Official Statistics

Background information and methodology: Health in the workplace

Published 15 July 2019

Purpose of the statistics

Context of the statistics

The Work and Health Unit (WHU) is a UK government unit which brings together officials from the Department for Work and Pensions (DWP) and the Department of Health and Social Care (DHSC), to lead the government’s strategy to supporting working age disabled people, or people with long term health conditions enter, and stay in, employment.

In Improving Lives: the future of work, health and disability, the government set out an ambitious and comprehensive programme of action to see 1 million more disabled people in work by 2027. It established the need for action in 3 key settings: the welfare system, the workplace and the healthcare system.

In July 2019 the government published Health is everyone’s business: proposals to reduce ill-health related job loss. This consultation focuses specifically on the workplace setting, and the vital role that employers play in helping disabled people and people with long-term health conditions to stay in, and thrive, in work.

Purpose of the statistics

Health in the workplace – patterns of sickness absence, employer support and employment retention has supported an evidence based process of policy development.

This publication provides detailed breakdowns of:

• the characteristics of people experiencing long-term sickness absence (LTSA)
• the difference in employment retention rates experienced by disabled and non-disabled people and how these vary by the type of employer an individual works for and the type of work they do
• how the provision of sick pay and occupational health vary according to the employer an individual works for and the type of job they do

This document sets out the methodology for producing each of these analyses. The data sources used and limitations of the analyses presented are described in the remainder of this document. Where appropriate, the comparability to previous analysis is discussed.

We intend that these statistics will be used by a wide variety of people in the Department for Work and Pensions (DWP) and the Department of Health and Social Care (DHSC), other central government departments, the National Health Service (NHS), Northern Ireland, Scotland and Wales devolved administrations, as well as local authorities and employers across the United Kingdom, amongst others.

The statistics will also be used to answer Parliamentary Questions and requests under the Freedom of Information Act, as well as by journalists and commentators.

Reporting of the statistics

Unless otherwise stated, all the differences reported are statistically significant at the 95% level.

Percentages may not add to 1 due to rounding.

Graphs and figures are based on unrounded figures.

Definitions and terminology within the statistics

Chapter 1 and 2 definitions

Disability: the Government Statistical Service (GSS) Harmonised Standard Definition of disability is used. In summary the core definition covers people who report (current) physical or mental health condition(s) or illnesses lasting or expected to last 12 months or more, and the condition(s) or illness(es) reduce their ability to carry out day to day activities.

In employment: people of working age who either did paid work in the reference week (as an employee or self-employed), had a job that they were temporarily away from, were placed with employers on government-supported training and employment programmes or doing unpaid family work.

Left work following a LTSA: “leaving work for now”, “early retirement” or “doing something else”.

Longitudinal surveys: repeated observations of the same variable over a period of time interviewing the same cohort of people.

Long-term health condition (LTHC): an individual reported having a physical or mental health condition or illness that lasts, or is expected to last, 12 months or more.

Long-term sickness absence (LTSA): a period of 4 weeks or more where an employed individual is prevented from working due to illness or injury.

Mental health (MH) condition: include stress depression and anxiety (common mental health problems) and manic depression, schizophrenia and other serious mental health problems.

Musculoskeletal (MSK) condition: include back pain, neck and upper limb problems and other musculoskeletal problems.

Other condition: includes other respiratory conditions (such as asthma, Chronic Obstructive Pulmonary Disease (COPD), bronchitis, pneumonia). It also includes other gastrointestinal problems (such as irritable bowel syndrome (IBS), piles, bowel cancer, stomach ulcer), headaches and migraines, genito-urinary, urine infections, menstrual problems, pregnancy problems, heart, blood pressure and circulation problems, eye, ear, nose & mouth or dental, sinusitis and toothache, diabetes and other (accidents, poisonings, infectious diseases, skin disorders and anything else not covered above).

Other employment: an individual was on a “government training scheme” or were an “unpaid family worker”.

Retention rate: the percentage of people who are in employment at year 1 and year 2 interviews.

Stayed in work following a LTSA: for employees at the time of the LTSA this was “returned to work for the same employer with adjustments”, “returned to work for the same employer without adjustments”, “returned to work for a different employer” or “became self-employed”. For the self-employed at the time of their LTSA, “returned to work for the same employer with adjustments”, “returned to work for the same employer without adjustments” or “went to work for an employer”.

Working age: respondents aged between 16 and 64 years old.

Workplace: the office or site which an individual works, not just the department. A small branch (less than 50 employees) of a large company (250+ employees) will be classified as a small workplace.

Workplace size: the total number of employees at the respondent’s workplace, not just the particular section or department. People employed by employment services who may work during the course of a week at a number of locations are required to refer to the place where they worked the longest number of hours during the reference week.

• small workplace: 1 to 49 employees at workplace
• medium workplace: 50 to 249 employees at workplace
• large workplace: 250+ employees at workplace

Chapter 3 definitions

Above Statutory Sick Pay (SSP): employer provides amounts or durations of sick pay above the statutory minimum either “company sick pay” or “both” (SSP and company sick pay).

Blue collar: employment involving manual labour.

Full time: employee works 30 or more hours per week.

Occupational Health (OH): occupational health services provide advice and practical support about how to stay healthy in the workplace and how to manage health retention conditions. For example, access to health care professionals, or support or advice for making workplace adjustments.

Occupational Sick Pay (OSP) or company sick pay: employers may choose to pay more than Statutory Sick Pay. Employees can potentially receive both types of pay at different times because employers may offer OSP at first and then begin paying SSP.

Organisation size: the amount of people that work for the organisation that pays your wages at all of its sites or offices in the UK.

• small organisation: 2 to 49 employees at organisation
• medium organisation: 50 to 249 employees at organisation
• large organisation: 250+ employees at organisation
• part time: employee works less than 30 hours per week

Sick pay: a payment in the event an individual is too ill to work. There are 2 types of sick pay which may be provided to employees – Statutory Sick Pay (SSP) and Ooccupational Sick Pay (also known as company sick pay).

Statutory Sick Pay (SSP): is the legal minimum qualifying employees are entitled to when they are unable to work due to ill health. At the time of the survey, employers were obligated to pay £87.55 per week, for up to 28 weeks. The first qualifying condition for SSP is an individual must be absent from work for 4 or more days in a row (including non-working days).

White collar: a person who performs professional, managerial, or administrative work.

Status of the statistics

Ad hoc publication

Health in the workplace – patterns of sickness absence, employer support and employment retention and the statistics contained within it are treated as a one off ad hoc statistical publication.

Quality statement

These statistics have been developed using guidelines set out by the UK Statistics Authority. Extensive quality assurance of each chapter has been undertaken including triangulation to published data sources where appropriate.

The analysis in chapter 1 has had an independent review conducted by colleagues in DWP who have compared these findings to those published on NOMIS, where appropriate.

The analysis in chapter 2 has been compared to previously published DWP findings, where appropriate.

The analysis in chapter 3 uses the data from the published DWP research report Health and wellbeing at work: a survey of employees. Where possible, the findings have been compared to those published including the unweighted bases.

Feedback

We welcome feedback

We welcome any feedback you have on the statistics. Please provide feedback by emailing: whu.ohanalysis@dwp.gov.uk

Useful links

Read the Health in the workplace: patterns of sickness absence, employer support and employment retention ad hoc release and data tables

Long-term sickness absence

Chapter 1 of Health in the workplace – patterns of sickness absence, employer support and employment retention looks at the occurrence of long-term sickness absence (LTSA) in the United Kingdom (UK).

Methodology

Data and sample

This analysis, as in chapter 2, uses the Annual Population Survey (APS) . The APS is a combined survey of households in the UK that provide information on key social and socioeconomic variables between the 10 yearly censuses. The APS is the largest household survey in the UK, with approximately 170,000 working age respondents (16 to 64 years old). The respondents are sampled to be representative of the UK population. The data is weighted to minimise non-response bias and provide estimates of the entire UK working age population. The Annual Population Survey is produced by the Office for National Statistics (ONS).

Chapter 1 analyses 12 months of survey data using the January 2018 to December 2018 APS. Only individuals of working age (16 to 64 years old) were included in this analysis. The LTSA questions are asked to respondents who are currently working, those who are currently on sickness absence or those who left their job less than 12 months prior to their 2018 interview.

The sample corresponds to approximately 30.4 million working age people living in the UK. This number is higher than the total in work because the LTSA module additionally asks those who have worked in the last 12 months.

The module asks individuals whether they have had a spell of LTSA in the last 12 months. If they did have a spell of LTSA in the 12 months prior to their interview, they are asked how many spells of absence they had in that period.

People who are currently on a LTSA and were scheduled work during their interview week, are asked how many spells they have had including the one they are currently on^{[footnote 1]}. The combination of these 2 variables determines how many people had a LTSA and how many spells they had.

Those who have had at least 1 spell of LTSA are then asked the duration of their LTSA, the main health condition causing it, whether they were an employee, self-employed or other at the time of the LTSA and also what they did following their LTSA (the “destination” of their LTSA for example a return to work with adjustments). For individuals with multiple spells of LTSA in the last 12 months, these questions are answered for their longest LTSA spell.

Some of the respondents who are still on a spell of LTSA answered their destination as something other than “still on LTSA”, this is because respondents report the destination of their longest spell of LTSA.

The analysis provides breakdowns of an individual’s characteristics by gender, age, ethnicity, marital status, highest qualification, disability status, whether they are an employee or self-employed and their occupation. These characteristics are not necessarily causing different absence rates, but should give a broad view of the sorts of people who experience such absences.

This analysis also looks at trends of LTSA of employees by the characteristics of their employer, sector, workplace size and industry. Again, we cannot assume that employer characteristics are causing the differences in absence rates.

These characteristics are asked at the time of the interview (not when the LTSA occurred), respondents answer the question as is the case at the point of the interview. Therefore, for some characteristics, the answers may be different than they were at the time of their LTSA, however we would expect gender, age, ethnicity, marital status and highest qualification to remain relatively stable over time. Disability status, employment type and employer characteristics are more likely to change due to a LTSA spell or over time.

Health

An individual’s main health condition causing the absence is grouped into “musculoskeletal” (MSK) condition, “mental health” (MH) condition and “other condition”. MSK is defined in the APS as either “back pain”, “neck and upper limb problems (such as arthritis in hand joints, stiff neck)” or “other MSK problems”. MH is defined in the APS as either “stress, depression, anxiety (common mental health problems)” or “manic depression, schizophrenia and other serious mental health problems”. All other health categories^{[footnote 2]} are combined into “other”.

Destination following LTSA

An individual is defined as “left work” immediately after their LTSA if they either. They left their job and stopped working for now, took early retirement or did something else.

For employees at the time of their LTSA, those that “remained in work” either returned to work for the same employer with a new or adjusted role or working pattern, returned to work for the same employer without a new or adjusted role or working pattern, returned to work for a different employer or, became self-employed.

For self-employed individuals at the time of their LTSA, those that “remained in work” either returned to work with a new or adjusted role or working pattern, returned to work without a new or adjusted role or working pattern or went to work for an employer.

86% of individuals return to the same employer following their LTSA spell. The 4% of individuals that returned to a different employer would have reported the characteristics of their new employer.

Cases with unknown LTSA

There are approximately 5,800 respondents who should have answered the LTSA questions but have not. Our investigations show that 96% of these approximately 5,800 respondents did not complete their fifth wave Labour Force Survey (LFS) and therefore their answers from the fourth wave have been carried over or ‘brought forward’. As the LTSA questions are only asked to respondents in wave 1 and 5 of the LFS, we do not know if these respondents have had a LTSA in the last 12 months. We have considered the characteristics of these individuals and found that they are broadly similar to the sample of people who did answer the LTSA module. We therefore assume they have similar levels of LTSA as those cases within our LTSA sample.

These approximately 5,800 people equate to roughly 2 million people when weighted to be representative of the UK population. As all of the ‘brought forward’ cases were in work in their wave 4 interview, we apportioned the in-work likelihood of LTSA to these individuals based on the sample that answered their wave 5 LFS. This is, in part, because it is an established practice within labour market statistics. This means that our adjusted volumes are higher than which can be obtained from the raw APS data.

Prior to the adjustment there were approximately 1.3 million people that had a spell of LTSA in the previous 12 months, our adjustment, accounting for the ‘brought forward’ cases increased this by approximately 85,000 people.

This analysis assumes the ‘brought forward’ cases are as likely to have multiple spells as those who answered the wave 5 interview. This increases the total amount of spells every year by 106,000.

The adjustment, when rounded, increases the total amount of people who have a spell of LTSA to 1.4m people every year which equates to 1.8 million spells.

The rest of the analysis in this chapter, including the percentages are based on the unadjusted spells or people who had a LTSA. As explained above, we do not see significant differences in the characteristics of the individuals that were ‘brought forward’ and hence do not expect any systematic biases.

Spells

There are approximately 32 respondents (9,000 individuals when weighted) that report having had a LTSA but do not answer how many separate spells they have had within the 12-month period. For these individuals, as we know they have had at least 1 spell of LTSA, we assume this was their only spell of LTSA in the 12 month period (88% of respondents had a single spell of LTSA in the last 12 months). As we assume only 1 spell of LTSA, this figure could be slightly under estimating the total volume of LTSA spells.

Duration

Respondents who are still on a spell of LTSA cannot accurately report the duration of their absence. We therefore exclude these when looking at the duration of LTSA spells. Respondents with multiple spells of LTSA report the duration of their longest spell.

Comparison with previous estimates

The LTSA estimates presented here should not be compared with previously published estimates of LTSA due to substantial methodological differences.

LTSA rates were presented in the Work, health and disability green paper: data pack published in October 2016. Since these estimations, the LTSA questions in the APS have been reviewed and amended to improve the quality and range of questions. For full detail on the previous methodology used in the 2016 Green Paper, refer to the background methodology document.

Caveats

Sample bias

When using a sample of individuals to understand an entire population, there is always an inherent sample error. This means that our random sample may not be indicative of the whole population and the results could be inaccurate. To give an idea of the potential scale of this problem we have included the 95% confidence intervals (CI) for the proportion of people who have a spell of LTSA among those who are working, worked in the last 12 months or are currently on a sickness absence. This CI displays the range of values that the figure can take for us to be 95% certain that the true value lies within.

The confidence intervals for the unscaled values (not accounting for the ‘brought forward’ cases) are as follows:

	Central	Lower 95% estimate	Upper 95% estimate
Number of people who had a spell of LTSA in the last 12 months (millions)	1.35	1.31	1.38
Proportion of all who are working, have worked in the last 12 months or currently on sickness absence (%)	4.4%	4.3%	4.6%

The confidence intervals around the volume estimates should be considered as illustrative as the central estimate do not contain the adjustment for the ‘brought forward’ cases which creates additional uncertainty. Additionally, confidence intervals do not account for all types of possible error.

Reporting error and questionnaire wording

As with any question on a survey, it is possible that the question may be misinterpreted and answered incorrectly.

LTSA questions are asked of those who are in work or left work in the last 12 months or are currently on a sickness absence. The employer characteristics questions are only asked of those in employment at the time of questionnaire.

To calculate the likelihood of experiencing a LTSA by employer characteristics, we have restricted the sample to those who are employees at the time of the questionnaire. This means that those who have had a LTSA but are not in work at the time of the interview are not included in employer characteristics analysis. This could be a cause of potential bias as individuals that work for employers with certain characteristics may be more likely to not return to work following a LTSA. However, if there is a bias we would expect it to be small as the majority of people (86%) return to work for the same employer following their LTSA.

Variations in employment retention rates for disabled people

Chapter 2 of Health in the workplace – patterns of sickness absence, employer support and employment retention looks at the employment outcomes for people with disabilities and how it varies depending on the employer they work for and the job they do. It additionally, compares the employment outcomes of disabled people to non-disabled people.

Methodology

Data and sample

This analysis uses the 2 year longitudinal APS which interviews the same individuals at 2 distinct time periods, 1 year apart^{[footnote 3]}. Longitudinal datasets allow respondent’s movement into and out of employment to be analysed at 2 points in time.

This analysis identifies the flows of people that move from employment to either unemployment or inactivity between 2 points in time. This analysis does not identify if an individual has experienced other changes in their employment status between these 2 time periods.

An individual has retained employment if they are in employment in both years (even if they have experienced a period of worklessness in between, not captured by the survey). An individual that changes their type of employment, for example an employee in the first period became self-employed in the second will be classified as “remained employed”. This analysis is not designed to capture every movement in and out of employment but give a picture of how people’s employment status can change from 1 year to the next.

This analysis is limited to people who are of working age (16 to 64 years old) in both year 1 and year 2 of the survey. People who were outside of working age at one of the interviews were excluded. This is to focus on employment changes during the course of people’s working lives, without the results being affected by the large number of ‘routine’ transitions that can occur as they move in or out of working age.

As this analysis seeks to look at the retention of disabled people in the workplace, it includes only those individuals who kept their disability status constant from 1 year to the next. This means those whose disability status changed from year 1 to year 2 were excluded. Our estimate of those in work, approximately 23 million, is therefore lower than the total in work population.

The reason this analysis was restricted to those individuals that did not experience a change in disability status at the time of the interviews was to identify trends in employment outcomes rather than changes in employment outcomes which could be driven by changes in an individual’s disability status.

This analysis uses the January to December 2016 to January to December 2017 longitudinal APS dataset.

This analysis looks at variations in the retention rates of individuals by their characteristics and the characteristics of their employer. The longitudinal dataset does not ask respondents the full range of employer and personal characteristics questions that the cross-sectional APS does. Therefore, this analysis merges the relevant cross-sectional datasets onto the longitudinal APS (January to December 2016 and January to December 2017).

This analysis uses the characteristics as reported in the first time period.

These characteristics are not necessarily causing the retention rates, but should give a broad view of the sorts of people who experience such employment outcomes.

Comparison to other estimates

The estimates presented here should not be compared with previously published estimates of retention due to the substantial methodological differences.

Characteristics of disabled people in employment: April to June 2017 and Improving Lives The Future of Work, Health and Disability Improving Lives The Future of Work, Health and Disability analyse the flow of disabled people and non-disabled people into and out-of-work. In both of the publications an individual’s disability status is defined in the first year, therefore, these 2 pieces of analysis should not be compared.

In 2016, Citizens Advice published Working with a health condition or disability containing similar analysis showing that people who are disabled or have a health condition can face a range of barriers to work. The publication shows disabled people are much less likely to be employed and are twice as likely to fall out of work.

Although these results are similar to this publication, these figures use a different methodology and only cover England and Wales, rather than the whole of UK.

Caveats

As this analysis is based on longitudinal survey data, the precision and accuracy of the estimates can be affected by response errors, sampling errors, and attrition bias (if people who were interviewed in all waves are not representative of the wider population). This analysis is weighted to account for attrition bias and ensure the analysis is representative of the UK population.

To give an idea of the potential scale of sampling error we have included the 95% confidence intervals (CI) for the proportion of people that retain employment. This CI displays the range of values that the figure can take for us to be 95% certain that the true value lies within.

The confidence intervals are as follows:

	Central	Lower 95% estimate	Upper 95% estimate
Number of disabled people who retained work (millions)	1.97	1.94	1.99
Proportion of all disabled people in work in year 1	89.2%	88.1%	90.2%
Number of non-disabled people who retained work	23.16	23.11	23.22
Proportion of non- disabled people in work in year 1 (millions)	95.2%	95.0%	95.4%

Employer variation in support

Chapter 3 of Health in the workplace – patterns of sickness absence, employer support and employment retention explores how the job an individual has, and the type of employer they work for, affects the individual’s likelihood of receiving services such as occupational health and sick pay above the statutory minimum. This is further analysis of the 2015 DWP research report Health and wellbeing at work: a survey of employees.

Methodology

Data and sample

This analysis uses data from the survey used for the 2015 DWP research report Health and wellbeing at work: a survey of employees, 2014. The survey interviewed a representative sample of 2,013 employees by telephone via ‘Random Digit Dialling’ (RDD). This included mobile numbers to capture ‘mobile only’ households and ex-directory numbers (because of the random generation of lists). Fieldwork took place between January and April 2014.

Due to the low prevalence of employees who had been off work for more than 2 weeks in the general population, a boost sample was necessary. The boost element was separate to the RDD approach and involved following up respondents from the Health Survey for England, the Scottish Health Survey and Welsh Health Survey, who had specific characteristics that made them more likely to have had time off sick. The boost provided 219 further interviews. Finally, a consumer access panel provided contact details for a further 139 cases.

The total number of employees interviewed was 2,371.

Eligibility for the survey:

• those aged 16 or over in Great Britain who had done any paid work as an employee for 7 hours or more in any week in the last month
• employees were defined as those working for organisations of 2 or more people. Consultants were not included, but casual staff and those on temporary contracts were
• those on maternity leave or sick leave who had a contract of employment were also included

A weight has been applied to the data that adjusts for selection probabilities and calibrates to Labour Force Survey Statistics for employees. The 2 samples, the RDD and boost, were brought together in this process in order to enable analysis of the boosted group in the context of the wider employee population. The weight’s main effect is to bring the level of those who have had more than 2 weeks off sick back to their population level following the boost.

The combined sample was weighted to provide estimates that are representative of the GB population of employees aged 16 and over.

“Above SSP” is defined here as when an employee has been or would be paid more than the statutory minimum at some point during their sickness absence, not necessarily for the entirety of the absence. “SSP only” is defined as when the employee was or would be paid the statutory minimum for the entirety of their sickness absence, if they were eligible for sick pay.

Comparison with previous estimates

The analysis uses the data analysed within the published 2015 DWP research report Health and wellbeing at work: a survey of employees, 2014.

A previous DWP survey Health and well-being at work: a survey of employees published in 2011 estimated the occurrence of SSP and access to OH services. As stated in the 2015 report, access to OH services significantly increased over this time period.

Caveats

Only those who stated they were paid sick pay by their employer were asked which type of sick pay they were paid.

1. Our analysis showed 914 respondents reported being “still on LTSA” at the time of the interview, the majority (869 people, 95%) of this group were not scheduled work in their reference week, that their scheduled working days were not relevant or their days scheduled to work “does not apply”. As they were not scheduled work, individuals were asked the LTSA questions as if they were not currently on a LTSA. Of the remaining 45, 22 had not had days off work because of sickness or injury during their reference week. The remaining 23 had not been off work for 4 weeks or longer. We removed these individuals from the destination analysis as their answers are not consistent with a completed spell of LTSA. As we remove “still on LTSA” when analysing the destination of individuals, we are confident that the sample used for this analysis does not contain an inherent bias. These 45 individuals are included in the remained of the analysis as we have no reason to suspect problems (such as arthritis in hand joints, stiff neck)” or “other MSK problems”. MH is defined in the APS as either “stress, depression, anxiety (common mental health problems)” or “manic depression, schizophrenia and other serious mental health problems”. All other health categories^{[footnote 2]} are combined into “other”. ↩

2. Other respiratory conditions (for example asthma, Chronic Obstructive Pulmonary Disease (COPD), bronchitis, pneumonia), other gastrointestinal problems (such as irritable bowel syndrome (IBS), piles, bowel cancer, stomach ulcer), headaches and migraines, genito-urinary to include urine infections, menstrual problems, pregnancy problems, heart, blood pressure and circulation problems, eye, ear, nose & mouth or dental to include sinusitis and toothache, diabetes and other (accidents, poisonings, infectious diseases, skin disorders and anything else not covered above). ↩ ↩²

3. Annual Population Survey QMI, 2012 ↩