Home Environment and Noise Annoyance in a National Sample of Multi-Family Buildings in Sweden- Associations with Stress-Related Symptoms


 Background: Poor acoustic condition at home can have negative health impact. The aim was to investigate home environment factors and stress symptoms associated with noise annoyance. Methods: All adults (≥18 y) registered in selected apartments in Sweden were invited to participate in a questionnaire survey including medical questions and personal factors. Totally 5775 adults participated (response rate 46%). Information on home environment was obtained through an indoor environment questionnaire. Two-level logistic regression models (individual, municipality) were performed to estimate associations.Results: Totally 11.2% reported any frequent noise at home. Voice/radio/TV/music/similar sounds from neighbours (13.2%) and scraping sound/footsteps/thumping from neighbours (16.5%) and road traffic (16.1%). Younger participants reported more noise annoyance and more stress-related symptoms. Females were more sensitive to noise from neighbours and traffic noise causing severe effects. Noise annoyance was related to tiredness, headache and difficulty concentrating (OR=1.70-8.19), and noise annoyance from many sources was related to a higher risk of symptoms. A warmer climate (OR=1.30), buildings constructed from 1961-1985 (OR=1.50-1.85), renting (OR=1.83) and living alone (OR=1.42) were related to more noise annoyance. A warmer climate (OR=1.95), higher municipality population density (OR=1.24), a longer living time (OR=1.34), construction year (1961-1975) (OR=2.42), renting (OR=1.80-2.32), living above ground floor (OR=1.45) and having a bathroom fan (OR=1.84) were associated with increased noise annoyance from neighbours. Factors associated with increased noise annoyance from installations or ventilation/fans/heat pumps included a warmer climate, higher municipality population density, construction year (1961-1995), renting and any mechanical ventilation. Higher municipality population density, construction year (especially 1961-1985) and renting were associated with more noise annoyance from traffic (OR=1.77-3.92). Renting (OR=1.73) and living above ground floor (OR=1.60) were related to more severe traffic noise effects.Conclusions: The present study shows that noise annoyance may cause stress-related symptoms. Younger age, female gender, living alone, a longer living time, a warmer climate, higher municipality population density and building factors (renting, construction period 1961-1975, living above ground floor and mechanical ventilation system) were related to more noise annoyance.

Research to date has tended to focus on outdoor environmental noise (i.e. road tra c and aircraft noise), while little attention has been paid to indoor noise. The predominant source of noise annoyance in residential quarters is tra c followed by neighbourhood noise. Neighbourhood noises usually include noise of footsteps, sounds with high information content such as language or music, etc. It is in the nature of humans to have their attention drawn to such informative sounds, even at relatively low level. The annoyance potential of neighbourhood noise is therefore relatively high. Most published studies focus on tra c noise caused health problems. Few studies exist on effects of neighbourhood noise on annoyance or health (24,25). One study from Sweden reported that the residents were twice more often annoyed from noise from installation of ventilation and air-conditioning systems than noise from tra c (25). One large European study found neighbourhood noise increased health risk for the cardiovascular system and increased risk of annoyance by depression and migraine among adults (24).
More studies are needed to investigate how annoyance of noise can affect occupants' health. Our study is part of the Building Energy, Technical Status and Indoor Environment (BETSI) study which consists of a strati ed random sample of all multi-family buildings in Sweden. The rst aim of our study was to estimate the prevalence of noise annoyance (tra c noise and neighbourhood noise) and stress-related symptoms. The second aim was to study associations between noise annoyance and stress-related symptoms. The third aim was to study home environment factors in relation to noise annoyance.

BETSI study
The Swedish National Board of Housing, Building and Planning commissioned the BETSI (buildings, energy use, technical status and indoor environment) study in 2006, aiming to obtain representative information of the status of Swedish buildings, as well as indoor environment in relation to occupants' health.
The selection of buildings was conducted by Statistics Sweden (SCB) by using a multi-stage sampling procedure (26). Totally 30 Swedish municipalities were selected from a total of 290 municipalities across Sweden through a strati ed random selection procedure, taking into account geographic and demographic characteristics (temperature zone and region), and degree of urbanization. Data on all buildings and their construction year from these 30 municipalities were obtained from the central building register in Sweden. Two samples of residential buildings were created: single-family houses and multi-family buildings.
Strati ed random sampling was applied to sample buildings based on the construction year in ve classes (before 1960, 1960-1975, 1976-1985, 1986-1995 and 1996-2005) aiming to get the same number of buildings in each age class. This leads to an over-sampling of new buildings, since most buildings in Sweden are old.

Study population and questionnaires
The present study consisted a subsample of 690 multi-family buildings (including 8841 apartments). The number of adults (≥ 18 years old) living in each apartment was identi ed by SCB from the Swedish civil registration register. All adults (≥ 18 years) registered in the selected apartments received a personal questionnaire including medical questions and personal factors. Moreover, one indoor environment questionnaire was sent to each apartment. The questionnaires were developed at Department of Medical Science, Uppsala University, based on previous studies (27)(28)(29)(30)(31)(32). The postal questionnaire was administered by SCB in the spring of 2008. Two reminders were sent to those who did not reply the rst time. Totally 5775 adults participated in the current study and returned the personal questionnaire (46%), and a total of 4369 indoor environment questionnaires (49%) were returned.

Questions on stress-related symptoms
One question included in personal questionnaire asked: In the last three months, have you had any of the following symptoms? This question was followed by a list of symptoms. The current study included three stress-related symptoms: (1) tiredness; (2) headache and (3) di culty concentrating. There were three alternatives for each symptom: yes, often (every week); yes, sometimes; no, never. Those questions originated from the MM-questionnaire from Örebro University Hospital in Sweden (31). Moreover, gender and age were asked in the personal questionnaire.

Questions on noise annoyance
In the personal questionnaire, there were ve questions about noise annoyance at home: (1) Have you been disturbed by noise at home in the last three months? (yes, often (every week); yes, sometimes; no, never); (2) How much have you been disturbed by sound/noise in the last three months from inside of the building, from: a) lines and pipes; b) ventilation/fans inside; c) voice/radio/TV/music/similar sounds from neighbours; d) scraping sound/footsteps/thumping/similar sounds from neighbours; e) amusement centre in the property; f) stairwell or elevators. There were sex alternatives for each sub-question: not at all being disturbed; being disturbed not very much; being disturbed a little bit; being disturbed much; being disturbed very much; not existed.
(3) How much have you been disturbed by sound/noise in the last three months from outside of the building, from: a) ventilation/fans/heat pumps; b) road tra c; c) train tra c; d) ight tra c. There were sex alternatives for each sub-question: not at all being disturbed; being disturbed not very much; being disturbed a little bit; being disturbed much; being disturbed very much; not existed.
(4) Does the tra c noise (road, train or ight tra c) cause some of the following effects (severe tra c noise effects)?: a) di cult to hear radio/TV; b) telephone calls being affected; c) normal conversations being affected; d) rest/relaxation being disturbed; e) di culties in sleeping; f) being woken up from tra c noise.
A continuous noise score (0-10) was created by adding the number of yes answers to question (2) noise annoyance from sources inside of the building (sex sub-questions) and question (3) noise annoyance from sources outside of the building (four sub-questions). The categorized noise score (0-3) was created: coded as 0 if the continuous score equals 0; coded as 1 if the continuous score equals 1; coded as 2 if the continuous score equals 2; coded as 3 if the continuous score equals 3 to 10.

Questions on home environment
Data on home environment factors were gathered through the questionnaires, including: (1) Time of been living in the current apartment (0-5 years; more than 5 years); (2) Total number of teenagers at home; (3) Total number of children at home; (4) Total number of persons (children or adults) at home; (5) Size of the home (m 2 ); (6) Ownership of the current apartment (self-owned; renting); (7) Location of the apartment (at ground oor/basement; above ground oor); (8) Any Mechanical ventilation (yes; no); (9) Bathroom fan (yes; no); (10) Window opening frequency (everyday; less often).
A variable "crowdedness" was calculated based on total number of persons at home and size of the home (persons/100 m 2 ). Another variable "living alone" was calculated by selecting homes with only one occupant.
Data on building construction year was collected from the National Building Register, Real Property Register and were categorized into ve groups: before 1960, 1961-1975, 1976-1985, 1986-1995 and 1996-2005. Data on population and area of each municipality were obtained from wikipedia website (33).

Statistical analysis
Chi-square test was applied to compare the prevalence of noise annoyance sources and stress-related symptoms between subgroups (strati ed for gender, age, ownership). Initially, factor analysis was performed for two noise factors, including noise from inside of the building (sex variables) and noise from outside of the building (four variables). Then, factor analysis was applied for sex severe tra c noise effects. Two-level logistic regression models (individual, municipality) were performed to estimate associations between noise annoyance sources and stress-related symptoms. Interaction analyses were performed to investigate how gender, age, smoking and ownership affect the associations between noise annoyance sources and respective stress-related symptom. Similar two-level models were conducted to estimate the associations between categorized noise score and stress-related symptoms. Furthermore, similar two-level models were applied to estimate associations between building and home environment factors and different noise annoyance sources. Moreover, mutual adjustment logistic regression model (two-level) was performed including all exposure variables (building and home environment factors) with p < 0.2, which were selected from the respective single exposure model.
The statistical analysis was conducted with STATA 15.1 (STATA Corp, Texas, USA) and SPSS 25.0 (SPSS Inc., Chicago, IL, USA). Associations were expressed as odds ratios (OR) with a 95% con dence interval (CI) for ordinal regression and logistic regression models. In all statistical analyses, two-tailed tests and a 5% level of signi cance were applied, except for the interaction analyses. For the interaction analyses, two-tailed tests and a 10% level of signi cance were applied.

Results
A total of 5775 personal (46%) and 4369 indoor environment (49%) questionnaires were returned. Among the returned personal questionnaires, 73 had no information on gender, and 49 had no information on age. The comparison between participants and non-participants was performed by SCB Sweden, and have been published previously (35). All differences between participants and non-participants were small, except for age and civil status. Older persons and married persons had a higher participation rate. Among the participants: 56.5% were females; 63.5% aged between 18 to 65 years old, 36.5% were older than 65 years old; 12.0% were current smokers.
Prevalence of noise annoyance and stress-related symptoms are shown in Table 1, strati ed for gender, age (18-40 y, 40-65 y and over 65 y) and ownership.
Totally 11.2% reported any frequent noise at home in the last three months. Most commonly reported noise from inside of the building were voice/radio/TV/music/similar sounds from neighbours (13.2%) and scraping sound/footsteps/thumping from neighbours (16.5%). Road tra c was the most commonly reported noise problem from outdoor sources (16.1%). The most common reported severe tra c noise effects were being di cult to hear radio/TV (3.3%), rest/relaxation being affected (3.8%), di culties in sleeping (3.3%) and being woken up from tra c noise (3.5%). Tiredness was the most common symptom among participants (23.1%). Reporting of headache (8.5%) and di culty concentrating (5.5%) were less common. Females had more symptoms than males (for all symptoms p < 0.001). In general, younger participants (≤ 65 y) and participants living in rented apartments reported more noise problems as well as more stress-related symptoms.
Associations between noise annoyance and stress-related symptoms are shown in Table 2. All types of noise annoyance were associated with tiredness, headache or di culty concentrating. Moreover, a higher noise score was related to increased risk of all stress-related symptoms (Table 3). e Any consequence caused by tra c noise was de ned as reporting "often" to any of the following tra c noise caused consequences: di cult to hear radio/TV, telephone calls being affected, normal conversations being affected, rest/relaxation being affected, di culties in sleeping and being woken up from tra c noise. Table 2 Associations between noise annoyance and weekly tiredness, headache and di culty concentrating OR(95%CI) a . a Two-level logistic regression models (individual, municipality). The odds ratios were adjusted for gender, age and smoking. b Any consequence caused by tra c noise was de ned as reporting "often" to any of the following tra c noise caused consequences: di cult to hear radio/TV, telephone calls being affected, normal conversations being affected, rest/relaxation being affected, di culties in sleeping and being woken up from tra c noise. Table 3 Associations between categorized noise score and weekly tiredness, headache and di culty concentrating OR(95%CI) a . a Two-level logistic regression models (individual, municipality). The odds ratios were adjusted for gender, age and smoking. b A continuous noise score (0-10) was created by adding the number of yes answers to noise from sources inside of the building (sex sub-questions) and noise from sources outside of the building (four sub-questions). The categorized noise score (0-3) was created: coded as 0 if the continuous score equals 0; coded as 1 if the continuous score equals 1; coded as 2 if the continuous score equals 2; coded as 3 if the continuous score equals 3 to 10.
Interaction analyses were applied to study if gender, age, smoking and ownership affect the associations between noise annoyance and respective stressrelated symptom. Females were more often tired than men if exposed to tra c noise causing severe effects (p for interaction 0.013). Older subjects got more often headache if exposed to noise at home (p for interaction 0.053 with age as continuous variable). Moreover, older subjects got more often headache if exposed to tra c noise causing severe effects (p for interaction 0.003 with age as continuous variable). Subjects owning their apartment were more often tired if exposed to noise at home, as compared to those renting their apartment (p for interaction 0.036). Moreover, subjects owning their apartment got more often headache if exposed to tra c noise causing severe effects (p for interaction 0.054). No interaction by smoking was observed on the associations between different types of noise annoyance and headache, tiredness or di culty concentrating (p for interaction > 0.1).
The initial factor analysis identi ed three factors: the rst factor included voice/radio/TV/music/similar sounds from neighbours, scraping sound/footsteps/thumping/similar sounds from neighbours and noise from stairwell or elevators; the second factor included noise from ventilation/fans inside and noise from ventilation/fans/heat pumps from outside of the building; the third factor included noise from road tra c, train tra c and ight tra c. Noise from lines and pipes and noise from amusement centre in the property were not included in any factor. The second factor analysis for severe tra c noise effects identi ed two factors: the rst factor included di cult to hear radio/TV, telephone calls being affected and normal conversations being affected; the second factor included rest/relaxation being disturbed, di culties in sleeping and been woken up from tra c noise.
The majority of the participants were living in warmer climate zones (median value for temperature zone is 3). The median value of municipality population density is 1179 person/km 2 and the median value of crowdedness is 2.3 person/100m 2 . Over half (52.6%) of the participants have been living in the current apartment more than 5 years. Totally 6.8% reported any teenager at home and 11.7% reported any children at home. Less than half (40.9%) of the participants were living alone. About half (50.7%) of the participants had rented apartments. The majority of the participants were living above ground oor (81.8%). Apartments with mechanical ventilation accounted for 45.4%. Only 14.6% had a bathroom fan. A total of 73.9% ventilated their homes on a daily basis by opening windows (Table 4).
Construction year 1961-1975, living alone and renting were all associated with any frequent noise at home. Any mechanical ventilation was related to less disturbing noise at home. Three risk factors were identi ed in the mutual adjustment model including all variables with p < 0.2 in Building and home environment factors in relation to noise annoyance from sources inside of the building were analysed (see Table S1 and Table S2 in Additional le 1). Mutual adjustment model including all variables with p < 0.2 in Table S1 (see Additional le 1) were performed (  ight tra c. Renting and living above ground oor were related to more reporting of any severe tra c noise effect.   Table S1 (see Additional le 1). The odds ratios were adjusted for gender, age and smoking.
b The ORs were expressed per 1000 increase for municipality population density (number of persons per km 2 , person/km 2 ). Table 6 Associations between environment factors and noise annoyance from sources inside (much vs. less/not at all) OR(95%CI) a .  Table S2 (see Additional le 1). The odds ratios were adjusted for gender, age and smoking.
b The ORs were expressed per 1 unit increase for temperature zone.
c The ORs were expressed per 1000 increase for municipality population density (number of persons per km 2 , person/km 2 ). Table 7 Associations between environment factors and noise annoyance from sources outside (much vs. less/not at all) OR(95%CI) a . The prevalence of any frequent noise at home and any severe tra c noise effect in different construction periods, strati ed by ownership (self-owned/renting) are shown in Fig. 1 and Fig. 4, respectively. The prevalence of severe tra c noise effects was shown in Fig. 3, strati ed by ownership. Participants living in buildings constructed during 1961-1975 and 1976-1985 reported more noise problems ( Fig. 1 and Fig. 4). Reported any frequent noise at home, severe tra c noise effects and any severe tra c noise effect were all higher among those living in rented apartments as compared to those living in self-owned apartments ( Fig. 1, Fig. 3 and Fig. 4). Increased municipality population density (indicated by four quartiles) was related to a higher prevalence of any frequent noise at home (Fig. 2) and a higher prevalence of any severe tra c noise effect (Fig. 5).
This study shows that noise annoyance from different noise sources and stress-related symptoms (including tiredness, headache and di culty concentrating) were common among occupants. Stress-related symptoms were associated with all types of noise annoyance. Exposure to many noise sources (a higher noise score) was related to more symptoms.
We found that personal factors, climate, degree of urbanization as well as building related factors can in uence noise annoyance. Personal factors, such as younger age, female gender, living alone and a longer living time in the current residence were related to more noise annoyance. Females reported more noise from neighbours and were more sensitive to tra c noise causing severe effects. Occupants living in a warmer climate were more likely to be annoyed from noise at home, noise from amusement centre and noise from stairwell/elevators. Higher degree of urbanization (higher municipality population density) was related to more noise annoyance from neighbours, installations (lines and pipes), ventilation/fans/heat pumps, train tra c and ight tra c. Occupants living in rented dwelling were more annoyed from noise at home, noise from neighbours, amusement centre, installations (lines and pipes and stairwell/elevators), ventilation/fans/heat pumps, road tra c and train tra c. More noise annoyance were reported among those living in buildings constructed from 1961-1985.
Apartment situated above ground oor was related to more complaints of noise from neighbours and tra c noise causing severe effects. Having mechanical ventilation at home was related to more complaints of noise from ventilation/fans/warm pumps. Daily window opening was related to less noise annoyance from ventilation/fans/warm pumps but more noise annoyance from ight tra c.

Prevalence of noise annoyance from building sources
Reporting of any frequent noise at home was common in our study (11.2%). We found only one previous study from Sweden investigating the prevalence of noise annoyance from different sources among residents. The study suggested that complaints of annoyance to tra c noise (8.7%) and sound from neighbours (7.8%) were most common among participants, followed by sounds from ventilation systems (3.9%), sounds from other installations (2.4%) and industry noise (1.1%) (36).
Totally 13.2% reported voice/radio/TV/music/similar sounds from neighbours and 16.5% reported scraping sound/footsteps/thumping/similar sounds from neighbours. The prevalence found in our study is somewhat higher as compared to one previous Swedish study (36). Reporting of noise from amusement centre in the property was less common in our study (2.1%).
A total of 8.9% reported noise from lines and pipes, and 10.5% reported noise from stairwell/elevators. The prevalence of noise from installations in our study are higher as compared to one previous Swedish study, which showed a prevalence of 2.4% for noise from installation (36). Another study from Sweden suggested that the proportion of persons who were annoyed of noise from installations was more than twice as high as for tra c noise (25). The noises generated by installations, such as ventilation and air-conditioning systems, often have a dominant portion of low frequencies  and may be a source of great annoyance and sleep disturbance. Thus, it can be important to regulate the noise exposure from sources such as installations.
We found that 7.6% were annoyed by noise ventilation/fans inside, and 4.5% were annoyed by noise from ventilation/fans/heat pumps from outside of the building. The prevalence found in our study is slightly higher than data from one previous Swedish study (36).

Prevalence of noise annoyance from tra c
We found that 16.1% reported noise from road tra c, 3.1% reported noise from train tra c and 3.3% reported noise from ight tra c. Road tra c and railway noise contribute signi cantly to the burden of disease in Sweden each year (use disability-adjusted life-years (DALY) measure) (37). The most important contributor to the disease burden (DALY) was sleep disturbances (54%), followed by annoyance (30%) and cardiovascular diseases (16%) (37). Aircraft noise can be the most annoying noise source among all transportation noise sources when standardized for noise exposure level (5, 38, 39).

Home environment factors in association with severe tra c noise effects
The factor analysis for severe tra c noise effects identi ed two factors: the rst factor included di cult to hear radio/TV, telephone calls being affected and normal conversations being affected; the second factor included rest/relaxation being affected, di culties in sleeping and being woken up from tra c noise.
We found that renting and living above ground oor were associated with severe tra c noise effects. Rented apartment buildings in Sweden were mostly built in 1960's and 1970's with poor internal sound insulation (thinner walls and only double glass windows). Those buildings are more likely to be located in poor areas such as major roads or train tracks. There can be noise remedies such as concrete walls to be placed between buildings and major roads/train tracks to reduce tra c noise. This kind of noise remedies can be effective to reduce noise levels for people living at ground oor or basement but less or not effective for people living above ground oor. Living at basement can be less disturbed by noise as small windows are often used in basement. Moreover, the ground connected to basement can help to reduce noise.
Noise is an environmental stressor that causes annoyance and interferes with daily activities (1). One Swedish study suggested that having access to quiet indoor and outdoor sections in one's dwelling was related to less activity interference among residents (40). Studies from Sweden and Brazil have investigated tra c noise in association with daily activity disturbances. Indoor activity disturbances (conversation, radio/TV, concentration, rest/relaxation, di culties in falling asleep, awakenings, does not keep window open as often as desired) due to road tra c noise were much higher among those living in areas close to a main road than those living in quite areas in Sweden (41). Complaints of activity disturbances decreased in both areas after a two year period due to reduced tra c volume in both areas (41). Living in areas exposed to tra c noise, being aware of tra c noise in the home and considering the place of residence a noisy environment were predictors of noise annoyance among residents in Brazil (42). Living in areas exposing to tra c noise was associated more annoyance during activities, including watching TV, resting, talking and concentration (42).

Noise annoyance from indoor and outdoor sources in association with stress-related symptoms
Tiredness was the most commonly reported symptom in our study (23.1%), followed by headache (8.5%) and di culty concentrating (5.5%). We found that noise annoyance from noise at home, from sources outside of the building (road tra c, train tra c, ight tra c and ventilation), from sources inside of the building as well as reported severe tra c noise effects were all associated with stress-related symptoms (tiredness, headache and di culty concentrating). Moreover, a higher noise score was related to increased risk of stress-related symptoms.
People annoyed by noise may experience a variety of negative responses, such as anger, disappointment, dissatisfaction, withdrawal, helplessness, depression, anxiety, distraction, agitation, or exhaustion (10,43). The symptoms included in our study can be considered as stress-related psychosocial symptoms. Few studies have shown that noise annoyance and its relation with headache/tiredness. One experiment study from Sweden found that noise annoyance among university students in laboratory environment was positively associated with self-reported headaches (44). Another study from Sweden found that annoyance from road tra c noise among residents was related to tiredness (41). Moreover, one Swedish study suggested that exposing to sound levels from road tra c ranging from 45-68 dBA at the most exposed side may induce tiredness among residents (40). Some studies indicated that tra c noise exposure or tra c noise induced annoyance were associated with mental stress or mental disorders among residents (40,41,45,46). The results in our study is in agreement with previous studies on noise annoyance from outdoor sources (mainly outdoor tra c) and stress symptoms/metal disorders. Our study indicate that noise annoyance from indoor sources (neighbours, installations and ventilation/fans/heat pumps) can affect stress-related symptoms, which is a new nding.

Age
Younger participants reported more noise annoyance as compared to older participants (> 65 y), but the interaction analyses found that older participants were more sensitive to noise than younger participants. Moreover, we found that stress-related symptoms were more prevalent among younger participants as compared to older participants (> 65 y). Few studies have investigated age in association with noise annoyance. One study from India found that middle-aged subjects were more annoyed by road tra c noise as compared to younger and older subjects (47). One study from France indicated that older participants were more likely to be highly annoyed by aircraft noise (11). Younger participants in our study can have higher demands on their home environment. Another explanation can be that the effect of noise-induced annoyance in older people is concealed by physical consequences of age, with a strong increase of illnesses such as hearing loss.

Gender
We found that males reported less voice/radio/TV/music/similar sounds as compared to females in the mutual adjustment model (OR 95%CI, 0.75(0.61,0.91), p = 0.004). No gender difference was found regarding other noise complaints. According the interaction analyses, females were more sensitive to any severe tra c noise effect than males. We found no previous studies investigating gender in relation to noise annoyance from indoor sources.

Living alone and living period in the current apartment
We found that living alone was related more complaints of noise at home. The reason can be that persons living alone are not adopted to hearing other persons in the family, so they can be more sensitive to noise. A longer living time in the current apartment (more than 5 years) was associated with more reports of noise from neighbours. People with a longer living period in the current home can have less tolerance to noise and can be more sensitive to noise from neighbourhood.

Climate zone
Living in a warmer climate was related to more noise annoyance from any frequent noise at home, noise from amusement centre in the property and noise from stairwell/elevators inside of the building in our study. We found no previous studies on climate and noise annoyance.

Urbanization
We found that increased municipality population density was associated with more noise annoyance from neighbours, installations (lines and pipes), ventilation/fans/heat pumps, train tra c and ight tra c. A higher municipality population density indicates a higher degree of urbanization. A higher degree of urbanization often associates with increased exposure to environmental noise, especially from more frequent human activities in neighbourhood, from installations/ventilation systems in buildings nearby and from increased tra c intesity. Therefore, our results implies that measures for reducing environmental noise exposure in high populated large cities are needed, even in country like Sweden with a low population density.

Ownership
Renting was associated with noise annoyance from different sources in our study, including any frequent noise at home, noise from neighbours, amusement centre, installations (lines and pipes, stairwell/elevators), ventilation/fans/heat pumps, road tra c and train tra c. Previous Swedish study indicated that a large amount of multi-family buildings are owned by the community and used for renting (48). Multi-family buildings constructed from 1961-1975 in Sweden are mostly rented and prioritized for renovation in nowadays. Those buildings often have poor sound insulation due to thinner walls, no triple glass windows and poorer insulation in oors. Moreover, they can be located in areas with more tra c noise because such areas have lower land price. Families with lower socio-economic status are more likely to live in those buildings. Moreover, people living in rented building areas are more likely to have less social control as compared to those living in self-owned building areas. We found that participants living in self-owned apartments were more sensitive to any frequent noise at home and any severe tra c noise effect. This indicates that the associations between renting and noise annoyance indicators in our study cannot be explained by high noise sensitivity among participants living in rented apartments. It is more likely that participants living in rented apartments were exposed to more noise than those living in self-owned apartments (worse sound insulation, more noise from outside, worse social control/less respect between tenants).

Building construction year
We found that construction year from 1961-1985 was more likely to be associated with noise annoyance in general. We have found previously that subjects living in multi-family buildings built during 1961-1985 in Sweden had more respiratory illnesses (35). Multi-family buildings from 1961-1975 were constructed during a building boom in large cities in Sweden in uenced by energy saving demands. New construction techniques and building materials were applied to reduce energy consumption during this period (48). It was reported that the prevalence of dampness problems was highest among those buildings (49). Those buildings were mostly poor quality high-rise buildings rented by poor people. They were constructed with thinner walls, poorer insulation in oors and only double glass windows (no triple glass windows). In Sweden, apartment buildings from 1976-1985 often have thinner walls with less sound insula and double glass windows. Newer apartment buildings have better insulation and tripple glazing. Thus, poor sound insulation can be the main reason to noise annoyance found in our study.

Other building related factors
Having mechanical ventilation was associated more complaints of noise from ventilation/fans/warm pumps. Noise from mechanical ventilation system can disturbs, especially in buildings with poor sound insulation. Reducing noise impact on residents when installing a mechanical ventilation system can be important.
Having a bathroom fan at home was related to noise annoyance from amusement centre. We have no explanation to this association which could be a chance nding.
Daily window opening was related to less complaints of noise from ventilation/fans/warm pumps. People with the habit of frequently opening windows can be less sensitive to constant noise from ventilation systems. In our study, daily window opening was associated with more complaints of noise from ight tra c. The reason is unknown. One study reported that residents who kept living room windows open were shown to be less annoyed by aircraft noise (50), which is opposite to our nding.

Strengths
The present study is one of a few studies on noise annoyance that covered a sample of a whole country. All subjects (≥ 18 years) living in multi-family buildings were included in the present study, with no prior information on their health status. There was no major difference in participation rate between different municipalities, between those being born in Sweden and foreign-born persons, and between Swedish citizens or non-Swedish citizens. Most of the differences in participation rate were small (35). Some factors discussed such as temperature zone and population density were not obtained from the selfadministered questionnaire, which reduces potential reporting bias.

Limitations
The response rate was not very high in our study (46%). The participation rate was higher among elderly and married persons, but slightly lower in larger cities, suburban municipalities and in older buildings (35). Information bias can occur in questionnaire surveys. Subjects may overestimate or underestimate environment risk factors as well as stress-related symptoms. We found some associations between speci c home environment risk factors and noise annoyance, not a similar association for all risk factors. Thus, our ndings are less likely to be seriously biased by selection or information bias.

Conclusions
As shown in the present study, noise annoyance from different indoor and outdoor sources were common among occupants living in multi-family buildings in Sweden. Noise annoyance was related to tiredness, headache and di culty concentrating. Exposure to noise from many sources was related to a higher risk of symptom.
Younger participants reported more noise annoyance and more stress-related symptoms as compared to older participants. Females can be more sensitive to noise from neighbours and tra c noise causing severe effects. Occupants living alone complained more often noise at home. Occupants having a longer living time in the current residence reported more often noise from neighbours. Living in a warmer climate was related to more noise annoyance from noise at home, noise from amusement centre and noise from stairwell/elevators. Higher municipality population density was associated with more noise annoyance from neighbours, installations (lines and pipes), ventilation/fans/heat pumps, train tra c and ight tra c. Renting was associated with noise annoyance from noise at home, noise from neighbours, amusement centre, installations (lines and pipes and stairwell/elevators), ventilation/fans/heat pumps, road tra c and train tra c. Occupants living in buildings constructed from 1961-1985 reported in general more noise annoyance. Living above ground oor was related to more complaints of noise from neighbours and tra c noise causing severe effects. Having mechanical ventilation was associated more noise annoyance from ventilation/fans/warm pumps. Daily window opening was associated with less complaints of noise from ventilation/fans/warm pumps but more complaints of noise from ight tra c. Our study suggested that personal factors (younger age, female gender, living alone, a longer living period in current apartment), a warmer climate zone, higher population density as well certain building factors (including renting, construction year from 1961-1985, living above ground oor, mechanical ventilation systems) were associated with more noise annoyance among occupants. The study was performed in accordance with the Declaration of Helsinki. The study and the consent procedure were approved by the Regional Ethical Committee in Uppsala, Sweden (reference number 2007-281). Informed consent was obtained from all participants. An information letter sent together with the questionnaire stated that if the subjects answered and returned the questionnaire it meant they had given informed consent.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Figure 2
Prevalence of any frequent noise at home in relation to urbanization, strati ed by ownership.

Figure 3
Prevalence of different severe tra c noise effects, strati ed by ownership.