The validity of a diagnosis of heart failure in a hospital discharge register

Transcription

1 The European Journal of Heart Failure 7 (2005) The validity of a diagnosis of heart failure in a hospital discharge register Erik Ingelsson a, *, Johan Ärnlfv a, Johan Sundstrfm a, Lars Lind b,c a Department of Public Health and Caring Sciences, Section of Geriatrics, Uppsala University, Box 609, SE Uppsala, Sweden b Department of Medical Sciences, Uppsala University, Sweden c Astra Zeneca R&D, Mölndal, Sweden Received 19 April 2004; received in revised form 25 October 2004; accepted 20 December 2004 Available online 23 May 2005 Abstract Background and aims: The accuracy of a diagnosis of heart failure () in hospital discharge registers is largely unknown. We aimed to determine the validity of such a diagnosis in the Swedish hospital discharge register. Methods and results: In a population-based study of 2322 middle-aged men (the ULSAM study), 321 participants were diagnosed with according to the Swedish hospital discharge register, during a median follow-up time of 29 years. A review board examined the validity of the diagnosis according to the European Society of Cardiology definition of. Eighty-two percent of the possible cases were classified as having definite. An echocardiographic examination increased the validity to 88%. For patients treated at an internal medicine or cardiology clinic the validity was 86% and 91%, respectively. If was the primary diagnosis, the validity was 95%, irrespective of clinic type. Conclusion: The diagnosis in the Swedish hospital discharge register appears slightly less precise than for acute myocardial infarction and stroke. For population-based research, only those with a primary diagnosis of in the hospital discharge register should be regarded as definite cases, or alternatively the cases should be validated individually. D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved. Keywords: Heart failure; Hospital discharge register; Epidemiology; Validity 1. Introduction In epidemiological studies, clinical disease end-points are often assessed using health registers. The reliability of such studies is dependent on the quality of the register data, which varies between different regions and different diagnoses. In many countries, data on hospitalizations are recorded in a national hospital discharge register. The overall quality of the Swedish hospital discharge register is commonly regarded as high, and the validity of the register has been evaluated for some diagnoses, e.g. acute myocardial infarction and acute stroke [1,2]. The reliability of these cardiovascular diagnoses in the register has been shown to be high, but to date, little is known about the * Corresponding author. Visiting address: K3ls7ngsgr7nd 10D, , Uppsala, Sweden. Tel.: ; fax: address: erik.ingelsson@pubcare.uu.se (E. Ingelsson). validity of the diagnosis of heart failure () in the Swedish, or any other, hospital discharge register. The prevalence and incidence of is rising with higher age in both men and women [3,4], and the age-adjusted mortality for patients is four to eight times that of the general population [5], although recent reports suggest an improving prognosis [6,7]. is commonly described as a syndrome, caused by cardiac dysfunction, and is recognized by a constellation of signs and symptoms, however, there is no unequivocal definition of. There are several diagnostic definitions for, which use clinical criteria [8 14], with varying sensitivities and specificities, depending on the severity of and the degree of certainty in the diagnosis [15]. These diagnostic schemes usually comprise combinations of clinical signs and symptoms of, laboratory blood and radiological examinations. Assessments of left ventricular filling pressures or systolic function indices may or may not be included in the diagnostic /$ - see front matter D 2005 European Society of Cardiology. Published by Elsevier B.V. All rights reserved. doi: /j.ejheart

2 788 E. Ingelsson et al. / The European Journal of Heart Failure 7 (2005) criteria. As left ventricular systolic function is normal in a large proportion of patients [16 19], schemes that do not rely on a measurement of systolic function are possibly to be preferred. The definition of proposed by the European Society of Cardiology (ESC), is intended to be used in clinical practice, clinical trials and epidemiological research [14], and is based on both symptoms of and objective evidence of cardiac dysfunction at rest. In cases where the diagnosis is in doubt, the response to treatment is a useful check of the diagnosis. Studies examining the validity of the diagnosis in different European and American patient samples have been published [20 22], but to our knowledge there is no published study of the validity of a diagnosis in any national hospital discharge register. Since the clinical definition of is less certain than that of acute myocardial infarction, we hypothesized that the hospital discharge register diagnosis of would be less valid. The primary aim of the present study was to determine the validity of a given diagnosis of in the Swedish hospital discharge register against the ESC definition, in order to be able to use this diagnosis as an end-point in future epidemiological studies of in Sweden. Secondary aims were to investigate if the presence of an echocardiographic examination, the type of hospitalization clinic or the position of the diagnosis code changed the accuracy of the diagnosis. 2. Methods 2.1. Study population This study uses the ULSAM (Uppsala Longitudinal Study of Adult Men) cohort, which is a health investigation aimed at identifying metabolic risk factors for cardiovascular disease. All 50-year-old men living in Uppsala, Sweden in were invited (2841 men) to participate in the ULSAM study. Of these, 82% (2322 men) participated in the investigation [23]. In addition to regular re-examinations, the data has been completed with annual updates on mortality and in-hospital morbidity using national registers. The ULSAM study is described in detail on the Internet ( All subjects gave written consent and the Ethics Committee of Uppsala University approved the study Registers The Swedish hospital discharge register is administered by the National Board of Health and Welfare, which started to collect data on individual patients who had been treated as in-patients at public hospitals, in the 1960s. For some 20 years, not all of the county councils reported their hospitalization data to the register since it was not compulsory. However, since 1987 when reporting was made compulsory, the hospital discharge register records all in-patient care in Sweden. It contains the dates of hospital admissions and discharges, hospital and clinic codes and up to six coded discharge diagnoses, the first being the principal cause of hospitalization. The register uses the codes of the International Classification of Diseases (ICD), edition eight (ICD-8) until the end of 1986, edition nine (ICD-9) from 1987 to the end of 1996 and edition ten (ICD-10) from 1997 onwards Selection of possible cases from the Swedish hospital discharge register The ULSAM participants were linked to the Swedish hospital discharge register using the unique personal identification number of all citizens of Sweden. The diagnosis of was allowed in any of the six possible diagnosis positions in the hospital discharge register. As a diagnosis of, we considered ICD heart failure codes , , (ICD-8), 428 (ICD-9), I50 (ICD-10) and hypertensive heart disease with heart failure, I11.0 (ICD- 10). Three hundred and twenty-one men had a hospital discharge register diagnosis of between entry into the ULSAM study and the end of 2001, with the first subject registered with a diagnosis in The median follow-up time was 29 years. If a subject had multiple hospital discharge register diagnoses of over the years, only the first occasion was considered in the present analysis Data collected for review Medical records, including referral notes, radiology reports, ECG reports, echocardiography reports if available, other journal records during the hospital stay and the discharge records were collected for each person with a diagnosis of in any diagnosis position in the hospital discharge register. Out of the 321 cases, two were excluded because of insufficient hospital and clinic coding in the hospital discharge register and two were excluded as their medical records could not be found, despite extensive searching in the archives Diagnostic classification The hospital discharge register cases were then classified by a review board consisting of two experienced doctors, who accessed all the journal records described above, and classified the cases as either, definite, questionable or miscoded. The classification relied on the definition proposed by the European Society of Cardiology [14]. Thus, to be classified as a definite case, there had to be symptoms and signs of and bobjective evidenceq of cardiac dysfunction at rest. In cases of doubt, the response to treatment was a useful check of the diagnosis. The required bobjective evidenceq was echocardiography, however, since the study commenced prior to the widespread availability of echocardiography, electrocardiography and

3 E. Ingelsson et al. / The European Journal of Heart Failure 7 (2005) X-ray were also considered acceptable when an echocardiography report was not available. The hospital discharge register cases where the review board could not find supporting evidence of according to the ESC definition were classified as questionable. For example, left ventricular dysfunction classified by echocardiography but without symptoms of, or breathlessness without objective evidence of were both classified as questionable. In questionable cases the review board also looked at any previous and subsequent admissions to clarify the diagnosis. In the few cases where a possible case was coded with an obviously incorrect ICD-code, the case was classified as miscoded Statistical methods The percentage of subjects with a register diagnosis of receiving a definite diagnosis in the review process was used to assess the validity of the hospital discharge register (positive predictive value). The hospital discharge register cases were divided into different sub-groups on the basis of which clinic they were hospitalized at (internal medicine, cardiology, lung medicine or other), if they underwent an echocardiographical examination during or within a few weeks of the hospital stay, and which position the diagnosis code was given (position 1, position 2 and position 3 6). All calculations were performed using JMP 3.2 (SAS Institute, Cary, NC, USA). 3. Results A total of 317 possible cases were enrolled in the study, based on the hospital discharge register diagnosis. Table 2 Validity of the heart failure diagnosis in the hospital discharge register over time using the European Society of Cardiology definition as gold standard Definite Questionable Miscoded All cases (n=317) Diagnosis year (n=108) 95 (88) 10 (9) 3 (3) Diagnosis year (n=101) 80 (79) 18 (18) 3 (3) Diagnosis year (n=108) 84 (78) 23 (21) 1 (1) Presence of echocardiography (n=154) Diagnosis year (n=32) 30 (94) 2 (6) 0 (0) Diagnosis year (n=53) 47 (89) 5 (9) 1 (2) Diagnosis year (n=69) 58 (84) 10 (14) 1 (1) Absence of echocardiography (n=163) Diagnosis year (n=76) 65 (86) 8 (11) 3 (4) Diagnosis year (n=48) 33 (69) 13 (27) 2 (4) Diagnosis year (n=39) 26 (67) 13 (33) 0 (0) Data is described in different subgroups with absolute numbers of cases (percentages in brackets). =Heart Failure. Using the ESC definition, 259 (82%) of the possible hospital discharge register cases were classified as having definite by the review board. The presence of an echocardiographical examination increased the validity to 88%, and the absence of an echocardiographical examination gave a diagnostic validity of 76%. For patients treated at an internal medicine or cardiology clinic the validity was 86% and 91%, respectively. Furthermore, if the diagnosis was in the first position in the hospital discharge register (primary diagnosis) the validity was 95%. The percentages and total numbers of cases that were classified as definite, questionable or miscoded in different sub-groups are listed in Table 1. For patients treated at an internal medicine or cardiology clinic with a primary diagnosis of, the Table 1 Validity of the heart failure diagnosis in the hospital discharge register using the European Society of Cardiology definition as gold standard Definite Questionable Miscoded Total (n=317) 259 (82) 51 (16) 7 (2) Echocardiography Presence of echocardiography (n=154) 135 (88) 17 (11) 2 (1) Absence of echocardiography (n=163) 124 (76) 34 (21) 5 (3) Clinic of hospitalization Hospitalization at an internal medicine clinic (n=203) 175 (86) 25 (12) 3 (1) Hospitalization at a cardiology clinic (n=65) 59 (91) 6 (9) 0 (0) Hospitalization at a lung medicine clinic (n=18) 9 (50) 9 (50) 0 (0) Hospitalization at any other hospital clinic (n=31) 16 (52) 11 (35) 4 (13) Diagnosis position diagnosis in position 1 of the hospital discharge register (n=140) 133 (95) 6 (4) 1 (1) diagnosis in position 2 of the hospital discharge register (n=112) 85 (76) 24 (21) 3 (3) diagnosis in position 3 6 of the hospital discharge register (n=65) 41 (63) 21 (32) 3 (5) Hospitalization at an internal medicine or cardiology clinic and primary 131 (96) 5 (4) 0 (0) diagnosis of hospitalization (n=136) The percentages refer to the total numbers in each row. Data is described in different subgroups with absolute numbers of cases (percentages in brackets). =Heart Failure.

4 790 E. Ingelsson et al. / The European Journal of Heart Failure 7 (2005) validity was 96%. When examining if the validity had changed over time, we could see that the highest validity was found in the earliest diagnosis dates, and had fallen from 88% to 78% during the last decade. This trend was not altered by the presence or non-presence of echocardiography (Table 2). 4. Discussion In the present study, we examined the validity of the diagnosis of in the Swedish hospital discharge register. Eighty-two percent of the cases in the hospital discharge register were classified as having definite according to the ESC definition. The validity of the hospital discharge register diagnosis was markedly higher in patients treated at an internal medicine or cardiology clinic, or when it was the primary diagnosis. We also found a tendency of decreasing diagnosis validity over time, and that the presence of an echocardiographical examination increased the validity only slightly. To our knowledge there is no published original study concerning the validity of the diagnosis in any national hospital discharge register, but in a recent preliminary report, the validity of the Scottish hospital discharge codes for was similar to our results [24]. In this report it was also concluded that the ICD codes alone failed to capture many admissions, and that using the hospital codes alone would underestimate the total burden of. Previous validation studies of other cardiovascular diseases, such as acute myocardial infarction [1,2,25] and stroke [2,26], have shown a validity of approximately 95% in the hospital discharge registers in different countries, including Sweden. Since is a more complex disease to diagnose, with symptoms and signs that can easily be misinterpreted, this lower validity is not unexpected. The presence of an echocardiographic examination increased the validity only slightly. This illustrates that is a disease defined by a constellation of clinical symptoms and signs, not just a measurement of left ventricular dysfunction. It may also reflect the clinician s uncertainty when facing the fact that echocardiographic left ventricular systolic function is normal in one out of two to three elderly patients with definite [16 19], in whom diastolic dysfunction may be the primary cause. The diagnosis in patients with a normal systolic function relies more heavily on clinical symptoms and signs, laboratory and X-ray findings and response to treatment. Interestingly, diagnostic validity appeared to decrease over the last two decades. The trend was not dependent of the presence of echocardiography. This pattern might be a result of changed routines for examination and investigation of patients in later years, with a greater reliance on echocardiography and lesser use of pulmonary X-ray and clinical picture, which could have diminished the doctors ability to correctly diagnose. More likely though, it is a result of less extensive medical records, i.e. not so carefully described signs and symptoms and shorter discharge records. This makes it harder to validate later hospital discharge register cases as definite, since this is a retrospective study. Another possible explanation of the decreasing validity over time may reflect that the study population has become 25 years older during the observation period and that diagnosing is more difficult in elderly people with a higher degree of co-morbidity. However, it should be noted that a longitudinal cohort study may not be the best study design to examine diagnosis validity at different time points. This observation therefore needs to be confirmed using cross-sectional studies at different time points. Of course, some of the hospital discharge register cases that were classified as questionable could in fact be true cases, if the evidence in the medical records was insufficient. However, in future utilization of this cohort for studies, as in retrospective epidemiological population studies examining etiology and risk factors for generally, it is better to have some false negative cases in the large referent group, than false positive cases in the smaller case group. However, in other types of studies, e.g. assessment of the economic burden of or studies of total incidence and prevalence in the community, such high specificity at the expense of some sensitivity, should not be preferable. A limitation of this study is that we only examined men of the same age with similar ethnic background (almost exclusively white), and from the same area of Sweden. Two hundred and eighty-nine (90%) of the discharges were from the local university hospital; hence the results could be biased by local routines. The results may therefore have limited generalisability to women and other age- and ethnic groups, as well as other settings. Strengths of the present study are the long follow-up time and the few cases lost to follow-up. In conclusion, this study shows that the validity of the diagnosis in the Swedish hospital discharge register appears less precise than for other recently investigated cardiovascular diagnoses, such as acute myocardial infarction and stroke [1,2,25,26], at least when including all clinics and all diagnosis positions. However, when including only cases from internal medicine and cardiology clinics or cases with a primary diagnosis of, the validity is comparable to the above diagnoses. Our findings imply that for population-based research, only those with a primary diagnosis of in the hospital discharge register should be regarded as definite cases, or alternatively the cases should be validated individually. Acknowledgements This study was supported by grants from Primary Health Care in Uppsala County, Royal Scientific Society

5 E. Ingelsson et al. / The European Journal of Heart Failure 7 (2005) Foundation (Kungliga vetenskapssamh7llets fond), Swedish Heart Lung Foundation (Hj7rt-Lungfonden) and Thuréus Foundation. References [1] Hammar N, Alfredsson L, Rosen M, Spetz CL, Kahan T, Ysberg AS. A national record linkage to study acute myocardial infarction incidence and case fatality in Sweden. Int J Epidemiol 2001; 30(Suppl. 1):S30 4. [2] Lindblad U, Rastam L, Ranstam J, Peterson M. Validity of register data on acute myocardial infarction and acute stroke: the Skaraborg Hypertension Project. Scand J Soc Med 1993;21(1):3 9. [3] Johansson S, Wallander MA, Ruigomez A, Garcia Rodriguez LA. Incidence of newly diagnosed heart failure in UK general practice. Eur J Heart Fail 2001;3(2): [4] Cowie MR, Mosterd A, Wood DA, Deckers JW, Poole-Wilson PA, Sutton GC, et al. The epidemiology of heart failure. Eur Heart J 1997;18(2): [5] Kannel WB, Belanger AJ. Epidemiology of heart failure. Am Heart J 1991;121(3 Pt. 1): [6] Cleland JG, Gemmell I, Khand A, Boddy A. Is the prognosis of heart failure improving? Eur J Heart Fail 1999;1(3): [7] Levy D, Kenchaiah S, Larson MG, Benjamin EJ, Kupka MJ, Ho KK, et al. Long-term trends in the incidence of and survival with heart failure. N Engl J Med 2002;347(18): [8] McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med 1971;285(26): [9] Wilhelmsen L, Eriksson H, Svardsudd K, Caidahl K. Improving the detection and diagnosis of congestive heart failure. Eur Heart J 1989;10(Suppl. C):13 8. [10] Walma EP, Hoes AW, Prins A, Boukes FS, van der Does E. Withdrawing long-term diuretic therapy in the elderly: a study in general practice in The Netherlands. Fam Med 1993;25(10): [11] Carlson KJ, Lee DC, Goroll AH, Leahy M, Johnson RA. An analysis of physicians reasons for prescribing long-term digitalis therapy in outpatients. J Chronic Dis 1985;38(9): [12] Schocken DD, Arrieta MI, Leaverton PE, Ross EA. Prevalence and mortality rate of congestive heart failure in the United States. J Am Coll Cardiol 1992;20(2): [13] Gheorghiade M, Beller GA. Effects of discontinuing maintenance digoxin therapy in patients with ischemic heart disease and congestive heart failure in sinus rhythm. Am J Cardiol 1983;51(8): [14] The Task Force on Heart Failure of the European Society of Cardiology. Guidelines for the diagnosis of heart failure. Eur Heart J 1995;16(6): [15] Mosterd A, Deckers JW, Hoes AW, Nederpel A, Smeets A, Linker DT, et al. Classification of heart failure in population based research: an assessment of six heart failure scores. Eur J Epidemiol 1997;13(5): [16] Vasan RS, Larson MG, Benjamin EJ, Evans JC, Reiss CK, Levy D. Congestive heart failure in subjects with normal versus reduced left ventricular ejection fraction: prevalence and mortality in a populationbased cohort. J Am Coll Cardiol 1999;33(7): [17] Vasan RS, Benjamin EJ, Levy D. Prevalence, clinical features and prognosis of diastolic heart failure: an epidemiologic perspective. J Am Coll Cardiol 1995;26(7): [18] Senni M, Tribouilloy CM, Rodeheffer RJ, Jacobsen SJ, Evans JM, Bailey KR, et al. Congestive heart failure in the community: a study of all incident cases in Olmsted County, Minnesota, in Circulation 1998;98(21): [19] Mosterd A, Hoes AW, de Bruyne MC, Deckers JW, Linker DT, Hofman A, et al. Prevalence of heart failure and left ventricular dysfunction in the general population; The Rotterdam Study. Eur Heart J 1999;20(6): [20] Goff Jr DC, Pandey DK, Chan FA, Ortiz C, Nichaman MZ. Congestive heart failure in the United States: is there more than meets the I(CD code)? The Corpus Christi Heart Project. Arch Intern Med 2000;160(2): [21] Marantz PR, Alderman MH, Tobin JN. Diagnostic heterogeneity in clinical trials for congestive heart failure. Ann Intern Med 1988; 109(1): [22] Remes J, Reunanen A, Aromaa A, Pyorala K. Incidence of heart failure in eastern Finland: a population-based surveillance study. Eur Heart J 1992;13(5): [23] Hedstrand H. A study of middle-aged men with particular reference to risk factors for cardiovascular disease. Ups J Med Sci Suppl 1975;19: [24] Khand AU, Shaw M, Gemmell I, Cleland JGF. Do discharge codes underestimate hospitalisation due to heart failure? Validation study of hospital discharge coding for heart failure. Eur J Heart Fail [25] Rapola JM, Virtamo J, Korhonen P, Haapakoski J, Hartman AM, Edwards BK, et al. Validity of diagnoses of major coronary events in national registers of hospital diagnoses and deaths in Finland. Eur J Epidemiol 1997;13(2): [26] Leppala JM, Virtamo J, Heinonen OP. Validation of stroke diagnosis in the National Hospital Discharge Register and the Register of Causes of Death in Finland. Eur J Epidemiol 1999;15(2):