Tobacco Documents Online

Report to CORESTA SURVEY OF INDOOR AIR QUALITY VENTILATION, AND SMOI~ING ACTIVITY IN RESTAURANTS July 1, 1999 322220856

Report to CORESTA SURVEY OF INDOOR AIR QUALITY, VENTILATION, AND SMOKING ACTIVITY IN RESTAURANTS H. R. Bohanon, Jrt, J.-J. Piad6*, M. Schorp*, Y. Saint-Jalm+ t R. J. Reynolds Tobacco Co. Winston-Salem, NC, USA, *Philip Morris Europe, Neuch~tel, Switzerland, + Seita, Fleury-les-Aubrais, France ABSTRACT As a CORESTA ETS Sub-Group activity, six different participants in six different countries conducted a survey of indoor air quality, ventilation and smoking activity in restaurants. This work was to assess methods and determine major hurdles to performing such studies in a sound and reliable way. A number of typical medium-priced restaurants were surveyed during lunch or dinner. The objectives were: 1. To survey restaurants in several different countries. 2. To determine acceptability of indoor air conditions as judged by occupants of the restaurant space. 3. To assess concentrations of environmental tobacco smoke (ETS) present in the selected spaces. 4. To appraise the ventilation system including operation and maintenance (O&M) of heating, ventilating, and air conditioning (HVAC) systems and estimate outdoor air ventilation rates. 5. To determine smoking rates in the restaurants observed. 6. To investigate correlation between the smoking rates, ventilation rates, ETS concentrations, and occupant perceptions of restaurant indoor air quality. This report discusses the results and conclusions from this survey. INTRODUCTION This project was designed to survey restaurants in a number of countries where smoking is permitted. The hypothesis is that most restaurateurs operate their restaurants so that 80% or more of the patrons perceive the air quality to be acceptable. Restaurants were selected to be typical of the country. The surveys were conducted during the lunch and/or dinner period. The protocol suggested selecting high occupancy days. Thirty-four restaurants in six countries were surveyed, adding valuable data to the published information regarding indoor environmental conditions or ventilation in existing restaurants. This pilot study was not designed to yield representative data for the individual countries due to the small number of samples. 322220857

Country Participating Organizations Principal Investigator(s) France Seita Mr. Yves Saint-Jalm L. H. P.* Louis Harris France Japan Japan Tobacco Inc. Korea Switzerland United Kingdom United States Korea Ginseng and Tobacco Research Institute Philip Morris Rothmans International British American Tobacco Co. Ltd. Imperial Tobacco Co. Philip Morris R. J. Reynolds Tobacco Co. Lorillard Tobacco Co. Dr. Masao Matsukura Mr. Isao Ishii Mr. Takuya Asai Mr. Takumi Nishina Dr. Moon Soo Rhe˘ Dr. J.-J. Piad~ Dr. Matthias Schorp Mr. L. Poget Mr. Barrio Frost Mr. Nigel Warren COVANCE Mr. Hoy Bohanon Mr. David Taylor Mr. Robin Wilson Table 1: Participants in the Survey * Laboratorie d'Hygi6ne de la ville de Paris (Dr. A. M. Laurant) METHODS A model protocol was issued for this survey. The test results from each country differ slightly in content due to local adaptations of the protocol. Model Protocol The model protocol was slightly modified in each country to address regional differences and to accommodate specific situations. The model protocol addressed the following subjects: 1) Selection of Buildings A minimum of five restaurants in different buildings should be tested. The restaurants and buildings should be dissimilar (e.g. not all hotel dining rooms). 2) Procedure for Selecting Sampling Spaces Within the Restaurant Where there are smoking and nonsmoking sections, one sample should be from each section. 3) Procedure for Selecting Sampling Locations in Spaces Sampling should be conducted in the breathing zone and not be unduly influenced by fans or ventilation systems, or or by direct exposure to sidestream or mainstream smoke plumes. 4) Qualitative Assessment of Ventilation System (Natural or Mechanical) 322220858

The protocol addressed both natural and mechanical ventilation systems. Mechanical systems should be assessed as to adequacy, cleanliness and structural integrity. 5) Determination of Outside Air Quality Local ambient monitoring station data on air quality should be obtained if aw~ilable from local air quality authority. Direct measurement of the lbllowing analytes should be obtained immediately outside of the establishment being tested: RSP, UVPM, FPM, solanesol, carbon monoxide, carbon dioxide, temperature and relative humidity (RH). 6) Determination of Indoor Air Quality Indoor air quality measurements should be made in duplicate in at least two sampling locations for each restaurant. If the restaurant has a smoking and a nonsmoking section, at least one measurement should be made in each section. Concentrations of RSP should be determined by the method described by Conner et al. Samples should be collected for at least tbur hours during testing. Concentrations of UVP~ should be determined by the method described by Conner et al. Samples should be collected for at least lbur hours during testing. Concentrations of FPM should be determined by the method described by Ogden et al. This analytical method utilizes the same sample as that collected for the determinations of RSP and UVPM described above. Concentrations of solanesol should be determined by the method described by Ogden et al. This analytical method utilizes the same sample as that collected for the determinations of RSP and UVPM described above. Concentrations of nicotine should be determined by EPA method I P-2A. Samples should be collected for at least four hours during testing. Concentrations of 3-EP should be determined by the method described by Ogden et al. This analytical method utilizes the same sample as that collected for the determinations of nicotine described above. In addition, measurements should be made for carbon monoxide, carbon dioxide, temperature and relative humidity. 7) Determination of Smoking Activity Two methods lbr determining smoking activity were recommended. The first method is collecting and counting the cigarette butts. The second method is visual observation of smoking. 8) Determination of Occupancy of Test Space The number of people in each room or smoking or nonsmoking area should be counted every thirty minutes separating counts of patrons and employees. 9) Determination of Quantity of Outside Air Supplied to Test Spaces Method i - [lave test and balance (T&B) engineering consultant measure HVAC and establish a ventilation rate. 322220859

Method 2 - Use CO.,. measurements along with counts of people in the space to estimate air exchange rates. Method 3 - Use tracer gas to determine air exchange rates. 10) Indoor Air Quality Questionnaire A questionnaire was designed that recorded gender, age, smoking status and asked about visitation to the establishment. The environmental questions addressed temperature, odor, drafts, noise, smoke, lighting and acceptability. Depending on the country, these questions were translated and adapted. In some cases questions were added to the original questionnaire. 11) Scheduling The testing should be scheduled during high occupancy times in the restaurant. 12) Analysis In this observational study, the analysis should use the observed data without modification. In some environmental studies, background correction is appropriate. Since this study was not designed to specifically collect background data, no corrections should be made in the analysis. 4 322220860

Testing in Each Country Table 2. Shows the actual testing that occurred in each location. France Japan Korea Switzerland United United Kingdom States Number of 5 4 5 5 10 5 Restaurants Number of days 3 2 2 2 1 2 Single or Sing. Dup. Dup. Dup. Dup. Dup. Duplicate Sample Outside RSP No Yes Yes Yes Yes No Outside UVPM Outside FPM No No Yes No Outside Solanesol No Yes Outside CO Yes Yes Outside C02 Yes Yes outside Temp. Yes Yes Outside Rft Yes Yes Indoor RSP Yes Yes Indoor UVPM Yes Yes Indoor FPM Yes No Indoor Solanesol Yes Yes Indoor CO Yes Yes Indoor CO_, Yes Yes Indoor Temp. Yes Yes Indoor RH Yes Yes Indoor Nicotine Yes Yes Indoor 3-EP Yes Yes Smoking Activity Yes Yes Occupancy Yes Yes Ventilation Rate Yes Yes Questionnaire Yes Yes Questionnaire Yes No modified/adapted Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Ycs No IYcs 322220861

RESULTS AND DISCUSSION Occupant Judgement Restaurant owners are motivated by business reasons to satisfy their customers. However, unlike comments on obvious items such as food, services and ambience, it is sometimes difficult for the owners to receive feedback specific to indoor environmental quality. The situation is further complicated by the fact that the owner strives to accommodate a wide range of individual prel~rences. Little guidance on how to provide acceptable indoor environmental conditions for restaurants is available. Some national or international ventilation standards or guidelines provide some general in|brmation, but that information is generally not specific for the hospitality sector. The Pilot Study Protocol provided a questionnaire, which sought basic information on the occupants, and assessed their perception towards a number of indoor environmental conditions, such as noise, temperature, draft, odor, humidity, freshness, tobacco smoke, and indoor environmental quality. In addition, France and Switzerland both added a question concerning indoor air quality, following what has recently been done in the frame of the European Building Audit (Bluyssen 1995). Questionnaires were completed on the same days as indoor air measurements ~vere performed. The time required to fill out a questionnaire was normally less than five minutes. In general, the wait staffadministered questionnaires, typically after they took orders for lunch or dinner. Some patrons declined to answer the questionnaires. The completed questionnaires were returned to the wait stall In total, 1370 questionnaires were returned by the restaurant patrons in five countries. No reliable estimate could be obtained for the non-response rate. The wait staff received a small gratuity for the additional workload. For the analysis of the questionnaires, the responses on the 5-point scales were converted to "Satisfied .... Rather Satisfied", ~'Neither", "Rather Dissatisfied", and "Dissatisfied." The responses marking the two lower ratings ("Rather Dissatisfied" and "Dissatisfied") were consolidated to estimate the "% Dissatisfied" with each of the indoor environmental parameters. When available, the 10-point scale was analyzed as in the European Building Audit (Bluyssen 1995) to derive the "% Dissatisfied" with indoor air quality (IAQ). This percentage was then averaged with the % dissatisfied fbr indoor environmental quality (IEQ) to derive the inverse estimate of"% Overall Acceptance." (figures 1 and 2). Excluding those restaurants that had less than 20 questionnaires returned or highly variable ventilation rates led to a subset of the data comprising 15 restaurants, which could be further analyzed. Table 3 presents the "'% Dissatisfied" assessments consolidated by country for each indoor environmental parameter. Dissatisfaction rates above 20% were observed in some countries for Air Draft, Air Freshness and Noise. Dissatisfaction rates with regard to indoor air quality and indoor environmental quality range between 3.1 and 12.0% and are not significantly different. 322220862

Table 3: Questionnaires - Summary Restaurants "Percent Dissatisfied" Country France Japan Switzerland USA UK Percent Dissatisfied Temp Draft llumid Fresh Smoke 5.7 8.7 5.0 37.9 4.3 17.4 6.5 19.6 3.9 38.5 12.0 15.8 8.0 15.8 9.2 7.7 7.2 l 1.8 n.d. n.d. 17.5 12.0 14.2 12.9 2.7 Noise Odor IAQ IEQ 27.0 10.5 7.8 6.9 31.5 2.2 n.d. 12.0 15.8 19.1 3.1 2.7 17.2 8.1 n.d. 3.0 n.d. n.d. 5.0 n.d. Figure 1 depicts the cumulative percentages for overall acceptance per country for all restaurants. Figure 2 shows the ratings for overall acceptance separating the data into smoker and nonsmoker responses for those restaurants having more than 20 questionnaires returned, and where the ventilation rate was found to be constant. In general, non-smokers tend to be slightly more dissatisfied than are smokers. oVERALL ACCF.PTANCE Indoor environmental design guidelines and Fig. 2: % Overall Acceptance among Sntokem v$ N(m-$mokem in 15 Restaur~nt~ 100 90 g5 1 2 3 4 5 6 7 8 ~ 10 11 12 13 14 15 Restaurant No. standards provide for "guidance values" of dissatisfaction in the range of 10 - 30% (EUR 14449 EN, 1992) or tbr more than 80% acceptance (ASHRAE 62-1989). These guidance values are based on experimental studies (Cain 1983) or field tests in some workplace environments (Fanger 1988). The data presented above indicate that physical stressors such as noise and draft challenge those limits. On the other hand, occupants in real-world environments appear to have diffi~rcnt expectations towards air quality parameters. The acceptance rate observed in this study is indeed substantially higher than what would have been predicted ti-om the results of the above studies, including the most recent one by Walker et al. (1997). 322220863

Concentration of ETS The CORESTA Restaurant Pilot Study combined the efforts from six different laboratories in as many different countries. A large amount of data was generated, reflecting the diverse situations that can be encountered when monitoring indoor air constituents in restaurants. Despite an initial agreement on the nature of the analytes and the methods to assess their levels, the difficulties of conducting this Pilot Study highlight the need for a commonly agreed methodology and validation of that methodology, including field tests. Detection of outliers Indoor air monitoring in restaurants often calls for environmental analyses at the ppb level. Despite the great care taken in performing the measurements, and even if the method is well controlled, (as checked by QA samples) erroneous results often do occur. They may be due to failures in the sampling, or may be the result of a strongly irdaomogeneous environment. In addition, ETS markers, such as UVPM and FPM, may yield implausibly high results due to interference from other combustion sources. It is very useful to obtain two or more replicate determinations per location. This helps to detect outliers and may give an indication of the most plausible result. In this data set, about 5% of the raw data were outliers. Additional information can be derived, however, from the auto- correlation that prevails among the set of ETS markers as discussed below. It is not a simple task to perform this two-dimensional cross-checking using statistical tools, therefore the outliers were identified by visual evaluation. Completion of the CORESTA inter-laboratory study that is on- going tbr all these methods should help in detecting outliers by providing an estimate of the method reproducibility. Results and correlation As an example, the scatter plots showing all the cross-correlation between the analytical parameters for the Swiss data set are shown in Figure 3. The monitored ETS-markers fall into two groups: the gas-phase compounds (nicotine SN and 3-ethenylpyridine EP3) and the estimates tbr ETS-derived particulate-phase concentrations(the UVPM and FPM estimations and the one obtained from solanesol, SPM). These estimations of the level of ETS-derived particulate matter (or ETS-RSP) were pertbrmed as recommended in Nelson et al., 1997. In addition, the concentrations of respirable suspended particles (RSP) are also reported. 322220864

Fig. 3 Scatter Plots between the 6 determinations (RSP, ETS-markers) Swiss Data, Z ILl FPM RSP UVPM SPM SN EP3 n=80 These plots, as well as those generated from the other countries' data sets, suggest the following observations: Within the particulate-.phase, there is a very good correlation between the two spectrometric estimations. Fhis is not surprising since both determinations are made from the same sample. Conversely, an outlying point should be investigated for possible intert~rence. There are two obvious cases (discussed below) in the Swiss data set, The solanesol concentration (or the derived SPM estimation) is also well correlated with UVPM or FPM There are, however, some points that show lower SPM than UVPM or FPM, mostly at low s~noke levels. This effect is illustrated in Figure 4, showing the ratio of SPM to UVPM as a function of UVPM. • O~er the whole database, it appears that the RSP level does not correlate well with ETS-RSP le~,els. The gas phase data are well correlated. There is a trend towards a lower nicotine / 3- ethenylpyridine ratio at lower smoke levels as illustrated in Figure 5, that shows this ratio as a function of nicotine levels. As nicotine exhibits much larger sorption effects, this ratio will 322220865