The importance of understanding epidemiology in order to inform financial decisions: A lesson from the Scottish Home Oxygen Service

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

Available online at www.sciencedirect.com

Public Health journal homepage: www.elsevier.com/puhe

Original Research

The importance of understanding epidemiology in order to inform financial decisions: A lesson from the Scottish Home Oxygen Service R. Wood, I. Grant, M. Bain* NHS National Services Scotland, Gyle Square, 1 South Gyle Crescent, Edinburgh EH12 9EB, UK

article info

summary

Article history:

Objectives: To ensure that decisions on the future planning of the Scottish Home Oxygen

Received 26 January 2011

Service reflect population needs by examining the epidemiology of the main conditions

Received in revised form

that require home oxygen therapy and trends in their management.

30 March 2012

Study design: Analysis of routinely available vital event and health service data supple-

Accepted 17 July 2012

mented by published literature. Use of linked data to provide person-based analyses.

Available online 14 September 2012

Methods: Consideration of trends in key risk factors, disease incidence, prevalence and mortality for chronic neonatal lung disease, cystic fibrosis, chronic interstitial lung disease

Keywords:

in adults and chronic obstructive pulmonary disease. Examination of trends in manage-

Home oxygen

ment of these conditions including hospital admissions, length of stay and re-admissions.

Epidemiology

Results: The prevalence of all the conditions studied has increased in Scotland over recent

Health service data

years due to a combination of increased incidence, increased survival, more active case

Needs assessment

finding and demographic changes. There have been changes in management with trends

Service planning

towards shorter hospital stays. Conclusions: The clinical need for home oxygen therapy is likely to continue to increase over the next 10e20 years. It will encompass all age groups and a complex range of conditions. Public health needs to be proactive in providing relevant needs assessment information to ensure that planning within financial constraints is appropriately informed on population needs. Crown Copyright ª 2012 Published by Elsevier Ltd on behalf of The Royal Society for Public Health. All rights reserved.

Introduction Home oxygen refers to the provision of an enriched oxygen supply for patients outwith a hospital setting (often via cylinders or an oxygen concentrator), and is used in a variety of ways by a range of patient groups. The potential range of conditions requiring provision of home oxygen is wide, but

a small number of key conditions account for the majority of demand.1e4 Key conditions include (in order of increasing age when usually affected/requiring home oxygen): chronic neonatal lung disease (CNLD), cystic fibrosis (CF), adult chronic interstitial lung disease (CILD) and chronic obstructive pulmonary disease (COPD). The requirement for home oxygen varies by condition. In CF, CILD and COPD, patients can

* Corresponding author. Tel.: þ44 (0) 131 275 6325; fax: þ44 (0) 131 275 7530. E-mail address: [email protected] (M. Bain). 0033-3506/$ e see front matter Crown Copyright ª 2012 Published by Elsevier Ltd on behalf of The Royal Society for Public Health. All rights reserved. http://dx.doi.org/10.1016/j.puhe.2012.07.006

948

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

become oxygen dependent in the latter stages of their disease as lung function declines.5e9 In contrast, premature babies with neonatal respiratory distress progressing to CNLD are oxygen dependent at the start of life.10,11 Most surviving infants eventually ‘outgrow’ their CNLD and oxygen dependence, although the time taken varies widely.12 The demand for and provision of home oxygen in Scotland had been increasing steadily over many years. In 2009, it was felt that this increasing demand was becoming unsustainable in financial terms, and discussions began about finding financial solutions. There was uncertainty about what was driving increasing demand. The prevailing assumption at the time was that the clinical need for home oxygen in the Scottish population would be static or falling due to declining smoking prevalence, hence the increasing provision must, at least partly, be due to overprescribing and inappropriate overprovision. Options for reducing costs were discussed. In order to ensure that future changes to the provision of home oxygen were focused on meeting population needs and maintaining high-quality care, as well as achieving a sustainable financial situation, the Medical Director of NHS National Services Scotland (which supplies part of the home oxygen service) asked the Scottish Public Health Network (ScotPHN) to conduct a national needs assessment. Needs assessment is a systematic process of gathering and making sense of information on the needs of a particular population for a particular aspect of health care. The overall

aims of needs assessment are to inform planning processes and ensure that services provided are congruent with the population’s needs.13 The Scottish home oxygen needs assessment analysed a range of information relating to: the epidemiology of the four conditions outlined above, the effectiveness of home oxygen therapy, trends in home oxygen service provision in Scotland, user and provider views on the Scottish Home Oxygen Service, and comparative levels of home oxygen provision elsewhere in the UK. The epidemiological information was critical in testing assumptions about trends in the clinical need for home oxygen. This paper sets out the approach taken within the needs assessment to gathering epidemiological data, presents selected results, and reflects on how the findings contributed to the overall needs assessment and subsequent development of the Scottish Home Oxygen Service.

Methods Gathering of data on the four conditions focused on analysis of routinely available health service and vital event data supplemented by review of the published literature. The analyses conducted for each of the conditions, and the data sources used, are summarized in Table 1. Trends in key risk factors for CNLD (prematurity) and COPD (smoking) were considered. CILD does not have a single over-riding risk

Table 1 e Analyses conducted relevant to the epidemiology and management of four key conditions requiring home oxygen provision. Analysis Chronic neonatal lung disease (CNLD) Premature births 1989e2008 Incidence of CNLD 1989e2008 Mortality from CNLD 1989e2009 Average length of stay for CNLD 1989e2008 Cystic fibrosis (CF) Incidence of CF 2004e2009 Prevalence of CF 1989e2009 Mortality from CF 1989e2009 Admissions for CF 1989e2009 Average length of stay for CF 1989e2009 Re-admission within 1 or 12 months of discharge with CF 1989e2009 Chronic interstitial lung disease (CILD) Prevalence of CILD 1989e2009 Mortality from CILD 1989e2009 Admissions for CILD 1989e2009 Average length of stay for CILD 1989e2009 Re-admission within 1 or 12 months of discharge with CILD 1989e2009 Chronic obstructive pulmonary disease (COPD) Prevalence of smoking 1974e2008

Prevalence of COPD 1989e2009 Mortality from COPD 1989e2009 Admissions for COPD 1989e2009 Average length of stay for COPD 1989e2009 Re-admission within 1 or 12 months of discharge with COPD 1989e2009

Data source Birth registrations and maternity hospital discharge data Special care baby unit discharge data Death and birth registrations Special care baby unit discharge data Neonatal bloodspot screening data and birth registrations Linked hospital discharge and death data Death registrations and mid-year population estimates Linked hospital discharge data Linked hospital discharge data Linked hospital discharge data

Linked hospital discharge and death data Death registrations and mid-year population estimates Linked hospital discharge data Linked hospital discharge data Linked hospital discharge data

General Household Survey, Scottish Schools Adolescent Lifestyle and Substance Use Survey, and Scottish Health Survey Linked hospital discharge and death data Death registrations and mid-year population estimates Linked hospital discharge data Linked hospital discharge data Linked hospital discharge data

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

factor,14e16 and CF has a genetic aetiology. For each condition, trends in disease incidence and/or prevalence, and mortality since 1989 were examined. Trends in the management of all the conditions were also examined through analysis of hospital admissions, average length of stay and re-admission rates. Absolute numbers and rates were calculated by age group and sex wherever possible. The resulting epidemiological and disease management information was used to build understanding of the past (and hence likely future) trends in need for home oxygen services within the Scottish population. Information on prevalence is of particular interest in needs assessment as it reflects the actual number of people in the population who may require services. The Information Services Division of NHS National Services Scotland maintains a linked database which holds all general hospital discharge records since 1981 for an individual, together with their death registration record if relevant. This linked dataset was used to estimate prevalence rates of CF, CILD and COPD by age and sex for each year from 1989 to 2009. Individuals were counted as a prevalent case in any one year if they had at least one prior admission for the condition of interest since the start of the linked database in 1981, and were still alive on 1 July of the year in question. In each year, any individual was counted once in the age and sex group relevant to them at that time. The linked database was also used to assess length of stay and re-admission rates for CF, CILD and COPD. A hospital discharge record is generated each time an individual is discharged from one episode of hospital care to another, hence one admission to hospital could generate multiple records if the patient requires transfer between different hospitals, facilities (e.g. high dependency unit and ward) or consultant teams. Using the linked file allows analysis of complete admissions, and hence meaningful analysis of overall length of stay, and also re-admissions for the same condition within a specified period of time following discharge home. Further details on the analyses conducted, including the ICD-9 and ICD-10 codes used to identify the conditions of interest, are available on request.

Results Chronic neonatal lung disease CNLD occurs predominantly in babies born prematurely. The proportion of births that are premature (<37 completed weeks of gestation) has increased over time, from around 6.5% in the late 1980s to almost 8.0% in 2008. The proportion of births that are extremely premature (<28 weeks of gestation) has also increased over that time period from around 0.35% to around 0.5%. The absolute number of babies born in Scotland at <28 weeks of gestation (the group at highest risk of CNLD) has fluctuated over time, but overall has increased from 209 in 1989 to 284 in 2008. In international studies, the incidence of CNLD amongst surviving premature infants has been found to range from 12% to 35%, with incidence increasing with increasing prematurity.11,17e20 In a UK observational study of babies weighing <1250 g, incidence of CNLD increased between 1987 and 1992. The study suggests that the risk of

949

developing CNLD may then have fallen between 1992 and 1997, but the change did not reach statistical significance.21 A recent US study, based on admissions to hospitals, suggests that the absolute incidence of CNLD fell 3.3% (P ¼ 0.0009) annually between 1993 and 2006.22 Trends in the number of babies treated in Scottish neonatal units for CNLD show that, over recent years e when the number of premature births has been increasing e the number of babies with CNLD has been relatively constant at approximately 45 per year. The number of babies in Scotland dying from neonatal CNLD, and the mortality rate, has shown a clear downward trend over the 1989e2009 time period from around 17 to three children per year. Trends in the average total length of stay for babies with CNLD in Scotland for the period 1996e2002 show overall length of hospital stay for babies falling substantially, from around 95 days to 57 days. More recent data suggest that this decline may now have plateaued.

Cystic fibrosis The incidence of CF in Scotland based on neonatal screening data has fluctuated from year to year due to the small numbers involved, but overall is broadly in line with the expected incidence from the literature of around 25 cases per year or 4 per 10,000 births. Prevalence estimates for CF are shown in Fig. 1 for illustration. The estimated prevalence has almost doubled over the period of study, from 602 people living with CF in 1989 to 1169 in 2009, with almost all the increase being due to increasing numbers of people aged 15 years, reflecting the recognized improvements in survival of CF patients achieved over the last 20 years.23,24 An analysis of an independent disease register suggested that there were around 750 people living with CF in Scotland in 2003,25 comparable to the estimates derived from the routine data. Mortality from CF fell over the early 1990s, although the decline appears to have plateaued over recent years. The average age at death has increased substantially over the past 20 years. In keeping with the prevalence data shown above, deaths in people aged <15 years are now unusual, but survival beyond 45 years remains uncommon; most deaths occur in the 15e34-years age group. The number of hospital admissions for CF has increased steadily over time. The average length of stay for CF admissions fell substantially over the early 1990s, but has been relatively constant over recent years. Twelve-month re-admission rates have been consistently very high for CF patients at over 80%.

Chronic interstitial lung disease Prevalence estimates derived from hospital admission and death data suggest that the overall prevalence of CILD has increased substantially over the past 20 years, from 1933 cases in 1989 to 6394 cases in 2009. All adult age groups have shown an increase in prevalence, but the oldest age groups have seen the highest absolute and relative increases in numbers of cases and prevalence rates. Consistent data on trends in the overall incidence and prevalence of CILD are difficult to find in the literature, and problems with inconsistent diagnostic classifications have

950

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

1,500 1,400 1,300 1,200

Number of prevalent cases

1,100 1,000 45+

900

35-44

800

25-34

700

15-24 5-14

600

0-4

500 400 300 200 100 0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year

Fig. 1 e Prevalent cases of cystic fibrosis, all Scotland by year and age group. Data source: ISD linked database.

been noted.26 Trends in particular subtypes of CILD can show different patterns. There is some evidence that incidence of and deaths from idiopathic pulmonary fibrosis have increased over recent years in the UK,7,14,27 whilst incidence of sarcoidosis has remained relatively constant.14 A study based on new cases of all forms of CILD identified through outpatient or hospital admission data in Denmark suggested that the overall incidence of CILD had increased by around 33% between 1995e2000 and 2001e2005,28 broadly similar to the increase in prevalence over that time period shown above. The number of admissions for CILD has more than doubled over the past 20 years to its current level of more than 2000 per year. The average length of stay for CILD admissions has fallen substantially over the period examined from around 18 to 8 days. Re-admission rates have increased slightly, with over one-third of patients now re-admitted due to their CILD within a year of discharge.

Chronic obstructive pulmonary disease Smoking is by far the most common cause of COPD,29 with at least 80% of COPD deaths thought to be attributable to smoking.30,31 There can be a long lag time of up to several decades between starting smoking and developing clinically apparent COPD. Overall, smoking rates in men (aged 16 years) have declined progressively from their peak at the end of the Second World War. At that time, over 60% of adult men smoked cigarettes and around a further 20% smoked tobacco in other forms (e.g. pipes). Cigarette smoking rates in women were relatively stable at around 40% from the end of the War to around 1970, and have subsequently declined progressively. In both men and women, the greatest decline in smoking prevalence occurred over the 1970s and 1980s, with the rate of decline slowing progressively since then.32

Prevalence estimates from the linked dataset suggest that the prevalence rate of COPD has actually increased over recent years, particularly in the oldest age groups. This, in conjunction with increasing numbers of older people, means that the estimated number of people living with COPD severe enough to have required at least one admission to hospital has increased substantially from 31,149 cases in 1989 to 83,314 cases in 2009. As yet, there is no evidence of this increase in prevalence slowing down. The increase in prevalence has been more marked in women than men. In 1989, 44% of prevalent COPD cases were women; in 2009, this had increased to 53%. COPD is the fifth leading cause of death in the UK and fourth worldwide.31 The number of deaths in Scotland from COPD increased slightly over the 1990s, but has been relatively constant over the last 10 years. In total, 5548 deaths were attributed to COPD in 2009. The increase in the number of deaths over the 1990s was predominantly due to an increase in deaths in women and in the oldest age groups. COPD is one of the most common causes of hospital admission in Scotland.8 The number of admissions for COPD has increased persistently over the past 20 years, and is now around 29,000 per year. The average length of stay for COPD admissions has fallen dramatically over the past 20 years, and both short and longer term re-admission rates have risen as shown in Fig. 2.

Discussion Routine data were analysed and a literature review was undertaken to explore trends in the epidemiology and management of four conditions that account for a high proportion of patients requiring home oxygen. This process provided insight into the clinical need for home oxygen in the

951

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

25

100

20

80 70

15

60 50

10

40 30

5

20 10

0

% of patients readmitted for same condition within specified period

Average length of stay per admission (days)

90

0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Average length of stay

% patients readmitted within 1 month

% patients readmitted 12 months

Fig. 2 e Chronic obstructive pulmonary disease: average length of stay and percentage of patients readmitted within 1 or 12 months of discharge, all Scotland by year. Data source: ISD linked database.

Scottish population, and made an important contribution to the overall needs assessment for this service. There are more premature babies being born, but due to improvements in the ventilation of very small babies and other advances, the number of babies developing CNLD is relatively constant and the proportion of babies with CNLD that survive is increasing. The average age at which babies with CNLD are discharged home has decreased markedly. The evidence shows that the incidence of CF is relatively constant but survival is increasing, hence the number of people living with CF is also increasing. There is some evidence that further improvements in survival of CF patients can be anticipated as Scotland still has higher mortality amongst CF patients than other high-income countries such as England and Wales and the USA.33 It is likely that some of the additional years of life that have been gained by CF patients are relatively healthy and independent years, but equally, an unknown proportion may be additional years lived with severe disease that requires support such as home oxygen. For CILD, there is unresolved debate over how much of the apparent rise in incidence and prevalence over recent years is due to genuine increase in disease, and how much can be explained by improved awareness, better diagnosis and more Overall, precise diagnostic classification/coding.28,34,35 however, there is evidence of upward trends in the prevalence of, admission for and mortality from at least some forms of CILD over the last 20 years. For COPD, there is evidence that the number of recognized prevalent cases and the total number of admissions continue to increase, although mortality from COPD has been relatively stable over the last 10 years. COPD has become equally common in women and men, and more common in older people. These trends reflect complex interactions between historical smoking patterns, population ageing and changes in diagnostic practice.

The four conditions studied vary greatly in the number of people they affect; for example, fewer than 100 babies develop CNLD in Scotland every year, whereas it is estimated that around 80,000 people are living with COPD that is sufficiently severe for them to have required at least one admission to hospital. However, the available evidence suggests that the prevalence of each of the conditions studied has increased in Scotland over recent years. A number of reasons lie behind the increases in prevalence: increased incidence (more new cases), increased survival (people living longer with their disease), more active case finding/more accurate diagnosis/ recording, and demographics (more premature babies being born and more older people in the population). These increases in prevalence are likely to continue at least over the next 10e20 years. In the longer term, the prevalence of disease primarily caused by smoking, in particular COPD, may stabilize or even start to fall. The disease management information shows a consistent pattern across the four conditions, with a trend towards considerably shorter (and in some cases, somewhat more frequent) hospital admissions. Patients are therefore increasingly being cared for in the community, which has implications for the provision of home-based care such as domiciliary oxygen. Taken together, epidemiological and disease management information suggests that the clinical need for home oxygen therapy will increase at least over the short to medium term, and that the range of patients served will continue to encompass all age groups, including the very young and very old. The analyses undertaken for the needs assessment relied heavily on routine data, particularly the use of linked hospital discharge and death data, in estimating prevalence and understanding changing patterns of hospital-based care for certain conditions. Estimating prevalence from linked

952

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

discharge and death data is useful for conditions such as CF, CILD and COPD that are likely to require hospital admission and, once contracted, are likely to be lifelong. It is worth emphasizing, however, that only individuals with disease severe enough to have required at least one admission will be counted as a prevalent case; individuals with early-stage disease still to be diagnosed or being managed in the community will not be counted. Scotland’s hospital discharge data are known to be of high completeness and accuracy36 but, as for all routine data, they are not error free. Some miscoding of the diagnosis that necessitated hospital admission does occur. When discharges are miscoded to or from a condition of interest, prevalence estimates for the condition derived from the data will be erroneously inflated or deflated. To minimize (but not eliminate) this error, only first admissions for CF occurring when an individual was <25 years of age, and only first admissions for CILD or COPD occurring when an individual was aged 25 years were included as indicating a prevalent case. Prevalence estimates derived from the routine data were compared with what would be expected from the literature where possible and, as noted, high levels of agreement were found. In general, therefore, the routine data proved themselves to be amply fit for the needs assessment purposes, and their analysis offered a pragmatic, timely and cost-effective way to provide a wide range of information. When these epidemiological findings were considered along with other aspects of the needs assessment, it was clear that the initial assumptions about declining need and overprovision were largely unfounded, and any planning based on reducing demand in the foreseeable future would be unsound. Without the detailed epidemiological investigation, these assumptions could certainly have led to wrong decisions being made. The home oxygen service in Scotland is now being redesigned with a focus on both reducing costs and improving quality of care for patients, with the needs assessment findings integral to the redesign planning. Public health must be proactive in asking the epidemiological questions, even when (or perhaps especially when) others have not considered this. In times when finances will increasingly drive decision-making, this is an important lesson.

Acknowledgements The authors wish to thank the analytical staff within the Information Services Division of NHS National Services Scotland and the Scottish Newborn Screening Laboratory for extraction and analysis of data.

Ethical approval None sought.

Funding No specific funding was used for this study but the Scottish Public Health Network provided support for the overall preparation and publication of the needs assessment report, which includes the findings reported in this paper. The Scottish

Public Health Network also reimbursed the Scottish Public Health Observatory for the time taken for the literature review for the needs assessment report.

Competing interests None declared.

references

1. National Health Service Executive. Burdens of disease: a discussion document. Leeds: Department of Health; 1996. 2. Sullivan SD, Ramsey SD, Lee TA. The economic burden of COPD. Chest 2000;117:5Se9. 3. Balfour-Lynn IM, Field DJ, Gringras P, Hicks B, Jardine E, Jones RC, et al. BTS guidelines: BTS guidelines for home oxygen in children. Thorax 2009;64:ii1e26. 4. Primhak RA, Hicks B, Shaw NJ, Donaldson GC, BalfourLynn IM. Use of home oxygen for children in England and Wales. Arch Dis Child 2011;96:389e92. 5. Douglass H, Potter H, Jarad N. Current practice in prescription, assessment and use of oxygen therapy in cystic fibrosis: a national UK survey (abstract). J Cystic Fibros 2008;7:S77. 6. TurnereWarwick M, Burrows B, Johnson A. Cryptogenic fibrosing alveolitis: clinical features and their influence on survival. Thorax 1980;35:171e80. 7. Gribbin J, Hubbard RB, Le Jeune I, Smith CJ, West J, Tata LJ. Incidence and morality of idiopathic pulmonary fibrosis and sarcoidosis in the UK. Thorax 2006;61:980e5. 8. Audit Scotland. Managing long-term conditions. Edinburgh: Audit Scotland. Available at: [last accessed 1.10.10], http:// www.auditscotland.gov.uk/docs/health/2007/nr_070816_ managing_long_term.pdf; 2007. 9. Calverly PMA. Supplementary oxygen therapy in COPD: is it really useful? Thorax 2000;55:537e8. 10. Banks-Randall BA, Ballard RA. Bronchopulmonary dysplasia. In: Taeusch HW, Ballard RA, Gleason CA, editors. Avery’s diseases of the newborn. 8th ed. Philadelphia: Saunders; 2005. Ch. 49, p. 723e736. 11. Tapia JL, Agost D, Alegria A, Standen J, Escobar M, Grandi C, et al. NEOCOSUR Collaborative Group. Bronchopulmonary dysplasia: incidence, risk factors and resource utilization in a population of South American very low birth weight infants. J Pediatr 2006;82:15e20. 12. Patrinos ME. Chronic lung disease of infancy (bronchopulmonary dysplasia). Chapter VIII.4. In: Yamamoto LG, Inaba AS, Okamoto JK, Patrinos MP, Yamashiroya VK, editors. Case based pediatrics for medical students and residents. Honolulu: Department of Pediatrics, University of Hawaii, John A. Burns School of Medicine [last accessed 1.10.10] Available at: http://www.hawaii.edu/ medicine/pediatrics/pedtext/s08c04.html; 2002. 13. Cavanagh S, Chadwick K. Health needs assessment: a practical guide. London: National Institute for Health and Clinical Excellence; 2005. 14. Hubbard R, Johnston I, Britton J. Survival in patients with cryptogenic fibrosing alveolitis: a population based cohort study. Chest 1998;113:396e400. 15. Hubbard R, Venn A, Britton J. Exposure to antidepressants and the risk of cryptogenic fibrosing alveolitis: a caseecontrol study. Eur Respir J 2000;16:409e13. 16. Baumgartner KB, Samet JM, Coultas DB, Stidley CA, Hunt WC, Colby TV, et al. Occupational and environmental risk factors for idiopathic pulmonary fibrosis: a multicenter

p u b l i c h e a l t h 1 2 6 ( 2 0 1 2 ) 9 4 7 e9 5 3

17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

caseecontrol study. Collaborating centers. Am J Epidemiol 2000;152:307e15. Egreteau L, Pauchard JY, Semama DS, Matis J, Liska A, Romeo B, et al. Chronic oxygen dependency in infants born at less than 32 weeks’ gestation: incidence and risk factors. Pediatrics 2001;2:26. Kamper J, Feilberg-Jorgensen N, Jonsbo F, PedersenBjergaard L, Pryds O. The Danish national study in infants with extremely low gestational age and birthweight (the ETFOL study): respiratory morbidity and outcome. Acta Paediatr 2004;2:225e32. Hentschel J, Berger TM, Tschopp A, Mu¨ller M, Adams M, Bucher HUSwiss Neonatal Network. Population-based study of bronchopulmonary dysplasia in very low birth weight infants in Switzerland. Eur J Pediatr 2005;164:292e7. Larroque BP, Ancel PY, Marret S, Marchand L, Andre´ M, Arnaud C, et al. EPIPAGE Study Group. Neurodevelopmental disabilities and special care of 5-year-old children born before 33 weeks of gestation (the EPIPAGE study): a longitudinal cohort study. Lancet 2008;371:813e20. Manktelow BN, Draper ES, Annamalai S, Field D. Factors affecting the incidence of chronic lung disease of prematurity in 1987, 1992, and 1997. Arch Dis Child Fetal Neonatal Ed 2001;85:33e5. Stroustrup A, Trasande L. Epidemiological characteristics and resource use in neonates with bronchopulmonary dysplasia: 1993e2006. Pediatrics 2010;126:291e7. Dodge JA, Morison S, Lewis PA, Coles EC, Geddes D, Russell G, et al. Incidence, population and survival of cystic fibrosis in the UK, 1968e95. Arch Dis Child 1997;77:493e6. Dodge JA, Lewis PA, Stanton M, Wilsher J. Cystic fibrosis mortality and survival in the UK: 1947e2003. Eur Respir J 2007;29:522e6. Gray R. Healthcare services for children with cystic fibrosis. Report to the Chief Medical Officer of Scotland. In: Scotland e an evaluation of current provision. Edinburgh: Scottish Government; 2003. Demedts M, Wells AU, Anto´ JM, Costabel U, Hubbard R, Cullinan P, et al. Interstitial lung diseases: an epidemiological overview. Eur Respir J 2001;18:2se16.

953

27. Johnston I, Britton J, Kinnear W, Logan R. Rising mortality from cryptogenic fibrosing alveolitis. BMJ 1990;301: 1017e21. 28. Kornum JB, Christensen S, Grijota M, Pedersen L, Wogelius P, Beiderbeck A, et al. The incidence of interstitial lung disease 1995e2005: a Danish nationwide population-based study. BMC Pulmon Med 2008;8:24. 29. Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. GOLD executive summary. Am J Respir Crit Care Med 2007;176:532e55. 30. NHS Health Scotland, ScotPHO. An atlas of tobacco smoking in Scotland: a report presenting estimated smoking prevalence and smoking attributable deaths within Scotland. Edinburgh: NHS Health Scotland; 2007. 31. National Clinical Guideline Centre. Chronic obstructive pulmonary disease: management of chronic obstructive pulmonary disease in adults in primary and secondary care. London: National Clinical Guideline Centre; 2010. 32. ScotPHO. Chronic obstructive pulmonary disorder. Available at: http://www.scotpho.org.uk/home/Healthwellbeinganddisease/copd/copd_data/copd_mortality.asp [last accessed 1.09.10]. 33. Fogarty A, Hubbard R, Britton J. International comparison of median age at death from cystic fibrosis. Chest 2000;117:1656e60. 34. Wells AU, Hirani NBritish Thoracic Society Interstitial Lung Disease Guideline Group. Interstitial lung disease guideline: the British Thoracic Society in collaboration with the Thoracic Society of Australia and New Zealand and the Irish Thoracic Society. Thorax 2008;63(Suppl. V):v1e58. 35. Raghu G, Nyberg F, Morgan G. The epidemiology of interstitial lung disease and its association with lung cancer. Br J Cancer 2004;91(Suppl. 2):S3e10. 36. ISD. An assessment of SMR01 and associated data 2004e2006 Scotland report. Edinburgh: ISD Scotland NHS National Services Scotland/Crown ; 2007.