310 Journal of Pain and Symptom Management Vol. 33 No.

3 March 2007

Original Article

A Brief Cognitive-Behavioral Intervention

Reduces Hospital Admissions in Refractory
Angina Patients
Roger K.G. Moore, MD, David G. Groves, PhD, John D. Bridson, BA (Hons), MSc,
Antony D. Grayson, BSc, Helen Wong, PhD, Austin Leach, MB BS, FFARCS, Robert
J.P. Lewin, MA (Hons), MPhil, and Michael R. Chester, MD, MRCP, FESC
Cardiothoracic Center (R.K.G.M., A.D.G.), Liverpool; National Refractory Angina Center (D.G.G.,
J.D.B., H.W., A.L., M.R.C.), Royal Liverpool and Broadgreen University Hospitals NHS Trust,
Liverpool; and Department of Health Sciences (R.J.P.L.), University of York, York, United Kingdom

Abstract
Chronic refractory angina is an increasingly prevalent, complex chronic pain condition, which
results in frequent hospitalization for chest pain. We have previously shown that a novel
outpatient cognitive-behavioral chronic disease management program (CB-CDMP) improves
angina status and quality of life in such patients. In the present study of 271 chronic refractory
angina patients enrolled in our CB-CDMP, total hospital admissions were reduced from 2.40
admissions per patient per year to 1.78 admissions per patient per year (P < 0.001). The rising
trend of total hospital bed day occupancy prior to enrollment fell from 15.48 days per patient per
year to a stable10.34 days per patient per year (P < 0.001). There were 32 recorded myocardial
infarctions prior to enrollment compared to eight in the year following enrollment (14% vs.
2.3%, P < 0.001) and overall mortality was lower that comparable groups treated with
surgery. This study shows that educating patients and demystifying angina using a brief
outpatient CB-CDMP produces an immediate and sustained reduction in hospital admission
costs that represents a major potential health care saving. This benefit accrues in addition to the
known effects of CB-CDMP on symptoms and quality of life. These data suggest that a
CB-CDMP approach to symptom palliation represents a low cost alternative to palliative
revascularization. J Pain Symptom Manage 2007;33:310e316. Ó 2007 U.S. Cancer
Pain Relief Committee. Published by Elsevier Inc. All rights reserved.

Key Words
Chronic refractory angina, refractory angina, stable angina, angina pectoris, myocardial
infarction, patient education, rehabilitation, hospital admissions, mortality, cognitive-
behavioral therapy, cost saving, quality of life, health economics

Introduction
Address reprint requests to: Michael R. Chester, MD, Chronic refractory angina (chronic stable
The UK NHS National Refractory Angina Center, angina where revascularization is no longer
Royal Liverpool & Broadgreen University Hospitals considered feasible) is an increasingly preva-
NHS Trust, Liverpool L14 3PE, United Kingdom.
E-mail: [email protected] lent, complex chronic pain condition, which
results in frequent hospitalization.1 Repeated
Accepted for publication: October 20, 2006. hospitalization reduces sufferers’ quality of

Ó 2007 U.S. Cancer Pain Relief Committee 0885-3924/07/$esee front matter

Published by Elsevier Inc. All rights reserved. doi:10.1016/j.jpainsymman.2006.10.009
Vol. 33 No. 3 March 2007 CBT Reduces Hospitalization and Myocardial Infarction in Angina 311

life and frustrates physicians facing patients management programs to reduce emergency
with complex problems, which are notoriously hospital bed days for people with long-term
difficult to alleviate. The American College of conditions by 5% by 2008.7 This pressure to
Cardiology (ACC)2 and the European Society reduce avoidable hospitalization reflects a
of Cardiology (ESC)3,4 emphasize the need general trend across major health economies.
for patient education and rehabilitation tai-
lored to individual circumstances at the outset
of angina management. However, the ACC has
observed that emphasis on technological inter- Methods
ventions in the management of chronic angina Most patients referred to NRAC are severely
has led many to overlook ‘‘critically important incapacitated by angina and experience
aspects of high quality care. Chief among these intense anxiety and apprehension, believing
neglected areas is the education of patients.’’2 that unsuitability for revascularization implies
A novel cognitive-behavioral chronic disease a bleak prognosis. Many of these experienced
management program (CB-CDMP), which re- patients have significant counter-therapeutic
flects ACC and ESC principles, was instituted misconceptions, which have been demon-
in Liverpool in the United Kingdom in 1997 strated to generate erroneous ideas about car-
for the treatment of chronic refractory angina diac risk factor reduction.8 Consequently, at
(CRA).5 It has previously been demonstrated the outset of the CB-CDMP, a bio-psychosocial
that this approach has a significant impact on diagnosis was made, which defined both the
symptoms, improving frequency and stability pathological cause of symptoms and the psy-
of angina and quality of life, and reducing anx- chological consequences for the patient and
iety and depression in CRA sufferers.6 It also immediate carers. We used a simple five-item
emerged from our experience in delivering questionnaire as a clinical tool to elicit
the program that an important additional ben- patients’ views on regular exercise, whether
efit to patients accrued from prioritizing symp- angina damages the heart, whether they
tom relief and improving functional and thought they ought to avoid provoking epi-
psychological status. Many patients reported sodes of angina, and if avoidance of angina
a reduction in the number of times they were would reduce their risk of MI.5 This facilitated
admitted to hospital with chest pain and exploration of patients’ cardiac misconcep-
suspected myocardial infarction (MI). They tions during the course of a two-hour system-
attributed this to the improvement in their atic interview with a consultant cardiologist
symptoms and psychological status, resulting and a consultant in pain medicine. We identi-
from a better understanding of their condi- fied specific individual misconceptions and an-
tion. We wanted to know if these individual pa- ginal threat avoidance behaviors by posing
tient experiences reflected a measurable effect open-ended questions about the patients’ and
across the population. carers’ understanding of the condition, the
We present an audit of the effect of the CB- meaning of symptoms and what constituted ap-
CDMP on hospitalization, on fatal and nonfa- propriate behaviors. During interviews,
tal MI, and on all-cause mortality in a cohort patients commonly stated that they thought
of consecutive CRA patients referred to the that coronary narrowings progressed gradually
UK National Refractory Angina Center and that progression to complete blockage and
(NRAC) and treated according to modern sudden death was inevitable. Most patients
multidisciplinary pain management principles. thought that the pain was due to back pressure
Data on hospital admissions and bed use are on the heart as it struggled to force blood past
symptom-specific, enabling us to distinguish coronary artery narrowings and believed that
emergency admissions with chest pain from their cardiac function was far worse than the
admissions for other reasons. These data are objective evidence suggested. It became obvi-
also objectively verifiable measures that were ous that although most patients recognized
of obvious relevance to health service man- that, in principle, exercise was a good thing,
agers planning services. For example, the they themselves were sufficiently intimidated
UK Government Department of Health has by their misconceptions about angina to avoid
announced that it expects chronic disease exercise for fear of damaging their hearts.
312 Moore et al. Vol. 33 No. 3 March 2007

Having identified misconceptions and associ- were not included in the analysis, as it was
ated maladaptive behaviors (i.e., those which in- not possible to have their admission history
crease rather than reduce the risk of MI), we independently verified. Formal approval was
challenged the patients’ beliefs and, using their obtained from each hospital’s ‘‘Caldicott
previous experiences, offered evidence-based Guardian’’ (the person in UK hospitals with
alternative explanations for their symptoms. A overall responsibility for confidentiality of
realistic, patient-defined objective was agreed.9 patient-identifiable data) to extract admissions
This was supported with written and audiotaped data on all patients referred locally between
educational material and homework based on 1/1/97 and 1/10/02. This time period ensured
the York Angina Plan.10 All patients received extraction of at least one year’s post-
stress management advice and relaxation train- enrollment follow-up data. Presenting com-
ing tapes and manuals, and all agreed to under- plaints for each admission were extracted
take a modest, symptom-limited, graduated using the hospital’s International Statistical
exercise program at home, at a level appropri- Classification of Diseases and Related Health
ate to achieving their personal objectives. New Problems (ICD 10) codes.
patients were followed up within eight weeks
at a second, hour-long combined pain and car- Statistical Analysis
diology consultation. Core beliefs were revisited Predefined periods of one year were used to
and any remaining cardiac misconceptions calculate hospital admission, bed day, and MI
were addressed and progress toward the pa- rates before and after enrollment in the CB-
tient’s objective was discussed. Individuals CDMP. The data were analyzed using nonpara-
requiring further pain or psychotherapeutic metric tests of significance. Death rates were
intervention were then offered treatment described using the product limit methodol-
according to the CRA guideline.5 ogy of Kaplan and Meier. Cox regression was
used to determine which parameters were signif-
Diagnosis icantly related to mortality during the follow-up
All patients were referred to NRAC following period. All analyses were performed using SAS
angiogram review by a consultant cardiologist for Windows version 8.2 and SPSS version 11.
and, where appropriate, a consultant cardiac
surgeon, who had agreed there was no revascu-
larizable disease or that the risks of intervention
were unjustifiable. A diagnosis of CRA was con- Results
firmed only if both consultants concurred at Four hundred thirty outpatients and three
the initial combined cardiology and pain clinic inpatients were referred between 1/1/97 and
interview. Patients whose symptoms did not ap- 1/10/02, of whom 383 were diagnosed with
pear to be of cardiac origin were not included CRA. The three inpatients were enrolled into
in the audit. the program directly during admission to
NRAC’s parent hospital, the Cardiothoracic
Mortality Center (CTC). These admissions were ex-
Patients were matched by their National cluded from the data, as they would have
Health Service number, name, date of birth, artificially increased the total number of ad-
and postal code, and vital status was confirmed missions in the month prior to enrollment at
by the National Strategic Tracing Service NRAC. Further, to ensure entirely conservative
(NSTS), which records all-cause mortality in pre-enrollment figures, all admissions to the
the UK. This enabled us to examine mortality CTC during the month prior to enrollment
in the whole cohort. at NRAC were excluded from the data. Demo-
graphic details are presented in Table 1 and
Admissions treatments received in the audit period are
NRAC receives local and national referrals. presented in Table 2.
The admissions data relate only to referrals
from 11 local hospitals where we were able to Mortality
obtain reliable admissions data. Patients re- Of the 383 CRA patients, 14 died within
ferred from outside the local referral area the one-year follow-up period (the majority
Vol. 33 No. 3 March 2007 CBT Reduces Hospitalization and Myocardial Infarction in Angina 313

Table 1 Of the baseline parameters, only impaired

Baseline Characteristics of the Study Group left ventricular function predicted all-cause
Percentage of Study Group mortality (univariate Cox regression P ¼ 0.001).
Entire CRA Admissions The Kaplan-Meier survival curves for CRA pa-
cohort cohort cohort tients categorized by left ventricular function
Characteristics (n ¼ 433) (n ¼ 383) (n ¼ 271)
are presented in Fig. 2.
Age (years) at recruitment
<46 5 3
46e60 36 33 32 Myocardial Infarctions
61e75 52 55 58
>75 7 8 7
In the year prior to enrollment in the CB-
CDMP, there were 32 recorded MIs compared
Gender
Male:female 76:24 80:20 79:21 to eight in the year following enrollment
Diagnosis
(14.7% vs. 2.3%, P < 0.001).
CRA 90 100 100
Syndrome X 9 d d
Other 1 d d Admissions
Previous MI 60 69 70 Forty-three of the 383 CRA patients seen dur-
Previous CABG 60 66 66 ing the study period were referred from out of
Previous PTCA 30 34 36
Left ventricular 31 39 39
area and, therefore, were unable to have their
ejection fraction <40% admission history independently verified and
Ex- or current smoker 67 81 83 so were not included in the analysis. Of the lo-
Hypertension 40 44 44
Elevated cholesterol 78 84 86
cally referred patients, 69 were excluded as
Diabetic 18 20 19 they had less than one year of follow-up data.
Two or more antiangina 88 89 92 This gave an admissions audit cohort of 271
medications
CRA patients.
CRA ¼ chronic refractory angina; MI ¼ myocardial infarction; The impact of enrollment into the program
CABG ¼ coronary artery bypass grafting; PTCA ¼ percutaneous
transluminal coronary angioplasty. on hospital admissions and bed day occupancy,
comparing one year prior to enrollment with
one year post-enrollment, is given in Table 3.
having moderate or severe left ventricular Following enrollment, total hospital admis-
impairment). The average one year, all-cause sions fell from 2.40 admissions per patient
mortality was 3.8% (95% confidence interval per year to 1.78 admissions per patient per
2.5%e5.1%). The cumulative survival of these year (P < 0.001). Total hospital bed day occu-
patients in comparison with survival seen in pancy fell from 15.48 days per patient per
a similar population of patients randomized year to 10.34 days per patient per year
to palliative surgery or spinal cord stimula- (P < 0.001). This improvement was mainly
tion11 and survival observed in all patients un- the consequence of significantly reduced ad-
dergoing revascularization in the North West missions and bed day occupancy for patients
region of the UK is shown in Fig. 1. suffering cardiac chest pain or MI. Total

Table 2
Pain Therapies Received By the Patient Cohorts in the First Year of Treatment
Percentage of Each Cohort Group Receiving Specific Pain Interventions
Algorithm Entire cohort (n ¼ 433) CRA cohort (n ¼ 383) Admissions cohort (n ¼ 271)

Outpatient counseling, cognitive 100 100 100

behavior therapy, and cardiac
rehabilitation
Transcutaneous nerve stimulation 51 50 49
Temporary sympathectomy 36 35 41
Oral opioids 6 5 5
Spinal cord stimulation 2 2 1
Epidural opioids 1 1 1
Intrathecal opioid pump 1 1 0
Enhanced external counterpulsation 1 0 0
CRA ¼ chronic refractory angina.
314 Moore et al. Vol. 33 No. 3 March 2007

£728,520, saving £362,440. This 33% reduction

in admissions represents a reduction in hospi-
talization costs of £1337 per patient per year.

Discussion
Our clinical experience shows that serious
misconceptions and maladaptive behaviors
are very common in ‘‘veteran’’ end-stage re-
fractory angina patients. A recent unpublished
audit of the angina beliefs of 197 consecutive
CRA patients completing an angina beliefs
Fig. 1. Kaplan-Meier survival curves for CRA questionnaire immediately prior to enrollment
patients (NRAC, ESBY) compared to patients foll- in our CB-CDMP revealed that 30% mistakenly
owing revascularization (NWQIP survival data thought that angina was a kind of small heart
obtained from www.nwheartaudit.nhs.uk, accessed attack, 53% erroneously believed that each ep-
October 18, 2004).
isode of angina does more damage to the
heart, and 82% thought that it was very impor-
noncardiac admissions did not change signifi- tant to avoid anything that might bring on an
cantly following enrollment (107 vs. 121, episode of angina. It is known that people
P ¼ 0.380). with misconceptions about angina are more
To exclude the possibility that the reduction anxious and more physically limited than
in admissions might be explained by regression those with fewer such misconceptions.8 This
to the mean, we recalculated admission rates study adds to previously published evidence
in monthly blocks for the study period (Fig. 3). that addressing these misconceptions in a CB-
CDMP improves symptoms and quality of life
for CRA sufferers.6 It demonstrates that rising
Reduction in Admission Costs rates of emergency hospital admissions can be
Calculations assumed a daily hospital bed
halted and reversed with simple educational in-
cost per patient of £260, estimated by the
terventions aimed at raising morale and break-
CTC Hospital Finance Department for the
ing the cycle of increasing inactivity.
relevant period. The total amount spent for
The decline in admissions was observed
12 months before NRAC enrollment (4,196
immediately following the first outpatient
admissions) was £1,090,960, and for the 12
appointment (Fig. 3), during which patients
months after (2,802 admissions) was
received education about angina and anxiety
was reduced by addressing patients’ miscon-
ceptions. This improvement was maintained
over the following year, during which approxi-
mately half of the patients received additional
pain interventions, of which 98% were nonin-
vasive or minimally invasive therapies (Table 2).
The reduction in hospitalization cannot rea-
sonably be explained by regression to the
mean, but could potentially be the result of
simple optimization of antianginal medica-
tion.12 However, all patients’ antiangina medi-
cation had been scrutinized by a consultant
cardiologist prior to referral and changes in
prescription at enrollment were minimal.
Even though cardiac risk factors were strin-
Fig. 2. Kaplan-Meier survival curves for NRAC gently controlled on entry to the program,
patients categorized by left ventricular function. the dramatic reduction in hospital admissions
Vol. 33 No. 3 March 2007 CBT Reduces Hospitalization and Myocardial Infarction in Angina 315

Table 3
Hospital Admissions (n ¼ 271)
1-Year 1-Year Relative P-value
Pre-NRAC Post-NRAC Reduction (%) (Wilcoxon)a

Total number of admissions 651 483 26 <0.001

Total days in hospital 4,196 2,802 33 <0.001
Number of admissions with chest pain 414 297 32 <0.001
Days in hospital with cardiac chest pain 2,806 1,739 38 <0.001
Total number of cardiac admissions 526 340 35 <0.001
Days in hospital with cardiac admissions 3,501 2,120 39 <0.001
Total number of MI admissions 31 8 74 0.002
Days in hospital with MI 282 79 72 0.007
Total number of noncardiac admissions 107 121 þ13 0.380
Days in hospital with noncardiac admissions 601 590 2 0.311
NRAC ¼ national refractory angina center.
a
Two-tailed analysis.

for MI was unexpected. It could be proposed for most patients and our CB-CDMP appears
that patients remaining at home rather than to offer an alternative for patients who are un-
attending hospital despite significant angina willing to accept the risks of revascularization.
symptoms could account for the reduction in
MI admission rates seen in this study. Although Limitations of Study
patients enrolled in the program were in- This was an observational study with no con-
structed in strategies to control the severity of trol group to compare admissions, mortality,
angina episodes, they were also encouraged and MI outcomes. However, designing an ethi-
to attend hospital if they suffered an exacerba- cal randomized trial presents its own difficulties
tion of their cardiac chest pain lasting longer when the primary intervention (patient educa-
than 20 minutes, in accordance with accepted tion and individually tailored rehabilitation)
guidelines. The death rate in this study was has already been shown to be effective and is rec-
consistent with the 2%e7% annual mortality ommended in nationally16 and internationa-
quoted in published CRA cohorts.11,13,14 lly2e4 accepted guidelines. Given that the audit
The prevailing view among cardiac special- looked at hospital admissions one year before
ists is that holistic pain management tech- and one year after enrollment, and the NRAC
niques for chronic angina should be CB-CDMP was the only significant new interven-
regarded as a last resort, only to be deployed tion for patients during the study period, pa-
once coronary artery bypass grafting (CABG) tients effectively acted as their own controls.
and angioplasty are no longer feasible.15 How- The definition of the primary reason for ad-
ever, redo-CABG and angioplasty are palliative mission was based on hospital coding rather

Fig. 3. Monthly total hospital admissions (admissions/month).

316 Moore et al. Vol. 33 No. 3 March 2007

than review of the clinical data. This may pro- before and one year after enrolment onto a refrac-
vide for a higher rate of error, but given the tory angina program. Eur J Pain 2005;9:305e310.
large size of the data set, should not influence 7. Department of Health. The National Service
the overall trends reported. framework for long-term conditions, 2005. Available
from. www.dh.gov.uk. Accessed October 18, 2006.
Conclusion 8. Furze G, Lewin R, Murberg TA, Bull P,
Thompson D. Does it matter what patients think?
The overall annual cost of hospitalization The relationship between changes in patients’ be-
for patients with angina in the UK was esti- liefs about angina and their psychological and func-
mated at £208 million in 2000.17 This study tional status. J Psychosom Res 2005;59:323e329.
demonstrates that, even in a cohort of patients 9. Bridson J, Hammond C, Leach A, Chester M.
with complex, chronic angina, implementa- Making consent patient centred. BMJ 2003;
tion of a CB-CDMP emphasizing rehabilitation 327(7424):1159e1161.
significantly reduces hospitalization and MI 10. Lewin RJP, Furze G, Robinson J, et al. A rando-
rates without increasing mortality. The effects mised controlled trial of a self-management plan
are immediate and sustained, suggesting such for patients with newly diagnosed angina. Br J
Gen Pract 2002;52:194e201.
programs may at least pay for themselves,
and that a more detailed health economic 11. Mannheimer C, Eliasson T, Augustinsson LE,
analysis is warranted. Use of such programs et al. Electrical stimulation versus coronary artery
bypass surgery in severe angina pectoris: the ESBY
has consistently been shown to produce im- study. Circulation 1998;97:1157e1163.
provements in health care use, disability, and
12. Spertus JA, Dewhurst TA, Dougherty CM, et al.
health-related quality of life similar to those
Benefits of an ‘angina clinic’ for patients with
we have observed.8,18e20 The principles of coronary disease: a demonstration of health status
such programs are easily implemented without measures as markers of health care quality. Am
the need for specialist psychological services. Heart J 2002;143:145e150.
Indeed, simple interventions, such as the 13. Sanderson JE, Ibrahim B, Waterhouse D, et al.
York Angina Plan,10 could be implemented Spinal electrical stimulation for intractable angina-
quickly by nurses and doctors in primary dlong-term clinical outcome and safety. Eur Heart
care, and by cardiologists and pain teams in J 1994;15:810e814.
secondary care, with specialist centers such as 14. Ten Vaarwerk IA, Jessurun GAJ, DeJongste JL,
NRAC taking on the more complex patients. et al. Clinical outcome of patients treated with spi-
nal cord stimulation for therapeutically refractory
angina pectoris. Heart 1999;82:82e88.
References 15. De Jongste MJL, Tio RA, Foreman RD. Chronic
therapeutically refractory angina pectoris. Heart
1. Chester MR. Chronic refractory angina: time to 2004;90:225e230.
sort out a neglected and growing problem. Br J
Cardiol 2000;7(3):108e111. 16. Department of Health. National Service frame-
work for coronary heart disease, 2000. Available
2. ACC/AHA 2002 guideline update for the man-
from. www.dh.gov.uk/publications. Accessed Octo-
agement of patients with chronic stable angina.
ber 18, 2006.
Available from. www.americanheart.org. Accessed
October 18, 2006. 17. Stewart S, Murphy N, Walker A, et al. The cur-
3. The European Society of Cardiology. Guidelines rent cost of angina pectoris to the National Health
on the management of stable angina pectoris: exec- Service in the UK. Heart 2003;89:848e853.
utive summary. Eur Heart J 2006;27:1341e1387. 18. Frizelle DJ, Lewin RJP, Kaye G, et al. Cognitive--
Available from. www.escardio.org. Accessed October behavioural rehabilitation programme for patients
18, 2006. with an implanted cardioverter defibrillator: a pilot
4. Mannheimer C, Camici P, Chester MR, et al. study. Br J Health Psychol 2004;9:381e392.
The problem of chronic refractory angina. Report 19. Lewin B, Cay E, Todd I, et al. The angina man-
from the ESC Joint Study Group on the Treatment agement programme: a rehabilitation treatment. Br
of Refractory Angina. Eur Heart J 2002;23(5): J Cardiol 1995;2(8):221e226.
355e370.
20. Lewin B, Robertson I, Cay EL, Irving J,
5. Available from. www.angina.org. Accessed Campbell MA. Effects of self-help post myocardial
October 18, 2006. infarction rehabilitation on psychological adjust-
6. Moore RK, Groves D, Bateson S, et al. Health-re- ment and use of health services. Lancet 1992;339:
lated quality of life of patients with refractory angina 1036e1040.