Research Letter |
Prevalence of chronic disease in the elderly based on a national pharmacy claims database
SIR—The elderly are the fastest growing sector of society with those
70 years of age accounting for nearly 8% of the Irish population [1, 2]. Yet there is little or no baseline information on the prevalence of chronic disease in this population [3]. This lack of information on health trends can inhibit accurate predictions for future health care needs in this age group.
The most recent data on prevalence of chronic disease were reported in the 2000 ‘Health and Social Services for Older People I’ (HeSSOP I) survey [4], which relied on self-reported illness and involved 937 community-dwelling adults 65 years of age. HeSSOP I provides insight into the prevalence of certain chronic conditions. However, this was not the primary purpose of the survey, and disease definition was vague, for example, ‘memory/concentration problems’. It is also likely to underestimate disease prevalence as it excludes those in nursing homes.
This article explores the potential for using a national pharmacy claims database to estimate prevalence of disease. The data are readily available, are collected on a continuous basis and cover the majority of prescribing in those aged 70 years. However, the data are not diagnosis linked; thus, combination of drug therapies are used as surrogate markers of disease [5]. This approach lacks specificity for certain conditions as some drugs have broad licensed indications; however, this method has been used and validated in other settings [6–11].
The aim of this study was to estimate chronic disease prevalence in an elderly national population using an existing prescribing claims database.
 |
Methods |
|---|
 Top Methods ResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
The Irish HSE–Primary care reimbursement services (HSE–PCRS) pharmacy database was used. The HSE–PCRS provides free health services, including medicines to 1.2 million people in Ireland. It is a means-tested scheme for people aged <70 years, but since July 2001 it is free to all those aged
70 years. It is estimated that over 97% of this age group avail of the medical card scheme [3].
A framework for identifying conditions from the pharmacy database using drug WHO ATC codes was developed from the chronic disease score classification I [5, 8, 12] (see Appendix 1 in the supplementary data on the journal website, http://www.ageing.oxfordjournals.org). The national HSE–PCRS prescribing database for 2004 was used. The study population was defined as all individuals aged 70years who were dispensed three or more items associated with a specific chronic condition during the 12-month period. Restricting to a minimum of three scripts was felt to reduce the likelihood of including prescriptions associated with misdiagnosis or incidental users.
Statistical analysis
Chronic disease prevalence rates were calculated for the elderly population based on the number of patients aged 70 years (n = 316,928). Cost of drug therapy was calculated as the sum of the ingredient drug costs for the combination of drugs used, and cost per patient was the total cost divided by the population with that condition. The age (70–74 years and 75 years) and gender (female is the reference group) distribution of chronic disease is examined; incidence ratios and 95% confidence intervals (CIs) are presented. All analysis was performed using SAS (v 9.1 SAS Institute, Cary, NC, USA).
|
Results |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
In 2004, 86% (271,518) of elderly patients received three or more drug items for at least one of the nine chronic conditions identified in Table 1. Those aged
75 years accounted for 62% (169,019) of the population, and 59% (161,114) were females.
|
Table 1 gives the prevalence data for chronic conditions among this elderly population. Cardiovascular disease (CVD) was the most common at 72%. This was followed by central nervous system (CNS) conditions at 37%, musculo-skeletal conditions at 28%, upper gastrointestinal (GI) at 24% and respiratory at 14%. Diabetes, thyroid disease and glaucoma occurred in 5–8% of this population, and cancer therapy was received by 4% but is likely to be an underestimate [13].
Overall, CVD was the highest costing condition to treat; however, the cost per patient was relatively low at 
390 per year compared with dementia at 907 or respiratory conditions at 540 per patient per year.
Co-morbidities
There was a high level of co-morbidity, with two chronic diseases experienced by 27% (86,514), three conditions by 19% (60,930) and four or more conditions by 14% (44,035) of the population. The average cost for a single condition was 316, which increased to 609 for two, 870 for three and 1164 per patient per year for those with four conditions. CVD and CNS conditions were the most frequently occurring combination at 30% (94,271), followed by 23% (72,986) for CVD and musculo-skeletal conditions, 20% (63,296) for CVD and upper GI conditions and 13% (42,430) for CNS and musculo-skeletal conditions.
Table 2 summarises the age and gender distribution for the six most prevalent conditions. There is a consistent pattern of diseases between the age groups, with CVD, respiratory disease and diabetes more prevalent in males. CNS, musculo-skeletal and GI conditions were more common in females.
|
|
Discussion |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
The demographic structure of the study population closely mirrored that of the
70 age group national population where women account for 58% and those aged 75 years 63% [1]. Eighty-six per cent of the elderly population had experienced at least one of the nine chronic conditions identified. Three-quarters were receiving CVD medication, of which >40% was for ischaemic vascular conditions. However, antihypertensive therapy accounted for the majority of prescriptions. There was an estimated prevalence of between 20% and 30% for CNS, musculo-skeletal or upper GI diseases.
Cardiovascular conditions were the highest costing disease group, but on a cost-per-patient basis respiratory conditions and dementia were the most expensive conditions to treat medically. The prevalence of co-existing diseases suggest that three-quarters of elderly people live with two or more significant chronic diseases; CVD and CNS conditions being the most common combination. There was also a clear gender difference in many of the diseases examined. CVD and respiratory conditions were more prevalent in men, which may reflect higher smoking prevalence, whereas CNS and musculo-skeletal conditions were more common in women, which may reflect their increased willingness to seek help for these illnesses.
Accurate disease prevalence data are required both from an epidemiological and strategic health care planning perspective. There is a strong association between chronic disease, especially arthritis, CVD and stroke, and a decline in locomotor activity and social participation among the elderly [14, 15]. In addition, life expectancy is increasing with an associated increase in years lived with some degree of disability [16–18]. Underestimating disease and co-morbidity prevalence could have far reaching implications for the provision of health care and elderly care services in the future. This includes the provision of community services to help maintain elderly people in their own homes, nursing home beds and specialist practitioners.
In Ireland, there is no regular national data collection on health care or pharmacy utilisation, with the exception of the HSE–PCRS database in those aged 70 years. This database represents a unique opportunity to estimate the national prevalence of chronic disease in the elderly population.
This study suggests that the prevalence of chronic disease in the elderly population may be higher than previously thought. In comparison with the HeSSOP I survey, 56% of the survey population reported CVD, 46% musculo-skeletal conditions, 14% respiratory conditions, 6% diabetes, whereas 12–17% reported sleep disturbance or depression/anxiety, but the overall prevalence of CNS conditions was difficult to estimate [4]. Differences are likely to reflect the strengths and limitations of both methodologies and the difficulty in accurately identifying national disease prevalence in specific populations. Self-reported illness relies on a person understanding the medical conditions they are been treated for, whereas the higher self-reported prevalence of musculo-skeletal conditions reflect the fact that this condition does not always result in a chronic prescription. Accurate clinical data with standardised disease definitions supplied by general practitioners are required to validate both methodologies, which are currently not collected in the Irish health care system.
Our study provides a valuable insight into the medical conditions experienced by the elderly population within one country. Pharmacy data are collected in several other countries, and there is the possibility of developing a consensus on surrogate markers for diseases, which could facilitate international comparison of disease states and treatments. This would require standardised methodology and validation against diagnosis data.
|
Study limitations |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
The limitations of pharmacy claims databases are readily acknowledged [5, 8, 19] and mainly concern the lack of definitive diagnosis information (see supplementary data available at the journal website, http://www.ageing.oxfordjournals.org). In addition, this methodology cannot identify conditions that are not managed by standard drug therapy or conditions for which there are no drug therapies, for example, some cancers or cataracts. The methodology also needs to be continuously revised as new drug treatments emerge.
|
Conclusion |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
The prevalence of chronic conditions, based on pharmacy claims data, in those aged
70 years in Ireland is high, with a significant level of co-morbidity. In a country with little routine collection of primary diagnosis data, estimating disease prevalence using pharmacy claims data may be a pragmatic means of monitoring disease prevalence in an elderly population. The methodology offers the possibility of international comparison of disease prevalence, prescribing and drug costs in managing chronic conditions. |
Key points |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
- There is a lack of accurate prevalence data on chronic diseases in the Irish elderly population.
- Over 85% of this population received regular medications to treat chronic conditions associated with ageing.
- The level of chronic disease and co-morbidity appears much greater than indicated in currently available national data.
- National pharmacy claims databases offer the potential to compare prevalence and treatment of chronic conditions on a regional and international level.
|
Conflicts of interest |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
There are no conflicts of interest.
Department of Pharmacology and Therapeutics, Trinity College Dublin, Dublin 8, Dublin, Ireland
* To whom correspondence should be addressed at: Email: naughtc{at}tcd.ie
|
Acknowledgements |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
We thank the Health Research Board, Ireland, for financial support. We also thank the HSE–PCRS for the data on which this study is based. The HRB or the HSE–PCRS did not influence the preparation of this article.
|
References |
|---|
|
TopMethodsResultsDiscussionStudy limitationsConclusionKey pointsConflicts of interestReferences |
|---|
- Central Statistics Office. Population and Migration Estimates. http://www.cso.ie (2005) (1 December 2005, date last accessed).
- National office Statistics: Women Outlive Men, Life Expectancy Continues to Improve. http://www.statistics.gov.uk (2005) (29 January 2006, date last accessed).
- O’Hanlon A, McGee H, Barker M et al. Health and Social Service for Older People II (HeSSOP II). Dublin: National Council on Ageing and Older People, 2005; 91.
- Garavan R, Winder RMH. Health and Social Services for Older People (HeSSOP I). Dublin: National Council on Ageing and Older People. http://www.ncaop.ie (2001) (1 February 2006, date last accessed).
- Von Korff M, Wagner EH, Saunders K. A chronic disease score from automated pharmacy data. J Clin Epidemiol 1992; 45: 197–203.[CrossRef][Web of Science][Medline]
- Fishman PA, Goodman MJ, Hornbrook MC, Meenan RT, Bachman DJ, O’Keeffe Rosetti MC. Risk adjustment using automated ambulatory pharmacy data: the RxRisk model. Med Care 2003; 41: 84–99.[CrossRef][Web of Science][Medline]
- Gray J, Majeed A, Kerry S, Rowlands G. Identifying patients with ischaemic heart disease in general practice: cross sectional study of paper and computerised medical records. BMJ 2000; 321: 548–50.
[Abstract/Free Full Text] - Maio V, Yuen E, Rabinowitz C et al. Using pharmacy data to identify those with chronic conditions in Emilia Romagna, Italy. J Health Serv Res Policy 2005; 10: 232–8.
[Abstract/Free Full Text] - Sartor F, Walckiers D. Estimate of disease prevalence using drug consumption data. Am J Epidemiol 1995; 141: 782–7.
[Abstract/Free Full Text] - Silwer L, Lundborg CS. Patterns of drug use during a 15 year period: data from a Swedish county, 1988–2002. Pharmacoepidemiol Drug Saf 2005; 14: 813–20.[CrossRef][Web of Science][Medline]
- van de Vijver DA, Roos RA, Jansen PA, Porsius AJ, de Boer A. Estimation of incidence and prevalence of Parkinson’s disease in the elderly using pharmacy records. Pharmacoepidemiol Drug Saf 2001; 10: 549–54.[CrossRef][Web of Science][Medline]
- WHO (World Health Organization). Anatomical therapeutics chemical (ATC) classification index with defined daily doses (DDDs). Geneva: WHO, 2001.
- National cancer registry Ireland. Cancer in Ireland 1994–2001. http://www.ncri.ie (2005) (3 March 2006, date last accessed).
- Adamson J, Lawlor DA, Ebrahim S. Chronic diseases, locomotor activity limitation and social participation in older women: cross sectional survey of British Women’s Heart and Health Study. Age Ageing 2004; 33: 293–8.
[Abstract/Free Full Text] - Ayis S, Gooberman-Hill R, Ebrahim S. Long-standing and limiting long-standing illness in older people: associations with chronic diseases, psychosocial and environmental factors. Age Ageing 2003; 32: 265–72.
[Abstract/Free Full Text] - National Office Statistics. Health Expectancy Living Longer, More Years in Poor Health. http://www.statistic.gov.uk (2004) (29 January 2006, date last accessed).
- Perenboom RJ, Van Herten LM, Boshuizen HC, Van Den Bos GA. Trends in disability-free life expectancy. Disabil Rehabil 2004; 26: 377–86.[CrossRef][Web of Science][Medline]
- EHEMU. Health Expectancy in Ireland. http://www.hs.le.ac.uk/reves/ehemutest (2005) (22 October 2005, date last accessed).
- Rector TS, Wickstrom SL, Shah M et al. Specificity and sensitivity of claims-based algorithms for identifying members of Medicare+Choice health plans that have chronic medical conditions. Health Serv Res 2004; 39: 1839–57.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||