Quality of emergency room care for atrial fibrillation

Abstract

Objectives. Atrial fibrillation (AF) is the most common arrhythmia in emergency rooms (ER). We surveyed the clinical characteristics and quality of care of AF patients in three emergency rooms in Helsinki, Finland. Design. Observational data of the treatment of 179 consecutive symptomatic AF patients were prospectively collected. The quality of care was analysed according to a predestined set of criteria. Results. Mean age of the patients was 63 years and 61% were men. The leading symptom was palpitation (86%). Sinus rhythm was achieved in 70%. New anticoagulation was initiated in 20% and cardiovascular medication modified in 42% of patients. Considering the overall quality of care, including documentation in the patient chart, it was classified as good in 53% of all patients, whereas the quality of therapeutic decisions and planning for follow-up was good in 77%. Conclusions. The ER visit results in extensive treatment modifications in two of three patients. Although inadequate care is rare, maintaining good quality requires adherence to clinical guidelines, careful documentation and plans for follow-up.

• Atrial fibrillation
• emergency care
• quality of care

Atrial fibrillation (AF) is one of the most common single diagnoses in the emergency room (ER) environment 1–5. Most patients in the ER seem to have acute AF with palpitation or chest pain as the leading symptoms 6, 7. Most studies concentrating on AF patients in the ER have evaluated the use of anticoagulation. In general, adherence to warfarin therapy is low in spite of available clinical guidelines 8–12. Cardioversion (CV) is performed in up to 62% of the patients with acute AF in the ER 2. In a previous study 13 we found that, in the case of persistent AF, the documented median time from diagnosis in the ER to elective cardioversion (CV) after anticoagulation was as long as 78 days. We also found that the longer delay from the diagnosis of AF to CV was associated with poorer restoration and maintenance of sinus rhythm 13.

Compared with the burden of AF in acute care, published data on the role of the ER in the treatment of these patients is limited. Besides the anticoagulation therapy, we don't know much of the procedures and quality of care provided to these patients in the ER. Therefore we conducted a survey on the clinical characteristics and quality of care of symptomatic AF patients in three ER's covering the population base of 560 000 inhabitants in Helsinki, Finland. Special attention was paid to presenting symptoms and background morbidity, ER procedures and planning of the patients’ care after the index visit.

Methods

The study design was prospective and observational. Data was collected from February 2003 through September 2003 as part of the AUTOMAATTI research project of the Helsinki University Central Hospital. All patients with ECG verified symptomatic AF who gave informed consent were included. Of the 179 patients in the final study group, 103 were recruited in primary hospitals (50 in Maria Hospital and 53 in Malmi Hospital) and 76 patients in the Helsinki University Central Hospital. In all ER's 24-hour anaesthesiology service was available, allowing electrical CV. A written guideline of referral to each of the hospitals was in action during the study, allowing free access to the primary hospitals to all patients and recommending referral of unstable patients and patients with recent invasive procedures to the University Hospital. Each patient was included only once in the study cohort. Patients were treated according to routine local protocols.

Four independent evaluators (HK-V, ML, JR and RP) collected data from the medical records and evaluated the acute treatment of AF in the ER's using a structured form. Background information included patient demography, previous medical history, particularly past and current medication including anticoagulation, and characteristics of the acute medical condition.

In the analysis of the quality of care, attention was paid to the appropriateness of evaluation and treatment of AF in the ER and programming of follow-up. No national clinical practice guideline for treatment of AF in acute care setting was available at the time of the study. However, the 2001 version of the ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation 14 had been widely quoted and this document was therefore used as the golden standard of evaluation in this study. A tool to assess the overall quality of ER care for AF was constructed. This tool focused on the documentation of the patient's medical history, clinical assessment, treatment decisions and planning for follow-up. Based on all available documentation, the performance was classified as good (fulfilling the guideline), fair (minor deviation) or inadequate (posing a potential threat or disadvantage to the patient). The model for quality assessment is described in detail in Appendix A.

The details of the criteria were extensively discussed and piloted by the authors before the study. Initially, one of the four independent evaluators assessed each ER treatment episode. In case of any uncertainty, the case was judged by a panel of at least two evaluators.

Statistical methods

Microsoft Access was used for data entry and data exploration. Continuous variables are presented as mean±SD and discrete variables as numbers of patients (percentages).Two sample t-test (two-tail) was used to compare continuous variables, Fisher's exact test or Fisher-Freeman-Halton test for unordered categorical variables and Kruskal-Wallis test for ordered categorical variables. Statistical calculations were performed either with SPSS (v 11.0, SPSS Inc, Chicago, Il) or with StatExact (v 4.0, Cytel Software Corporation, Cambridge, MA). Statistical significance was set at a value of p < 0.05.

Results

Patient characteristics

Patient characteristics are summarized in Table I. Men with acute symptomatic AF were on average ten years younger than the women. Hypertension was the most common background disease in both sexes but valvular disease was more common in women (p = 0.015).

Table I.  Patient characteristics.

	Male, n = 110 n (%)	Female, n = 69 n (%)	p
Age, years, mean (range)	58.7±12.9 (21–84)	68.9±10.9 (35–86)	<0.001
AF duration < 2 days	83 (75.5)	53 (76.8)	0.860
Previous AF	87 (79.1)	51 (73.9)	0.467
Previous cardioversion	75 (68.2)	41 (59.4)	0.262
Cardiological work-up including echocardiography	45 (40.9)	16 (23.2)	0.016
Hypertension	52 (47.3)	37 (53.6)	0.445
Ischaemic heart disease	26 (23.6)	11 (15.9)	0.258
Valvular disease	10 (9.1)	16 (23.2)	0.015
Thyroid disease	1 (0.9)	1 (1.4)	1.000
Prior stroke	5 (4.5)	7 (10.1)	0.218
Diabetes	9 (8.2)	5 (7.2)	1.000
Alcohol abuse	10 (9.1)	4 (5.8)	0.571
Retired	55 (50.0)	53 (76.8)	0.001
Living alone	20 (18.2)	32 (46.4)	<0.001

Of the 179 patients, 138 (77.1%) had experienced at least one previous AF episode. Notably, only 34.1% of the patients with previous AF had undergone cardiological work-up, including echocardiography.

Of the whole study group, 35.2% had no diagnosed cardiovascular disease including hypertension, ischaemic heart disease, valvular disease, heart failure or cardiomyopathy. Thus, lone fibrillation was assumed in approximately 1/3 of all patients.

Presenting symptoms and signs

By definition, all patients had symptoms associated with AF. The main symptom was palpitation in 86.0%. Other AF related symptoms were shortness of breath (12.8%), chest pain (7.8%) and non-specific symptoms (e.g. dizziness and fatigue) in 17.9% of the patients. Women experienced chest pain more often than men (15.9% vs. 2.7%, p = 0.03). ECG evidence of myocardial ischaemia was, however, found in only nine patients (5.0%). Signs of decompensated heart failure were rare (5.0%).

The mean systolic and diastolic blood pressure in the ER was 140/90 mmHg (range 80–222 mmHg and 46–128 mmHg, respectively). Figure 1 shows the ventricular response rate and proportion of patients who received rate-slowing medication in the ER. Heart rate over 120 bpm was found in 32.8% of the patients. Only two patients presented with heart rate less than 60 bpm.

Figure 1.  Ventricular response rate at the time of admission to ER and use of rate-slowing medication given in the ER (digitalis or/and beta-adrenergic blocking agent).

Preceding medication

Preceding cardiovascular medication (including antihypertensive therapy) is presented in Table II. Before this visit to the ER, 93 (52.0%) of the patients used antithrombotic medication. Of the 138 patients who had experienced AF previously, 64 (46.4%) had no antithrombotic medication and 54 (39.1%) of them were on antiarrhythmic medication (Table II).

Table II.  Preceding cardiovascular medication before and the modifications in these drug groups during the emergency room visit (including significant dose adjustments,% of all patients).

	Before, n (%)	Started or modified during the ER visit, n (%)
Warfarin	27 (15.1)	32 (17.9)
Aspirin	61 (34.1)	1 (0.6)
Other	5 ( 2.8)	1 (0.6)
β-blocker	100 (55.9)	28 (15.6)
Digitalis	14 (7.8)	9 (5.0)
ATII/ACE	60 (33.5)	3 (1.7)
Flecainide	23 (12.8)	3 (1.7)
Sotalol	17 (9.5)	4 (2.2)
Disopyramide	6 (3.4)	0 (0.0)
Propafenone	4 (2.2)	2 (1.1)
Amiodarone	3 (1.7)	3 (1.7)
Quinidine	1 (0.6)	1 (0.6)

β-blocker = a beta-adrenergic blocking agent.

ATII/ACE = an angiotensin II receptor antagonist or an angiotensin-converting enzyme inhibitors.

Treatment in the ER

Seventy-nine patients (44.1%) received rate-slowing medication. In 30 (16.8%) cases AF rhythm was normalized spontaneously, in 13/30 of them after rate-slowing medication (Figure 2).

Figure 2.  Treatment procedures for AF and the outcomes in the emergency room.

Of the 179 patients, 109 (60.9%) underwent cardioversion. The overall success rate of electrical CV was 95.8% whereas flecainide infusion resulted in sinus rhythm in 66.7% of the attempts (Figure 2). Altogether 125 patients (69.8%) achieved sinus rhythm during the ER visit. No complications of electrical or pharmacological CV were observed.

Cardiovascular, antithrombotic or antiarrhythmic medication was modified in 83 (46.4%) of the patients (Table II). The most common changes were initiation of new anticoagulation and initiation or dose adjustment of a beta-receptor blocking agent. Of the 32 patients with new warfarin treatment, 24 (75.0%) were discharged home from the ER. On the other hand, 48.7% of patients with their first documented AF (n = 41) received new anticoagulation and CV was scheduled to be performed later due to duration of AF more than 24–48 hours. At the time of leaving the ER, 60.3% of all the patients had an antithrombotic; 34.1% had anticoagulant and 26.2% aspirin.

Discharge

Of all the patients, 147 (82.1%) were discharged home from the ER and 94 (63.9%) of them had underwent cardioversion. Ischemia or cardiac decompensation were the most common reasons for hospital ward admission.

Quality of AF care in the ER

Figure 3 shows the quality of care assessed in five dimensions: classification of AF (a), recording medical history (b), clinical assessment (c), therapeutic decisions (d) and plans for follow-up (e). Criteria and representative examples of the categories “good”, “fair” and “inadequate” in each of the dimensions are given in the Appendix.

Figure 3.  Assessment of the quality of AF care in the ER. For detailed criteria, see Appendix. a) classification of AF b) recording medical history c) clinical assessment d) therapeutic decisions e) plans for follow-up

Taken together, the overall (from classification to planning for follow-up) quality of care was good in 94/179 patients (52.5%). If only therapeutic decisions and planning for follow-up were considered, the corresponding proportion was 138/179 (77.1%). In most cases, less than good performance was observed in only one (26.8%) or two (14.5%) dimensions. The distribution of good performance was not affected by sex, age (under or over 75 years) or whether this was the patient's first or recurrent AF. However, decisions to refrain from CV in the ER were judged to be more appropriate than decisions to perform one (p = 0.042; Table III).

Table III.  Quality of care with (n = 109) or without (n = 70) cardioversion

	Inadequate n (%)	Fair n (%)	Good n (%)	p
Classification of AF
CV	0 (0.0)	12 (11.0)	97 (89.0)	0.167
No CV	2 (2.9)	1 (1.4)	67 (95.7)
Recording medical history
CV	4 (3.7)	20 (18.3)	85 (78.0)	0.672
No CV	0 (0.0)	14 (20.0)	56 (80.0)
Clinical assessment
CV	1 (0.9)	16 (14.7)	92 (84.4)	0.837
No CV	0 (0.0)	12 (17.1)	58 (82.9)
Therapeutic decisions
CV	3 (2.8)	13 (11.9)	93 ( 85.3)	0.042
No CV	1 (0.9)	2 (2.9)	67 (95.7)
Plans for follow-up
CV	3 (2.8)	21 (19.3)	84 (77.1)	0.468
No CV	3 (4.2)	9 (12.9)	58 ( 82.9)

Discussion

In this study we outline the present role of the ER in the care of symptomatic AF patients. Treatment of symptomatic AF in the ER was found to be active and decision-intensive. Moreover, we present a novel method to evaluate the quality of care provided to AF patients in the ER setting. Using this method, the quality of therapeutic decisions and planning for follow-up was qualified as good in 3/4 in patients.

Previously, anticoagulation therapy has been widely used as a golden standard or primary quality indicator when the treatment of AF has been evaluated. The underutilization of anticoagulation is well documented in AF patients 15 although it has improved since the early 1990's 16, 17.In our study 60.3% of the patients were discharged on either anticoagulant or aspirin. However, the number of patients treated with anticoagulants is only one side of appropriate care for AF and the fact that the proportion of INR tests within the therapeutic range can be as low as 50% 9, 15 must be taken into account.

The primary observation in our study was that current treatment provided for AF in the ER is active and plays a crucial role in the management of patients with symptomatic first-time or recurrent AF. The mainstream therapy is rhythm control: altogether 70% of all patients achieved sinus rhythm in the ER (Figure 2). The success rate of CV in our patients was comparable to previous studies 2, 6 and the treatment was found to be safe. However, our findings are different from those observed in the GEFAUR-1 study 18 in which AF was accepted as such in 60% of all patients and rhythm control was attempted in only 42% of eligible patients. Different patient profiles probably explain these findings: our study included only patients who presented to the ER because of symptomatic AF whereas del Arco and co-workers included a wide selection of ER patients with ECG demonstration of AF.

In addition to active CV policy, cardiovascular, antithrombotic or antiarrhythmic medication was modified in almost half of the patients. Notably, two thirds of patients with their first episode of AF had delays more than 24–48 hours and were therefore put on new anticoagulation and CV was scheduled at a later time point. Such a crucial role of ER decision-making in managing AF patients sets high requirements to the documentation, communication and organization of care after the ER visit. Patient education is another major challenge in the busy ER environment. We did not assess the patients’ experience of care, nor did we study the impact of organizational and educational factors on outcome. These issues as well as the impact of access to transesophageal echocardiography in the ER should be addressed in further studies in order to assess the quality of AF care in a comprehensive manner.

We were not able to find any published tool to assess the overall quality of care provided to AF patients in an acute setting. Developing such a tool for acute cardiac patients can be an exhausting task 19, particularly if the relative impact of each quality dimension on outcome is to be determined 20. Quality indicators should be in line with evidence-based clinical guidelines as adherence to such guidelines will by definition lead to better outcome 21. We developed a five-dimension three-grade quality assessment tool for acute AF care on the basis of the extensively disseminated ACC/AHA/ESC guidelines 14. This study demonstrates the feasibility of such an instrument in the assessment of routine care for unselected symptomatic AF patients. The assessment tool is relatively simple and suitable for routine quality surveillance. Further studies are currently underway to validate it against patient experience and various measures of treatment outcome.

Using this tool the overall quality of care was found to be acceptable, with shares of inadequate care ranging from 1 to 3% in the five dimensions (Figure 3). An interesting and somewhat surprising difference was found between the quality of care in patients with vs. without attempted cardioversion (Table III). In most cases this difference was due to insufficient plans for future medication and follow-up in patients with restored sinus rhythm. This observation further emphasizes the need to establish seamless care pathways and common information platforms between the ER and the primary care facilities.

In conclusion, we found that current treatment of symptomatic AF in the ER is active and decision-intensive. In two thirds of patients the ER visit lead to significant treatment modifications, in spite of the fact that most patients had already been treated for AF. Treatment in the ER represents a window of opportunity to improve the treatment of AF in the community. Although inadequate care is rare, maintaining good quality requires adherence to clinical guidelines, careful documentation and plans for follow-up after the ER visit.

This study was supported by the EVO funds of the Helsinki University Central Hospital. Ms Liisa Käppi, RN, and the personnel of the participating emergency rooms are acknowledged for skilful help.

Appendix: Quality of AF care in the ER

The quality of care was assessed in five dimensions: classification of AF, recording medical history, clinical assessment, therapeutic decisions and plans for follow-up.

Table

Good:	guideline requirements fulfilled
Fair:	minor deficiency or non-justified deviation
Inadequate:	posing a potential disadvantage or threat to the patient
Typical examples: Classification of AF (paroxysmal, persistent, permanent)
Inadequate:	AF not listed as diagnosis
	Wrong diagnosis
Fair:	AF listed as diagnosis but missing or wrong classification
Good	Appropriate classification
Recording medical history
Inadequate:	Background morbidity, medication, possible previous AF or duration of current AF not listed
Fair:	One of the above missing
Good:	Sufficient information available
Clinical assessment
Inadequate:	Cardiovascular assessment (vital functions, decompensation, ischemia) missing or erroneous
Fair:	High initial blood pressure or ventricular response rate not appropriately recognized
Good:	Careful clinical examination with appropriate interpretation
Therapeutic decisions
Inadequate:	Inappropriate polypharmacy, anticoagulation against guidelines, discharge of unstable patient
Fair:	Minor deviation from guidelines in a typical case of AF
	Insufficient care of myocardial ischemia or heart failure
Good:	Treatment in line with guidelines
Plans for follow-up
Inadequate:	Need of anticoagulation not recognized, insufficient follow-up of new medication, responsibility for further treatment poorly defined
Fair:	Deviation from guidelines without explicit reason, e.g. cardiac work-up not scheduled, need to modify drug dosing not mentioned
Good:	Appropriate plan distinctly communicated