Sick building syndrome and associated risk factors among the population of Gondar town, northwest Ethiopia

Background Sick building syndrome (SBS) consists of a group of mucosal, skin, and general symptoms temporally related to residential and office buildings of unclear causes. These symptoms are common in the general population. However, SBS symptoms and their contributing factors are poorly understood, and the community associates it with bad sprits. This community-based cross-sectional study was, therefore, conducted to assess the prevalence and associated factors of SBS in Gondar town. Methods A community-based cross-sectional study was conducted from March to April 2017. A total of 3405 study subjects were included using multistage and systematic random sampling techniques. A structured questionnaire and observational checklists were used to collect data. SBS was assessed by 24 building-related symptoms and confirmed by five SBS confirmation criteria. Multivariable binary logistic regression analysis was used to identify factors associated with SBS on the basis of adjusted odds ratio (AOR) with 95% confidence interval (CI) and p < 0.05. The Hosmer and Lemeshow goodness of fit test was used to check model fitness, and variance inflation factor (VIF) was also used to test interactions between variables. Results The prevalence of SBS in Gondar town was 21.7% (95% CI = 20.3–23.0%). Of this, the mucosal symptoms account for 64%, the general symptoms account for 54%, and the skin symptoms account for 10%. From study participants who reported SBS symptoms, 44% had more than one symptom. Headache (15.7%), asthma (8.3%), rhinitis (8.0%), and dizziness (7.5%) were the commonest reported symptoms. SBS was significantly associated with fungal growth in the building [AOR = 1.25, 95% CI = (1.05, 1.49)], unclean building [AOR = 1.26, 95% CI = (1.03, 1.55)], houses with no functional windows [AOR = 1.35, 95% CI = (1.12, 1.63)], houses with no fan [AOR = 1.90, 95% CI = (1.22, 2.96)], utilization of charcoal as a cooking energy source [AOR = 1.40, 95% CI = (1.02, 1.91)], cooking inside the living quarters [AOR = 1.31, 95% CI = (1.09, 1.58)], and incensing and joss stick use [AOR = 1.48, 95% CI = (1.23, 1.77)]. Conclusion The prevalence of SBS in Gondar town was high, and significant proportion of the population had more than one SBS symptom. Headache, asthma, rhinitis, and dizziness were the commonest reported SBS symptoms. Fungal growth, cleanliness of the building, availability of functional windows, availability of fan in the living quarters, using charcoal as a cooking energy source, cooking inside the quarters, and incensing habit or joss stick use were identified as factors associated with SBS. Improving the sanitation of the living environment and housekeeping practices of the occupants is useful to minimize the prevalence of SBS.


Background
Housing is one of the basic needs of a human being and fundamental for wellbeing [1,2]. People spend more than 90% of their time indoors [3]. Housing increasingly becomes a major public health concern. For many years, the housing environment has been acknowledged as one of the main settings that affect human health. Indoor air quality, home safety, noise, humidity and mold growth, indoor temperatures, asbestos, lead, radon, volatile organic compounds (VOC), lack of hygiene and sanitation equipment, and crowding are some of the most relevant possible health threats in dwellings [4][5][6].
Physical, mental, and social health is affected by the living conditions. The quality of housing conditions plays a decisive role in the health status of the residents. Many health problems are either directly or indirectly related to the building itself, because of the construction materials that were used and the equipment installed or the size or design of the individual dwellings [4,7]. These building-related health problems are categorized in building-related diseases and sick building syndrome (SBS) [8]. SBS consists of a group of mucosal, skin, and general symptoms that are temporally related to residential and office buildings. SBS comprises a group of symptoms of unclear causes divided into mucous membrane symptoms related to the eyes, nose, throat, and dry skin, together with what are often called general symptoms of headache and lethargy [7][8][9]. SBS can be influenced by a variety of factors, like building-related factors (air-conditioned building, fresh air ventilation rates, indoor temperature, poor building service maintenance and cleaning, relative humidity) [7,[10][11][12][13], environmental factors and pollutants [VOCs (formaldehyde, solvents, etc.), carbon monoxide (stoves, heaters, and furnaces), dust and fibers (asbestosis, fiberglass, dirt), bio-aerosols (bacteria, molds, viruses, pollen, dust, mites, animal danders, animal excreta), trapped outdoor pollutants (vehicle or industrial exhausts), physical factors (lighting, vibration, noise, temperature, crowding, photo duplication)] [14][15][16][17], and personal factors (gender, history of being allergic, job dissatisfaction, cigarette smoke, increased use of computers) [18][19][20][21][22].
SBS symptoms are common in the general population. However, SBS symptoms and their contributing factors are poorly understood, especially in developing countries including Ethiopia. The community associates SBS with bad spirit. This community-based cross-sectional study was, therefore, conducted to assess prevalence and contributing factors of SBS in Gondar town.

Study design and description of study settings
A community-based cross-sectional study was conducted from March to April 2017 in Gondar town.
Gondar town is located in the northern part of Ethiopia in Amhara National Regional State, North Gondar Zone, at a distance of 747 km from Addis Ababa and 170 km from Bahirdar at 12°45′ north latitude and 37°45′ east longitudes. Gondar was founded in 1643. Based on the 2016 population estimate, Gondar had a total population of 621,168 with 3200/km 2 [23].

Sample size determination
Single-population proportion formula [24] was used to determine the sample size (n) with the following assumptions: p (prevalence of SBS = 50%, hence there are no other similar studies in Ethiopia), 95% confidence interval (standard normal probability = 1.96); 3% margin of error or maximum error to commit, z = the standard normal tabulated value, and α = level of significance.
We used a design effect of 3 based on the recommendation of the Population Services International research tool kit [25] and 7% non-response rate; the final sample became 3429.

Sampling procedures
Multistage sampling technique was used to select study participants. Seven kebeles (the lowest administrative unit in Ethiopia) were selected from a total of 22 kebeles using simple random sampling technique. Residential buildings found in the selected kebeles were chosen using systematic random sampling technique in every seven interval. The first residential building was selected from seven houses by lottery method.

Data collection procedures
A structured questionnaire and observational checklist were used to collect data. The questionnaire was pre-tested out of the study area in a community which had similar characteristics prior to the actual data collection. Twelve graduating class environmental health students were involved in the data collection process. Training was given for the data collectors and supervisors. The data collectors visited all systematically selected households and interviewed all household members. For under-5-year-old household members, data collectors interviewed mothers or caregivers. Data collectors asked the study subjects to recall the presence of 24 SBS symptoms and related information in 3 months prior to the survey. Collectors also observed the housing and living environment condition. The overall interview process was supervised by supervisors. The collected data were checked and corrected by the data collectors immediately after finalizing the questionnaire before they left the house. Supervisors daily checked the completeness, quality, and consistency of information collected.

Measurement of study variables
SBS, the primary outcome variable of this study, was defined as the presence of at least one symptom associated with housing condition in the last 3 months prior to the survey. SBS was assessed by asking have you had any (or more) of the following symptoms during the last 3 months: (i) general symptoms including fatigue, headache, dizziness, reduced attention, hyperactivity, fever, chills, and eye strain; (ii) mucosal symptoms including rhinitis, nasal congestion, wheezing, asthma, dyspnea, severe lung disease, epistaxis, upper respiratory tract irritation, chest tightness, dry throat, cough, and eye irritation; and (iii) skin symptoms including skin rashes, dry or flushed facial skin, scaling/itching scalp or ears, and lip dryness [9,26,27]. Five criteria were used to confirm whether the symptoms are SBS or not. The criteria were as follows: (a) symptoms aggravate when staying at home, (b) symptoms either immediately or gradually disappear when leaving the house, (c) symptoms recur when returning home, (d) symptoms aggravate during the night, and (e) symptoms disappear when the room is ventilated or cleaned. The wealth index of households was determined using principal component analysis (PCA). As health and demographic surveys (DHS) recommended, we used asset and service variables to determine the wealth index. Initially, the wealth index was classified into very poor, poor, moderate, rich, and very rich. But there was no significant difference between very poor and poor, and moderate, rich, and very rich. Based on this fact, the wealth index was classified into poor and rich. Residential buildings were taken as clean if the physical structures (floors, walls, ceilings, or roofs) have no any visible dirt, soot, spider's wrap, crack, and dampness. The living compound was taken as clean if the living environment is free from wastes, vectors, and unpleasant odor. The illumination system of the living quarters was taken as adequate if the light energy is constant, free from glare, uniformly distributed to the entire room, and suitable to perform daily activities inside without strain.

Data management and statistical analysis
Data were entered using EPI-INFO version 7 and exported into SPSS version 20 for further analysis. For most variables, data were presented by frequencies and percentages. Univariable binary logistic regression analysis was used to choose variables for the multivariable binary logistic regression analysis, variables with p value less than 0.05 by the univariable analysis were then analyzed by multivariable binary logistic regression for controlling the possible effect of confounders (like age, sex, family size, economic status, and education status of the family), and finally, variables which had significant association with SBS were identified on the basis of AOR with 95% CI and p < 0.05. The Hosmer and Lemeshow goodness of fit test was used to check model fitness. VIF was also used to test interactions between variables.

Socio-demographic information
A total of 964 residential buildings were visited, and a total of 3405 residents in these houses were included in this study with 99.4% response rate. One thousand eight hundred thirty-eight (54%) study subjects were female. Three fourth, 2620 (76.9%), of the study subjects were aged between 15 and 64 years. Nine hundred thirty-two (27.4%) participants graduated from colleges or universities. One thousand three hundred ninety-one (40.9%) study subjects were not engaged at the time of the survey, and 1179 (34.6%) participants were self-employed. Two thousand twenty-three (59.4%) study subjects were economically poor (Table 1).

Housing condition
The majority, 2389 (70.2%), of the study subjects lived in houses constructed from wood and mud. One thousand seven hundred eighty (52.3%) study subjects reported that the floor of their house is earthen floor. Two thousand eighty-six (61.3%) study subjects said that they had only one bedroom. The majority, 3217 (94.5%), of the study participants reported that they had no fan in their house. Nearly two thirds, 2211 (64.9%), of the study subjects said that their residential building has no functional windows. One thousand seven hundred sixty-six (51.9%) study subjects lived in houses in which the illumination system was not adequate. Almost all, 3317 (97.4%), of the study participants reported that they got light from electricity. A quarter, 879 (25.8%), of the study participants had pets in the home. Fungal growth and dampness was observed among 1462 (42.9%) participants' house. One thousand three hundred seventy-six (40.4%) study subjects reported that they recently used pesticides, paints, and solvents. One thousand eight hundred eighty-five (55.4%) participants reported that cooking inside the living quarters is a common practice. One thousand three hundred twenty-seven (39%) study subjects said that outdoor air pollutant sources (like garages, and metal and woodwork houses) were found around their home (within a 200-m radius). Incensing and utilization of a joss stick was a common habit in 1329 (39%) participants' house. Nearly one tenth, 335 (9.8%), of the study subjects reported that at least one of the family members smokes cigarette. The houses of 1196 (35.1%) study subjects were newly constructed (within 10 years), and the houses of 1295 (38%) subjects were clean ( Table 2).

Sanitation of the living environment
Nearly two thirds, 2192 (64.4%), of the study participants used traditional pit latrine, and three fourths, 2600 (76.4%), of study subjects got drinking water from an in-compound tap. Three thousand sixteen (88.6%), 43 (1.3%), and 1892 (55.6%) of the study participants reported that they use charcoal, kerosene, and electricity, respectively, as household energy sources. A quarter, 887 (26%), of the participants reported vector infestation in their living environment, and the living environment of 2255 (66.2%) study subjects was clean ( Table 3).

Prevalence of sick building syndromes
From a total of 3405 participants included in this study, 738 participants reported one or more symptoms associated with poor housing condition. The prevalence of SBS in Gondar town was therefore found to be 21.7% (95% CI = 20.3-23.0%). Ninety (12.2%) occupants who had one or more symptoms reported that the symptoms always occurred, and 648(87.8%) occupants who had SBS said that the symptoms occurred sometimes in the last 3 months. From study participants who reported symptoms related with housing, 414 (56%) and 324 (44%) had one and more than one symptom/s, respectively. Four hundred seventy-four (64%), 401 (54%), and 72 (10%) of the reported symptoms were mucosal, general, and skin symptoms, respectively. Headache, 221 (15.7%); asthma, 116 (8.3%); rhinitis, 112 (8.0%); and dizziness, 106 (7.5%), were the commonest reported symptoms (Table 4).

Factors associated with sick building syndromes
Univariable binary logistic regression was used to choose variables for the final model on the basis of p values less than 0.05. Types of wall; fungal growth; cleanliness of the building; presence of functional windows; presence of fan in the quarters; cooking inside the quarters; charcoal use; habits of cigarette smoking; incensing habit and joss stick use; outdoor pollutant sources near the building; recent utilization of pesticides, paints, and solvents; and cleanliness of the living compound were variables selected for the final model. VIF was calculated considering one independent variable as the dependent variable turn by turn to test interactions between variables. The test result shows that VIF for all variables was below 3, threshold for collinearity diagnostics. This showed that there is no multicollinearity effect between independent variables. Table 5 shows variables associated with SBS. SBS was statistically associated with fungal growth in the building. The probability of having SBS was 1. 25

Discussion
This study found that prevalence of SBS was 21.7% (64% mucosal symptoms, 54% general symptoms, and 10% skin symptoms). The prevalence of mucosal, general, and skin symptoms reported by this study is higher than the findings of other studies in China which reported mucosal symptoms (7.1%), general symptoms (11.4%), and skin symptoms (4.4%) [27]. Another study in China reported lower prevalence (mucosal symptoms (35.1%), general symptoms (39.4%,) and skin symptoms (43.4%)) [9] compared to the current study with exception of skin symptoms. The current prevalence is lower than that report in a study in three North European cities. The prevalence reported by the later study was 30.8% (20% mucosal, 10% general, and 8% dermal symptoms) [28].  The finding of the current study is also lower than that of a study in China that reported 74.3% mucosal symptoms, 78.7% general symptoms, and 47.5% skin symptoms [29]. The variations of prevalence among different studies may be due to differences in housing and environmental conditions. Most of the houses in the current study area were substandard. The outdoor air in the settings of the other studies is polluted by industrial emissions compared with the current setting. This study depicted that cleanliness of residential buildings was statistically associated with SBS. The prevalence of SBS was higher among occupants who lived in unclean buildings compared with occupants who lived in clean buildings. This finding is in line with the findings of other similar studies [7,9,19,30]. This may be due to the fact that unclean building surfaces including carpets accumulate dust and dirt, which are reservoirs for chemicals, allergens, and diseases causing pathogens [31,32].
SBS was significantly associated with infestation of fungus or molds in the living building. Occupants who lived in buildings where fungal growth was observed reported SBS compared with their counterparts. The finding of this study is supported by other studies [7,16,33]. This is due to the fact that fungus or molds cause health problems in the mechanisms of either infection or allergy or toxin. Fungal spores are generally recognized as important causes of respiratory allergies [16,34].
The current study explored that availability of functional windows is statistically associated with SBS. A wide range of literature also reported the effect of general ventilation on the health of the occupants [19,[35][36][37]. This fact can be explained that presence of functional windows as means of ventilating a building naturally helps the external fresh air to the living quarters and removes the internal exhausted air which in turn reduces the amount of contamination with chemicals or microorganisms, so that increased ventilation can be seen as an effective treatment of SBS [19,[37][38][39]. This study also revealed that availability of fan in the living quarters was significantly associated with SBS. The prevalence of SBS was higher among occupants who lived in houses with no fan. Other studies also reported similar findings [40,41]. This is because fan-assisted ventilation improves the quality of the indoor air [42,43].
This study reported that SBS was associated with household cooking energy sources, cooking practice, and incensing habits of occupants. The prevalence of SBS was higher among occupants who used charcoal as cooking energy source. Occupants who used charcoal and cooked inside the living quarters [44][45][46] and who used incensing and joss stick [27,29] had more chance to develop SBS. This can be justified that charcoal use and incensing habits are incomplete combustion processes that can generate gracious pollutants. Generally, cooking energy sources and cooking practices are the main sources for gracious pollutants to the indoor air [47,48].

Limitation of the study
This research did not assess the condition of office buildings in this study though the condition of office buildings is a covariate for SBS associated with residential buildings. Moreover, we did not measure indoor air quality, thermal condition, and light intensity using instruments. However, we used standardized checklists to assess these parameters.

Conclusion
The prevalence of SBS in Gondar town was found to be high, and a significant proportion of the population had more than one SBS symptom. Headache, asthma, rhinitis, and dizziness were the commonest reported SBS symptoms. Fungal growth, cleanliness of the building, availability of functional windows, availability of fan in the living quarters, using charcoal as cooking energy source, cooking inside the quarters, and incensing habits or joss stick use were identified as factors associated with SBS. Improving the sanitation of the living environment and housekeeping practice of the occupants is useful to minimize the prevalence of SBS.