2 General Approach

2.1 Introduction
The approach that we have adopted in the health benefit calculations for the review of the NAQS is consistent with the approach adopted in the report published by the Department of Health's Committee on the Medical Effects of Air Pollutants early in 1998 (Quantification of the Effects of Air Pollution on Health in the United Kingdom, COMEAP, 1998). The results of this type of calculation are a refinement of the crude estimates that might be produced by assuming that all the population is exposed to some national average concentration of pollutants. A method that is equivalent to assessing the population weighted mean concentrations of air pollutants across the country is required because both the concentrations air pollutants and the population density are variable across the country.

The method adopted by COMEAP (1998) can be summarised as follows.

1. The country has been divided into 1 km grid squares and the annual average concentration of pollutants and resident population has been estimated for each square. The former has been derived from the national mapping of the UK pollution climate undertaken at NETCEN and the latter from census data .
   
2. A baseline level of the given health-related and pollution affected events e.g., daily deaths, hospital admissions for the treatment of respiratory diseases has been assigned to each grid square.
   
3. By combining the data from (1) and (2) and applying a coefficient linking pollutant concentrations with the relevant effects the estimated health impact of each pollutant can be calculated for each grid square.
   
4. Summing the results obtained in (3) gives the relevant totals for the UK.
   

The results presented for particles, SO₂ and NO₂ by COMEAP were based on the population in urban areas only because the epidemiological studies on which the dose response coefficients were based, were done in cities. The calculations for ozone included both the urban and rural populations but were performed for the summer only. This was because ozone concentrations are generally higher in rural than in urban areas and concentrations are also higher in the summer. The health benefit calculations for the review of the NAQS have followed these conventions.

2.2 Dose response coefficients
The literature on the effects of air pollutants on health is extensive and was not reviewed for the IGCB report. This was felt to be unnecessary as the COMEAP report had examined the relevant evidence and had produced a series of dose-response coefficients linking concentrations of three major pollutants with effects on health. The exposure (dose)-response coefficients used in the current analysis are the same as those used by COMEAP (1998) and are presented in Table 2.1.

Table 2.1. Dose Response Coefficients

Pollutant	Health Outcome	Dose-response coefficient
PM10	Deaths brought forward (all causes) Respiratory hospital admissions	+ 0.75% per 10 µg/m3 (24 hour mean) + 0.80% per 10 µg/m3 (24 hour mean)
Sulphur dioxide	Deaths brought forward (all causes) Respiratory hospital admissions	+ 0.6% per 10 µg/m3 (24 hour mean) + 0.5% per 10 µg/m3 (24 hour mean)
Ozone	Deaths brought forward (all causes) Respiratory hospital admissions	+ 0.6% per 10 µg/m3 (8 hour mean) +0.7 % per 10 µg/m3 (8 hour mean)
NO2	See note below	See note below
Notes: For NO2 a coefficient of 0.5% per 10 µg/m3 was used to estimate the effect on respiratory hospital admissions in a sensitivity analysis.
Source: COMEAP (1998)

Table 2.1 shows that dose-response coefficients were specified for particulate matter, ozone and sulphur dioxide. The Committee also examined nitrogen dioxide and carbon monoxide but felt that the evidence was not sufficiently strong to allow firm estimates of total effects on health to be made. In the case of nitrogen dioxide, a dose-response coefficient was, however, defined for respiratory hospital admissions and this coefficient can be used for a sensitivity analysis (see EAHEAP (1999) for more information on the implications of this).

The health effects that were considered were daily deaths and admissions to hospital for the treatment of respiratory diseases. In both cases the COMEAP report made clear that the numbers of events calculated as related to exposure to air pollution, could not be simply interpreted as extra events. Deaths are brought forward and hospital admissions may be either brought forward or caused de novo. The extent of advancement of deaths and hospital admissions cannot yet be calculated and estimates from a few days or weeks to a year have been produced. This inability to calculate the extent of advancement of these events is due to the time-series nature of the epidemiological studies upon which the estimates are based.


2.3 Baseline rates for deaths brought forward and respiratory hospital admissions
The baseline rates of deaths brought forward and respiratory hospital admissions used in the current work have been updated from those used in COMEAP (1998). They are listed in Table 2.2 along with the ones used by COMEAP for comparison. The baseline rates that are used should match those used in the studies generating the dose response coefficients. For example, most studies of air pollution and mortality have excluded accidents. Revised baseline rates, as recommended by EAHEAP (1999), have been used in order to take this into account.

Table 2.2. Baseline death rates and respiratory hospital admissions rates per 100,000 people (Figure in brackets are those used by COMEAP, 1998)

Pollutants	Deaths	Respiratory hospital admissions
Particles, SO2 (and NO2)	10741 (1106.4)5	8302 (1342.3)6
Ozone	491.83 (506.8)7	3604 (345)8
Notes 1 deaths excluding external causes, per year, 1995 2 emergency respiratory admissions, per year, 1994/5 3 deaths excluding external causes, per summer (April to September 1995) 4 emergency respiratory admissions, per summer (April to September 1995) 5 deaths including external causes, per year, 1995 6 respiratory admissions, per year, 1994/5 7 deaths including external causes, per summer (April to September 1995) 8 emergency respiratory admissions, per summer (April to September 1993)

2.4 Population statistics
The population statistics used in the COMEAP report were based on the 1981 census and included people living in Great Britain only. The urban population represented within this census data was estimated by assigning areas as urban if the sum of urban and suburban land cover classes within the Land Cover Map of Great Britain (Fuller et al, 1994) for a given 1 km square was greater than 20%. This gave an urban population of 42,500,000. We have used data from the 1991 census for the whole of the UK for the calculations presented in this report and the urban population represented within this dataset is 40,700,000. The change in urban population between the 1981 and 1991 based datasets may be due, at least in part, to differences in the way that the census data has been aggregated from enumeration district to grid square totals.

---