ARTICLE IN PRESS

0360-1323/$ - se

doi:10.1016/j.bu

�Correspondand Energy, De

of Denmark, N

Tel.: +4545 25

E-mail addr

Building and Environment 43 (2008) 1647–1657

www.elsevier.com/locate/buildenv

Is the use of particle air filtration justified? Costs and benefits offiltration with regard to health effects, building cleaning

and occupant productivity

Gabriel Bekoa,b,�, Geo Clausena, Charles J. Weschlera,c

aInternational Centre for Indoor Environment and Energy, Department of Mechanical Engineering, Technical University of Denmark,

Nils Koppels Alle, Building 402, 2800 Lyngby, DenmarkbFaculty of Civil Engineering, Department of Building Services, Slovak University of Technology, Radlinskeho 11, 81368 Bratislava, Slovakia

cEnvironmental and Occupational Health Effects Institute (UMDNJ-RW Johnson Medical School and Rutgers University), Piscataway, NJ 08854, USA

Received 28 February 2007; received in revised form 10 October 2007; accepted 15 October 2007

Abstract

Estimates of costs and the corresponding benefits of particle filtration have been derived for a standard office building. Reduction inoccupants’ exposure to particles during their workday is anticipated to reduce their morbidity and mortality. Filtration may also reducethe costs associated with building and HVAC cleaning. Conversely, losses of occupant productivity due to sensory offending pollutants

emitted from used ventilation filters can lead to significant economic losses. The results of the present analysis are strongly dependent onseveral key input parameters; consequently, the sensitivity of the results to these parameters was evaluated as part of this study. Thestudy also acknowledges that the benefits-to-costs ratio depends on the perspective of the stakeholder: the employer renting the building

is impacted by occupant performance and building energy costs; the building owner is impacted by maintenance of the building and itsHVAC system; society is impacted by the employees’ health and welfare. Regardless of perspective, particle filtration is anticipated tolead to annual savings significantly exceeding the running costs for filtration. However, economic losses resulting from even a small

decrease in productivity caused by sensory pollutants emitted from used ventilation filters have the potential to substantially exceed theannual economic benefits of filtration. Further studies are required to determine if meaningful benefits can be obtained from morefrequent filter replacement or application of different filtration techniques that limit the emission of offending pollutants into theventilation air.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Cost–benefit evaluation; Particle filtration; Used filters; Productivity; Morbidity; Mortality

1. Introduction

Numerous epidemiological studies [1–4] have reportedassociations between outdoor airborne particles and bothmorbidity and mortality [5]. Most of our exposure to suchparticles occurs inside buildings. Outdoor air is a sig-nificant and often dominating source of indoor particles,especially in mechanically ventilated structures such as

e front matter r 2007 Elsevier Ltd. All rights reserved.

ildenv.2007.10.006

ing author. International Centre for Indoor Environment

partment of Mechanical Engineering, Technical University

ils Koppels Alle, Building 402, 2800 Lyngby, Denmark.

40 18; fax: +4545 93 21 66.

ess: gb@mek.dtu.dk (G. Beko).

offices and schools [6–10]. In such premises supply airfilters reduce the outdoor-to-indoor transport of particlesand the consequent concentration of indoor particles ofoutdoor origin [10–13].Air filtration has the potential to improve the health of

occupants, and thus reduce productivity loss. Fisk andRosenfeld [14] have estimated that, when the filtrationsystem in an office building is upgraded with more efficientfilters, the financial benefits resulting from an improvedindoor environment may exceed the costs of filtration by asmuch as a factor of twenty. Although more efficient filtershave higher energy penalties and total costs, these tend to benegligible relative to salaries, rent or health insurance costs[12]. Improved particle filtration can also be cost-effective in

ARTICLE IN PRESS

Table 1

Input data used in the analysis

Parameter Assumed value

Total building ventilation rate 10m3/s

Outdoor air supply rate per

occupanta10L/s/person

Occupancya 0.07 occupant/m2

Number of occupants 1000

Floor area of building 14,300m2

Average annual outdoor PM10 level 30 mg/m3

HVAC operating hours 3000 h/yr ( ¼ 12 h/day,

250 days/yr)

Occupants’ time at work 8 h/day; 250 days/yr

Face velocity in HVAC unit 2.5m/s

Initial pressure drop across filters 100 Pa

Final pressure drop across filters 200 Pa

Fan total efficiency 70%

Motor efficiency 65%

Cost of electricity $0.15/kWh

Filter life 3000 h

Average daily salary per occupantb $184

Average annual salary per occupantb $46,000

Discount rate 3%

aFrom CEN CR 1752 [24].bBased on the average hourly office worker salary in the US in 2005,

which was $23 [26].

G. Beko et al. / Building and Environment 43 (2008) 1647–16571648

terms of reduced failures in telephone switching andcomputing equipment [15]. Additional benefits may accruefrom the protection of the individual components of HVACsystems, which is often the primary reason filtration systemsare installed.

One factor that has been neglected in previous economicevaluations of filtration is the potential release of sensorypollutants from used filters. Loaded filters have been foundto be a serious source of sensory pollutants [16–20], withthe potential to contribute to symptoms characteristic ofsick-building syndrome (SBS) [21] and a consequentdecrease in occupant performance [22,23]. Even a smalldecrease in productivity leads to meaningful economicallosses.

The purpose of the present paper is to present a broadevaluation of the economic costs and benefits of standardparticle filtration in office buildings. We recognize that formany of the parameters in this evaluation, the availableinput data are imprecise and limit the accuracy of theresulting estimates. However, our intent is to provide asense of the more important parameters affecting aggregatecosts and benefits associated with building air filtration,leading to more informed decisions when choosing afiltration strategy in office buildings.

2. Approach

A standard office building was used as the setting for thefollowing estimates of economic costs and correspondingbenefits of a single-pass particle filtration system. Westarted with the assumption that the building contained1000 occupants; the size of the building and the airhandling system was then determined based on a designcriteria of 10L/s/occupant of outdoor supply air and anoccupant density of 14.3m2/occupant [24]. The resultingtotal airflow, absent recirculation, is 10m3/s. Operatingwithout recirculated air is typical in northern Europe andwas chosen to keep the model relatively simple.

We assumed yearly replacement of F7/EU7 filters. Thisfilter efficiency is the minimum recommendation ofEN13779 [25] for one-pass filtration. The reference condi-tion in the present analysis was an identical building withno filtration system. The direct costs associated with airfiltration include both initial costs and annual runningcosts. An indirect cost is lost productivity due to SBSsymptoms caused by pollutants from soiled filters. Benefitsfrom filtration include reduced morbidity and mortality forbuildings occupants resulting from lower indoor levels ofparticles that originated outdoors. An additional benefit offiltration is a decreased soiling rate and less frequentcleaning of the occupied space and of the HVAC ducts.

The results of the analysis have been expressed in 2005US dollars per standard year, using a yearly discount rateof 3%, where applicable. The standard year assumes thatno additional investments have to be made for obtainingand installing a new ventilation/filtration system (i.e.,adding a filter rack or auxiliary equipment). Additionally,

the results have been normalized in US dollars peroccupant. Some of the input parameters are impreciseand thus are sources of uncertainty. Therefore, wherepossible, we have performed parametric investigationswithin a relevant range of the input parameters. The basicinput data used for the calculations are listed in Table 1.

3. Methods and results

3.1. Estimate of costs

The costs associated with a filtration system include theinitial purchase of the filters, racks and fans, as well as theannual costs for energy and maintenance. We have notincluded the cost for filtration housing and the incrementalcost of a larger central air unit to handle the increasedstatic pressure due to the presence of filters as well asincreased cooling coil capacity to address additional heatfrom the larger air handler. We have simply focused on theannual running costs and compared them with the annualbenefits. The annual running costs associated with fibrousbag filters include their replacement (new filters, labor,disposal of old filters) and power consumption resultingfrom the energy needed to move air through the filters.We have assumed that the building in this analysis uses

0.6� 0.6m EU7 bag filters. Given the total air flow rateand the 2.5m/s face velocity of air passing through thefilters, we calculated the cross-sectional area of the filterbank (4m2) and, thus, the number of filters needed (11filters). We adjusted this number to a value that wasconsistent with a realistic filter bank (12 filters). The cost

ARTICLE IN PRESSG. Beko et al. / Building and Environment 43 (2008) 1647–1657 1649

input data were taken from Camfil Farr’s publicly availableLife Cycle Cost software [27], which is based on theEurovent document [28]:

�

Tab

An

Item

En

Re

Re

Dis

To

To

cost of filter: $80/filter;

� labor for filter replacement: $12/filter; � filter disposal: $5/filter.

The power consumption needed to overcome the pressuredrop across the installed filters increases as the filtersbecome more loaded. The average pressure drop over thefilter’s lifetime was determined from a linear approxima-tion of the pressure drop increase over time. The life spanof the filters as well as their initial and final pressure drop,were estimated according to Hangstrom [29] for an urbanenvironment. From the corresponding air power (airflowrate multiplied by average pressure drop) and the given fanand motor efficiencies, the required fan power wascalculated:

Fan power ¼ðAirflow rateÞ � ðAverage pressure dropÞ

Total fan and motor efficiency.

(1)

Finally the price of power consumption per filter lifetimewas obtained:

Energy cost ¼ ðFan powerÞ � ðOperating hoursÞ

� ðElectricity priceÞ. ð2Þ

For the type of comparison we are trying to make in thispaper, it is sufficient if the annual cost of air filtration(Table 2) represent one time/1 yr estimates in 2005 prices.We judge it to be an unnecessary refinement to determinean average of all annual expenses over a longer time period(HVAC system lifetime) corrected to the present valuethrough the respective discount rate.

3.2. Estimate of benefits

3.2.1. Health endpoints

Numerous studies have demonstrated associations be-tween particulate air pollution and adverse health effects.Earlier epidemiological studies were based on data fromPM10 monitoring stations, while more recent epidemiolo-gical studies have focused on PM2.5 data. Nonetheless, thecoarse fraction (PM2.5–PM10) also appears to contribute

le 2

nual costs of air filtration (fibrous bag-filters)

Cost ($/yr)

ergy 1480

placement—filter 960

placement—labor 140

posal 60

tal 2640

tal per occupant 2.64

to both morbidity and mortality [30]. The present studyaddresses particles smaller than 10 mm diameter (PM10),recognizing that fibrous filters are more efficient inremoving coarse particles than fine particles.We decided to use 30 mg/m3 for our value of annual

average outdoor PM10 concentration. Standards currentlyin place in North America and Western Europe for theannual average outdoor concentration of PM10 lie between30 and 60 mg/m3 [31]. Although the new WHO air qualityguideline for the annual mean concentration of particulatematter recommends 20 mg/m3 [32], only about 30% of theurban population worldwide experience annual PM10concentrations smaller than this value [31].The determination of the concentration of indoor PM10,

with and without filtration, requires several estimates.Indoor particles come from both outdoor and indoorsources. Particles that originate outdoors are removed byventilation-system filters and by deposition onto indoorsurfaces [33]. These loss processes depend strongly onparticle-size and vary with building design and operation.The current analysis assumes that indoor sources such as

smoking and other combustion processes (e.g., cooking)are not present. In today’s office-type buildings, this isusually the case. Nor are the effects of particle resuspensionon various health outcomes included in the currentanalysis. Since the epidemiological literature focuses onassociations between outdoor particles and health con-sequences, we have focused on the ‘‘indoor proportion ofoutdoor particles’’ (IPOP) as defined by Riley [8]. Based onthis study and the study of Jamriska [13], we estimated theannual average indoor PM10 concentration withoutparticle filtration to lie between 65% and 95% of theoutdoor PM10 level. We used 65%, 80% and 95% in ourparametric investigations of the impact of reduction ofindoor PM via filtration on various health outcomes. Forthese evaluations we assumed that removal efficienciesremained unchanged throughout a filter’s lifetime. This is aconservative assumption since removal efficiencies tend toincrease as a filter loads.Cohort and cross sectional studies have found associa-

tions between particle levels measured over a period of ayear or more and mortality. We have used results fromsuch studies as the basis for our estimates instead of resultsfrom studies that focus on acute particle exposures.Regarding morbidity, health effects were evaluated forthe following: respiratory hospital admissions, asthmarelated emergency room (ER) visits, minor restrictedactivity days (MRAD) and work loss days (WLD). Otherhealth and welfare benefits were excluded due to theinability to appropriately monetize them and to avoiddouble counting.Most studies express the health effects as a function of a

measured change in air pollutant levels. The calculation ofthe corresponding relative risk of health impact depends onthe concentration–response (C–R) functions from epide-miological studies. We have based our estimates on thegenerally accepted linear dose–response model. Hence, any

ARTICLE IN PRESSG. Beko et al. / Building and Environment 43 (2008) 1647–16571650

exposure reduction leads to a proportionate reduction inPM-induced mortality and other health effects [10,31].

3.2.1.1. Premature mortality. The baseline incidence rateis the number of cases of the health effect per year in theassessment location corresponding to baseline pollutantlevels in that location [34]. The US national averagebaseline incidence rates for non-accidental mortality byage-group were taken from a CDC compressed mortalityfile [35] as described in Hubbell [34]. The population-weighted average mortality rate per 100 people per year forthe 20–64 yr age group was determined using the US censuspopulation data for the year 2000 [36]. Such an approachassumes that the building occupants represent the age-specific population stratification in the presently assumedrange of workforce age (20–64 yr). For this group, theaverage annual mortality rate was 0.29 deaths/100 people.

Most of the recent epidemiological studies estimate aC–R function for the health effects of fine particles—PM2.5 [1–3]. The 1999 US EPA Report to Congress on theBenefits and Costs of the Clean Air Act [37] and numerousother evaluations use a C–R coefficient of 6.6% (i.e., anincrease in mortality rates of 6.6% for a 10 mg/m3 increaseof PM2.5) taken from the American Cancer Society study[2], which was based on a large sample size and extensivegeographic coverage. Using this study, other investigators[38,39] have translated the C–R coefficient for PM2.5 toone for PM10. We have based our calculations on a C–Rcoefficient of 3.8% increase in mortality rates for each10 mg/m3 increase of PM10 [39]. The decline of the annualmortality rate in our standard office building resultingfrom the addition of particle filtration was estimated as

Decline of mortality ¼ ððIndoor PM10 decreaseÞ=10Þ

� ðC2R functionÞ

� ðAverage annual

mortality rate per personÞ

� ðNo: of occupantsÞ. ð3Þ

The period that the occupants spend in the filtration-protected environment is limited to their working hours.The results for the health effects were therefore correctedby a factor of 0.228 ((8 work hours/24)� (250 work days/365)), assuming a linear C–R correlation.

To estimate the economic value of reductions in airpollution-related premature mortality, we have used amethodology referred to as the ‘‘value of statistical life-years’’ or VSLY [37]. In this approach, age-specific lifeexpectancy and death rates are taken into account togetherwith the age distribution and size of the exposed popula-tion to estimate the average number of years of life lost(YLL) per death. YLL per death was calculated from theWHO World Health Report 2002 [5] for the AMR-A sub-region, which includes Canada, United States and Cuba.The YLL attributable to urban air pollution (152,000)divided by the corresponding mortality (28,000) results in5.4 YLL per death. Hence, the number of saved life years in

the hypothetical office building is calculated as

ðNumber of saved life yearsÞ ¼ ðDecline of mortalityÞ

�YLL: ð4Þ

Note that the YLL and mortality values obtained fromWHO [5] are representative of the sub-region’s entirepopulation. Less premature deaths occur and more years oflife are lost per death among people between 20 and 64 yrof age than among older citizens.There are several ways to assign a dollar value to

premature mortality. One of them is by the associateddecrease in earnings—the human capital loss (HCL)approach:

Value of avoided mortality ¼ ðNumber of saved life yearsÞ

� ðAnnual salaryÞ. ð5Þ

The value of a person’s expected future earnings lost whena person dies prematurely is a narrow definition of benefitsand tends to underestimate the economic value ofpremature mortality. ‘‘Willingness to pay’’ (WTP), theamount that people are willing to pay to reduce the risk ofloosing a year of life, is an economic measure that isconsidered more comprehensive than the human capitalapproach:

Value of avoidedmortality ¼ ðNumber of saved life yearsÞ

� VSLY; ð6Þ

where VSLY is the value of a statistical life year based onWTP (VSLY ¼ $457,000 [37], adjusted with 3% discountrate to 2005 dollars).Ideally, WTP should capture the loss in satisfaction—

from consumption, leisure time, interaction with friendsand family—that occurs when life is shortened. However,such an approach to benefit valuation reflects the employ-ee’s perspective and is not representative of the party(building owner or employer) directly burdened by thecosts of filtration.In the present study, the value of avoided premature

mortality was calculated via both (i) HCLs (Eq. (5)) and (ii)WTP to avoid a lost year of life (Eq. (6)).The results for the various annual average indoor PM10

concentrations with and without filtration are summarizedfor the two valuation methods in Fig. 1.

3.2.1.2. Morbidity. The methodology of estimating thehealth benefits of filtration with regard to morbidity issimilar to that for mortality. The population-weightedaverage baseline incidence rates for the 20–64 yr age groupwere determined for respiratory hospital admissions andasthma ER visits based on the age-group-specified baselineincidence rates [34] and the US Census population data forthe year 2000. The average baseline incidence rates forMRAD and WLD were taken from US EPA [37].Given the purpose of the present analysis, we felt it was

adequate to use approximate C–R functions determined inprevious studies. We have chosen to use the estimates of

ARTICLE IN PRESSG. Beko et al. / Building and Environment 43 (2008) 1647–1657 1651

C–R functions derived from a meta-analysis [39]; these aresummarized in Table 3. We acknowledge that health effectrelationships are region-specific and applying a C–Rfunction from a single study or from several studies to allof the US or Europe introduces additional uncertainty. Thedecline in annual morbidity endpoints in the currently

0

5

10

15

20

25

30

35

0 20 40 60 80 100

Indoor PM reduction with filtration (%)

Ben

efi

t-H

CL

($/y

r/o

ccu

pan

t)

65% of outdoor PM w/out filter

80% of outdoor PM w/out filter

95% of outdoor PM w/out filter

0

20

40

60

80

100

120

140

160

180

0 20 40 60 80 100

Indoor PM reduction with filtration (%)

Ben

efi

t-V

SL

Y (

$/y

r/o

ccu

pan

t) 65% of outdoor PM w/out filter

80% of outdoor PM w/out filter

95% of outdoor PM w/out filter

Fig. 1. Economic benefits derived from decreased mortality rates as a

function of average indoor PM10 concentration reduction with filtration.

Benefits are depicted for scenarios that assume indoor PM10 concentra-

tion without filtration to be 65%, 80% or 95% of the outdoor PM level.

The results determined using (a) HCLs and (b) value of statistical life year

are shown.

Table 3

Baseline rates, C–R functions and unit values of morbidity endpoints

Health endpoint % change per 10mg/m3

daily average PM10aBaseline rates

(per 100 people

Resp. hosp. admission 1.39 1.32

Asthma ER visit 3.11 0.57

MRAD 4.92 780

Work loss days 7.74 237

aFrom Cesar [39].bFrom US EPA [37]; adjusted with 3% discount rate to 2005 dollars.

analyzed standard office building with filtration wasestimated as

Decline of morbidity endpoint

¼ ððIndoor PM10 decreaseÞ=10Þ � ðC2R functionÞ

� ðAverage baseline incident rate per personÞ

� ðNo: of occupantsÞ: ð7Þ

The monetary value associated with this decline wascalculated as

Value of health endpoint ¼ ðDecline ofmorbidity endpointÞ

� ðUnit value per incidentÞ.

ð8Þ

The unit value of a WLD was based on the average dailysalary of an office worker estimated as described earlier.The other unit values used in Eq. (8) were taken from a USEPA report [37] and adjusted from 1990 to 2005 dollarvalues by a 3% discount rate (Table 3). Wherever possible,the report uses the ‘‘WTP’’ approach to estimate the valueof avoided morbidity. However, WTP estimates are notavailable for some health effects. In such cases the cost oftreating or mitigating the effect is used as an alternativeestimate. The cost of illness (COI) usually understates thetrue value as it only captures the estimates of medical costsand costs of lost work-time; it does not reflect the value ofavoiding associated pain, suffering and lost leisure time.The results (Fig. 2) were again obtained by linear scaling

adjusting for the time period of the day and year that theoccupants spend in the hypothetical office building. Resultsfor respiratory hospital admissions and asthma ER visitsare not displayed separately. Their maximum value for thechanges in indoor PM concentrations examined in thisanalysis are below $1.5 and $0.05 per year per occupant,respectively, and thus they make a negligible contributionto the overall benefits associated with reduced morbidity.

3.2.2. Building and HVAC duct cleaning

Although cleaning products themselves can be a sourceof pollutants [40–42], indoor air quality may be moreadversely impacted by a lack of cleaning and inadequatecleaning programs [43–45]. Cleaning costs depend on thecleaning program. The annual cleaning cost in an averageUS building is about $15.7 perm2 per year [43] (adjusted by

per year)

Unit value per incident ($) Derivation of unit value

estimates

10 750b COI

302b COI

59b WTP

184 Daily wage

ARTICLE IN PRESS

0

5

10

15

20

25

30

35

0 20 40 60 80

Indoor PM reduction with filtration (%)

Ben

efi

t-M

RA

D (

$/y

r/o

ccu

pan

t) 65% of outdoor PM w/out filter

80% of outdoor PM w/out filter

95% of outdoor PM w/out filter

0

5

10

15

20

25

30

35

0 20 40 60 80 100

Ben

efi

t-W

LD

($/y

r/o

ccu

pan

t) 65% of outdoor PM w/out filter

80% of outdoor PM w/out filter

95% of outdoor PM w/out filter

0

5

10

15

20

25

30

35

0 20 40 60 80 100

Indoor PM reduction with filtration (%)

Ben

efi

t-T

ota

l ($

/yr/

occu

pan

t)

65% of outdoor PM w/out filter

80% of outdoor PM w/out filter

95% of outdoor PM w/out filter

100

IndoorPM reduction with filtration (%)

Fig. 2. Economic benefits from decreased morbidity rates as a function of

average indoor PM10 concentration reduction with filtration. Benefits are

depicted for scenarios that assume indoor PM10 concentration without

filtration to be 65%, 80% or 95% of the outdoor PM level. Results are

shown for decreased rates of (a) MRAD, (b) work loss days (WLD) and

(c) total morbidity.0

30

60

90

120

150

180

210

0 20 40 60

Decrease of cleaning costs due to filtration (%)

Be

ne

fit

($/y

r/o

cc

up

an

t)

Initial cost w/out filters $5/m2/yrInitial cost w/out filters $10/m2/yrInitial cost w/out filters $15/m2/yrInitial cost w/out filters $20/m2/yr

Fig. 3. Economic benefits as a function of decrease of cleaning costs due

to particle filtration. The correlation is shown for scenarios that assume

$5, $10, $15 or $20/m2 per year cleaning cost when no filtration is used.

G. Beko et al. / Building and Environment 43 (2008) 1647–16571652

3% discount rate to 2005 dollars). Cleaning has its greatestimpact on coarse particles deposited on horizontalsurfaces. The transport of such particles from outdoors

to indoors can be partly prevented by filtration. However,there is insufficient information to accurately estimate howchanges in indoor airborne particle concentrations influ-ence cleaning programs, the required degree of cleaningand the associated costs. For the purpose of our analyseswe have assumed that filtration could decrease thesecleaning costs anywhere from 5% to 60%. Hence, we haveconducted analyses for the impact of filtration on theseeconomic costs using four different estimates of thecleaning costs when no filtration is in place ($5, $10, $15or $20 perm2 per year). For each of these initial cleaningcost estimates, we have then estimated the benefit resultingfrom filtration over a range from 5% to 60% in thedecrease in the cleaning costs (see Fig. 3).Estimating the benefits with regard to HVAC unit

cleaning, is not a simple task, since such cleaning dependson several parameters and, in common practice, it is veryirregular and often neglected. This is the case in spite ofcurrent standards in some countries (see Finnish Society ofIndoor Air Quality and Climate [46], Pasanen [17]). CamfilFarr’s Life Cycle Cost software [27] assumes, for contin-uous HVAC operation, a cleaning interval of 5 yr with aEU4 filter and 20 yr with a EU7 filter in the ventilationsystem. The cost of ventilation duct cleaning can beestimated assuming an approximate duct area of 0.05m2

perm3/h airflow [27]. The cleaning cost per square meter ofduct area varies from country to country. In Sweden thecost is estimated to be $5/m2 [27]. The price of one-timecleaning in the present calculation, $9000, is derived fromsuch estimates. If we suppose that a system withoutfiltration would be cleaned once in 5 yr (rough cost of$1800/yr; disregarding discount rate for simplification) andwith a EU7 filtration this period would increase to 15 yr($600/yr; disregarding discount rate), we roughly save upto $1200 per year. That corresponds to a saving of $1.2 peryear per occupant, which is a negligible contributioncompared with some of the other economic benefitsconsidered in this study.

ARTICLE IN PRESS

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 1 2 3 4 5 6

Mo

neta

ry lo

ss (

$/y

r/o

ccu

pan

t) pollution 75% of time

pollution 50% of time

pollution 33% of time

Productivity loss (%)

Fig. 4. Monetary loss as a function of productivity loss due to pollutants

from used ventilation filters. The correlation is depicted for three different

scenarios: decreased occupant productivity occurs over 33%, 50% or 75%

of the filter’s service-life.

G. Beko et al. / Building and Environment 43 (2008) 1647–1657 1653

3.3. Estimate of productivity losses

Several recent studies have examined the adverse impactof indoor air pollution on performance of typical officework and the negative economic consequences [47–49]. Itis, however, difficult to estimate the extent to whichpollutants from loaded ventilation filters contribute tosuch an impact.

Wargocki [23] observed that at a high outdoor air supplyrate (80% of total airflow), replacing a 6 months old supplyair filter with a clean one improved call-center operatorperformance by a 10% reduction in average talk-time.However, at a low outdoor air supply rate (8% of totalairflow), filter replacement had no significant effect onperformance. In a study by Wyon [22], replacing usedsupply air pre-filters with new ones increased the self-estimated productivity in an office building by 5.7%.

In a laboratory experiment by Alm [50], 47% of thehuman subjects were dissatisfied with the air quality uponentering a room supplied with air that had passed througha used filter, whereas only 16% were dissatisfied when theair had passed through a new filter. Previous studies [51,52]suggest that productivity increases between 1% and 1.5%when the percentage dissatisfied with the air quality isdecreased by 10%. However, it is not obvious that theresults of these studies translate in a straightforwardfashion to average office work. Moreover, at the resolutionof Alm’s [50] experiments, there was no significant effect onthe measured performance of office work. Based on variousstudies and assumptions, Fisk and Rosenfeld [14] estimatedthat the productivity decrease caused by typical SBSsymptoms (such as can be caused as well by used filters[18,21]), can be between 1% and 4%.

An additional factor to consider is the period of a filter’sservice-life during which it noticeably emits pollutants.This may differ with location, season, HVAC operationand filter type. Pasanen [16] concluded that during the first3 months of intermittent operation of filters in actualbuildings, the odor emissions increased to a level thatevery third person would consider the resulting indoor airquality unacceptable. In the present analysis we haveexamined three different scenarios: decreased occupantproductivity attributable to pollutants from a used filteroccurs 3, 6 or 8 months after its installation (i.e., during33%, 50% or 75% of its service-life for a filter that isreplaced yearly).

Given the rather large range of possible effects ofemissions from used filters on occupant productivity (seeabove), we conducted a sensitivity analysis for the impactof productivity loss on the associated monetary loss usingEq. (9).

Monetary loss per occupant

¼ ð%productivity loss=100Þ � ðAnnual salaryÞ

� ðfraction of occupants’ work

period that the filters polluteÞ: ð9Þ

It has not been determined whether the initial negativeimpact of used filters on productivity changes as the filterfurther loads. Therefore we consider the chosen productiv-ity loss to be an average value over the entire period duringwhich the filter pollutes. The results are shown in Fig. 4 foreach of the three scenarios noted in the previousparagraph.The results indicate that even a small adverse effect of

used filters on the occupants’ productivity results inmeaningful economic losses. Such losses have the potentialto be one to two orders of magnitude higher than theannual running costs of filtration and may substantiallyexceed its benefits. This is an area that warrants furtherinvestigation.

4. Discussion

4.1. Sample calculation

The results presented in the previous paragraphs must beinterpreted cautiously. The input data for a number ofparameters depend on factors such as building design,climatic zone and specific maintenance programs. We have,therefore, calculated lower, central and upper estimates ofthe benefits and costs associated with particle filtrationusing lower, central and upper values for key inputvariables (see Table 4), as determined from the literatureand our judgment.Moreover, different stakeholders perceive costs and

benefits of filtration differently. For instance, a buildingowner who sublets a building to an employer is not directlyimpacted by changes in morbidity and mortality rateswithin the building; the employer renting the building isimpacted by changes in morbidity and mortality rates inone way; the employee is impacted by changes in morbidityand mortality rates in another way. Hence, in the overall

ARTICLE IN PRESS

Table 4

Input values used for the lower, central and upper estimates of endpoints associated with the use of particle filtration

Estimate Input variables

Mortality and morbidity Building cleaning Productivity

Indoor PM level

w/out filters

(% of outdoor PM)

Indoor PM

reduction with

filters (%)

Cleaning cost

decrease due to

filtration (%)

Initial cost w/out

filtration ($/m2/yr)

Productivity loss

(%)

Polluting period of

filter service time

(%)

Lower 65 30 5 5 0 –

Central 80 60 30 10 0.5 50

Upper 95 80 40 15 1.0 75

G. Beko et al. / Building and Environment 43 (2008) 1647–16571654

analysis presented in Fig. 5, we consider four differentperspectives:

(a)

Perspective of owner: subletting the building; pays forbuilding maintenance, but not for energy use.

(b)

Perspective of employer: renting the building; does not payfor building maintenance, but does pay for energy use.

(c)

Perspective of employer that owns the building;responsible for building maintenance and energy use.

(d)

Perspective of Society: total costs and benefits arerelevant.

This approach has its own weaknesses. For example, anemployer encounters economic losses as a consequence ofthe premature death of an employee (lost productivityresulting from disruption, loss of knowledge and trainingof a replacement); however, such losses are not properlycaptured via the ‘‘human capital’’ or ‘‘WTP’’ approaches.Given the inadequacy of these metrics, benefits fromdecreased mortality rates were not included in the estimatespresented in Figs. 5b and c. Similarly, since medical costsare typically not directly borne by the employer, benefitsfrom decreased morbidity rates (other than WLD) were notincluded in the estimates presented in Figs. 5b and c.

Although the evaluations presented in Fig. 5 are onlyrough estimates, they indicate that, regardless of perspec-tive, removal of particulate air pollution from the supplyair is beneficial if the negative aspects of loaded particlefilters can be avoided. However, if loaded filters pollute theair that passes through them and cause even a smallproductivity loss in the work environment, potentialeconomic benefits can easily be overwhelmed by decreasedworker efficiency. Alternatives to current filtration prac-tices may be able to avoid this pitfall. Investigating suchalternatives using economic analyses analogous to thoseused in this work, supported by experimental studies, iswarranted. It is anticipated that more frequent filterchanges would partially mitigate the effect. This impliesthat filter changes should be based on odors emanatingfrom the filter as well as pressure drop across the filter (seeFitzner [53], EN13779 [25]). Best practice should movetowards the development and use of efficient, low-polluting

filtration systems that are easy to maintain, have low lifecycle costs and minimal environmental impact.Additional benefits, not included in the present analysis,

include reduced soiling of heat exchangers and energyrecovery units. Soiling of HVAC systems not only reducesthe efficiency of heat exchangers and energy recovery units,but can also be a source of sensory offending emissions(similar to those emanating from loaded filters). We havenot included these endpoints in the model since thenecessary input data are not available. However, ourrough estimates of the benefits obtained from reducedsoiling of HVAC components (not presented here) indicatethat the value of these benefits are somewhere betweennegligible and the same order of magnitude as the annualrunning costs of filtration. Other factors that have not beenconsidered in the present analysis include taxes, insurance,the environmental impact of filter disposal, and the factthat premature death is more likely to occur among moresenior personnel who tend to have higher salaries.It should be remembered that the reference condition for

the current comparison was an identical HVAC systemwith no filter in place. That is, the present study does notaddress the costs and benefits of improved filtration.However, it does provide a sense of the endpoints thatwould be most influenced by upgrading existing particlefiltration.

4.2. Influence of recirculation

The present analysis is for a one-pass ventilation system.However, in many parts of the world (e.g., Europe outsideof Scandinavia, the United States, Singapore) as much as80–90% of the supply air is recirculated. In a scenariowhere 80% of the total airflow is recirculated air and theamount of outside air is 10 L/s/occupant, the total airflowthrough the ventilation unit would be 50m3/s (i.e., fivetimes larger than in a system with no recirculation). Hence,for an identical face velocity, the size of the filter bank andconsequently the energy, filter replacement and disposalcosts would increase. However, these increases would besomewhat less than a factor of five since the pressure dropwould be anticipated to increase at a slower rate despite theadded filtration of particles of indoor origin.

ARTICLE IN PRESS

43

3.6

86

1.2

0

50

100

150

200

250

300

350

Building cleaning Running costs (w/out energy)

Mo

neta

ry v

alu

e (

$/y

r/o

ccu

pan

t)

Lower estimate Central estimate Upper estimate

Lower estimate Central estimate Upper estimate

Lower estimate Central estimate Upper estimate

Lower estimate Central estimate Upper estimate

Benefits Costs

0114.5

115

18

345

1.50

50

100

150

200

250

300

350

Work loss days Running costs(energy only)

Productivity loss

Mo

neta

ry v

alu

e (

$/y

r/o

ccu

pan

t)

Benefits Costs Losses

3.6 011

43

4.5

115

345

18

86

2.6

0

50

100

150

200

250

300

350

Work lossdays

Buildingcleaning

Runningcosts-total

Productivityloss

Mo

neta

ry v

alu

e (

$/y

r/o

ccu

pan

t)

Benefits Cost s Losses

3.6 0

91

197.8

37

115

43

345

144

86

30

2.6

0

50

100

150

200

250

300

350

MortalityWTP-VSLY

Morbidity-total

Buildingcleaning

Runningcosts-total

Mo

neta

ry v

alu

e (

$/y

r/o

ccu

pan

t)

Benefits Cost s Losses

Productivityloss

Fig. 5. Lower, central and upper estimates of the aggregated benefits and

costs of particle filtration from four perspectives: (a) building owner, (b)

employer, (c) employer/building owner and (d) society (see text for

details).

G. Beko et al. / Building and Environment 43 (2008) 1647–1657 1655

We may assume that the relatively large benefits resultingfrom the filtration of outdoor particles would be similar withand without the use of recirculating air. Moreover, since theparticle filters are commonly placed downstream of themixing box in systems with recirculation, the concentrationof indoor generated particles would be less in such a system.This is expected to be beneficial in terms of both health andfrequency of surface cleaning. We should be mindful of thefact that the filter cake accrued on filters in a recirculatingsystem differs from that in a one-pass system, and thetemperature of the air passing through the filters is likely tobe different in the two systems. These differences couldinfluence pollutants emanating from a used filter andpotential effects on occupant productivity. Taken together,this cursory examination suggests that the benefits stilloutweigh the running costs in a system that recirculates alarge fraction of the supply air. However, further analysesare necessary to support or refute this conclusion.

5. Conclusions

The present study indicates that the overall benefits ofusing particle air filtration in office buildings are severaltimes larger than the associated running costs. Substantialsavings are obtained from decreased occupant morbidityand mortality (resulting from reduced exposure to particlesof outdoor origin) as well as from less frequent buildingcleaning. However, the magnitude of the net benefits varieswith the perspective of the stakeholder and the valuationapproach. For society as a whole, a major portion of thebenefits derives from reduced occupant morbidity andmortality. For both building owners and society, savingsrelated to less frequent building cleaning are substantial.The employer obtains significant economic benefits mainlyfrom the reduced number of days of work lost related tothe adverse health effects of particles. On the other hand,the benefits obtained from lower indoor particle concen-trations may easily be overwhelmed by even a smalldecrease in occupant productivity as a consequence ofsensory offending pollutants emanating from used particlefilters. Future research would be valuable in two areas withlarge uncertainties and also a large influence on the overallresults of this assessment: (i) quantifying the impact ofpolluting ventilation filters on worker productivity and(ii) quantifying the benefit to the employer delivered fromdecreased occupant mortality.

Acknowledgements

This work has been supported by the Danish TechnicalResearch Council (STVF) as part of the research programof the International Centre for Indoor Environment andEnergy established at the Danish Technical University forthe period 1998–2007. The authors would like to thankProfessors Kirk Smith and William W. Nazaroff, bothfrom University of California, Berkeley, for providinghelpful literature and valuable comments. We thank Camfil

ARTICLE IN PRESSG. Beko et al. / Building and Environment 43 (2008) 1647–16571656

Farr Denmark for providing us with the Life Cycle Costsoftware.

References

[1] Dockery DW, Pope III CA, Xu X, Spengler JD, Ware JH, Fay ME,

et al. An association between air pollution and mortality in six US

cities. New England Journal of Medicine 1993;329(24):1753–9.

[2] Pope III CA, Thun MJ, Namboodiri MM, Dockery DW, Evans J,

Speizer FE, et al. Particulate air pollution as a predictor of mortality

in a prospective study of US adults. American Journal of Respiratory

and Critical Care Medicine 1995;151:669–74.

[3] Pope III CA, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K,

et al. Lung cancer, cardiopulmonary mortality and long-term

exposure to fine particulate air pollution. Journal of American

Medical Association 2002;287(9):1132–41.

[4] Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL,

et al. Fine particulate air pollution and hospital admission for

cardiovascular and respiratory diseases. Journal of American

Medical Association 2006;295(10):1127–34.

[5] WHO. Reducing risks, promoting healthy life. The World Health

Report 2002. World Health Organisation; 2002.

[6] Clayton CA, Perrit RL, Pellizzari ED, Thomas KW, Whitmore RW,

Wallace LA, et al. Particle total exposure assessment methodology

(PTEAM) study: distributions of aerosol and elemental concentra-

tions in personal, indoor and outdoor air samples in a southern

California community. Journal of Exposure Analysis and Environ-

mental Epidemiology 1993;3:227–50.

[7] Wallace L. Indoor particles: a review. Journal of the Air and Waste

Management Association 1996;46:98–126.

[8] Riley WJ, McKone TE, Lai AK, Nazaroff WW. Indoor particulate

matter of outdoor origin: importance of size-dependent removal

mechanisms. Environmental Science and Technology 2002;36:200–7.

[9] Weschler CJ, Shields H, Shah B. Understanding and reducing the

indoor concentration of submicron particles at a commercial building

in Southern California. Journal of the Air and Waste Management

Association 1996;46:291–9.

[10] Hanninen OO, Palonen J, Tuomisto JT, Yli-Tuomi T, Seppanen O,

Jantunen MJ. Reduction potential of urban PM2.5 mortality risk

using modern ventilation systems in buildings. Indoor Air

2005;15:246–56.

[11] Mendell MJ, Fisk WJ, Petersen M, Dong MX, Hines CJ, Faulkner

D, Deddens JA, Ruder AM, Sullivan D, Boeniger MF. Enhanced

particle filtration in non-problem office environment: summary

finding from a double-blind crossover intervention study. In:

Proceedings of indoor air 1999, vol. 4, Edinburgh; 1999. p. 974–5.

[12] Fisk WJ, Faulkner D, Palonen J, Seppanen O. Performance and costs

of particle air filtration technologies. Indoor Air 2002;12:223–34.

[13] Jamriska M, Morawska L, Ensor DS. Control strategies for sub-

micrometer particles indoors: model study of air filtration and

ventilation. Indoor Air 2003;13:96–105.

[14] Fisk WJ, Rosenfeld AH. Estimates of improved productivity and

health from better indoor environments. Indoor Air 1997;7:158–72.

[15] Weschler CJ. Predictions of benefits and costs derived from

improving indoor air quality in telephone switching companies.

Indoor Air 1991;1:65–78.

[16] Pasanen P, Teijonsalo J, Seppanen O, Ruuskanen J, Kalliokoski P.

Increase in perceived odor emissions with loading of ventilation

filters. Indoor Air 1994;4:106–13.

[17] Pasanen P. Emissions from filters and hygiene of air ducts in the

ventilation systems of office buildings. Doctoral dissertation, Kuopio

University Publications C. Natural and Environmental Sciences, vol.

80; 1998. p. 1–77.

[18] Clausen G. Ventilation filters and indoor air quality: a review of

research from the International Centre for Indoor Environment and

Energy. Indoor Air 2004;14(S7):202–7.

[19] Beko G, Halas O, Clausen G, Weschler CJ. Initial studies of

oxidation processes on filter surfaces and their impact on perceived

air quality. Indoor Air 2006;16:56–64.

[20] Hyttinen M, Pasanen P, Bjorkroth M, Kalliokoski P. Odors and

volatile organic compounds released from the ventilation filters.

Atmospheric Environment 2007;41(19):4029–39.

[21] Clausen G, Alm O, Fanger PO. The impact of air pollution from used

ventilation filters on human comfort and health. In: Proceedings of

the ninth international conference on indoor air quality and climate,

indoor air 2002, vol. 1, Monterey, USA; 2002. p. 338–43.

[22] Wyon DP, Tham KW, Croxford B, Young A, Oreszczyn T. The

effects on health and self-estimated productivity of two experimental

interventions which reduced airborne dust levels in office premises.

In: Proceedings of healthy buildings 2000, vol. 1, Espoo, Finland;

2000. p. 641–6.

[23] Wargocki P, Wyon DP, Fanger PO. The performance and subjective

responses of call-centre operators with new and used supply air filters

at two outdoor air supply rates. Indoor Air 2004;14(S8):7–16.

[24] CEN CR1752. Ventilation for buildings—design criteria for the

indoor environment. European Committee for Standardization; 1998.

[25] EN13779. Ventilation for non-residential buildings. Performance

requirements for ventilation and room-conditioning systems. Eur-

opean standard; 2004.

[26] US Department of Labor. Bureau of Labor Statistics. National

compensation survey: occupational wages in 2005; 2005. /http://

www.bls.gov/ncs/home.htmS.

[27] Camfil Farr. Life Cycle Cost software 1.47EU. Camfil Farr;

2001–2002.

[28] EUROVENT. Recommendation concerning calculation of life cycle

costs for air filters. European committee of air handling, air

conditioning and refrigeration equipment manufacturers; 1999.

[29] Hangstrom K, Holmber R, Lehtimaki M, Rilio J, Siitonen E.

Hinkkassuodatukson perusteet ja optimaalinen valinta. INVENT/

Suodatin-projekti, Vaihe 1, Raportti 48, MET; 1996. p. 1–54 (in

Finnish).

[30] Sandstrom T, Nowak D, Van Bree L. Health effects of coarse

particles in ambient air: messages for research and decision-making.

European Respiratory Journal 2005;26:187–8.

[31] Cohen AJ, Anderson HR, Ostro B, Pandey KD, Krzyzanowski M,

Kunzli N, et al. Urban air pollution. In: Ezzati M, Lopez AD,

Rodgers A, Murray CJL, editors. Comparative quantification

of health risks. Global and Regional Burden of Disease Attributable

to Selected Major Risk Factors. WHO, vol. 1; 2004. p. 1354–433.

/http://www.who.int/publications/cra/chapters/volume2/1353-1434.

pdfS.

[32] WHO. WHO air quality guidelines global update 2005. Report on a

working group meeting. World Health Organization, Bonn, Ger-

many, 18–20 October 2005.

[33] Wallace L. Correlations of personal exposure to particles with

outdoor air measurements: a review of recent studies. Aerosol Science

& Technology 2000;32(1):15–25.

[34] Hubbell BJ, Hallberg A, McCubbin DR, Post E. Health-related

benefits of attaining the 8-h ozone standard. Environmental Health

Perspectives 2005;113(1):73–82.

[35] CDC. CDC WONDER (1996–1998). Atlanta, GA: Center for

Disease Control and Prevention; 2005. /http://wonder.cdc.govS.

[36] US Census Bureau. Accessed via CDC WONDER on-line database,

Center for Disease Control and Prevention; 2005. /http://wonder.

cdc.govS.

[37] US EPA. The benefits and costs of the clean air act. Report to

Congress. Environmental Protection Agency; 1999.

[38] Cifuentes L, Borja-Aburto VH, Gouveia N, Thurston G, Davis DL.

Assessing the health benefits of urban air pollution reductions

associated with climate change mitigation (2000–2020): Santiago, Sao

Paulo, Mexico City, and New York City. Environmental Health

Perspectives 2001;109(S3):419–25.

[39] Cesar H, Borja-Aburto VH, Cicero-Ferandez P, et al. Improving air

quality in metropolitan Mexico City. An economic valuation. Policy

ARTICLE IN PRESSG. Beko et al. / Building and Environment 43 (2008) 1647–1657 1657

research working paper 2785, The World Bank; 2002. /http://www-

wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/

2002/03/08/000094946_02022604025131/Rendered/PDF/multi0page.

pdfS.

[40] Nazaroff WW, Weschler CJ. Cleaning products and air fresheners:

exposure to primary and secondary air pollutants. Atmospheric

Environment 2004;38:2841–65.

[41] Kildesø J, Tornvig L, Skov P, Schneider T. An intervention study of

the effect of improved cleaning methods on the concentration and

composition of dust. Indoor Air 1998;8:12–22.

[42] Weschler CJ, Shields H. Indoor ozone/terpene reactions as a

source of indoor particles. Atmospheric Environment 1999;33:

2301–12.

[43] Kildesø J, Schneider T. Prevention with cleaning. In: Spangler JD,

McCarthy JF, Samet JM, editors. Indoor air quality handbook. New

York: McGraw-Hill; 2001. p. 64.1–64.18.

[44] Franke DL, Cole EC, Leese KE, Foarde KK, Berry MA. Cleaning

for improved indoor air quality: an initial assessment of effectiveness.

Indoor Air 1997;7:41–54.

[45] Skulberg KR, Skyberg K, Kruse K, Eduard W, Djupesland P, Levy

F, et al. The effect of cleaning on dust and the health of office

workers: an intervention study. Epidemiology 2004;15(1):71–8.

[46] FISIAQ. Classification of indoor climate, construction and finishing

materials. Finnish society of indoor air quality and climate, Espoo,

Finland; 1995.

[47] Wargocki P, Wyon DP, Baik YK, Clausen G, Fanger PO. Perceived

air quality, sick building syndrome (SBS) symptoms and productivity

in an office with two different pollution loads. Indoor Air 1999;

9:165–79.

[48] Wargocki P, Wyon DP, Sundell J, Clausen G, Fanger PO. The effects

of outdoor air supply rate in an office on perceived air quality, sick

building syndrome (SBS) symptoms and productivity. Indoor Air

2000;10:222–36.

[49] Djukanovic R, Wargocki P, Fanger PO. Cost–benefit analysis of

improved air quality in an office building. In: Proceedings of Indoor

Air 2002, vol. 1, Monterrey, USA; 2002. p. 808–13.

[50] Alm O. Ventilation filters and their impact on human comfort, health

and productivity. Doctoral dissertation (PhD thesis MEK-I-Ph.D.

01.02). International Centre for Indoor Environment and Energy,

Technical University of Denmark, Lyngby; 2001.

[51] Wargocki P, Wyon DP, Fanger PO. Productivity is affected by the air

quality in offices. In: Proceedings of healthy buildings 2000, vol. 1,

Espoo, Finland; 2000. p. 635–40.

[52] Wargocki P, Wyon DP, Fanger PO. Pollution source control and

ventilation improve health, comfort and productivity. In: Proceedings

of the third international conference on cold climate heating,

ventilating and air conditioning—cold climate HVAC 2000, Sapporo,

Japan; 2000. p. 445–50.

[53] Fitzner K. Control of pollutants in air handling systems. In: Proceedings

of healthy buildings 2000, vol. 2, Espoo, Finland; 2000. p. 21–33.