Thermal Comfort Applied in Hospital Environments: A Literature Review

: The predicted mean vote (PMV) is the most widely used model around the world to assess thermal comfort in indoor environments. The year 2020 marks the 50th anniversary of the PMV model and also the year in which the World Health Organization (WHO) declared the COVID-19 outbreak a pandemic. In this context, hospital environments and health professionals are at the center of attention, and a good indoor environment for those professionals to develop their activities is essential. Thus, considering the PMV model and focusing on hospital environments, this study performed a literature review of studies published between 1968 and August 2020. The research identiﬁed 153 papers on thermal comfort and its application in hospitals, health centers, and elderly centers. Speciﬁc inclusion and exclusion criteria were adopted to determine the most relevant studies for the four research questions proposed in this study. After applying the exclusion criteria, 62 studies were included in order to identify their main characteristics. In the universe of the 62 studies, this review identiﬁed 24 studies that applied the PMV model and 12 where there was a comparison of PMV and the thermal sensation votes (TSV) reported by people. The main ﬁndings of this research are: (i) A good thermal environment for professionals and patients is important, and more studies are needed; (ii) there are little explored topics, such as productivity related to thermal comfort in hospital environments; (iii) in addition to thermal comfort, other indoor environmental quality (IEQ) parameters have also been evaluated, such as indoor air quality (IAQ); (iv): the COVID-19 pandemic has highlighted how the quality of indoor spaces is important in order to ensure occupant’s health.


Introduction
The predicted mean vote (PMV) is an index that shows the average thermal sensation of a large group of people exposed to the same environment [1]. This thermal comfort index was proposed by P.O. Fanger in 1970 and is used to evaluate the thermal sensation in moderate environments [2]. Currently, there is a growing need to evaluate indoor environments, given that in an environment with good thermal comfort, there is a significant improvement in people's health, wellbeing, and productivity [3,4]. Although thermal comfort is extremely relevant for occupants, buildings must be prepared not only to offer comfort to their users, but also to operate efficiently, since buildings are responsible for approximately one third of the total energy consumption throughout the world [5].
The year 2020 marks the 50th anniversary of the PMV model, which has been applied in different areas in recent decades to assess thermal comfort: the automotive sector [6], naval sector [7], construction [8], schools [9], universities [10], offices [11], and industry [12]. This year is also important paper performed a literature review with papers published from 1968 until August 2020, which apply the concepts of thermal comfort in hospitals, health centers, and elderly centers, aiming to answer four research questions (RQs) proposed in this paper, as well as to verify the main characteristics of these studies.

Methods
This literature review was performed considering a three-step methodology: the proposition of research questions (RQs) to guide the literature review, a method to search and select the studies, and a tool to perform content analysis. In the following subsections, this three-step methodology is explained in detail.

Research Questions (RQs)
The main goal of this study was to verify and summarize the studies that apply thermal comfort in hospital environments. In order to achieve this goal, 4 research questions (RQs) are proposed: (a) According to Djongyang et al. [23] and de Dear et al. [3], thermal comfort is required in indoor environments because it directly affects people's perception, in terms of health/wellbeing and productivity. On the other hand, Thapa et al. [24] claim that optimizing the energy used in buildings, whether for heating or for cooling, is a reality today, because there is a need for energy saving. Based on these premises, RQ1 is formulated: RQ1: Considering studies on thermal comfort in hospital environments, what are the main aspects that are taken into account: health/wellbeing, productivity or energy saving? (b) According to Humphreys and Nicol [25], the PMV model does not consider the adaptive actions that people undertake in indoor environments in order to maintain their comfort, leading PMV to underestimate or overestimate the real thermal sensation felt by people in buildings. Based on this premise, RQ2 is formulated: RQ2: Considering studies on thermal comfort in hospital environments, which relate PMV and real thermal sensation, studies indicate that PMV predicted well, underestimates or overestimates the real thermal sensation? (c) Different levels of activity require specific environmental conditions for people, in order to achieve thermal comfort. Thus, it is important to define the target group of the research [26]. In addition to this factor, the type of environment in which people are inserted in the hospitals usually has its own standardized environmental requirements, determined by the type of activity to be performed [21]. Based on these premises, RQ3 is formulated: RQ3: Considering studies on thermal comfort in hospital environments, what was the most evaluated group and which area within the hospital was most evaluated? (d) The hospital environment is complex, since it can change from waiting rooms to operating rooms and intensive care units (ICUs), which demand different requirements for environmental parameters due to the type of activity/care. The concern with planning the environment must go far beyond simply offering thermal comfort to its occupants [27]. Thus, the hospital environment should be prepared to offer a good indoor environmental quality (IEQ). Based on these premises, RQ4 is formulated: RQ4: What other parameters of indoor environmental quality (IEQ) were evaluated, in addition to thermal comfort?

Method for Bibliographic Search
Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) [28] guidelines were adapted to apply in this study. This method combines keywords and performs research in scientific information databases. Then, through a specific screening, it is possible to reduce the number of studies found, through specific and defined selection criteria for the research. Over the last few years, literature review studies on thermal comfort have been published using this method [29,30]. It has 4 steps to reduce the number of articles that will be selected: identification (step 1), screening (step 2), eligibility (step 3), and inclusion (step 4) for analysis.
As a search strategy and to identify the articles (step 1), it was decided to combine the following keywords, using Boolean operators, in the SCOPUS database ("thermal comfort" OR "thermal conditions" OR "predicted mean vote" OR "predicted percentage of dissatisfied") AND ("hospital" OR "health centers" OR "elderly centers"). The search occurred in the titles, abstracts, and keywords of the published studies, considering the time period from 1968 until August 2020. The final search was conducted on 25 August 2020. The SCOPUS database was chosen to perform this research to attend to the objective of this review, since the authors understand that all major journals publishing in thermal comfort are indexed there.
After identification, the screening phase started (step 2), where inclusion and exclusion criteria were applied in order to delimit the studies found and align with the RQs. Table 1 shows the inclusion and exclusion criteria that were adopted: Table 1. Inclusion and exclusion criteria.

Papers in English Papers in other languages
Articles published until 2020 Papers that do not provide the complete basic information (author, title, year of publication or source) Papers that focus on the relation of thermal comfort and hospital environments Papers in thermal comfort, but not in hospital environments Papers that might answer RQ1, RQ2, RQ3, and RQ4 Repeated papers After this screening, the next step consisted of a preliminary analysis of the selected articles with complete and accessible texts. Eligibility (step 3) consisted of reading the abstracts to verify if the selected articles might answer RQ1, RQ2, RQ3, and RQ4, this being a second refinement. After this second refinement, we were able to obtain the portfolio of articles included to perform the review (step 4).

Tool for Content Analysis
An open code environment developed to carry out thorough bibliometric analyses, Bibliometrix [31] was used. The Bibliometrix package, written in R language, provides a set of tools for scientific research using bibliometrics, and it was used in this study for data content. Also, provides several routines to import bibliographic data from the SCOPUS database, implementing bibliometric analysis and setting up data matrixes for co-citation, coupling, scientific collaboration analysis, and co-word analysis. Based on the selected articles, their main characteristics were analyzed, such as the number of articles per year of publication; most used keywords; most relevant authors; and most relevant sources/journals. Additionally, Critical Appraisal Skill Programme(CASP) checklists [32,33] were used in order to focus on studies that are relevant to obtain the answers to the proposed research questions. Each one of the articles was verified in order to verify it would be suitable to be included in the final portfolio for performing the review.

Preliminary Results
The results of the search strategy through the combinations of keywords in SCOPUS can be found in Table 2: Table 2. Results of search strategy.

Search Strategy Keywords and Combinations Results
Database: SCOPUS. Search in: Title, Abstract and Keywords.
Type of Article: Review and Article ("thermal comfort" OR "thermal conditions" OR "predicted mean vote" OR "predicted percentage of dissatisfied") AND ("hospital" OR "health centers" OR "elderly centers") 153 articles Based on the total of 153 articles found, the method described in Section 2.2 was applied. Figure 1 shows the application of the method: Thus, there are a total of 62 articles. Two articles are literature reviews on the topic and 60 are research articles. The two literature reviews that were previously published are available in Table 3, being the citations obtained from Google Scholar in August 2020: The first study [22], published in 2012, proposes to fill a gap on thermal comfort, since to date, no literature reviews on thermal comfort in hospitals have been published. With a focus only on hospitals, some health-related buildings fall outside of its scope, such as health centers and elderly centers. At the time, the authors considered the number of original works insufficient to determine the relationship between hospital staff productivity and thermal comfort, and they also considered it important to carry out comparative studies in more than one hospital.
The second and most recent study [34] focuses on the energy-saving aspect and its relationship with comfort in hospitals. The authors aimed to review only technologies to achieve energy efficiency. Research of this nature is important as energy demand for buildings has gained international prominence. The authors found that hospitals alone account for approximately 6% of total energy consumption in the public service sector. This review sought, unlike its predecessors, to evaluate articles related to thermal comfort in all kinds of hospital/healthcare environments, not just hospitals, to perform an updated review on the topic. Therefore, health centers and elderly centers were also included.
The 60 research articles are shown in Table 4, organized in chronological order. The table presents the article's title and journal, as well as the country where the research was performed. The citations were obtained from Google Scholar:

General Considerations of Studies
In this section, all 62 articles obtained were considered for analysis. Figure 2 shows the number of publications per year that relate thermal comfort and hospital environments: The publication of the first literature review in 2012 appears to have influenced the production of articles with the theme, with 2013 being the year with a great increase, eight in total. It is noted that in the interstice between 1968 and 2020, there were years when articles with the theme were not published. Figure 3 shows the most used keywords in combination with thermal comfort over the years, with greater emphasis on the size for the words more used:  Table 5 shows the most used words in the published papers, with the occurrences of each one: It is visible that the most used keyword, in combination with thermal comfort, is "hospitals". It can be seen that this information is deducible since most studies are carried out in hospitals. Only four studies differ from this reality, the studies in [19,20] which were carried out in elderly centers, as well as the studies in [21,55] which were carried out in health center facilities.
According to Google Scholar, four articles appear as the most cited: Lomas and Giridharan [53], 124 citations; Hwang et al. [45], 123 citations; Chow and Yang [41], 122 citations and Sodha et al. [38] with 120 citations. Authors with more publications are shown in Figure 4. The author with the greatest number of published papers on the subject is "K.J. Lomas", with 3 articles. All other authors have 2 publications each. The journals that most published on the topic are shown in Table 6: Building and Environment is the journal that concentrates the majority of publications in just one source, with 14 publications. The rest of the sources that did not appear in Table 6 have only one published article each and were not presented here to avoid a very large table. For the other sources of publication, please see Table 4.

RQ1: Considering Studies in Thermal Comfort in Hospital Environments, What Are the Main Aspects
That Are Taken Into Account: Health/Wellbeing, Productivity or Energy Saving? Table 7 shows, for each study, which main aspect was taken into account by the authors that published the paper: Of the 60 research articles, only in [37] was it not possible to determine which aspect the work focuses on, leaving 59 articles. Although the energy-saving aspect is quite relevant and can be considered a research trend in the area, it corresponds to only 17% of the research on the theme that relates thermal comfort and hospital environments. Most works (81%) focus on the health/wellbeing aspect, and this research returned only one work that relates productivity and thermal comfort in hospital environments.
Two possible explanations are raised in order to answer why only one study was performed taking into account the productivity aspect: according to [75,78], temperature control is vital in hospitals, as it can indirectly influence the condition of the patient in addition to contributing to the onset of infections. Thus, research that attempts to ascertain the influence of thermal comfort on employee productivity becomes impracticable as it would be necessary to change these factors. Further, the difficulty in establishing parameters to measure the productivity of a team of doctors or nurses may be another reason.

PMV Season or Period of Data Collection Sample Predicted Well
Overestimated Underestimated Not Suitable for Predict [18] x November 2017 55 pregnant women [45] x Winter and Spring 933 respondents [51] x Summer 114 occupants [21] x Winter and Spring 99 patients [56] x May and June (2011) 110 questionnaires [57] x May 2011 to February 2012 188 questionnaires [59] x 31st March to 7th June 2011 55 staff members 35 patients [64] x Not available Not available [17] x October and November, 2011 58 subjects interviewed [69] x Summer 37 respondents [74] x July to November, 2015 and March to May 2016 451 patients, 331 visitors and 146 medical staff [20] x Summer and winter 509 participants The PMV predicted well the real thermal sensation in [21,69], whereas in the other studies, the PMV was somehow unable to accurately measure the actual mean vote reported by people. Thus, 10 out of the 12 studies concluded that, in some way, PMV was not suitable, either overestimating or underestimating when applied in hospital environments. A possible solution would be the use of adaptive comfort models to assess thermal comfort in hospital environments, in order to obtain better results when comparing real and calculated votes of thermal sensation. However, it is necessary to take into account that the sample is considerably small and more studies are required.

RQ3: Considering Studies in Thermal Comfort in Hospital Environments, What Was the Most Evaluated Group and Which Area Within Hospital Was Most Evaluated?
In order to answer this question, Table 9 was created. This table contains the specifications of the target group and the site within the hospital environment in which the study was conducted, also presenting the main findings: Regarding the most explored group in the literature, "staff" and "patients" are tied as the most evaluated, and each one is presented in 30 (71.43%) and 29 (69.05%) articles, respectively out of the 42 studies. Twenty articles bring both together. The "visitants" group is present in eight works. This distribution was already expected, since both staff and patients are the means and the ends of a hospital environment and are the protagonists of the typical activity performed in these environments.
In a more in-depth way, the following distribution for the "staff" group of works that brought some specification in relation to the type of employee includes: the surgical team (8); nurses (5), doctors (2), and nursing assistants (2). For the "patients" group, only two studies specify the type of patient studied. Wheldon and Hull [36] studied full-term babies and premature babies; more recently, Fabbri, Gaspari, and Vandi [18] concentrated efforts to assess the relationship between thermal comfort and pregnant women. This finding shows a gap in thermal comfort studies, since several types of patients and their relationship with the thermal environment have not been explored.
Regarding the site within the hospital environment in which the articles were conducted, most of the articles have specifications. The most explored places in the literature were the wards (17) and the operating rooms (12), corresponding to 40.47% and 28.57%, of the total studies, respectively. The other places, such as the administrative part or waiting rooms, have been little studied, showing that new research can be done in these places. Fanger's indices were not within the thermal comfort range in most rooms (PPD = 52% in winter and PPD = 62% in summer) Cheong and Chong [27] Office's workers Administration offices The temperature of the dry air bulb recorded varied between 22.1 and 22.4 • C. This is slightly below the recommended range for acceptable indoor air quality of 22.5-25.5 • C from the local indoor air quality guideline; 49% of respondents complain of feeling cold.
Hashiguchi et al. [43] Nurses and nursing assistants * Wards The temperature was found to be between 20 and 23 • C, corresponding to previous studies carried out in Japan.
Skoog, Frasson and Jagemar [44] Nurses and nursing assistants * Wards Mean Air Temperature in summer and winter were, respectively, 21.5 and 21.8 • C.
Hwang et al. [45] * Wards The neutral temperature observed for a TSV = 0, was 23.2 • C. The variation of the observed percentages of the dissatisfaction model, 20.7-25.4 and 21.8-26.2 • C for winter and summer, respectively, was wider than the predicted percentages of the dissatisfaction model, being 21.9-25.0 and 24.2-26.9 • C for winter and summer, respectively.
Mazzacane et al. [46] Surgical Team Operating rooms Nurses claim to be comfortable 75% of the time, while assistants experience mild discomfort 90% of the time. The results confirm the existence of a thermal difference between professionals and patients, the latter constantly subjected to cold thermal stress.

Yau and Chew [51] * Several Departments
The comfortable temperature range that satisfied 90% of the occupants in the space was in the range of 25.3-28.2 • C.

Verheyen et al. [21] * Wards
No significant difference between the predicted mean vote (PMV), obtained from objective measurements, and the actual mean vote (AMV), obtained subjectively, for all different wards, except for neurology department.
Pourshaghaghy and Omidvari [54] * * * Wards and operating rooms PPD in men is higher than one verified in women in both winter and summer seasons, although the PPD difference is less than 5%.
Azizpour et al. [56] * * * Facility Department When analyzing the linear regression between TSV and PMV, neutrality was found around +0.75 instead of 0, as given in the Fanger model. The neutral temperature found was 26.8 • C, 1.8 • C higher than that calculated by the Fanger model (25 • C).

Azizpour et al. [57] *
Lobby, office, praying room, kindergarten, and catering area The new PMV limit corresponding to the neutrality range in this field study was −0.22 and +1.73 as opposed to −1 and +1 in the PMV model, and the operative temperature was 26.8 • C.  Indoor environmental quality (IEQ) takes into consideration visual comfort (light), sound (noise), thermal comfort (temperature), and indoor air quality (carbon dioxide concentration and volatile organic compounds) [85]. It plays an important role in influencing the comfort and productivity of occupants in buildings, as people remain indoors a significant part of their time [86]. Eleven studies went beyond thermal comfort and analyzed other environmental parameters, as shown in Table 10: Table 10. Additional aspects evaluated in parallel to thermal comfort.

Reference Parameter Main Findings
Chen, Jiang and Moser [40] Particle concentration Particle concentration between 2 and 2.55 m (particles/m 3 ). Very small particle concentration in the operating room.
Berardi and Leoni [16] Microbiological irborne contamination and CO 2 concentration Air microbial amount was higher in the wards and operating rooms than in the hospital offices.
Skoog, Frasson and Jagemar [44] CO 2 concentration Dust concentration The maximum values measured during the winter were 576 ppm for CO 2 concentration and 6.1 × 10 −6 g/m 3 for dust concentration.
Ho, Rosario and Rahman [48] Contaminant removal effectiveness (CRE) and the mean contaminant concentration The parameters were used to evaluate the ventilation performance of the room through simulation.
Azmoon et al. [58] Light intensity The average light intensity for all hospital workstations was 296 lux.
De Giuli et al. [59] Indoor environmental quality (IEQ) Medium illuminance values have been established. Employees complained about lack of privacy, size of rooms, and acoustic discomfort.
Lawrence, Jayabal and Thirumal [76] Indoor air quality (IAQ) A patient room was studied. By monitoring the air quality system, it was demonstrated how different types of ventilations systems might benefit patients.
Wu et al. [83] Sound pressure level (LAeq) and acoustic satisfaction The mean value of the equivalent continuous A-weighted sound pressure level in the wards was 59.2 dB, this being a satisfactory value.
In the studies analyzed, it was found that some studies took into account another parameter, in addition to thermal comfort. Azmoon et al. [58] measured the relationship between thermal comfort and light intensity with the quality of sleep and eye strain. Wu et al. [83] studied noise in the environment and how it affects the perception of comfort. Alfa and Öztürk [80] assessed patients' perceptions of the indoor environment in terms of architectural design, thermal comfort, indoor air quality (IAQ), lighting and acoustic parameters.
However, the parameter that appears most associated with thermal comfort in hospitals is indoor air quality (IAQ). Berardi and Leoni [16], Cheong and Chong [27], Ho, Rosario, and Rahman [48], Lawrence, Jayabal, and Thirumal [76], and Chen, Jiang, and Moser [40] studied IAQ inside hospitals. It is important to note that studies on IAQ are now extremely relevant, due to the pandemic caused by the new coronavirus . Studies on the impacts (IAQ) are necessary given that inadequate ventilation or low air quality in the environment can increase the risk of airborne transmission diseases [87]. According to Correia et al. [88], adequate ventilation reduces the amount of microorganisms suspended in the air, thus reducing the possibility of infection.
The Federation of European Heating, Ventilation, and Air Conditioning Associations (REHVA) [89] updated its guide on the operation and use of services in buildings in areas with a COVID-19 outbreak, in order to prevent the spread of COVID-19, proposing changes in heating, ventilation, and air conditioning systems. The main recommendation is to stop air recirculation and increase intake of external air. The internal environment must be strongly ventilated, exclusively with fresh air, to reduce the concentrations of the virus, in case of eventual contamination by suspended droplets.
According to Zhang [90], in order to reduce the risk of SARS CoV-2 infection, the outdoor ventilation rate must be increased to a level closer to the capacity of the building ventilation system. The required quality level of buildings is increasing, and it is mandatory to acknowledge solutions that facilitate maximized thermal comfort and indoor air quality, while energy consumption is minimized.

Final Considerations
It was found in the articles analyzed that there are still some little explored topics, such as productivity in hospital environments. Comfort conditions in specific patients, such as patients with chronic diseases and children, are also little explored. Only two articles were identified, one of which studied pregnant women [18] and another premature babies [36]. Regarding the comfort of specific patients, given the importance of the theme, it is necessary that research of this nature be performed in order to define standards for these types of patients for different types of environments.
The review [22] makes two criticisms addressed here: (a) Studies in the area were conducted in only one hospital and (b) focused on only one group of people. Through the analysis of the current articles, this research adds, relative to criticism (a), that most of the new studies have been carried out in more than one hospital, thus overcoming this judgment; and to criticism (b), that there was no significant improvement since 15 articles focused on more than 1 group while 11 focused on only 1. It should be noted that the universe of thermal comfort in hospital environments is very little explored, especially relative to PMV.
This article proposes four research questions. When answering RQ1, it is shown that around 81% of the published studies deal with the health/wellbeing aspect, 17% deal with energy savings, and around 2% address the productivity aspect. The response from RQ2, on the other hand, shows that the PMV model was not effective when applied in hospital environments since 10 of the 12 articles that relate PMV to TSV pointed out that the Fanger index does not represent the real thermal sensation felt by people in hospital environments. RQ3 shows that the groups "staff" and "patients" are the most studied, as can be expected; however, only the staff group has well-defined specifications, and the medical team is the most approached type. In RQ4, results showed that the IEQ parameter in addition to thermal comfort is IAQ, very important today due the pandemic caused by coronavirus.
Finally, it is important to emphasize the importance of the relationship between buildings and the thermal comfort of the occupants, since discomfort can affect not only patients, but the entire staff. It is believed that this relationship has so far been little explored and that there are still questions to be answered through further studies conducted in these environments.

Study Limitations
The search for papers was limited to the combination of keywords. Further limitations lay in bias risk assessment factors, which were not considered in the included articles in the literature review performed in this research.

Conflicts of Interest:
The authors declare no conflict of interest.