A new definition of fuel poverty in Scotland: review of recent evidence

Chapter 5 Indoor temperatures

' Standards are too often treated as if they possessed some kind of separate and independent status or viability. They seem to exist more or less indefinitely - long after their origins and rationale have been lost in the mists of time and the words of countless reports and debates. They are compared, averaged, raised or lowered in this or that respect. The Committee suggests that housing standards are not of this order; they are means to the ends of public health - the physical, mental and social well-being of man. They should be examined and periodically re-examined to determine whether or not they are fulfilling their objective.' ( WHO, 1961).

5.1. 'Satisfactory heating regimes' in Europe - timeline

5.1.1. Housing Commission of the League of Nations 1936

The earliest technical reports on optimal indoor temperatures were published by the Housing Commission of the League of Nations. Starting in 1936, their reports on public health and housing covered a wide remit:

'These reports deal with the thermal environment of housing, noise, natural and artificial lighting, town planning, air pollution, water supply, sewage treatment and the collection and treatment of domestic refuse'. ( WHO, 1961).

5.1.2. Parker Morris and World Health Organization Reports 1961

The Parker Morris standards for public housing in the UK were published in 1961 and noted:

' We therefore think it is time to recognise that a home without good heating is a home built to the standards of a bygone age'.

When the temperature was −1 °C outside, the Parker Morris Standards recommended that heating systems installed in homes be capable of maintaining 13 °C in most rooms and halls, with living and dining spaces able to be kept at 18 °C. However, the need for topping rooms up with secondary heating sources was noted, on occasions when they were being used for anything other than passing through or sleeping in. Where these occasions became routine, more extensive primary heating systems would be required:

'… they can easily be topped up for use as studies or bed-sitting rooms. But where family requirements are clearly going to demand such use as a matter of course - and this may become widespread within the life of the building - a more expensive installation capable of heating the bedroom to 65 °F [18 °C] as well will represent greater value for money'.

Whilst these temperature thresholds were modest, they were not considered ideal, but rather practicable during an era of mass public housing investment:

' The aim here, as with all the other minimum standards which we are recommending, should be a progressive raising of standards through the years'.

However, by 1978, when the Building Research Establishment monitored indoor temperatures in 1,000 homes, average living room temperatures were still at the Parker Morris minimum standard from 1961 (18.3 °C); bedrooms were slightly warmer (14.6 °C) (Hunt & Gidman, 1982).

5.1.3. The 1968 WHO report

The pioneering role of the League of Nations was taken over in 1961 by the World Health Organization ( WHO), with the publication of Public Health Aspects of Housing in 1961. Thereafter, WHO alone became the main protagonist in setting temperature thresholds for satisfactory heating regimes in Europe. It has remained so for almost 50 years, (Ormandy & Ezratty, 2012).

The WHO's initial recommendation was that homes in Europe be kept between 15 ºC and 25 ºC, on the basis that this range protected people from illness and disease.

5.1.4. The 1984 WHO report

The 1984 Working Group recommended that this range be altered to between 18 ºC and 24 ºC, a range that should apply to all non-vulnerable households. No explicit reasons were given for this change (Ormandy & Ezratty, 2012). However, reports from other Working Groups (working on related topics) had started to recommend a minimum indoor temperature of 20 ºC for at-risk groups (e.g. WHO, 1979), which is likely to have influenced their deliberations, although they did not ultimately adopt a different regime for vulnerable groups. Instead, they felt there was insufficient evidence on the matter.

The Group concluded by recommending a series of studies across Europe, with a view to revisiting the issue once these were complete. Many of these studies were subsequently commissioned.

5.1.5. The 1987 WHO report

A Working Group convened again in 1985, and updated their evidence review. They did so, at least in part, on the basis of reactions to the energy crises of the late 1970's, which had led to some European Governments introducing policies that constrained indoor heating regimes to a range between 15 ºC and 19 ºC :

' WHO has therefore to reconsider its former recommendations regarding indoor climate in order to discourage excessively low temperatures which may have detrimental effects on the health of at-risk groups'.

The Group published its report in 1987, and focused particularly on vulnerable people living in cold homes:

'[A]… main objective was to consider whether sufficient evidence from scientific studies and epidemiological investigations was available to demonstrate an adverse effect on health of low indoor temperatures, especially in high-risk groups.'( WHO, 1987).

Vulnerable groups were defined as ' the elderly, the sick, disabled, and preschool children' ( WHO, 1987).

The 1987 publication is the WHO's seminal report on indoor temperature guidelines. It is a concise 20-page summary of relevant studies, citing 56 different sources. Almost all the evidence base cited remains valid now, despite 30 more years of further experimental research and field study.

Among other recommendations:

• it retained the 18 ºC - 24 ºC range, but only for non-vulnerable groups;
• for vulnerable groups, they recommended 20 °C as a minimum temperature, hence narrowing the range still further to 20 °C - 24 ºC.

5.1.6. The latest WHO report - 2007

This retained these recommendations.

5.1.7. Post-1987

Since that time, the only notable adaptation to these recommendations concerns rooms where infants sleep. Here there is a general consensus in the UK that " an ambient room temperature of 16-20°C, combined with light bedding or a lightweight well-fitting sleeping bag, offers a comfortable and safe environment for sleeping babies", although it is also noted that further research is necessary to establish this with confidence (Lullaby Trust, 2017). This consensus currently applies to infants in the first 12 months of life. The NHS currently recommends a sleeping temperature for infants of 18 ºC (e.g. NHS Sheffield, 2017).

5.2. 'Satisfactory heating regimes' and UK Fuel Poverty Strategies

The earliest Fuel Poverty policies deployed in Scotland, Wales, England and Northern Ireland were all predicated on the belief that a link existed between cold homes and poor health. These early policies were, first and foremost, public health policies. In the UK Fuel Poverty Strategy of 2001, for example, the word ' health' features 238 times in a document of 158 pages. The rationale for focusing on health derived from the WHO series of reports outlined earlier in this Chapter, as well as from a (then) very small body of research evidence linking, for example, cardiovascular illnesses and exposure to cold (e.g. Donaldson & Keatinge, 1997; Eurowinter Group, 1997). In the same year as the UK Fuel Poverty Strategy was published, the first study linking cold temperatures to energy inefficient homes was also published (Wilkinson et al., 2001).

5.2.1. England and Northern Ireland

Fuel Poverty Strategies were published in 2001 and 2015 for England, and in 2004 and 2011 for Northern Ireland. In specifying a satisfactory heating regime, they did not fully rely on the range of temperatures which WHO recommended. Rather, they used:

• the WHO lower temperature threshold ( 18 ºC) to set a temperature that should be maintained in most ^[20] rooms;
• with the exception of living rooms, where a temperature of 21 ºC was set (the midpoint of WHO's 18 ºC -24 ºC range).

As early as 1983, the Royal Society of British Architects were recommending that future homes be built in a manner that permitted temperatures of 18 ºC for bedrooms and 21 ºC for living rooms, providing some degree of precedent ( RIBA, 1983). Although the temperature for 'other rooms' ^[21] complied with the WHO lower limit, there may have been relatively little scientific evidence supporting 21 ºC as the ideal for living rooms. According to Healy and Clinch (2002), for example:

' The Building Research Establishment has generally considered 18-21 ºC as a comfortable temperature for a living room during wintertime, and Boardman has generally advocated this temperature range as 'comfortable'. (bold font added)

The BRE Domestic Energy Factfile (2003) states:

' It would be expected that the average temperature would stabilise as more households move towards their desired comfort levels. For most people, a living room temperature during occupied periods of 21 ºC would be regarded as comfortable. A temperature perhaps 2ºC below this would generally be considered adequate elsewhere in the dwelling so that the overall comfort level might be around 19-20 ºC.' ( BRE, 2003). (bold font added)

The fixing of 2 temperature set points, broadly within the WHO recommended range, may have been deemed a prerequisite for estimating domestic energy needs, including fuel poverty prevalence. In both original Fuel Poverty Strategies, England and Northern Ireland set targets which required monitoring of domestic energy need over time. It would have been more feasible to estimate need on the basis of specified temperatures in specified types of room, rather than relying on a temperature range which could vary by as much as 6 ºC.

Another factor contributing to the decision to set 21 ºC as the temperature for living rooms may have been a series of modelled indoor temperature data which were available from BRE at the time. For the 4 years preceding England's 2001 Strategy, centrally heated homes were estimated to maintain average indoor temperatures around 19 ºC (Uttley & Shorrock, 2008). It could be argued that selecting the narrower band of 18 ºC in most rooms and 21 ºC in living rooms meant that the temperatures recommended were within the realms of what might be deemed ' reasonably practicable'. However, these modelled internal temperatures were the result of energy balance calculations using building physics data and DUKES energy consumption figures. They were not drawn from the monitored performance of homes. More recently, independent studies monitoring actual temperatures suggest that these modelled estimates were optimistic i.e. homes were in fact maintaining significantly lower temperatures at that time.

5.2.2. Scotland and Wales

Scotland and Wales opted for a more nuanced temperature regime which was more closely aligned to the WHO's 18 ºC to 24 ºC range. According to the Scottish Fuel Poverty Statement (Scottish Executive, 2002):

'… a 'satisfactory heating regime ' would use the levels recommended by the World Health Organisation. For elderly and infirm households, this is 23 °C in the living room and 18 °C in other rooms, to be achieved for 16 hours in every 24. For other households, this is 21 °C in the living room and 18 °C in other rooms for a period of 9 hours in every 24 (or 16 in 24 over the weekend); with two hours being in the morning and seven hours in the evening'.

Elderly is defined as people over 60 years old. In short:

• elderly and infirm people should be able to maintain their living room temperature at 23 °C, and all other rooms in their home at 18 °C - for most of every day;
• all other households should be able to maintain their living room temperature at 21 °C, and all other rooms a temperature of 18 °C - for most of weekend days, and shorter periods of time during the week.

Wales adopted a similar definition of satisfactory heating regimes in their more recently published Fuel Poverty Strategy ( WAG, 2010).

A concern with these recommendations is that they do not permit provision of 24-hour levels of safe temperatures, assuming perhaps that homes will retain sufficient heat to be safe, even in the 8-15 hour unheated intervals when indoor temperatures are likely to fall. For fuel poor households living in energy inefficient homes, levels of heat loss from the building fabric could mean that temperatures fall well below these safe levels during these unheated intervals, some of which could occur when people are at home.

5.2.3. Satisfactory heating regimes for UK homes: summary of timeline

The temperatures required to maintain satisfactory heating regimes in the UK are broadly based on the WHO recommendations of 1987, thirty years ago. Contrary to the assertion that many have made, WHO did not stipulate either 21 ºC or 23 ºC as set points for different rooms. Rather, they specified a range of indoor temperatures within which human health was likely to be protected, with this range being narrower for at-risk groups.

The adoption of set points may have been based on the need to have specific temperatures from which to calculate domestic energy need and fuel poverty prevalence, without which BRE could not have generated statistics that could be easily understood and tracked over time. The set points chosen had due regard for the WHO's decision on a minimum safe temperature, with an upper set point positioned mid-way between the 18 ºC and 24 ºC range.

The narrower range of 18 ºC and 21 ºC may have reflected modelled data on room temperatures thought by BRE and others to be prevailing in UK homes at the time (19 ºC), although it is now recognised that these estimates were probably inflated.

In the context of fuel poverty prevalence, the issue of whether a recommended temperature regime could reasonably be altered has important consequences. For example, the current recommendation yields a prevalence rate of 31%, but 26% if vulnerable groups are expected to maintain the same temperatures as standard occupants (see Table 7.1).

Table 7.1.: Variants of temperature regime and effects on fuel poverty prevalence

Occupant Types	Temps.	Heating Pattern	Fuel Poverty Rates
			Older	Families	Other	All
Standard occupants	21°C/18°C	9 hrs weekdays, 16 else	45%	16%	29%	31%
Any occupant aged 60+ or LTSD [22]	23°C/18°C	16 hours per day
Standard occupants	21°C/18°C	9 hrs weekdays, 16 else	36%	14%	26%	26%
Any occupant aged 60+ or LTSD	21°C/18°C	9 hrs weekdays, 16 else
Standard occupants	21°C/18°C	9 hrs weekdays, 16 else	39%	14%	26%	27%
Any occupant aged 65-79 or LTSD	21°C/18°C	9 hrs weekdays, 16 else
Any occupant aged 80+	23°C/18°C	16 hours per day

5.3. New studies on temperatures in UK homes (2014 - 2017)

This section contains a brief review of studies which have been published since the Scottish Fuel Poverty Strategic Working Group (2016) examined evidence related to ' satisfactory heating regimes'. The studies described are all published in established peer-reviewed journals, have acceptable sample sizes, use conventional scientific methodologies, and interpret results using advanced analytic tools.

5.3.1. Public Health England's 2014 Cold Weather Plan

In 2011, the Department of Health and Health Protection Agency, in association with the Meteorological Office, published a Cold Weather Plan ( DOH, 2011) aimed at protecting the health of those most vulnerable to the impacts of cold temperatures ( DHHPA, 2011). It recommended that:

• 21 °C should be the minimum recommended daytime temperature for rooms occupied during the day;
• 18 °C should be the minimum recommended night-time temperature for bedrooms.

In 2014, three years later, Public Health England ( PHE) published a review of evidence on indoor temperatures supporting a change; it recommended:

• a single minimum temperature threshold of 18 ºC.

PHE and other stakeholders concluded that:

' The review suggests that 18 ºC is a reasonable minimum threshold to adopt for future guidance in the Cold Weather Plan for England based on the limited evidence available and discussion with experts. The evidence did not support a threshold of 21 °C in living rooms, and there was additional concern that if widely implemented by the whole population it could potentially result in excessive use of household energy and lead to higher carbon emissions. Neither of these are desirable outcomes from a broad health and wellbeing perspective….A single minimum threshold of at least 18 ºC for all rooms was considered the most appropriate advice'. (Jevons, Carmichael, Crossley & Bone, 2016).

The Jevons et al. meta-analysis acknowledged that more still needed to be learned about the health impacts of temperatures on specific at-risk groups, such as people with long term conditions and frail older people. For these people, ' temperatures slightly above this threshold may be beneficial to health'. This somewhat undermines a previous assertion in their paper, namely that a one-size fits all recommendation is most appropriate since it ' means that messages are kept as simple as possible whilst protecting the health of all'.

While aiming for a single generic message, PHE downplayed a wealth of existing evidence regarding significant differences in both preferred and actual ambient temperatures by age and vulnerability. Since the 2014 Cold Weather Plan was published, that evidence base has grown, and some of it is reviewed below.

5.3.2. The Energy Follow-Up Survey (Hamilton et al., 2017)

This study used temperature data from 823 English homes, all of which were part of the English Survey's Energy Follow-Up sample (2010-2011). The research comprises a large cross-sectional survey of indoor temperatures and dwelling/household characteristics. Among key results:

• during daytime heating periods (standardised to an outdoor temperature of 5 ºC), living room temperatures ranged between 18.4 ºC in homes located in NW England, and 19.2 ºC in the East and East Midlands;
• bedroom temperatures ranged between 17.8 ºC in SW England and Yorkshire/Humber, and 18.6 ºC in the East and the South East;
• households categorised as vulnerable (low income or in receipt of specified benefits) maintained warmer living rooms than did non-vulnerable households (19.3 ºC and 18.7 ºC respectively); there was no significant difference in bedroom temperatures for the two groups (18.3 ºC and 18.2 ºC respectively);
• using the Boardman definition of fuel poverty, households that were fuel poor had colder living rooms than households not in fuel poverty (18.2 ºC and 18.9 ºC respectively); their bedrooms were colder (17.2 ºC and 18.4 ºC respectively);
• by contrast, comparisons using the LIHC definition of fuel poverty found no difference between the living- and bed-room temperatures of fuel poor households and those not in fuel poverty.

The authors conclude that:

' Policies in England that focus on households in lower socio-economic conditions will likely find an unmet need for improved indoor temperature conditions'.

5.3.3. The 2015 Leicester Study (Kane et al., 2015)

Another comprehensive study of winter temperatures (December to February) in UK homes is based on monitoring 249 homes in Leicester. There was a strong positive association between outdoor and indoor temperatures, indicating that households ran their heating systems in a manner largely responsive to the weather outside: in living rooms the correlation between outdoor and indoor temperature changes was high (r = 0.87), and broadly the same in bedrooms (r = 0.91).

The sample included a range of housing types and income brackets, but had a higher than representative group of retired households. Among the key results:

• living rooms were maintained at an average of 18.5 ºC;
• bedrooms were maintained at 17.4 ºC;
• both rooms had a standard deviation of 3 ºC, which indicates a substantial variation in maintained temperatures between households; as noted below, vulnerability status was significantly implicated in these variations.

The authors note that these temperatures are:

• lower than what is generally assumed in the current BRE energy models used to estimate fuel poverty prevalence;

• lower than fuel poverty recommended temperatures of 21 ºC in living rooms (a 2.5 ºC shortfall) and 18 ºC in bedrooms (a 0.6 ºC shortfall).

Shortfalls of the same order of magnitude have been reported in many other recent (2014-2017) UK studies, some using more geographically dispersed samples than Leicester. In summarising these, Kane and colleagues (2015) conclude:

' Overall, the high level of agreement between the findings of these different studies, both with regard to the heating patterns and the temperatures achieved, is rather remarkable; especially given the different samples, monitoring equipment used, periods of measurement, parameter definitions and methods of calculation'.

In the Leicester study:

• living rooms of the over 60's were maintained at 19.3 ºC, almost 3 ºC higher than those of people in their 20's (16.4 ºC);
• living rooms of people unable to work were maintained at 20.6 ºC, almost 2 ºC higher than living rooms of working people (18.9 ºC);
• bedrooms, on the other hand, were maintained at significantly cooler temperatures among those who might be vulnerable to the impacts of cold.

Here, the two vulnerable groups averaged a mean difference between living room and bedroom of 2 ºC, whereas non-vulnerable households maintained living- and bed-rooms at broadly similar temperatures.

The two potentially vulnerable groups (older and unable to work) also used their heating in a different way from non-vulnerable people, with heating:

• starting earlier in the year;
• coming on earlier in the afternoon;
• running for longer each day;
• with secondary heating sources used more often;
• and fewer under-heated days.

Data on central heating timings indicate too that some of the significant disparity between living- and bed-room temperatures among those more vulnerable arose because of secondary heating systems being used in living rooms through the day. Many of these (e.g. electric fan heaters, log burners, and open coal fires) are known to be particularly expensive ways of generating extra heat (Sutherland Tables, 2017). This was also found in households located in small areas of Northern Ireland which had a high prevalence of fuel poverty (78% of homes in fuel poverty, with 19% of these in extreme fuel poverty). Among the 2,145 homes surveyed in these areas, more than two-thirds relied on secondary heating systems almost throughout the year (Liddell & McKenzie, 2013). In a second in-depth survey of 40 fuel poor households who were using wood pellet boilers, two thirds were also using open coal fires (Liddell & Guiney, 2017).

The tendency for people on lower incomes to maintain warmer than average temperatures in living rooms but lower temperatures in bedrooms was also noted in the Energy Follow-Up Survey (see 5.3.2.). It is possible that those more vulnerable to the impacts of cold elect to invest in a warmer than average room to spend their day in, often using more expensive forms of secondary heating, and perhaps trading this extra expense off by running their bedrooms at colder than average temperatures.

Kane and colleagues conclude:

' Based on the evidence generated here, those over 60 and those unable to work, are likely to use more energy and have the highest heating bills, and so energy efficiency measures could benefit them most. They are also a sector of the population which might experience most difficulty in paying fuel bills, yet benefit most from maintaining a relatively high indoor temperature over the whole day'.

Broadly speaking, almost all post-2000 field studies indicate that temperatures in English homes average around 18 ºC to 19 ºC in living rooms, with bedrooms running at around 1.5ºC cooler than that. There is, however, a large inter-household variation, even in homes of similar design, age, and energy efficiency, with vulnerability status being a significant mediating variable (Rupp, Vásquez & Lamberts, 2015).

5.3.4. The Portsmouth Study (Teli, Dimitriou & James, 2016).

This was a relatively small but carefully monitored study of 18 flats in Portsmouth, which gathered data in a month straddling March and April. Unlike the Leicester study, which was based on a wide variety of homes and demographics, this focused on low income tenants in a social housing tower block. Temperature shortfalls were greater than those reported in Leicester:

• more than half of the tenants failed to achieve bedroom temperatures of 18 ºC;
• a third of occupants used no bedroom heating at all during this time, though outdoor temperatures ranged between 2 ºC and 12 ºC;
• over 80% failed to attain 21ºC in their living rooms, and those that did only attained these temperatures during the night through availing of cheaper night-time tariffs;
• more than two-thirds of occupants relied on portable electric heaters, and for half of these electric heaters were their sole source of heating.

Whilst it is known that many households fail to meet the 18 ºC - 21 ºC range of temperatures in their homes, the Portsmouth study highlights the fact that - on average - households more likely to be in fuel poverty maintain temperatures further below that range than do households who are not in fuel poverty. As evidence cited in this Chapter indicates, this is almost certainly more often for reasons to do with affordability than idiosyncratic perceptions of thermal comfort. The study added to a growing accumulation of evidence documenting lower ambient temperatures in the homes of people experiencing fuel poverty.

An important consequence, as Teli and colleagues point out, is that the potential of fuel poor households to contribute to carbon reduction targets, even after deep retrofits, will sometimes be minimal. For households in fuel poverty, there is an increasing expectation that many will need to produce more rather than less carbon post-retrofit, since they are currently unable to heat their homes to temperatures that even approximate 18 ºC and 21 ºC. The extent of shortfall from these thresholds may mean that even the deepest retrofits may not allow them to achieve safe temperatures without using more energy and expending more carbon.

This finding lends weight to the argument that fuel poverty is a social justice, rather than a carbon-reduction issue. It also points to the need for caution in setting up measurable outcomes for fuel poverty policies and programmes; whilst these are often conceptualised in terms of people in fuel poverty having 'lower fuel' bills and homes that 'emit less carbon' (Scottish Government, 2012) such goals may be neither realistic nor ideal.

5.3.5. Scottish solid-walled homes (Herrera & Bennadji, 2015)

This small study of 24 solid granite walled buildings located in the North-East of Scotland compared homes which had been upgraded with improved energy efficiency measures with homes which had not. The study found that:

• average summer temperatures indoors were moderately low: 19.8 ºC (non-retrofitted) and 18.8 ºC (retrofitted); and
• in winter, non-retrofitted homes averaged 14.4 ºC, and the retrofitted homes 15.4 ºC. These temperatures are well below the levels that the WHO considers safe.

Notably, though, in terms of 'thermal comfort', the occupants of these homes did not report dissatisfaction with indoor temperatures. Even in winter, most of the occupants described the temperature as ' comfortable'.

5.3.6. Spatial shrink and temperatures (Tweed, Humes & Zapata-Lancaster, 2015)

A qualitative study, interviewing older people in England, Scotland & Wales, illustrates the extent to which ambient temperatures in different rooms shape older people's everyday use of space. Respondents may:

• avoid certain rooms because they are not kept at a comfortable temperature;
• move furniture into rooms which are kept more comfortably warm so that core activities can still be pursued;
• divide rooms into smaller spaces with curtains or more structural room dividers;
• remain close to a radiator when carrying out more sedentary activities.

All of these common practices represent forms of spatial shrinkage which are likely to reduce mobility, as well as the everyday options people have for variety and stimulation in their home. Both of these effects have implications for health and wellbeing (Liddell & Guiney, 2015).

5.3.7. Cold homes and older people's health (Shiue, 2016)

Nurses gathering data for the English Longitudinal Study of Ageing visited the private homes of 7,997 participants - all over 50 years old - in order to collect a wide range of physiological data. In each home, they recorded the temperature of the room they worked in whilst they were visiting, a crude measure of ambient temperature in people's homes, but compensated for by the very large random sample.

Using the WHO minimum threshold of 18 ºC, the study found:

• among 50-64 year olds, 19% were in rooms with temperatures of less than 18 ºC;
• among 65-79 year olds, 16% were below the 18 ºC threshold;
• among those 80 years and over, 9% were below the 18 ºC threshold.

Where rooms were below the 18ºC threshold, residents showed a wide range of adverse health markers including:

• higher blood pressure readings;
• higher cholesterol levels;
• lower white blood cell counts;
• poorer lung function.

Although the study did not evaluate potential confounding variables or covariates (such as income, smoking, etc.) the results showed a significant linear association between room temperature and adverse outcomes. In other words, there was a dose-response effect, such that the further below 18ºC a room temperature was, the more adverse the biomarker results were.

Summarising the range of studies discussed here, 5 conclusions are reached.

1. Relatively few UK households attain temperatures which WHO consider safe.

2. This is especially so in the more northerly regions of the UK.

3. Households in fuel poverty live in colder homes than households which are not.

4. Older households and households where people are at home all day opt for warmer temperatures.

5. At least among over 50's, colder temperatures are significantly correlated with poorer health, and in a dose-response fashion.

5.4. ' Satisfactory heating regimes' and thermal comfort

5.4.1. Defining thermal comfort

Thermal comfort is a key focus of energy efficiency programmes, largely because perception of thermal comfort (or the lack of it) is the primary driver of how residents choose to use their heating systems. After the energy efficiency status of a home has been improved, studies suggest that as much as 50% of potential energy savings can be lost as a result of people experiencing poor thermal comfort post-retrofit, and seeking their own solutions to remedy this (Teli et a., 2016). Thermal comfort is also of interest when exploring the range of temperatures which could be considered ' satisfactory' because, as previously noted, the range of indoor temperatures which people maintain in their homes is considerable. This suggests that thermal comfort might have an important part to play in agreeing what can be considered a ' satisfactory heating regime'.

There are 3 international standards of assessment for measuring thermal comfort, the most common of which is ASHRAE 55-2013. This defines thermal comfort as:

' that condition of mind that expresses satisfaction with the thermal environment and is assessed by subjective evaluation'. (in Rupp, et al., 2015).

The critical elements of this definition are that thermal comfort is:

• a condition of mind…
• …assessed by subjective evaluation.

This makes thermal comfort a complex yardstick by which to assess what a safe set of indoor temperatures is likely to be, since at least part of thermal comfort derives from to a person's subjective preferences. The potential for adverse health impacts from living in conditions that are too cold (or too warm) may not feature in decisions people make about thermal comfort. As with some other aspects of health and wellbeing, what people prefer is not always what best protects them from harm.

In this context, the WHO (2010) recommended that the concept be replaced with an evidence-based definition of ' thermal health' in the coming years. Little progress on this new concept has been made, and none in the field of fuel poverty.

Thermal comfort preferences are generally assessed in experimental conditions rather than field studies, which afford participants the opportunity to vote for their optimal room temperature under a variety of conditions (see Rupp et al.'s review, 2015). Two aspects of these studies are relevant here, namely thermal comfort and gender, and thermal comfort and age.

5.4.2. Thermal comfort and gender

The Rupp et al. review of preferred indoor temperatures (2015) notes that studies largely support the commonly held perception that women:

• are more sensitive to ambient temperature than men;
• prefer higher room temperatures than men do.

This is unlikely to be entirely because of gender differences in ' conditions of mind' or ' subjective evaluation'. Physiological differences between men and women also underpin these preferences. Women, on average:

• have a lower skin temperature than men;
• are more sensitive to skin temperature, particularly in their hands and feet.

5.4.3. Thermal comfort and age

The Rupp review also notes evidence of age differences in temperature preferences, though these differences are not quite as consistent as the evidence for gender differences. People aged 67 to 73 years show more distinct physiological reactions to cooler temperatures (e.g. peripheral vasoconstriction) than do younger people (20-25 years). Distal vasoconstriction is implicated in many of the adverse health impacts of indoor cold (particularly raised blood pressure), which makes this physiological sensitivity in older people an important consideration when assessing the range of indoor temperatures to recommend for older people (Liddell & Morris, 2010).

van Hoof et al. (2017) also note that one of the more common medications prescribed to older people are beta-blockers (mainly for the treatment of heart disease and high blood pressure). These are known to alter thermoregulatory responses, with common symptoms including cold hands and feet. Thermal discomfort arising from this is frequently reduced when people experience warmer temperatures.

Among older people living at home with Alzheimer's Disease and related dementias ( ADRD), thermal comfort levels show wide variability. Until recently, the two most common causes of excess winter deaths in the UK were respiratory events and cardiovascular events. Over the past few years, however, the two most common causes have become respiratory events and events related to ADRD (Liddell, Morris, Gray, et al., 2016). Here too, a blend of adverse physiological reactions and a change in subjective sensations regarding thermal comfort is implicated.

5.4.4. Thermal comfort in England

Using DECC's 2013 Energy Follow Up Survey data (n = 2,616 households), there was no correlation between levels of thermal comfort and energy consumption, or between thermal comfort and indoor temperatures (Huebner et al, 2016).

5.4.5. Thermal comfort in Scotland

The most recent Scottish House Condition Survey (2015) indicates that 27% of fuel poor households say their home is never or only sometimes warm in winter. For people not in fuel poverty, the figure is only slightly less at 21%. This means that more than 1 in 5 households in Scotland are presently living in homes they say are too cold in winter.

5.4.6. Summary

' Thermal comfort is a complex topic and we are far from understanding all its related aspects. It has become evident that there is a gap in thermal comfort studies in relation to interdisciplinary research. The association with other professionals like psychologists, physiologists, sociologists, philosophers and even with other building related ones (architects and engineers) could be of great value for the development of an integral (systemic/holistic) research approach that may help to a better comprehension about sensation, perception and thermal comfort…We certainly need a better understanding of thermal comfort to face climate change and the demands for more energy efficient buildings.' (Rupp et al., 2015).

The Panel recognised the importance of people being able to achieve thermal comfort at home, even if comfort was reached at temperatures outside those recommended by WHO. There is a wealth of evidence, for example, that many vulnerable people prefer bedrooms to be cooler than 20 ºC finding this conducive to better rest and wellbeing. However, in calculating the prevalence of fuel poverty, the Panel thought that estimates of prevalence should be based on temperature regimes thought to be protective of human health. Where thermal comfort levels deviated substantially from those regimes (and these would need to be at the extremes of heat or cold), the issue e could be a matter for targeted public health messaging rather than a review of safe temperature thresholds.

5.5. Unintended consequences: Overheating

Adverse physiological reactions to heat in England are thought to become evident when ambient temperatures exceed 26 ºC in bedrooms and 28 ºC in living rooms (Gupta & Kapsali, 2016). There is growing concern in the UK about the prospect of high-grade energy efficiency measures leading to homes being overheated in summer, particularly houses built to passivhaus or similarly high energy efficiency standards.

Overheating in Scotland

Climatic variations in the UK mean there is more potential for overheating in some parts of the UK than others, although colder regions are not exempt from risk. A recent study carried out in Scotland demonstrated that:

'… low energy buildings, through a combination of effective heat retention and recovery, occupant behaviour and poor design or installation essentially override the northerly location'. (Morgan, Foster, Sharpe & Poston, 2015).

The authors concluded that 54% of 26 new homes built to high standards of energy efficiency were overheating for more than 6 months of the year. Whilst relatively few fuel poor households in Scotland are likely to be living in homes of passivhaus or similar highly efficient standards, the issue is likely to becoming of increasing importance over time. However, the authors also note that actual overheating (as measured by sophisticated temperature loggers) and people's self-report of overheating bore little correlation with one another, leading them to conclude that ' what is being defined as overheating is simply the desired comfort range of some occupants'. This adds to the considerable body of evidence highlighting the variable nature of what people consider to be ideal indoor temperatures.

5.6. Options for change

The temperatures required to maintain satisfactory heating regimes in Scotland are broadly based on the WHO recommendations of 1987, published 30 years ago. These recommendations were primarily aimed at protecting human health from the adverse effects of cold and damp homes. A review of their present-day appropriateness seemed timely, especially since a large number of well conducted studies have been published in the last 3 years.

In recommending an 18 ºC - 23 ºC temperature range for estimating fuel poverty prevalence, Scotland has remained closer to longstanding WHO recommendations than England or Northern Ireland. It has set higher standards and more demanding targets, especially in a time of the UK's increasingly ageing population.

That having been said, there are still several options for changing how a satisfactory heating regime is defined in the Scottish fuel poverty context. These include:

(i) weighing up the advantages and disadvantages of using actual temperatures versus thermal comfort as the most valid yardstick;

(ii) raising the temperature set points currently in use;

(iii) lowering the set points;

(iv) changing the wording used to define a ' satisfactory heating regime'.

These options are considered in this section.

5.6.1. The metric: Temperatures or thermal comfort?

When examining evidence on actual ambient temperatures in people's home, and comparing this with evidence related to people's estimation of thermal comfort, it seems evident that any future recommendations regarding safe temperatures at home should be based on actual temperatures. Thermal comfort is a state of mind, is subjective, and though bound up with physiological reactions to cold, there are many other influences on how people rate ambient temperature, not all of which are fully understood. Furthermore, temperatures which people think (or say) are comfortable are not necessarily temperatures which the WHO would consider safe. For all of the above reasons, thermal comfort assessments are not recommended as a primary tool for identifying households in fuel poverty.

Based on the most recent Scottish House Condition Survey (2015), a thermal comfort metric would do little to reduce the prevalence of fuel poverty in the region, since more than 1 in 5 households not in fuel poverty under the current measure say their homes are too cold in winter, with this figure rising to more than 1 in 4 among the fuel poor.

Thermal comfort is, however, an appropriate metric by which to measure the outcomes associated with energy efficiency retrofits, so that we can be assured of householder satisfaction with measures taken, and so that we can better understand the relationship between thermal comfort and different types and depths of retrofits. The concept is probably valid as an outcome metric, but not as part of how fuel poverty is defined - it is used as an outcome metric in Chapter 7.

5.6.2. Non-vulnerable households: lower the 21 ºC threshold in living rooms?

It is widely recognised that this would save considerable money for households and governments, would lower fuel poverty prevalence estimates, and reduce carbon emissions. Of all the options for changing Scotland's temperature recommendations it is both the easiest option and carries the greatest positive impact at policy level. There is also nothing in the WHO recommendations which would challenge this, and since UK living rooms tend to be run at around 18 ºC to 19 ºC, the status quo is (on average) still within the WHO's 18 ºC - 24 ºC range.

On the other hand, it is also where the greatest temperature shortfall occurs. Bedrooms are generally maintained much nearer the current 18 ºC threshold, with living rooms falling much shorter of the 21 ºC threshold. Reducing living room recommendations still further could be a controversial matter. There is no hard evidence to support the change, and we do not yet fully understand why living rooms are run with such a large temperature shortfall. In some cases, it will be a matter of thermal comfort (i.e. choice), in others of inaffordability and hardship.

The ratio of one to the other is - though unknown - worth consideration. Among the majority of people in fuel poverty, it is more likely that the shortfall results primarily from issues of affordability, rationalised by an ' I'm not complaining' attitude which people frequently invoke to account for living in a cold home (Boardman, 2010). This is supported by the fact that, after an energy efficiency upgrade, most people elect to raise their room temperatures rather than spending less on energy. ^[23] (Teli et al., 2016).

For these reasons, and because the question of what constitutes a ' satisfactory heating regime' is concerned mostly with protecting the fuel poor, there seems little justification for lowering the living room temperature threshold from 21 ºC.

5.6.3. Vulnerable households: raise or lower the 23ºC temperature in living rooms?

To stay within the WHO's 18 ºC to 24 ºC range, living room temperatures could be raised to 24 ºC among vulnerable households. However, there could be risks associated with overheating in the event of the threshold for vulnerable people being raised from 23 ºC to 24 ºC, particularly bearing in mind that adverse physiological reactions become evident at 26 ºC (Gupta & Kipsali, 2016). The margin between 24 ºC and 26 ºC leaves little room for real-world error in managing home temperatures.

On the other hand, lowering the living room temperature recommended for vulnerable people to that for non-vulnerable groups (as done in England and Northern Ireland) would be counter to the consistently expressed view of WHO panels since 1982, namely that vulnerable people should be able to live in temperatures around 2 ºC warmer than able-bodied people.

In the absence of evidence to the contrary, there seems little to support lowering living room temperatures recommended for vulnerable people.

5.6.4. Non-vulnerable households: lower than 18 ºC in other occupied rooms?

The 18 ºC seems difficult to lower, in view of these considerations:

(i) this represents the lowest safe temperature set out by the WHO;

(ii) it is the minimum threshold of room temperatures recommended by Public Health England;

(iii) it broadly approximates the temperatures which the average British household already maintains in rooms other than their living room.

In many respects it represents the commonly accepted status quo for a minimum threshold, and there is no substantive evidence to support lowering it.

5.6.5. Vulnerable households: lower than 18 ºC in other occupied rooms?

On the basis of the 1987 WHO recommendations, a minimum threshold of 18 ºC could be viewed as insufficient for vulnerable households, since WHO set a minimum threshold of 20 ºC for this group, which WHO define as ' the sick, the handicapped, the very old and the very young' ( WHO, 1987).

Minimum temperatures for the ' the sick, the handicapped, the very old' ^[24] should perhaps be raised from 18 ºC to the WHO recommendation of 20 ºC, since there is sound evidence to support that change. Under the current guidelines, if these at-risk groups were to comply with Scottish temperature recommendations, they would be moving from a living room at 23 ºC, to a bathroom or kitchen at 18 ºC, which is a drop of 5 ºC. Medical evidence has consistently demonstrated the level of physiological effort required to regulate temperature when people move rapidly from warm to cool conditions (and vice versa). Core body temperature in humans seldom changes by more than 0.3 ºC, despite even radical changes in ambient temperature. This stability is achieved through physiological effort, which involves most of the cardiovascular system.

Older people are less effective than younger ones at maintaining core body temperature when exposed to thermal stress, since they usually have a lower metabolic rate from which to generate heat or cool down; they also have less effective vasoconstriction responses because of hardening of the blood vessels (DeGroot and Kenney, 2007). Other factors that impede rapid heating up or cooling down include low or high body fat levels, sedentary lifestyle, medication use, and chronic conditions (heart conditions, circulatory disease & diabetes) (Marmot Review Team, 2011).

In that context, an observational study of 17,000 participants from England, Scotland and Wales showed that a 2 ºC drop in outdoor temperature had a threefold greater impact on systolic blood pressure in those aged 65 or older than in those aged 35 or younger (Brennan, et al., 1982).

Among vulnerable groups, therefore, a temperature regime which means that recommended temperatures in different parts of the home vary by 5 ºC could represent a significant risk to their health, particularly if shifts between thermal zones happen several times each day, for example as older people rise in the morning, wash and get dressed, prepare meals, and undress for bed. It should be noted that the WHO 1987 Report on indoor temperatures also remarked on the adverse health impacts of at-risk groups having to ' move repeatedly between warm and cold indoor environments'.

For vulnerable groups there seems to be sufficient evidence to raise the 18 ºC minimum threshold for all occupied rooms other than the living room, to a higher threshold of 20 ºC:

(i) since it is the WHO recommendation for vulnerable groups;

(ii) because of medical evidence on the physiological stressors caused by rapid warming and cooling, which increase with age and infirmity.

5.6.6. Age thresholds

The evidence cited here supports the decision to acknowledge that older people, people with disabilities, and those unable to work often require higher temperatures at home in order to protect their health and wellbeing. However, there is no clear and safe indication of whether vulnerable groups could be differentiated in terms of heating needs. For example, there is as yet no convincing evidence that people over 70 need more heating than people over 60, or that people with physical disabilities need higher temperatures than people living with mental illness.

In the context of fuel poverty, the current temperature threshold change (from 21 ºC to 23 ºC) is recommended for people over 60 years old in Scotland. On the one hand, it could be rational to raise the age limit denoting a need for warmer homes to 65 years, based on the fact that this is what WHO originally used as the threshold for defining people as 'elderly' (Ormandy & Ezratty, 2015). This is especially so in cases where people between 60 and 65 are still in good health and in full-time work.

But several other alternatives for segmenting older age groups have also been explored, one of which is a needs-based segmentation (e.g. Shekarriz & Spinelli, 2012). This focuses on transition points in people's lives, which are the points in a lifespan when support is most needed. These could include retirement, at whatever age that occurs, since it generally leads to a change in income and how much time people are at home (both of which will affect the proportion of income needed to heat a home adequately).

Another point of view on the matter, as outlined in Chapter 3, is that in the absence of any long term illness or disability, temperature recommendations should not be automatically raised on an age threshold, since age alone is not always a satisfactory proxy for vulnerability, except in the much later years.

5.6.7. Secondary heating systems and their implications for how a ' satisfactory heating regime' is defined

Some of the studies cited in this report indicate that, in fuel poor households particularly, extra heat is often generated from inefficient and expensive secondary heating sources, such as paraffin heaters and open coal fires. This greatly increases the cost of keeping a home warm, and hence the risk (and severity) of fuel poverty. In this context, perhaps more consideration could be given to not only what temperatures people are able to maintain in their homes, but also how they achieve those temperatures. ' Being able to maintain a satisfactory heating regime through the sole use of a home's primary heating system' is, perhaps, worth considering as an amendment to the current definition of what constitutes a ' satisfactory heating regime'.

5.7. Summary

It seems prudent to retain existing temperature recommendations except in instances where evidence strongly supports change. This seems particularly apposite when current recommendations for temperatures in Scotland so closely resemble those originally set out by the World Health Organization. Although the Working Groups who set the WHO temperature range could not have known their future impact, they met regularly to review previous decisions, commissioned work they knew would guide them, and have retained their 1985 recommendations ever since. In the course of 30 years there have been almost no substantive challenges to these recommendations.

When viewed historically, the WHO's 18 ºC - 24 ºC temperature range has been considered the gold standard for European homes since the 1980's, because of its concern with minimising adverse health impacts. Yet for most of that time, and for the vast majority of UK households, these levels of warmth have been aspirational rather than normative. The narrower band of 18º C (most rooms) and 21ºC (living room) temperatures still exceed what the average UK household is achieving at present. Since most UK experts who were involved in fixing the 18 ºC and 21 ºC set points relied on a weak evidence base (including modelled temperature data that overestimated ambient temperatures) it has been difficult to attain what looked, at the time, like a reasonable thermal increase.

To use the terminology of fuel poverty, 18 ºC - 23 ºC (for vulnerable) and 18 ºC - 21 ºC (for non-vulnerable households) remain ' needs to achieve' temperatures, rather than temperatures which approximate those in the average home. A decision to lower any of the recommended ambient temperatures in Scotland will distance fuel poverty strategies even further from the gold standard that WHO set out.

There may, however, be substantive justification for raising one of the temperatures in the ' satisfactory heating regime', namely the temperature set point for all rooms other than the living room in the homes of vulnerable people. In line with WHO's recommendation, it is argued that this should increase from 18 ºC to 20º.

In sum, any decision about whether to alter the existing definition of what constitutes a ' satisfactory heating regime' in Scotland hinges on:

• whether to remain close to a 30-year old set of WHO recommendations; or
• whether to move towards a set of temperatures that may more accurately reflect what energy efficiency improvements can be practically made in Scottish homes, given the status quo of a cold climate and high fuel prices.

According to previous WHO Panels, they faced a similar dilemma - recommend what an evidence base suggests is ideal, or limit recommendations to what is pragmatic? ( WHO 1961; WHO 2010). Since addressing fuel poverty in Scotland remains, principally, a matter of concern for the nation's health and wellbeing, then applying the principle of ' first do no harm' could reasonably be invoked. In that context, with one exception (which required an upward adjustment of temperature), recent evidence broadly supports the existing temperature recommendations which Scotland uses in defining fuel poverty.

Key Conclusions on indoor temperatures

Temperature should remain the metric used to define what constitutes 'a ' satisfactory heating regime'. Thermal comfort lacks objectivity and is not associated with a concern for protecting human health.

Among non-vulnerable households, the 21 ºC threshold for living rooms should be retained, and also the 18 ºC threshold for all other occupied rooms.

Among vulnerable households, the 23 ºC threshold for living rooms should be retained too, but the 18 ºC threshold for all other occupied rooms should be raised to 20 ºC in order to more fully meet WHO recommendations associated with vulnerability.