Overheating in care settings: magnitude, causes, preparedness and remedies

ABSTRACT Research in UK and elsewhere has highlighted that older people are particularly vulnerable to negative health effects of overheating. This paper examines the magnitude, causes, preparedness and remedies for addressing the risk of summertime overheating in four case study residential care and extra-care settings across the UK, spanning different building types, construction and age. An interdisciplinary approach is adopted, drawing from building science and social science methods, including temperature monitoring, building surveys, and interviews with design and management teams. The findings suggest that overheating is a current and prevalent risk in the case study schemes, yet currently little awareness or preparedness exists to implement suitable and long-term adaptation strategies (e.g., external shading). There was a perception from designers to managers, that cold represents a bigger threat to older occupants’ health than excessive heat. A lack of effective heat management was found across the case studies that included unwanted heat gains from the heating system, confusion in terms of responsibilities to manage indoor temperatures, and conflicts between window opening and occupant safety. Given that care settings should provide protection against risks from cold and hot weather, design, management and care practices need to become better focused towards this goal.


Introduction
Climate change is expected to result in hotter, drier summers in the UK with increased frequency, intensity and duration of high external temperatures (DEFRA, 2011). This is expected to have a significant impact on internal temperatures within buildings, causing overheating which can affect the thermal comfort of the occupants (Zero Carbon Hub (ZCH), 2015; Hames & Vardoulakis, 2012) and result in negative impacts on the health and wellbeing of the population (DEFRA, 2012). Furthermore, many new buildings have high levels of thermal insulation and airtightness in order to minimize heat loss which can prevent the dissipation of unwanted heat, particularly in summer (NHBC, 2012;ZCH, 2014). This problem will become more prevalent if energy-efficiency agendas are pursued to support climate change mitigation without due regard to the risks of unwanted heat during summer (DEFRA, 2012;NHBC, 2012).
The risk of excessive heat for the vulnerable population (elderly, disabled, socially isolated etc.) has been recognized by the UK Climate Change Risk Assessment (CCRA) (DEFRA, 2012). Older people are generally at greater risk of increased high temperatures, with physiological studies showing that the body's response to heat is impaired with age (Kenny, Yardley, Brown, Sigal, & Jay, 2010) as well as chronic or severe illnesses such as heart conditions, respiratory disease or severe mental illness (Gasparrini, Armstrong, Kovats, & Wilkinson, 2012;Koppe, Kovats, Jendritzky, & Menne, 2004;PHE, 2014). Epidemiological evidence indicates that older people are particularly vulnerable to the effects of excessive heat (Åström, Fosberg, & Rocklöv, 2011). Whilst health and age can impede a person's capacity to adapt, socio-cultural and personal factors also affect a person's adaptability. Older healthy persons do not necessarily perceive themselves to be vulnerable (Abrahamson et al., 2009) and therefore do not prepare for extreme weather events effectively (Wolf, Adger, & Lorenzoni, 2010). Older people also tend to be more sedentary than younger people. Analysis of English House Condition Survey data suggests that people aged over 65 years spend more than 80% of their time at home, and people aged over 85 years more than 90% (Adams & White, 2006). As such, they are more susceptible to higher temperatures within buildings.
Studies (Fouillet et al., 2006;Holstein, Canouï-Poitrine, Neumann, Lepage, & Spira, 2005;Kovats, Johnson, & Griffith, 2006;Mackenbach & Borst, 1997) indicate that heat-related mortality during heatwaves (short periods of higher-than-seasonally-expected temperatures) was highest in relative terms amongst occupants of residential and nursing homes, despite the presence of care staff who could act to protect vulnerable residents. A German study also found an increased heat-related mortality risk amongst all nursing-home residents regardless of age (Klenk, Becker, & Rapp, 2010); while in the heatwave experienced across Europe in 2003, a study indicated that in France mortality was highest in the least physically frail residents (Holstein et al., 2005). With the UK's ageing population projected to continue (Office of National Statistics, 2014), resulting in an increase in population aged over 75 from 8% of the total population in 2012 to 13% in 2035, overheating in buildings inhabited by generally older and more vulnerable people, such as residential care and extra-care schemes, is a significant area of concern.
Care and extra-care housing schemes are generally hybrid building types, simultaneously functioning as long-term residences, sometimes nursing environments and workplaces (Walker, Brown, & Neven, 2015). This hybridity can impact (positively and negatively) on the building's risk of summertime overheating; including safety issues, diverging needs and preferences (particularly between staff and residents), user-technology interaction and questions about who is responsible for thermal conditions (van Hoof, Kort, Hensen, Duijnstee, & Rutten, 2010). Recent research (Brown, 2010;Neven, Walker, & Brown, 2015;Walker et al., 2015) also indicates that the regulatory context and business considerations of a care scheme focus on the provision of good care, which is associated with ensuring no resident is too cold and that they are secure and safe (Walker et al., 2015). These considerations reinforce the idea that care settings should be 'warm' places.
Since the European heatwave of summer 2003, there has been considerable attention paid to national preparations and responses to periods of hot weather across European countries, including in the UK, particularly in relation to vulnerable persons such as those who live in care and extra-care schemes. Amongst other things, this has culminated in the Heatwave Plan for England (PHE, 2015), which is linked to the UK Met Office Heat-Health Watch Service (HHWSa system that provides early warning of periods of high temperatures which may affect the health of the UK public). The heatwave plan provides practical advice on what should be done to prepare for (long-term, all year round) and deal with (short-term actions) hot weather within health and social care settings, including providing a 'cool' room that remains below 26°C before, during and after a period of weather above the HHWS regional threshold temperatures. Despite this, there is some evidence that new-build care and extra-care housing schemes are already too warm for occupants and are overheating (Barnes et al., 2012;Burns, 2008;Guerra Santin & Tweed, 2013;Lewis, 2014). However, the scale of the issue for existing care settings is relatively unknown due to most heat-related health risk studies focusing on the relationship between external temperatures and heat-related excess deaths during the summer months. Yet, understanding the relation between indoor temperature and health is probably more critical (CCC, 2014) due to the range of factors mediating the relation between indoor and outdoor temperature, including building design and occupants' thermal comfort practices (Dengel & Swainson, 2012).
Within this context, this paper investigates the magnitudes, likely causes, preparedness and remedies for addressing the risk of summertime overheating in four case study care schemes (two residential and two extracare), located across the UK. This is achieved by: . assessing the magnitude (prevalence) of overheating through physical monitoring of indoor (covering residential, communal and office spaces) and outdoor temperatures over one summer period across the four case studies . evaluating the potential causes of overheating and preparedness for tackling overheating through building surveys and interviews with key members of the design and management teams . identifying remedies in terms of appropriate recommendations for practitioners (designers, care providers, housing providers), regulators (Care Quality Commission (CQC)), policy-makers (Department of Communities and Local Government (DCLG) and Department for Environment, Food and Rural Affairs (DEFRA)) and care staff.

Research study and approach
A case study-based approach was adopted in this research study, focusing on two residential care homes and two extra-care facilities to demonstrate the risk of overheating in environments with different levels of care provision. While a care home is generally for older people with frailties (physical and cognitive), providing them with single private bedrooms and access to communal social spaces and onsite care services with meals provided and staff on call 24 hours a day, an extra-care scheme is designed to accommodate older people who are becoming frailer and less able, but who still require and/or desire some level of independence. Extra-care housing schemes provide varying levels of care and support; at a minimum, there will be some kind of on-call assistance for people in an emergency, but not necessarily a physical presence 24 hours a day, as is available in residential care homes. Extra-care schemes also usually provide self-contained units, consisting of a kitchen, living/dining area, bathroom and one or two bedrooms, in addition to communal social facilities. Such differences in care provision and physical facilities mean that residential care homes are operated in a different manner to extra-care facilities; the more independent residents in extra-care schemes are generally afforded greater responsibility and control over their thermal environment than residents in residential care schemes, where it is expected that due to their frailties the staff are more likely to exert control. Also as confirmed by Flyvbjerg (2006), a detailed examination of a single or a few case examples can provide reliable information about the broader class.
The four case studies are located in North England (one care home), South West England (one extra-care), and two in South East England (one care home and one extra-care) ( Figure 1). They were selected based on ownership (public and private care), variation in built age (and related building regulation context) and location. All but one are managed by not-for-profit organizations. The average age of the residents ranged from 85 to 89 years. Table 1 outlines the key characteristics of the case studies and other important criteria (ventilation, construction type) considered during the selection process. It must be noted that due to issues with recruitment (schemes simply being unable to provide adequate time and access), the case studies were relatively self-selecting which may mean that they have some degree of preexisting interest in overheating and climate change.
The research approach was interdisciplinary, drawing from building science and social science methods, and involved conducting primary research across the four case study schemes as follows: . Monitoring of indoor and outdoor temperatures at 15-minute intervals over the summer months from June to September 2015. Thirty-three rooms across the four case studies were monitored, which included communal areas, offices and resident rooms/flats. . Building surveys of the case studies were undertaken to identify building design features that may enable or prevent occupants (staff, residents) to control their thermal environment during periods of hot weather. . Semi-structured interviews were conducted during September 2015 with five designers and four asset managers involved in the four case study buildings. The interviews focused on understanding the impact of briefing, building design and management of the schemes on overheating.

Overheating metrics and care settings
Overheating, whilst a widely used term, is currently neither precisely defined nor understood and can be assessed in relation to thermal comfort, health or productivity (ZCH, 2015). This is in part linked to the complexities of assessing individuals' adaptability to external temperatures, depending upon the climatic conditions they face, and are used to, as well as assessing thermal comfort. Hence there is inconsistency regarding what particular conditions may constitute overheating (Dengel & Swainson, 2012), although the ZCH (2015) has recently attempted to address this issue for housing. This lack of definition of overheating means that there is a multitude of overheating assessment metrics (Table 2), generally based on either temperature-health effects or thermal comfort indicators. Within the health sector, and more specifically the care sector, there is guidance on (outdoor) threshold temperatures at which heat-related deaths are expected to increase, such as 24.5°C (see PHE heatwave plan guidance; PHE, 2014). However, apart from heatwave plan guidance indicating that at least one room in care schemes should be  kept below 26°C in order to provide a 'cool area', there is a lack of guidance or standards in terms of indoor temperatures at which overheating occurs in care settings, and the level of associated risk to health. The 'building' sector has several overheating metrics with different internal temperature thresholds that focus mainly on thermal comfort comprising both 'static' and 'adaptive' approaches. Whilst CIBSE has adopted the adaptive approach in recent years, there is much discussion (ZCH, 2015(ZCH, , 2016 as to whether or not this is wholly appropriate, particularly in buildings where the occupants are less able to adapt to their local environment, such as care homes and accommodation for vulnerable occupants.
The static approach enables simple calculations to be undertaken when assessing the performance of building. The main criteria for identifying the overheating risk according to the static approach are: . overheating in (non-air-conditioned) bedrooms occurs when the indoor operative temperature is over 26°C for at least 1% of occupied hours . overheating in (non-air-conditioned) living rooms and offices occurs when the indoor operative temperature is over 28°C for at least 1% of occupied hours In contrast, the adaptive approach as described in CIBSE (2015) accounts for the adaptation of occupants to their environmental context within free-running buildings. It is based on the presumption that the occupants have adapted to external temperatures during the preceding few days, i.e., the running mean (T rm ) to create an allowable indoor operative temperature in relation to the external temperature: In terms of specific overheating criteria, three overheating criteria are provided in CIBSE TM52 (2013), of which if two are failed overheating is deemed to have occurred (Table 3).
The adaptive approach also takes into consideration the sensitivities of occupants, and differing levels of thermal expectations, such as category I (a high level of expectation only used for spaces occupied by very sensitive and fragile persons), where the suggested acceptable comfort range is ±2 K from the temperature calculated from the running mean of the outdoor temperature, T rm : In terms of this study, category I is used for residential and communal areas and category III (acceptable, moderate level of expectation and used for existing buildings) is used for the office areas.
In terms of assessment of overheating and thermal comfort within the care sector specifically, there have been relatively few studies, particularly in temperate climates similar to that of the UK's, undertaken in order to ascertain what constitutes 'thermal comfort' within the care sector specifically (CCC, 2014). While one study showed that elderly people require higher temperatures in order to achieve thermal comfort (Mendes et al., 2013), further studies indicate that this may be in part due to a prevalent perception that older people require warmer conditions, rather than an accurate representation of this demographic (Walker et al., 2015). This is why it is important that health-related thresholds are used when evaluating care settings for overheating due to the specific vulnerabilities of this demographic group, as highlighted through both physiological and epidemiological studies (Åström et al., 2011;Gasparrini et al., 2012;Koppe et al., 2004;PHE, 2014), and the potential inability to gather information on thermal comfort from physically and cognitively frail residents. To complicate matters further, care sector buildings are general hybrids in nature (being both residential and work places), which means that the thermal comfort needs of the occupants will vary hugely, and as such make it difficult to determine the most appropriate overheating metrics to use.
Although the static approach has its issues, in that whilst it establishes the occurrence of overheating, it does not necessarily indicate the severity of the overheating (Nicol, Hacker, Spires, & Davies, 2009), the adaptive approach was developed from research in non-domestic settings and there are doubts as to its appropriateness within a care setting, where the capacities of the residents are somewhat unique; the more vulnerable the occupants, the less likely they are able to adapt to changes in temperature. In addition, CIBSE Guide A (2015 p. 1-16), whilst adopting the adaptive approach to assessing overheating risk, states that a static threshold temperature is still appropriate for bedrooms: 'It is desirable that bedroom temperatures at night should not exceed 26°C unless there is some means to create air movement in the space, e.g. ceiling fans.' A recent review by the ZCH (2016) that sought to provide a starting point for developing a national policy or standard on overheating indicates that for bedrooms specifically a static threshold temperature should be used, rather than an adaptive approach as, generally, a person's ability to adapt and cool down when sleeping is more limited and the available evidence, although limited, indicates that not only is sleep affected when operative temperatures are above 25°C, but also heat-related deaths and illnesses are more likely above this temperature.
Given this context, the main metrics used to assess overheating risk within this study include the static CIBSE Guide A (2006) overheating and thermal comfort criteria (referred to as the static approach) for all rooms; the adaptive overheating and thermal comfort approach outlined in CIBSE TM52 (2013) and CIBSE Guide A (2015), and which is based on BS EN 15251:2007 for all rooms except bedrooms; and the PHE's recommended maximum internal temperature threshold of 26°C to be maintained before, during and after a heatwave.
It must also be noted that whilst the authors acknowledge the requirement for the operative temperature (T op ) to be calculated in order to undertake the above overheating risk methodologies, due to practical constraints, air temperature (dry bulb) was used as a proxy for T op . According to Mavrogianni, Taylor, Davies, Thoua, and Kolm-Murray (2015), this is a common limitation of monitoring studies due to cost constraints. The operative temperature and dry bulb temperatures mainly differ in indoor spaces with higher levels of exposed thermal mass or high indoor air velocity (Mavrogianni et al., 2015), neither of which were prevalent in the case study buildings within this study.

Monitoring of thermal conditions
External and internal data loggers were installed in the four case study schemes. The locations of the data loggers in each case study are identified on the floor plans (see the supplemental data online). In addition, the onsite managers were informed of their locations and asked to ensure staff knew of their locations. The loggers recorded dry bulb air temperature and/or relative humidity at 15-min intervals for three months during the summer (mid-June 2015-end of September 2015). Unfortunately, due to participant availability and practical restrictions, it was not possible to install all loggers across the four case studies on the same date, so the temporal coverage varies slightly for each case study.
The external data loggers used (Onset HOBO U23 Pro v2) measured external temperature (accuracy = ±0.21°C; range = 0-50°C) and relative humidity (accuracy = ±2.5%; range = 10-90%RH). They were placed in convenient and secure locations, generally just above ground level (0.5-1.0 m above) and away from sources of direct and reflective heat and light sources. The internal data loggers used were Onset HOBO U12 (temperature (accuracy = ±0.35°C; range = 0-50°C) and relative humidity (accuracy = ±2.5%; range = 10-90% RH)) and Maxim Integrated ibutton DS1922L (temperature only; accuracy = ±0.5°C; range = −10-65°C). In total, 34 internal data loggers were installed. They were placed in convenient and secure locations to prevent removal by either staff and/or residents. The internal data loggers were placed at around 1.80 m from floor level and away from sources of direct light and heat such as light bulbs, radiators or large electronic appliances). However, two were lost (one in the communal area in case study A; one in a bedroom in case study B), apparently due to their removal by either staff or residents. The loss of the data loggers highlights the difficulties of monitoring spaces with remote researchers and active on-site occupants. No data were retrieved from these loggers. Despite this, across the four case studies data were available for 17 residential rooms including six living rooms (extra-care units only; case studies C and D) and 11 bedrooms, eight communal areas (lounges and dining areas) and eight offices.
During their installation, information relating to the different occupancies of the rooms was gathered (Table 4). Information on construction materials (including building types, insulation levels and glazing types) and heating/cooling/ventilation systems and controls installed were also gathered during a building survey and through a desktop review of technical specifications and architectural drawings.

Magnitude of overheating risk
Overall, the summer of 2015 was cool and wet, with the Met Office (2016) reporting that despite the mean annual temperature being 0.4°C above the 1981-2010 long-term average, the monthly mean temperatures from May to September were below average (e.g., July mean temperature was 14.4°C, 0.7°C below the 1981-2010 average). However, a new UK temperature record of 36.7°C (Heathrow, London) was set on 1 July and external temperatures across all regions of the UK were particularly high over this short period of time. Although the Met Office refers to this as a one-day heatwave, it is worth noting that there is no official UK definition of a heatwave except for the following: 'a heatwave is an extended period of hot weather relative to the expected conditions of the area at that time of year' (Met Office, 2016). Generally, the Met Office uses the World Meteorological Organisation (WMO) definition of a heatwave which is 'when the daily maximum temperature of more than five consecutive days exceeds the average maximum temperature by 5°C, the normal period being 1961-1990'. There were no heatwave periods, as defined by the WMO during the monitoring period. However, there were periods in which localized external temperatures in case studies A and D met the trigger temperature thresholds of the PHE's HHWS, upon which the PHE's heatwave plan is based, and which are referred to as heatwaves in PHE guidance and documentation. Table 5 presents the average mean and maximum temperatures over the monitoring period across different spaces, and results from the overheating analysis. Relatively, the residential areas (private bedrooms and living rooms) were the most susceptible to overheating. About 16 out of the 17 residential rooms (across all case studies) overheated according to the static approach, including all bedrooms. In terms of the adaptive approach, two living rooms in case study C overheated.
Five out of the eight communal areas overheated during the monitoring period according to the static approach, whilst only three (all in case study D) overheated according to the adaptive approach. Four out of the eight offices monitored overheated according to the static approach, whilst only one (manager's office, case study B) overheated according to the adaptive approach.
Due to the small sample size, it is difficult to ascertain the impacts of orientation, size and location of individual rooms on the overheating risk, particularly as the differences in temperatures could be due to individual occupant  Notes: a Case study C Communal 1 and case study D office 2 both have air-conditioning installed and as such were not assessed using either the adaptive or the static approaches, as recommended in CIBSE guidance.  Note: the horizontal grey band in the offices graph indicates summer comfort temperature range for air-conditioned offices (office 2 is air-conditioned).
behaviour, which was not recorded in detail during the study. However, the fact that the temperatures in similar rooms (such as case study D flats with the same orientation, window opening and floor area) varied suggests that overheating is as much to do with heat management within the individual rooms as the overall design. Across the monitored spaces, the average mean indoor temperatures were relatively high. CIBSE (2015) guidance on thermal comfort indicates that in bedrooms, thermal comfort and quality of sleep decreases in temperatures above 24°C. Overall, nine out of the 11 bedrooms monitored had an average mean temperature of 24°C or above, and the average mean temperature across all the bedrooms was 24.5°C. In the other room types (private living rooms, communal areas and offices), the average mean temperatures across the monitored rooms were 25.5, 24.7 and 25.7°C respectively. This is also significant as PHE (2015) guidance indicates that at 24.5°C excess heat-related deaths become apparent, suggesting that the temperatures within all the case study buildings could be resulting in both thermal discomfort and increased health risks. In addition, as Table 5 indicates, all rooms have maximum temperatures above 26°C (the PHE indoor threshold temperature for 'cool rooms'). Indoor temperatures appear highest in case study D, where five of the 10 rooms monitored have average mean temperatures above PHE indoor threshold temperature. Figure 2 indicates when the risks are occurring and when static threshold temperatures (CIBSE, 2006;PHE, 2015) are being reached across the different room types in case study D. Temperatures in the bedrooms do not fall to recommended summer comfort temperatures until September, when external temperatures (day and night) have also reduced. Furthermore, for a significant period, the internal temperatures of the flat living rooms and communal areas are around or above the PHE's recommended temperature threshold for 'cool rooms'. This is particularly noteworthy as the 'cool room' threshold is reached even when the external temperatures have not breached the HHWS thresholds (day maximum = 31°C; night minimum = 16°C) that indicate action is required. In addition, Figure 2 demonstrates that both offices, despite office 2 being air-conditioned, are above CIBSE (2006) recommended static comfort levels for the majority of the monitored period. It also highlights that except for the office areas, there are distinctive 'spikes' in the internal temperatures that correspond with the period of high external temperatures around 1 July. Figure 3 demonstrates the correlation in case study A between external and internal temperatures during a period in which the HHWS trigger temperature thresholds (day maximum = 29°C; night minimum = 15°C) were achieved; 29 June-2 July 2015. During this period, indoor temperatures within all rooms rose by approximately 2°C during the first day alone. Furthermore, on the second day, both the indoor and the external temperatures rose again, resulting in all the monitored rooms experiencing temperatures above the PHE heatwave plan recommended threshold temperature for 'cool rooms' of 26°C. Whilst this is likely to have health-risk implications for the most vulnerable residents, it must also be noted that in the period leading up to the short-term high external temperatures, temperatures in all rooms were already above static indoor summer comfort temperatures during occupied hours (23°C in bedrooms, 25°C in offices and living areas; CIBSE, 2006), indicating a high likelihood of thermal discomfort for all occupants (staff and residents) during this period. The communal lounge was the only room in which overnight temperatures dropped to similar levels as those prior to the hot weather period. The temperatures in the residential and office areas remained relatively high and only resumed previous levels after one day and two nights. This suggests that the existing design measures (such as the thermal mass of the building retaining heat) and heat-management strategies in these areas (such as ventilation) were not enough to bring down indoor temperatures during short periods of high external temperatures.
Although thermal comfort surveys were not undertaken (as the majority of residents were physically and cognitively frail), informal discussions with both staff and residents during the building survey indicated that three case studies (A, C and D) were generally considered to be very warm during summer; particularly C and D. This is worth noting as these two case studies are the extra-care facilities, and control over ventilation and cooling is split between staff and residents, in comparison with the two residential care homes in which thermal control appeared, generally, to be the responsibility of the staff only. A number of staff also commented Table 6. Assessment of design features of the case study buildings in relation to their potential impact on overheating risk.
Positive characteristics (aspects that can help mitigate overheating risk) Negative characteristics (aspects that can help exacerbate overheating risk) Case study A + Enclosed courtyard with green cover and shrubbery + Office areas (high internal gains) face in a northerly orientation + Heavyweight materials used in construction a + Balconies on some southerly facing rooms provide shading to rooms below + Internal blinds and curtains present in most rooms + Openable windows in corridors to enable cross-ventilation + Low-energy light fittings + Simple heating controls (thermostatic radiator valves (TRVs) at the top of radiators) -Communal heating and hot water system with a distribution pipework throughout the building -Low reflective roof (low albedo) -Single -aspect bedrooms -Trickle vents painted over (maintenance issue) -Window restrictors present; no control on balcony doors (open or shut only) Case study B + Enclosed garden area with significant green cover, planting and mature trees + Relatively heavyweight wall and floor materials used + Internal blinds and curtains present in most rooms + Low-energy light fittings + Simple heating controls present (only TRVs in rooms) -Communal heating and hot water system with a distribution pipework throughout the building -Heavy sash windows difficult to open, with little fine control -Non-reflective roof (low albedo) -Single-aspect rooms Case study C + Secure green space around the building with low shrubbery, and minimal hard paving + Where large areas of hard paving are present, it is northerly facing + Relatively heavyweight wall and floor materials used + Internal blinds and curtains present in most rooms + Brise soleil (fixed louvres) and overhanging eaves to provide additional shading in the main south-facing communal area + Low-energy light fittings + Openable windows in corridors to enable cross-ventilation + Trickle vents and openable windows present in all rooms + Simple heating controls present (zoned thermostats and individual radiator TRVs) -TRVs at low level (poor accessibility for physically frail) -Communal heating and hot water system with a distribution pipework throughout the building -Window restrictors present -Low-reflective roof (low albedo) -Single-aspect flats Case study D + Balconies with vertical panels for shading + In-built planters on balconies for additional green cover + Internal courtyard with raised planting beds (open to south-west) + Mature tree retained on site (south-west) + White roof (high albedo) + Heavyweight wall and floor materials used + Internal blinds and curtains present in most rooms + Low-energy light fittings + Openable windows in corridors to enable cross-ventilation + Trickle vents and openable windows present in all rooms except one office -Communal heating and hot water system with a distribution pipework throughout the building -Complex heating controls in residential flats -Lever handles on windows not suitable for some residents with physical frailties (adaptations required) -Window restrictors present -Single-aspect flats -Exposed car park on west of site Note: a Unless adequate overnight ventilation is provided, heavyweight materials can increase the night-time overheating risk as the materials may release heat captured during the day into the indoor spaces.
on the disparity between their perception of thermal comfort with that of the residents who were much more sedentary. Since 'keeping warm' was perceived to be related to good care, staff expected to experience higher levels of thermal discomfort, particularly in terms of being too hot, in order to ensure the thermal comfort of residents. Although staff tolerated high indoor temperatures to be part of their job, it raises concerns about risks to their health. Interestingly, some staff members noted that they actually felt cooler in the summer than the winter, in part because they were more able to encourage the opening of windows and use of electric mobile fans.

Potential causes of overheating in care settings
It is well-recognized that building design plays a significant role in terms of exacerbating or mitigating high temperatures, particularly in terms of its ability to minimize heat gain (solar and internal), maximize excess heat loss during hot weather periods (summer), and enable effective heat management by occupants (Gale, Fitzsimmon, Gartner, & Gale, 2011;Gething & Puckett, 2013;McHugh & Keefe, 2012;Tregenza & Wilson, 2011). All four case studies had some design features that could either exacerbate or reduce the risk of overheating (summarized in Table 6). The key design features that were designed to tackle overheating included brise soleil (case study C; Figure 4), overhanging eaves (case study C) and large balconies (with further in-built space for planting and green vegetation) to provide additional shading on south-facing facades (case study D; Figure 5). Building surveys (a combination of walk-through and inspection of buildings from outside and inside) helped to uncover the likely reasons for the occurrence of overheating which were not apparent otherwise. For example, conflicts were discovered between design strategies (for passive cooling) and other priorities such as resident requirements, safety and security that hindered effective management of heat. Across all the case studies, residential areas were found to be mostly single-aspect spaces lacking through ventilation due to practical, spatial and care requirements. Internal shading (blinds,   curtains) were common but keeping blinds closed during the day as a remedial measure was found to be feasible only where rooms were unoccupied, as residents needed to see out and have access to daylight ( Figure 6). Window restrictors were installed to maintain safety and security of residents, although it limited occupants' ability to open windows to provide adequate ventilation (Figure 7). The design of heating and ventilation controls also appeared to impact upon the occupants' ability to manage their thermal environment effectively. In case study D lever handles were considered to be not appropriate for a care setting due to the physical frailties of some residents. As a result staff had adapted the lever handles in one flat using bicycle handlebars to make them longer for a resident with severe arthritis to still be able to open and close their windows. Although all three recently built case studies (A, C and D) had trickle vents installed in windows to ensure continuous background ventilation, they only appeared to be in regular use in case study D. In the flats in case study C, even on a hot day, the trickle vents remained closed (Figure 8). This appeared to be due in part to the occupants being unaware they were there, and also an expectation that they had already been opened by the staff. In case study A trickle vents had been painted over (most likely by maintenance staff) (Figure 9), highlighting the need for communication with building management and maintenance staff about the purpose of such strategies.
The building surveys also uncovered a lack of effective heat management practices. Across all four case studies it was found that the centralized heating and hot water system remained on, and in use, throughout the year, resulting in unwanted summer heat gains. In part this was due to the need for hot water in individual bedrooms (care units) and flats (extra-care units) as well as the varied heating requirements of individual residents; some   were reported to want additional heating even during the summer months, whereas others did not. Due to this variety in requirements, there was evidence of heating controls being adapted to remove access from residents, particularly in the resident rooms of care settings ( Figure  10) and communal areas (care and extra-care settings) to ensure more effective management of heating. Despite this, even in areas under staff 'control' (such as communal areas), thermostats settings were set very high ( Figure  11). Furthermore, in case study D, installation issues with the heating system itself (the exact cause was unknown) meant that residents had been asked by management to keep the thermostat in their bathrooms on 'max' (over 30°C), which they subsequently were doing.
Semi-structured interviews with designers and asset managers highlighted the impact of common procurement methods such as design and build that involves a single main contractor undertaking all aspects of the work (who may appoint several disparate subcontractors), while the initial designer of a care scheme may not be involved in the ongoing design and specification process. This can lead to decisions, mainly cost driven, that conflict with the original design intent for the building, e.g., in one of the case studies, roof design and specification was changed from concrete (high thermal mass which can absorb excess heat within the building) to timber (low thermal mass that cannot absorb excess heat as effectively) without assessing its effect on overheating. Also insufficient communication of design intent from design teams through contractors and care providers to end users led to inadequate user understanding for operating heating and ventilation systems. Moreover, it was often the building management team that undertook the handover process, rather than the on-site end users themselves (care staff).
In addition, lack of adequate internal communication was also discovered within the care organizations, which was in part due to the separation of building management teams (usually based off-site) from care staff, with the result that responsibility for heating control was removed from the daily users (care staff and residents), and they were not always able to alter temperatures. Several of the asset managers also commented on the practical difficulties in achieving full communication with on-site staff, acknowledging that there was a relatively high turnover of care staff. This led to lack of agency as well as confusion surrounding responsibilities within on-site staff, subsequently resulting in contradictory actions (windows left open with heating on) or even inaction by staff.

Preparedness for tackling overheating
To assess the preparedness for tackling overheating in care settings, semi-structured interviews were conducted with the designers and asset managers (involved in the case studies) who highlighted a number of factors relating to the design, briefing and management of care scheme developments that are likely to impede preparedness. An underlying culture of 'warmth' was prevalent, with the predominant attitude within both the designers and the managers that being 'too cold' was the issue, rather than being 'too warm'. 'We focus on keeping people warm in the winter that's our main focus' (manager).
Such views have a strong factual basis: cold is strongly associated with mortality. Recent estimates indicate there are currently 41,000 premature deaths caused by cold weather in the UK annually compared with just 2000 premature heat-related annual deaths (CCC, 2014). Furthermore, future projections indicate that cold-related deaths in the UK are expected to remain high (Hajat, Vardoulakis, Heaviside, & Eggen, 2014). However, most studies (e.g., Vardoulakis & Heaviside, 2012; Figure 11. A thermostat in a corridor set to 27°C in summer in case study C. CCC, 2014) also indicate that excess heat-related deaths in the UK are expected to increase significantly, with one study (Hajat et al., 2014) suggesting an increase to approximately 7000 a year by the 2050s.
There was also a lack of awareness of the heat-related risks of climate change, and subsequent low prioritization of design measures for avoiding overheating. One designer stated that when designing and developing the briefing for care schemes, overheating was seen as 'the poor sister … to other aspects of climate change'. This appeared to be in part due to the 'warmth culture' as well as a relatively unconcerned attitude, particularly amongst some managers, towards heatwaves, which were seen as something that only occurs rarely in the UK, and as such could be managed through short-term adaptation practices, such as mobile electric fans. Both the designers and the asset managers also indicated that there was a lack of understanding of long-term measures to mitigate overheating: We need to understand it a little bit more … we're not as familiar with the solutions … it's not just us I think that's the [building] industry as a whole. (designer) The lack of standardized advice, calculations and standards in relation to the assessment of the overheating risk during the design stage was also felt to exacerbate the lack of awareness, particularly as one designer pointed out that modelling of thermal environment for building regulations focused on energy and carbon savings rather than on overheating specifically. Furthermore, all but one asset manager (case study A) stated that overheating was not considered a risk within the lifetime of the schemes they were developing and commissioning, and as such was not part of long-term strategic planning: I don't know … the impact of what's going to be over what sort of timescale for a business like this. … I struggle I guess to anticipate that in the lifetime of this business that it's going to become a huge issue.
The interviewees also noted that where there were conflicts due to other priorities, more often than not, the other priorities took precedent. An example of this was the need for care organizations with several developments to run their individual schemes efficiently. Two of the asset managers highlighted this had led to an increase in the use of building management systems (BMS) and centralized heating and hot water systems that could be managed off-site, which took away responsibility for heating control from the daily users (care staff and residents), and they were not always able to alter temperatures.
During periods of hot weather, most of the reported measures undertaken were relatively short-term and reactive, such as providing mobile electric fans or localized air-conditioning units as well as care practices as outlined in the PHE's (2015) heatwave plan for England such as: keeping them [windows] closed, keeping your blinds down, getting your fans, pushing your fluids, all that sort of stuff, putting people in light clothing and all the things that you would typically do to keep the building nice and shady and as cool as we can … . (asset manager) The managers indicated that they felt this was sufficient in terms of tackling current overheating risk, particularly as they had not had significant overheating problems reported to them by on-site staff. In addition, the management of overheating was generally left to the carers (frontline care staff) and there were no organizationwide strategic management plans, except for the PHE's heatwave plan, which the managers expected carers to implement. As such, the approaches to heat management taken by the case studies appeared to be reactive rather than proactive, and indicate a lack of preparedness for addressing overheating risks.

Discussion and recommendations
The environmental monitoring and overheating analysis has revealed the occurrence of overheating in summer 2015 across all four schemes, which raises concerns regarding the future risk of overheating in a warming climate. The study also highlighted the lack of monitoring and awareness of localized external and internal temperatures as part of building management. This emphasizes the need to monitor indoor (and outdoor) temperatures in care settings regularly, with feedback to management, frontline care staff and residents to identify any occurrence of overheating and support timely action.
The differences between the results of the static and adaptive approach analyses suggest that the adaptive approach could be underestimating the overheating risk, particularly in relation to buildings and rooms occupied mainly by vulnerable persons (or those less able to adapt). Whilst there is some overlap between static threshold temperatures in building sector guidance (as in CIBSE Guide A) and health-related guidance (as set out in the PHE's heatwave plan), fundamentally there is a lack of evidence on appropriate temperature thresholds (for health and thermal comfort) within the care sector and, specifically, for older people. Combined, this is likely to lead to confusion and lack of prioritizing of the risk and understanding of how to identify when and where overheating may occur. This issue was reflected in the prioritization of other design, spatial, cost and care requirements and needs over overheating, and a lack of long-term strategic planning and preparedness for overheating mitigation.
Throughout the study there was a prevalent perception, from care scheme designers to managers, that older people 'feel the cold' and that cold represents a bigger threat than heat to older occupants' health. While cold is still more prevalent as a health risk, there is less recognition that heat can also present a significant health risk. Heatwaves were seen as something that only occur rarely in the UK, and as such can be managed through short-term adaptation practices, such as mobile electric fans. This is why design for overheating was not found to be commonplace and innovative design solutions for overheating were not widespread within the design of care schemes. Even planning for future overheating was not perceived to be 'top of the agenda' as care and housing providers tend to plan for the near future, rather than the longerterm. The majority of the asset managers interviewed did not anticipate the effects of climate change to be large enough to impact upon operations within the next 30 years or sothe lifespan for which buildings in the care sector are intended to cater.
A key finding in terms of the causes of overheating related to the management of heat in care facilities. The heating was left on throughout the summer in all the case studies due to differing requirements of occupants and different levels of control, and capacity and separation of roles (between building management and care) particularly within the medium-sized care organizations (case studies A, C and D) creating confusion in terms of responsibilities to manage heating controls and indoor temperatures. These findings suggest that neither design nor management will be sufficient responses on their own. Design measures cannot necessarily be wholly protective of vulnerable residents during periods of hot weather, whilst improved management cannot fully compensate for inappropriately designed buildings that overheat significantly already, or outside extreme hot weather periods. Given that vulnerable residents are within settings that should be providing care and therefore protection against thermal risks (arising from both cold and hot weather), building design, management and care practices need to become better focused towards this goal.
Against this backdrop of research findings, key recommendations for policy-makers, regulatory/guidance bodies, care/housing providers, designers and care staff are proposed (Table 7). Most importantly, given that there is no statutory maximum internal temperature for care settings, collaboration across key care and building sector bodies, such as PHE, the CQC, Chartered Institution of Building Services Engineers (CIBSE), DEFRA, DCLG and Department of Health (DoH) is critical for the standardization of health-related and thermal comfort-related temperature thresholds for overheating in care settings. This will enable effective adaptation solutions specific to the care sector to be developed and implemented.

Study limitations and suggestions for future research
Whilst the overheating analysis was based on individual occupancy profiles for each monitored room, a limitation of the study was the lack of thermal comfort surveys with residents, although some insight on actual thermal comfort experiences was gained through discussions with staff and residents during building surveys and interviews with management. Undertaking thermal comfort surveys in care settings is challenging due to the frail nature of residents, the majority of whom suffer from mild physical and cognitive disabilities. This would make any survey undertaken of the residents unreliable. In addition, the majority of the management were also wary of asking care staff to participate in such surveys due to time pressures, in part because the majority of the care staff were not employed by their organization directly. Despite this practical limitation, it is recommended that further studies seek to undertake thermal comfort surveys to provide a more complete picture of the impact of overheating on the various types of occupants within care facilities. Furthermore, the study results need to be used with caution due to the small sample size and large differences between both the building characteristics and individual occupants.
Despite this, the study offers valuable insights into current summertime temperatures of the case study care facilities during a relatively 'cool' summer period, and raises a number of questions relating to the preparedness of the care sector against hot weather that could be addressed by a larger-scale monitoring and thermal comfort study of care facilities.

Conclusions
Whilst the study findings are more illustrative than conclusive due to the small sample size, they do add important new evidence to the current overheating risk in care and extra-care settings, given that there is currently little research on the prevalence of summertime overheating in care settings in the UK. It is also found that the deployment of adaptive comfort for assessing the risk of overheating is likely to be inappropriate in spaces where residents are less able to adapt to their local environment (such as bedrooms in care homes), as it might create a false sense of reduced risk of heat stress to inhabitants. This in turn might reduce either or both the policy side's and/or the design and management side's focus on overheating as a problem that needs to be addressedor at least diminish the true extent of the problem. The criteria for assessing overheating risk in spaces inhabited by residents having limited opportunity for adaptation forms an important area for research, if adequate facilities are to be provided for the ageing and vulnerable population in the UK. The findings also suggest that overheating is a current risk in the care sector that is likely to be exacerbated in future due to climate change, yet there is currently little awareness and implementation of suitable and long-term adaptation approaches (such as external shading, provision of cross-ventilation). Such strategies require input from designers, development teams, care providers, care managers and frontline staff. Yet such fundamental change also requires support, in terms of enhanced and focused regulations, standards and guidance, from key care sector bodies and government departments or agencies. Perhaps most urgently, there needs to be a culture change within the care sector itself, so that risks posed by elevated temperatures are prioritized alongside risks from cold.