who are aware they are participating in a clinical trial may be more vigilant about monitoring anticoagulation care than in their routine practice. A well designed retrospective study may be the next best method. In our study, all eligible patients were included in the data collection, and the two groups were comparable, reducing any selection biases. This study evaluated the anticoagulation management practices in one family medicine clinic. Our patient population was similar to that in other studies evaluating anticoagulation management; consequently, our findings may be generalizable to other similar populations.
Due to the nature of the retrospective cohort study design, a number of confounders were not controlled for. The amount of time spent with patients in each model was not recorded, neither was the format or content of education provided. We did observe a subgroup of patients who crossed over from UC to PC. Although in subgroup analysis we still found a significant difference in TTR, we cannot necessarily assume this effect was due to the pharmacists' anticoagulation management. We found significant difference, but more research is needed in this area. Future research would allow for control of confounders such as time, education, protocol driven care, multivariable regression through prospective randomized controlled or prospective cohort design.