Indicators and determinants of the years of working life lost: a narrative review

Working life expectancy

WLE is defined as the number of years a person is expected to work over a lifetime [18, 22, 23]. Studies have analysed working life expectancies from two analytical perspectives: the cohort perspective, that is, the average lifetime spent in work of an actual birth cohort; and the period perspective, that is, the expectancy of a fictional cohort who would experience stable levels of work participation rates. Some studies defined WLE as the number of years a person is expected to be in paid employment, others as the number of years a person is economically active, including years spent in both employment and unemployment [18, 24].

In some official statistics the length of working life is defined as WLE at the age of 15 [24]. However, studies focusing on the older population have defined it, for example, as average time spent in employment between 50 and 74 [23]. WLE estimates vary substantially, depending on analytical approach (period vs. cohort perspective and Sullivan vs. multi-state modelling) [16, 22], country context, sex, and socioeconomic factors. For example, in Europe, from 1995 to 2001, average WLE at the age of 50 using the multi-state life table approach was 9.4 years for men and 6.4 years for women [16]. For men this ranged from 7.0 to 7.3 years in Austria, Belgium and France and from 10.9 to 11.4 years in Denmark, Greece, Portugal and the United Kingdom; for women, WLE was 3.3 years in Italy and ranged from 8.8 to 9.4 years in Denmark, Finland and Germany [16]. Only in Finland were women found to work a few months longer than men [16]. In 2009, WLE at the age of 50 for men using the Sullivan method ranged from 7.4 years in Hungary to 13.0–13.9 years in Norway, Sweden and Switzerland to 16.5 years in Iceland, and for women it ranged from 5.6 to 5.9 years in Italy, Luxembourg and Poland to 11.6–11.9 years in Denmark, Norway and Sweden to 13.6 years in Iceland [24]. In Estonia, Finland and Latvia, WLE at the age of 50 for men was a few months longer for women than men. The difference in WLE between men and women in 2009 was largest (3.7–4.2 years) in Ireland, Italy and Spain [24]. In Finland, using the Sullivan method, WLE at the age of 50 was 10 years for women and 9 years for men in 2012 [22]; it was 10 years for both men and women between 2005 and 2014 using the multi-state life table approach [25]. WLE at the age of 50 has been increasing across Europe and the United States since the mid-90s [22-24]. However, the increase has been larger in women than in men, leading to a decrease in the sex difference in WLE [24].

Healthy working life expectancy

Healthy WLE is defined as the number of years on average a person can expect to work in good health (without a limiting long-standing illness) between 50 and 70 years of age [4, 16]. Healthy WLE is higher for men than for women. In Europe, it ranges between 5.5 and 9.7 years for men and between 2.9 and 8.3 years for women [4, 16, 26, 27]. During 1995–2001, average healthy WLE at the age of 50 in Europe was 7.5 years for men and 4.8 years for women [16]. This ranged from 5.5 years in France to 9.7 years in Greece for men and from 2.9 years in Italy to 6.1–6.2 years in Denmark and Finland for women [16]. It was 9.4 years for men and 8.3 years for women in England during 2002–2013 [4].

Years of working life lost

YWLL is defined as a summary measure of different exits from the labour market [18]. However, the majority of previous studies have not used YWLL as a summary measure of different exits from employment; they determine YWLL to be a single premature exit route such as disability retirement, early retirement or unemployment. Only two studies included three [17] or several [18] exit routes from the labour market in their calculation of YWLL. Disability retirement is a common cause of YWLL [17-18].

Occupationally active life expectancy

Occupationally active life expectancy measures the average age at which work capacity ends [28]. With no occurrence of disability pension or death before retirement, the occupationally active life expectancy would be equal to statutory retirement age [28].

Disability-adjusted working life years

Disability-adjusted working life years (DAWLYs) estimate the number of working years lost due to a disease or its consequences [29]. One DAWLY represents one lost year of heathy working life. The calculation of DAWLYs is similar to that of disability-adjusted life years [30]. DAWLYs for a disease are calculated as the sum of working years lost due to premature death and the number of working years lost due to disability.

Potential gains in life expectancy

By not considering competing risks, the years of potential life lost to premature death can be overestimated [25]. Potential gains in life expectancy, which does incorporate competing risks into the model, estimates the potential gains if a particular cause of deaths can be eliminated [31]. This indicator can also be used for working population or as potential gains in WLE.

Working life expectancy scale

The working life expectancy scale (WLE scale) uses a visual analogue scale to assess a worker’s expectation of maintaining their current job [32]. Workers are required to respond to a single statement: ‘I can keep working in this job for 5 more years’. The response scale ranges from 0 (never) to 100 (always) at intervals of 10, higher scores indicating a higher expectation of maintaining the current job [32].

Lost-work opportunity score

The lost-work opportunity score consists of unemployment, forced retirement and earlier than planned retirement [33]. It considers early retirement due to temporary lay-offs, company buy-outs, forced relocations, and so forth [33].

Potential of years of working life lost

Potential of years of working life lost (PYWLL) is estimated as YWLL plus expected losses occurring from the absence of return-to-work. It describes the difference between official and actual retirement ages [19].

Inactivity ratio

The inactivity ratio is calculated as YWLL or PYWLL divided by active age range (15–65 years) [19]. It gives the percentage of time in inactivity. The inactivity ratio was 16% for YWLL and 30% for PYWLL in patients with osteoarthritis [19], and 13% for YWLL and 25% for PYWLL in patients with rheumatic diseases in Portugal [34].

Work environment disability-adjusted life year

Work environment disability-adjusted life year (WE-DALY) assesses the number of fatal and non-fatal injuries and illnesses occurring in industry because of exposure to workplace hazards [35]. It is the sum of the number of years of life lost by premature mortality in the worker population and the number of years of life lived with a disability for each non-fatal injury or illness [35].

The index of potential years of work tenure lost

The index of potential years of work tenure lost (IPYWL) is calculated by dividing the sum of potential years of work tenure lost (PYWL) (sum of the number of deaths in a working age group × remaining working years of that age group) by expected PYWL (PYWL × standard proportion [calculated by numbers of death in a working age group divided by all workers]) [36].

The work lost rate

The work lost rate is a measure of absenteeism [37]. It is calculated as ‘hours unpaid/(work hours paid + hours unpaid) × 100’. Workers younger than 30 years old had a higher work lost rate, while workers older than 60 years old had a lower work lost rate due to a small number of workers with longer absences [37]. Unpaid hours included sick days and absent days due to personal reasons.

Sociodemographic factors

Education was positively associated with labour force participation and WLE [24]. People with higher levels of education had a higher WLE than people with lower levels of education [18, 38]. In the Netherlands, low-educated men had a 7-year and less educated women had a 10-year lower WLE at the age of 30 than highly educated people [18]. This difference was 2.5 years for men and 3.4 years for women at the age of 50 [18]. In the Netherlands and the United States, educational inequalities in WLE are larger for men than for women [18, 38].

Healthy WLE was also inversely associated with level of education, which was longest for people with a tertiary education (11.3 years) in England [4].

Older workers (50 years+) with a low level of education more frequently exit from paid employment due to ill health than workers with a high level of education [39]. In the Netherlands, low-educated men lost 6.3 years due to unemployment and 3.4 years due to disability retirement, while highly educated men lost 1.9 and 0.8 years, respectively [18]. Low-educated women lost 7.0 years due to unemployment and 3.0 years due to disability retirement, whereas highly educated women lost 1.8 and 1.4 years, respectively [18]. Low-educated workers more commonly suffer from diabetes, cardiovascular diseases and musculoskeletal disorders than highly educated workers [40], and they lose more working years due to disability retirement than highly educated people [40, 41], particularly disability retirement due to musculoskeletal diseases [41].

Older workers (50 years+) with a low occupational grade more frequently exit from paid employment due to ill health than workers of a high occupational grade [39]. In Sweden, blue-collar workers lost more working years than white-collar workers [5]. In South Korea, manual workers were at increased risk of disability retirement [42]. In Finland, WLE at the age of 50 was 10.5 years for upper-level non-manual workers and 9.5 years for manual workers [25]. Self-employed Finnish workers had the largest WLE at the age of 50 [25], and in 2012, upper non-manual workers had a 3.5-year higher WLE than manual workers [22]. Healthy WLE (11.9 years) [27] and occupationally active life expectancy (62 years) [28] were highest for Finnish male executives, while occupationally active life expectancy was lowest for unskilled workers (52 years) [28]. In England, healthy WLE was longest for self-employed (11.8 years) and non-manual occupations (10.3 years) and was shortest for people with manual occupations (8.7 years) [4].

The 2008–2009 financial crisis reduced WLE [38, 43]. In Spain, the recession reduced WLE in men by 12 years and in women by 7 years [43]. WLE changed more among manual and unskilled workers than among non-manual and skilled workers [43]. The loss of working years was mainly due to unemployment or being out of the labour market (inactivity) [43]. In people aged 50, WLE declined by 1.5 years in men and by 0.5 years in women in the United States [38]. In a study conducted among Finnish men [28], marital status was associated with occupationally active life expectancy. Mean occupationally active life expectancy was shortest for single men (50 years) and longest for ever-married men (59 years).

In the United States, foreign-born people had a lower WLE at the age of 50 than native-born people [23]. The gap in the duration of working life at the age of 50 between native- and foreign-born people in the United States has increased over time [23]. Healthy WLE was inversely associated with area-level deprivation, and in England healthy WLE was longest for people living in the least deprived areas (10.5 years) and shortest for those in the most deprived area (6.8 years) [4].

Lifestyle risk factors

In South Korea, workers who did not exercise regularly or were underweight or overweight were more likely to retire before the age of 65 than workers who exercised regularly or were normal weight [44]. People who were physically active for at least 2 days a week were more likely to work beyond retirement age [45]. Disability retirement is more common in people with overweight or obesity [46], in smokers [42, 47] and in those who are inactive during their leisure time [48] compared with normal weight, never smokers or active individuals, respectively.

Work-related risk factors

In Denmark, at the age of 30, men exposed to high physical work demands had 2 years lower WLE than low-exposed men, and women exposed to high physical work demands had 3 years lower WLE than low-exposed women [49]. Workers exposed to high physical work demands lost more working years due to unemployment, sickness absence and disability retirement than low-exposed workers [25, 49]. In the Netherlands, workers who felt underappreciated at work or had a low focus on their developing skills and knowledge were more likely to retire early [50]. In older workers (50 years+) in England, psychosocial demands predicted preference for early retirement, while decision authority predicted preference for late retirement and reduced work exits [51]. Manual workers were at increased risk of disability retirement [42], and physically heavy jobs [52] or physical workload [25, 53] increased the number of working years lost due to disability retirement. The lifting or carrying of heavy loads, hand tool vibrations and hard physical work during working life increased the number of working years lost due to long-term sickness absence or disability retirement in Danish older workers [54].

In workers with a chronic disease, low supervisor and co-worker social support, and burnout predicted exits from paid employment [55]. In workers aged 45–63 with a chronic disease, a favourable change in physical workload reduced the rate of exit from paid employment [56].

Chronic diseases

Workers with a chronic disease exit paid employment due to disability retirement, unemployment and early retirement more frequently than workers without such conditions [55]. Cardiovascular disease, chronic mental illness, diabetes and chronic musculoskeletal disease are among the most frequent causes [55]. People with depressive symptoms spend more time unemployed and absent through sickness than those without depressive symptoms, particularly women [57]. In patients with osteoarthritis, early retirement contributed the most to YWLL, followed by unemployment and disability retirement [19]. Older Canadian workers suffering from mental- or musculoskeletal disorders were at increased risk of non-employment, while older workers suffering from diabetes or cardiovascular disease were at increased risk of early retirement [12]. A South Korean study reported an average reduction in WLE of 9.7 years because of disability retirement [42]. In Finland, impaired glucose metabolism reduced participation in working life [58], and disabling shoulder lesions reduced working life by 2 to 8 years compared with the general population [59]. In Canada, arthritis or rheumatism at the age of 15 were found to reduce WLE of men by 4 years and of women by 3 years [60].

Evidence from Northern European countries shows that disability retirement is awarded at a younger age for a mental disorder than for a musculoskeletal disorder or any other somatic disorder [21]. Thus, mental disorders cause more YWLL due to disability retirement than musculoskeletal disorders [21]. In Norway, depression and anxiety are the most common causes of working years lost due to mental disorders [21]. In Sweden, disability retirement due to psychiatric disorders was highest in men aged 20–29 and lowest in men aged 60–64 [52]. In the Netherlands, workers with poor physical health and those in a financial position to be able to stop working before statutory retirement age were more likely to retire early [50]. In workers with a chronic disease, both poor self-rated health and extent of activity limitation from the disease predicted exit from paid employment [55]. The number of pain sites and the severity of pain were associated with lower scores on the WLE scale [32].