Monitoring of sickness absences is an essential part of occupational health care. Even though sickness absence is not a simple function of ill health, and includes psychosocial factors and coping behaviors, it still remains a valuable tool to assess the impact of disease or other factors on a worker’s capacity and health status [1,2]. The relationship between occupational health risks and sickness absence has been well documented [3,4,5], supporting the need to monitor and evaluate sick-leave spells for both prognosis and early intervention. Absenteeism as a phenomenon has been the focus of research for several decades [6], with an increasing body of literature over the years [7], establishing the need for monitoring which is performed mainly by national statistical services and social security organizations across countries [8,9,10,11].

The definition of absenteeism has been an issue of debate for several conferences of the International Labor Organization (ILO) [12], troubling researchers in the field to the point that today this makes cross-country comparisons difficult. ILO defines absence due to illness as the absence from work, usually of short duration, that may be work related or not, which is attributed by the employee to illness or injury and is accepted by the employer as such [13]. The European Council and most organizations calculate on a yearly basis the compensated working days per employee (days for which employees receive salary compensation) lost due to sickness or injury [9,10,11,12]. Other reasons such as vacations, maternity and paternity leave, educational leave or strikes can also contribute to absenteeism, but are usually not monitored [10,12,14].

Several indicators have been used to monitor absenteeism [15], like: (i) Absence from work; (ii) Frequency of occurrence of absence; (iii) Cumulative incidence (individual frequency); and (iv) Absenteeism rate (%). The International Social Security Association (ISSA) in order to facilitate international comparisons, proposed in 1981 to use absenteeism rate, which express the percentage of working time lost, as a monitor tool since it takes into account the differences in working time per country or occupation [11,16].

International comparative studies are few, with several methodological problems mainly because of lack of widely accepted standard measures of absenteeism, and trends in absenteeism have not been explored. The CESifo DICE Report [17] is such an effort, but the data extracted by OECD (Organization for Economic Co-Operation and Development) Health [9] and WHO Health for all databases [10] were incomplete, allowing several arguments. In the CESifo DICEreport [17] Greece was ranked in the most favorable position for the period 1996–2005, with an average of 4.9 sick leave days per employee per year, while Germany, UK and Netherlands had shown 16.7, 7.6 and 5.1, respectively. In the Fourth European Working Conditions Survey [18], Greek employees reported both the most unfavorable working conditions but the lowest sickness absence (<4 days per year).

The IMF staff paper on work absence in Europe [19] for the period 1983–2003, has shown that the percentage of employees who were absent daily was <0.5% in Greece, showing a declining trend, whereas in most European countries the percentage exceeded 2.5%. Similarly in the European Labour Force Survey the absenteeism rate for the period 2004–2006 for Greece does not exceed 0.2%, while for most countries it exceeded 1.5% [20]. These studies used available data from Eurostat, OECD Health database and WHO Health for all databases and despite some inconsistencies, Greek employees seem to be less frequently absent compared to the most industrialized countries worldwide [7].

Many of the differences on sickness absences across countries may be attributed to the respective sickness insurance policies. In the Greek private sector, a worker on sick leave due to illness will be compensated by the company with 50% of his wage; occupational injuries (accidents) are fully compensated. In fact this lasts till the third day of a sick-leave spell, and afterwards the employee will be paid (50%) by the social insurance coverage scheme. Employees can get sickness benefits for up to one year and in some cases two years, before getting a disability pension, or even be fired if the employer or the insurance company argues that the employee is not willing to follow instructions to improve his health. In general an employee is entitled of sickness benefits for as many months as the years of employment completed. Sick leaves up to 2 days are commonly accepted by employee’s report of sickness (not exceeding 3–4 different spells per year). For sick-leave spells longer than 3 days a physician’s certificate is required and if the sick-leave spell is longer than 10 days it has to be validated by the insurance medical committee.

In general, there are relatively few comparative studies on sickness absence, with several limitations (variations in the average annual working days; incompleteness in the monitoring networks and in the methodologies used) and it is considered that sickness absence has not been the subject of a systematic investigation in Greece so far [21]. Based on these, we undertook this research in order to estimate the levels of sickness absence in the private sector in Greece, using shipyard and national insurance data.

Table 1 presents various indicators of the insured employees at the Social Insurance Institute (IKA) for the period 1987 to 2006. Total compensated sick-leave days reached 7.65 million at 2004. A steady increase in the insured population by 1% per year was monitored since 1987. In the 20-years national data, the results also showed a 7-year wave in sickness absence indexes (a decrease during the period 1991–1997 and an increase in 1998–2004) combined with a small yet significant decline as a general trend. The annual sick-leave rate during the last decade hardly exceeded 2.5% and the mean sick leave duration per employee was well below 6 days (Table 1).

Table 2 and Table 3 present the characteristics and main results from the shipyard employees (92% males; mean age 38 ± 10 years) for the period 1999–2006. Approximately 60% of the employees who were absent at least once appear to have short term sick leaves. Out of total spells every year 75% were short term, representing approximately 25% of the total sick-leave days. The percentage of total employees with at least one sick-leave spell per year increased from 36% in 1999 to 52% in 2006 (Table 2).

During the 8-year study period the duration of sickness absence per employee ranged between 4.6 and 8.7 days/year, with an increasing tendency, while the mean duration of each sick-leave spell ranged between 5.9 and 8.6 days, demonstrating a declining trend. Around 1% of total employees were absent due to illness every day (Table 3). Frequency of sickness absence occurrence along with its cumulative incidence were steadily rising, indicating that over the years more employees tend to be more frequently absent due to illness. Sick leave rate ranged between 1.42 and 2.68% with an increasing trend after 2001.

Shipyards data have shown that an average of 6.2 days lost per employee due to sick-leave compared to the estimated 5.9 days of IKA employees. Our data confirm that Greece has one of the lowest sickness absence rates in the European Union [18]. However the level of sickness absence seems to be higher than the estimated rates in previous reports [17,19,20,22] and this may attributed in the more frequent rather than longer sick-leave spells.

Several reasons could explain the low sickness absence rate in Greece, as well as the diversity that exists across countries. Osterkamp and Röhn have reported several of these reasons in 2007; many applied in Greek case [7]. Firstly, the low compensation rate (50%) in Greece, which has been proven to play a key role [23,24]. Countries like Sweden where social insurance schemes provided generous sickness absence compensation have been shown to have high levels of sickness absence [24,25]. A second reason is the difference in unemployment rates. The unemployment rate has been rather high (about 10% of the population) in Greece during the study period compared to most countries in the EU according to Eurostat [26,27].

The employment rate in Greece during the study period according to the Fourth European Survey was 59.6%, one of the lowest in the European Union. Retirement pensions in the private sector typically start at 60 after 35 years of employment but early, nondisabled retirees included men at age 58 years who completed 35 years of work and those who work in hazardous occupations (“heavy and unhealthy jobs”) at ages 55 or 53 years for full or partial retirement, respectively [28]. In theory higher unemployment rates mean greater difficulties in finding a new job in case of dismissal, and therefore greater effort by the employee to stay at work even with impaired health (i.e., “presenteeism”) but the real effects of these parameters (unemployment or employment rate and retirement schemes) on sick leave duration are not well documented [7].

Another parameter could be the differences in the educational level among employees. According to the 4^th European Working Conditions Survey [18] around 35% of employees in Greece are of primary or lower secondary education, one of the highest proportion in the EU, and less than 20% have received any special training. In addition, Greek employees reported one of the lowest levels of autonomy at work [18,29] which is possibly related to supervision and motivation. It is important to investigate the role of the psychosocial working environment on sickness absence. Paradoxically, work load, which is considered to affect sick leave rate, is reported as heavier in Greece and it is related to a lengthy working week (over 40 h) [18,19], with 40% of employees (including self-employed) working even more.

Furthermore, diversity in health-ill patterns among populations might partly explain differences in sickness absence rates. However, it is not very plausible to explain more than a small fraction since morbidity patterns did not vary so much at least in industrialized countries [7]. Several reports have shown higher morbidity levels among Greek employees or at least similar to the rest European population [25,30] while according to the 4^th European Working Conditions Survey [18] Greek employees are not only exposed to occupational hazards to a greater extent compared to other Europeans, but they report far more frequently that are not very satisfied with their work. These facts are in contrast with the low sick leave rate in Greece, which besides the low compensation rate is hardly excusable and remains an interesting case to be explored, perhaps in the direction of personal, work and socio-cultural related factors. For example, the increased family support in Greek society [31] enables employees to keep working, even under moderate health conditions [29,31] or the management toleration in some unofficial (and not recorded) sick leaves.

We have to acknowledge that the results of this study cannot be applied to the public sector which it has structural differences and data on sickness absence are not available. Even though shipyard data cannot be considered as representative of private employment in Greece, the comparable indexes were very much alike with the national data from the major insurance scheme (IKA). Although higher than previously recorded the fact remains; Greece ranks last or first in the list with six or less days lost annually per employee in private sector.

In this study, we have described and compared time series of data on sickness absence originated by (i) shipyard employees; and (ii) the largest social insurance fund in Greece, which covers the majority (70–90%) of employees in the private sector.

The low level of sickness absences in private sector in Greece was confirmed in our study, although it was found higher than the suggested in previous international studies. Sick leave rate was 2.5% and the total duration per employee was 5.8 days/year. Interestingly, short spells (<4 days) of sickness absences accounted at least for 25% of the total duration which currently is not recorded in national statistics. In the 20-years national data, results have shown a 7-year wave in sickness absence indexes (a decrease the period 1991–1997 and an increase in 1998–2004) combined with a small yet significant decline as a general trend. These observations deserve detailed attention and could only partly be attributed to the compensation and unemployment rates in Greece, while other factors should be explored in field studies whereas sick-leaves should always be monitored and evaluated as an assessment tool of the working environment.