Symptom patterns, duration and responses in newly diagnosed patients with heart failure

7
RESEARCH PAPER Symptom patterns, duration and responses in newly diagnosed patients with heart failureRobyn Gallagher RN PhD Associate Professor, Chronic and Complex Care, Faculty of Nursing, Midwifery and Health, University of Technology, Sydney, New South Wales, Australia Anne Sullivan RN Clinical Nurse Specialist, MACARF Program, Royal North Shore Hospital, Sydney, New South Wales, Australia Susan Hales RN Clinical Nurse Specialist, MACARF Program, Ryde Hospital, Sydney, New South Wales, Australia Geraldine Gillies RN Clinical Nurse Specialist, MACARF Program, Hornsby Hospital, Sydney, New South Wales, Australia Rhonda Burke RN Clinical Nurse Specialist, MACARF Program, Manly Hospital, Sydney, New South Wales, Australia Geoffrey Tofler MBBS Professor Director, MACARF Program, Royal North Shore Hospital, Sydney, New South Wales, Australia Accepted for publication June 2011 Gallagher R, Sullivan A, Hales S, Gillies G, Burke R, Tofler G. International Journal of Nursing Practice 2012; 18: 133–139 Symptom patterns, duration and responses in newly diagnosed patients with heart failure This study investigated the symptom patterns and duration and associated predictors occurring prior to first heart failure (HF) admission. Data from the Managing Cardiac Function (MACARF) program from January to December 2007 were reviewed in relation to preadmission symptoms and contacts with health professionals. Patients (n = 242) were aged 78.7 years (SD 12 years), male (54%) and married (45%). Patients experienced up to seven symptoms (Mean 2.7, SD 1.4) for a median of 4.47 days (range 1–7) before admission, most commonly increased dyspnoea on exertion (88%), and for the shortest duration chest discomfort. Less than half (48%) contacted a health professional before hospitalization, most often a general practitioner (37%). The duration patients experienced before presenting to hospital was increased if they presented during business hours (beta = 2.68) or the evening (beta = 1.88) (and therefore less from midnight to 8:30 am), or with a change in symptom (beta = 1.4), whereas duration was reduced by chest discomfort (beta =-2.01) and older age (beta =-0.07). There is a significant time window during which patients and health professionals may detect and act on worsening HF symptoms. Newly diagnosed patients with HF need support to recognize and respond to these symptoms to avoid hospital admission. Correspondence: Robyn Gallagher, Chronic and Complex Care, Faculty of Nursing, Midwifery and Health, University of Technology, PO Box 123, Broadway, Sydney, NSW 2007, Australia. Email: [email protected] International Journal of Nursing Practice 2012; 18: 133–139 doi:10.1111/j.1440-172X.2012.02010.x © 2012 Blackwell Publishing Asia Pty Ltd

Transcript of Symptom patterns, duration and responses in newly diagnosed patients with heart failure

Page 1: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

R E S E A R C H P A P E R

Symptom patterns, duration and responses innewly diagnosed patients with heart failureijn_2010 133..139

Robyn Gallagher RN PhDAssociate Professor, Chronic and Complex Care, Faculty of Nursing, Midwifery and Health, University of Technology, Sydney,

New South Wales, Australia

Anne Sullivan RNClinical Nurse Specialist, MACARF Program, Royal North Shore Hospital, Sydney, New South Wales, Australia

Susan Hales RNClinical Nurse Specialist, MACARF Program, Ryde Hospital, Sydney, New South Wales, Australia

Geraldine Gillies RNClinical Nurse Specialist, MACARF Program, Hornsby Hospital, Sydney, New South Wales, Australia

Rhonda Burke RNClinical Nurse Specialist, MACARF Program, Manly Hospital, Sydney, New South Wales, Australia

Geoffrey Tofler MBBSProfessor Director, MACARF Program, Royal North Shore Hospital, Sydney, New South Wales, Australia

Accepted for publication June 2011

Gallagher R, Sullivan A, Hales S, Gillies G, Burke R, Tofler G. International Journal of Nursing Practice2012; 18: 133–139

Symptom patterns, duration and responses in newly diagnosed patients with heart failure

This study investigated the symptom patterns and duration and associated predictors occurring prior to first heart failure(HF) admission. Data from the Managing Cardiac Function (MACARF) program from January to December 2007 werereviewed in relation to preadmission symptoms and contacts with health professionals. Patients (n = 242) were aged 78.7years (SD 12 years), male (54%) and married (45%). Patients experienced up to seven symptoms (Mean 2.7, SD 1.4) fora median of 4.47 days (range 1–7) before admission, most commonly increased dyspnoea on exertion (88%), and for theshortest duration chest discomfort. Less than half (48%) contacted a health professional before hospitalization, most oftena general practitioner (37%). The duration patients experienced before presenting to hospital was increased if theypresented during business hours (beta = 2.68) or the evening (beta = 1.88) (and therefore less from midnight to 8:30am), or with a change in symptom (beta = 1.4), whereas duration was reduced by chest discomfort (beta = -2.01) andolder age (beta = -0.07). There is a significant time window during which patients and health professionals may detect andact on worsening HF symptoms. Newly diagnosed patients with HF need support to recognize and respond to thesesymptoms to avoid hospital admission.

Correspondence: Robyn Gallagher, Chronic and Complex Care, Faculty of Nursing, Midwifery and Health, University of Technology, PO Box 123,Broadway, Sydney, NSW 2007, Australia. Email: [email protected]

International Journal of Nursing Practice 2012; 18: 133–139

doi:10.1111/j.1440-172X.2012.02010.x © 2012 Blackwell Publishing Asia Pty Ltd

Page 2: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

Key words: delay, health professional contact, heart failure, hospital admission, symptoms.

INTRODUCTIONHeart failure (HF) is a global health issue, affecting anestimated 5.7 million people worldwide.1 People with HFexperience major reductions in their health-relatedquality of life2 and frequent hospital admissions.3 Hospitaladmissions have serious consequences for people with HFincluding the potential for further disease progression. Upto 50% of hospital admissions for HF may be preventableif worsening symptoms are detected promptly becausetimely and appropriate interventions by health profession-als can occur.4,5 However, failure to monitor HF symp-toms and delay in seeking help for worsening symptoms iscommon.6–10 Patients are reported to delay an average ofalmost 3 days with key HF symptoms before presenting tohospital.11 The extent of this delay varies according towhether the study examined acute symptoms, in whichcase delay ranges from 13 h through to 4 days,12–15 orsubacute or non-specific symptoms in which case delayranges from 7 to almost 10 days.12,15

Delay is also influenced and potentially exacerbated bythe diversity of HF symptoms. Patients with HF experi-ence a range of symptoms that indicate deterioration oftheir condition, with the most common symptomsreported prior to hospitalization including dyspnoea,oedema, fatigue and angina, and less often dyspnoea onexertion, orthopnoea, paroxysmal nocturnal dyspnea,cough, nausea, palpitations and weight gain.11,13,15,16

However, the meaning of these symptoms can be difficultfor patients to interpret, having many potential causes.15

For instance, patients are reported to delay longer inpresenting to hospital if they had dyspnoea or oedema, butdelay less if they have chest pain.11 Furthermore, experi-ence with HF, such as a previous admission, has beennoted to contribute to a reduction in delay.11–13,15

HF symptoms also present in different patterns. Some-times symptoms develop insidiously, accumulating innumber over many days, or in other cases, presentingas a cluster of symptoms which develop almostsimultaneously.13,17–19 These different types and patternsof symptoms appear to influence patients’ treatment-seeking responses. However, the effects reported areinconsistent at best with multiple symptoms,12

simultaneous onset pattern13 and nonsimultaneous (i.e.gradual) symptom onset pattern15 all reported to increasedelay.

Symptoms are subjective experiences and patientsthemselves are in the best position both to detect theappearance of HF symptoms as well as the trajectory ofchange in symptoms against the background of any exist-ing subacute symptom variability.20

This means that patients with HF must learn the symp-toms that indicate decompensation of their HF and rou-tinely monitor, detect and respond to changes in thesesymptoms to avoid admission.21 For instance, patientsmay respond by contacting their primary health physicianrather than presenting to the hospital Emergency Depart-ment (ED). This area has had little investigation, althoughone study reported that contacting a physician contributedto delay in presenting to hospital.11 Whether this is anappropriate delay resulting from patients managing theirsymptoms well is unclear; particularly as patients maychoose to delay seeking treatment until they can see theirphysician during business hours.12 Nonetheless, the poten-tial consequence of this delay is that the physician’s onlyappropriate option is to then send the patient to the ED.22

It is therefore vital that patients learn skills associatedwith detecting changes in HF symptoms and appropriatetreatment-seeking responses as early in the course of HFas possible and preferably from the time of diagnosis.21,23

An ideal opportunity to promote these activities occursfollowing the patients’ first presentation to hospital forHF, when the patient’s recent symptom experiences anddecision-making responses can be used as a platform fordiscussion of self-care by HF nurses and primary carephysicians. This study set out to investigate in relation tofirst hospital admission of patients with HF:

• The symptom patterns, duration of symptoms and pre-dictors of duration of symptoms occurring prior tohospital admission,

• The patterns of presentation to ED prior to hospitaladmission, including whether patients had contacted ahealth professional, and the actions that health profes-sionals assumed.

METHODSSetting and sample

The patients in this study were registered with the HFsupport program (Managing Cardiac Function program(MACARF) ) in one area health service in Sydney, Aus-tralia. To be eligible for this program, patients must have

134 R Gallagher et al.

© 2012 Blackwell Publishing Asia Pty Ltd

Page 3: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

a diagnosis of HF, be community-dwelling, free frommajor cognitive or psychiatric disorders and not requiringpalliative or end-of-life care. For the current study, aninitial review of 1300 hospital admission records from theMACARF database for the period January to December2007 was conducted. Records were excluded when theadmission was not through ED or were not for a primarydiagnosis of HF, which means that cardiac revasculariza-tion procedures or myocardial infarction were alsoexcluded. Of the original 434 admission records that wereeligible after this screening, 242 individuals were retainedfor data analysis on the basis that this was their first admis-sion for HF.

The post hoc power calculation determined that atalpha level of 0.05, with 12 predictor variables, an r2 of0.19 and a sample size of 242, that the power was 0.95(http://www.danielsoper.com/statcalc/calc09.aspx).24

Additional ethics approval was not required as patientsconsented to the use of their data for quality assuranceactivities and data were de-identified and routinelygathered by the MACARF nurses as part of usual care.25

Data collection and instrumentsData were collected by the MACARF nurses from thepatients during their hospital stay on sociodemographicfactors including age, gender, marital status, whether theperson lived alone or with other people, clinical charac-teristics including whether the HF diagnosis was new andfunctional status. Functional status was determined usingassessment of the New York Heart Association (NYHA)functional class, which ranges from I to IV, with higherclassification indicating lower functional status. Thesedata plus specific data on admission date and time werecollected or clarified from the patient’s medical record.

Data on symptom type and duration before presentingto ED were collected directly from the patient using a tooldeveloped by the MACARF team. Patient report wasconsidered more useful than medical records as symptomsand duration of symptoms have been reported to varyaccording to different types of health professionals.15 Thetool guides a series of structured questions related tothe HF symptoms patients had experienced preceding thecurrent admission. Ten symptoms had been selected forthe list on the basis of previous studies11,13,15 and clinicalexpertise. Specifically, patients were asked whether theyhad experienced increased dyspnoea on exertion, cough,awakening at night short of breath, feeling tired or lackingenergy, needing to take more frequent rest during the

day, weight gain, swelling of ankles, legs or stomach, lossof appetite or nausea, increased palpitations or chest dis-comfort. When patients indicated that they had experi-enced a particular symptom, they were then asked toidentify how long they had experienced this symptombefore presenting to the ED. Patients were then askedwhether they had contacted a health professional regard-ing the change in symptom, the type of health professionalcontacted and the response or action the health profes-sional took prior to their presentation to the ED.

AnalysisData were analysed using the IBM Statistical Package forthe Social Sciences (SPSS 19 for Windows, Chicago, IL,USA). Two variables were created from the symptomdata based on the work of Friedman13 and Evangelista.11

The variable duration of acute symptoms was created toidentify the length of time patients had whatever acuteHF symptom or symptoms that most directly pre-ceded admission. A second variable, simultaneous onsetsymptom pattern, was also created to distinguish thosepatients whose onset of symptoms occurred over time vs.those patients whose symptoms occurred all at once. Fre-quencies, percentages, medians, means and standarddeviations were used to describe the sample sociodemo-graphic and clinical characteristics, symptoms, symptomduration and patterns.

The independent predictors of symptom duration priorto ED presentation were determined by multiple linearregression analysis using the variables age, gender, maritalstatus, number of symptoms, experienced a symptomchange, presence of chest discomfort, increased dyspnoeaand oedema, simultaneous symptom onset pattern, con-tacted a health professional, and presentation time andwinter season. A backwards method was used to producethe most parsimonious model with the critical level set at0.05.26 All assumptions of regression analysis were testedincluding multicollinearity and no assumption wasviolated.27

RESULTSPatients (n = 242) were aged 78.7 years (� SD 12 years),male (54%), and married (45%) as detailed in Table 1.38% had NYHA class II function at baseline and stayed inhospital an average of 9.48 days (SD 7.9 days), followingwhich the majority were discharged to their homes(86%). Admissions occurred more commonly during the

Heart failure symptoms and hospitalization 135

© 2012 Blackwell Publishing Asia Pty Ltd

Page 4: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

winter months (44%), on a Monday (24%) and duringbusiness hours (52%) (Table 2).

Prior to hospitalization, patients experienced up toseven symptoms (Mean 2.7, SD 1.4) for a median of 4.47days (range 1–7) (Table 3). Approximately one third(33%) experienced simultaneous onset of symptoms

rather than a gradual onset. The most common symptompatients experienced was increased dyspnoea on exertion(88%), followed by swelling of feet and stomach (49%),cough (27%) and fatigue (24%) (Table 4). The symptomparticipants experienced the shortest duration beforebeing admitted to hospital was chest discomfort (median1 day), followed by increased dyspnoea on exertion(median 3 days) and waking at night short of breath(median 3.5 days).

In response to symptoms, just less than half (48%) ofthe patients contacted a health professional prior to pre-senting to the ED, who most often was a GP (37%) asdetailed in Table 4. Of the health professionals contacted,

Table 1 Sample characteristics (n = 242)

Characteristic N %†

Age, mean (SD) 78.7 (12)Male 131 (54)Married 109 (45)Live with someone 159 (66)English first language 206 (85)NYHA Class (n = 177)

I 21 (8)II 91 (38)III 59 (24)IV 6 (3)

Length of stay, mean (SD) 9.5 (7.9)Discharged home 208 (86)Died within year 13 (5)

† %Ages may not total 100 due to missing values.

SD, standard deviation; NYHA, New York Heart Association.

Table 2 Patterns of admission and presentation to ED (n = 242)

Characteristic N (%)

Most frequent months of

presentation to EDAugust 58 (24)July 29 (12)June 22 (9)

Most frequent days of

presentation to EDMonday 58 (24)Friday 37 (15)Wednesday 34 (14)

Time of presentation to ED†

Business hours (8:30–17:30) 125 (52)Evening (17:30–24:00) 54 (22)Morning (24:01–08:30) 40 (17)

† %Ages may not total 100 due to missing values.

ED, Emergency Department.

Table 3 Heart failure symptom prevalence and duration prior to

hospital admission (n = 242)

Symptom type N (%) Duration

median

Increased dyspnoea on exertion 212 (88) 3Swelling of ankles, legs or stomach 118 (49) 7Cough 65 (27) 7Feeling tired or lack of energy 58 (24) 10Waking at night short of breath 48 (20) 3.5Chest discomfort 47 (19) 1Needing more rest during the day 32 (13) 10Weight gain 30 (12) 8.5Lack of appetite or nausea 27 (11) 7Increased palpitations 16 (7) 5

Table 4 Interaction with health professionals prior to hospital

admission (n = 242)

Action N %

Contacted a health professional prior 111 48GP 90 37Specialist 11 5GP and specialist 9 4Community nurse 7 3

Health professional action 96 40Sent to ED 49 20Change CCF medication 24 10Antibiotic 15 6Investigations 8 3

ED, Emergency Department; CCF, congestive cardiac failure.

136 R Gallagher et al.

© 2012 Blackwell Publishing Asia Pty Ltd

Page 5: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

40% then undertook an action, which most often was tosend the patient to ED (20%).

The significant predictors of symptom duration prior topresentation to the ED were determined by multiplelinear regression analysis and the model developedexplained 19% of the variance (Table 5). Factors thatincreased the time a patient experienced symptoms beforepresenting to hospital included time of day with increasedtime to presentation occurring during business hours(beta = 2.68) or the evening (beta = 1.88) (as comparedwith from midnight to 8:30 am), and experiencing asymptom change (beta = 1.4). In contrast, the length oftime a patient experienced symptoms before presentingwas reduced by having chest discomfort (beta = -2.01)and being of older age (beta = -0.07).

DISCUSSIONThis study demonstrated that patients with HF experi-ence a number of diverse symptoms which tend to accu-mulate over several days before patients present to theED for treatment. The pattern and duration of symp-toms is consistent with previous reports, although thetime to presentation was slightly longer in the currentstudy,12,13,15 most likely because the new HF diagnosismeant that patients were unable to infer a meaning totheir symptoms.11–13,15 The most common HF symptomsincluded dyspnoea, oedema, cough and fatigue, and ithas been argued that patients may find it difficult to

interpret and to potentially distinguish these symptomsfrom other comorbid conditions, especially in elderlypeople like our sample.15 Furthermore, when patientshad a symptom that was readily attributable to a cardiaccause, such as chest pain, treatment was sought earlierand earlier in our study than others.11 Confusion aboutHF symptoms and the symptom trajectory, including achange in symptom as occurred in this sample, has pre-viously been identified as a contributor to delay aspatients may be uncertain if the changes in their symp-toms infer deterioration in their health.28 It is alsopossible that younger patients may have more difficultydetermining the meaning or seriousness of their symp-toms, as younger patients delayed longer.12,15

Regardless of symptom pattern and duration prior toHF admission, admission has serious consequences forthis population. Patients were hospitalized for more than9 days on average, and for these elderly patients, thislength of stay can cause serious reduction in muscle toneand function.29 Therefore avoiding hospitalization byself-management and timely treatment by health profes-sionals is vital. Our study and others demonstrated thatthere is a significant time window when these interven-tions could occur.12,13,15 However, other factors thansymptom pattern are clearly at play in the patient’s deci-sion to seek help. It appears that many patients with HFadopt a ‘wait and see’ approach to symptoms, whereasothers may only seek help when family insists.28 Thismeans that by the time patients contact a health profes-sional, symptoms have progressed to a degree thathospital admission is the only option. Although theavailability of primary care physicians may be influentialas in the early hours of the morning, the only course ofaction may be to attend the ED. Multiple other factorsmay also come into play, such as depression and cogni-tive impairment30,31 and future studies should includethese factors.

For health professionals working with patients withHF, the results of the current study suggest that newlydiagnosed/admitted patients need to be provided witheducation and support for self-management which is per-sonally relevant. One way to do this would be to discussthe symptom experience that brought them to hospitaland then to teach methods to readily detect and respondto key symptoms. The most appropriate intervention ortreatment-seeking response should then be discussed inlight of their existing interactions with family and healthprofessionals.21

Table 5 Predictors of symptom duration prior to admission for

heart failure

Predictors B† 95% CI P-value

Chest discomfort -2.01 -3.43–-0.62 0.005Simultaneous symptom onset -1.05 -2.26–-0.16 0.089Age -0.07 -0.12–0.02 0.005Presenting during business

hours (8:30–17:30)‡

2.68 1.10–4.26 < 0.001

Presenting in the evening

(17:30–24:00)‡

1.88 0.01–3.67 0.01

Experienced symptom change 1.40 0.23–2.58 0.02Present in winter months 1.01 0.08–0.12 0.08

Model statistics F test = 7.74, d.f. = 232 P < 0.001, adj

R2 = 0.19. † Unstandardized beta. ‡ Vs. 24:00–8:30 h.

CI, confidence interval.

Heart failure symptoms and hospitalization 137

© 2012 Blackwell Publishing Asia Pty Ltd

Page 6: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

LIMITATIONSThe study results cannot be generalized to all patients withHF because the sample was limited to newly diagnosedpatients with HF and excluded patients who had symp-toms but avoided hospitalization by self-managing ortreatment from health professionals. As symptom type,duration, self-management and health professional inter-action were self-reported, the results may be limited bysocial desirability and ability to accurately recall.

CONCLUSIONPatients having their first HF admission endure a constel-lation of symptoms for several days before being hospital-ized. Therefore during admission patient support needs tobe directed at symptom recognition and management, andappropriate and timely seeking of health professionalsupport. The HF nurse and primary care physician repre-sent the major candidates to provide this support forpatients with HF.

REFERENCES1 World Health Organization. Global Burden of Disease Report.

2004. Available from URL: http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004.Accessed July 2010.

2 Alehagen U, Rahmqvist M, Paulsson T, Levin LA. Quality-adjusted life year weights among elderly patients with heartfailure. European Journal of Heart Failure 2008; 10: 1033–1039.

3 Aranda JM Jr, Johnson JW, Conti JB. Current trends inheart failure readmission rates: Analysis of Medicare data.Clinical Cardiology 2009; 32: 47–52.

4 Opasich C, Rapezzi C, Lucci D et al. Precipitating factorsand decision-making processes of short-term worseningheart failure despite ‘optimal’ treatment (from the IN-CHFRegistry). The American Journal of Cardiology 2001; 88: 382–387.

5 Ditewig JB, Blok H, Havers J, van Veenendaal H. Effective-ness of self-management interventions on mortality, hospi-tal readmissions, chronic heart failure hospitalization rateand quality of life in patients with chronic heart failure: Asystematic review. Patient Education and Counseling 2010; 78:297–315.

6 Moser DK, Doering LV, Chung ML. Vulnerabilities ofpatients recovering from an exacerbation of heart failure.American Heart Journal 2005; 150: 984e7–984e13.

7 Ni H, Nauman D, Burgess D et al. Factors influencingknowledge of and adherence to self-care among patientswith heart failure. Archives of Internal Medicine 1999; 159:1613–1619.

8 Gary R. Self-care practices in women with diastolic heartfailure. Heart & Lung 2006; 35: 9–19.

9 Schnell-Hoehn KN, Naimark BJ, Tate RB. Determinants ofself-care behaviours in community-dwelling patients withheart failure. The Journal of Cardiovascular Nursing 2009; 24:40–47.

10 Gallagher R. Self management, symptom monitoring andassociated factors in people with heart failure living in thecommunity. European Journal of Cardiovascular Nursing 2010;9: 153–160.

11 Evangelista LS, Dracup K, Doering LV. Treatment-seekingdelays in heart failure patients. The Journal of Heart and LungTransplantation 2000; 19: 932–938.

12 Goldberg RJ, Goldberg JH, Pruell S et al. Delays in seekingmedical care in hospitalized patients with decompensatedheart failure. The American Journal of Medicine 2008; 121:212–218.

13 Friedman M. Older adults’ symptoms and their durationbefore hospitalization for heart failure. Heart & Lung 1997;26: 169–176.

14 Parshall MB, Welsh JD, Brockopp DY, Heiser RM,Schooler MP, Cassidy KB. Dyspnea duration, distress, andintensity in emergency department visits for heart failure.Heart & Lung 2001; 30: 47–56.

15 Jurgens CY. Somatic awareness, uncertainty, and delay incare-seeking in acute heart failure. Research in Nursing &Health 2006; 29: 74–86.

16 Evangelista LS, Dracup K, Doering LV. Racial differences intreatment-seeking delays among heart failure patients.Journal of Cardiac Failure 2002; 8: 381–386.

17 Jurgens CY, Moser DK, Armola R, Carlson B, Sethares K,Riegel B. Symptom clusters of heart failure. Research inNursing & Health 2009; 32: 551–560.

18 Lee KS, Song EK, Lennie TA et al. Symptom clusters inmen and women with heart failure and their impact oncardiac event-free survival. The Journal of CardiovascularNursing 2010; 25: 263–272.

19 Hertzog MA, Pozehl B, Duncan K. Cluster analysis ofsymptom occurrence to identify subgroups of heart failurepatients: A pilot study. The Journal of Cardiovascular Nursing2010; 25: 273–283.

20 Webel AR, Frazier SK, Moser DK, Lennie TA. Daily vari-ability in dyspnea, edema and body weight in heart failurepatients. European Journal of Cardiovascular Nursing 2007; 6:60–65.

21 Riegel B, Moser DK, Anker SD et al. American Heart Asso-ciation (AHA) council on cardiovascular nursing; AHAcouncil on clinical cardiology; AHA council on nutrition,physical activity, and metabolism; AHA interdisciplinarycouncil on quality of care and outcomes research. State ofthe science: Promoting self-care in persons with heartfailure: A scientific statement from the American HeartAssociation. Circulation 2009; 120: 1141–1163.

138 R Gallagher et al.

© 2012 Blackwell Publishing Asia Pty Ltd

Page 7: Symptom patterns, duration and responses in newly diagnosed patients with heart failure

22 Weintraub NL, Collins SP, Pang PS et al. Acute heartfailure syndromes: Emergency department presentation,treatment, and disposition: Current approaches and futureaims. A scientific statement from the American HeartAssociation. Circulation 2010; 122(19): 1975–1996. Avail-able from URL: http://circ.aha.journals.org. AccessedOctober 2010.

23 Chriss PM, Sheposh J, Carlson B, Riegel B. Predictors ofsuccessful heart failure self-care maintenance in the firstthree months after hospitalization. Heart & Lung 2004; 33:345–353.

24 Soper DS. Statistics calculator: Post-hoc Statistical PowerCalculator for Multiple Regression. 2004. Available fromURL: http://www.danielsoper.com/statcalc/calc09.aspx. Accessed July 2010.

25 National Health and Medical Research Council of Australia.When does quality assurance in health care requireindependent ethical review? 2003. Available from URL:http://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/e46.pdf. Accessed March 2012.

26 Katz M. Multivariable Analysis: A Practical Guide for Clinicians,2nd edn. Cambridge: Cambridge University Press, 2008.

27 Cohen J, Cohen P. Applied Multiple Regression/CorrelationAnalysis for the Behavioural Sciences. Hillsdale, NJ, USA:Lawrence Erlbaum Associates, 1975.

28 Patel H, Shafazand M, Schaufelberger M, Ekman I. Reasonsfor seeking acute care in heart failure. European Journal ofHeart Failure 2007; 9: 702–708.

29 Berg HE, Eiken O, Miklavcic L, Mekjavic IB. Hip, thigh andcalf muscle atrophy and bone-loss after 5-week bedrestinactivity. European Journal of Applied Physiology 2007; 99:283–289.

30 Johansson P, Nieuwenhuis M, Lesman-Leegte I, vanVeldhuisen DJ, Jaarsma T. Depression and the delaybetween symptom onset and hospitalization in heart failurepatients. European Journal of Heart Failure 2011; 13(2): 214–219.

31 Cameron J, Worrall-Carter L, Page K et al. Cognitiveimpairment predicts poor self-care in patients with heartfailure. European Journal of Heart Failure 2010; 12: 508–515.

Heart failure symptoms and hospitalization 139

© 2012 Blackwell Publishing Asia Pty Ltd