You are here

Article Text

Article menu

Download PDFPDF

Heart failure and cardiomyopathies
Original research article
Routes to diagnosis of heart failure: observational study using linked data in England
  1. Alex Bottle1,
  2. Dani Kim1,
  3. Paul Aylin1,
  4. Martin R Cowie2,
  5. Azeem Majeed1,
  6. Benedict Hayhoe1
  1. 1 Department of Primary Care and Public Health, Imperial College London, London, UK
  2. 2 National Heart and Lung Institute, Imperial College London, London, UK
  1. Correspondence to Dr Alex Bottle, Department of Primary Care and Public Health, Imperial College, London, UK; robert.bottle{at}imperial.ac.uk

Abstract

Objective Timely diagnosis and management of heart failure (HF) is critical, but identification of patients with suspected HF can be challenging, especially in primary care. We describe the journey of people with HF in primary care from presentation through to diagnosis and initial management.

Methods We used the Clinical Practice Research Datalink (primary care consultations linked to hospital admissions data and national death registrations for patients registered with participating primary care practices in England) to describe investigation and referral pathways followed by patients from first presentation with relevant symptoms to HF diagnosis, particularly alignment with recommendations of the National Institute for Health and Care Excellence guideline for HF diagnosis.

Results 36 748 patients had a diagnosis of HF recorded that met the inclusion criteria between 1 January 2010 and 31 March 2013. For 29 113 (79.2%) patients, this was first recorded in hospital. In the 5 years prior to diagnosis, 15 057 patients (41.0%) had a primary care consultation with one of three key HF symptoms recorded, 17 724 (48.2%) attended for another reason and 3967 (10.8%) did not see their general practitioner. Only 24% of those with recorded HF symptoms followed a pathway aligned with guidelines (echocardiogram and/or serum natriuretic peptide test and specialist referral), while 44% had no echocardiogram, natriuretic peptide test or referral.

Conclusions Patients follow various pathways to the diagnosis of HF. However, few appear to follow a pathway supported by guidelines for investigation and referral. There are likely to be missed opportunities for earlier HF diagnosis in primary care.

  • heart failure
  • electronic medical records
  • health care delivery
  • quality and outcomes of care

https://doi.org/10.1136/heartjnl-2017-312183

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    In heart failure (HF), accumulation of fluid in lungs and peripheries and reduction in blood flow to the muscles result in classic symptoms of breathlessness, ankle swelling and fatigue.1 HF has considerable impact on patients and healthcare systems,2 with high morbidity and mortality.3 The Global Burden of Disease study estimated that some 40 million people had HF worldwide in 20154; the estimated UK total is 550 000.5

    Early identification is key to effective management.1 Professionals rely initially on symptom recognition; the use of serum natriuretic peptide (either B-type natriuretic peptide (BNP) or N-terminal pro-B-type natriuretic peptide (NTproBNP)) as a screening test in symptomatic patients has been recommended for over 10 years.6–8

    Patients with HF often present first in primary care.9 Identification and diagnosis in primary care is supported by professional guidelines1 8 10 recommending minimum standards of investigation,11 as well as medications improving symptoms and prognosis (beta-blockers and ACE inhibitors). In the National Health Service (NHS) in England, some elements of investigation and management in primary care (echocardiographic imaging, referral for ‘specialist assessment’, referral for exercise-based rehabilitation and treatment with an ACE inhibitor or angiotensin II receptor blocker (ARB) and a beta-blocker) have for over 10 years been supported by performance related payments under an incentivisation scheme—the Quality and Outcomes Framework.

    National Institute for Health and Care Excellence (NICE) guidance recommends the following for patients presenting with HF symptoms in primary care8:

    1. history and examination

    2. for patients with history of acute myocardial infarction (AMI): referral for echocardiograph and ‘specialist’ assessment, usually by a consultant cardiologist or HF nurse specialist, for confirmation of diagnosis, classification and planning of management, within 2 weeks

    3. for patients without history of AMI: serum natriuretic peptide and ECG, followed by echocardiograph and referral for specialist assessment within 6 weeks subject to a BNP level of 100–400 pg/mL or NTproBNP 400–2000 pg/mL (patients with higher levels, BNP >400 pg/mL or NTproBNP >2000 pg/mL, require referral within 2 weeks).

    Diagnosis and management of HF in primary care has been criticised, with evidence of both underdiagnosis12 13 and overdiagnosis,14 and no improvement in survival over nearly 15 years.15 Professionals may fail to adhere to guidance and lack confidence in investigations and treatments.12 16 17 This has changed little, despite new guidance and incentivisation.15 16 18 However, a clear picture of identification and management pathways for HF in primary care is currently lacking. Using linked primary and secondary care routinely collected data, we sought to characterise the journey of patients with HF symptoms in NHS primary care in England through investigation, treatment and referral, assessing potential for earlier diagnosis.

    Methods

    Data sources

    The Clinical Practice Research Datalink (CPRD) is a database of pseudonymised electronic records from about 7% of UK general practices from 1987 to the present, considered representative of the UK population.19 Primary care records are linked nationally to hospital admissions (Hospital Episode Statistics (HES)) and the death registry (Office for National Statistics (ONS)).

    Patient cohort and subgroups

    We searched for patients with HF diagnosed between 1 January 2010 and 31 March 2013. For each patient, diagnosis date was the first coded record of HF, either in the primary care record or in hospital admission data. We identified primary care consultations and hospital admissions for HF using codes by Hawkins et al 20 augmented by our local clinicians (see online appendix). In HES, the earliest record of HF in International Statistical Classification of Diseases, 10th revision, codes, recorded as either primary or secondary diagnoses, was used.

    The following exclusion criteria applied: CPRD records at practices not linked to HES, patients not registered in a CPRD practice for the whole 10-year period and standard CPRD data quality exclusions regarding ‘up to standard’ practices and patients with ‘acceptable data quality’ (see online appendix).

    Published sets of clinical codes for patient characteristics, tests, medications and referrals were used whenever available (see online appendix), tracking back 5 years before the date of diagnosis. For example, comorbidities from the Charlson set were defined as per Khan et al,21 with some extra comorbidities defined by our team. Otherwise, codes were identified using the CPRD medical and product dictionaries.

    Pathway definitions

    We divided patients into two main groups according to primary care consultations in the 5 years before HF diagnosis: those with a primary care record of HF symptoms and those with a primary care consultation but no recorded HF symptoms.

    In the first group, we classified management into seven ‘pathways’ depending on tests, medications and referrals recorded. Pathway #1 comprised patients investigated in accordance with NICE recommendations (echocardiogram and/or serum natriuretic peptide test and a referral to a specialist): we refer to this as the ‘NICE recommended pathway’ as a shorthand, although for the most part we did not incorporate the guideline’s timeframe. Partial concordance with guidelines was defined as pathway #2 (echocardiogram and/or serum natriuretic peptide test but no referral) or pathway #4 (referral only). Patients who had no investigations recorded but were on appropriate medications for HF were categorised as pathway #3a (put on a new type of medication) or pathway #3b (already on medication). Medications of interest were ACE inhibitors, ARBs and HF-specific beta-blockers. Pathway #5 (‘Other pathway’) covered those who had received other relevant management options, such as smoking cessation advice, influenza vaccination, an ECG and/or a non HF-specific beta-blocker. Finally, pathway #6 (‘No pathway followed’) covered patients with recorded symptom(s) who did not appear to receive any management.

    For the second group, who saw their general practitioner (GP) but had no record of HF symptoms, the same set of pathways was applied using consultation records in the 5 years before HF diagnosis.

    Statistical analysis

    Patient characteristics were summarised for all patients, while comorbidities were examined for those presenting to their GP with an HF symptom before diagnosis. Data were reported as absolute numbers and proportions with p values from χ2 tests. Analyses were conducted using SAS V.9.4.

    Results

    Patient characteristics

    Between 1 January 2010 and 31 March 2013, 36 748 patients had a diagnosis of HF recorded in CPRD practices that met the inclusion criteria and were included in the analysis. A total of 29 113 (79.2%) had this first recorded in hospital data (HES) and 7635 (20.8%) in primary care. Half were female and nearly one in three was aged 85 or over; 9 in 10 had at least one potentially relevant comorbidity, with nearly half having three or more, especially hypertension (69.9%), chronic pulmonary disease (33.9%) and coronary artery disease (31.6%) (table 1).

    View this table:
    Table 1

    Patient characteristics: at diagnosis or before presentation of first symptom by diagnosis source

    A total of 15 057 patients (41.0%) had a primary care consultation with one of the three HF symptoms recorded before diagnosis (table 2), while 21% presented with any two and just 2% presented with all three. The most common symptom was breathlessness, present in 80% of the 15 057. Half had a primary care consultation prediagnosis but with some other symptom recorded, and the remaining 11% had no primary care consultation prior to diagnosis.

    View this table:
    Table 2

    Proportion of patients with heart failure symptom before diagnosis by symptom type

    Figures 1 and 2 show patients in each of the two main groups split by the elements of the NICE-recommended pathway; table 3 provides more detail on their management. The remaining 3967 patients had no contact with primary care in the 5 years before diagnosis and were hospitalised for HF, giving this hospital admission date as the date of their diagnosis. We did not consider them further in the analysis.

    Figure 1
    Figure 1

    Pathways to diagnosis for patients presenting to their GP before their HF diagnosis with HF symptoms recorded. BNP, B-type natriuretic peptide; GP, general practitioner; HF, heart failure.

    Figure 2
    Figure 2

    Pathways to diagnosis for patients presenting to their GP before their HF diagnosis with only non-HF symptoms recorded. BNP, B-type natriuretic peptide; GP, general practitioner; HF, heart failure.

    View this table:
    Table 3

    Management pathways after first symptom*, overall and by diagnosis source

    In addition to categorising patients based on HF symptom recorded in primary care, we also examined whether HF was first recorded in primary or secondary care; we use ‘GP-diagnosed’ and ‘hospital-diagnosed’ as shorthand for this. One in five patients were GP diagnosed, of whom almost all (97%) had seen the GP for some reason in the previous year and just over half (55%) had presented with HF symptoms in the previous 5 years. Nearly four in five patients were hospital diagnosed, of whom most (89%) had seen the GP for some reason in the previous year and a third (37%) had an HF symptom recorded in a GP consultation in the previous 5 years.

    Patients with a record of HF symptoms in a primary care consultation

    Nearly one in four (23.6%) of the 15 057 patients with HF symptoms recorded before diagnosis followed the NICE-recommended pathway (#1) of an echocardiogram and/or serum natriuretic peptide test and a specialist referral; another 14.4% were referred without either investigation (figure 1, table 3). GP-diagnosed patients were more than twice as likely to be on the NICE-recommended pathway (#1) than hospital-diagnosed ones. Of those not referred for any element of the NICE pathway, 3705 (=926+2779) or a quarter of this group were already taking or started relevant medication (pathways #3a and #3b) on presentation of HF symptoms. In the 9% of this group on no pathway, subsequent hospital diagnosis was much more likely than GP diagnosis.

    Considering adherence to NICE recommended timeframes, 1991 (13.2% of the 15 057) had had a previous AMI at the time of their first symptom, of whom only 23 (1.2%) completed the pathway within the recommended 2 weeks. A total of 13 066 (86.8% of the 15 057) had not had a previous MI at the time of their first symptom, of whom 511 (3.9%) completed the pathway within the recommended 6 weeks.

    Patients with only a record of non-HF symptoms in a primary care consultation

    Nearly one in eight (13.2%) of the 17 724 with only non-HF symptoms recorded in primary care before diagnosis followed the NICE-recommended pathway (#1); another 12.5% were referred without recorded investigation (figure 2, table 3). Of those 3922 (=1606+134+1918+264 from figure 2) patients referred for an echo without recorded HF symptoms, 1461 (37%) had circulatory disorders recorded at the consultation at which the referral was made, including 786 (20%) that were actually HF. One in four of this group followed no pathway: again, subsequent hospital diagnosis was much more likely than GP diagnosis.

    Discussion

    Summary of findings

    Patients took many routes to a HF diagnosis. Only one in five patients with recognised HF had it first recorded in a primary care consultation, although almost all of them had seen a GP in the year before this diagnosis date, almost half with recorded HF symptoms. The remaining four out of five patients had HF recorded for the first time during a hospital admission; the vast majority of these had also seen a GP in the previous year, 37% with recorded HF symptoms.

    In the 5 years before diagnosis, only 11% had no primary care consultation. Of the 41% of our sample with a primary care consultation with one of the three key HF symptoms, management aligned with NICE guidance was documented for nearly one in four (pathway #1): only 4% followed it within 6 weeks. Another one in seven were referred without record of either investigation (pathway #4).

    For the 48% of patients who had a primary care consultation with only non-HF symptoms recorded, the picture was also mixed. One in nine followed the NICE-recommended pathway, suggesting that the GP suspected HF but did not record an HF symptom. This is supported by the fact that HF was recorded as a diagnosis for 20% of these patients during the consultation when the referral was made.

    Comparison with other studies

    CPRD has recently been used to look at prognosis following HF diagnosis in primary care and secondary outpatient care.22 Of nearly 78 000 patients identified with HF in primary or secondary care in that study between 1997 and 2010, 42% were first recorded in primary care and 58% on hospital admission. Our data, covering patients diagnosed between 2010 and 2013, suggest a more extreme differential of 21% and 79%. Inspection of our full CPRD extract dating back in 1997 reveals a steady rise in the proportion of patients diagnosed in hospital from about 50% in 2003 to about 80% in 2013. Reasons for this are unclear. It is possible that increasing familiarity with guidelines has led GPs to be cautious in diagnosing and treating HF in primary care based on clinical assessment: they may be more likely to await recommended investigations and specialist assessment. Pressure on secondary care diagnostic services may lead to some patients then presenting to hospital acutely while still under investigation. Patients admitted to hospital for HF without a prior diagnosis of HF in primary care have worse prognosis and management.22 Much less is known about what happens before diagnosis, and ours is the first study to our knowledge mapping out the routes to a diagnosis of HF.

    Strengths

    Almost 100% of England’s population has access to primary and secondary care. The large data source provides sufficient sample size to describe the varied routes to a diagnosis of HF in real-world practice. The CPRD population is broadly representative of the UK population in terms of age, sex and ethnicity, and the database is well used in research.19 HES covers all NHS hospitals in England with excellent coverage and highly accurate primary diagnosis and procedure coding.23 The CPRD–HES–ONS linkage allows researchers to track patient pathways through the whole system, rare in other countries.

    Limitations

    CPRD data are entered by GPs during routine consultations and not for the purpose of research. Their use allows auditing for clinical quality improvement and supports claims relating to financially incentivised activities.24–26 Therefore, those required by schemes such as the UK’s Quality and Outcomes Framework may be more likely to be used than, for example, symptom codes.27 However, it is recognised that coding in primary care remains highly variable.24

    We divided our cohort based on recording of HF symptoms in GP consultations. The validity of this approach depends both on patients reporting symptoms and, more importantly, GPs recording them as Read codes; free text is no longer available in CPRD. We found a number of patients following the NICE HF pathway despite having none of the three main symptoms entered for that consultation. Instead, a variety of codes for procedures, examination and diagnoses were listed, with over half the patients having circulatory system disorders.

    Another apparent anomaly is that 18% of patients referred for an echocardiogram actually had a diagnosis of HF recorded in the same consultation. Some GPs likely wished to code a working diagnosis while following the pathway for investigation. Of course, other GPs may wish to have the diagnosis confirmed by investigations before ‘committing themselves’ by entering the code for HF in the notes, leading to the variation we found.

    As with many epidemiological studies, case designation relies primarily on physician diagnosis rather than objective investigations. While there is evidence of underdiagnosis,28 a UK study that used an expert panel found that perhaps a third of patients may be incorrectly labelled as having HF.16 Nevertheless, the working diagnosis is arguably the most relevant definition when examining prescribing behaviour.20

    Early identification and classification of HF is key to effective management,1 with an important distinction to be made between patients who have HF with reduced ejection fraction, where conventional medical therapies are effective in improving outcomes, and HF with preserved ejection fraction, where they generally are not.28 This distinction is generally not well made in primary care,28 and relevant Read codes are rarely used. We were therefore unable to distinguish between the two.

    Our definitions of the various pathways following GP consultation depended on several elements. Both referrals and echocardiograms may be made for other cardiac problems and are not necessarily specific for HF, but we have made the assumption that in our cohort, consisting of patients with HF recorded at some point, these were made for suspected HF. Serum natriuretic peptide testing was recommended in the 2005 European Society of Cardiology and the August 2010 NICE guidance but only available for a minority of practices during our study period. It is now more widely used, though threshold values are still under discussion.9

    Implications for policy and practice

    There are significant challenges in the identification, diagnosis and management of HF in primary care.9 HF manifests itself in different ways, resulting in variation in patients’ health-seeking behaviour and GPs’ decision making. Sudden onset, severe symptoms will lead the patients straight to the emergency department, bypassing earlier diagnosis by the GP. In contrast, patients with more insidious symptoms of HF may not see their GP in the early stages but wait instead until symptoms worsen and then attend the emergency department, perhaps influenced by actual or perceived problems with GP access. Others may see their GP with several different problems and fail to mention HF symptoms until the GP’s suspicions are at last raised.

    From the GP’s perspective, the diagnosis is made harder by the lack of specificity of HF symptoms and the confusion with respiratory conditions12 (one in three of our cohorts had chronic obstructive pulmonary disease compared with just 2% for the general population29 and 11% for patients aged 75–79),30 limited time availability, limited access to investigations and low confidence in interpretation of investigation results.9While clear clinical guidelines exist, perceived information overload as well as a belief that they do not apply to all patients mean that GPs do not always use them9 and not all will be aware of them. Our results suggest that NICE guidelines (currently being updated) are not followed in primary care for the majority of patients who go on to be diagnosed with HF, with a significant majority being diagnosed in the acute, secondary care setting. Finally, uncertainty about pharmacological management strategies, and concerns about comorbidity and polypharmacy, especially in elderly patients,9 complicates initial management; long waiting lists and limited availability of serum natriuretic peptide testing and specialist clinics may also prove a disincentive to GPs in referring patients.9

    Conclusions

    Patients follow various routes through the NHS to diagnosis of HF. Most continue to be diagnosed in hospital despite many presenting earlier in primary care with suggestive symptoms. Further research is needed to understand how GPs react to patients presenting with these symptoms and how they use current guidelines: there appears to be considerable opportunity to improve the early management of these patients in primary care.

    Key messages

    What is already known on this subject?

    • Early identification of heart failure (HF) is key to effective management, but diagnosis can be challenging, especially in primary care.

    • National Institute for Health and Care Excellence (NICE) guidelines recommend investigation of patients with suspected HF with echocardiogram and/or serum natriuretic peptide, followed by specialist referral; degree of adherence to this recommendation is not known.

    What might this study add?

    • Almost 80% of patients with HF appear to be diagnosed first in secondary care.

    • Of those with HF symptoms recorded in primary care, only 24% followed a pathway aligned with NICE guidelines and rarely within the strict timeframe.

    How might this impact on clinical practice?

    • There may be missed opportunities for earlier HF diagnosis in primary care.

    Supplemental material

    Acknowledgments

    The Dr Foster Unit at Imperial is affiliated with the National Institute of Health Research (NIHR) Imperial Patient Safety Translational Research Centre. The NIHR Imperial Patient Safety Translational Centre is a partnership between the Imperial College Healthcare NHS Trust and Imperial College London. The Dr Foster Unit at Imperial College are grateful for support from the NIHR Biomedical Research Centre funding scheme. The views and opinions expressed here in are those of the authors and do not necessarily reflect those of the NHS, the NIHR,MRC, CCF, NETSCC, the HSR programme or the Department of Health.

    References

    1. ↵
      2. Ponikowski P ,
      3. Voors AA ,
      4. Anker SD , et al
      . ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: the task force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). Eur Heart J 2016;37:2129–200.
      OpenUrlCrossRefPubMed
    2. ↵
      2. Hobbs FDR ,
      3. Doust J ,
      4. Mant J , et al
      . Diagnosis of heart failure in primary care. Heart 2010;96:1773–7.
      OpenUrlFREE Full Text
    3. ↵
      2. Roger VL
      . Epidemiology of heart failure. Circ Res 2013;113:646–59.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016;388:1545–602.
      OpenUrlCrossRefPubMed
    5. ↵
      2. Bhatnagar P ,
      3. Wickramasinghe K ,
      4. Williams J , et al
      . The epidemiology of cardiovascular disease in the UK 2014. Heart 2015;101:1182–9.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      2. Yancy CW ,
      3. Jessup M ,
      4. Bozkurt B , et al
      . 2013 ACCF/AHA guideline for the management of heart failure. J Am Coll Cardiol 2013;62:e147–239.
      OpenUrlFREE Full Text
    7. ↵
      2. Swedberg K ,
      3. Cleland J ,
      4. Dargie H , et al
      . Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005): the task force for the diagnosis and treatment of chronic heart failure of the European Society of Cardiology. Eur Heart J 2005;26:1115–40.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      National Institute of Health and Care Excellence. Chronic heart failure: national clinical guideline for diagnosis and management in primary and secondary care: National Clinical Guideline Centre, 2010.
    9. ↵
      2. Smeets M ,
      3. Van Roy S ,
      4. Aertgeerts B , et al
      . Improving care for heart failure patients in primary care, GPs' perceptions: a qualitative evidence synthesis. BMJ Open 2016;6:e013459.
    10. ↵
      Scottish Intercollegiate Guidelines Network. SIGN 147 Management of chronic heart failure: a national clinical guideline: SIGN, 2016.
    11. ↵
      2. Cleland JG ,
      3. Cohen-Solal A ,
      4. Aguilar JC , et al
      . Management of heart failure in primary care (the improvement of Heart Failure Programme): an international survey. Lancet 2002;360:1631–9.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      2. van Riet EES ,
      3. Hoes AW ,
      4. Limburg A , et al
      . Prevalence of unrecognized heart failure in older persons with shortness of breath on exertion: unrecognized HF in older persons with shortness of breath on exertion. Eur J Heart Fail 2014;16:772–7.
      OpenUrlCrossRefPubMed
    13. ↵
      2. Morgan S ,
      3. Smith H ,
      4. Simpson I , et al
      . Prevalence and clinical characteristics of left ventricular dysfunction among elderly patients in general practice setting: cross sectional survey. BMJ 1999;318:368–72.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      2. Valk MJ ,
      3. Mosterd A ,
      4. Broekhuizen BD , et al
      . Overdiagnosis of heart failure in primary care: a cross-sectional study. Br J Gen Pract 2016;66:e587 –92.
      OpenUrl
    15. ↵
      2. Taylor CJ ,
      3. Ryan R ,
      4. Nichols L , et al
      . Survival following a diagnosis of heart failure in primary care. Fam Pract 2017;34:161-168.
      OpenUrl
    16. ↵
      2. Hancock HC ,
      3. Close H ,
      4. Fuat A , et al
      . Barriers to accurate diagnosis and effective management of heart failure have not changed in the past 10 years: a qualitative study and national survey. BMJ Open 2014;4:e003866.
    17. ↵
      2. Dahlstrom U ,
      3. Hakansson J ,
      4. Swedberg K , et al
      . Adequacy of diagnosis and treatment of chronic heart failure in primary health care in Sweden. Eur J Heart Fail 2009;11:92–8.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      2. Brettell R ,
      3. Soljak M ,
      4. Cecil E , et al
      . Reducing heart failure admission rates in England 2004-2011 are not related to changes in primary care quality: national observational study. Eur J Heart Fail 2013;15:1335–42.
      OpenUrlCrossRefPubMed
    19. ↵
      2. Herrett E ,
      3. Gallagher AM ,
      4. Bhaskaran K , et al
      . Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 2015;44:827–36.
      OpenUrlCrossRefPubMed
    20. ↵
      2. Hawkins NM ,
      3. Scholes S ,
      4. Bajekal M , et al
      . Community care in England: reducing socioeconomic inequalities in heart failure. Circulation 2012;126:1050–7.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      2. Khan NF ,
      3. Perera R ,
      4. Harper S , et al
      . Adaptation and validation of the charlson index for read/OXMIS coded databases. BMC Fam Pract 2010;11:11:1.
      OpenUrl
    22. ↵
      2. Koudstaal S ,
      3. Pujades-Rodriguez M ,
      4. Denaxas S , et al
      . Prognostic burden of heart failure recorded in primary care, acute hospital admissions, or both: a population-based linked electronic health record cohort study in 2.1 million people. Eur J Heart Fail 2017;19:1119–27.
      OpenUrl
    23. ↵
      2. Burns EM ,
      3. Rigby E ,
      4. Mamidanna R , et al
      . Systematic review of discharge coding accuracy. J Public Health 2012;34:138–48.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      2. Jordan K ,
      3. Porcheret M ,
      4. Croft P
      . Quality of morbidity coding in general practice computerized medical records: a systematic review. Fam Pract 2004;21:396–412.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      2. Gray J ,
      3. Orr D ,
      4. Majeed A
      . Use of read codes in diabetes management in a south London primary care group: implications for establishing disease registers. BMJ 2003;326:1130.
      OpenUrlAbstract/FREE Full Text
    26. ↵
      2. Doran T ,
      3. Fullwood C ,
      4. Gravelle H , et al
      . Pay-for-performance programs in family practices in the United Kingdom. N Engl J Med 2006;355:375–84.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      2. Deaton C ,
      3. Benson J
      . Time for correct diagnosis and categorisation of heart failure in primary care. Br J Gen Pract 2016;66:554–5.
      OpenUrlFREE Full Text
    28. ↵
      2. Fuat A ,
      3. Hungin AP ,
      4. Murphy JJ
      . Barriers to accurate diagnosis and effective management of heart failure in primary care: qualitative study. BMJ 2003;326:196.
      OpenUrlAbstract/FREE Full Text
    29. ↵
      Quality and Outcomes Framework 2015-16. Recorded disease prevalence, achievements and exceptions: NHS Digital, 2016. http://content.digital.nhs.uk/pubs/qof1516. (accessed 1 Sep 2017).
    30. ↵
      2. Melzer D ,
      3. Correa Delgado J ,
      4. Winder R , et al
      . The age UK almanac of disease profiles in later life. A reference on the frequency of major diseases, conditions and syndromes affecting older people in England, 2015. http://www.ageuk.org.uk/Documents/EN-GB/For-professionals/Research/Age_UK_almanac_FINAL_9Oct15.pdf?dtrk=true (accessed 1 Sep 2017).

    Footnotes

    • Contributors AB, DK and BH conceived and designed this study. AB and DK prepared the data; DK carried out the analysis, overseen by AB and BH. All authors took part in interpreting the data for this study, commented on and helped to revise drafts of this paper and have approved the final version. AB is the guarantor.

    • Funding The Dr Foster Unit at Imperial College London is partially funded by a grant from Dr Foster, a Telstra Health private healthcare information company. The Dr Foster Unit at Imperial College London is partly funded by research grants from the National Institute for Health Research Health Services Research. Prof Cowie’s salary is supported by the NIHR Cardiovascular Biomedical Research Unit at the Royal Brompton Hospital, London.

    • Ethics approval ISAC of CPRD; we also have LREC approval from NRES London – South East committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement Due to information governance rules applicable to CPRD, no data are available for sharing.

    • Correction notice Since this paper was first published online an acknowledgements statement has been added.