Keywords
INTRODUCTION
Atrial fibrillation (AF) is the most frequent arrhythmia in clinical practice and its prevalence is expected to increase due to the increased survival in the general population and of patients with comorbidities that predispose to the development of AF.1,2 The wide range of clinical manifestations and seriousness of presentation1,3 means patients with AF are attended in many different healthcare settings by a variety of specialists.
Atrial fibrillation is a potentially serious condition given that it doubles mortality and has a high level of morbidity.1,3-7 However, AF-related complications can largely be prevented or efficiently controlled by early, adequate treatment of arrhythmia.7-9 Similarly, the earlier therapy is begun, the more likely it is that sinus rhythm can be restored and maintained.7,9 Consequently, adequate treatment in each healthcare setting is crucial to the subsequent optimal evolution of the patient.
However, treatment of AF and prophylaxis for complications often vary according to the specialist responsible for each patient despite the availability of clinical guidelines and consensus documents.10-13 Furthermore, these are applied differently in different settings. In a recent study, we determined the inadequacy of some aspects of AF treatment in half the patients with this arrhythmia and the reluctance of specialists to rectify inadequacies in most cases.14 These findings have been corroborated elsewhere.15-19
The objective of the present study is to evaluate the impact of implementing a common AF treatment protocol in different settings within a healthcare area, on improving treatment and increasing specialists' commitment to rectifying preexisting inadequacies.
METHODS
Settings
The study was conducted in an urban, tertiary hospital and a primary care clinic (PCC) in the catchment area. The contexts in which patients were included were the following: a) family physician clinics in the PCC; b) PCC and hospital outpatient cardiology clinics; c) hospital emergency room (ER); and d) conventional hospital wards.
Patients attending ER and subsequently admitted were included in ER setting when evaluating before the visit, and in the hospital setting at discharge, as final treatment depends on the physicians on the wards.
Different healthcare settings are independent and autonomous in their treatment of these patients and face no restrictions in the application of treatments recommended in the clinical guidelines below (electric cardioversion, antiarrhythmic drugs, anticoagulants, antiplatelet treatment).
Periods
The study was conducted in 2 periods, as follows:
Period 1 or preintervention. We included all adult patients (>18 years) with antecedents of AF demonstrated by electrocardiogram attending any of the healthcare settings mentioned during 14 consecutive days in June 2004 whether their reason for attending was directly AF-related or not. We excluded patients first diagnosed with AF at this consultation (first episodes), as they did not receive treatment prior to AF and we could not therefore evaluate treatment adequacy before the visit or the change in treatment if it were inadequate.
Period 2 or postintervention. We included all adult patients (>18 years) with antecedents of AF demonstrated by electrocardiogram attending clinic in any of the healthcare settings mentioned during 14 consecutive days in June 2005 (the same period as in 2004), whether the reason for consultation was AF-related or not. We again excluded first episodes.
In the 2 periods, all patients were treated according to the criteria of the physician responsible and followed the care routines established at the time of the study. The same specialists were responsible for patients in the 2 periods, although there were changes among the interns who collaborated in attending patients.
Intervention
Between the 2 periods, we designed a protocol for action in the face of AF agreed by physicians from all the healthcare settings involved through the participation of 2 or 3 representatives from each. The protocol covered both treatment of different types of AF and procedures for referral between fields. To create this, the physicians responsible based themselves on the current clinical guidelines specified below10-13 and adapted them to the realities of the healthcare area.
The protocol was publicized in 2 ways: a) a pocket-sized brochure summarizing the protocol was distributed to all physicians working in the healthcare settings involved, both interns (n=81) and specialists (n=53), and b) in each setting, clinical sessions were led by 2 of the physicians responsible for preparing the protocol. In the sessions, they explained AF treatment, data obtained on adequacy in the first period, the reason for preparing the agreed protocol and described the protocol. Although sessions were aimed at the same physicians who had received the summary, only 70% attended.
Variables Studied
In the 2 periods, we conducted a personal interview with each patient and/or their close family, and reviewed clinical records.
We obtained the following data on each patient:
Age
Gender
Social-family support (good, partial, none)
Quality of life/degree of dependence (Barthel index, Annex 1)
Arterial embolism risk factors (number and type)
AF classification (paroxystic, persistent, permanent)
Reason for visit (see text)
Antiarrhythmic treatment. Type of treatment and adequacy according to current clinical guidelines (see text)
Prophylactic treatment for arterial embolism. Type of treatment and adequacy according to current clinical guidelines (see text)
Reasons for attending clinic were classified in 2 sub-groups: a) directly AF-related, and b) non AF-related.
Treatment was considered adequate when it corresponded to that in clinical guideline recommendations current at the time of the study, even though they might differ on certain points. In the first period, we accepted: "Guidelines of the Spanish Society of Cardiology on Cardiac Arrhythmias," "Guidelines for the treatment of atrial fibrillation in hospital emergency services," "ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation."10-12 In the second period, we also accepted the more recent "7th Consensus of the American College of Chest Physicians."13 This was not available for inclusion in the first period evaluation but could be included in the second period and in the preparation of the protocol.
Evaluation of adequacy of treatment was conducted before and after the study visit. We considered treatment inadequate when it was not specified in guidelines for the use it was being given, when it was not the first choice treatment and there was no justifiable reason for not using the first choice treatment, or when there was a contraindication to its use. However, we did consider adequate, prophylactic treatment for arterial embolism, in patients theoretically indicated for anticoagulation, according to guidelines, but not receiving it because of a medical or social contraindication detailed in their clinical records.
In the first and second periods, the physicians responsible for the patients were only informed about the existence of an epidemiologic study which some of their patients had been included in. They were asked to give permission for inclusion, but, to avoid bias, were not informed in advance about the exact nature of data gathered or about the period when analysis would be conducted.
Data on the variables studied were gathered by interns (n=8) and of specialists in family and community medicine (n=2), internal medicine (n=3, including 2 sub-specialists in ER), and cardiology (n=1). All data was recorded by interns under the supervision of the specialists, who then judged the adequacy or inadequacy of treatment based on the specified clinical guidelines. The same specialists were involved in data collection in the 2 periods.
Statistical Analysis
Values are expressed as mean (SD) for continuous variables and absolute values or percentages for discontinuous variables. For comparisons we used the Student's t test and one-way ANOVA with Bonferroni correction for normally distributed continuous variables and χ2 and the Fisher test for qualitative variables. A P value less than .05 was considered significant. Calculations were with software SPSS 12.0.
RESULTS
We included 293 patients in the first period and 267 in the second. Table 1 shows the settings patient came from and percentages of patients with AF versus total number of visits in each setting, as well as clinical-epidemiologic characteristics of patients included. In the first period, 2 patients admitted to ER or in-hospital died: 1 of an AF-related cause (antiarrhythmic treatment). In the second period, 5 patients died: 2 of AF-related causes (1 antiarrhythmic treatment, 1 anticoagulation treatment) (P=.27).
Antiarrhythmic Treatment
In the first period, antiarrhythmic AF treatment varied greatly in all settings (Table 2). In 71% of patients, antiarrhythmic treatment was adequate according to clinical guideline recommendations current at the time of the study. We found no differences in drug regimens between patients attending in the different settings studied. However, significant differences existed in adequacy of treatment, with greater adequacy in patients attending ER and lower adequacy in patients attending family physician clinics and in-hospital (P=.007).
In 31 patients, antiarrhythmic treatment drugs were modified at the study visit. Treatment was most frequently changed following medical attention in ER (P<.001 vs other settings). The principal reasons for attending clinic that led to modification of antiarrhythmic treatment were directly AF-related: complications of the illness or treatment and symptoms caused by AF.
After the visit, overall adequacy of antiarrhythmic treatment rose from 71% to 73% (nonsignificant). The increased percentage of adequate antiarrythmic treatments was not significant in any setting. We continued to record significantly greater inadequacy of antiarrhythmic treatment in primary care (Table 3).
The principal reasons for inadequate antiarrhythmic treatment before and after the visit (Tables 2 and 3, respectively) were: a) digoxin prescribed as the only drug and as first choice to control heart rate in patients with moderate or high physical activity (guideline Class IIb recommendation only in patients in repose); b) digoxin in patients with paroxysmal AF (Class III recommendation both to control heart rate, with Level of Evidence B, and for cardioversion or prophylaxis of new episodes, with Level of Evidence A); c) amiodarone for chronic control of heart rate in patients with permanent AF (this does not appear as a guideline recommendation) and d) other, includeding:
Amiodarone as first choice therapeutic for chronic treatment in patients without structural cardiomyopathy with paroxysmal AF and very isolated episodes (<1/year), clinically well-tolerated (antiarrhythmic treatment in these cases is a Class IIb recommendation with level of evidence C, and amiodarone is the second choice drug)
ß-blockers in patients with chronic obstructive pulmonary disease and bronchial hyper-reactivity (Class III recommendation, level of evidence C)
Combined calcium antagonists and ß-blockers in patients with history of bradycardia
After implementing the protocol (period 2), we also found substantial clinical variety in antiarrhythmic drug regimens before the visit although differences arose when compared with the settings of inclusion (Table 4). Overall adequacy increased significantly with respect to the previous year (80% vs 71%; P=.009). This was reflected in all healthcare settings although it was not significant in individual cases (Figure 1A). Family physician clinics continued to record a lower level of adequacy.
Figure 1. Comparison of adequacy of antiarrhythmic (A) and antithrombotic treatments (B) before visit between preintervention and postintervention periods. FP indicates family physician; ER, emergency room.
In the second period, antiarrhythmic treatment was modified in 35 patients. Treatment was most frequently modified in ER and in-hospital (P<.001 vs other settings), and reasons for attending clinic leading to changes were the same as in the first period. The percentage of corrective interventions to remedy previous inadequacy was similar to the previous period, both overall and in individual settings (Figure 2A). Reasons for inadequacy were also similar, both before and after the visit, although the frequency of each changed (Tables 4 and 5, respectively).
Figure 2. Comparison of percentage of correction in case of inadequacy of antiarrhythmic (A) and antithrombotic (B) treatments between preintervention and postintervention periods. FP indicates family physician; ER, emergency room.
Antithrombotic Prophylaxis
In the first period, 53% of patients received anticoagulation treatment before medical attention, and 30% received antiplatelet treatment (4% received both). Overall adequacy of arterial embolism prophylaxis was 67% before the study visit and we found no significant differences between settings (Table 2). After the visit, overall adequacy was 70% (nonsignificant increase) with no significant differences between settings.
The principal reasons for inadequacy of antithrombotic prophylaxis before and after the visit (Tables 2 and 3, respectively), were: a) no anticoagulation in patients with AF and high risk of arterial embolism without contraindication (non-fulfillment of Class I recommendation and Level of Evidence A); b) no anticoagulation or antiplatelet treatment in patients in who, due to lower risk of arterial embolism, the 2 choices of treatment are accepted (non fulfillment of Class I recommendation and Level of Evidence A); and c) anticoagulation in patients without medical indication (Class III recommendation and Level of Evidence C) or with medical or social contraindications.
Prophylaxis was less adequate in patients with paroxysmal AF both before the visit (58% vs 74% in permanent AF and 85% in persistent AF; P=.006) and after the visit (61% vs 78% in permanent AF and 85% in persistent AF; P=.005). Adequacy was greater in patients with a previous neurologic complication, both before the visit (83% vs 64%; P=.01) and after the visit (87 vs 67%; P<.007), although we found no relationship between adequacy of treatment and other risk factors such as advanced age (anticoagulated patients were younger, although the difference was nonsignificant), diabetes, high blood pressure or presence of heart failure. Adequacy at discharge was lower in family physician clinics than other settings (P=.05).
Overall adequacy of arterial embolism prophylaxis before the visit was greater in the second period than the first (81% vs 67%; P<.001). This was reflected in all settings although it was only significant in ER (Figure 1B). The percentage of corrective interventions in cases of previous inadequacy was also greater than in the first period (Figure 2B), both overall (35% vs 9%; P<.001) and in ER and outpatient cardiology. We also recorded a nonsignificant increase in family physician clinics.
The reasons for inadequacy were similar to those in the first period, both before and after the visit, although the frequency of each changed (Tables 4 and 5, respectively). In the second period, we found no relation between adequacy of arterial embolism prophylaxis and AF type, or with any of the risk factors for arterial embolism, either before or after the visit.
DISCUSSION
This study shows the value of training and of unifying criteria for AF treatment. Between the 2 periods in the study, no new ideas or technological approaches were developed. We simply publicized and reinforced the application of known concepts in our context.
The benefit of implementing the protocol was two-fold. In the postintervention period, initial adequacy was notably greater in all healthcare settings, reflecting the use of the consensus protocol and the training given during the year between the 2 periods analyzed. More importantly, in the second period we also found a significant increase in corrective interventions in cases previously found to be inadequate. This reflects physicians' commitment to improving the situation of patients receiving inadequate treatment.
Although the results of implementing a combined treatment protocol in different healthcare settings may seem obvious, little previous experience exists of the value of similar training programs. However, results reported elsewhere are also positive20,21: Zimetbaum et al20 showed the usefulness of providing information about clinical guidelines for AF in ER on improvements in cost-effectiveness, although they did not determine intrinsic adequacy. In Spain, previous experience of programs to improve AF treatment has also been reported by Ruiz et al,21 who achieved 90% anticoagulation in patients without contraindications. However, these researchers implemented their own protocol and no data were given about anticoagulation prior to the intervention. The present study includes scientific tests of prior status of treatment and treatments were prescribed by a number of physicians other than the researchers and those responsible for the protocol. Application of the protocol to a non-selected sample of physicians objectively increases the value of the level of acceptance.
The commitment of physicians to remedying preexisting inadequacies was evident in arterial embolism prophylaxis, but this was not the case with antiarrhythmic treatment, despite an overall improvement in the initial adequacy of antiarrhythmic treatment in the second period. The explanation may lie in the multifactorial nature of the issue. To date we have found no other study that determines the level of physicians' commitment to improving inadequacies existing before consultation. Most studies of AF treatment adequacy are descriptive and only refer to anticoagulation.15-19,22,23 Adequacy of antiarrhythmic treatment is barely studied24 and there may be less awareness of incorrect use. Alternatively, in cases when the principal reason for a change of treatment was a complication in treatment or illness, fear or reluctance to start a new treatment without a specific reason to do so, may be the explanations. Finally, the great improvement found in initial adequacy in the second period may have been influenced by less obvious corrective interventions.
However, despite increased adequacy in primary care before the visit, in comparison with the previous year, and at discharge, in comparison with before the visit, patients in primary care continued to be the least likely to receive anticoagulation treatment or to have it changed because of its inadequacy. Family physicians are, or should be, the doctors who best know the personal, social, and family situation of patients in order to determine whether to start and/or maintain anticoagulation treatment and to control adequate follow-up and possible harmful effects or events that might justify interrupting it. The present study found increased, but nonsignificant, adequacy of treatment in primary care as did Martín et al,22 who reported 81% adequacy in antithrombotic prophylaxis in a primary care clinic. These studies demonstrate anticoagulation treatment can be started and controlled from primary care although figures for adequacy of treatment in most studies conducted in this and other settings do not usually surpass 60%.15-19,24 In primary care, many patients with AF attending family physician clinics are controlled by cardiologists. This may explain the lack of commitment and reluctance to change treatment on the part of family physicians, patients, or patients' families. Whatever the case may be, increased training sessions in primary care seem indicated.
To date, numerous multicenter studies have been conducted to evaluate the adequacy of coordinated treatment of AF in different centers within the same setting, whether in primary care,9,10 ER,12-14 or in-hospital,15-17 but multidisciplinary analysis within a single healthcare area, as in the present study, has not been conducted. This enables us to make reliable comparisons and even design future strategies. Moreover, given that the structure of healthcare areas and settings studied is similar to other parts of Spain, the intervention we conducted could be applied elsewhere. However, the peculiarities of each area, especially the intensity and type of relationship between different settings and their capacity and autonomy for therapeutic action, could substantially influence results. In our case, improved results in ER, in-hospital and outpatient cardiology can probably be explained by closer relationships between these settings, favored by physical proximity and a certain flow of physicians from one setting to another. Future improvements in communication with primary care will depend on each hospital.
Limitations of the Study
One limitation of the present study is the inclusion of the "7th Consensus of the American College of Chest Physicians" in the second period as an accepted clinical guideline. While inclusion in this period was obligatory, as it was already current, it could have increased the range of patients considered as receiving adequate treatment. In any case, all patients believed to need antithrombotic prophylaxis as defined by the consensus document were also in need of prophylaxis according to at least one of the guideline documents accepted in the first period.
CONCLUSIONS
The present study shows that the consensus of physicians involved in attending patients with AF, following the implementation of a specific protocol derived from this consensus and adapted to all the settings, can significantly improve antiarrhythmic treatment in daily clinical practice. However, greater emphasis should be placed on specific issues, especially the commitment of all physicians to improving treatment.
ABBREVIATIONS
AF: atrial fibrillation
Financed by a grant from the 3M Foundation, awarded by the Sociedad Española de Cardiología (2004).
Correspondence: Dra B. Coll-Vinent.
Servicio de Urgencias. Hospital Clínic.
Villarroel, 170. 08036 Barcelona. España.
E-mail: bcvinent@clinic.ub.es
Received June 23, 2006.
Accepted for publication January 25, 2007.