Epidemiology and disease burden from allergic disease in Scotland: analyses of national databases

Abstract

Summary

Background

There are ongoing concerns about the quality of care provided to patients with allergic disorders in Scotland, but there are relatively few reliable data on the overall disease burden. We sought to: (1) describe the incidence, prevalence and outcome of allergic disorders; (2) estimate healthcare burden and costs; and (3) investigate ethnic variations in the epidemiology and outcomes from allergic disorders in Scotland.

Methods

Data sources: national surveys; primary care data; prescribing and medication data; hospital admissions data and mortality data.

Results

Allergic disorders are extremely common in Scotland, affecting about one in three of the population at some time in their lives. Incidence was highest for eczema (10.2 per 1000 registered patients). Over 4% of all GP consultations and 1.5% of hospital admissions were for allergic disorders. There were 100 asthma deaths in 2005 (20 per million people). Direct healthcare costs for allergic disorders were an estimated £130 million per year, the majority of these being incurred in primary care and related to asthma.

Conclusions

Allergic disorders are common in Scotland and given the very high proportion of children now affected, the high disease burden associated with these conditions is likely to persist for many decades.

Introduction

In recent decades, there has been a dramatic increase in the prevalence of a number of allergic disorders in the UK, resulting in considerable challenges to an ill-prepared National Health Service (NHS). 1, 2, 3, 4 Concerns regarding the quality of allergy care were particularly highlighted in the Royal College of Physicians report Allergy: the unmet need, 5 which in turn precipitated parliamentary enquiries into allergy care provision in both England 6 and Scotland. 7, 8

Much of the epidemiological work underpinning the Royal College of Physicians report was based on detailed secondary analyses of available healthcare data-sets, with a focus on those data-sets most relevant to England and Wales. 1 Important data-sets of particular relevance to the Scottish context were, because of time and resource constraints, in the main overlooked. 9

The aim of this study was to describe the epidemiology, morbidity, mortality and economic costs to NHS Scotland of allergic disorders. In seeking to answer these questions, we focused on the following conditions: allergic conjunctivitis; allergic rhinitis; anaphylaxis; angioedema; asthma; drug allergy; eczema/atopic dermatitis; food allergy; urticaria; and allergies not classified elsewhere. Our secondary aim was to investigate ethnic variations in the epidemiology and outcomes from these allergic disorders. In addition, Scotland's position in international allergy rankings was considered.

Methods

Data-sets

Data were collected from routine health information sources and large, high quality national and international surveys. Surveys included the 1995, 1998 and 2003 Scottish Health Surveys (SHS), 10 the 2001 Health Survey for England (HSE) for comparative purposes, 11 and the International Study of Allergies and Asthma in Childhood (ISAAC). 12 Appendix 1 shows questions on allergy used in these surveys. Primary care healthcare data were obtained through the Information Services Division (ISD) of NHS National Services Scotland, for the Practice Team Information programme 13 and the Quality and Outcomes Framework 14 of the General Medical Services contract, from the Primary Care Clinical Informatics Unit, and from the QRESEARCH programme. 15 Data on prescribing were provided by the Prescription Cost Analysis programme 16 (part of ISD) and from over-the-counter sales 17 for prescribing and medication. Data for continuous inpatient stays (the same continuous spell of inpatient treatment) were provided by ISD from the Scottish Morbidity Record (SMR01) 18 and mortality data from the General Register Office for Scotland (GROS). 19 A description of the salient features of these data-sets is given in Appendix 2.

Definitions

Allergic disease definitions were based on those used in the SHS, HSE and ISAAC surveys (Appendix 1), Read codes for primary care data (Appendix 3) and the World Health Organization's (WHO) International Classification of Diseases (ICD 10) codes for SMR01 (Appendix 4).

Statistical methods

Incidence rates from primary care were calculated as the number of new cases of disease diagnosed in a specific year divided by the total number of patients registered with the study practices for that year (to give person years of exposure). These rates were multiplied by 1000 to give rates per 1000 registered patients per year. Lifetime prevalence was estimated from the number of GP patients with a recorded diagnosis of the disease at any point since being registered with a practice divided by the total number of patients registered with the study practices. These rates were standardized by sex and five-year age bands using the estimated mid-year Scottish population (obtained from GROS population estimates). Rates were multiplied by 1000 to give rates per 1000 registered patients. These rates were also standardized by sex and five-year age bands using the estimated mid-year (2004) English population and multiplied by 1000 to give rates per 1000 registered patients to compare with English primary care (QRESEARCH) data. All analyses were undertaken using SPSS v 13.0.

Assessing costs

The cost of GP consultations, hospital inpatient stay and day cases were calculated using unit costs and national estimates for measures of healthcare utilization. 20, 21 Community prescription costs were estimated using ISD prescription data using the standard Defined Daily Dose (DDD) and Gross Ingredient Cost (GIC). The DDD, as defined by the WHO is the assumed average maintenance dose per day for a drug when used for its main indication in adults. 22 The GIC is the cost of an item before any discounts that may be made by the supplier to pharmacies; it does not include dispensing costs or fees. Fees paid by the recipient are also excluded; this equates with the Net Ingredient Cost. 23

Results

Incidence, prevalence and outcomes of allergic disorders

There were 12,210 incident cases of allergy in primary care in 2003–2004. The overall incidence rate of a primary care diagnosis of any allergic disorder was 32 per 1000 registered patients (95% CI 31–33) in 2003–2004 (

Table 1

Incidence of new primary care diagnoses; 95% confidence interval (CI) by gender and diagnosis; rate per 1000 registered patients per year, Scotland 2003–2004 (Source: PCCIU)

Disease	Incidence in men		Incidence in women		Total incidence
	(95% CI)		(95% CI)		(95% CI)
All allergy	27	(26–28)	37	(36–38)	32	(31–33)
Allergic rhinitis	5.4	(5.0–5.7)	6.8	(6.5–5.4)	6.1	(5.9–6.3)
Anaphylaxis	0.02	(0.01–0.03)	0.04	(0.01–0.07)	0.02	(0.01–0.04)
Angio-oedema	0.03	(0.00–0.06)	0.05	(0.02–0.08)	0.04	(0.02–0.06)
All asthma	4.2	(3.9–4.5)	5.4	(5.1–5.7)	4.8	(4.6–5.0)
Allergic asthma	0.04	(0.01–0.07)	0.06	(0.03–0.09)	0.05	(0.03–0.07)
Conjunctivitis	0.84	(0.71–0.97)	1.6	(1.5–1.8)	1.2	(1.1–1.4)
Drug allergy	1.7	(1.5–1.9)	3.2	(2.9–3.5)	2.5	(2.3–2.6)
Eczema	8.8	(8.4–9.2)	11.6	(11.1–12.1)	10.2	(9.9–10.6)
Food allergy	0.17	(0.11–0.23)	0.22	(0.15–0.28)	0.19	(0.15–0.24)
Urticaria	2.7	(2.5–2.9)	4.1	(3.8–4.4)	3.4	(3.2–3.6)
Other allergies	1.7	(1.5–1.9)	3.2	(2.9–3.5)	2.5	(2.3–2.6)

Table 1). Incidence was higher in women (37 per 1000 registered patients) than men (27 per 1000 registered patients). The highest incidence in 2003–2004 was for eczema at 10.2 per 1000 registered patients (95% CI 9.9–10.5) for people of all ages and for both sexes combined. Incidence of eczema was higher in women (11.6 per 1000 registered patients) than men (8.8 per 1000 registered patients). The lowest incidence in 2003–2004 was for anaphylaxis at 0.02 per 1000 registered patients (95% CI 0.01–0.04) for people of all ages and for both sexes combined and was higher in women (0.04 per 1000 registered patients) than men (0.02 per 1000 registered patients). The recording of ethnicity in primary care was insufficient to permit analysis.

During the 12 months ending in March 2006, 8.3% of men and 10.5% of women consulted their practice with any diagnosis of allergic disease (PCCIU, data not shown). The lifetime prevalence for all allergies in Scotland (in 2004), based on general practice consultations, was substantially higher at 350 (95% CI 348–352) per 1000 registered patients, compared to 243 (95% CI 243–245) per 1000 registered patients for England (in 2005) (Appendix 5). Prevalence was higher in Scotland when compared to England for allergic rhinitis, asthma and eczema, but not for anaphylaxis (Table 1, Appendix 5).

Asthma posed a particularly high disease burden. The proportion of respondents aged 16 years or over in the 2003 SHS who reported ever having wheezed was 27% in men and 26% in women (Table 2, Appendix 6). Among adult respondents to the 2003 survey, 13% of men and 14% of women reported that they had ever had asthma diagnosed by a doctor.

The lifetime prevalence of doctor-diagnosed asthma among children under 16 years in 2003 was 20% in boys and 12% in girls. The prevalence of wheeze within the last 12 months was 16% in boys and 12% in girls. The corresponding figures for lifetime wheeze were 29% and 20% (Appendix 6). Although ethnic group is recorded in the SHS the number of respondents from ethnic minorities was too small to allow separate analysis.

The 2002 ISAAC survey of 12–14-year-olds in Scotland (based on self-reports) found an overall prevalence of wheezing in the past 12 months of 28% and the lifetime prevalence of asthma was 26%, similar to the rest of the UK (data not shown). 12 The lifetime prevalence of diagnosed asthma increased slightly from 20% in 1995 to 24% in 2002. 12

The 2002 ISAAC survey in Scotland showed an overall prevalence of a rhinoconjunctivitis in the previous 12 months of 17%, with a lifetime prevalence of hayfever of 36%, slightly lower than the rest of the UK (data not shown). Between 1995 and 2002 there was a slight decrease in the prevalence of rhinoconjunctivitis in the previous 12 months among 12–14-year-olds in Scotland from 19% to 17%. 12

The 2002 ISAAC survey in Scotland showed an overall prevalence of a flexural rash in the past 12 months of 13% while the lifetime prevalence of eczema was 24%, similar to the rest of the UK (data not shown). ISAAC data showed that between 1995 and 2002 there was a considerable decrease in the prevalence of flexural rash in the previous 12 months among 12–14-year-olds in Scotland from 17% to 13%. 12 No data on ethnicity were reported for the Scottish ISAAC data.

The only deaths recorded for allergic disease were for asthma. There were 100 asthma deaths (ICD10 J45-J46) in Scotland in 2005, these occurring at a rate of approximately 20 per million of the population. There were no recorded deaths from any other allergic problem.

Health service utilization

Primary care

There were 60,553 GP consultations for allergy in 2003–2004. The consultation rate for all allergies was 39 (95% CI 39–39) per 1000 registered patients in 2003–2004 (

Table 2

GP consultation rates per 1000 GP consultations per year, (95% confidence intervals, source, PCCIU) Scotland 2003–2004

Disease	GP consultations per 100,000 per year (95%CI)
All allergic conditions	39 (39–39)
Allergic rhinitis	4.3 (4.3–4.3)
Anaphylaxis	0.03 (0.03–0.003)
Angioedema	0.03 (0.03–0.03)
Asthma	20 (20–20)
Allergic asthma	0.34 (0.34–0.34)
Conjunctivitis	0.53 (0.53–0.53)
Drug allergy	0.72 (0.72–0.73)
Eczema	10 (10–10)
Food allergy	0.16 (0.16–0.17)
Urticaria	1.6 (1.6–1.6)
Other allergy	1.2 (1.2–1.2)

Table 2) which was nearly 4% of all consultations for that year. The allergic conditions for which patients most commonly consulted GPs were asthma (20 [95% CI 20–20] consultations per 1000 registered patients annually, or nearly half of all GP consultations for allergic disease) and eczema (10 [95% CI 10–10] per 1000 registered patients).

Over 7.7 million community prescriptions were dispensed for allergic conditions in Scotland in 2003–2004, 13% of all prescriptions dispensed for that year. These were mainly for eczema (1.8 million) and asthma (1.7 million).

Secondary care

There were 24,189 admissions for all allergic disorders in 2004–2005, accounting for 1.5% of the total admissions for all conditions in that year. Of these, asthma was the commonest reason and accounted for 83% of allergic disease continuous inpatient stays. In the financial year 2004–2005 there were 1455 continuous inpatient stays for other allergic problems (urticaria=286; anaphylaxis=370; atopic dermatitis=245; food allergy=296; allergic rhinitis=93; angioedema=156 and conjunctivitis=9). The highest rates were for asthma in children under 15 (35.5 per 1000,

Table 3

Continuous inpatient stays for allergic conditions in Scotland (rate of admissions per year, per 1000 population) in 2004–2005

Disease	Total	Men	Women	0–14	15–44	45+
All allergic disease	12	10	15	44	15	6
Allergic rhinitis	0.13	0.14	0.12	0.39	0.26	0.3
Anaphylaxis	0.52	0.45	0.60	0.74	0.74	0.39
Angioedema	0.22	0.18	0.27	0.13	0.25	0.22
Asthma	9.9	8.0	12.2	35.5	12.1	0.47
Conjunctivitis	0.01	0.02	0.00	0.07	0.01	0.00
Eczema	0.34	0.35	0.34	1.83	0.36	0.09
Food allergy	0.42	0.42	0.42	0.258	0.26	0.12
Urticaria	0.40	0.36	0.45	1.37	0.53	0.18

Table 3), followed by eczema (1.83 per 1000) in the same age group. Around 10% of hospital discharges included information on ethnic group, insufficient to allow meaningful analysis.

Healthcare costs

GP consultations for allergic problems in Scotland cost the NHS an estimated £1.5 million per year, hospital admissions £10.2 million per year and community prescribed treatments for allergic conditions nearly £120 million (

Table 4

GP consultations, 2003–2004 (source PCCIU), hospital inpatient costs and community prescriptions 2004–2005 (source ISD) for Scotland. Cost in £1000s

Primary/secondary care		Cost (£1000s)
GP	Allergic rhinitis	167
	Anaphylaxis	1
	Angioedema	1
	Asthma	786
	Conjunctivitis	20
	Drug allergy	28
	Eczema	397
	Food allergy	6
	Urticaria	61
	Other allergy	47
	All allergies	1514
Hospital, admissions	Asthma
	Inpatient	2391
	Day case	122
	All allergy excluding asthma Inpatient
	Allergic rhinitis	454
	Anaphylaxis	220
	Angioedema	105
	Conjunctivitis	6
	Eczema	223
	Food allergy	167
	Urticaria	223
	Other allergy	4496
	All allergy excluding asthma Day case
	Allergic rhinitis	175
	Anaphylaxis	2
	Angioedema	1
	Conjunctivitis	1
	Eczema	63
	Food allergy	51
	Urticaria	34
	Other allergy	1486
	Total hospital admissions	10,220
Community prescriptions	Asthma	94,867
	Adrenaline	733
	Topical corticosteroids	6572
	Emollients	5898
	Antihistamines	11,887
	Total community prescriptions	119,9577
	Overall total	131,691

Table 4). Overall, allergic disease costs the NHS over £130 million per year in direct costs. Asthma was the major contributor across all cost domains (Table 4).

Primary care prescribing costs for all conditions amounted to nearly £1 billion, of which prescriptions for all allergic problems accounted for 13% (nearly £120 million) of this budget; this compared to 11% for gastrointestinal problems (nearly £104 million).

Discussion

Main findings of this study

This is the most comprehensive and detailed review of the disease burden posed by allergic disorders in Scotland ever undertaken. We have found that allergic disorders are extremely common, affecting over one in three of the Scottish population at some point in their lives. The majority of these individuals are affected by one or more of the organ-specific allergic disorders, namely eczema, allergic rhinitis and/or asthma. Also, the lifetime prevalence for all allergic disease was higher in Scotland than England, namely for allergic rhinitis, asthma and eczema, but not for anaphylaxis. This analysis also reveals that these conditions are associated with significant costs to NHS Scotland (over £130 million), which are predominantly associated with the provision of asthma care and community-based prescribing.

Strengths and limitations of this work

The main strengths of this work include the use of data from a number of large and representative data-sets, the inclusion of a broad range of allergic conditions enabling us to include possible diagnostic transfer (for example, between angioedema and urticaria), and the use of a range of relevant epidemiological, health services utilization and cost-related outcomes. In addition, we were able to draw on our collective understanding and experience of working with these data-sets, and, more specifically, data relating to allergic disorders, in order to interpret findings.

This work does however have a number of important potential limitations, these in the main relating to the available data sources. These include, most notably, the fact that we were dependent on recorded clinical diagnoses in both primary and secondary care, and also in relation to coding of deaths. This latter issue is, for example, known to be a potential problem in relation to the underestimate of deaths from anaphylaxis, which are often coded as asthma deaths. 24 In addition, the discordant time periods of each data-set made comparisons between data-sets slightly challenging. There are also potential problems in relation to interpreting data on trends as these can be affected by incentives to improve quality of care (for example, the Quality and Outcomes Framework), 14 healthcare changes (for example, bed availability, commissioning priorities and other policy changes), changes in perceptions (for example, greater public awareness of food allergy) and data artefacts.

There are also limitations imposed by information gaps, such as the lack of data on allergic reactions in dentistry, utilization of out-of-hours primary care (NHS24), accident and emergency attendances, outpatient care, inpatient prescribing, some over-the-counter purchases of drugs for eczema and allergic rhinitis in particular, and in many cases regional, socioeconomic and ethnic variations in allergic disease risk and outcomes. As a consequence of this, our estimates of costs to the NHS are likely to be a substantial underestimate.

Considering the findings of this work in relation to the wider literature

Prospective cohort studies are needed to allow a more accurate characterization of the epidemiology of allergic disorders in Scotland. In the absence of such studies, which are of necessity time-consuming and expensive to mount, it is important that secondary analysis of cohorts generated using routine data-sets is undertaken 25 and furthermore that in-depth work is undertaken to assess the validity of routine clinical records as has, for example, been undertaken in relation to anaphylaxis records from the General Practice Research Database (GPRD). 26 Similar work is now needed for hospital-based data-sets. Consideration also needs to be given to including a much broader range of allergy questions in the SHS.

Given the lack of data on ethnicity, it is important to consider bridging this gap using the data linkage techniques recently developed by Bhopal et al., 27 which allow linkage of healthcare records with census ethnicity codes using a probabilistic matching technique. Similarly, to obtain a broader picture of costs to society associated with allergic disorders, it is important to consider, for example, the impact on school and work performance, 28 and this should also prove possible with greater investment in data-linkage techniques. 9

The gaps in recoding of emergency contacts within the NHS – NHS24 and A&E attendance – also need to be filled and this may prove possible in due course when the electronic health record is introduced into NHS Scotland. Outpatient recording by diagnosis has begun but is at an early stage. Given the likely future move to the Systematised Nomenclature of Medicine Clinical Terms (SNOMED-CT) coding system, it is important that serious consideration is given to ways in which allergy codes need to be developed to facilitate secondary analyses of the type undertaken in this study. 29

More specifically, given the relatively rapid changing epidemiology of allergic disorders, it is important to continue to monitor allergic disease trends and use the variation discovered therein for hypothesis generation and to provide the data needed to inform assessments of the feasibility of mounting much needed clinical studies of allergic disease prevention and management.

Conclusions

Allergic diseases are extremely common in Scotland and are responsible for substantial morbidity, healthcare utilization and costs to the NHS. Evidence suggests that the previous increases in allergic disorder prevalence reported throughout the latter half of the 20th century may be stabilizing, but given the very high proportion of children now affected, allergic conditions are likely to represent a major strain on the NHS in Scotland for many years to come. Although Scotland has some excellent data-sets, these are in relation to allergic conditions at present disparate, incomplete and complex to analyse and interpret. Given these limitations, consideration needs to be given to setting up a long-term Working Group to ensure that some of the methodological limitations uncovered by this study can be overcome and maximum use is made from the opportunities associated with introduction of SNOMED-CT and the electronic health record.

Footnotes

DECLARATIONS

Competing interests None declared

Funding The Scottish Government's Chief Scientist Office (CSZ/2/252) funded this work and The University of Edinburgh acted as the sponsor

Ethical approval Not applicable

Guarantors CA and AS

Contributorship CA was the researcher employed on this project and undertook data analysis and led the writing of this report. RG, CS and CF were grant-holders on this project and contributed to data interpretation andcommenting on draft versions of this report. AS conceived the study and was the principal investigator for this work overseeing all aspects of the project and contributed to the editing and writing of this report