Advertisement

Diabetologia

, Volume 62, Issue 7, pp 1195–1203 | Cite as

The impact of diabetes on productivity in China

  • Thomas R. HirdEmail author
  • Ella Zomer
  • Alice Owen
  • Lei Chen
  • Zanfina Ademi
  • Dianna J. Magliano
  • Danny Liew
Article

Abstract

Aims/hypothesis

Diabetes increases the risk of premature death and reduces work productivity. We estimated the impact of diabetes in China in terms of mortality, years of life lost, and productivity-adjusted life years (PALYs) lost in the Chinese population.

Methods

Life table modelling was used with simulated follow-up of those with diabetes in the Chinese population of working age (20–49 years in women and 20–59 years in men) until retirement age (50 years for women and 60 years for men). Data regarding the prevalence of diabetes, as well as excess mortality, labour force dropout and productivity loss attributable to diabetes, were taken from published sources. Models were constructed for the cohort with diabetes and repeated for the same cohort assuming that they had no diabetes. The differences in number of deaths, years of life lived and PALYs lived between the two models reflected the impact of diabetes. The WHO standard 3% annual discount rate was applied to years of life and PALYs lived.

Results

In 2017, an estimated 56.4 million people of working age in China (7.1%) had diabetes. With simulated follow-up until retirement, those with diabetes were predicted to experience an estimated 4.1 million more deaths, the loss of an additional 22.7 million years of life (3.7%) and the loss of an additional 75.8 million PALYs (15.1%). This was equivalent to an average of 1.3 PALYs lost per person with diabetes. Based on gross domestic product (GDP) per full-time worker in 2017, the loss in PALYs equated to a total of Chinese ¥17.4 trillion (US$2.6 trillion) in lost GDP owing to reduced productivity, with an average of ¥307,925 (US$45,959) lost per person with diabetes.

Conclusions/interpretation

Our study demonstrates the significant cumulative impact of diabetes on productivity across the working lifetime in the Chinese population, highlighting the potential economic benefits of diabetes prevention in the longer term.

Keywords

Cost Diabetes Productivity 

Abbreviations

GDP

Gross domestic product

ILO

International Labour Organization

OECD

Organisation for Economic Co-operation and Development

PALY

Productivity-adjusted life years

QALY

Quality-adjusted life years

UN

United Nations

WPP

World Population Prospects

Introduction

The People’s Republic of China is the epicentre of the worldwide diabetes epidemic, with an estimated 114.4 million people with diabetes in 2017 [1]. This equates to one in four of people with diabetes worldwide living in China and follows a rapid increase in the prevalence of diabetes in China, from 0.7% in 1980, to 2.5% in 1994, 5.5% in 2000 and 10.9% in 2013 [2, 3, 4, 5]. While the prevalence of diabetes is highest in older age groups, it continues to rise among younger people in China [6]. Furthermore, there is evidence to suggest Asian populations develop symptoms at a younger age and experience greater severity of complications and risk of premature mortality, compared with Europid populations [7, 8, 9]. The burden of disease is potentially greater in younger populations owing to increased years lived with disease and higher risk of chronic complications [10].

Diabetes-related morbidity can lead to reduced workforce participation and productivity while at work, including more work days lost to ill health (absenteeism) and reduced efficiency at work (presenteeism) [11, 12]. The resulting loss of productivity can impose an economic burden on individuals, employers and governments through reduced earnings, tax revenue and gross domestic product (GDP) [13, 14]. In the USA, an estimated $89.9 billion was lost owing to diabetes-related productivity losses in 2017, including diabetes-related absenteeism ($3.3 billion), presenteeism ($26.9 billion), reduced labour force participation ($37.5 billion), and premature deaths attributed to diabetes ($19.9 billion) [13]. In China, this has been less well studied, but the IDF estimated that diabetes cost China approximately US$109.8 billion in 2017 [1]. However, these estimates were based on the ‘direct’ costs of diabetes relating to healthcare expenditure, and did not incorporate ‘indirect’ costs, including diabetes-related productivity losses. Estimates of productivity loss are important to capture the broader economic burden of diabetes and to inform the case for investment in its prevention and control [15]. In the present study, we sought to estimate the impact of diabetes on the Chinese population, both in terms of years of life lost and productivity-adjusted life years (PALYs) lost due to diabetes [16, 17].

Methods

Our analyses utilised multistate life table models [18], constructed for separate sex and age (in 5-year age groups) cohorts of the Chinese population aged 20 to 49 years in women and 20 to 59 years in men, with follow-up until 60 years in men and 50 years in women (retirement age) [19]. Official retirement age for female ‘professionals’ is 55 years (including medical personnel and other professions) and 50 years for all other female workers [19, 20]. However, owing to a lack of data regarding diabetes prevalence within professions, retirement age was conservatively assumed to be 50 years for all women.

Age-specific mortality rates, workforce statistics and measures of productivity were used to simulate the progression of these cohorts until death or retirement age, measuring cumulative years of life and PALYs lived. Data were derived from a combination of publicly available datasets and published sources, shown in electronic supplementary material (ESM) Table 1. First, the life table model estimated these variables for the working-age population who had diabetes. The cohort was then re-simulated with the hypothetical assumption that they did not have diabetes, with relevant changes to mortality rates, labour force rate and productivity indices (see below). The differences in total years of life lived and PALYs lived between the two cohorts reflected the impact of diabetes. The WHO standard 3% annual discount rate was applied to all years of life and PALYs lived [21].

Population and mortality rates

The demographic profile of the cohort was derived from the 2017 China Statistical Yearbook, stratified by sex and 5-year age groups [22]. Diabetes prevalence estimates from the 2017 IDF Diabetes Atlas were used to calculate the number of people with diabetes in the population by sex and age group [1]. Sex- and age-group-specific mortality data for 2017 were extrapolated from the 2010 census data using temporal trends in adult mortality rates for China from the United Nations (UN) World Population Prospects (WPP) [23, 24]. These were attributed to those with and without diabetes based on age- and sex-specific diabetes prevalence and the RR of all-cause mortality associated with diabetes in Chinese populations derived from a national prospective study of adults with diabetes (ESM Table 2) [25]. Mortality rates were obtained for 5-year age bands, and extrapolated using exponential functions to provide rates for age in single years (chosen for best fit, R2 = 0.96–0.99), assuming that the rate for a 5-year age group applied to people in the midpoint of that age band. We projected temporal trends in population mortality risk across the model time horizon using average annual proportional reduction in adult mortality in China (1.0% per year) from the UN WPP forecast [24]. Annual age- and sex-specific mortality rates were applied to the model in yearly cycles with deaths assumed to have occurred at the midpoint of the year.

Labour force participation

Sex- and age-specific labour force participation in China were drawn from International Labour Organization (ILO) estimates for 2017 [26]. Labour force participation was lowest in those aged 20–24 years in both men (69.4%) and women (65.0%) and highest in men aged 30–34 years (97.0%) and women aged 25–29 years (84.2%).

Productivity indices

Diabetes-related productivity loss was characterised using two productivity indices: diabetes-related labour force dropout, which captures the shortfall in labour force participation in those with diabetes compared with those without diabetes, and a productivity index, which reflects the productivity of an individual as a proportion, ranging from 0 (entirely unproductive) to 1 (entirely productive), and captures impairment to productivity due to a health condition [16, 17]. These inputs were derived from estimates of absenteeism, presenteeism, and labour force participation in those with diabetes compared with those without diabetes [27]. Diabetes-related labour force dropout was expressed as labour force participation percentage shortfall and ranged from 7.0% in women and 5.2% in men with diabetes aged 20–29 years to 12.8% in women and 8.3% in men with diabetes over 40 years, respectively [27]. These relative reductions were applied to 2017 ILO sex- and age-group-specific population labour force participation rates to derive the labour force participation rates in those with and without diabetes. In the absence of data on the division of the labour force into full- and part-time employment by disease status, all employees were assumed to be in full-time employment. In the model, years of life lived by the cohort were multiplied by the labour force participation rate to calculate years lived in the labour force.

To estimate PALYs lived by the diabetes cohort, each year lived in the labour force by the cohort was multiplied by a productivity index derived from estimates of diabetes-related absenteeism and presenteeism [16, 17]. This is akin to multiplication of years of life lived by utilities to derive quality-adjusted life years (QALYs) [28]. Absenteeism was defined as the number of lost work days per year owing to diabetes and was expressed as a percentage of the total working days per year, while presenteeism was defined as self-assessed productivity loss while at work and expressed as a percentage of total productivity. Absenteeism was estimated to be 10.2 days per year in women and 1.9 days in men [27], which, as a proportion of the 245 maximum working days per year in China, represents a 4.1% and 0.8% reduction in productivity, respectively. The shortfall in productivity due to diabetes-associated presenteeism was 1.0% in women and 0.6% in men [27]. The available evidence did not allow for stratification of absenteeism and presenteeism estimates by age group. The combined productivity diabetes-related shortfall owing to absenteeism and presenteeism was thus assumed to be 5.1% in women (productivity index = 0.95) and 1.8% in men (productivity index = 0.98). The productivity index in those without diabetes was assumed to be 1.0.

Data on the GDP per worker were drawn from the 2018 Organisation for Economic Co-operation and Development (OECD) Compendium of Productivity Indicators, and in China in 2017, the figure was ¥179,486 (US$26,789) [29]. We assumed that the economic value of each PALY was equivalent to annual GDP per worker. We projected temporal trends in GDP across the model time horizon using the OECD long-term GDP forecasts [30].

Sensitivity and scenario analyses

First, the individual contribution of absenteeism, presenteeism, labour force dropout and premature mortality to productivity loss were assessed. Second, deterministic sensitivity analyses were undertaken to assess the impact of uncertainty around diabetes-related mortality risk, productivity indices, and economic data inputs on the model and PALYs lost in those with diabetes in the Chinese population. These include: upper and lower uncertainty bounds around estimates of all-cause mortality risk associated with diabetes based on the upper and lower 95% CI around estimates of RR, respectively [25], and the upper and lower uncertainty bounds around productivity indices based on decreasing and increasing estimates of absenteeism, presenteeism, and labour force dropout by 25%, respectively [27]. Finally, scenario analyses were undertaken to explore other model assumptions and compared with the base case, including: varying population mortality risk, by doubling the average annual reduction in mortality risk from the UN WPP (1.0% per year) to a 2.0% reduction per year; and by removing the temporal trend and maintaining 2017 mortality risk across the model time horizon. Similarly, trends were varied in GDP per worker, by doubling the annual GDP growth rate from 3.2% per year (OECD forecast average annual GDP growth rate) to 6.4% per year; and by removing the temporal trend and maintaining 2017 GDP per worker estimates across the model time horizon [31]. To assess the impact of the assumption of the WHO standard annual discount rate of 3.0%, scenario analyses were performed in which the discount rate applied was 5.0% or 1.5% [22].

Results

The prevalence of diabetes in the Chinese working age population was 7.1% (9.6% in men and 4.7% in women), equating to 56.4 million people (41.4 million men and 15.1 million women) between 20 years and retirement age living with diabetes (Table 1).
Table 1

 The age- and sex-specific population and number of people living with diabetes in China in 2017

Five-year age group

Men

Women

Population a

Prevalence of diabetes (%) b

Number of men with diabetes

Population a

Prevalence of diabetes (%) b

Number of women with diabetes

20–24

49,362,747

2.1

1,047,558

45,076,346

1.1

495,924

25–29

64,710,828

3.4

2,200,168

62,636,763

1.8

1,155,245

30–34

52,681,251

5.2

2,758,241

51,989,896

3.0

1,550,205

35–39

48,947,934

7.7

3,750,163

47,150,411

4.6

2,179,783

40–44

57,659,007

10.6

6,094,422

55,446,671

6.8

3,793,198

45–49

63,466,389

13.7

8,696,194

61,392,324

9.6

5,912,172

50–54

59,041,717

16.7

9,863,741

   

55–59

36,088,731

19.2

6,939,624

   

Total

431,958,604

9.6

41,350,111

323,692,411

4.7

15,086,527

aAge- and sex-specific population estimates were based on the 2017 China Statistical Yearbook [22]

bAge- and sex-specific prevalence of diabetes based on estimates by age and sex from the 2017 International Diabetes Federation Diabetes Atlas (8th Edition) [1]. Number of men and women with diabetes calculated based on prevalence of diabetes but, due to rounding of data presented in this table, values may not precisely match

Excess mortality and years of life lost to diabetes

Until each cohort reached retirement age, there were an estimated 4.1 million more deaths in those with diabetes than in the same cohort assuming no diabetes (Table 2). We estimated that years of life lived by the current cohort of people living with diabetes in China would be reduced by an estimated 22.7 million years (3.7%) over their working lifetime, compared with the same cohort assuming no diabetes (Table 2). This equated to an average of 0.2 years of life lost per person with diabetes (0.5 in men and 0.1 in women) over the working lifetime.
Table 2

 Excess deaths and years of life lived in those with diabetes, and in the same cohort assuming no diabetes, over the working lifetime of the Chinese population simulated from life table modelling

Five-year age group

Deaths in cohort with diabetes

Deaths in ‘diabetes cohort’ assuming no diabetes

Excess deaths in diabetes cohort

Years of life lived in cohort with diabetes

Years of life lived in ‘diabetes cohort’ assuming no diabetes

Years of life lost (%)

Men

  20–24

233,673

93,055

140,618

23,068,663

24,137,075

1,068,412 (4.4)

  25–29

491,025

199,620

291,405

44,595,058

46,819,697

2,224,639 (4.8)

  30–34

606,503

252,329

354,174

50,479,125

53,125,434

2,646,309 (5.0)

  35–39

794,274

338,998

455,276

60,347,927

63,562,440

3,214,513 (5.1)

  40–44

1,199,885

526,579

673,306

82,915,298

87,189,008

4,273,710 (4.9)

  45–49

1,497,072

676,916

820,156

93,781,293

98,110,211

4,328,918 (4.4)

  50–54

1,366,637

639,541

727,096

74,209,042

76,935,304

2,726,262 (3.5)

  55–59

501,187

241,477

259,710

25,645,750

26,133,906

488,156 (1.9)

  Total

6,690,256

2,968,515

3,721,741

455,042,156

476,013,075

20,970,919 (4.4)

Women

  20–24

32,738

10,329

22,409

9,601,854

9,748,303

146,449 (1.5)

  25–29

73,911

23,937

49,974

19,713,555

20,027,822

314,267 (1.6)

  30–34

92,616

30,925

61,691

22,362,213

22,717,911

355,698 (1.6)

  35–39

114,090

39,455

74,635

24,822,935

25,188,598

365,663 (1.5)

  40–44

153,089

55,088

98,001

29,880,056

30,234,422

354,366 (1.2)

  45–49

128,472

48,322

80,150

22,401,833

22,549,120

147,287 (0.7)

  Total

594,916

208,056

386,860

128,782,446

130,466,176

1,683,730 (1.3)

Total

7,285,172

3,176,571

4,108,601

583,824,602

606,479,251

22,654,649 (3.7)

Calculation of years of life lived were modelled in life tables with a half cycle correction and were subject to an annual discount rate of 3%

Productivity-adjusted life years lost to diabetes

Diabetes was estimated to reduce PALYs lived by the current cohort of people living with diabetes in China by 75.8 million PALYs (56.3 million in men and 19.5 million in women) over the working lifetime or by 15.1% (14.1% in men and 18.6% in women) (Table 3). This equated to 1.3 PALYs lost per person with diabetes (1.4 in men and 1.3 in women). Assuming a constant GDP per full-time worker of ¥179,486 (US$26,789), productivity lost to diabetes in China would be associated with a ¥17.4 trillion (US$2.6 trillion) loss in GDP. This is equivalent to an average GDP loss of ¥307,925 (US$45,959) per person with diabetes over the working lifetime.
Table 3

PALYs lived in those with diabetes, and in the same cohort assuming no diabetes, over the working lifetime of the Chinese population simulated from life table modelling

Five-year age group

PALYs lived in cohort with diabetes

PALYs lived in ‘diabetes cohort’ assuming no diabetes

PALYs lost (%)

PALYs lost per person with diabetes

Men

  20–24

18,174,785

20,808,177

2,633,392 (12.7)

2.5

  25–29

36,552,611

41,998,670

5,446,059 (13.0)

2.5

  30–34

41,785,698

48,233,987

6,448,289 (13.4)

2.3

  35–39

49,109,290

56,983,885

7,874,595 (13.8)

2.1

  40–44

64,612,176

75,376,378

10,764,202 (14.3)

1.8

  45–49

68,090,522

79,840,754

11,750,232 (14.7)

1.4

  50–54

48,713,250

57,417,480

8,704,230 (15.2)

0.9

  55–59

14,684,793

17,354,773

2,669,980 (15.4)

0.4

  Total

341,723,125

398,014,104

56,290,979 (14.1)

1.4

Women

  20–24

6,471,378

7,757,910

1,286,532 (16.6)

2.6

  25–29

13,618,347

16,457,544

2,839,197 (17.3)

2.5

  30–34

15,467,109

18,851,847

3,384,738 (18.0)

2.2

  35–39

16,826,762

20,694,590

3,867,828 (18.7)

1.8

  40–44

19,447,052

24,143,268

4,696,216 (19.5)

1.2

  45–49

13,710,929

17,187,259

3,476,330 (20.2)

0.6

  Total

85,541,577

105,092,418

19,550,841 (18.6)

1.3

Total

427,264,702

503,106,522

75,841,820 (15.1)

1.3

Calculation of PALYs were modelled in life tables and subject to an annual discount rate of 3%

Sensitivity and scenario analyses

Figure 1 shows the contribution of the four causes of diabetes-related productivity loss considered in our models. Labour force dropout (62.1%) and mortality (24.7%) were the major contributors to productivity loss, followed by absenteeism (9.0%) and presenteeism (4.2%). Accordingly, the majority of costs associated with productivity losses were caused by diabetes-related labour force dropout (¥10.8 trillion, US$1.6 trillion) and mortality (¥4.2 trillion, US$640.8 billion), followed by absenteeism (¥1.6 trillion, US$232.9 billion) and presenteeism (¥728.4 billion, US$108.7 billion). The proportion of PALYs lost to diabetes-related mortality was higher in men (24.7%) than women (7.1%), while the proportion of PALYs lost to absenteeism was higher in women (21.1%) than in men (9.0%).
Fig. 1

 Economic burden of productivity loss in those with diabetes owing to diabetes-related premature mortality, labour force dropout, absenteeism and presenteeism over the working lifetime in the Chinese population

The model was sensitive to a number of inputs such as productivity indices, diabetes-related labour force dropout, and mortality risk, and model assumptions, including temporal trends in mortality risk and the annual discount rate (Table 4). Compared with the base case, at upper and lower uncertainty bounds of absenteeism and presenteeism estimates, PALYs lost to diabetes were reduced and increased by 3.1%, respectively; and by 15.6% at the upper and lower bounds of estimates of diabetes-related labour force dropout, respectively. Applying the upper and lower bounds of 95% CI around estimates of all-cause mortality risk associated with diabetes, PALYs lost were increased by 2.7% and decreased by 3.2%, respectively. In scenario analyses, doubling the annual reduction in population mortality risk to 2% reduced PALYs lost by 1.0%, while removing all temporal trends in population mortality risk increased PALYs lost by 1.2%. Doubling the annual GDP growth rate to 6.4% led to an increase in the estimate of GDP lost to ¥21.1 trillion (US$3.2 trillion), while removing all temporal trends in GDP decreased the estimate of GDP lost to ¥13.6 trillion (US$2.0 trillion). Finally, increasing the annual discount rate to 5% corresponded to a 14.5% reduction in PALYs lost, and a reduction in annual discount rate to 1.5% led to a 14.2% increase in PALYs lost (Table 4).
Table 4

Sensitivity and scenario analyses to assess the impact of the uncertainties around productivity, mortality and economic data inputs on PALYs lost in those with diabetes in the Chinese population and the associated economic impact

Analysis

PALYs lost owing to diabetes

% change in PALYS lost compared with base case

GDP lost (US$

trillion)

GDP lost per person with diabetes (US$)

Base case

75,841,820

 

2.6

45,959

1. Productivity indices upper uncertainty bound a

78,197,619

+3.1

2.7

47,322

2. Productivity indices lower uncertainty bound a

73,486,021

−3.1

2.5

44,596

3. Labour force dropout upper uncertainty bound b

87,686,339

+15.6

3.0

52,994

4. Labour force dropout lower uncertainty bound b

63,997,300

−15.6

2.2

38,974

5. Upper uncertainty bound of all-cause mortality risk associated with diabetes c

77,894,655

+2.7

2.7

47,320

6. Lower uncertainty bound of all-cause mortality risk associated with diabetes c

73,396,102

−3.2

2.5

44,338

7. Temporal trend in population mortality risk is doubled to a 2% reduction per year d

75,057,463

−1.0

2.6

45,375

8. No temporal trend in population mortality risk d

76,718,554

+1.2

2.6

46,615

9. Annual GDP growth rate is doubled to 6.4% per year e

  

3.2

55,918

10. No temporal trend in GDP e

  

2.0

36,000

11. Annual discount rate increased to 5% f

64,838,665

−14.5

2.2

38,343

12. Annual discount rate reduced to 1.5% f

86,619,273

+14.2

3.0

53,597

aSensitivity analysis 1 and 2 apply (1) a 25% increase and (2) a 25% reduction in absenteeism and presenteeism estimates, holding all other model inputs constant

bSensitivity analysis 3 and 4 apply (3) a 25% increase and (4) a 25% reduction in diabetes-related labour force dropout estimates, holding all other model inputs constant

cSensitivity analysis 5 and 6 apply (5) the upper bound of the 95% CI and (6) the lower bound of the 95% CI around the estimate of RR of all-cause mortality associated with diabetes, holding all other model inputs constant

dScenario analysis 7 and 8 apply (7) double the annual reduction in mortality risk to 2% per year and (8) no temporal trend in population mortality risk, holding all other model inputs constant

eSensitivity analysis 9 and 10 apply (9) double the annual growth rate in GDP to 6.4% per year and (10) no temporal trend in GDP across the model, holding all other model inputs constant. These sensitivity analyses do not affect the number of PALYs lived but do affect their assumed value and therefore the resulting GDP lost

fSensitivity analysis 11 and 12 apply an annual discount rate (11) increased to 5% (in line with the WHO standard annual rate) and (12) reduced to 1.5%

Discussion

Our study highlights the considerable impact of diabetes on the years of life lived and productivity in China. Among the working age Chinese population with diabetes followed to retirement age, diabetes was predicted to cause 4.1 million excess deaths, 22.7 million years of life lost and a 15.1% loss of PALYs, associated with a significant economic impact over the working lifetime.

Productivity losses accumulated from a combination of premature mortality and diabetes-related labour force dropout, absenteeism and presenteeism while at work. Over the working lifetime of the diabetes cohort, higher all-cause mortality risk in those with diabetes resulted in a 3.7% reduction in years of life lived; this was higher in men (4.4%) than women (1.3%). This is consistent with previous studies showing higher mortality risk in working age Chinese men than women [32]. Despite a lower prevalence of diabetes among younger age groups, the relative impact of diabetes on years of life lost was greater among younger people. This is consistent with the strong association between duration of diabetes and mortality risk, and evidence for high risk of diabetes complications and mortality among younger age groups in East Asian populations [8, 33, 34].

Our findings estimate that diabetes will cause a 15.1% reduction in the total number of PALYs lived by the current Chinese population with diabetes, or 1.3 PALYs per person, over a working lifetime. This was similar to the average number of PALYs lost per person with diabetes in a recent study in Australia [16]. However, retirement age in China is 10 years and 20 years lower in men and women, respectively, than in Australia, and therefore people living with diabetes in China incurred similar productivity losses over a shorter timeframe, and in younger age groups. This is likely to be the result of higher mortality risk and labour force dropout in the Chinese population with diabetes compared with the Australian population with diabetes [16]. We further estimated that the lost productivity incurred a loss of ¥17.4 trillion (US$2.6 trillion) in GDP, demonstrating the significant economic impact of diabetes-related productivity losses. Furthermore, as our model did not take into account the considerable direct costs of diabetes (including diagnosis, treatment and care), this is likely to be a highly conservative estimate of economic impact. This is supported by recent ADA research which found that productivity losses only accounted for 27.5% of the total economic costs of diabetes in the USA in 2017, and a global study which reported that 34.7% of the total economic burden was due to productivity losses [13, 27]. These findings suggest that the wider economic burden of diabetes in China could be three to four times greater than our estimates.

The absolute number of PALYs lost over the working lifetime in our model was greater in men. This is because the prevalence of diabetes is higher in men than women and increased time at risk of diabetes-related productivity losses owing to the higher retirement age in men than women in China [20]. However, the relative reduction in productivity was greater in women with diabetes (18.6%) than men with diabetes (14.1%), and in all age groups driven by greater labour force dropout and absenteeism in women with diabetes compared with men. This is reflected in the high proportion of productivity losses due to labour force dropout and absenteeism in women (66.6% and 21.1%, respectively) compared with men (60.7% and 5.0% respectively). There is a wealth of evidence for employment shortfall in people with diabetes compared with those without diabetes [13, 35, 36]. For example, diagnosis of diabetes in the USA was associated with approximately double the labour force participation shortfall and more days of work lost in women with diabetes compared with men with diabetes [37, 38].

We estimated an average GDP loss of ¥307,925 (US$45,959) per person with diabetes over the working lifetime. In theory, this amount could be spent per person in the current diabetes cohort in China to prevent diabetes as a break-even investment. However, this assumes 100% effectiveness of prevention. If an intervention was able to prevent 10% of diabetes, the break-even investment amount would be ¥30,793 (US$4596) per working age person. These figures are based on saved productivity alone and therefore likely to be a conservative estimate with savings from reduced direct costs of diabetes adding considerable economic benefit [15]. Future studies on diabetes prevention in Chinese populations that incorporate both direct and indirect costs of diabetes would more accurately characterise the potential cost benefit of these interventions.

Our study adds information to previous estimates by the IDF and others of the economic burden of diabetes, by quantifying these in terms of missed production opportunities, rather than health expenditure alone [1]. Another strength of our study was the use of contemporary sex-specific and age-group-specific estimates of diabetes prevalence, mortality risk, labour force participation and in-work productivity. Life table modelling allowed us to capture the impact of diabetes-related productivity losses across the working lifetime. We found that the majority of diabetes-related productivity losses were due to labour force dropout (62.1%) and premature mortality (24.7%) in those with diabetes. This suggests that increased labour force retention and improved diabetes treatment leading to reduced premature mortality in those with diabetes could reduce diabetes-related productivity losses in China. This highlights the trade-off between direct and indirect costs, where increased spending on direct costs may reduce indirect costs through productivity gains. In our analyses, we used PALYs to calculate productivity losses, and an advantage of this approach is that PALYs can be ascribed a financial value (GDP in our study) and net costs calculated.

Our study had several limitations that warrant mention. Data on the productivity effects of diabetes in Chinese populations were not available, and hence estimates from a multi-country meta-analysis of the effects of diabetes on absenteeism, presenteeism and labour force dropout were used instead, which may not have been generalisable to Chinese populations. Uncertainty around productivity indices was explored in sensitivity analyses; varying absenteeism and presenteeism by 25% had a small effect on estimates of PALYs lost (±3.1%), whereas the model was more sensitive to equivalent variation in labour force dropout (±15.6%). Our findings were based on modelled estimates from life tables which simulated the progress of the current cohort of people living with diabetes in China through to retirement, but we did not account for incident diabetes arising in the cohort, and hence would have underestimated the potential return on investment from prevention [15]. We also assumed that current projections in temporal trends in mortality rates and GDP growth held true across the model time horizon. However, in scenario analyses, the doubling and removal of the trend in population mortality rate affected the model output by <2%, although estimates of GDP lost were more sensitive to the equivalent changes in GDP growth rate. Other limitations of this study were that: (1) the contribution of comorbidities of diabetes (particularly obesity, and cardiovascular disease risk factors) to productivity loss could not be distinguished; (2) in the absence of available data, the assumption was made that those working were in full-time employment, and the impact of diabetes upon unpaid work was not included; and (3) diabetes might impact on GDP in ways other than through productivity losses [12, 14]. While these limitations may affect the accuracy of the estimate produced by our model, the overall conclusion of our study is unlikely to have changed.

Our findings highlight the significant productivity losses owing to diabetes in the current cohort of people living with diabetes in China. Given the considerable economic impact of these productivity losses, prevention of diabetes and of the complications of diabetes through adequate management of glucose levels should be considered an investment with potentially large economic benefits in the longer term. To inform relevant interventions and their potential social and economic returns on investment, further research is needed to describe the dynamic trade-off between the costs of prevention and treatment strategies and their net economic consequences, taking into account future productivity gains.

Notes

Contribution statement

TRH, DL and DJM conceived and designed the study and analyses. TRH, EZ, AJO, LC and ZA made substantial contributions to analysis and interpretation of data. All authors made substantial contributions to drafting the article and approved the final version. TRH and DL are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Funding

DJM is supported by a National Health and Medical Research Council (NHMRC) Senior Research fellowship. DL has received honoraria or study grants from AbbVie, Astellas, AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, Novartis, Pfizer, Sanofi and Shire. EZ has received study grants from AstraZeneca, Pfizer and Shire. ZA has received research funding support from NHMRC, Swiss Medical Board, and Swiss Network for Health Technology Assessment, Commission for Technology and Innovation Switzerland, Novartis, Pfizer, AstraZeneca and Vifor. The above research funding was not utilised in the design of the study; the collection, analysis and interpretation of data; writing the report; or the decision to submit the report for publication.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Supplementary material

125_2019_4875_MOESM1_ESM.pdf (41 kb)
ESM (PDF 40.8 kb)

References

  1. 1.
    International Diabetes Federation (2017) IDF Diabetes Atles (8th). International Diabetes Federation, Brussels, BelgiumGoogle Scholar
  2. 2.
    National Diabetes Research Group (1981) A mass survey of diabetes mellitus in a population of 300,000 in 14 provinces and municipalities in China. Zhonghua nei ke za zhi 20:678–683Google Scholar
  3. 3.
    Pan XR, Yang WY, Li GW, Liu J, National Diabetes Prevention and Control Cooperative Group (1997) Prevalence of diabetes and its risk factors in China, 1994. Diabetes Care 20(11):1664–1669.  https://doi.org/10.2337/diacare.20.11.1664 CrossRefPubMedGoogle Scholar
  4. 4.
    Gu D, Reynolds K, Duan X et al (2003) Prevalence of diabetes and impaired fasting glucose in the Chinese adult population: International Collaborative Study of Cardiovascular Disease in Asia (InterASIA). Diabetologia 46(9):1190–1198.  https://doi.org/10.1007/s00125-003-1167-8 CrossRefPubMedGoogle Scholar
  5. 5.
    Wang L, Gao P, Zhang M et al (2017) Prevalence and ethnic pattern of diabetes and prediabetes in China in 2013. JAMA 317(24):2515–2523.  https://doi.org/10.1001/jama.2017.7596 CrossRefPubMedPubMedCentralGoogle Scholar
  6. 6.
    Chan JC, Malik V, Jia W et al (2009) Diabetes in Asia: epidemiology, risk factors, and pathophysiology. JAMA 301(20):2129–2140.  https://doi.org/10.1001/jama.2009.726 CrossRefPubMedGoogle Scholar
  7. 7.
    Ramachandran A, Ma RC, Snehalatha C (2010) Diabetes in Asia. Lancet 375(9712):408–418.  https://doi.org/10.1016/S0140-6736(09)60937-5 CrossRefPubMedGoogle Scholar
  8. 8.
    Reynolds K, Saydah SH, Isom S et al (2018) Mortality in youth-onset type 1 and type 2 diabetes: The SEARCH for Diabetes in Youth study. J Diabetes Complicat 32(6):545–549.  https://doi.org/10.1016/j.jdiacomp.2018.03.015 CrossRefPubMedPubMedCentralGoogle Scholar
  9. 9.
    Ma RCW (2018) Epidemiology of diabetes and diabetic complications in China. Diabetologia 61(6):1249–1260.  https://doi.org/10.1007/s00125-018-4557-7 CrossRefPubMedGoogle Scholar
  10. 10.
    Huo X, Gao L, Guo L et al (2016) Risk of non-fatal cardiovascular diseases in early-onset versus late-onset type 2 diabetes in China: a cross-sectional study. Lancet Diabetes Endocrinol 4(2):115–124.  https://doi.org/10.1016/S2213-8587(15)00508-2 CrossRefPubMedGoogle Scholar
  11. 11.
    Sculpher M (2001) The role and estimation of productivity costs in economic evaluation. In: Drummond MF, McGuire A (eds) Economic Evaluation in Health Care: Merging Theory with Practice. Oxford University Press, Oxford, pp 94–112Google Scholar
  12. 12.
    Krol M, Brouwer W (2014) How to estimate productivity costs in economic evaluations. Pharmacoeconomics 32(4):335–344.  https://doi.org/10.1007/s40273-014-0132-3 CrossRefPubMedGoogle Scholar
  13. 13.
    American Diabetes Association (2018) Economic costs of diabetes in the U.S. in 2017. Diabetes Care 41(5):917–928.  https://doi.org/10.2337/dci18-0007 CrossRefPubMedCentralGoogle Scholar
  14. 14.
    Rasmussen B, Sweeny K, Sheehan P (2016) The Impact of Wellness on Workforce Productivity in Global Markets. A Report to the U.S. Chamber of Commerce’s Global Initiative on Health and Economy. In. Victoria Institute of Strategic Economic Studies, Melbourne, AustraliaGoogle Scholar
  15. 15.
    Bertram MY, Sweeny K, Lauer JA et al (2018) Investing in non-communicable diseases: an estimation of the return on investment for prevention and treatment services. Lancet 391(10134):2071–2078.  https://doi.org/10.1016/S0140-6736(18)30665-2 CrossRefPubMedGoogle Scholar
  16. 16.
    Magliano DJ, Martin VJ, Owen AJ, Zomer E, Liew D (2018) The productivity burden of diabetes at a population level. Diabetes Care 41(5):979–984.  https://doi.org/10.2337/dc17-2138 CrossRefPubMedGoogle Scholar
  17. 17.
    Owen AJ, Maulida SB, Zomer E, Liew D (2018) Productivity burden of smoking in Australia: a life table modelling study. Tob Control:tobaccocontrol-2018-054263.  https://doi.org/10.1136/tobaccocontrol-2018-054263 CrossRefGoogle Scholar
  18. 18.
    Briggs AD, Wolstenholme J, Blakely T, Scarborough P (2016) Choosing an epidemiological model structure for the economic evaluation of non-communicable disease public health interventions. Popul Health Metrics 14(1):17.  https://doi.org/10.1186/s12963-016-0085-1 CrossRefGoogle Scholar
  19. 19.
    Organisation for Economic Co-operation and Development (OECD) (2017) Pension At A Glance 2017: OECD and G20 Indicators. OECD Publishing, ParisGoogle Scholar
  20. 20.
    Feng Q, Yeung W-JJ, Wang Z, Zeng Y (2019) Age of retirement and human capital in an aging China, 2015–2050. Eur J Popul 35(1):29–62CrossRefGoogle Scholar
  21. 21.
    World Health Organization, Baltussen RMPM, Adam T, Tan-Torres Edejer T et al (2003) Making choices in health: WHO guide to cost-effectiveness analysis. World Health Organization http://www.who.int/iris/handle/10665/42699
  22. 22.
    National Bureau of Statistics of The People’s Republic of China (2017) China Statistical Yearbook 2017. Beijing, ChinaGoogle Scholar
  23. 23.
    National Bureau of Statistics of the People’s Republic of China (2010) China Statistical Yearbook 2010. Beijing, ChinaGoogle Scholar
  24. 24.
    United Nations Department of Economic and Social Affairs Population Division (2017) World Population Prospects: The 2017 Revision. United Nations, New YorkGoogle Scholar
  25. 25.
    Bragg F, Holmes MV, Iona A et al (2017) Association between diabetes and cause-specific mortality in rural and urban areas of China. JAMA 317(3):280–289.  https://doi.org/10.1001/jama.2016.19720 CrossRefPubMedPubMedCentralGoogle Scholar
  26. 26.
    International Labour Office (2017) ILO labour force estimates and projections. International Labour Office, Geneve, Switzerland, pp 1990–2030Google Scholar
  27. 27.
    Bommer C, Heesemann E, Sagalova V et al (2017) The global economic burden of diabetes in adults aged 20–79 years: a cost-of-illness study. Lancet Diabetes Endocrinol 5(6):423–430.  https://doi.org/10.1016/S2213-8587(17)30097-9 CrossRefPubMedGoogle Scholar
  28. 28.
    Torrance GW, Feeny D (1989) Utilities and quality-adjusted life years. Int J Technol Assess Health Care 5(04):559–575.  https://doi.org/10.1017/S0266462300008461 CrossRefPubMedGoogle Scholar
  29. 29.
    Organisation for Economic Co-operation and Development (OECD) (2018) OECD Compendium of Productivity Indicators 2018. OECD Publishing, ParisGoogle Scholar
  30. 30.
    Organisation for Economic Co-operation and Development (OECD) (2018) OECD Economic Outlook: Statistics and Projections: Long-term baseline projections, Real GDP long-term forecast (indicator). OECD publishing, ParisGoogle Scholar
  31. 31.
    The World Bank (2018) World Development indicators, 2017. The World Bank, Washington DCGoogle Scholar
  32. 32.
    He J, Gu D, Wu X et al (2005) Major causes of death among men and women in China. N Engl J Med 353(11):1124–1134.  https://doi.org/10.1056/NEJMsa050467 CrossRefPubMedGoogle Scholar
  33. 33.
    Ma RC, Chan JC (2013) Type 2 diabetes in East Asians: similarities and differences with populations in Europe and the United States. Ann N Y Acad Sci 1281(1):64–91.  https://doi.org/10.1111/nyas.12098 CrossRefPubMedPubMedCentralGoogle Scholar
  34. 34.
    Al-Saeed AH, Constantino MI, Molyneaux L et al (2016) An inverse relationship between age of type 2 diabetes onset and complication risk and mortality: the impact of youth-onset type 2 diabetes. Diabetes Care 39(5):823–829.  https://doi.org/10.2337/dc15-0991 CrossRefPubMedGoogle Scholar
  35. 35.
    Kahn ME (1998) Health and labor market performance: the case of diabetes. J Labor Econ 16(4):878–899.  https://doi.org/10.1086/209909 CrossRefGoogle Scholar
  36. 36.
    Seuring T, Archangelidi O, Suhrcke M (2015) The economic costs of type 2 diabetes: a global systematic review. PharmacoEconomics 33(8):811–831.  https://doi.org/10.1007/s40273-015-0268-9 CrossRefPubMedPubMedCentralGoogle Scholar
  37. 37.
    Minor T, MacEwan JP (2016) A comparison of diagnosed and undiagnosed diabetes patients and labor supply. Econ Hum Biol 20:14–25.  https://doi.org/10.1016/j.ehb.2015.10.003 CrossRefPubMedGoogle Scholar
  38. 38.
    Tunceli K, Bradley CJ, Nerenz D, Williams LK, Pladevall M, Elston Lafata J (2005) The impact of diabetes on employment and work productivity. Diabetes Care 28(11):2662–2667.  https://doi.org/10.2337/diacare.28.11.2662 CrossRefPubMedGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Authors and Affiliations

  1. 1.School of Public Health and Preventive MedicineMonash UniversityMelbourneAustralia
  2. 2.Department of Clinical Diabetes and EpidemiologyBaker Heart and Diabetes InstituteMelbourneAustralia

Personalised recommendations